134
Multivalent bioconjugates for the inhibition of anthrax toxin, influenza virus, and HIV by Jacob T. Martin A Thesis Submitted to the Graduate Faculty of Rensselaer Polytechnic Institute in Partial Fulfillment of the Requirements for the degree of DOCTOR OF PHILOSOPHY Major Subject: Chemical and Biological Engineering Approved by the Examining Committee: _________________________________________ Dr. Ravi Kane, Thesis Adviser _________________________________________ Dr. Peter Tessier, Member _________________________________________ Dr. Steven Cramer, Member _________________________________________ Dr. Shekhar Garde, Member _________________________________________ Dr. Marlene Belfort, Member Rensselaer Polytechnic Institute Troy, New York April 2014 (For Graduation May 2014)

Multivalent bioconjugates for the inhibition of anthrax toxin, influenza virus, and HIV

  • Upload
    mit

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Multivalent bioconjugates for the inhibition of anthrax toxin, influenza virus, and HIV

by

Jacob T. Martin

A Thesis Submitted to the Graduate

Faculty of Rensselaer Polytechnic Institute

in Partial Fulfillment of the

Requirements for the degree of

DOCTOR OF PHILOSOPHY

Major Subject: Chemical and Biological Engineering

Approved by the Examining Committee:

_________________________________________ Dr. Ravi Kane, Thesis Adviser

_________________________________________ Dr. Peter Tessier, Member

_________________________________________ Dr. Steven Cramer, Member

_________________________________________ Dr. Shekhar Garde, Member

_________________________________________ Dr. Marlene Belfort, Member

Rensselaer Polytechnic Institute Troy, New York

April 2014 (For Graduation May 2014)

ii

© Copyright 2014

by

Jacob T. Martin

All Rights Reserved

iii

CONTENTS

CONTENTS .......................................................................................................................... iii

LIST OF TABLES ................................................................................................................. viii

LIST OF FIGURES ................................................................................................................. ix

ACKNOWLEDGMENT ....................................................................................................... xiii

ABSTRACT ......................................................................................................................... xv

1. Introduction ................................................................................................................. 1

1.1 Anthrax ............................................................................................................... 1

1.1.1 Overview ................................................................................................ 1

1.1.2 Intoxication pathway ............................................................................. 1

1.2 Influenza ............................................................................................................. 3

1.2.1 Overview ................................................................................................ 3

1.2.2 Virion morphology and infection pathway ............................................ 4

1.2.3 Antigenic variation ................................................................................. 5

1.2.4 Infection pathway .................................................................................. 5

1.3 Human Immunodeficiency Virus (HIV) ............................................................... 6

1.3.1 Overview ................................................................................................ 6

1.3.2 Infection pathway .................................................................................. 6

1.3.3 Current treatment strategies ................................................................. 7

1.4 Multivalent therapeutics .................................................................................... 8

1.4.1 Multivalency ........................................................................................... 8

1.4.2 Scaffolds for multivalent therapeutics ................................................... 9

1.4.3 A multivalent anthrax toxin inhibitor with in vivo efficacy .................. 10

1.4.4 Multivalent anthrax toxin inhibitors with control over average ligand spacing .................................................................................................. 11

1.4.5 Multivalent anthrax toxin inhibitors with predefined ligand spacing . 13

iv

1.5 Motivation ........................................................................................................ 18

2. Preclinical development of polyvalent inhibitors of anthrax toxin ........................... 19

2.1 Therapeutic suitability ..................................................................................... 19

2.1.1 Biocompatibility ................................................................................... 19

2.1.2 PGA inhibitor synthesis ........................................................................ 20

2.1.3 PGA inhibitor efficacy in vitro .............................................................. 20

2.1.4 Storage stability.................................................................................... 22

2.2 Efficacy in vivo .................................................................................................. 22

2.2.1 Toxin challenge .................................................................................... 22

2.2.2 Spore challenge .................................................................................... 23

2.3 Pharmacokinetics ............................................................................................. 25

2.3.1 Serum stability ..................................................................................... 25

2.3.2 PEGylation of PGA-based inhibitors to improve blood residence time26

2.4 Conclusions ...................................................................................................... 28

3. Radial heptavalent inhibitors of anthrax toxin .......................................................... 30

3.1 Abstract ............................................................................................................ 30

3.2 Introduction ..................................................................................................... 30

3.3 Materials and methods .................................................................................... 32

3.3.1 Preparation of toxin components ........................................................ 32

3.3.2 Cytotoxicity assay ................................................................................. 32

3.3.3 Serum stability ..................................................................................... 33

3.3.4 Rat intoxication .................................................................................... 33

3.4 Results and discussion...................................................................................... 33

3.4.1 Structure-based design of heptavalent inhibitors ............................... 33

3.4.2 Characterization of heptavalent inhibitors .......................................... 35

3.5 Conclusions ...................................................................................................... 38

v

4. Multivalent inhibitors of influenza virus.................................................................... 39

4.1 Rationale for the design of multivalent inhibitors of influenza virus entry ..... 39

4.1.1 Targeting a conserved epitope on the hemagglutinin spike ............... 39

4.1.2 De novo design of protein ligands that target a conserved epitope on the hemagglutinin spike ....................................................................... 41

4.1.3 Design of divalent inhibitors that target a conserved epitope on the hemagglutinin spike ............................................................................. 43

4.1.4 Design of polyvalent inhibitors that target a conserved epitope on the hemagglutinin spike ............................................................................. 46

4.2 Production of multivalent bioconjugates of hemagglutinin-binding proteins 46

4.2.1 Modification of hemagglutinin-binding proteins from previous studies .............................................................................................................. 46

4.2.2 Cloning and bacterial production of modified hemagglutinin-binding proteins ................................................................................................ 49

4.2.3 Preparation of hemagglutinin-binding proteins for thioether bioconjugation reactions ..................................................................... 53

4.2.4 Synthesis of divalent hemagglutinin-binding protein bioconjugates .. 54

4.2.5 Synthesis of polyvalent hemagglutinin-binding protein bioconjugates .............................................................................................................. 55

4.2.6 Purification of multivalent hemagglutinin-binding protein bioconjugates ....................................................................................... 59

4.3 Characterization of multivalent bioconjugates of hemagglutinin-binding proteins ............................................................................................................ 61

4.3.1 In vitro microneutralization assay of influenza virus infection inhibition .............................................................................................................. 61

4.3.2 Inhibition of influenza infection with multivalent hemagglutinin-binding protein conjugates ............................................................................... 62

4.4 Conclusions ...................................................................................................... 64

5. Multivalent HIV entry inhibitors ................................................................................ 65

5.1 Rationale for the design of multivalent HIV entry inhibitors ........................... 65

5.1.1 Targeting the cellular receptor CCR5 ................................................... 65

vi

5.1.2 Leukotoxin E is a protein ligand of CCR5 .............................................. 67

5.1.3 Design of multivalent inhibitors that target CCR5 ............................... 68

5.2 Production of polyvalent bioconjugates of leukotoxin E ................................. 69

5.2.1 Cloning and bacterial production of leukotoxin E variants and fusion proteins ................................................................................................ 69

5.2.2 Synthesis of polyvalent leukotoxin E bioconjugates ............................ 73

5.2.3 Purification of polyvalent leukotoxin E bioconjugates ........................ 74

5.3 Characterization of polyvalent conjugates of leukotoxin E ............................. 75

5.3.1 Flow cytometry assay of CCR5-binding inhibition ............................... 75

5.3.2 Inhibition of CCR5-binding with polyvalent bioconjugates of leukotoxin E ............................................................................................................ 76

5.4 Conclusions ...................................................................................................... 78

6. Multivalent oligonucleotide aptamer bioconjugates ................................................ 80

6.1 Rationale for the use of oligonucleotide aptamers as ligands ........................ 80

6.2 Production of polyvalent bioconjugates of the ssDNA aptamer sgc8c ........... 81

6.2.1 Synthesis of polyvalent bioconjugates of the ssDNA aptamer sgc8c .. 81

6.2.2 Purification of polyvalent bioconjugates of the ssDNA aptamer sgc8c84

6.3 Characterization of polyvalent bioconjugates of the ssDNA aptamer sgc8c ... 85

6.4 Conclusions ...................................................................................................... 86

7. Suggestions for future work ...................................................................................... 87

7.1 Identification of oligonucleotide aptamers that target cell receptors ............ 87

7.1.1 Rationale for the search for oligonucleotide aptamers that target cell receptors .............................................................................................. 87

7.1.2 Screening libraries of oligonucleotide aptamers ................................. 87

7.1.3 Characterization of screened oligonucleotide aptamers ..................... 92

7.2 Alternative designs for multivalent HIV entry inhibitors ................................. 92

7.2.1 Peptide ligand derivatives of CCL5 ....................................................... 92

vii

7.2.2 Rationale for the design of divalent CCR5-targeting HIV inhibitors .... 95

7.2.3 Synthesis of homodivalent bioconjugates of aptamers that bind CCR5 .............................................................................................................. 97

7.2.4 Synthesis of heterodivalent inhibitors that bind CCR5 ........................ 98

7.3 Bioengineered protein polymer scaffolds ........................................................ 99

7.3.1 Using proteins as monodisperse polymer scaffolds ............................ 99

7.3.2 Bioengineering scaffolds for precise control over multivalent architecture ........................................................................................ 100

8. Conclusions .............................................................................................................. 102

REFERENCES ................................................................................................................... 103

APPENDIX A – List of abbreviations ............................................................................... 118

APPENDIX B – Additional contributions ........................................................................ 120

viii

LIST OF TABLES

Table 1.1: Polymerization of NAS at various [M]/[CTA] ratios.a ...................................... 14

Table 3.1: Inhibition of anthrax toxin action in a rat intoxication model........................ 37

Table 4.1: Homobifunctional PEG linkers purchased for homodivalent conjugation of

hemagglutinin-binding proteins. ............................................................................. 46

ix

LIST OF FIGURES

Figure 1.1: Representation of the anthrax intoxication pathway. .................................... 2

Figure 1.2: Synthesis of a polyvalent anthrax toxin inhibitor of controlled molecular

weight. ..................................................................................................................... 12

Figure 1.3: Influence of peptide density on the potency of a polyvalent anthrax toxin

inhibitor. .................................................................................................................. 13

Figure 1.4: Design of polyvalent inhibitors with control over molecular weight and ligand

spacing. .................................................................................................................... 15

Figure 1.5: Inhibitory activity of polyvalent inhibitors of anthrax toxin derived from a

controlled molecular weight homopolymer, pNAS. ................................................ 16

Figure 1.6: Characterization of polyvalent inhibitors based on poly(AAm-co-NAS)

copolymers. ............................................................................................................. 17

Figure 2.1: Synthesis of a PGA-based polyvalent anthrax toxin inhibitor. ...................... 20

Figure 2.2: Inhibition of cytotoxicity in vitro by PGA-based polyvalent inhibitors. ......... 21

Figure 2.3: Inhibition of cytotoxicity in vitro by PGA-based polyvalent inhibitors

synthesized in bulk for preclinical studies. .............................................................. 21

Figure 2.4: Inhibitory activity of PGA-based anthrax toxin inhibitors after long-term

storage under various conditions. ........................................................................... 22

Figure 2.5: Inhibitory activity of a PGA-based polyvalent inhibitor against anthrax lethal

toxin in vivo. ............................................................................................................. 23

Figure 2.6: Inhibitory activity of a PGA-based polyvalent inhibitor against anthrax toxin in

vivo generated by infection with B. anthracis spores. ............................................ 24

Figure 2.7: Inhibitory activity of PGA-based inhibitors after incubation in serum. ........ 25

Figure 2.8: ELISA assay of PGA-based anthrax toxin inhibitor blood residence time. .... 26

Figure 2.9: Silver-stain (left) followed by barium iodide-stain (right) of a 4-12% acrylamide

SDS-PAGE of PGA-based anthrax toxin inhibitors conjugated with monofunctional

PEG20k or PEG40k. ...................................................................................................... 28

Figure 3.1: Structure-based design of heptavalent anthrax toxin inhibitors. ................. 32

Figure 3.2: Synthesis scheme of heptavalent anthrax toxin inhibitor. ............................ 34

x

Figure 3.3: Characterization of a well-defined heptavalent anthrax toxin inhibitor. ..... 36

Figure 3.4: Influence of linker length on the activity of heptavalent inhibitors. ............ 37

Figure 4.1: Combined image of HA from influenza A/South Carolina/1/1918 (SC1918/H1)

with structures of several strain-specific antibodies and the broadly-neutralizing

antibody CR6261. ..................................................................................................... 40

Figure 4.2: Comparison of crystal structure of the broadly-neutralizing Fab CR6261 in

complex with HA to the crystal structures of the designed hemagglutinin-binding

proteins in complex with HA. .................................................................................. 42

Figure 4.3: Estimated distance between HB protein binding sites on the HA trimer stalk.

................................................................................................................................. 45

Figure 4.4: SDS-PAGE of IMAC fractions from FLAG-HB36.5 purification revealing non-

FLAG-tagged HB36.5 co-production. ....................................................................... 48

Figure 4.5: Predicted net protein charge at various pH values of three HB36.5 variants.

................................................................................................................................. 49

Figure 4.6: SDS-PAGE of IMAC fractions from DDDG-FLAG-HB36.5_A341C_ΔHM

purification. .............................................................................................................. 52

Figure 4.7: UV chromatogram of DDDG-FLAG-HB36.5_ΔHM on a GE HiLoad 16/600

Superdex 200 SEC column. ...................................................................................... 52

Figure 4.8: Synthesis scheme for homodivalent PEG-protein bioconjugation. ............... 55

Figure 4.9: Synthesis schemes for polyvalent polymer-maleimide-protein bioconjugation.

................................................................................................................................. 56

Figure 4.10: NMR spectra of three different PGA samples functionalized with increasing

percentages of maleimide. ...................................................................................... 57

Figure 4.11: Silver-stained non-reducing SDS-PAGE of conjugation reactions of HBA (lanes

1-12) and HBB (lanes 13-25) before purification, 5.0 µg total HB protein per lane. 59

Figure 4.12: Influenza virus infection inhibition efficacy of homodivalent bioconjugates

of HBB. ...................................................................................................................... 62

Figure 4.13: Influenza virus infection inhibition efficacy of polyvalent bioconjugates of

HBA. .......................................................................................................................... 63

xi

Figure 5.1: Silver-stained reducing SDS-PAGE of IMAC fractions from LukE-H6C

purification. .............................................................................................................. 72

Figure 5.2: Silver-stained reducing SDS-PAGE of IMAC fractions from CH6-GFP-LukE

purification. .............................................................................................................. 72

Figure 5.3: Silver-stained non-reducing SDS-PAGE of LukE-H6C conjugation reaction

products before purification, 5.0 µg total LukE protein per lane. .......................... 73

Figure 5.4: Silver-stained non-reducing SDS-PAGE of LukE-H6C conjugation reaction

products after SEC purification on a Superdex200 column. ................................... 75

Figure 5.5: CCR5-binding inhibition efficacy of polyvalent PGA-LukE bioconjugates. .... 77

Figure 5.6: CCR5-binding inhibition efficacy of polyvalent HyA-LukE bioconjugates. .... 77

Figure 6.1: SYBR Gold-stained 15% acrylamide TBE-Urea PAGE of ssDNA model aptamer

polyvalent conjugation reactions with PGA-maleimide after 88 hours at ambient

temperature. ............................................................................................................ 83

Figure 6.2: Heat map of SYBR Gold-stained 4-20% acrylamide TBE-PAGE of ssDNA

aptamer sgc8c conjugation reactions with PGA-maleimide. .................................. 84

Figure 6.3: SEC multi-chromatogram of the PGA35k-sgc8c reaction product. ............... 85

Figure 6.4: Enhanced binding of polyvalent PGA-sgc8c conjugates to lymphocyte cell

receptors, assayed by flow cytometry. .................................................................... 86

Figure 7.1: General SELEX procedure for screening RNA aptamers. ............................... 88

Figure 7.2: The sequence of the initial DNA library and a representative mfold prediction.

................................................................................................................................. 89

Figure 7.3: Schematic representation of the new library design for screening of short

oligonucleotide aptamers. ....................................................................................... 90

Figure 7.4: Structure of CCL5 monomer (PDB ID 1HRJ), with regions important for binding

CCR5 highlighted. ..................................................................................................... 93

Figure 7.5: Inhibition of HIV-1Ba-L infection by Ac-CFAYIARPLPRA-Am-functionalized

liposomes. ................................................................................................................ 95

Figure 7.6: Divalent conjugates for targeting CCR5. ........................................................ 96

Figure 7.7: Synthesis scheme for homodivalent PEG-aptamer bioconjugates. .............. 97

xii

Figure 7.8: SYBR Gold-stained 15% acrylamide TBE-Urea PAGE of ssDNA model aptamer

homodivalent PEG conjugation reactions after 88 hours at ambient temperature.

................................................................................................................................. 98

Figure 7.9: Synthesis scheme for heterodivalent PEG-aptamer bioconjugates. ............. 99

xiii

ACKNOWLEDGMENT

This dissertation was made possible with the support of several people whom I would

like to acknowledge here.

First and foremost, I would like to thank Prof. Ravi Kane, my Ph.D. advisor. Ravi has

not only provided the majority of funding for my studies, but he has almost always been

available for feedback at any time of day, including late at night and even sometimes

while on personal vacations. He has also taught by example, especially with his patience

and careful approach to problems. Furthermore, I have been happy to find that we have

many shared interests, both academic and unrelated to our research. Ravi has always

treated me with respect, and I have greatly appreciated that throughout the years.

I would also like to thank my other doctoral committee members, especially Prof.

Pete Tessier, who has been very kind to me throughout the years. Despite not being my

primary advisor, Pete has almost always been willing to discuss my research and my

career. He has also inspired me with his teaching style and with the productivity of his

research group; I aspire to someday be as successful as he has been both as a primary

investigator and as a teacher in the classroom. Prof. Steve Cramer and Prof. Shekhar

Garde have also been substantial influences on me. Steve has been very supportive of

me and I really admire his consistent involvement with the students in the department.

Shekhar was a really great instructor to have for Thermodynamics, and he also always

provided very astute guidance to the department’s graduate student association during

the years that I participated. Shekhar has been a great model of a leader for me.

My colleagues in the Kane lab deserve special recognition for helping me with much

of the research that is presented herein. I am grateful to Dr. Dhananjoy Mondal for

teaching me organic synthesis techniques, Dr. Manish Arha and Dr. Isil Severcan for

helping me with cloning and aptamer screening, and Dr. Sunit Srivastava and Dr. Marc

Douaisi for testing countless samples for me in a variety of assays. I am also glad to have

spent time in lab with Dr. Indrani Banerjee, Dr. Sanket Patke, Dr. Mohan Boggara, and

Dr. Jeff Litt, as they were fun and friendly colleagues to work with. I also appreciate the

xiv

help of the RPI staff, some of whom I got to know very well, especially Lee Vilardi and

Rose Primett. Dorota, Nancy, and Sharon were also very friendly and helpful for me.

I would like to express my gratitude for the support of my family and friends as well.

My parents have been especially generous in helping me financially whenever I needed

it, and have always given me good advice in times when I doubted myself, so I am

extremely grateful for that. My friends from Delaware (especially Ricky Komdat, Matt

O’Brian, Steve Newth, Pat Kerrane, Dan Kerrigan, Steve Smith, and Tiffany Garber) and

Connecticut (especially Harrison Paine, Ray Naclerio, Aman Kidwai, Andrew Harris, James

Halperin, Phil Nizzardo, Steve Papen, Steve Ferketic, Ryan Notti, Kevin Duffy, and Jim

Warren) have made more trips to visit me in the past six years than I could keep track of,

and I am so glad that those friendships have persisted despite me living a drive of several

hours away for so many years. I hope that the new friends that I have made locally will

also remain friends for many years to come. Some of these friends (especially Dr. Mike

Riley, Erin Eldeen, Hannah White, Erin Keys, Ana Hoyos, Ashlee Vilardi, Ashley Cress,

Krunal Mehta, James Woo, Jeremy Sauer, Corey Lemley, Dr. Eric Sterner, Clay Albracht,

Panos Karampourniotis, Jasmine Shong, and Emily Gnacik) have really helped me to stay

sane and socially balanced, and others (especially Brady Cress, Dr. Sayaka Masuko, and

Dr. Leyla Gasimli) have not only been great friends but have also been genuine

inspirations to me with their dedication to their research. I am thankful for all of these

friendships; they definitely helped me to stay motivated throughout the years.

Finally, I would also like to acknowledge the funding I received from the Chauncey

and Doris Starr Fellowship during my first two years at RPI, as well as the NIH grants

which enabled my advisor, Prof. Kane, to continue to support me during my studies.

xv

ABSTRACT

Here I describe the structure-based design of macromolecular bioconjugates for the

effective inhibition of three different diseases with broad public interest: anthrax,

influenza, and HIV/AIDS. To that end, I have synthesized multivalent arrangements of

bioactive macromolecules (i.e., peptides, proteins, and oligonucleotides) on

biocompatible scaffolds in order to enhance the inhibitory efficacy of those

biomolecules. Multivalency is the simultaneous interaction of multiple binding elements

with multiple target receptors, and this phenomenon underlies a powerful strategy for

controlling the potency of bioactive molecules by manipulating their context. For

example, some multivalent ligand-receptor interactions are known to exhibit binding

avidities that are several orders of magnitude stronger than the corresponding

monovalent receptor-ligand binding affinities. An important aspect of this research was

the identification of appropriate targets from the etiology of each of these diseases. Once

suitable targets were identified, I used strategically-chosen scaffolds to control the

multivalent arrangements of targeting ligands. Specifically, the anthrax toxin protective

antigen heptamer was the target of multivalent anti-toxin inhibitors that display peptide

ligands on either linear or radial scaffolds. Similarly, a conserved region of the influenza

hemagglutinin glycoprotein was targeted by a series of linear divalent and polyvalent cell

entry inhibitors. Finally, a recently-identified protein ligand for CCR5, a co-receptor for

HIV cell attachment and entry, was used for polyvalent display on linear polymer

scaffolds, and the ability to obstruct CCR5 binding was shown. The results support the

potential for multivalent bioconjugates such as these to lead to very promising

therapeutic applications.

1

1. Introduction1

1.1 Anthrax1

1.1.1 Overview

The disease anthrax is caused by infection with the bacterial species Bacillus anthracis,

which can occur through the inhalation of bacterial spores.2 For this reason, B. anthracis

is well known to the general public as a potential bioterrorism and/or biowarfare agent.2,3

There are two major virulence factors which contribute to the pathogenicity of B.

anthracis. First, the bacteria produces a poly-D-glutamic acid capsule which protects

against macrophage phagocytosis4 and allows the bacteria to spread systemically.

Second, B. anthracis produces anthrax toxin, which attacks the cells of the immune

system and causes death of the host.5,6 Ciprofloxacin is an antibiotic that has been

approved by the Food and Drug Administration (FDA) in the United States for post-

inhalational treatment of anthrax spores.7 Antibiotics such as ciprofloxacin may help to

reduce the proliferation of the pathogen within the body, but they do not prevent the

damage caused by the release of bacterial toxins. Thus there is a need for an antitoxin

treatment.

1.1.2 Intoxication pathway

Anthrax toxin is comprised of individually nontoxic monomeric proteins known as

protective antigen (PA), lethal factor (LF), and edema factor (EF).5,6,8 The combination of

PA and LF is called lethal toxin (LeTx) and the combination of PA and EF is called edema

toxin (EdTx). The individual components assemble into toxic complexes upon interacting

with receptors on the surfaces of target cells, as represented in Figure 1.1. The detailed

mechanisms governing the entry of LF and EF into target cell cytosol has been reported

in the literature.9–12

Portions of this chapter previously appeared as: Martin, J. T.; Kane, R. S. Design of Polyvalent Polymer Therapeutics. In Functional Polymers by Post-Polymerization Modification; Theato, P.; Klok, H.-A., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2013; pp. 267–290.

2

Figure 1.1: Representation of the anthrax intoxication pathway.

(1) Protective antigen (PA) binds the cellular receptors ANTXR1 and ANTXR2. (2) The amino-terminal 20 kDa fragment of PA (PA20) is cleaved off by a furin-like protease, while (3) the remaining fragment (PA63) self-associates into heptamers ([PA63]7). Alternatively, [PA63]7 can form in the blood and bind ANTXR1 or ANTXR2. (4) Edema factor (EF) and lethal factor (LF) bind [PA63]7 and (5) the complexes internalize by receptor-mediated endocytosis. (6) The low pH of the endosome triggers the insertion of [PA63]7 into the membrane and translocation of EF and LF. Reprinted with permission from Collier, R. J.; Young, J. A. T. Annu. Rev. Cell Dev. Biol. 2003, 19, 45. Copyright 2003 Annual Reviews.

First, PA binds to the human cellular receptors anthrax toxin receptor 1 (ANTXR1, also

known as ATR and TEM8)13 or anthrax toxin receptor 2 (ANTXR2, also known as CMG2),14

which are widely expressed in epithelial cells.15 Cell-surface proteases cleave a 20 kDa

fragment off of PA that does not play a further role in the intoxication pathway.16,17 The

remaining 63 kDa fragment (PA63) subsequently self-assembles into toroid-shaped

prepores18 that can function as octamers,19 but are usually primarily ([PA63]7)

heptamers.20,21 Either way, the oligomerization of PA63 is necessary for LF and EF binding22

and for clustering the ANTXR on the cell surface, which then induces clathrin-mediated

endocytosis.23 Once the toxin complexes are in the acidic environment of the endosome,

3

[PA63]7 inserts seven loops into the membrane to form a 14-stranded beta-barrel channel,

allowing the enzymatic LF and EF to translocate via a charge state-dependent Brownian

ratchet mechanism.11,12,24,25

LeTx and EdTx exhibit various immunosuppressive effects, and the bacterial

expression of either toxin independently can be fatal to the host organism.26–29 EF binds

to calmodulin in the cytosol and exhibits potent adenylate cyclase activity.30,31 The

increased cAMP affects cytokine expression, leading to inhibition of neutrophil

phagocytosis of B. anthracis.32,33 LF is a zinc-dependent protease that cleaves mitogen-

activate protein kinase kinases (MAPKKs) 1-4, 6 and 7; this in turn downregulates signaling

of extracellular-signal-regulated kinases, p38 mitogen-activated protein kinases, and c-

Jun N-terminal kinases.34,35 This activity of LF affects a multitude of cell types, including T

cells, dendridic cells, macrophages, and endothelial cells, by inhibiting cytokine

expression. Inhibition of the production of Type IIA phospholipase A2 by macrophages, of

differentiation of monocytes into macrophages, of production of immunoglobulin by B

cells, of production of superoxide by neutrophils, and of neutrophil motility are also the

result of LeTx exposure.36–38 Finally, LeTx exhibits a cytotoxic effect on macrophages,

dendritic cells, and some endothelial cells, and may eventually cause morbidity due to

vascular leakage and organ failure.28,29

1.2 Influenza

1.2.1 Overview

Influenza, commonly known to the general public as simply “the flu,” is a disease caused

by a virus that infects humans and a variety of other mammals and birds. There are three

types (analogous to species) of influenza virus, known as influenza A, B, and C,39 which

are classified based on the viral proteins on the exterior of the virions and the

characterization of the RNA genomes.40–42 Phylogenetically, influenza A and B are more

closely related than C, which is an indicator that influenza C diverged from the

evolutionary tree in the more distant past.42 Another difference between the types is that

influenza B and C are essentially limited to infecting human hosts, with only a couple of

4

known exceptions.39 In contrast, the native hosts of influenza A are waterfowl.39,43 While

influenza A is normally asymptomatic in aquatic birds, it is typically the most virulent type

in humans; it should be noted that the factors which contribute to influenza pathogenicity

in avian hosts may be different from those in humans.43 The natural avian hosts sustain

the reservoir of viral genetic variation from which the virus evolves to infect other animal

species, a process known as zoonosis.42 This process often involves reassortment of viral

genomes and can lead to the sudden appearance of new, pandemic-causing influenza A

strains. Genome reassortment is restricted by the influenza type,44 and I will be focused

on influenza type A for the remainder of this document.

1.2.2 Virion morphology and infection pathway

Influenza virions are known to be pleomorphic,40,41 which means that the virions can exist

in more than one general shape. In vitro, the viral particles are typically characterized as

being spherical with a diameter of roughly 100 nm, but in vivo the virions have been

reported to take on more filamentous appearances, with diameters of ~100 nm and

lengths of up to 20 µm.41 Another three classes of virions have been described for a total

of five different classes.40 Although the functional significance of this pleomorphism is

unknown, it has been found that all forms of the virions contains just one copy of the viral

genome.45,46 The genome is packaged inside the virion as a ribonucleoprotein complex, in

which the RNA segments are coated with nucleoprotein and an RNA-dependent RNA

polymerase.44 This complex and another protein called the nuclear export protein are

enclosed in a viral capsid formed by the M1 protein.41 Enveloping the viral capsid of the

mature virion is a lipid bilayer membrane, in which three transmembrane proteins are

embedded. Two of the three proteins are hemagglutinin (HA) and neuraminidase (NA),

which form the glycoprotein “spikes” that point outward from the viral envelope, and the

third envelope protein is the matrix ion channel protein (M2). HA exists as a homotrimer

and NA exists as a homotetramer; these are immediately formed during the process of

viral protein expression. The typical center-to-center distance between the glycoprotein

spikes on the virion surface is about 11 nm on average, and a spherical virion of average

size typically displays about 350 of these spikes in a roughly 6:1 ratio of HA to NA.40

5

1.2.3 Antigenic variation

The genome of influenza consists of eight separate negative-sense, single-stranded RNA

segments.44 The ability of these segments to recombine in a host cell that has been

infected with more than one strain of influenza can create an antigenic shift, which can

sometimes result in the sudden emergence of a pandemic-causing strain. Antigenic drift

is another type of antigenic variation, but describes the more continuous changes in the

antigenic sites of the viral glycoproteins over time.47,48 Proteins which are characterized

by the ability to consistently perform their primary functions while at the same time

tolerating a wide range of antigen-altering mutations are known as having structural

plasticity. The HA and NA of influenza are some of the best examples of proteins with a

high degree of structural plasticity, and this feature enables influenza to quickly develop

mutants that escape antibody neutralization by the mammalian adaptive immune

system.

1.2.4 Infection pathway

The first step in viral infection is cell attachment. The influenza virus attaches to cells by

the binding of HA to N-acetylneuraminic (sialic) acid on the cell surface.44 Sialic acids are

monosaccharides that are common on the ends of many kinds of glycosylated proteins

and exist in the airways and lungs of mammals and birds, which explains the mechanism

of transmission. Once attached to a cell, the virion is endocytosed.44 In the endosome,

acidic pH triggers a conformational change in the hemagglutinin stem/stalk region that

exposes a cell membrane fusion peptide and draws the viral and cell membranes

together.44 Meanwhile, the M2 ion channel pumps protons from the endosome into the

viral interior in order to release the ribonucleoprotein complex from the viral matrix. After

entering the cytosol, the nuclear localization signal of the ribonucleoprotein complex

enables delivery of the viral RNA and RNA-dependent RNA polymerase into the nucleus.44

The negative-sense RNA genes are then transcribed in the nucleus into positive-sense

messenger RNA and positive-sense complementary RNA for transcribing additional copies

of the negative-sense genome.44 The messenger RNAs are translated in the endoplasmic

6

reticulum, and the viral envelope proteins are apically sorted by the golgi apparatus to

specific locations on the cell surface for virion assembly.44 The virions begin to bud off

from the host cell once a sufficient amount of viral M1 protein has accumulated along the

lipid bilayer and a full set of the segmented viral genome has been packaged.44 The NA

then cleaves any extracellular sialic acid which would otherwise prevent the virions from

disengaging from the initially-infected host cell.44

1.3 Human Immunodeficiency Virus (HIV)

1.3.1 Overview

Despite over thirty years of intense scientific effort to develop cures and vaccines, the

Joint United Nations Programme on HIV/AIDS estimates that there were 2.5 million new

HIV infections in 2011.49

1.3.2 Infection pathway

HIV targets human cells primarily via a host cell surface glycoprotein called CD4 molecule.

Interaction with CD4 tethers the HIV virion to the cell surface and allows for binding to

co-receptors C-C chemokine receptor type 5 (CCR5) or C-X-C chemokine receptor type 4

(CXCR4).50,51 The interaction with CD4 induces a conformational change in gp120 that

allows it to complex with the co-receptor, and further conformational rearrangement

causes the release of gp41 into a fusogenic hairpin structure.52–54 This glycoprotein

structure pierces the host cell membrane and brings it and the viral membrane together,

resulting in the fusion of the viral and cell membranes.

Once the virion has entered the cell, the RNA from HIV is internalized, reverse

transcribed into double-stranded DNA, and integrated into the host cell chromosome by

a combination of viral and host enzymes.50,51 The integration into the host chromosome

allows the provirus to lie dormant for some time post-infection. When the viral DNA is

transcribed, several different viral mRNAs are produced and then translated into both

regulatory and structural proteins. The self-assembly of structural proteins with the viral

genomic RNA forms a core,51 which translocates to the cell surface and buds off, acquiring

7

the viral envelope by taking from the host cell membrane. During the final stage of viral

budding, a glycosaminoglycan polyprotein precursor is cleaved by a virally-encoded

protease, which results in the morphological progression into mature virions.51

1.3.3 Current treatment strategies

Pharmaceutical companies have discovered multiple classes of small molecule drugs that

target each of the various steps in the HIV replication pathway. Currently available classes

of antiretroviral drugs include viral entry/fusion inhibitors, nucleoside and nucleotide

reverse transcriptase inhibitors, non-nucleoside reverse transcriptase inhibitors,

integrase inhibitors, protease inhibitors, and maturation inhibitors. More experimental

methods based on RNA interference have also been shown to inhibit HIV replication,55,56

but unfortunately, each of these strategies is susceptible to the emergence of HIV strains

that are resistant.

HIV resistance is just one of several serious medical concerns with current

antiretroviral drugs. Drug resistant strains evolve due to the error-prone nature of the HIV

reverse transcriptase enzyme and the high rate of viral turnover in infected individuals.

The evolution of resistance can be greatly affected by patient compliance to the

treatment regimen, such as improper administration or missed doses. In order to combat

the emergence of drug resistance, the current treatment strategy is to prescribe a

combination of several classes of drugs (at least three) into one “cocktail,” a strategy that

is known simply as combination antiretroviral therapy (cART) or highly active

antiretroviral therapy (HAART). However, many of these drugs can produce dangerous

side effects.57 Intolerance to side effects is a serious issue that can make it difficult to find

a satisfactory combination of therapies. In addition, the need to administer multiple drugs

simultaneously can also be extremely costly, especially for individuals in the countries

where populations are at the greatest risk, because of the high rate of poverty in those

areas.

Due to the concerns associated with current HIV treatments, the best solution to

reducing the numbers of HIV infections may be to prevent the transmission of the virus

to uninfected individuals. This goal continues to drive efforts to develop an effective HIV

8

vaccine,58–60 but there have been only two unsuccessful trials conducted to date.61–63

Recent research has shown that proper adherence to a HAART regimen can significantly

help reduce the risk of spreading HIV, but this finding is still subject to the problems

mentioned above. With these factors in mind, a prophylactic microbicide that could be

applied prior to intercourse may be a very worthwhile objective. Such a prophylactic must

not only be potent, but also able to impede the evolution of viral resistance.

1.4 Multivalent therapeutics

1.4.1 Multivalency

Multivalency is one of the most notable strategies used by nature to modulate

interactions between biological machinery.64 A multivalent system is characterized by an

arrangement of multiple binding elements on one entity interacting with another (often

biological) entity. Relative to the binding strength of an individual ligand-receptor pair (a

monovalent interaction), the corresponding multivalent or polyvalent (multivalency of a

very high valency) interaction is frequently enhanced by many orders of magnitude,64,65

and there have been a multitude of reviews in the past fifteen years highlighting the

potential benefits of utilizing multivalency.64,66–73 However, multivalency can play a role

in a wide variety of biological interactions that are not merely limited to the enhancement

of binding strength. For example, multivalent interactions can affect biological

signaling64,70,74 by establishing scaled interactions, by creating combinations of biological

interactions, and by maintaining contact between two surfaces over a large area. Such

multivalent interactions may play an important role in determining cell fate.75 In addition,

potent multivalent binders can be used to inhibit undesired interactions such as viral

attachment to cells76–82 and bacterial toxin assembly.83–96 Finally, multivalency can be

used to improve the specificity of therapeutics for certain applications such as tumor cell

targeting.97,98 Due to the fact that these cells have overexpressed receptors on the cell

surface, the best specificity for a therapeutic may be one that does not use targeting

ligands that are monovalently strong binders; in this case, using a polyvalent therapeutic

in which the monovalent contacts are only very weakly interacting should provide the

9

greatest contrast between desired and off-target binding. As is evident from these

examples, many of these different types of multivalent interactions have therapeutic

implications, and there has been growing interest in implementing multivalency for

therapeutic purposes.71–73

1.4.2 Scaffolds for multivalent therapeutics

Some of the structures that have been used to support multivalent or polyvalent displays

include nanoparticles, often gold or iron oxide, and derivatives of naturally occurring

constructs, such as viral particles and liposomes. However, synthetic polymers have

advantages in that they can be used in a broad range of structures (scaffolds) to

manipulate the presentation of biomolecules, and the details of the presentation can be

sequentially modified, thus resulting in the optimization of activity through multivalency.

For all types of linear polymer therapeutics (including end-functional, side-chain-

functional linear polymers, or branched (brush) polymers), the molecular weight of the

polymer backbone is one of the primary variables in the scaffold design. Whether using

pre-synthesized polymers or controlled polymerization techniques for custom synthesis,

the molecular weight, the DP, and the polydispersity index (PDI) are measures of how

well-defined the polymer sample is. These factors affect the hydrodynamic radius and

therapeutically relevant qualities such as the enhanced permeation and retention (EPR)

effect in tumors, other pharmacokinetic and biodistribution effects, and biodegradability

or toxicity.

There are a multitude of good reviews on the synthesis of polymer conjugates,99–108

but the main strategies relevant for synthesis of multivalent polymer therapeutics are

“grafting through” and “grafting to”. The “grafting through” strategy involves the

polymerization of monomers that have been pre-conjugated with the desired ligand,

which can yield polymers that display a high density of ligands along the polymer

backbone. However, there are several disadvantages of this strategy. Lack of reaction

modularity and lack of reproducibility are major disadvantages that arise from the need

to tune the reaction conditions of each polymerization and the difficulty in handling or

accurately measuring the minuscule quantities of expensive, modified-biomolecule

10

ligands, respectively. Furthermore, the difficulties involved in characterization and

analysis of qualities such as polydispersity and ligand composition can hinder the

optimization of the product multivalent presentation. In contrast, the “grafting to”

strategy involves the attachment of ligands to the polymer scaffold post-polymerization.

By using controlled polymerization and traditional polymer characterization techniques

to reveal accurate descriptors of the pieces involved pre-conjugation, product analysis is

much easier. That information can then be used in turn to guide subsequent multivalent

designs, which can often be synthesized without complication of reaction conditions.

1.4.3 A multivalent anthrax toxin inhibitor with in vivo efficacy

One of the first examples of a polymeric multivalent conjugate that was effective in vivo

was the anthrax toxin inhibitor reported by Mourez et al.86 This polyvalent polymer

therapeutic inhibited the assembly of anthrax toxin. The researchers used phage display

to identify a peptide sequence (HTSTYWWLDGAP) that binds the [PA63]7 heptamer and

inhibits toxin assembly with a half-maximal inhibitory concentration (IC50) of 150 µM

(weakly). Multiple copies of a modified peptide containing the [PA63]7-binding sequence

and a C-terminal lysine, Ac-HTSTYWWLDGAPK-Am, were attached pendant to a

polyacrylamide polymer backbone. The polyvalent conjugation was performed by

reacting poly(N-acryloyloxysuccinimide) (pNAS), an activated ester form of polyacrylic

acid, with the amine-functionalized peptide ligand. This protocol was based on previous

reports64,78–80 on the synthesis of polyvalent displays of sialic acid along polyacrylamide

backbones. The polyvalent anthrax toxin inhibitor thus created was 7,500-fold more

potent than the monovalent peptide at inhibiting anthrax toxin complex formation in

vitro, on a per-peptide basis. Furthermore, the polyvalent inhibitor was shown to have in

vivo efficacy against a toxin challenge, preventing symptoms in a rat intoxication model.

The successful inhibition of the toxin in vivo demonstrated that the therapeutic

design strategy was sound. The study86 showed that it is possible to construct a

therapeutically effective conjugate from weakly-binding, previously-unknown ligands

without extensive modification or optimization. Ligands which have been screened from

a random library for a desired target molecule can be made significantly more efficacious

11

by multivalent display on a suitable scaffold. Our group has placed a particular emphasis

on controlling the architecture of multivalent scaffolds in order to better understand the

influence of the design parameters on the efficacy of multivalent therapeutics. Here, we

will primarily highlight our studies on polymer-based scaffolds. Specifically, we will discuss

controlling the molecular weight and pendant ligand spacing for linear polymeric

scaffolds91,109,110 We will also include some discussion on increasing the biocompatibility

and biodegradability of the conjugate through the use of known biocompatible polymers,

such as N-(2-hydroxypropyl)methacrylamide (HPMA), or synthetic polypeptide-based

scaffolds, such as poly(glutamic acid) (PGA).92 While there have been many studies on the

structure-activity relationships of potential multivalent therapeutics (see recent reviews

on multivalent therapeutics111,112 and more general reviews of polymer therapeutics113–

116), the discussion here will focus primarily on the previous work of the Kane research

group towards developing polyvalent inhibitors of anthrax toxin.

1.4.4 Multivalent anthrax toxin inhibitors with control over average ligand spacing

In the interest of investigating the role of the geometry of a multivalent display (such as

the arrangement and spacing between ligands) on the resulting therapeutic efficacy, we

pursued linear polymer modification. In the context of multivalent therapeutic design, the

molecular weight of a polymer scaffold is particularly important for determining the

potential valency and density of a conjugated ligand. Moreover, with greater precision

over DP and PDI, the degree of uncertainty in conjugate characterization can be reduced

so that structure-activity relationships and optimized therapeutic properties can be

elucidated with greater confidence. For these reasons, we91,109 and others80,117 have

investigated the effects of scaffold molecular weight on post-polymerization functional

polymers by minimizing the effects of PDI.

Two different approaches were used to create multivalent inhibitors using polymer

scaffolds with low PDI. The work by Gujraty et al.91 took advantage of the fact that

poly(tert-butyl acrylate) (pTBA) was commercially available in a range of molecular

weights (e.g., 28.4, 69, 100, and 150 kDa) with low PDIs (ranging from 1.03 to 1.2). For

each sample, the pTBA was treated to produce low-PDI pNAS (Figure 1.2). The resulting

12

polymer contained the same activated esters and polymer backbone that were used to

make the anthrax toxin inhibitor reported by Mourez et al.,86 except that in this case the

starting materials had molecular weights that were more well-defined, leading to

products that could be compared with greater confidence. An alternative approach

toward low-PDI polymer scaffolds by Yanjarappa et al.109 relied on the RAFT is a

controlled, living polymerization technique known as reversible addition-fragmentation

chain transfer (RAFT) polymerization to create an activated polymer backbone with well-

controlled molecular weight. The resulting heteropolymers of N-

methacryloyloxysuccinimide (NMS) and N-(2-hydroxypropyl)methacrylamide (HPMA)

were characterized by 1H NMR spectroscopy and size-exclusion chromatography (SEC)

confirmed the ability to control the composition and size of the polymer chains.

Figure 1.2: Synthesis of a polyvalent anthrax toxin inhibitor of controlled molecular weight.

(a) TFA/CH2Cl2. (b) N,N’-Carbonyldiimidazole, N-hydroxysuccinimide, pyridine, 110°C. (c) i. peptide, DMF. ii. NH4OH. Reprinted with permission from Gujraty, K. V; Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Biomacromolecules 2006, 7, 2082. Copyright 2006 American Chemical Society.

The controlled molecular weight of the resulting activated polymers was exploited to

compare the efficacies of a range of polyvalent inhibitors that varied in both the density

of the conjugated ligands on the backbone and in the backbone molecular weight. While

the use of commercially available pTBA allowed for the synthesis of low-PDI

homopolymers of pNAS, the use of RAFT to create activated copolymers of HPMA with

controlled molecular weight and PDI demonstrated that various polymer backbones are

generally effective for use in polyvalent therapeutics. The results (Figure 1.3) revealed

that for polyvalent inhibitors prepared from the same size backbone and various degrees

13

of ligand conjugation, there existed a ligand coupling percentage corresponding to a peak

in potency, and that higher ligand density led to a plateau in potency (Figure 1.3a).

Furthermore, a similar pattern was observed when comparing a range of backbone sizes

for a specified ligand density. That is, potency increased with increasing backbone length

up to a point, beyond which the potency plateaued (Figure 1.3b). For this latter set, the

increase in backbone length at constant peptide density basically meant there was a

greater number of ligands per chain on average but that the average spacing between

ligands along the chain was the same. These results showed that although polyvalency

can provide orders of magnitude enhancements of potency over monovalent ligands,

beyond a certain point, increasing valency may not correlate with further increases in

potency.

Figure 1.3: Influence of peptide density on the potency of a polyvalent anthrax toxin inhibitor.

(a) Number of monomer repeat units in the backbone, N ca. 200. (b) Influence of number of monomer repeat units on the inhibitory potency (peptide density ca. 3%). IC50 values are reported on a per-peptide basis. Reprinted with permission from Gujraty, K. V; Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Biomacromolecules 2006, 7, 2082. Copyright 2006 American Chemical Society.

1.4.5 Multivalent anthrax toxin inhibitors with predefined ligand spacing

Well-defined polymer scaffolds with low PDI were needed to accurately characterize the

polyvalent inhibitors that were described in the previous section. However, due to the

nature of the subsequent ligand grafting reactions, the grafted ligands on all of the

inhibitors were randomly distributed along the portions of the scaffold which had been

14

pre-activated. While statistical interpretations of the characterization data could lead to

deductions about the average valency and average inter-ligand spacing, it was

hypothesized that more precise control over the placement of the activated monomers

would help to gain an even more comprehensive understanding of the relationship

between structure and activity in polyvalent systems. With this goal in mind, the semi-

batch RAFT polymerization approach described previously was modified to construct

polymers with defined spacing between reactive monomers.110

Table 1.1: Polymerization of NAS at various [M]/[CTA] ratios.a

Reprinted with permission from Gujraty, K. V; Yanjarappa, M. J.; Saraph, A.; Joshi, A.; Mogridge, J.; Kane, R. S. J. Polym. Sci. A. Polym. Chem. 2008, 46, 7246. Copyright 2008 Wiley Periodicals, Inc.

The first step was to determine what distances between ligands would likely provide

a reasonable range of activity in the designed polyvalent inhibitor. To solve this problem,

a range of homopolymers of NAS with low PDIs were prepared by RAFT. The samples

differed in their DP as determined by SEC (Table 1.1). The Ac-HTSTYWWLDGAPK-Am

peptide was conjugated with each polymer sample such that the number of copies of the

ligand that were attached was controlled, and the remainder of the polymer backbone

was “quenched” by reaction with ammonium hydroxide. An average of three copies of

the peptide were conjugated to each individual polymer chain (as determined by 1H NMR

spectroscopy), but due to the differences in the backbone chain length of each sample, it

was inferred that the resulting polyvalent inhibitors would have different averages for the

number of unconjugated monomers separating each attached peptide. As illustrated in

Figure 1.4, the difference in ligand spacing could conceivably lead to a measurable

15

variation in sample efficacy. Indeed, for the inhibitor sample with DP = 120, the inhibition

IC50 of the binding of LF to [PA63]7 was over an order of magnitude lower than the IC50 of

the polyvalent inhibitor with DP = 80 (Figure 1.5a). Furthermore, as shown in Figure 1.5b,

the polyvalent inhibitor with DP = 120 was able to inhibit the cytotoxicity of anthrax lethal

toxin on RAW264.7 cells in vitro with an IC50 of 36 nM, yet the inhibitor with DP = 80 was

unable to inhibit cytotoxicity even at concentrations as high as 1 µM on a per-peptide

basis. With these results in hand, it was possible to design the polymers with controlled

spacing between reactive monomers so that the controlled spacing matched that of the

average spacing achieved by the homopolymer conjugation protocol. For a polyvalent

inhibitor with a number of peptides “𝑖” conjugated randomly along the polymer

backbone, the average spacing between adjacent conjugated peptides was calculated by

DP (𝑖 + 1)⁄ . Accordingly, for the polyvalent inhibitor with DP = 120 and 𝑖 = 3, the average

number of monomers of separation was estimated to be roughly 30. This spacing thus

served as the target number of non-active acrylamide (AAm) monomers to incorporate

between reactive NAS monomers in a controlled RAFT polymerization of a

heteropolymer, poly(AAm-co-NAS).

Figure 1.4: Design of polyvalent inhibitors with control over molecular weight and ligand spacing.

The linear polyvalent inhibitors displaying peptides (black ovals) are shown bound to the PA63 heptamer at the peptide-binding sites (circles). The spacing between peptides on the linear scaffold is either too short (left panel) or is sufficient (right panel) to allow a polyvalent interaction. Reprinted with permission from Gujraty, K. V; Yanjarappa, M. J.; Saraph, A.; Joshi, A.; Mogridge, J.; Kane, R. S. J. Polym. Sci. A. Polym. Chem. 2008, 46, 7246. Copyright 2008 Wiley Periodicals, Inc.

16

Figure 1.5: Inhibitory activity of polyvalent inhibitors of anthrax toxin derived from a controlled molecular weight homopolymer, pNAS.

(a) Inhibition of the binding of the anthrax lethal factor. (b) Inhibition of toxin-induced cytotoxicity at various concentrations of polymeric inhibitors with DP = 80 (open squares) and DP = 120 (filled circles). Control polymer (polyacrylamide) did not show any inhibitory activity at concentrations tested (open triangles). Reprinted with permission from Gujraty, K. V; Yanjarappa, M. J.; Saraph, A.; Joshi, A.; Mogridge, J.; Kane, R. S. J. Polym. Sci. A. Polym. Chem. 2008, 46, 7246. Copyright 2008 Wiley Periodicals, Inc.

After analysis of the homopolymers was complete, the procedure that was used to

synthesize controlled heteropolymers of NMS and HPMA109 was adapted for the purpose

of creating a specific arrangement of reactive NAS monomers within a low-PDI poly(AAm)

scaffold. Exploiting knowledge of the rate of homopolymerization of AAm and the fact

that the NAS monomers have much faster reaction kinetics of attachment to the growing

polymer chain during RAFT than the AAm monomers, small injections of NAS were made

at calculated time points during the RAFT polymerization of AAm. As can be seen in Figure

1.6a, the NAS was added to the reaction 90, 126, and 162 minutes into the polymerization

in order to generate a spacing of ca. 31 AAm monomers between each pair of NAS

monomers. In a second sample (Figure 1.6b), the time interval between successive

additions of NAS was shorter, leading to a spacing of ca. 18 AAm monomers. By using the

reactive NAS monomers to conjugate the amine-functional Ac-HTSTYWWLDGAPK-Am

peptide, polyvalent inhibitors with controlled ligand spacing were created. The effect of

this controlled ligand display on inhibitor performance is shown in Figure 1.6c; in the in

vitro cytotoxicity assay, the polyvalent inhibitor with ca. 31 AAm monomers between

17

peptide ligands exhibited an IC50 that outperformed the inhibitor with a spacing of ca. 18

AAm monomers by two orders of magnitude.

Figure 1.6: Characterization of polyvalent inhibitors based on poly(AAm-co-NAS) copolymers.

Characterization of molecular weight during semibatch RAFT copolymerization of polymers with an average spacing of (a) 31 and (b) 18 units between adjacent blocks. (c) The inhibitory activity of the resulting peptide-functionalized polyvalent inhibitors in a cytotoxicity assay. Reprinted with permission from Gujraty, K. V; Yanjarappa, M. J.; Saraph, A.; Joshi, A.; Mogridge, J.; Kane, R. S. J. Polym. Sci. A. Polym. Chem. 2008, 46, 7246. Copyright 2008 Wiley Periodicals, Inc.

The polyvalent inhibitor with increased space between ligands presumably had

better performance due to a combination of the following attributes: more adequate

matching of the distance between the ligands along the polymer backbone and the

distance between binding sites on [PA63]7 (as illustrated in Figure 1.4, page 15) and

increased steric blocking of lethal factor from binding [PA63]7 due to increased polymer

presence at the heptamer surface. Although the steric stabilization of surfaces bound by

18

a polyvalent inhibitor is expected to play a role in inhibition of biological interactions,79

factors that enhance the binding strength of the inhibitor account for the majority of the

inhibitor’s efficacy.64 Thus, it is desirable to adjust the architecture of the polyvalent

inhibitor such that the complementarity to the target’s binding epitopes is maximized. As

an additional example, the target of the anthrax toxin inhibitors described thus far has

been [PA63]7, a heptameric protein oligomer with seven binding sites for the peptide

ligand containing the sequence HTSTYWWLDGAP. For the polyvalent inhibitors of varying

controlled molecular weights that were prepared by modifying pTBA, the sample that

displayed the greatest inhibitory activity was the one that had an average of 6 ± 2 peptide

ligands per polymer chain,91 which means the valency of the inhibitor and target were

nearly matched.

1.5 Motivation

The initial work by Mourez et al.86 showed that it is possible to construct a therapeutically

effective conjugate from weakly-binding, previously-unknown ligands without extensive

modification or optimization. While such a study was encouraging because of the

implication that the road to developing new constructs that are therapeutically active

may not be too difficult, the actual translation of such approaches into clinically relevant

treatments has yet to be realized. In addition, the previous research on multivalent

inhibitors had begun to recognize the importance of the design parameters of ligand

spacing and valency, as described in Section 1.4. Therefore I was motivated to study the

design parameters of multivalent bioconjugate synthesis with greater precision, as well

as to try to expand the approach of designing multivalent inhibitors to new applications:

inhibition of influenza virus and HIV.

19

2. Preclinical development of polyvalent inhibitors of anthrax toxin2

The Kane research group has gained a great deal of experience the field of

multivalent/polyvalent conjugate design by studying a wide range of multi/polyvalent

anthrax toxin inhibitors that show potent activity in in vitro and in vivo toxin challenges

using rodent models. However, we firmly believe that this therapeutic strategy could find

real use in the medical field. To that end, I have put considerable time and effort into

preclinical development studies of polyvalent anthrax toxin inhibitors, as described in this

chapter.

2.1 Therapeutic suitability

2.1.1 Biocompatibility

When designing a polymeric substance to be administered within the human body or

within the environment, some of the primary concerns are biocompatibility and

biodegradability. For a polymer therapeutic to be biocompatible, the polymer and all of

its metabolites should be non-toxic, and administration should not induce excessive

immunological reactions. Biodegradability describes the ability of a material to be

decomposed naturally either due to natural chemical or enzymatic processes, and is

desirable in the context of most therapeutics because the ability of a material to be

naturally cleared from the body is correlated with good biocompatibility. The ability to

use pHPMA, a proven biocompatible polymer,118,119 as an effective multivalent or

polyvalent scaffold has already been described in a previous section.109 However, HPMA

is not biodegradable. On the other hand, the polypeptide-based polymer, poly(L-glutamic

acid) (PGA), possesses several properties desirable for drug carriers, including high water

solubility, biodegradability, and low toxicity (biocompatibility).120 PGA backbones have

been used to prepare a variety of polyvalent displays.81,93,121,122 Furthermore, a paclitaxel-

PGA conjugate named XYOTAX™ has shown encouraging outcomes in clinical studies.114,123

Portions of this chapter previously appeared as: Martin, J. T.; Kane, R. S. Design of Polyvalent Polymer Therapeutics. In Functional Polymers by Post-Polymerization Modification; Theato, P.; Klok, H.-A., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2013; pp. 267–290.

20

2.1.2 PGA inhibitor synthesis

The synthesis of anthrax toxin inhibitors based on PGA was originally described by Joshi

et al.92 This study used an ester activation strategy similar to that used for pAA to modify

the carboxyl groups on the side chains of this polypeptide with the amine-functional

[PA63]7-binding peptide ligand, Ac-HTSTYWWLDGAPK-Am, as shown in Figure 2.1.

Figure 2.1: Synthesis of a PGA-based polyvalent anthrax toxin inhibitor.

(a) EDC, H2O; (b) HOBt, DMF; (c) peptide, DMF, TEA (d) NH4OH, dialysis. The polyvalent inhibitor is a random copolymer composed of peptide conjugates (x repeat units) glutamine (y repeat units), and glutamic acid (z repeat units). Reprinted with permission from Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Bioconjug. Chem. 2006, 17, 1265. Copyright 2006 American Chemical Society.

2.1.3 PGA inhibitor efficacy in vitro

The resulting inhibitors synthesized by Joshi had similar activity to the corresponding pAA-

based inhibitors in an in vitro cytotoxicity assay (Figure 2.2). For the purpose of several

preclinical studies assaying the therapeutic potential of these PGA-based anthrax toxin

inhibitors, I have successfully been able to scale up the synthesis protocol to generate

hundreds of milligrams of polyvalent inhibitors per batch. We ensured that the inhibitors

I have created also show potent activity by testing in additional in vitro assays of

cytotoxicity inhibition (Figure 2.3).

21

Figure 2.2: Inhibition of cytotoxicity in vitro by PGA-based polyvalent inhibitors.

(a) PGA-based polyvalent inhibitor (open squares), polyacrylamide-based polyvalent inhibitor (filled triangles), and control polymer (ammonium hydroxide quenched activated PGA) (open triangles). (b) Influence of peptide density on the potency of polyvalent inhibitors with different molecular weights Mw = 4.5 kDa (filled circles), 12 kDa (filled squares), and 29 kDa (open squares); average number of repeat units ca. 27, 70, 170, respectively. Reprinted with permission from Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Bioconjug. Chem. 2006, 17, 1265. Copyright 2006 American Chemical Society.

Figure 2.3: Inhibition of cytotoxicity in vitro by PGA-based polyvalent inhibitors synthesized in bulk for preclinical studies.

PGA-based inhibitor with 5% peptide coupling (red line) and PGA-based inhibitor with 10% peptide coupling (blue line). Concentration is reported with respect to peptide.

22

2.1.4 Storage stability

To ensure that PGA-based anthrax toxin inhibitors would make good pharmaceutical

candidates, their shelf life stability was assayed under a range of different storage

conditions (relative humidity and temperature) and formulations (as a dry powder or

dissolved in aqueous solution). As can be seen from Figure 2.4, PGA-based inhibitors

retained their activity up to six months at room temperature, and the harsh conditions

did not appear to affect the inhibitors in a consistently deleterious manner. The shelf life

stability test is currently ongoing, and sample stability will be re-assayed at 9 and 12

month time points.

Figure 2.4: Inhibitory activity of PGA-based anthrax toxin inhibitors after long-term storage under various conditions.

The IC50 of PGA-based inhibitors with 5% peptide coupling stored dry (red) or in solution (yellow) and PGA-based inhibitor with 10% peptide coupling stored dry (blue) or in solution (green) and after long-term storage at various temperatures and relative humidities are shown. Samples were stored at the indicated temperature and relative humidity for up to 6 months prior to assaying the inhibitory activity in an in vitro cytotoxicity test. IC50 is reported with respect to peptide concentration. The storage conditions were maintained by Boston Analytical, Inc.

2.2 Efficacy in vivo

2.2.1 Toxin challenge

The PGA-based polyvalent inhibitors originally reported by Joshi et al.92 were shown to be

100% effective at preventing the intoxication symptoms of anthrax lethal toxin in vivo for

23

Fisher 344 rats (Figure 2.5). In this experiment, rats were injected with doses of anthrax

lethal toxin that were immediately active. PGA-based inhibitors were injected

immediately afterward at a separate injection site.

Figure 2.5: Inhibitory activity of a PGA-based polyvalent inhibitor against anthrax lethal toxin in vivo.

Rats were injected with lethal toxin alone (filled triangles, n = 14), lethal toxin plus Ac-HTSTYWWLDGAPK-Am peptide (empty circles, n = 12), or lethal toxin plus PGA-based inhibitor (filled squares, n = 7). The log-rank test indicates a statistically significant difference between the group that received lethal toxin alone and the group that received the toxin plus the PGA-based inhibitor (P = 0.0002), but no significant difference between the group that received lethal toxin alone and the group that received the toxin along with peptide (P = 0.54). Reprinted with permission from Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Bioconjug. Chem. 2006, 17, 1265. Copyright 2006 American Chemical Society.

2.2.2 Spore challenge

Although the ability of the inhibitors to protect against anthrax toxin in an in vivo model

is important, it is not necessarily clinically relevant. In the toxin challenge, both the toxin

and the inhibitor were injected at the same time, and so the concentration of inhibitor in

the blood stream would be at its highest at the same time as the toxin concentration in

the blood stream was highest. Therefore, a more stringent spore challenge test was

designed, in order to more closely model the ability of the PGA-based inhibitors to protect

patients that have been infected with B. anthracis from the effects of the lethal and

edema toxins. In this assay, mice were inoculated with a subcutaneous injection of B.

24

anthracis Sterne strain spores, which would then develop into a full body infection. The

dose of spores was designed such that approximately 50% of the mice would survive if

the antibiotic ciprofloxacin was administered as a treatment. Although administration

with antibiotic can help by killing the bacterial infection, it does not help to alleviate the

symptoms caused by the toxin. In this way, the experiment would provide opportunity to

measure the effect of the inhibition of the toxin produced by the infection of B. anthracis

Sterne. However, because the spores needed time to germinate and spread, the time

scale of the experiment was much different than for the toxin challenge (several days

versus a few hours), and so inhibitor concentrations in the blood stream may not coincide

with the concentrations of toxin. Inhibitor and ciprofloxacin doses were administered at

2 hours, 48 hours, and 96 hours following the inoculation of spores. The ability of the

PGA-based inhibitors to improve the survivability of mice in this spore challenge

experiment is shown in Figure 2.6.

Figure 2.6: Inhibitory activity of a PGA-based polyvalent inhibitor against anthrax toxin in vivo generated by infection with B. anthracis spores.

Groups of 10 A/J mice were inoculated with 2000 spores of B. anthracis Sterne strain on day 0. Ciprofloxacin (red line), ciprofloxacin + PGA-based inhibitor (blue line), or the delivery vehicle (PBS, black line) as a negative control were administered as injections via the tail vein at 2, 48, and 96 hours post-inoculation. The experiment was performed by Biophage Pharma, Inc.

As seen in Figure 2.6, there was only a mild increase in the survivability percentage

of mice injected with the PGA-based anthrax toxin inhibitor and antibiotic over the mice

25

that only received antibiotic treatment. Statistical analysis of the results (not shown)

showed that this difference was not large enough to be statistically significant. These

results showed that there may be some problems with the current design of the

polyvalent PGA-based anthrax toxin inhibitor. The next step was to try to identify what

factors might be affecting the efficacy of the PGA-based toxin inhibitor in this experiment.

2.3 Pharmacokinetics

2.3.1 Serum stability

The first hypothesis was that there may be some protease activity that is degrading the

PGA-based inhibitor in vivo on a time scale that is slow enough that allowed it to remain

efficacious in the toxin challenge (over the course of a few hours), but fast enough to

reduce inhibitor efficacy in the spore challenge (over the course of a few days). In order

to test this hypothesis, we performed a serum incubation assay. PGA-based inhibitors

were pre-incubated in serum for up to three days prior to assaying the inhibitory efficacy

in a cell culture cytotoxicity assay (Figure 2.7). From this experiment, it was evident that

the PGA-based inhibitors retain their activity at least over the course of a few days in

biological fluids.

Figure 2.7: Inhibitory activity of PGA-based inhibitors after incubation in serum.

26

Inhibition of anthrax toxin cytotoxicity in vitro by PGA-based inhibitor with 5% peptide coupling (red line) or PGA-based inhibitor with 10% peptide coupling (blue line) after incubation in serum for up to 72 hours. IC50 is reported with respect to peptide concentration.

2.3.2 PEGylation of PGA-based inhibitors to improve blood residence time

The next hypothesis was that the PGA-based inhibitors were susceptible to rapid

clearance from the circulation. In order to test the amount of inhibitor that remained in

the blood without significantly modifying the inhibitors with fluorescent dyes or

radioactive labels, our collaborators at the University of Toronto developed an ELISA test,

in which blood samples would be withdrawn at various times and exposed to plates with

[PA63]7 adsorbed in the wells. Any inhibitor that remained in the blood sample would then

bind to the adsorbed heptamer, and the presence of inhibitor was tested by the

subsequent addition of an antibody that was generated to bind to the Ac-

HTSTYWWLDGAP-Am peptide. The binding of that antibody would then be measured by

the binding of a secondary antibody fused with HRP. The results of the ELISA assay of

inhibitor residence time in the blood are shown in Figure 2.8.

Figure 2.8: ELISA assay of PGA-based anthrax toxin inhibitor blood residence time.

PGA-based inhibitor dissolved in PBS was first administered to mice via tail vain injection, and blood samples were withdrawn at the times indicated. An ELISA was performed on the samples and compared to controls with known inhibitor concentration. This work was performed by our collaborators in the Mogridge lab at the University of Toronto.

27

As we hypothesized, a large percentage of the PGA-based inhibitor has been cleared

from the circulation within one hour post-injection. These results may help explain the

lack of success during the spore challenge experiment, because the low concentration of

inhibitor remaining in the blood (if any) past this time would probably not be providing

adequate inhibition of anthrax toxin being continually released by B. anthracis.

Since analyzing these results, we have decided to try to improve the

pharmacokinetics of the PGA-based inhibitor by further modification with PEG. The use

of PEG conjugation is a commonly used method for increasing the hydrodynamic volume,

and by extension the blood residence time, of a variety of therapeutic molecules. Because

our inhibitors are roughly 40 kDa in size, we used mono-functional methoxy-PEG-amines

of 20 kDa and 40 kDa. These PEGs react with the HOBt-activated esters of the glutamate

side chains of our activated PGA in the same way that the amine of the lysine side chain

of Ac-HTSTYWWLDGAPK-Am does. Due to the more difficult solubilization and

significantly larger size of these PEGs relative to the peptide, we expected a slower

conjugation rate of the methoxy-PEG-amines to the activated PGA. Thus we added the

PEG first and allowed 5-10 min reaction prior to addition of the peptide to the reaction

mixture. Successful PEGylation could be visualized by SDS-PAGE with barium iodide

staining to show the PEG attachment (see Figure 2.9).124,125 Polyvalent PGA-peptide

conjugates cannot typically by visualized by either Coomassie blue or silver staining due

to the highly negatively-charged polypeptide backbone, as shown by the silver-stained

image of the gel (left). PEG is also not stained by common gel-staining methods, but after

treatment with barium chloride followed by an iodine solution, PEG stains dark brown

against a light brown background.124,125 As is evident by the multilple bands on the gel,

there are multiple different species of PGA-PEG/peptide in the reaction mixture. The

mixture of these PEGylated PGA-peptide anthrax toxin inhibitors have been tested for

activity in the in vitro inhibition of cytotoxicity assay, and they retain their inhibitory

activity.

28

Figure 2.9: Silver-stain (left) followed by barium iodide-stain (right) of a 4-12% acrylamide SDS-PAGE of PGA-based anthrax toxin inhibitors conjugated with monofunctional PEG20k or PEG40k.

Lanes left to right: (1) Reaction targeting average attachment of three PEG20k chains per PGA35k scaffold with 10% peptide coupling, (2) Reaction targeting average attachment of three PEG40k chains per PGA35k scaffold with 10% peptide coupling, (3) Reaction targeting average attachment of three PEG20k chains per PGA120k scaffold with 10% peptide coupling, (4) Reaction targeting average attachment of one PEG40k chain per PGA120k scaffold with 10% peptide coupling, (5) Reaction targeting average attachment of three PEG40k chains per PGA120k scaffold with 10% peptide coupling, (6) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled. Both images are of the same gel, first after silver staining (left) and then after barium iodide staining of PEG (right).

2.4 Conclusions

The fact that the PEGylated polyvalent PGA-peptide anthrax toxin inhibitors retain their

inhibitory activity in vitro is promising. We are currently pursuing methods for purifying

the individual species as shown in the gel image on the right in Figure 2.9, because the

polydispersity of the PGA and the lack of control over the number of PEG chains

complicates characterization efforts. Unfortunately, purification by SEC does not provide

adequate resolution for separation of the various species of PEGylated PGA-based

anthrax toxin inhibitors. To address this difficulty, we have turned to a less commonly-

used method for purification, large-scale gel electrophoresis. The apparatus allows for

size-based and/or charge-based fractionation of material by directing a flow of liquid

along the bottom of the gel to a fractionator.126–128 In this way, large samples can be run

29

through the gel and analyzed based on the time of electrophoretic elution. Once this

purification method has been optimized, the well-fractionated, PEGylated, PGA-based

anthrax toxin inhibitors can be characterized by 1H NMR to confirm the average percent

coupling of PEG. In addition, the fractionated, PEGylated inhibitors can be compared by

in vitro cytotoxicity assay and ELISA to determine blood residence time. If the blood

residence time is significantly improved, these PEGylated anthrax toxin inhibitors may be

much more promising candidates for clinical application. In that case, another in vivo

spore challenge experiment similar to the one described in Section 2.2.2 should be

performed with the PEGylated anthrax toxin inhibitors.

30

3. Radial heptavalent inhibitors of anthrax toxin3

3.1 Abstract129

The design of multivalent molecules, consisting of multiple copies of a biospecific ligand

attached to a suitable scaffold, represents a promising approach to inhibit pathogens and

oligomeric microbial toxins. Despite the increasing interest in structure-based drug

design, few multivalent inhibitors based on this approach have shown efficacy in vivo.

Here we demonstrate the structure-based design of potent biospecific heptavalent

inhibitors of anthrax lethal toxin. Specifically, we illustrate the ability to design potent

multivalent ligands by matching the pattern of binding sites on the biological target. We

used a combination of experimental studies based on mutagenesis to identify the binding

site for an inhibitory peptide on the heptameric subunit of anthrax toxin. We developed

an approach based on copper-catalyzed azide-alkyne cycloaddition (click-chemistry) to

facilitate the attachment of seven copies of the inhibitory peptide to a β-cyclodextrin core

via a polyethylene glycol linker of an appropriate length. The resulting heptavalent

inhibitors neutralized anthrax lethal toxin both in vitro and in vivo and showed

appreciable stability in serum. Given the inherent biocompatibility of cyclodextrin and

polyethylene glycol, these potent well-defined heptavalent inhibitors show considerable

promise as anthrax anti-toxins.

3.2 Introduction

Anthrax toxin, responsible for the major symptoms and death associated with anthrax, is

composed of three proteins: lethal factor (LF) and edema factor (EF) are enzymes that act

in the mammalian cell cytosol; protective antigen (PA) binds to cell-surface receptors and

mediates toxin internalization.130 PA is proteolytically processed into a 63 kDa fragment,

PA63, which oligomerizes into heptamers ([PA63]7) that bind EF and LF. The assembled

toxin complex is internalized by receptor-mediated endocytosis and the enzymes are

This chapter previously appeared as: Joshi, A.; Kate, S.; Poon, V.; Mondal, D.; Boggara, M. B.; Saraph, A.; Martin, J. T.; McAlpine, R.; Day, R.; Garcia, A. E.; Mogridge, J.; Kane, R. S. Biomacromolecules 2011, 12, 791. Copyright 2011 American Chemical Society.

31

translocated into the cytosol when the toxin reaches an acidic compartment. The

enzymatic activities of these proteins cause a variety of cellular dysfunctions that

contribute to disease progression.5 Anthrax toxin is, therefore, an important therapeutic

target and inhibitors have been described that block different steps in this intoxication

pathway.10,13,14,86,89,90,92,94,96,131–135

A promising approach to designing potent ligands for oligomeric targets involves the

design of inhibitors that are multivalent, as multivalency or polyvalency can enhance the

affinity of interactions by several orders of magnitude.64,67,69 Some oligovalent inhibitors

have been designed without prior knowledge of the spatial relationships among the

binding sites on the target64,86 and others have been designed with architectures that

matched the geometry of binding epitopes on various bacterial toxins.83,84,88,89,131,136,137

The rational structure-based design of oligovalent inhibitors that provide structurally and

compositionally well-defined molecules may, however, be better suited for pre-clinical

and clinical development. There are but a few examples of the structure-based design of

oligovalent inhibitors that have shown efficacy in vivo.88,89

Here, we describe the rational structure-based design of an anthrax anti-toxin based

on the heptavalent display of a biospecific ligand. Our strategy to optimize the design of

the inhibitor was to attach peptides to a defined scaffold that would present the peptides

in a spatial orientation matching the peptide-binding sites of the toxin. For the core of the

scaffold, we considered using cyclodextrins, because they are cyclic oligomers of

glucopyranose138 that have a defined molecular weight and a given symmetry, and they

are commercially available in high purity. β-cyclodextrin was chosen because it has 7-fold

symmetry, as does [PA63]7 (Figure 3.1), and because cyclodextrins are widely used as

pharmaceutical agents to enhance the solubility, bioavailability, and stability of drug

molecules. Furthermore, cyclodextrin has been found to partially block the pore of

oligomeric staphylococcal alpha-hemolysin139 and small cationic β-cyclodextrin

derivatives can bind to and block the [PA63]7 pore via electrostatic interactions.89 To graft

the inhibitory peptides to the β-cyclodextrin core, we selected polyethylene glycol as a

linker. We chose an appropriate linker length using information on the location of the

32

peptide-binding sites on [PA63]7, obtained as described below. The well-defined

heptavalent inhibitor not only neutralizes anthrax lethal toxin in vitro, but also protect

animals from a toxin challenge.

Figure 3.1: Structure-based design of heptavalent anthrax toxin inhibitors.

(A) Structure of the LF-binding face of [PA63]7. Residues 184, 187, 197, and 200, which form part of the peptide-binding site are shown in purple. (B) Structure of the core, β-cyclodextrin. (C) Scheme illustrating the binding of a heptavalent inhibitor, synthesized by the attachment of seven inhibitory peptides to the β-cyclodextrin core via an appropriate polyethylene glycol linker, to [PA63]7

3.3 Materials and methods

3.3.1 Preparation of toxin components

Mutations in PA were generated using Quickchange mutagenesis according to the

manufacturer’s instructions (Stratagene). PA and LF were purified as described

previously.140

3.3.2 Cytotoxicity assay

RAW264.7 cells were seeded in 96-well plates and incubated overnight. The cells were

treated with 10-8 M PA and 10-9 M LF in the absence or presence of inhibitors. After an

incubation period of 4 h, cell viability was assessed using the MTS assay according to the

manufacturer’s instructions (Promega).

33

3.3.3 Serum stability

The heptavalent inhibitor was dissolved in PBS and incubated with mouse serum (Sigma,

St. Louis, MO, USA) (80% v/v) at 37 °C. Samples were withdrawn at different time intervals

and tested in a cytotoxicity assay as described above.

3.3.4 Rat intoxication

Animal experiments were carried out by our collaborator, Prof. Jeremy Mogridge at the

University of Toronto, and were performed under University of Toronto ethical

guidelines. A mixture of purified PA (40 µg) and LF (8 µg) mixed with either PBS,

heptavalent inhibitor (300 nmol on a per-peptide basis), or control thioglycerol-

functionalized heptavalent molecules, was used for the rat intoxication experiments.

Fisher 344 rats (Charles River Laboratories) were injected intravenously in the tail vein.

Seven rats were used per group, and the appearance of symptoms of intoxication was

monitored over 4 h. The rats were euthanized to avoid unnecessary distress when the

symptoms became pronounced.

3.4 Results and discussion

3.4.1 Structure-based design of heptavalent inhibitors

We developed a synthetic strategy for attaching the peptide to the β-cyclodextrin core

via a polyethylene glycol linker (Figure 3.2). Given the difficulties inherent in attaching

seven polymeric linkers to a single core with high yield, we developed an approach based

on “click chemistry” in which the copper-catalyzed 1,3-dipolar cycloaddition of an azide

to a terminal alkyne results in the production of a triazole with high yield.141–144 Briefly,

the secondary hydroxyl groups of the β-cyclodextrin core were first methylated and seven

terminal alkyne groups were introduced by the reaction of the primary hydroxyl groups

with propargyl bromide. Seven polyethylene glycol linkers were attached to the core by

the reaction of the alkynes with O-(2-Aminoethyl)-O′-(2-azidoethyl)decaethylene glycol

(H2N-CH2CH2-(OCH2CH2)11-N3, MW = 570.7 Da) (PEG11) – a commercially available,

monodisperse, azide-functionalized polyethylene glycol derivative. Chloroacetylation of

34

the free terminal amine, followed by reaction with the peptide HTSTYWWLDGAPC, a

cysteine-derivatized version of the inhibitory 12-mer peptide, yielded a heptavalent

inhibitor (6, Figure 3.2).

Figure 3.2: Synthesis scheme of heptavalent anthrax toxin inhibitor.

(a) TBDMSCl, pyridine, 0°C - rt. (b) NaH, MeI, THF (c) NH4F, MeOH, reflux (d) NaH, propargyl bromide, DMF, 0°C – room temperature. (e) CuSO4, sodium ascorbate, THF:H2O:BuOH (0.5:1:1), 80°C (f) Chloroacetic anhydride, triethylamine (g) Peptide, DMF, DBU, triethylamine.

35

The selection of the polyethylene glycol linker, PEG11, was based on the location of

the peptide-binding site. Mourez et al.86 used phage display to identify an inhibitory 12-

mer peptide, HTSTYWWLDGAP, that competes with LF for binding [PA63]7. A subsequent

screen identified a related peptide, HYTYWWLD, that also contains the TYWWLD

sequence, which we’ve shown is both necessary and sufficient for competing with LF.145

These experiments suggested that the peptides interact with [PA63]7 at or near the LF-

binding site. This information was used to guide the design of heptavalent inhibitors. The

distance between the peptide-binding residues (P184, L187, K197, and R200) and the

center of the lumen of [PA63]7 is ca. 30-40 Å. Recognition and inhibition of the toxin would

be facilitated by choosing a linker such that the root-mean-square distance from the

center of the cyclodextrin core to the end of the linker matched the distance from the

center of the lumen of [PA63]7 to the peptide-binding residues,83,85,146–148 i.e., by

statistically matching the heptavalent inhibitor with the heptavalent target.94 For our

inhibitor, the linker consisted of a rigid cyclodextrin core, a flexible region (consisting

primarily of ethylene glycol repeat units149), and the amino acids CPAG (since the

TYWWLD residues of the HTSTYWWLDGAPC are necessary and sufficient for binding145).

Using the method of Krishnamurthy et al.,148 we estimated that the root-mean-square

distance from the center of the cyclodextrin core to the end of the linker was ca. 30 Å for

inhibitors based on the PEG11 linker, which was consistent with the distance from the

center of the lumen to the peptide-binding residues on [PA63]7.

3.4.2 Characterization of heptavalent inhibitors

We tested the ability of this well-defined heptavalent inhibitor (6, Figure 3.2) to neutralize

anthrax lethal toxin in vitro by incubating RAW264.7 cells with a mixture of PA and LF in

the presence of several concentrations of the inhibitor. The heptavalent molecule could

inhibit cytotoxicity with a half-maximal inhibitory concentration (IC50) of ca. 10 nM on a

per-peptide basis (Figure 3.3A). Heptavalent molecules presenting only thioglycerol

showed no inhibitory activity (Figure 3.3A), and the monovalent peptide did not inhibit

cytotoxicity at concentrations as high as 2 mM. The heptavalent inhibitor therefore

provided a more than 100,000-fold enhancement in the activity of this peptide. To test

36

whether the well-defined heptavalent inhibitor based on the PEG11 linker was resistant to

proteolytic degradation, we also incubated the inhibitor with 80% serum at 37 ºC.

Samples were withdrawn at various time intervals and their inhibitory activity was

determined using the cytotoxicity assay. As seen in Figure 3.3B, the heptavalent inhibitor

did not show any significant loss in activity over a three day period.

Figure 3.3: Characterization of a well-defined heptavalent anthrax toxin inhibitor.

(A) Inhibition of anthrax toxin-induced cytotoxicity of RAW264.7 cells by heptavalent inhibitors presenting HTSTYWWLDGAP (open circles) or control thioglycerol-functionalized molecules (closed circles). (B) Stability of heptavalent inhibitor in serum. Heptavalent inhibitor was incubated in 80% serum for indicated times and the IC50 was determined in a RAW264.7 cell cytotoxicity assay.

Next, we decided to probe the effect of the structure of the heptavalent inhibitor on

its potency. To that end, we synthesized a series of inhibitors with polyethylene glycol

linkers of different molecular weights (and therefore, different lengths), and tested their

activity in the cytotoxicity assay (Figure 3.4). As seen in Figure 3.4, heptavalent inhibitors

based on short PEG linkers (n ranging from 2 - 6) did not show significant activity in the

cytotoxicity assay. In contrast, the inhibitory activity did not significantly change with

further increase in the number of polyethylene glycol units beyond 11 (i.e., for n > 11), in

the range of molecular weights tested (Figure 3.4). When the root-mean-square distance

from the center of the cyclodextrin to the end of the linker was greater than or equal to

the distance from the lumen of the heptamer to the peptide-binding site, effective

37

inhibition was observed, consistent with both experimental and theoretical

studies.85,146,148,150

Figure 3.4: Influence of linker length on the activity of heptavalent inhibitors.

The indicated PEG linkers were used to join the HTSTYWWLDGAP peptide to β-cyclodextrin and the IC50 values of the resulting inhibitors were measured in a RAW264.7 cell cytotoxicity assay. The error bars represent the standard deviation from four separate experiments. Asterisks indicate that inhibitory activity was not detected.

Finally, our collaborator, Prof. Jeremy Mogridge at the University of Toronto, tested

the ability of the well-defined heptavalent inhibitors to neutralize anthrax toxin in vivo, in

Fisher 344 rats. Six of seven rats that were injected intravenously with anthrax lethal toxin

(a mixture of 40 µg of PA and 8 µg of LF) and six of seven rats that were co-injected with

toxin and thioglycerol-functionalized heptavalent molecules became moribund (Table

3.1). Co-injection of the heptavalent inhibitor based on the PEG11 linker (6, Figure 3.2)

with the toxin, however, prevented six of seven animals from becoming moribund.

Table 3.1: Inhibition of anthrax toxin action in a rat intoxication model.

Inhibitor Amount of peptide (nmol) Moribund/ Total

None 0 6/7

Negative Control 0 6/7

Heptavalent inhibitor 300 1/7

38

3.5 Conclusions

In conclusion, we have demonstrated the structure-based design of potent biospecific

oligovalent inhibitors of anthrax lethal toxin. We used experimental studies to identify

the binding site for an inhibitory peptide on the heptameric subunit of anthrax toxin. We

developed a click-chemistry-based approach for the efficient attachment of a suitable

polymeric linker to a cyclodextrin core. Subsequent functionalization with the peptide

yielded a well-defined heptavalent inhibitor that neutralized anthrax lethal toxin both in

vitro and in vivo and showed appreciable stability in serum. Given the inherent

biocompatibility of cyclodextrin and polyethylene glycol, these potent well-defined

heptavalent anti-toxins might serve as valuable adjuncts to antibiotics for the treatment

of anthrax. The approach outlined in this work might also be broadly applicable to

designing well-defined oligovalent molecules that inhibit pathogens or other microbial

toxins in vivo.

39

4. Multivalent inhibitors of influenza virus

4.1 Rationale for the design of multivalent inhibitors of influenza virus entry

4.1.1 Targeting a conserved epitope on the hemagglutinin spike

The current standard influenza vaccines, which include both live attenuated influenza

vaccine (LAIV) and inactivated influenza vaccine (IIV) formulations, were only 62%

effective at preventing acute respiratory infections of influenza in 2013.151 Analysis has

shown that LAIVs are consistently more effective than IIVs, particularly in children,152 but

the rate of protection is still relatively low when compared to that of standard vaccines

against other diseases, such as measles (confers 95-99% immunity after one dose).153

Nonetheless, the Center for Disease Control reports that vaccination has been shown to

reduce the severity and repercussions of influenza-like illness, even when prevention

rates are only moderate.154

One of the primary causes of the inability of standard influenza vaccines to prevent

a greater proportion of confirmed influenza infections is the antigenic drift of the virus,155

as discussed in Section 1.2.3 Antigenic variation on page 5. Seasonal influenza vaccines

take time to prepare, and so the selection of influenza strains which are included in

vaccine formulations are not always accurate predictions of what end up being the

dominant circulating strains.156 By the time flu season actually hits, the circulating

influenza viruses may have undergone enough antigenic drift to escape the protective

effects of the neutralizing antibodies that were generated by the immune system in

response to the administered vaccinations.155 This disconnect has been a driving factor in

recent years behind the interest in broadly neutralizing antibodies. These antibodies are

characterized by cross-reactivity to a broad spectrum of influenza subtypes.155–159 In the

last five years, crystal structures of some of these antibodies have been published, and

the structures have shown that the antibodies bind to regions of the viral glycoproteins

that are not as structurally plastic and therefore not prone to the same rates of antigenic

drift.155,157,158 In the case of HA, these so-called “conserved” epitopes are typically found

on the stem or directly at the receptor binding site, as opposed to other parts of the

40

globular head region that are more typically targeted (see Figure 4.1).155 In the case of the

antibodies CR6261 and F10, the binding epitopes on the stem were conserved across

almost all group 1 HA subtypes. Interestingly, the mechanism by which stem-binding

antibodies such as these neutralize the virus is not by blocking attachment of HA to the

sialic acid receptor; rather, the antibodies appear to inhibit fusion of the viral and cellular

membranes by constraining the action of the fusion peptide and stabilizing the pre-fusion

conformation of the HA stalk.155,157,158

Figure 4.1: Combined image of HA from influenza A/South Carolina/1/1918 (SC1918/H1) with structures of several strain-specific antibodies and the broadly-neutralizing antibody CR6261.

Strain-specific Fabs BH151 (PDB code 1EO8, in green), HC63 (PDB ID 1KEN, in pink), HC45 (PDB ID 1QFU, in dark red), and HC19 (PDB ID 2VIR, in blue) are shown at their respective binding sites on the globular head domain of the HA trimer, shown as a surface representation with the HA1 and HA2 portions of one HA protomer colored in pink and cyan, respectively. In contrast, the Fab of the broadly-neutralizing IgG1 CR6261 (PDB ID 3GBN, heavy chain in yellow, light chain in orange) binds to an epitope of the HA stem. From Ekiert, D. C.; Bhabha, G.; Elsliger, M.; Friesen, R. H. E.; Jongeneelen, M.; Throsby, M.; Goudsmit, J.; Wilson, I. A. Science 2009, 324, 246. Reprinted with permission from AAAS.

41

4.1.2 De novo design of protein ligands that target a conserved epitope on the hemagglutinin spike

While influenza vaccine research has been reenergized by the discovery of the conserved

epitopes of the HA stem region and the new quest to develop antigens that stimulate

antibody generation against these conserved targets, researchers in the Baker lab at the

University of Washington took a different approach and used their expertise in modeling

protein-protein interactions to computationally design new proteins to bind the

conserved HA stem epitope.160,161

By taking into consideration features such as the average solvent-exposed area on

protein surfaces that becomes buried during a typical protein-protein interaction, the

researchers began searching the protein data bank for proteins amenable to

manipulation (lacking disulfide bonds, expressed in E. coli, predicted to form monomers)

and which had structures with a high degree of shape complementarity to the

hydrophobic groove of the target epitope.161 In parallel, three key “hotspot” interactions

were identified by allowing disembodied amino acids to probe the target epitope,

optimizing potential electrostatic interactions, hydrogen binding, and Van der Waals

interactions.161 The fact that one of the three hotspots identified in this manner

resembled the interaction between the same amino acid (Tyr) observed in the crystal

structures of both the CR6261 and F10 antibodies was an early indication of the validity

of their technique.161 After the candidate protein scaffolds were identified, an iterative

process was developed for adding two or all three hotspot interactions to the scaffold. A

total of 88 new proteins based on 79 original scaffolds were designed in this way.161 Yeast

display was used to test the designed proteins for activity, which significantly narrowed

the pool down to just two candidates which reproducibly bound the HA stem. These

candidate proteins were HB36 and HB80, with HB standing for hemagglutinin-binder.

HB36 was one of the two-hotspot-binding designs, and was derived from a protein of

unknown function identified from Bacillus stearothermophilus.161 HB80 was a three-

hotspot-binding design, derived from a transcription factor from Antirrhinum majus.161

Neither of the original protein scaffolds were able to bind HA without the modifications.

42

These scaffolds were then affinity-matured in a variety of ways, including using a

novel deep sequencing approach.160,161 The researchers reported that most of the

modifications that were made involved optimizing the electrostatic interactions between

the HB proteins and the target epitope.160 The resulting affinity-matured variants, HB36.5

and HB80.4, were both able to bind HA with dissociation constants of ~1 nM, and the

crystal structures showed that the HB proteins bound in nearly the exact conformation in

which they had been designed to bind (see Figure 4.2). Furthermore, HB80.4 was shown

in an in vitro microneutralization assay to inhibit infection of Madin-Darby Canine-Kidney

(MDCK) cells by two different H1N1 influenza strains with half maximal effective

concentrations in the 100-200 nM range.160 This efficacy was comparable to that of the

original CR6261 IgG1 (IC50 = ~120 nM).159,160

Figure 4.2: Comparison of crystal structure of the broadly-neutralizing Fab CR6261 in complex with HA to the crystal structures of the designed hemagglutinin-binding proteins in complex with HA.

All three structures are in complex with SC1918/H1 displayed in approximately the same orientation. The crystal structure on the left is of the Fab CR6261 (PDB ID 3GBN, Fab heavy chain in yellow, Fab light chain in orange, HA trimer in grey, with one HA1 chain in purple and the corresponding HA2 chain in cyan), adapted from Ekiert, D. C.; Bhabha, G.; Elsliger, M.; Friesen, R. H. E.; Jongeneelen, M.; Throsby, M.; Goudsmit, J.; Wilson, I. A. Science 2009, 324, 246. Reprinted with permission from AAAS. The crystal structure in the middle is of HB36 (PDB ID 1U86, HB computational design in blue, actual HB structure in red, HA1 chain in pink, HA2 chain

43

in light blue), adapted from Fleishman, S. J.; Whitehead, T. A.; Ekiert, D. C.; Dreyfus, C.; Corn, J. E.; Strauch, E.-M.; Wilson, I. A.; Baker, D. Science. 2011, 332, 816. Reprinted with permission from AAAS. The crystal structure on the right is of HB80 (PDB ID 4EEF, HB computational design in green, actual HB structure in orange, HA1 in gold, HA2 in light blue). Adapted by permission from Macmillan Publishers Ltd: Nat. Biotechnol. 2012, 30, 543, copyright 2012.

4.1.3 Design of divalent inhibitors that target a conserved epitope on the hemagglutinin spike

Given our experience in designing multivalent conjugates which improve binding by

orders of magnitude, we were naturally interested in seeing if we could further enhance

the efficacy of the Baker lab’s computationally-designed monovalent HB proteins by

displaying them on appropriate multivalent scaffolds. The fact that HA is itself a trimer of

HA protomers and that these trimers are expressed polyvalently on the surface of

influenza virions made these HB proteins seem to us to be especially applicable for the

design of multivalent influenza inhibitors. While we could envision multiple different

strategies for creating multivalent conjugates that are effective at spanning multiple HB

protein binding sites, we decided to start with the most simple design first. That is, we

decided to create homodivalent conjugates by using bifunctional PEG linkers. Several

different chemistries can be used for creating homodivalent scaffolds. Commercially

available terminal functionalities include amine-reactive activated esters such as NHS,

TFP, and PFP, thiol-reactive maleimides, alkyne-reactive azides, and primary amines and

hydroxyls. Furthermore, the primary amines or terminal hydroxyl groups can be

chloroacetylated to introduce another form of thiol-reactivity, as described in Chapter 3

for the synthesis of β-cyclodextrin-based anthrax inhibitors.

In order to decide which PEG length might be appropriate for this application, I

consulted a thermodynamic analysis on the use of a flexible linker such as PEG in

divalently-binding constructs that was recently published by Kane.146 Although the

analysis specifically addresses the slightly more complex case of a heterodivalent binder,

the findings can easily be applied to homodivalent binders such as those we were

interested in designing. Consider a system consisting of a heterodivalent ligand composed

of two separate ligands, A and B, connected by a flexible linker, and interacting with a

receptor with a separate binding site specific to each ligand. The binding avidity of such a

44

heterodivalent ligand can be optimized by choosing a linker of such a length that will

maximize the probability that the two ligands are separated by an appropriate distance –

the distance, 𝑟, between the two ligand binding sites on the receptor. The derived

expression for the free energy of heterodivalent binding, ∆𝐺div0 , is given by the equation

∆𝐺div0 = ∆𝐺mono, A

0 + ∆𝐺mono, B0 − 𝑅𝑇 𝑙𝑛[𝐶eff(𝑟)], where the free energy of binding for

monovalent ligands A and B is represented by ∆𝐺mono, A0 and ∆𝐺mono, B

0 , respectively. The

term 𝐶eff(𝑟) represents an “effective concentration,” which is a function of 𝑟, and is

proportional to the probability that the two ends of the linker are a distance, 𝑟, apart.146

The behavior of a highly flexible linker in a good solvent approaches a random walk;

therefore the most probable distance separating the ends of the linker (and thus the two

ligands) is proportional to the root-mean-square (RMS) of the end-to-end distance of the

fully extended linker. Accordingly, the magnitude of 𝐶eff(𝑟) is maximized when the RMS

end-to-end distance of the linker matches the distance separating the two target binding

sites. This analysis applies for homodivalent linkers and the linear backbone between

ligands on a linear, polyvalent scaffold, and therefore the conclusions can be used to guide

the designed spacing between ligands for all types of one-dimensional multivalent

scaffolds.

To apply this analysis for designing multivalent displays of HB proteins, I first needed

to estimate the distance between binding sites. By using the PDB structure published by

the Baker group160 of HB80.4 in complex with SC1918/H1, I estimated that 6 nm was

approximately the distance that a linker would need to span in order to enable

simultaneous binding to two of the three epitopes. I calculated this distance as the arc

length between two of the binding sites, formed by a circle circumscribing the HA trimer

in the plane formed by the three binding epitopes (see Figure 4.3). Therefore, according

to the thermodynamic analysis discussed above, PEG linkers with RMS lengths in solution

that are in that range would maximize 𝐶eff and thus provide the greatest enhancement to

avidity. Discrete lengths of oligo(ethylene glycol) linkers (i.e. monodisperse PEG) are

commercially available (Quanta Biodesign) in a range of sizes from 4 to 29 ethylene glycol

repeats ((EG)n, n = 4-29), which correspond to extended lengths of 1.4 – 101.5 nm.

45

According to published correlations148,149 between the number of EG repeats (i.e. the

value of n) and the RMS length of PEG, I estimated the RMS lengths of these potential

linkers to be 1.2 – 3.1 nm. Longer homobifunctional PEGs are also available (Jenkem USA),

with MWs of 2.0, 3.5, 5.0, and 7.5 kDa, corresponding to estimated RMS lengths of

approximately 4, 5, 6, and 7.5 nm, respectively. It is noteworthy that the change in RMS

end-to-end distance of a flexible polymer like PEG is much more significant when

comparing linkers with smaller extended end-to-end distances. That is, as the distance

between target binding sites increases, the need to precisely match the RMS end-to-end

distance of the linker becomes less critical.

Figure 4.3: Estimated distance between HB protein binding sites on the HA trimer stalk.

The distance for an appropriately-sized flexible linker to span was estimated by calculating the arc length based on a circle circumscribing the HA trimer stalk in the plane of a cross-section of the trimer as shown, from the PDB ID 4EEF. Adapted by permission from Macmillan Publishers Ltd: Nat. Biotechnol. 2012, 30, 543, copyright 2012.

We decided to test three different PEG linkers differing in estimated RMS end-to-end

distance by ~2.5 nm for each size, as shown in Table 4.1. The smallest of the linkers has

an extended length that is only slightly larger than the predicted required minimum

distance, so we expected the 𝐶eff for this linker to be approximately an order of magnitude

46

less than 𝐶eff for each of the other two linkers. From the equation 𝐾𝑑𝑖𝑣 = 𝐾𝐴𝐾𝐵𝐶eff , 𝐶eff

is directly proportional to the resulting association constant of binding. Since 𝐾𝐴 = 𝐾𝐵 for

the case of a homodivalent linker, we expected the measurement of inhibitory efficacy

for our inhibitors to primarily be a reflection of the 𝐶eff difference.

Table 4.1: Homobifunctional PEG linkers purchased for homodivalent conjugation of hemagglutinin-binding proteins.

PEG-(maleimide)2

code Commercial Source P(EG)n

Approx.

MW (Da)

Extended

length (nm)

RMS length

(nm)

PEG1.1k Quanta Biodesign 25 1100 8.8 2.9

PEG3.5k Jenkem USA 79 3500 27.7 5.2

PEG7.5k Jenkem USA 170 7500 59.5 7.6

4.1.4 Design of polyvalent inhibitors that target a conserved epitope on the hemagglutinin spike

While we were interested in homodivalent PEG-based because of the specific geometry

of the target hemagglutinin trimer, we reasoned that polyvalent polymer-based HB

protein conjugates might also be useful for binding to multiple hemagglutinin spikes on a

single or multiple influenza virions.

4.2 Production of multivalent bioconjugates of hemagglutinin-binding proteins

4.2.1 Modification of hemagglutinin-binding proteins from previous studies

Through a collaboration with the Baker lab at the University of Washington, we received

a set of five plasmids encoding hemagglutinin-binding proteins. The coding sequences for

the proteins HB36.5 and HB80.4, as described in the previous section, were shared with

us in modified pET29 plasmids. In addition, three variants of HB36.5 and one variant of

HB80.4 were included. Each of the variants encodes a version of the original protein in

which one of the codons for an amino acid distal to the binding face of the protein was

mutated to a cysteine codon. These HB variants thus featured a chemical handle by which

they might be attached to a scaffold. The three cysteine variants of HB36.5 were named

47

HB36.5_Q265C, HB36.5_A276C, and HB36.5_A341C, and the cysteine variant of HB80.4

was named HB80.4_K315C. The nomenclature for the mutations was based on the

residue number of the original protein scaffold that was the basis for the HB protein

design, and should not be confused with the residue number of the final HB proteins, as

the HB36.5 and HB80.4 constructs provided to us only coded for 115 and 72 amino acids,

respectively.

After some initial work using these constructs as provided, we decided that some

further modifications would be beneficial for our use. The coding sequence of the HB

protein constructs we received from the Baker lab featured the following general design:

start codon – FLAG tag – start codon – HB protein sequence (with or without cysteine) –

polyhistidine tag. The FLAG tag is an eight-residue polypeptide tag which was

trademarked by Sigma-Aldrich Co., LLC and the sequence (DYKDDDDK) was specifically

designed to be a hydrophilic, antigenic, cleavable fusion peptide for use in protein

purification and detection.162 The Baker lab reported adding the tag to their HB protein

designs because it enhanced expression during yeast display and improved protein

solubility. However, the researchers did not remove the original start codon, and I found

that the presence of this additional start codon consistently resulted in the co-production

of a contaminating amount of non-FLAG-tagged proteins, as shown in Figure 4.4. While

there is no indication that the inhibition efficacy of the non-FLAG-tagged proteins is

significantly different from the full-length FLAG-tagged proteins, these proteins could not

be removed by SEC, and interfered with characterization and purification of downstream

reaction products. Therefore, we decided to engineer the HB constructs by removing the

original start codon.

48

Figure 4.4: SDS-PAGE of IMAC fractions from FLAG-HB36.5 purification revealing non-FLAG-tagged HB36.5 co-production.

Lanes left to right: (1) 0.1 µL of lysate from 1 L culture of E. coli Rosetta™ 2 (DE3) transformed with FLAG-HB36.5-coding pET29 vector, (2-6) 5.0 µL each of 1st to 5th column volumes 100 mM imidazole in HBS wash, (7-8) 5.0 µL each of 1st and 2nd column volumes 400 mM imidazole in HBS elution, (9) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (10-12) 5.0 µL each of 3rd to 5th column volumes 400 mM imidazole in HBS elution. Expected molecular mass of FLAG-HB36.5 is 13.3 kDa, expected molecular mass of non-FLAG-tagged HB36.5 is 11.9 kDa. Non-FLAG-tagged HB36.5 contamination can be seen as the lowest band in lanes 7, 8, 10, and 11.

The problem of co-producing contaminants was observed for both HB36.5 and

HB80.4 protein sets, but there were additional coinciding difficulties with maintaining the

solubility of the HB36.5 proteins in particular. Precipitation was observed at pH 7.4 in PBS,

but was especially evident at pH 6.5 in MES buffer, which was the condition suggested by

the protocol for the thioether conjugation reaction of protein thiols to maleimides which

I was using as a guide.163–165 After doing some analysis of the HB36.5 sequences with an

online “protein calculator” tool,166 I determined that the isoelectric point of the non-

FLAG-tagged HB36.5 protein was predicted to be at pH 6.7, while the isoelectric point of

the FLAG-HB36.5 proteins was predicted to be at pH 6.3 (see Figure 4.5). Since the

49

protocol for reactions of protein thiols with maleimides called for the use of pH 6.5 buffer

(see Section 4.2.3 Preparation of hemagglutinin-binding proteins for thioether

bioconjugation reactions), the net charge on the HB36.5 protein variants was likely at or

near neutral. It is well-known that proteins are most prone to precipitation at their

isoelectric points due to the decrease in electrostatic repulsion forces between the

proteins in solution. Inspired by the work of Rathnayaka et al.,167 we sought to add an

additional solubility-enhancing peptide of the form DDDG to the N-terminus of the FLAG-

tagged HB36.5 proteins in order to reduce the isoelectric point of the HB36.5 protein

variants to pH ~5.5 and add an additional three negative charges to the proteins at pH

6.5. We hoped that doing so would enable the HB36.5 protein variants to remain soluble

during the desired reaction conditions.

Figure 4.5: Predicted net protein charge at various pH values of three HB36.5 variants.

4.2.2 Cloning and bacterial production of modified hemagglutinin-binding proteins

The modifications described in the previous section were performed by Dr. Manish Arha

on the HB36.5 protein variants as follows. The DNA sequence present immediately after

50

the N-term FLAG tag, which coded for “His - Met” in the coding sequences of HB36.5,

HB36.5_A276C and HB36.5_Q265C, and “Ile - Met” in the coding sequence of

HB36.5_A341C, was removed by PCR amplifying the respective coding sequences from

their original plasmid. The PCR amplification also incorporated a “DDDG” tag at the 5′ end

of each of the coding sequences (after the start codon). Additionally, a new cysteine

variant of HB36.5 was created by adding a cysteine to the N-terminus of the solubility-

enhancement tag, yielding a “CDDDG” tag. For CDDDG-FLAG-HB36.5_ΔHM, the forward

primer for these modifications was (5′-CAT ATG TGT GAC GAC GAT GGA GAT TAC AAG

GAT GAC GAC GAT AAA GGA TCC TCT AAC GCA ATG GAC GGT CAA-3′) and the reverse

primer was (5′-CTC GAG TCA GTG GTG GTG GTG GTG GTG GGA TCC-3′). For DDDG-FLAG-

HB36.5_Q265C_ΔHM, DDDG-FLAG-HB36.5_A276C_ΔHM, and DDDG-FLAG-

HB36.5_A341C_ΔIM, the forward primer used was (5′-CAT ATG GAC GAC GAT GGA GAT

TAC AAG GAT GAC GAC GAT AAA GGA TCC TCT AAC GCA ATG GAC GGT CAA-3′) and the

reverse primer was (5′-CTC GAG TCA GTG GTG GTG GTG GTG GTG TTC GAG-3′). Similarly,

to remove the DNA sequence coding for the amino acids “His - Met”, which are present

immediately after the N-term FLAG tag of HB80.4 and HB80.4_K315C, the coding

sequences were PCR amplified from their respective plasmids using the forward primer

(5′-CAT ATG GAT TAC AAG GAT GAC GAC GAT AAA GGA TCC GCA TCC ACC AGA GGT TCC

GGT AGA-3′) and the reverse primer (5′-CTC GAG TCA GTG GTG GTG GTG GTG GTG TTC

GAG-3′). Each PCR-amplified sequence was then ligated into the NdeI and XhoI restriction

sites of the pET28b plasmid (modified). The pET28b vectors also contained a gene for

kanamycin resistance. I was then able to use the resulting vectors to transform the E. coli

Rosetta™2(DE3) strain (Novagen), which contains a chromosomal copy of T7 RNA

polymerase under the control of the lacUV5 promoter.

To produce the HB36.5 and HB80.4 proteins, the transformants were inoculated from

starter cultures at 1:100 dilution into multiple baffled Fernbach flasks each containing 1 L

of auto-induction media supplemented with 100 µg/mL kanamycin, as described by

Studier.168 After 6 hours of incubation at 37°C followed by 24 hours at 18°C and a constant

215 RPM, the cells were pelleted by centrifugation at 4°C for 30 minutes, and then

51

resuspended with a homogenizer into lysis buffer (pH 8.5, 100 mM HEPES, 100 mM NaCl,

25 mM imidazole, 1 mM TCEP, 1 mM PMSF, 2 mg/mL lysozyme, 0.2 mg/mL DNase) and

lysed by passing through a microfluidizer twice. The resulting lysate was clarified of

insoluble material by centrifugation at 4°C and 25,000 g for 30 minutes, and finally filtered

through a 0.45 µm filter. A disposable plastic column was packed with ~2.5 mL Ni-NTA

agarose resin for each liter of bacterial culture, and the clarified, filtered lysate was passed

through by gravity flow. The His-tagged HB protein variants exhibited strong affinity for

the immobilized metal resin, and were cleared of most contaminants by washes with a

step gradient of imidazole (25 mM, 50 mM, 100 mM for 5 column volumes each) in pH

8.5 HBS. The HB variants were then eluted from the columns with up to 10 column

volumes of 400 mM imidazole in pH 8.5 HBS. Reducing SDS-PAGE with silver staining was

performed on the IMAC fractions to confirm the successful expression and elimination of

the majority of contaminating native bacterial proteins, as shown in Figure 4.6. The results

of the gels were used to guide the choice of fractions to be pooled and concentrated by

3 kDa MWCO centrifugal filtration. These were subsequently purified by SEC on a

prepacked GE Tricorn 10/300 GL Superdex75 column in pH 7.5 MBS with TCEP and EDTA

for preventing the formation of disulfide bonds between proteins. The chromatogram of

UV absorbance at 280 nm featured two major peaks, even for non-cysteine-containing

variants of HB36.5, as shown in Figure 4.7. As reported by the Baker research group, the

HB36.5 variants exhibited a tendency to associate non-covalently as dimers during SEC,161

and this was confirmed by running the apparent dimer peak on non-reducing SDS-PAGE.

The fractions corresponding to the pure HB protein, as revealed by non-reducing SDS-

PAGE, were pooled and sterilized by vacuum filtration through 0.2 µm filters. Finally, the

protein solutions were concentrated by 3 kDa MWCO centrifugal filtration and stored at

4°C until further use in various assays and conjugation reactions.

52

Figure 4.6: SDS-PAGE of IMAC fractions from DDDG-FLAG-HB36.5_A341C_ΔHM purification.

Lanes left to right: (1) 5.0 µL of 5th column volume 100 mM imidazole in HBS wash, (2-5) 5.0 µL each of 1st to 4th column volumes 400 mM imidazole in HBS elution, (6) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (7-12) 5.0 µL each of 5th to 10th column volumes 400 mM imidazole in HBS elution. Expected molecular mass of DDDG-FLAG-HB36.5_A341C_ΔHM is 13.2 kDa.

Figure 4.7: UV chromatogram of DDDG-FLAG-HB36.5_ΔHM on a GE HiLoad 16/600 Superdex 200 SEC column.

SEC of a non-cysteine variant of HB36.5 shows the formation of dimers in solution, eluting between 60 and 80 mL, with monomers eluting between 80 and 100 mL.

53

4.2.3 Preparation of hemagglutinin-binding proteins for thioether bioconjugation reactions

I have optimized a protocol for synthesizing multivalent conjugates of any biomolecule

containing a single reactive thiol functionality with a scaffold presenting maleimide

functionalities. The important considerations for this type of conjugation reaction are

keeping the thiols from forming disulfide side products and preventing the hydrolysis of

the reactive maleimide during the conjugation reaction. First, the biomolecules of interest

are prepared for the reaction by reduction of disulfide-bonded biomolecules in the stock

solution to free thiols. This can be accomplished by incubating the biomolecules with a

high concentration of a reducing agent for at least 30 minutes. The use of dithiothreitol

(DTT), β-mercaptoethanol (ME), and other thiol-containing reducing agents is to be

avoided because these small-molecule thiols will directly compete with the biomolecule-

thiol during the subsequent conjugation reaction. Tris(2-carboxyethyl)phosphine (TCEP)

and tris(hydroxypropyl)phosphine (THP) are trialkylphosphine reducing agents that do

not contain thiols and that catalyze the conversion of disulfides into thiols by nucleophilic

attack of the phosphine towards the disulfide.169–172 Trialkylphosphines are not stable in

phosphate buffer, so MBS should be used.169

Once all disulfide bonds have been reduced to free thiols, it is important to prevent

the oxidation reaction that converts the thiols back into disulfides. The degassing of

dissolved oxygen from the buffer is an easily-overlooked consideration, but it is an

important step for that purpose.171 The easiest method for degassing oxygen is simply to

bubble high purity nitrogen or argon gas through the buffer for at least 30 minutes

immediately prior to its use.171 The pH of the buffer should be slightly acidic (~pH 6.5) to

keep the thiols protonated, which also helps prevent the reformation of disulfides.171

Trace metal ions may be another catalytic source of thiol oxidation,171,173 therefore the

MBS should also contain 5-10 mM EDTA in order to chelate any such metal ions. However,

the presence of chelating agents has also been shown to catalyze the oxidation of

trialkylphosphine reducing agents,170,171 so the combination of TCEP or THP with EDTA

should be avoided.

54

Although TCEP is frequently reported as being compatible with conjugation reactions

of thiols,170 I have found in multiple situations that concentrations of TCEP greater than 1

mM actually inhibit reactions with thiols. Therefore I remove the excess TCEP prior to the

biomolecule conjugation reaction by passing the reduced biomolecule stock through a

desalting column which has been equilibrated with buffer that contains 10 mM EDTA.

I have found that biomolecules prepared in this manner will react to the fullest extent

possible in a matter of minutes when combined at sufficient concentrations with the

chosen maleimide-activated scaffold in degassed pH 6.5 MBS with EDTA. If the

reformation of disulfide-bonded side products appears to be a significant limitation on

the reaction success, as determined by non-reducing SDS-PAGE, I recommend spiking the

reaction with a small amount (50 – 500 µM) of trialkylphosphine reducing agent after ~1

hour.

4.2.4 Synthesis of divalent hemagglutinin-binding protein bioconjugates

The synthesis scheme for homodivalent PEG-HB protein conjugates is shown in Figure 4.8.

PEG linkers used for homodivalent inhibitor synthesis were named according to their

molecular masses: PEG1.1k, PEG3.5k, and PEG7.5k (Table 4.1). Purified HB36.5 variants

CDDDG-FLAG-HB36.5_ΔHM and DDDG-FLAG-HB36.5_A341C_ΔHM, which for simplicity

will be referred to as HBA and HBB, respectively, from this point forward, were prepared

as described above for reactions of biomolecule-thiols with scaffold-maleimides. The

reduced HB proteins were then added to the scaffold at a 10% molar excess relative to

the linker maleimides (i.e. 220 mol% relative to the linker), in order to promote

attachment at both ends of the PEG. The reaction buffer was degassed pH 6.5 MBS with

EDTA, and the concentration of HB proteins was maintained as high as possible, typically

in the range of 300-500 µM. The reaction was left at ambient temperature for a few hours

before checking the conjugation efficiency by non-reducing SDS-PAGE. If a significant

amount of disulfide-bonded dimer appeared in the PEG reaction, the reaction was spiked

with 500 µM TCEP (pH 6.5) and left for an additional few hours. Under these conditions,

the homodivalent products were produced in significant quantity as characterized by non-

reducing SDS-PAGE (Figure 4.11 lanes 1-3 and 16-18) and subsequent SEC purification.

55

Figure 4.8: Synthesis scheme for homodivalent PEG-protein bioconjugation.

4.2.5 Synthesis of polyvalent hemagglutinin-binding protein bioconjugates

4.2.5.1 Activation of biocompatible polymer scaffolds by reaction with aminoethyl-maleimide

The synthesis of polyvalent protein conjugates is more complicated than the PGA-peptide

conjugation methods described in Chapters 1 and 2 for multiple reasons. First, the use of

organic solvents is not conducive to the attachment of proteins due to the denaturing

effects they have on protein folding. While some proteins may exhibit accurate refolding

behavior when returned to an aqueous buffer, this is not a given, and the refolding

process could be further complicated by the proximity of the scaffold and other nearby

unfolded proteins post-conjugation. Furthermore, proteins feature many free amines and

carboxylates on their surfaces and at their termini, which complicates the strategy of

using activated ester chemistry. Successful polyvalent conjugation requires the

attachment of multiple large biomolecules (each on the order of tens to hundreds of

kiloDaltons) to a single scaffold, which can be inherently difficult from a steric viewpoint.

Prima facie, using the highest possible concentrations of each of the components is the

most logical way to address this issue, but that approach may be difficult in and of itself

due to issues with high viscosity or precipitation.

In order to activate the biocompatible polymeric PGA and HyA scaffolds, I combined

and modified previously-published protocols for the polyvalent attachment of amido-PEG

to PGA174 or sonic hedgehog protein to HyA via thio-ether coupling.163 The synthesis

scheme is shown in Figure 4.9. Using degassed pH 6.5 MBS with 10 mM EDTA as a buffer,

I prepared stocks of various MW PGA at concentrations of 13.2 mg/mL (100 mM

glutamate monomer) or 3 mg/mL HyA (8 mM disaccharide monomer). I then added 4-

(4,6-Dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMTMM) up to 150

56

mol% relative to glutamate monomer or up to 1000 mol% relative to disaccharide

monomer, followed by up to 33% mol% AEM relative to glutamate monomer or up to 500

mol% relative to disaccharide monomer. DMTMM is a water-soluble amide-bond

coupling reagent that creates activated esters in situ,175,176 similar to the more commonly-

used EDC-NHS coupling chemistry. While a more basic pH might enhance the rate of

nucleophilic attack of the AEM amine on the activated esters, maleimides are known to

be prone to high rates of hydrolysis at high pH, so the reaction was performed at pH

6.5.177–180 The activated polymer scaffolds were either dialyzed extensively against water

for subsequent lyophilization and characterization by NMR (Figure 4.10), or dialyzed

extensively against pH 6.5 MBS with 10 mM EDTA. In the latter case, the concentration of

AEM in the sample was measured by UV absorbance at 290 nm in order to estimate the

percent maleimide coupling to the scaffold.

Figure 4.9: Synthesis schemes for polyvalent polymer-maleimide-protein bioconjugation.

(a) DMTMM, AEM, H2O; (b) dialysis; (c) pH 6.5 MBS, EDTA, thiol-biomacromolecules; (d) 100 mM MoO4 1-2 h. The polyvalent conjugate is a random copolymer composed of unconjugated monomers (x repeat units) and protein bioconjugates (y repeat units) (omitted from HyA scheme for simplicity).

57

Figure 4.10: NMR spectra of three different PGA samples functionalized with increasing percentages of maleimide.

Peaks with integration highlighted in green correspond to the two aromatic protons of the maleimide. Analysis shows the red spectrum corresponds to 2.3% coupling, the black spectrum corresponds to 15.0% coupling, and the blue spectrum corresponds to 37.5% coupling.

4.2.5.2 Thioether bioconjugation with scaffolds presenting reactive maleimides

For proof of concept, I first synthesized polyvalent PGA-based anthrax toxin inhibitors by

using PGA activated with maleimide and the Ac-HTSTYWWLDGAPC-Am [PA63]7-binding

peptide used in the cyclodextrin-based anthrax toxin inhibitors. Prior to peptide

conjugation, the percentage of PGA monomers featuring reactive maleimide

functionalities on their side chains was 10%, as analyzed by 1H NMR in D2O. Peptide

conjugation was performed in 50% DMF to dissolve and maintain the solubility of the

peptide. The reaction was left overnight, and the following morning, both unconjugated

and peptide-conjugated maleimide rings were hydrolyzed by dialyzing in the presence of

58

100 mM molybdate oxyanion.180 Hydrolysis of the maleimide ring prevents further

reaction of both unconjugated and biomolecule-conjugated maleimides with small-

molecule thiols such as free cysteine or glutathione.178,179,181,182 Analysis of the 1H NMR

spectra of polyvalent peptide-based anthrax toxin inhibitors synthesized by this method

showed attachment in the expected range (5% or 10% of PGA monomers featured

peptide conjugation), and these conjugates were as effective at inhibiting cytotoxicity in

an in vitro lethal toxin assay as inhibitors synthesized using the previously-established

approach with the same peptide coupling percentages (IC50 ~50 nM).

The synthesis of polyvalent conjugates of HBA and HBB with PGA60k, PGA120k, and HyA

scaffolds was performed under similar conditions to the syntheses of homodivalent PEG-

HB protein conjugates in Section 4.2.4. A bioengineered protein polymer, named

[(SE)10K]10W3H10 to convey the details of its repetitive amino acid sequence, was used as

an additional scaffold as well. This bioengineered polypeptide is described in further

detail in Section 7.3.2 Bioengineering scaffolds. It features 11 reactive amines (10 lysines

and the N-terminus) which were first activated by reaction with a 500 mol% excess (with

respect to amine) amount of the bifunctional crosslinking agent SMCC (Thermo Scientific

Pierce). This crosslinker is a short (8.3 Å) molecule with an amine-reactive sulfo-NHS

activated ester at one end and a thiol-reactive maleimide at the other end. A silver-

stained non-reducing SDS-PAGE of the reaction mixtures after overnight incubation is

shown in Figure 4.11. The observed formation of discrete bands in the gel in the size range

of 30 – 200 kDa (by comparison to the protein standard ladder) is typical for all kinds of

polyvalent conjugation reactions involving the attachment of large molecules to PGA. For

further proof of this phenomenon, compare the multivalent conjugation of large,

monofunctional PEGs to PGA in Chapter 2 (Figure 2.9 on page 28), the similar polyvalent

conjugation of proteins to PGA in Chapter 5 (Figure 5.3 on page 73 and Figure 5.4 on page

75), and the polyvalent conjugation of ssDNA aptamers to PGA in Chapter 6 (Figure 6.1

on page 83 and Figure 6.2 on page 84). In my experience, such results are typically

observed by SDS-PAGE even when the samples are loaded on the gel only 1 hour after the

initial reaction setup. It should be noted that although lane 5 appears at first glance to

59

have been an unsuccessful polyvalent conjugation of HBA, in fact, the HyA-HBA conjugate

is too large to enter the gel (assuming the target of 13% percent of HyA disaccharide

monomers were conjugated with HBA, the resulting polyvalent conjugate MW would be

~9,100 kDa). Upon close inspection, there is a bit of staining at the very top of lane 5,

which might actually be due to some HyA-HBA conjugate being stuck to the bottom of the

well.

Figure 4.11: Silver-stained non-reducing SDS-PAGE of conjugation reactions of HBA (lanes 1-12) and HBB (lanes 13-25) before purification, 5.0 µg total HB protein per lane.

Lanes left to right: (1-3) Reactions of HBA with PEG1.1k, PEG3.5k, or PEG7.5k (expected homodivalent conjugate MWs = 28.2 kDa, 37.8 kDa, or 41.6 kDa), (4) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (5) Reaction targeting 13% coupling of HyA monomers with HBA, (6) Reaction targeting coupling 11 copies of HBA to a bioengineered scaffold named [(SE)10K]10W3H10, (7-9) Reactions targeting 2%, 5%, or 13% coupling of PGA60k monomers with HBA, (10-12) Reactions targeting 2%, 5%, or 13% coupling of PGA120k monomers with HBA, (13) HBB-AEM (HBB with the reactive thiol quenched by reaction with excess AEM to prevent disulfide bond formation – expected MW = 13.4 kDa), (14) HBB in the presence of 5 mM DHA (attempt to force disulfide bond formation – expected MW = 26.5 kDa), (15) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (17-19) Reactions of HBB with PEG1.1k, PEG3.5k, or PEG7.5k (expected homodivalent conjugate MWs = 27.6 kDa, 37.1 kDa, or 41.0 kDa), (19-21) Reactions targeting 2%, 5%, or 13% coupling of PGA60k monomers with HBB, (22-24) Reactions targeting 2%, 5%, or 13% coupling of PGA120k monomers with HBB, (25) Reaction targeting coupling of 11 copies of HBB to a bioengineered scaffold named [(SE)10K]10W3H10. Molecular mass of unconjugated HBA is 13.6 kDa and unconjugated HBB is 13.2 kDa.

4.2.6 Purification of multivalent hemagglutinin-binding protein bioconjugates

After the conjugation has been deemed complete, any remaining unreacted maleimide

functionality can be easily quenched by reaction with excess ME or thioglycerol. However,

60

due to the reported susceptibility of maleimide conjugates to exchange with small-

molecule thiols,178,179,182 I prefer to hydrolyze the maleimide rings with molybdate.

Therefore, after overnight incubation at ambient temperature, all HB-protein thioether

conjugation reaction products were treated with 100 mM molybdate and 20 mM TCEP in

MBS for 1-2 hours to hydrolyze the maleimide ring and reduce any disulfide-bonded HB

proteins that formed during the reaction. By hydrolyzing the maleimide, further

conjugation was quenched.

As can be seen by SDS-PAGE of the reaction product, there is often a significant

amount of unreacted biomolecule leftover in the sample (Figure 4.11, lanes 1-12 and 16-

25). In order to most accurately determine the effect of polyvalent attachment of the

biomolecule of interest, the remaining unconjugated biomolecule must be removed. If

the MW of the scaffold is sufficiently larger than the MW of the conjugated biomolecule

(e.g. for HyA conjugates – the HyA backbone MW is ~1,500 kDa), the easiest purification

strategy is definitely centrifugal filtration. To facilitate removal of the unconjugated

biomolecule, the largest possible MWCO pore size filter should be used that will not

permit loss of the conjugation product to the filtrate. Therefore, the HyA-based

polyvalent HB protein conjugates were purified by centrifugal filtration on 100 kDa

MWCO filters (Millipore) by using ~50-100 times the original reaction volume of pH 7.4

PBS for three times each.

SEC is another size-based method for separating the conjugation products from

unconjugated reagent biomolecule. Less volume of eluent is required to elute larger

molecules from the SEC column due to the fact that their large size “excludes” them from

accessing much of the column volume. Smaller molecules are able to diffuse into the

pores of the SEC resin and access a larger percentage of the column volume, and therefore

they require a larger volume of eluent to be eluted from the column. Prior to SEC

purification, the bioconjugation reactions were spiked with a high concentration of TCEP

to again ensure the elimination of disulfide-bonded dimers. PGA-based polyvalent

conjugates and PEG-based homodivalent conjugates were then purified by SEC on a 24

mL Superdex75 column with pH 7.4 PBS as the eluent. After the appropriate SEC fractions

61

were pooled, the samples were concentrated so that an accurate HB concentration could

be detected with a NanoDrop UV-vis spectrophotometer. Each sample was then stored

at 4°C until further testing.

4.3 Characterization of multivalent bioconjugates of hemagglutinin-binding proteins

4.3.1 In vitro microneutralization assay of influenza virus infection inhibition

In order to assess the ability of the modified HB proteins and conjugates to inhibit

influenza infection, an in vitro assay was adapted from established protocols.183,184 First,

serial dilutions of HB protein-containing samples were prepared over a range of

concentrations of several orders of magnitude; in a typical experiment, half-log dilutions

of the samples were prepared in duplicate in 50 µL PBS per well in 96-well microtiter

plates. These dilutions were then incubated at 37°C for 1 h with a fixed amount of virus,

typically 25 times the TCID50 (50% tissue culture infective dose) in 50 µL culture medium

(DMEM + 10% FBS), so that the final concentration of HB protein spanned the range from

10 µM down to 100 pM. The virus strain tested was H1N1 A/CA/4/2009 (Virapur, CA), a

pandemic strain isolated from a child patient in San Diego, CA during the H1N1 outbreak

of 2009. After the incubation with virus particles, approximately 15,000 MDCK cells in 100

µL culture medium were added to each well and further incubated for another 18-20

hours (overnight) at 37°C under an atmosphere of 5% CO2. During that time the cells

adhered to the plate bottom, and if infected, began the expression of viral proteins. After

the overnight incubation, the cells were fixed with a mixture of 4:1 acetone:PBS for 10-

15 minutes, the fixing solution was removed, and the plates were left to dry. Finally, an

ELISA was performed to detect the relative concentration of influenza nucleoprotein in

each well. An anti-nucleoprotein murine antibody (MAB8257, Millipore) and an anti-

mouse goat antibody-HRP conjugate (115-035-146, Jackson Immunoresearch, PA) were

used in the ELISA, and the absorbance of each well relative to the negative control

(uninfected cells) and positive control (cells + virus without inhibitor) was used to

calculate percent viral inhibition.

62

4.3.2 Inhibition of influenza infection with multivalent hemagglutinin-binding protein conjugates

Purified and concentrated HBA and HBB conjugates were prepared in pH 7.4 PBS so that

the concentration of HBA and HBB in each sample was 20 µM. The microneutralization

assay described in Section 4.3 was performed by Dr. Marc Douaisi of the Kane research

group, and the results are shown in Figure 4.12 and Figure 4.13.

Figure 4.12: Influenza virus infection inhibition efficacy of homodivalent bioconjugates of HBB.

Results shown are for HBB conjugated to various sizes of PEG scaffolds. Dose-response curves for HBB monomer or homodivalent PEG conjugates are shown on the left. The resulting IC50 for each sample is shown on the right. Homodivalent conjugates of HBB with PEG1.1k, PEG3.5k, or PEG7.5k exhibit 39.6-fold, 15.2-fold, and 14.9-fold improved activity, respectively.

As shown in Figure 4.12, homodivalent PEG-based conjugates of HBB all showed

significant improvements in inhibitory potency, although the dependence on linker length

was not clear. In fact, the results were likely confounded by incomplete purification. As

was noted during the initial purification of the HB proteins from the bacterial lysate, the

HB36.5 variants tend to spontaneously associate as non-covalent dimers. Due to the

similar size of these non-covalent dimers and the divalent PEG conjugates, the non-

covalent dimers tend to co-elute with the divalent PEG conjugates on an SEC column. This

phenomenon has made complete removal of unconjugated HB36.5 protein monomer

from the divalent conjugation products very difficult by this method. In fact, none of the

samples tested so far have been completely purified, and so the degree to which

differences in contaminating unconjugated monomer content affected the measured

inhibitory activity is unknown.

63

Figure 4.13: Influenza virus infection inhibition efficacy of polyvalent bioconjugates of HBA.

HBA was conjugated with AEM or polymeric scaffolds as indicated at the top. The legend shows the target percent coupling for the polymeric scaffolds. The blue samples indicate a target coupling of 100%. PGA60k, PGA120k, and [(SE)10K]10 scaffolds do not exhibit significant differences in inhibition efficacy relative to AEM-quenched HBA monomer, but HyA scaffolds exhibit approximately 30-fold improvements in inhibition efficacy.

The polyvalent conjugation of HBA to a range of polymeric scaffolds yielded

inconsistent results. Polyvalent attachment of HBA to PGA60k, PGA120k, or the PGA-like

bioengineered scaffold [(SE)10K]10 appeared to have no significant effect on the inhibition

efficacy of HBA relative to the AEM-quenched HBA monomer. However, the conjugation

to the HyA scaffold appeared to have approximately the same enhancement effect on a

per-protein basis as the PEG conjugation did. These bioconjugates were significantly

larger than the divalent PEG conjugates, so the problem of monomer contamination was

not an issue in this case. Perhaps the highly negatively-charged PGA and (SE)n scaffolds

are having some deleterious interaction with the glycoproteins on the virion surface

which has counteracted the expected enhancement due to polyvalency. Although HyA is

also a highly negatively-charged scaffold, I used a large excess of AEM relative to HyA

disaccharide monomers in these maleimide activation reactions. This was done in order

to compensate for the relatively higher viscosity of HyA in solution compared to PGA,

which otherwise significantly reduced the yield of AEM-activated monomers. As a result,

64

it is likely that nearly all of the carboxylates on the HyA scaffold were replaced by

maleimide functionalities, which would have neutralized the polyanionic character of the

scaffold.

4.4 Conclusions

As we hoped, the principles of multivalent inhibition were successfully adapted to a new

target, and over an order of magnitude enhancement in inhibitory efficacy was observed

with two different types of scaffolds, PEG and HyA. However, there were limitations that

hampered the ability to accurately assess the effects of the scaffold on the inhibition

efficacy. The activity of the divalent PEG-HB bioconjugates may have been strongly

affected by differences in amounts of contaminating unconjugated monomer, which co-

elutes when purified by SEC. To address this issue, other methods of purification are

currently being considered, such as hydrophobic interaction column chromatography. In

addition, production of the other HB protein inhibitor, HB80.4, is underway.

Unconjugated HB80.4 has already similar shown inhibitory activity to HB36.5 by our

microneutralization assay, and this protein is not known to form dimers in solution.

Therefore, divalent PEG bioconjugates of HB80.4 may be much easier to purify, and thus

to determine the effect of the size of the PEG linker on inhibition efficacy.

Continued work on the synthesis of polyvalent HyA-bioconjugates of HB-proteins is

also ongoing. To help to elucidate the effect of HB protein spacing, a commercially-

available, enzymatically-synthesized, monodisperse HyA scaffold will be used, and a series

of bioconjugates with a range of HB-protein coupling percentages will be synthesized. A

worthwhile extension of this project in future work may also be to bioengineer a single

protein polymer with several tandem repeats of the HB protein included in-line (similar

to the recent work with bioengineered anthrax toxin inhibitors,185 see Appendix B).

Furthermore, it would be exciting to see if the inhibitors could work in vivo, in an animal

model of influenza infection.

65

5. Multivalent HIV entry inhibitors

5.1 Rationale for the design of multivalent HIV entry inhibitors

5.1.1 Targeting the cellular receptor CCR5

HIV infections are particularly difficult to treat and prevent, due in large part to the

propensity of the virus for mutation. As mentioned in Section 1.3.3 on page 7, current

treatment and prevention strategies often involve the administration of small-molecule

antiretroviral drugs that target viral proteins such as reverse transcriptase, integrase, and

protease. Due to the selection pressure of these treatments and the high mutation rate

of retroviruses, HIV can develop resistance to these drugs over a short time of exposure.

On a related note, the explosion of the problem of antibiotic resistance has demonstrated

that the evolution of resistant microorganisms is virtually inevitable when therapeutic

strategies are used that directly target the proteins and other structural features of

pathogens with short life cycles. While the simultaneous administration of multiple

different antiretroviral drugs (a drug “cocktail”) can decrease the rate of evolution to viral

drug resistance, the strong selection pressure and propensity for mutation remains, so

the emergence of multi-drug-resistant viral strains is still possible.

The strategy of targeting host receptors may help overcome these problems. Host

cellular receptors are conserved targets, not prone to the same high mutation rate or

variability between different serotypes that may be the case for bacterial and viral

proteins. Moreover, targeting host receptors may be more conducive to preventative

medical strategies. These factors have led me to pursue the design of HIV entry inhibitors

that target the host cell receptors involved in the infection pathway, specifically CCR5 (See

Section 1.3.2 Infection pathway on page 6).

HIV infection requires that the viral envelope glycoprotein gp120 first binds to the

human glycoprotein CD4 and then to a co-receptor.52–54 The chemokine receptors CCR5

and CXCR4 are the primary co-receptors used by HIV; viral strains that use the CCR5 co-

receptor are referred to as R5 strains, whereas those that use the CXCR4 co-receptor are

referred to as X4 strains. Typically, CCR5-tropic viruses are the dominant strains during

66

the initial stages of infection, while CXCR4 (X4) tropism is a hallmark of viral evolution

during prolonged infection. A noteworthy corollary is that individuals who harbor a

genetic mutation in the gene for CCR5, known as CCR5-∆32, do not express CCR5 and are

highly resistant to HIV infection,186 but are otherwise healthy. Furthermore, the known

natural chemokine ligands of CCR5 all show the ability to inhibit HIV infection, and the

data suggest that the mechanism of inhibition may be primarily due to inducing receptor

internalization rather than by steric inhibition.187 Based on the health of individuals with

the CCR5-∆32 phenotype, the role of CCR5 in normal immune system operation appears

to be non-essential, and some studies actually suggest that the natural functions of CCR5

actually facilitates HIV infection by inducing immune system activation in ways that

attract additional HIV-permissive cells to the site of infection.188 As CCR5 is a host cell

receptor rather than a viral protein, CCR5 is a static target and is not prone to the

variations that occur naturally or that may evolve between different viral strains and

mutants. Furthermore, inhibitors that target CCR5 may also be effective when

administered prior to HIV exposure, and would therefore be a particularly useful means

for inhibiting viral transmission.

CCR5 is a G protein-coupled receptor that forms a 7-transmembrane-domain

structure, making it inherently difficult to analyze via X-ray crystallography or NMR

spectroscopy. Consequently, there has been a lack of high resolution structural

information that has prevented the rational design of CCR5-targeting HIV entry inhibitors.

However, there are a few known natural chemokine ligands of CCR5, and each has been

referred to in the literature by multiple different naming conventions. The most well-

known CCR5 ligand was called RANTES, an acronym standing for “regulated on activation,

normal T-cell expressed and secreted,” and is now officially named C-C chemokine ligand

5 (CCL5). Macrophage inflammatory protein 1α and 1β (MIP-1α and MIP-1β, respectively)

are also natural ligands, officially named CCL3 and CCL4, respectively. Although these

ligands of CCR5 are naturally able to inhibit HIV infection, their primary roles are to

activate various functions of the immune system, which may prohibit their direct use as

therapeutics.

67

5.1.2 Leukotoxin E is a protein ligand of CCR5

Many bacterial species of the genus Staphylococcus produce pore-forming toxins,

including γ-hemolysin and various enterotoxins and leukotoxins (also called leukocidins).

The leukotoxins help the bacteria evade the host immune system response during an

infection by causing osmotic imbalances that lead to cell death and lysis of leukocytes,

especially the neutrophils that engulf invading bacteria.189 Several types of leukotoxins

have been described in the literature, all comprised of two separately-secreted water-

soluble proteins of slightly different molecular mass, which self-assemble into octameric

toroids with alternating subunits.189–194 The naming convention for these bi-component

leukotoxins has been to refer to the two components as one unit, with each component

represented by a single letter so that the subunit with the lower molecular mass is

reported first. Leukotoxin AB is one example, where subunit A is the slower-eluting

component on a gel filtration column (“S-component”) and subunit B is the faster-eluting

component (“F-component”), therefore leukotoxin subunit A is of lower molecular mass

than subunit B.

In general, the fact that leukotoxins kill leukocytes has been known for decades, but

it has only been very recently that some light has been shed on which host factors

function as the toxin receptors. For example, Leukotoxin ED (LukED) was originally

described in 1998,195 but it wasn’t until 2013 that it was reported by the Torres research

group that host cell expression of CCR5, CXC chemokine receptor 1 (CXCR1, formerly

known as Interleukin 8 receptor alpha, or IL8RA), or CXC chemokine receptor 2 (CXCR2,

formerly known as Interleukin 8 receptor beta, or IL8RB) are required for LukED

cytotoxicity.196,197 While both toxin subunits were required for the cytotoxic effect, it was

shown by the use of green fluorescent protein (GFP) fusions that only subunit E (LukE)

mediates binding to CCR5, CXCR1, and CXCR2.196,197 In addition, the researchers were able

to use the wide array of known CCR5-binding antibodies to begin to probe which region

of CCR5 interacts with LukE. Of particular interest to my research was the fact that LukE

binding to CCR5 was inhibited by the antibody 45531, which has also been shown to

inhibit HIV infection in vitro.196,198 Furthermore, the researchers showed that LukE did not

68

activate CCR5-mediated cell signaling and even slightly suppressed signaling by CCL5.196

This result is consistent with the leukotoxin function enabling immune system evasion,

but it is also evidence that LukE might serve as a useful ligand for inhibition of HIV cell

entry.

5.1.3 Design of multivalent inhibitors that target CCR5

Although synthetic inhibitors of HIV entry via the CCR5 co-receptor have already been

identified, these have primarily been small molecule inhibitors that are prone to

complications associated with inhibiting protein-protein interactions. This is because the

large contact areas that are typical of protein-protein interactions may allow for partial

retention of the binding interface even when a competing small molecule drug or peptide

ligand is present. The retention of a partial binding interface may even enable the

acquisition of mutations that allow the virus to recognize the receptor-inhibitor complex.

Moreover, if the two interacting proteins are already bound, it may be unlikely that a

small ligand will gain access to the target site, even if the drug has a lower dissociation

constant. In the case of HIV, these types of concerns may be more problematic because

the virion is coated with a polyvalent display of the gp120 glycoprotein. In addition,

studies have shown that viral resistance to such small-molecule CCR5-targeting entry

inhibitors occurs through the acquisition of mutations that increase gp120’s affinity for

CCR5 and/or that bind to the CCR5-drug complex.199

Several factors suggest that multivalent inhibitors might be well suited to

counteracting the challenges of HIV antigenic drift and resistance to other methods of

entry inhibition. First, the inherently larger structure of multivalent binders should be

more effective than a monovalent ligand at creating steric hindrance to block access of

the HIV glycoprotein to CCR5. This larger size, along with the fact that CCR5 can diffuse

laterally on the cell surface, should allow for a multivalent inhibitor to bind to multiple

CCR5 molecules simultaneously and with high avidity. The ability to tightly bind multiple

CCR5 receptors may play an important role in thermodynamically outcompeting HIV for

access to the receptors, thus preventing the emergence of resistant strains and reducing

the consequences of viral mutations. Additionally, the ability to bind multiple CCR5

69

receptors and draw them together may induce endocytosis, effectively downregulating

the expression of CCR5 on the cell surface and producing a virtual CCR5-∆32 phenotype.

The Kane research group92 and others75,81,93,121,164 have demonstrated the potential

value of polyvalent ligands based on PGA81,93,121 or hyaluronic acid75,164 (HyA) scaffolds.

We reasoned that these biocompatible scaffolds should also support the synthesis of

polyvalent conjugates of LukE or GFP-LukE. Based on our previous research experience in

designing polyvalent inhibitors, we expected polyvalent conjugates of LukE or GFP-LukE

to be more efficacious on a per-ligand basis than the monovalent LukE or GFP-LukE at

binding CCR5. Furthermore, we hypothesized that these polyvalent conjugates would be

able to compete with other known CCR5-binders and potentially inhibit HIV attachment

and entry into CCR5+ cells. Therefore I have synthesized polyvalent bioconjugates using

PGA and HyA as the scaffolds for polyvalent display of LukE. The polyvalent conjugates

were synthesized by using the free thiol of LukE engineered with a C-terminal cysteine for

attachment to PGA or HyA scaffolds pre-activated by the polyvalent attachment of

aminoethylmaleimide (AEM). The resulting conjugates were characterized by SDS-PAGE

and SEC to determine the quality of the polyvalent attachment, and flow cytometry to

determine the ability to bind and compete for CCR5 on HEK-293T cells.

5.2 Production of polyvalent bioconjugates of leukotoxin E

5.2.1 Cloning and bacterial production of leukotoxin E variants and fusion proteins

The bacterial genome of the S. aureus strain Newman was purchased from ATCC, and the

gene encoding LukE was PCR amplified using the forward primer (5’-TCT AGA AAT AAT

TTT GTT TAA CTT TAA GAA GGA GAT ATA CAT ATG AAT ACT AAT ATT GAA AAT ATT GGT

GAT GGT G-3’) and either the reverse primer (5’-CTC GAG TTA GTG ATG ATG ATG ATG

ATG ATT ATG TCC TTT CAC TTT AAT TTC GT-3’) to add a C-terminal hexahistadine tag (His-

tag), or (5’-CTC GAG TTA ACA GTG ATG ATG ATG ATG ATG ATT ATG TCC TTT CAC TTT AAT

TTC GT-3') to add an additional cysteine to the end of the His-tag. The addition of the His-

tag facilitates purification by an immobilized metal affinity column (IMAC). It should be

noted that the wild type LukE sequence does not contain any cysteines, so the addition

70

of a cysteine at the C-terminus provides the LukE variant with a single thiol as a chemical

handle for conjugation reactions. The two variants were each cloned into pET28b vectors

between the XbaI and XhoI restriction sites. The pET28b vectors also contained a gene for

kanamycin resistance. The resulting constructs were then used to transform the E. coli

Rosetta™2(DE3) strain (Novagen), which contains a chromosomal copy of T7 RNA

polymerase under the control of the lacUV5 promoter.

In order to reproduce the GFP-LukE fusion proteins used by Alonzo III et al., a GFP

coding sequence from the plasmid pCIBN(deltaNLS)-pmGFP (Addgene Plasmid #26867)

was amplified by PCR. During amplification, the forward primers (5’-CAT ATG CAT CAT

CAT CAT CAT CAC GTG AGC AAG GGC GAG GAG C-3’) or (5’-CAT ATG TGT CAT CAT CAT

CAT CAT CAC GTG AGC AAG GGC GAG GAG C-3’) and the reverse primer (5’-CTC GAG TTG

TTG TTG TTG GGA TCC CTT ATA CAG CTC GTC CAT GCC GAG AGT GAT CC-3’) were used to

introduce a BamHI restriction site at the 3’ end and either a hexahistidine or a Cys-His6

tag at the N-terminus. These were ligated into the NdeI and XhoI restriction sites of a

modified pET28b plasmid, in which the N-terminal hexahistidine coding sequence and

NcoI restriction site had been previously removed. The LukE coding sequence was again

PCR amplified from the S. aureus genomic DNA, this time with the forward primer (5’-

GGA TCC AAT ACT AAT ATT GAA AAT ATT GGT GAT GGT G-3’) and the reverse primer (5’-

CTC GAG TTA ATT ATG TCC TTT CAC TTT AAT TTC GT-3’), and ligated downstream of GFP

into the BamHI and XhoI restriction sites of the modified pET28b plasmid, which also

conferred kanamycin resistance. The LukE and GFP-LukE cloning work described in this

section was performed by Dr. Manish Arha, and I used the resulting constructs to

transform E. coli Rosetta™2(DE3) for bacterial production of the proteins.

The transformants were inoculated from starter cultures at 1:100 dilution into

multiple baffled Fernbach flasks, each containing 1 L of auto-induction media supple-

mented with 100 µg/mL kanamycin, as described by Studier.168 After 6 hours of incubation

at 37°C followed by 24 hours at 18°C and a constant 215 RPM, the cells were pelleted by

centrifugation at 4°C for 30 minutes, then resuspended with a homogenizer into lysis

buffer (pH 8.5, 100 mM HEPES, 100 mM NaCl, 25 mM imidazole, 1 mM TCEP, 1 mM PMSF,

71

2 mg/mL lysozyme, 0.2 mg/mL DNase) and lysed by passing through a microfluidizer

twice. The resulting lysate was clarified of insoluble material by centrifugation at 4°C and

25,000 g for 30 minutes, and finally filtered through a 0.45 µm filter. A disposable plastic

column was packed with ~2.5 mL Ni-NTA agarose resin for each liter of bacterial culture,

and the clarified, filtered lysate was passed through by gravity flow. The His-tagged LukE

variants exhibited strong affinity for the immobilized metal resin, and were cleared of

most contaminants by washes with a step gradient of imidazole (25 mM, 50 mM, 100 mM

for 5 column volumes each) in HEPES buffered saline (HBS). LukE variants were eluted

from the columns with 5 column volumes of 400 mM imidazole in HBS. Reducing SDS-

PAGE with silver staining was performed to confirm the successful expression and

elimination of the majority of contaminating native bacterial proteins, as shown in Figure

5.1 for LukE-H6-C and Figure 5.2 for C-H6-GFP-LukE. The results of the gels were used to

guide the choice of fractions to be pooled and concentrated by 10 kDa molecular weight

cut-off (MWCO) centrifugal filtration and subsequent size-based purification on a

prepacked GE HiLoad 16/600 Superdex200 prep grade SEC column. The cysteine-

containing protein variants were purified in pH 7.5 MES buffered saline (MBS) with TCEP

and EDTA for preventing the formation of disulfide bonds between proteins, and the non-

reactive protein variants were purified in pH 7.4 PBS. The chromatogram of UV

absorbance at 280 nm produced a single peak in the expected elution volume for each

protein. The purity of the fractions corresponding to the LukE or GFP-LukE variant peak

was confirmed by non-reducing SDS-PAGE, and the corresponding fractions were pooled

and sterilized by vacuum filtration through 0.2 µm filters. Finally, the protein solutions

were concentrated by 10 kDa MWCO centrifugal filtration and stored at 4°C until further

use in various assays and conjugation reactions.

72

Figure 5.1: Silver-stained reducing SDS-PAGE of IMAC fractions from LukE-H6C purification.

Lanes left to right: (1-4) 5.0 µL each of 2nd to 5th column volumes 25 mM imidazole in HBS wash, (5-6) 5.0 µL each of 1st and 2nd column volumes 50 mM imidazole in HBS wash, (7) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (8-10) 5.0 µL each of 3rd to 5th column volumes 100 mM imidazole in HBS wash, (11-18) 5.0 µL each of 1st to 8th column volumes 100 mM imidazole in HBS wash, (19) 5.0 µL of 1st column volume 400 mM imidazole in HBS elution, (20) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (21-24) 5.0 µL of 2nd to 5th column volumes 400 mM imidazole in HBS elution. Expected molecular mass of LukE-H6C is 32.9 kDa.

Figure 5.2: Silver-stained reducing SDS-PAGE of IMAC fractions from CH6-GFP-LukE purification.

Lanes left to right: (1) 5.0 µL of 2nd column volume 25 mM imidazole in HBS wash, (2-3) 5.0 µL each of 2nd and 4th column volumes 50 mM imidazole in HBS wash, (4-5) 5.0 µL of 1st and 2nd column volumes 100 mM imidazole in HBS elution, (6) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (7-9) 5.0 µL

73

each of 3rd to 5th column volumes 100 mM imidazole in HBS elution, (10-12) 5.0 µL each of 2nd, 3rd, and 5th column volumes 400 mM imidazole in HBS elution. Expected molecular mass of CH6-GFP-LukE is 59.9 kDa.

5.2.2 Synthesis of polyvalent leukotoxin E bioconjugates

Polyvalent bioconjugates of LukE-H6C were synthesized by the same polyvalent

bioconjugation protocols described previously in Section 4.2.5 Synthesis of polyvalent

hemagglutinin-binding protein bioconjugates. The silver-stained non-reducing SDS-PAGE

of a typical conjugation set of LukE-H6C with two different MW PGA scaffolds is shown in

Figure 5.3. Each lane contains the same quantity of LukE (5.0 µg, or 152 pmol). Lanes 2

and 3 are useful guides to show the positions of LukE monomer and disulfide-bonded

dimer, respectively. The high molecular mass bands in lanes 4 through 8 are indicative of

the formation of large polyvalent conjugates. Before purifying out the unreacted LukE

from these bioconjugate reaction products, the maleimide rings were hydrolyzed by

treatment with 100-200 mM molybdate oxyanion for 1-2 hours.

Figure 5.3: Silver-stained non-reducing SDS-PAGE of LukE-H6C conjugation reaction products before purification, 5.0 µg total LukE protein per lane.

74

Lanes left to right: (1) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (2) LukE-H6C-AEM (LukE with the reactive thiol quenched by reaction with excess AEM to prevent disulfide bond formation – expected MW = 33.1 kDa), (3) LukE-H6C in the presence of 5 mM DHA (attempt to force disulfide bond formation – expected MW = 65.8 kDa), (4-5) Reactions targeting 5% or 13% coupling of PGA60k monomers with LukE-H6C, (6-7) Reactions targeting 5% or 13% coupling of PGA120k monomers with LukE-H6C. Molecular mass of unconjugated LukE-H6C is 32.9 kDa.

5.2.3 Purification of polyvalent leukotoxin E bioconjugates

Each of the PGA-LukE conjugation reaction samples from Figure 5.3 were treated with 20

mM TCEP in MBS with 100 mM molybdate for 1-2 hours and then run on a 24 mL

Superdex200 column with pH 7.4 PBS as the eluent. As is evident from the

chromatograms, there was a significant amount of unreacted LukE in each sample,

corresponding to a large peak relative to the low shallow hump from the polyvalent

conjugate. However, by pooling all elution fractions collected up until the unreacted

monomer peak and running SDS-PAGE (Figure 5.4), it was clear that the unreacted LukE

was successfully removed (compare the position of the AEM-quenched LukE monomer in

lane 10 to lanes 1-4 and 6-9). After confirming the successful removal of unconjugated

LukE from the desired conjugation products, I next concentrated the samples by

centrifugal filtration with a 3 kDa MWCO filter, measured the concentration of LukE in

each sample by recording the UV absorbance at 280 nm on a NanoDrop

spectrophotometer, and adjusted each sample to a set concentration in preparation for

the flow cytometry assay of CCR5-binding inhibition.

75

Figure 5.4: Silver-stained non-reducing SDS-PAGE of LukE-H6C conjugation reaction products after SEC purification on a Superdex200 column.

Lanes left to right: (1) Highest MW SEC fractions of LukE conjugated to PGA60k (target 5% coupling), (2) Second highest MW SEC fractions of LukE conjugated to PGA60k (target 5% coupling), (3) Highest MW SEC fractions of LukE conjugated to PGA60k (target 13% coupling), (4) Second highest MW SEC fractions of LukE conjugated to PGA60k (target 13% coupling), (5) 5.0 µL of Thermo Scientific PageRuler™ Plus Prestained Protein Ladder with approximate molecular mass markers labeled, (6) Highest MW SEC fractions of LukE conjugated to PGA120k (target 5% coupling), (7) Second highest MW SEC fractions of LukE conjugated to PGA120k (target 5% coupling), (8) Highest MW SEC fractions of LukE conjugated to PGA120k (target 13% coupling), (9) Second highest MW SEC fractions of LukE conjugated to PGA120k (target 13% coupling), (10) Monomer SEC fractions of LukE-H6C-AEM (LukE with the reactive thiol quenched by reaction with excess AEM, to prevent disulfide bond formation – expected MW = 33.1 kDa).

5.3 Characterization of polyvalent conjugates of leukotoxin E

5.3.1 Flow cytometry assay of CCR5-binding inhibition

The LukE variants were tested by flow cytometry for their ability to bind to CCR5-

expressing HEK-293T cells (Human embryonic kidney cells containing the SV40 large T

antigen, which enables the cell to replicate transfected plasmids that contain the SV40

origin of replication during cell division, thus allowing for the inherited expression of a

76

non-native gene such as CCR5). Approximately 100,000 cells were incubated with 300 nM

LukE or GFP-LukE at 4°C for 30 minutes, followed by the addition of one of a variety of

competing CCR5-specific binders or a control antibody (HLA-ABC, APC-conjugate, 562006,

Becton Dickinson, CA) that binds to MHC class I molecules (which are also expressed by

the cells). The phycoerythrin-conjugated monoclonal antibodies 2D7 (555993, Becton

Dickinson, CA) or 45531 (FAB182P, R&D Systems, MN), which also bind to the ECL2 of

CCR5, or 3A9 (560635, Becton Dickinson, CA), which binds to the N-terminus of CCR5,

were added at a concentration of 1 nM, 4.2 nM, or 8.3 nM, respectively. Alternatively,

FLSC-Ig, a fusion protein mimic of the HIV gp120 protein in complex with CD4 (i.e. HIV

gp120 in its CCR5-binding conformation) was added at a concentration of 23 nM. FLSC-Ig

stands for full length single chain immunoglobulin, and consists of the HIV gp120 from the

strain Ba-L, the D1 and D2 domains of the human glycoprotein CD4, and a human

immunoglobulin Fc fragment IgG1. Therefore, FLSC-Ig might represent the binding

capability of HIV for CCR5 more accurately. In the presence of each competing binder (or

control antibody), the LukE or GFP-LukE remained at an excess of 13-300 fold, and the

combined solution was incubated with the cells at 4°C for an additional 30 minutes. After

two washes with 250 µL of PBS + 1% FBS, FLSC-Ig was detected with 2 µg/mL

AlexaFluor647-conjugated anti-human IgG goat antibody (A21445, Invitrogen, CA) at 4°C

for another 20 minutes. Finally the cells were washed twice in 250 µL PBS + 1% FBS and

then resuspended and fixed in 1% paraformaldehyde (in PBS) prior to acquisition with a

LSRII flow cytometer (Becton Dickinson, CA).

5.3.2 Inhibition of CCR5-binding with polyvalent bioconjugates of leukotoxin E

Purified LukE and GFP-LukE conjugates with PGA60k, PGA120k, and HyA were prepared in

pH 7.4 PBS so that the concentration of LukE or GFP-LukE in each sample was 2 µM. AEM-

quenched LukE and GFP-LukE were also prepared at the same concentration. The assay

described in Section 5.3 was performed by Dr. Marc Douaisi of the Kane research group,

and the results are shown in Figure 5.5 and Figure 5.6. In both figures, the amount of

CCR5-binding antibody detected when in competition with the LukE sample is reported

as a percentage of the amount of antibody detected when no LukE was present. In

77

addition, the binding of the FLSC-Ig was detected by a secondary antibody against the Fc

portion of the fusion protein. The concentration of the FLSC-Ig and antibodies used are

reported along the horizontal axis of both figures.

Figure 5.5: CCR5-binding inhibition efficacy of polyvalent PGA-LukE bioconjugates.

Figure 5.6: CCR5-binding inhibition efficacy of polyvalent HyA-LukE bioconjugates.

45531, 2D7, and 3A9 are codes for CCR5-binding antibodies. These antibodies were

competing with the polyvalent polymer-LukE conjugates for binding CCR5 on CCR5+

78

HEK293T cells. 45531 and 2D7 bind specifically to separate epitopes on the second extra-

cellular loop of CCR5 (ECL2), and 3A9 binds specifically to the N-terminus of CCR5.198 LukE

was originally reported to bind to ECL2 based on the fact that antibody 45531 could

directly outcompete LukE for binding CCR5.196 In contrast, neither 3A9 nor 2D7 were

shown to inhibit LukE binding, despite the fact that 2D7 binds to the ECL2 of CCR5 near

the binding epitope of 45531.196 HLA-ABC is the generic name for a control antibody that

binds to the unrelated MHC class I proteins that are also expressed by the HEK293T cells.

The binding of HLA-ABC thus serves as a control. In Figure 5.5, none of the PGA-LukE

conjugates appear to have any effect on the ability of any of the antibodies or the FLSC-

Ig to bind to CCR5. Unconjugated LukE does not seem to have much effect either, other

than perhaps a ~10% inhibition of antibody 45531, the antibody reported to compete

most directly with LukE for binding CCR5. In contrast, HyA-LukE13% (13% of disaccharide

monomers were targeted for LukE coupling) appears to be able to inhibit all three CCR5-

binding antibodies in the range of 30-60% (Figure 5.6). Surprisingly, the binding of HLA-

ABC to CCR5+ cells was also inhibited, although to a slightly lesser extent. However, the

binding of this antibody was not inhibited on CCR5- cells, indicating that the inhibition of

HLA-ABC is not due to non-specific binding of the HyA-LukE conjugate to the cells. I

hypothesize that the inhibition of HLA-ABC on CCR5+ cells might be due to steric

hindrance by the large HyA backbone. On the other hand, the binding of FLSC-Ig was not

inhibited by HyA-LukE either.

5.4 Conclusions

The ability of HyA-LukE to block three different CCR5-binding antibodies is a very

promising result, especially because each of these antibodies is itself able to inhibit HIV

infection. As with the HyA-HB bioconjugates described in Chapter 4, additional work

synthesizing HyA-LukE bioconjugates with more well-defined HyA scaffolds and coupling

percentages is underway. This work should help to elucidate the optimal spacing between

individual, polyvalently-conjugated LukE proteins on a linear scaffold. In addition, GFP-

LukE fusion proteins have been successfully produced, and polyvalent HyA bioconjugates

79

are in the process of being synthesized. The CCR5-binding capacity of these HyA-GFP-LukE

bioconjugates should be detectable by flow cytometry directly via the GFP fluorescent

signal. Once the binding inhibition of these HyA-LukE and HyA-GFP-LukE bioconjugates

has been optimized, testing them directly in HIV infection inhibition assays is the next

logical step.

80

6. Multivalent oligonucleotide aptamer bioconjugates

6.1 Rationale for the use of oligonucleotide aptamers as ligands

A potential downside of the bioconjugates described up to this point is the use of peptides

or proteins as ligands. These types of ligands might suffer from problems in therapeutic

applications due to the tendency of non-native peptides and proteins to provoke an

immunogenic response. Such undesired side-effects can be a concern especially if the

therapeutics are to be administered as prophylactics. In addition, peptide ligands in

particular often exhibit only weak affinity for their targets, partly due to their small size.

Although this can be overcome by multivalent presentation, peptides may not be the best

choice of ligand in applications where very strong avidity with the target is desired.

Furthermore, peptide ligands may share some of the same difficulties associated with

inhibiting protein-protein interactions as small-molecule drugs. Finally, manufacture of

synthetic peptides is typically limited to short sequences, partly due to the high cost of

reagents and partly due to the increased likelihood of errors during the synthesis of longer

peptide sequences. Recombinant production of protein ligands can be even more

complicated and costly. Proteins which require post-translational modifications such as

glycosylation or the formation of internal disulfide bonds can be especially difficult, and

proteins often require special storage conditions to prevent their misfolding, aggregation,

or precipitation.

Oligonucleotide aptamers have a number of advantages over peptide and protein

antagonists, including their low cost of production and predisposition for synthesis by

“good manufacturing practices”. Oligonucleotide aptamer-based therapeutics may also

have a better outlook for approval by the US FDA, because nucleic acids are typically non-

immunogenic. A case in point is the FDA-approved RNA aptamer, Macugen, which is

prescribed for the treatment of neovascular age-related macular degeneration. In

addition, even short oligonucleotides are typically much larger in size and molecular mass

than peptides, and this increased size may enhance their ability to inhibit protein-protein

interactions. The increased contact area between oligonucleotide aptamers and their

targets also allows for increased affinity and specificity relative to peptide ligands. For

81

example, some oligonucleotide aptamers have been characterized as binding their targets

so strongly that even the dissociation constants of monovalent complexes are in the

picomolar range. Taken together, these characteristic features of oligonucleotide

aptamers make them ideal ligands for the design of multivalent inhibitors.

6.2 Production of polyvalent bioconjugates of the ssDNA aptamer sgc8c

Our previous work and the literature on polyvalent therapeutics suggested that we would

be able to significantly enhance the efficacy of the monovalent aptamers using

polyvalency.

To test this hypothesis, I synthesized polyvalent conjugates of a previously-

published200–202 ssDNA aptamer called sgc8c with PGA scaffolds for polyvalent ligand

display (see Figure 6.1, Figure 6.2, Figure 6.3, and Figure 6.4). The sgc8c aptamer is 42

bases long and was first identified by whole cell-SELEX. It binds to a cellular receptor on

lymphocytes,203 so it is a great model for the kind of targeting ligand we are interested in

moving to in the future. Although we (and others) have already reported the successful

use of polyvalent inhibitors based on PGA scaffolds,81,93,121,122 these previously

synthesized inhibitors have not made use of oligonucleotide aptamer ligands.

6.2.1 Synthesis of polyvalent bioconjugates of the ssDNA aptamer sgc8c

One of the challenges of polyvalent conjugation of oligonucleotide aptamers to PGA is the

fact that each of the biomolecules in question is polyanionic. The behavior of

polyelectrolytes in aqueous solution is extremely complex204 (especially for polyanions in

solution with a mixture of monovalent and divalent cations, e.g. Na+, Ca2+, and Mg2+), and

charge repulsion between the scaffold and oligonucleotide ligands may restrain the ability

to use some conjugation strategies.

Anticipating these challenges, I first optimized the bioconjugation conditions using

much less expensive single-stranded DNA (ssDNA) oligonucleotides, 25 bases long, as

model nucleic acid aptamers, and then synthesized polyvalent PGA-based conjugates of

the ssDNA aptamer sgc8c. Both model ssDNA oligonucleotides were custom-synthesized

by a commercial supplier (Integrated DNA Technologies, Inc.) with a 5’-thiol functionality,

82

which was provided protected by a disulfide-linked hexane cap. In order to expose the

free thiol, the ssDNA stock was incubated with 10 mM trialkylphosphine reducing agent

for 1 hour at pH 6.5, similar to the preparation of protein-thiols for maleimide conjugation

described in Section 4.2.3. The cleaved hexane-thiol cap is then removed from the ssDNA-

thiol solution by passing through a size-exclusion spin column.

After preparing ssDNA oligonucleotides by the deprotection/purification protocol,

the ssDNA and functionalized PGA scaffold are combined in a reaction buffer the same as

that described in Section 4.2.5.2 for polyvalent HB conjugation to PGA. The final

concentration of oligonucleotide was adjusted to 200 µM in order to be consistent with

the concentration we expect to be able to achieve with our 5’ modified, transcribed RNA

aptamers. As shown in Figure 6.1 for the ssDNA model oligonucleotides and in Figure 6.2

for the sgc8c aptamer, under these conditions, my reactions result in high conjugation

efficiency. Keeping in mind that the molecular weight of the polymer backbone may

influence inhibitory potency, I have functionalized PGA scaffolds from two different

commercially available molecular weights (Mw 35 kDa and 120 kDa).92 After the

conjugation was completed, the remaining maleimide functionality were hydrolyzed by

incubation with 100 mM MoO4 for 1-2 hours.

83

Figure 6.1: SYBR Gold-stained 15% acrylamide TBE-Urea PAGE of ssDNA model aptamer polyvalent conjugation reactions with PGA-maleimide after 88 hours at ambient temperature.

Lanes left to right: reaction mixture with maleimide-functionalized PGA (2.3% of PGA monomers exhibit maleimides, 23 µM maleimide, 200 µM ssDNA-thiol), reaction mixture with maleimide-functionalized PGA (15% of PGA monomers exhibit maleimides, 150 µM maleimide, 200 µM ssDNA-thiol), reaction mixture with maleimide-functionalized PGA (37.5% of PGA monomers exhibit maleimides, 375 µM maleimide, 200 µM ssDNA-thiol), ssDNA oligomer ladder (27, 53, 75, 101mers), ssDNA-thiol oligomer reagent stock after deprotection and purification through BioRad spin column. The bands in the conjugation reaction lanes from top to bottom are presumably the polyvalent conjugate(s) with varying degrees of ssDNA coupling per PGA chain, disulfide-bonded (ssDNA-SS-ssDNA) divalent oligomers, and ssDNA-SS-cap that remained protected. There appears to be no remaining unreacted deprotected ssDNA-thiol in the reaction mixture in which there was an excess of available maleimides.

84

Figure 6.2: Heat map of SYBR Gold-stained 4-20% acrylamide TBE-PAGE of ssDNA aptamer sgc8c conjugation reactions with PGA-maleimide.

Lanes left to right: (1) ssDNA oligomer ladder with corresponding number of bases indicated to the left, (2-3) Reactions targeting 2.2% or 47% coupling of PGA35k monomers with sgc8c-thiol, (4-5) Reactions targeting 2.2% or 47% coupling of PGA120k monomers with sgc8c-thiol, (6) unreacted sgc8c-thiol with thiol protection cap uncleaved.

6.2.2 Purification of polyvalent bioconjugates of the ssDNA aptamer sgc8c

The reaction mixture corresponding to lane 5 in Figure 6.2 (PGA120k-sgc8c) was purified by

centrifugal filtration with a 100 kDa MWCO filter, as described in Section 4.2.6,

Purification of multivalent hemagglutinin-binding protein bioconjugates. The reaction

mixture corresponding to lane 3 in Figure 6.2 (PGA35k-sgc8c) was purified by SEC on a GE

Tricorn 10/300 GL Superdex 200 column, as shown in Figure 6.3. Fractions 4 and 5,

corresponding to the elution volume from 8 to 12 mL, were combined and concentrated

on 3k MWCO centrifugal filters prior to testing.

85

Figure 6.3: SEC multi-chromatogram of the PGA35k-sgc8c reaction product.

The PGA35k-sgc8c reaction product (lane 3 in Figure 6.2) was run on a GE Tricorn 10/300 GL Superdex200 column, with 2 mL fractions indicated as vertical lines. The absorbance of the eluent at wavelengths of 230 nm and 260 nm are shown as green and orange traces, respectively. Unreacted sgc8c monomer elutes between 14 and 16 mL, sgc8c disulfide-bonded dimer elutes between 12.5 and 14 mL, and polyvalent conjugates elute from the void volume through ~12 mL. Fractions 4-5 (4-8 mL elution time) were collected and tested as shown in Figure 6.4.

6.3 Characterization of polyvalent bioconjugates of the ssDNA aptamer sgc8c

We tested the ability of monovalent unconjugated sgc8c and polyvalent PGA-sgc8c

conjugates to inhibit the binding of a biotinylated version of the monovalent aptamer to

the lymphoblast T cell line MOLT-4. Sgc8c-biotin binding to lymphoblasts was detectable

by flow cytometry with the aid of a commercially available streptavidin-phycoerythrin

bioconjugate, and a dose-response assay of the sgc8c-biotin binding is shown in Figure

6.4. This curve yielded an EC50 of 17.9 nM. The ability of monovalent non-biotinylated

sgc8c or polyvalent PGA-sgc8c to inhibit sgc8c-biotin binding was tested by first

incubating serial dilutions of the inhibitors with the MOLT-4 cells for 30 minutes, followed

by washing and incubation with 100 nM sgc8c-biotin for 30 minutes more. The cells were

86

then washed again and incubated with streptavidin-phycoerythrin for another 30

minutes, before a final wash and fixation with paraformaldehyde. The phycoerythrin

signal was detected by flow cytometry, and the dose-response curves are shown in Figure

6.4. The IC50 of the monovalent, non-biotinylated sgc8c aptamer was 17.5 nM, which

matches the EC50 of the sgc8c-biotin (17.9 nM). By comparison, the polyvalent PGA35k-

sgc8c and PGA120k-sgc8c bioconjugates inhibited sgc8c-biotin binding with IC50s of 1.6 nM

and 1.9 nM, respectively. Therefore, the polyvalent bioconjugation of the ssDNA aptamer

enhanced its inhibitory activity by about an order of magnitude in both cases.

Figure 6.4: Enhanced binding of polyvalent PGA-sgc8c conjugates to lymphocyte cell receptors, assayed by flow cytometry.

Sgc8c-biotin, labeled as “agonist” in the graph on the left, was exposed to lymphocytes at a wide range of concentrations and binding to the cell receptors was detected by the secondary binding of streptavidin-PE. PGA35k-sgc8c (green curve in the antagonist graph on the right) and PGA120k-sgc8c (red curve in the antagonist graph on the right) purified from reactions in lanes 3 and 5, respectively, of Figure 6.2, showed nearly equivalent ability to compete with sgc8c-biotin for cell binding. Both polyvalent conjugate samples exhibited nearly two orders of magnitude enhancement in the ability to compete with sgc8c-biotin, relative to non-biotinylated sgc8c (blue curve in the antagonist graph on the right).

6.4 Conclusions

These results are extremely promising, as they strengthen the case for pursuing the use

of oligonucleotide aptamers as ligands in the design of multivalent inhibitors. If aptamers

can be identified against targets of interest, we could use these protocols for the

production of biocompatible multivalent therapeutics with low propensity for provoking

undesired immunogenic responses in patients.

87

7. Suggestions for future work4

7.1 Identification of oligonucleotide aptamers that target cell receptors

7.1.1 Rationale for the search for oligonucleotide aptamers that target cell receptors

We have already shown that targeting [PA63]7 can be a successful therapeutic strategy for

inhibiting the pathogenicity of B. anthracis.205,206 Additionally, I have more recently shown

that we can successfully inhibit influenza infection in vitro by targeting the viral HA

protein. However, there are some potential downsides for these strategies due to their

reliance on peptide and protein targeting ligands.

As was already discussed in Section 6.1 Rationale for the use of oligonucleotide

aptamers as ligands, oligonucleotide aptamer ligands have many advantages over peptide

and protein ligands. Furthermore, as was discussed in Section 5.1.1 Targeting the cellular

receptor CCR5, the host receptors that are exploited by a pathogen may be the most

logical targets when designing inhibitors of that pathogen. These factors have led me to

pursue biopanning for new anthrax toxin inhibitors that target the host cell receptors

involved in the intoxication pathway, specifically ANTXR2 (See Section 1.1.2 Intoxication

pathway, page 1).

7.1.2 Screening libraries of oligonucleotide aptamers

7.1.2.1 Systematic Evolution of Ligands by EXponential enrichment (SELEX)

Protocols for the identification of oligonucleotide aptamers were first reported in 1990

by two research groups independently.207,208 Systematic Evolution of Ligands by

EXponential enrichment (SELEX) is a type of iterative biopanning procedure for identifying

aptamers with desired binding properties. Many different adaptations of this protocol

have been reported in the past ~20 years,209–211 but the general procedure (depicted in

Figure 7.1) of almost every variant involves a progressive cycling of the following steps:

Portions of this chapter previously appeared as: Martin, J. T.; Kane, R. S. Design of Polyvalent Polymer Therapeutics. In Functional Polymers by Post-Polymerization Modification; Theato, P.; Klok, H.-A., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2013; pp. 267–290.

88

exposure of a large, randomized, undefined “library” of unique sequences of potential

oligonucleotide ligands to a target molecule, isolation of binders, and amplification of the

binder sequences to regenerate the oligonucleotide pool. This amplification step (and

additional translation steps in the case of RNA aptamer panning) requires the inclusion of

predefined primer sequences in the oligonucleotide sequence (see Figure 7.2). The

primers enable the enzymatic amplification of the oligonucleotide pool by polymerase

chain reaction (PCR), as well as forward and reverse in vitro transcriptions of the pool

between RNA and DNA forms, for the cases in which RNA aptamers are the desired

product. As a result, the standard library design involves a string of ~30-40 randomized

nucleotides, flanked on either end by fixed primer sequences (Figure 7.2).

Figure 7.1: General SELEX procedure for screening RNA aptamers.

Reprinted from Int. J. Parasitol., 33, Göringer, H. U.; Homann, M.; Lorger, M., In vitro selection of high-affinity nucleic acid ligands to parasite target molecules, 1309-1317, 2003, with permission from Elsevier.

89

Figure 7.2: The sequence of the initial DNA library and a representative mfold prediction.

The sequence of the initial DNA library for a typical SELEX screen with 40 random nucleotides flanked by primer-derived sequences is above an mfold-predicted secondary structure of a model aptamer, showing the interference from the flanking regions.

7.1.2.2 A modified SELEX protocol for identifying short oligonucleotide aptamers

Aptamers which have been identified by screening standard SELEX libraries as described

above may not be feasible for use in therapeutics due to their excessive length. The fixed

primer sequences can contribute to the RNA aptamer secondary structure (Figure 7.2)

and consequently influence target binding; thus, the random portion of the sequence

probably cannot be isolated and used independently. As a result, it may be necessary to

continue using the entire sequence, primers included, that was present for the screen

(typically ~75 bases or more). In order to reduce the high synthesis costs of such long

aptamers that may be prohibitive for therapeutic applications, it might be possible to

perform 5’ and 3’ truncations, but these are essentially time-consuming trial-and-error

studies, and are unlikely to yield significantly shorter aptamers that retain their binding

character. These limitations have motivated our lab group to develop a modified SELEX-

based approach to overcome the limitations of oligonucleotide aptamers panned from a

standard SELEX library.

The goal of developing a modified SELEX protocol was to enable the identification of

short oligonucleotide aptamers. A published technique for analyzing RNA structure212,213

90

provided some inspiration. This technique takes advantage of the ability to design RNA

sequences that self-hybridize into stable hairpins. Single-stranded DNA libraries are

commercially synthesized such the primer sequences are contained within 5’ and 3’

cassettes (Figure 7.3) that flank the random portion of the oligonucleotide sequence. The

hairpins in the cassettes are individually stable secondary structures and thus significantly

diminish the probability of interfering with the secondary structure of the random region.

We opted for random sequences that are 25 or 30 nucleotides long, depending on the

library; this should enable the identification of RNA aptamers no longer than 30

nucleotides that can be isolated from the flanking cassettes and that are well-suited for

synthesizing polyvalent bioconjugates (Figure 7.3). I applied this method in order to

screen large libraries of nuclease-resistant RNA for short aptamer sequences that bind to

ANTXR2, and my colleagues in the Kane group performed screens against CCR5.

Figure 7.3: Schematic representation of the new library design for screening of short oligonucleotide aptamers.

(a) Random region is flanked by 5’ and 3’ cassettes, which have a stem-loop structure where the tetraloop autonomously folds onto itself and eliminates any alternative folding.

Inhibitors composed of aptamer ligands will need to function in vivo, therefore the

conditions used in the binding screens were chosen to match the conditions in the blood

91

serum. A temperature of 37°C and a solution of 20 mM HEPES-NaOH buffer, pH 7.4,

containing 150 mM NaCl, 1.5 mM CaCl2, and 0.5 mM MgCl2 were maintained for each of

the steps in which the binding capacity of the pool was important. In order to protect the

RNA against nuclease degradation, 2’-fluoropyrimidines were used in the in vitro

transcription reactions. Yeast t-RNA was also added to the buffered RNA aptamer pool in

order to reduce the effects of potential non-specific (e.g. electrostatic) RNA binding

interactions during screening.

The folded RNA pool was exposed to one positive and one negative selection during

each round of the screen. The positive screen was against ANTXR2, which was incubated

overnight in protein-adsorbing tubes. Potentially uncoated surfaces were “blocked” by

exposure to a high concentration of bovine serum albumin (BSA) before exposure to the

RNA pool. The presence of BSA necessitated a negative screen against tubes that had only

BSA adsorbed. The negative screen was carried out first, in order to deplete the pool of

aptamers that bound BSA (or that bound nonspecifically).

Each round of screening consisted of in vitro RNA transcription, negative and positive

selections, reverse transcription, and amplification. The two single-stranded DNA libraries

were combined and amplified by polymerase chain reaction (PCR), then transcribed in

vitro (using 2’-fluoropyrimidines) to generate the initial RNA library. The starting RNA pool

was unfolded at 96°C and then allowed to refold at 37°C in the presence of the buffer

described above. The refolded RNA pool was allowed to incubate in the BSA-coated tubes

for 10 minutes, and the supernatant, containing aptamers that did not bind, was then

transferred to the tubes coated with ANTXR2. After a second incubation period, the

ANTR2-coated tubes were extensively washed with buffer to remove unbound or weakly

bound RNA. A 20 mM EDTA dissociation solution was then used to remove the divalent

cations that help give the RNA its secondary structure, thereby eluting the remaining

binders. Phenol-chloroform extraction was used to separate the recovered RNA from any

protein in the solution. Finally, the RNA pool was regenerated by reverse transcription,

PCR amplification and in vitro transcription. The process was repeated for five rounds.

92

7.1.3 Characterization of screened oligonucleotide aptamers

The final pool of aptamer sequences should be enriched with sequences that bind

specifically to the target, in this case ANTXR2. In order to determine which sequences

have been enriched, we have submitted the round 5 aptamer pool and the initial library

for Illumina deep sequencing. We are currently analyzing the sequencing results to

identify sequences which are promising candidates for further testing. Ideally, up to 10

unique aptamer sequences should be ordered and tested for their ability to bind

specifically to ANTXR2 and for their ability to inhibit the binding of PA. The aptamers that

show the best monovalent activity could then be used further for the synthesis of

polyvalent inhibitor constructs.

Modification of the oligonucleotide with an attachment chemistry functionality at

either the 5’ or 3’ end can be accomplished synthetically, by ordering commercially

available modified oligonucleotides. However, in the case of RNA aptamers, which often

include 2’-fluoropyrimidines for nuclease-resistance, even short aptamer sequences can

become quite costly. For this reason, RNA aptamers of interest could be functionalized at

the 5’-end with a modified GMP, using established procedures.214–216 Essentially, RNA

aptamers can be generated by in vitro transcription with T7 RNA polymerase in the

presence of a large molar excess of a modified GMP relative to GTP, such that the GMP is

preferentially incorporated at the 5’-end of the transcript. GMP-S and GMP-NH2 are

commercially available, and I have synthesized a modified GMP-alkyne from GMP using a

published synthesis strategy.217 The availability of these various options should allow for

attachment to a scaffold using whichever chemistry proves to be most convenient for the

exact application at hand.

7.2 Alternative designs for multivalent HIV entry inhibitors

7.2.1 Peptide ligand derivatives of CCL5

The natural ability of the C-C chemokine ligands to inhibit HIV infection has fueled a

number of studies with the aim of adapting CCL5 to improve its therapeutic potential. An

important thrust of much of this research was the attempt to uncouple the

93

agonistic/inflammatory activity of CCL5 from the binding affinity/inhibitory efficacy.218 Of

particular interest to us were the studies aimed at identifying the structural determinants

of CCR5 recognition and antiviral activity.218–220 Although three different research groups

performed similar peptide scanning studies using sequential CCL5-derived peptide

fragments, they produced conflicting reports as to which domains of CCL5 showed

binding activity and HIV inhibitory activity. However, it now seems that the field supports

the work of Lusso and colleagues,218,221–225 who initially identified the N-loop and β1-

strand (residues 11-29, CFAYIARPLPRAHIKEYFY) as the structural determinants of CCL5

that are most important for CCR5 binding and HIV inhibition (Figure 7.4).218,224

Interestingly, they also reported that an analogous region of CCL4 (residues 12-30,

CFSYTARKLPRNFVVDYYE) showed similar binding and inhibitory activity.221 A later study

on the binding surfaces of full-length CCL5 also found a second region they called the 40’

loop (residues 43-48, TRKNRQ) that was close to the key hydrophobic regions of CCL5 and

that was important for binding to the N-terminus of CCR5 (Figure 7.4).225

Figure 7.4: Structure of CCL5 monomer (PDB ID 1HRJ), with regions important for binding CCR5 highlighted.

The sequence of the N-loop / β1-strand, CFAYIARPLPRAHIKEYFY (residues 11-29), is highlighted in yellow and green. Yellow indicates the hydrophobic regions that appear crucial for CCR5 binding. The 40’ loop region, identified as an additional part of the binding surface by Duma et al.225 is highlighted in blue. Reprinted from Chem. Biol. 19, Secchi, M.; Longhi, R.; Vassena, L.; Sironi, F.; Grzesiek, S.; Lusso, P.; Vangelista, L., 1579-1588, 2012, with permission from Elsevier.

94

Since their initial report,218 the researchers have worked extensively to develop their

lead peptide candidate by sequential residue mutations and truncations.221–223 Additional

work on the structural interactions between CCL5 and CCR5 showed that the natural

chemokine ligands and the CCL5-derived peptides bind to the solvent-exposed N-terminal

region of CCR5. Furthermore, when the cysteine at the N-terminus of the peptide ligand

was used to create peptide dimers (via the formation of a disulfide bond), the key binding

surface area was increased. Accordingly, the activity of peptide dimers was enhanced 10-

fold relative to the corresponding monomers.221 However, NMR showed no interactions

between the monomers,222 which suggests that the efficacy enhancement by

dimerization may be purely a function of multivalency. Although Lusso and colleagues

stress the importance of peptide dimer formation for their reported inhibitory activities,

the researchers make no references to investigating the addition of a linker for increased

binding range or incorporating these peptide monomers into higher-order multivalent

ligand conjugates.221

In prior unpublished work, our group (in collaboration with the Gray-Owen group at

the University of Toronto) found that liposomal polyvalent conjugates of one of the initial

lead peptides (Ac-CFAYIARPLPRA-Am) reported by the Lusso group218 showed some

ability to inhibit HIV infection, but the activity was not very impressive, especially for a

polyvalent inhibitor (Figure 7.5). Recently, however, an improved version of the peptide

ligand with the sequence Ac-CFPYITRPGTYHDWWYTRKNRQ was reported to have three

to four orders of magnitude better activity.221 This peptide was able to inhibit HIV

envelope protein-mediated cell fusion with an IC50 of 20 nM and was able to inhibit HIV

infection of peripheral blood mononuclear cells (PBMCs) with an IC50 of 167 nM. Also

noteworthy is the fact that the authors found that removing the C-terminal amidation

improved solubility without affecting activity.221 Since lack of peptide water-solubility is a

factor that we have noticed in our preliminary work with Ac-CFAYIARPLPRA-Am, using

peptides without C-terminal amidation may be especially important for the performance

of our polyvalent inhibitors.

95

Figure 7.5: Inhibition of HIV-1Ba-L infection by Ac-CFAYIARPLPRA-Am-functionalized liposomes.

Infection of MAGI-CCR5 cells by HIV-1Ba-L in the presence of polyvalent liposome inhibitors, at per-peptide concentrations of 10 µM (FAYI 10 µM) and 1 µM (FAYI 1 µM), in the absence of inhibitor (Infected), and in the absence of virus (Uninfected). (Unpublished results obtained by former members of the group in collaboration with the Gray-Owen group at the University of Toronto).

7.2.2 Rationale for the design of divalent CCR5-targeting HIV inhibitors

While we expect that there may be multiple different strategies for creating multivalent

conjugates that are effective at inhibiting HIV entry to some extent, it may be especially

instructive to create divalent conjugates based on the ligands that bind to CCR5. There

are two types of divalent conjugates which may be interesting to study in this context.

Homodivalent conjugates are structures in which two copies of the same ligand are

tethered together, and heterodivalent conjugates consist of two different ligands

attached by a tether (Figure 7.6). Both types of divalent conjugates are likely to exhibit

enhanced binding affinity relative to that of the monovalent ligands separately, which can

often be a worthwhile pursuit in and of itself. That being said, homodivalent (and

polyvalent) conjugates may also be potentially useful for systems in which it may be

convenient or desirable to target the same binding site multiple times in close proximity,

as was the case for the hemagglutinin-binding influenza inhibitors in Chapter 4. On the

other hand, heterodivalent conjugates may be especially useful when it is desirable to

96

achieve greater physical coverage of a receptor (e.g. for inhibiting a protein-protein

interaction) or to achieve greater specificity for a specific receptor out of a class of

receptors with similar structure (e.g. to reduce the potential for side effects).

Figure 7.6: Divalent conjugates for targeting CCR5.

(a) Monovalent ligands targeting separate, distinct binding epitopes on CCR5 (e.g. the N-terminus and ECL2). (b) Heterodivalent conjugate binding two separate, distinct epitopes on CCR5 simultaneously. (c) Homodivalent conjugate binding two separate CCR5 receptors (CCR5 has been found to form constitutive dimers autonomously on the cell surface).

In the context of targeting CCR5 for HIV inhibition, both types of divalent inhibitors

would be interesting to study. For example, the constitutive dimerization of CCR5 on the

cell surface is a documented phenomenon,226,227 which would clearly enhance the binding

of a divalent ligand. Further clustering, aided by a homodivalent inhibitor, may lead to

internalization of the receptor and depletion of the CCR5 from the cell surfaces, which

could produce an additional downstream inhibitory effect.186,227,228 Heterodivalent

inhibitors, comprised of ligands that bind to more than one of the extracellular regions of

CCR5, may be more effective than monovalent inhibitors because of several reasons. Such

inhibitors are expected to exhibit increased affinity for CCR5 and also a larger steric

hindrance effect, both of which should increase the effectiveness of the inhibitor for

competing with the binding of gp120 envelope protein. Furthermore, heterodivalent

inhibitors that bind to multiple regions of CCR5 may exhibit inhibition against a broader

range of HIV subtypes, including those that may evolve resistance to one of the

monomeric ligands. Perhaps even more interesting is the fact that comparing the

efficacies of homodivalent and heterodivalent inhibitors of HIV infection may provide

instruction as to which of these modes of inhibition is more important for the optimal

inhibitor design. The additional level of understanding of the processes that contribute to

97

HIV entry inhibition that can be gained by comparing and contrasting these two types of

divalent inhibitors will undoubtedly be important for the intelligent design of conjugates

with higher levels of multivalency or complexity. For these reasons, I have completed

preliminary work on the synthesis and characterization both of homodivalent and

heterodivalent conjugates (see Figure 7.6), as well as polyvalent conjugates of the various

ligands identified in Section 7.2.1 above.

7.2.3 Synthesis of homodivalent bioconjugates of aptamers that bind CCR5

In preliminary work using the model 5’-thiol-functionalized ssDNA sequences first

described in Section 6.1, I have confirmed that the protocol for conjugation to a

homobifunctional PEG linker can be applied for thiol-functionalized oligonucleotide

aptamers. The typical conjugation strategy is depicted in Figure 7.7 below.

Figure 7.7: Synthesis scheme for homodivalent PEG-aptamer bioconjugates.

The PEG linker used for the conjugation optimization experiment was a maleimide-(EG)11-

maleimide linker. The purified, de-protected ssDNA was then added to the reaction buffer

at a 2.2:1 molar ratio relative to the linker. The reaction buffer is the same as for the

synthesis of homodivalent PEG-protein conjugates described in Section 4.2.4, degassed

pH 6.5 MBS with 10 mM EDTA at ambient temperature. The concentration of the ssDNA

in the reaction buffer was adjusted to 200 µM. As shown in Figure 7.8, under these

conditions, there is evidence for nearly complete divalent conjugation within 3.5 days.

98

Figure 7.8: SYBR Gold-stained 15% acrylamide TBE-Urea PAGE of ssDNA model aptamer homodivalent PEG conjugation reactions after 88 hours at ambient temperature.

Lanes left to right: ssDNA ladder (27, 53, 75, 101mers), ssDNA-thiol oligomer reagent stock after deprotection and purification through BioRad spin column, conjugation reaction mixture after 88 hours at ambient temperature. The bands in the conjugation reaction lane from top to bottom are presumably the homodivalent conjugate, disulfide-bonded (ssDNA-SS-ssDNA) divalent oligomers, monovalent conjugate, and ssDNA-SS-cap that remained protected. There appears to be no remaining unreacted deprotected ssDNA. Stained with SYBR Gold.

7.2.4 Synthesis of heterodivalent inhibitors that bind CCR5

CCR5-targeting heterodivalent inhibitors can be created by the combination of two site-

specific ligands (either peptides or oligonucleotides) connected by an appropriately-sized

PEG linker (Figure 7.6 and Figure 7.9). For optimizing reaction conditions for

heterodivalent inhibitors, I chose to experiment with the heterobifunctional linker

sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (sulfo-SMCC) as

proof of concept. Sulfo-SMCC is a commonly used, water-soluble, short (8.3 Å) spacer

with amine and thiol reactivity (sulfo-NHS and maleimide functionalities). The same

deprotection protocol and buffer conditions as used for homodivalent conjugation were

used here, and both 5’-thiol and 5’-amine functional ssDNA oligonucleotides were added

into the reaction buffer with sulfo-SMCC, in a 1:1:1 molar ratio. The formation of the

heterodivalent conjugate was confirmed by mass spectrometry. These kinds of divalent

99

conjugates will be interesting in future studies if multiple aptamers against the same

screening target are identified, especially if it can be determined that the aptamers bind

to different regions on the same target molecule.

Figure 7.9: Synthesis scheme for heterodivalent PEG-aptamer bioconjugates.

(a) aptamer 1 (5’-amine), aptamer 2 (5’-thiol), TCEP, NaCl, pH 6.5 MBS.

7.3 Bioengineered protein polymer scaffolds

7.3.1 Using proteins as monodisperse polymer scaffolds

Monodisperse, functional polymer scaffolds are readily available from nature in the form

of proteins. Amine (lysine), carboxylic acid (aspartic acid and glutamic acid), thiol

(cysteine), and hydroxyl (serine and threonine) are naturally occurring functional handles

on the side chains of peptides. The research of Yang et al.229 demonstrated as a proof of

concept the ability to acylate lysine side chains with acetic anhydride in three different

proteins: ubiquitin (8.6 kDa, 7 lysines, 1 N-terminal amine), lysozyme (14.3 kDa, 6 lysines,

1 N-terminal amine), and bovine carbonic anhydrase II (BCA) (29 kDa, 18 lysines, post-

translationally acetylated N- terminal amine). The ability to acylate the lysine side chains

depended on the size of the protein; denaturation with SDS was required for complete

acylation of the lysine residues in BCA, but there was no reported difference between the

native or denatured reactivity of the lysine residues in ubiquitin. Furthermore, the ability

to react ubiquitin with hydrophilic triethylene glycol carboxylate, hydrophobic benzoate,

anionic glutarate, and chemically reactive iodoacetate demonstrated both the generality

of the procedure and the range of chemical properties that might be introduced into

protein scaffolds. The approach is interesting, but the propensity for the proteins to fold

into tertiary structures should still be considered.

O

n

Na

O

O

O

O

N

O

O

O

n

N

O

O

HN

O

aptamer S

aptamer

100

7.3.2 Bioengineering scaffolds for precise control over multivalent architecture

Recombinant protein engineering methods can be used to take advantage of nature’s

polymer synthesis machinery in order to create scaffolds that are both monodisperse and

designed to have functional handles at predetermined locations.230 The Kiick research

group has employed these methods95,231,232 to make well-defined scaffolds into

polyvalent inhibitors of the cholera toxin B pentamer subunit. The researchers

demonstrated that by altering the composition of the polypeptide, they could control the

inter-ligand spacing, the valency, and even the conformation (α-helical versus random-

coil) of the inhibitors. In addition to observing improvements in inhibitors with inter-

ligand spacing and valency that matched the target toxin, the results revealed more

potent inhibition using α-helical scaffolds than random-coil scaffolds. The authors

hypothesized that the α-helical conformation may have increased the accessibility of the

pendant ligands and/or reduced the loss of conformational entropy upon binding, thus

leading to the measured improvement in potency. Using oligovalent proteins as

supramolecular templates for pre-arranging the displays of heterobifunctional ligands on

a polyvalent polymer scaffold have also been shown to yield vast improvements in

potency.233,234

As a demonstration of the power of using bioengineering techniques to design

multivalent binders, we have designed and tested monodisperse anthrax toxin inhibitors

based on polypeptide scaffolds. A manuscript describing this research185 was very

recently accepted for publication in Angewandte Chemie (see Appendix). Specifically, my

colleagues in the Kane group created polypeptides with the sequence

H10(SE)[HTSTYWWLDGAP(SE5)]n(SE)4, in which an (SE)5 linker separates adjacent copies of

the peptide HTSTYWWLDGAP that binds to the heptameric receptor-binding subunit of

anthrax toxin. These multivalent inhibitors were expressed in E. coli and purified using

immobilized metal affinity chromatography. The activity of the best polypeptide inhibitor

(n=8) had an IC50 of ~2 nM on a per-peptide basis.

We have also designed and begun expression and purification of well-defined

polypeptide scaffolds with the sequence [(GE)mK]n(GE)mH10 and [(SE)mK]n(SE)mH10,

101

where the K (lysine) residues will serve as reactive sites for ligand conjugation (Fig 5).

(GE)m or (SE)m serve as linkers whose length can be tuned by varying the value of “m”, “n”

represents the number of reactive sites in the scaffold (and provides control over the

valency of the multivalent ligand), and the terminal decahistidine sequence will help with

purification. Our prior results and a recent thermodynamic model emphasize the

importance of the root-mean-square (RMS) end-to-end distance of the linker as a design

parameter. The RMS end-to-end distance for the linker ((GE)m or (SE)m), however, will be

significantly larger in multivalent conjugates with bulky protein ligands (e.g., HB36.5 or

LukE) than in “free” unconjugated polypeptide scaffolds due to steric effects, particularly

for short linkers with an extended length comparable to the diameter of the protein

ligand. We have therefore initially designed scaffolds with 3 values of m – 10, 20, and 30

– corresponding to linkers composed of 20, 40, and 60 amino acids respectively. The

smallest value of m corresponds to a linker with an extended length between 7 and 8 nm.

Our simulations indicate that the intermediate linkers (m=20) would have RMS end-to-

end distances of ~6 nm, whereas the longest linkers (m=30) would have RMS end-to-end

distances of ~8.5 nm and an extended length of more than 21 nm. Three different values

for the valency, n, were initially chosen: 10, 20, and 30. These and other bioengineered

scaffolds should prove extremely useful for elucidating the most efficacious designs of

future polyvalent bioconjugates.

102

8. Conclusions

As we have shown in extensive work on the inhibition of a variety of different pathogenic

processes, multivalency can be a powerful tool for enhancing the efficacy of bioactive

molecules. This document specifically described the design principles that guided the

synthesis of multivalent biomolecular conjugates with enhanced activity in three different

applications with broad public interest: the inhibition of anthrax intoxication, the

inhibition of influenza infection, and the inhibition of HIV infection. Each of these

inhibitors has been designed to target a stage in the pathogenesis that is conserved or

that will not engender strong selection pressure. The protocols optimized during the

course of these studies could be used to synthesize the next-generation of multivalent

therapeutics, which may use a wide variety of new binding ligands, including

oligonucleotide aptamers. The successful use of protein engineering to bioengineer

scaffolds that have built-in chemical functionality, allowing for the attachment of

bioactive molecules at precisely defined locations and spacing, will be a crucial tool for

synthesizing bioconjugates that can clarify the complex mechanisms behind the actions

of various biological systems. By using biocompatible materials and methods for

bioconjugate synthesis such as those described here, there will be great potential for

clinical use of multivalent therapeutics in the near future.

103

REFERENCES

(1) Martin, J. T.; Kane, R. S. Design of Polyvalent Polymer Therapeutics. In Functional Polymers by Post-Polymerization Modification; Theato, P.; Klok, H.-A., Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2013; pp. 267–290.

(2) Coleman, M. E.; Thran, B.; Morse, S. S.; Hugh-Jones, M.; Massulik, S. Biosecur. Bioterror. 2008, 6, 147.

(3) Jernigan, D. B.; Raghunathan, P. L.; Bell, B. P.; Brechner, R.; Bresnitz, E. A.; Butler, J. C.; Cetron, M.; Cohen, M.; Doyle, T.; Fischer, M.; Greene, C.; Griffith, K. S.; Guarner, J.; Hadler, J. L.; Hayslett, J. A.; Meyer, R.; Petersen, L. R.; Phillips, M.; Pinner, R.; Popovic, T.; Quinn, C. P.; Reefhuis, J.; Reissman, D.; Rosenstein, N.; Schuchat, A.; Shieh, W.-J.; Siegal, L.; Swerdlow, D. L.; Tenover, F. C.; Traeger, M.; Ward, J. W.; Weisfuse, I.; Wiersma, S.; Yeskey, K.; Zaki, S.; Ashford, D. A.; Perkins, B. A.; Ostroff, S.; Hughes, J.; Fleming, D.; Koplan, J. P.; Gerberding, J. L. Emerg. Infect. Dis. 2002, 8, 1019.

(4) Makino, S.; Uchida, I.; Terakado, N.; Sasakawa, C.; Yoshikawa, M. J. Bacteriol. 1989, 171, 722.

(5) Collier, R. J.; Young, J. A. T. Annu. Rev. Cell Dev. Biol. 2003, 19, 45.

(6) Rainey, G. J. A.; Young, J. A. T. Nat. Rev. Microbiol. 2004, 2, 721.

(7) Meyerhoff, A.; Albrecht, R.; Meyer, J. M.; Dionne, P.; Higgins, K.; Murphy, D. Clin. Infect. Dis. 2004, 39, 303.

(8) Dixon, T. C.; Meselson, M.; Guillemin, J.; Hanna, P. C. N. Engl. J. Med. 1999, 341, 815.

(9) Abrami, L.; Lindsay, M.; Parton, R. G.; Leppla, S. H.; van der Goot, F. G. J. Cell Biol. 2004, 166, 645.

(10) Rainey, G. J. A.; Wigelsworth, D. J.; Ryan, P. L.; Scobie, H. M.; Collier, R. J.; Young, J. A. T. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 13278.

(11) Krantz, B. A.; Melnyk, R. A.; Zhang, S.; Juris, S. J.; Lacy, D. B.; Wu, Z.; Finkelstein, A.; Collier, R. J. Science 2005, 309, 777.

(12) Pentelute, B. L.; Sharma, O.; Collier, R. J. Angew. Chem. Int. Ed. Engl. 2011, 50, 2294.

104

(13) Bradley, K. A.; Mogridge, J.; Mourez, M.; Collier, R. J.; Young, J. A. Nature 2001, 414, 225.

(14) Scobie, H. M.; Rainey, G. J. A.; Bradley, K. A.; Young, J. A. T. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 5170.

(15) Bonuccelli, G.; Sotgia, F.; Frank, P. G.; Williams, T. M.; de Almeida, C. J.; Tanowitz, H. B.; Scherer, P. E.; Hotchkiss, K. A.; Terman, B. I.; Rollman, B.; Alileche, A.; Brojatsch, J.; Lisanti, M. P. Am. J. Physiol. Cell Physiol. 2005, 288, C1402.

(16) Molloy, S. S.; Bresnahan, P. A.; Leppla, S. H.; Klimpel, K. R.; Thomas, G. J. Biol. Chem. 1992, 267, 16396.

(17) Gordon, V. M.; Klimpel, K. R.; Arora, N.; Henderson, M. A.; Leppla, S. H. Infect. Immun. 1995, 63, 82.

(18) Petosa, C.; Collier, R. J.; Klimpel, K. R.; Leppla, S. H.; Liddington, R. C. Nature 1997, 385, 833.

(19) Kintzer, A. F.; Thoren, K. L.; Sterling, H. J.; Dong, K. C.; Feld, G. K.; Tang, I. I.; Zhang, T. T.; Williams, E. R.; Berger, J. M.; Krantz, B. A. J. Mol. Biol. 2009, 392, 614.

(20) Milne, J. C.; Furlong, D.; Hanna, P. C.; Wall, J. S.; Collier, R. J. J. Biol. Chem. 1994, 269, 20607.

(21) Ezzell, J. W.; Abshire, T. G. J. Gen. Microbiol. 1992, 138, 543.

(22) Mogridge, J.; Cunningham, K.; Lacy, D. B.; Mourez, M.; Collier, R. J. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 7045.

(23) Abrami, L.; Liu, S.; Cosson, P.; Leppla, S. H.; van der Goot, F. G. J. Cell Biol. 2003, 160, 321.

(24) Krantz, B. A.; Finkelstein, A.; Collier, R. J. J. Mol. Biol. 2006, 355, 968.

(25) Basilio, D.; Juris, S. J.; Collier, R. J.; Finkelstein, A. J. Gen. Physiol. 2009, 133, 307.

(26) Firoved, A. M.; Miller, G. F.; Moayeri, M.; Kakkar, R.; Shen, Y.; Wiggins, J. F.; McNally, E. M.; Tang, W.-J.; Leppla, S. H. Am. J. Pathol. 2005, 167, 1309.

(27) Lovchik, J. A.; Drysdale, M.; Koehler, T. M.; Hutt, J. A.; Lyons, C. R. Infect. Immun. 2012, 80, 2414.

105

(28) Cui, X.; Moayeri, M.; Li, Y.; Li, X.; Haley, M.; Fitz, Y.; Correa-Araujo, R.; Banks, S. M.; Leppla, S. H.; Eichacker, P. Q. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2004, 286, R699.

(29) Moayeri, M.; Haines, D.; Young, H. A.; Leppla, S. H. J. Clin. Invest. 2003, 112, 670.

(30) Guo, Q.; Shen, Y.; Zhukovskaya, N. L.; Florián, J.; Tang, W.-J. J. Biol. Chem. 2004, 279, 29427.

(31) Leppla, S. H. Proc. Natl. Acad. Sci. 1982, 79, 3162.

(32) O’Brien, J.; Friedlander, A.; Dreier, T.; Ezzell, J.; Leppla, S. Infect. Immun. 1985, 47, 306.

(33) Hoover, D. L.; Friedlander, A. M.; Rogers, L. C.; Yoon, I. K.; Warren, R. L.; Cross, A. S. Infect. Immun. 1994, 62, 4432.

(34) Vitale, G.; Bernardi, L.; Napolitani, G.; Mock, M.; Montecucco, C. Biochem. J. 2000, 352, 739.

(35) Duesbery, N. S.; Webb, C. P.; Leppla, S. H.; Gordon, V. M.; Klimpel, K. R.; Copeland, T. D.; Ahn, N. G.; Oskarsson, M. K.; Fukasawa, K.; Paull, K. D.; Vande Woude, G. F. Science 1998, 280, 734.

(36) Banks, D. J.; Ward, S. C.; Bradley, K. A. Expert Rev. Mol. Med. 2006, 8, 1.

(37) Turk, B. E. Biochem. J. 2007, 402, 405.

(38) Baldari, C. T.; Tonello, F.; Paccani, S. R.; Montecucco, C. Trends Immunol. 2006, 27, 434.

(39) Suzuki, Y. Biol. Pharm. Bull. 2005, 28, 399.

(40) Harris, A.; Cardone, G.; Winkler, D. C.; Heymann, J. B.; Brecher, M.; White, J. M.; Steven, A. C. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 19123.

(41) Rossman, J. S.; Lamb, R. A. Virology 2011, 411, 229.

(42) Webster, R. G.; Bean, W. J.; Gorman, O. T.; Chambers, T. M.; Kawaoka, Y. Microbiol. Rev. 1992, 56, 152.

(43) Zambon, M. C. Rev. Med. Virol. 2001, 11, 227.

(44) Bouvier, N. M.; Palese, P. Vaccine 2008, 26, D49.

106

(45) Calder, L. J.; Wasilewski, S.; Berriman, J. A.; Rosenthal, P. B. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 10685.

(46) Noda, T.; Sagara, H.; Yen, A.; Takada, A.; Kida, H.; Cheng, R. H.; Kawaoka, Y. Nature 2006, 439, 490.

(47) Smith, D. J.; Lapedes, A. S.; de Jong, J. C.; Bestebroer, T. M.; Rimmelzwaan, G. F.; Osterhaus, A. D. M. E.; Fouchier, R. A. M. Science 2004, 305, 371.

(48) Garten, R. J.; Davis, C. T.; Russell, C. A.; Shu, B.; Lindstrom, S.; Balish, A.; Sessions, W. M.; Xu, X.; Skepner, E.; Deyde, V.; Okomo-Adhiambo, M.; Gubareva, L.; Barnes, J.; Smith, C. B.; Emery, S. L.; Hillman, M. J.; Rivailler, P.; Smagala, J.; de Graaf, M.; Burke, D. F.; Fouchier, R. A. M.; Pappas, C.; Alpuche-Aranda, C. M.; López-Gatell, H.; Olivera, H.; López, I.; Myers, C. A.; Faix, D.; Blair, P. J.; Yu, C.; Keene, K. M.; Dotson, P. D.; Boxrud, D.; Sambol, A. R.; Abid, S. H.; St. George, K.; Bannerman, T.; Moore, A. L.; Stringer, D. J.; Blevins, P.; Demmler-Harrison, G. J.; Ginsberg, M.; Kriner, P.; Waterman, S.; Smole, S.; Guevara, H. F.; Belongia, E. A.; Clark, P. A.; Beatrice, S. T.; Donis, R.; Katz, J.; Finelli, L.; Bridges, C. B.; Shaw, M.; Jernigan, D. B.; Uyeki, T. M.; Smith, D. J.; Klimov, A. I.; Cox, N. J. Science 2009, 325, 197.

(49) Global report: UNAIDS report on the global AIDS epidemic 2012; Joint United Nations Programme on HIV/AIDS (UNAIDS): Geneva, Switzerland, 2012.

(50) Eckert, D. M.; Kim, P. S. Annu. Rev. Biochem. 2001, 70, 777.

(51) Garg, R.; Gupta, S. P.; Gao, H.; Babu, M. S.; Debnath, A. K.; Hansch, C. Chem. Rev. 1999, 99, 3525.

(52) Berger, E. A.; Murphy, P. M.; Farber, J. M. Annu. Rev. Immunol. 1999, 17, 657.

(53) Doms, R. W.; Peiper, S. C. Virology 1997, 235, 179.

(54) Siciliano, S. J.; Kuhmann, S. E.; Weng, Y.; Madani, N.; Springer, M. S.; Lineberger, J. E.; Danzeisen, R.; Miller, M. D.; Kavanaugh, M. P.; DeMartino, J. A.; Kabat, D. J. Biol. Chem. 1999, 274, 1905.

(55) Jacque, J.-M.; Triques, K.; Stevenson, M. Nature 2002, 418, 435.

(56) Pomerantz, R. J. Nat. Med. 2002, 8, 659.

(57) Joshi, D.; O’Grady, J.; Dieterich, D.; Gazzard, B.; Agarwal, K. Lancet 2011, 377, 1198.

(58) Letvin, N. L. Science 2009, 326, 1196.

107

(59) Baltimore, D. Science 2002, 296, 2297.

(60) Richman, D. D.; Little, S. J.; Smith, D. M.; Wrin, T.; Petropoulos, C.; Wong, J. K. Trans. Am. Clin. Climatol. Assoc. 2004, 115, 289.

(61) Munier, C. M. L.; Andersen, C. R.; Kelleher, A. D. Drugs 2011, 71, 387.

(62) Uberla, K. PLoS Pathog. 2008, 4, e1000114.

(63) Vaccari, M.; Poonam, P.; Franchini, G. Expert Rev. Vaccines 2010, 9, 997.

(64) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew. Chemie Int. Ed. 1998, 37, 2754.

(65) Rao, J.; Lahiri, J.; Isaacs, L.; Weis, R. M.; Whitesides, G. M. Science 1998, 280, 708.

(66) Kiessling, L. L.; Pohl, N. L. Chem. Biol. 1996, 3, 71.

(67) Kiessling, L. L.; Strong, L. E.; Gestwicki, J. E. Annu. Rep. Med. Chem. 2000, 35, 321.

(68) Mulder, A.; Huskens, J.; Reinhoudt, D. N. Org. Biomol. Chem. 2004, 2, 3409.

(69) Badjić, J. D.; Nelson, A.; Cantrill, S. J.; Turnbull, W. B.; Stoddart, J. F. Acc. Chem. Res. 2005, 38, 723.

(70) Kiessling, L. L.; Gestwicki, J. E.; Strong, L. E. Angew. Chem. Int. Ed. Engl. 2006, 45, 2348.

(71) Joshi, A.; Vance, D.; Rai, P.; Thiyagarajan, A.; Kane, R. S. Chemistry 2008, 14, 7738.

(72) Vance, D.; Shah, M.; Joshi, A.; Kane, R. S. Biotechnol. Bioeng. 2008, 101, 429.

(73) Vance, D.; Martin, J.; Patke, S.; Kane, R. S. Adv. Drug Deliv. Rev. 2009, 61, 931.

(74) Maheshwari, G.; Brown, G.; Lauffenburger, D. A.; Wells, A.; Griffith, L. G. J. Cell Sci. 2000, 113, 1677.

(75) Wall, S. T.; Saha, K.; Ashton, R. S.; Kam, K. R.; Schaffer, D. V; Healy, K. E. Bioconjug. Chem. 2008, 19, 806.

(76) Matrosovich, M. N.; Mochalova, L. V; Marinina, V. P.; Byramova, N. E.; Bovin, N. V. FEBS Lett. 1990, 272, 209.

(77) Spevak, W.; Nagy, J. O.; Charych, D. H.; Schaefer, M. E.; Gilbert, J. H.; Bednarski, M. D. J. Am. Chem. Soc. 1993, 115, 1146.

108

(78) Mammen, M.; Dahmann, G.; Whitesides, G. M. J. Med. Chem. 1995, 38, 4179.

(79) Choi, S. K.; Mammen, M.; Whitesides, G. M. Chem. Biol. 1996, 3, 97.

(80) Sigal, G. B.; Mammen, M.; Dahmann, G.; Whitesides, G. M. J. Am. Chem. Soc. 1996, 118, 3789.

(81) Kamitakahara, H.; Suzuki, T.; Nishigori, N.; Suzuki, Y.; Kanie, O.; Wong, C.-H. Angew. Chemie Int. Ed. 1998, 37, 1524.

(82) Honda, T.; Yoshida, S.; Arai, M.; Masuda, T.; Yamashita, M. Bioorg. Med. Chem. Lett. 2002, 12, 1929.

(83) Fan, E.; Zhang, Z.; Minke, W. E.; Hou, Z.; Verlinde, C. L. M. J.; Hol, W. G. J. J. Am. Chem. Soc. 2000, 122, 2663.

(84) Kitov, P. I.; Sadowska, J. M.; Mulvey, G.; Armstrong, G. D.; Ling, H.; Pannu, N. S.; Read, R. J.; Bundle, D. R. Nature 2000, 403, 669.

(85) Gargano, J. M.; Ngo, T.; Kim, J. Y.; Acheson, D. W.; Lees, W. J. J. Am. Chem. Soc. 2001, 123, 12909.

(86) Mourez, M.; Kane, R. S.; Mogridge, J.; Metallo, S.; Deschatelets, P.; Sellman, B. R.; Whitesides, G. M.; Collier, R. J. Nat. Biotechnol. 2001, 19, 958.

(87) Nishikawa, K.; Matsuoka, K.; Kita, E.; Okabe, N.; Mizuguchi, M.; Hino, K.; Miyazawa, S.; Yamasaki, C.; Aoki, J.; Takashima, S.; Yamakawa, Y.; Nishijima, M.; Terunuma, D.; Kuzuhara, H.; Natori, Y. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 7669.

(88) Mulvey, G. L.; Marcato, P.; Kitov, P. I.; Sadowska, J.; Bundle, D. R.; Armstrong, G. D. J. Infect. Dis. 2003, 187, 640.

(89) Karginov, V. A.; Nestorovich, E. M.; Moayeri, M.; Leppla, S. H.; Bezrukov, S. M. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 15075.

(90) Basha, S.; Rai, P.; Poon, V.; Saraph, A.; Gujraty, K.; Go, M. Y.; Sadacharan, S.; Frost, M.; Mogridge, J.; Kane, R. S. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 13509.

(91) Gujraty, K. V; Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Biomacromolecules 2006, 7, 2082.

(92) Joshi, A.; Saraph, A.; Poon, V.; Mogridge, J.; Kane, R. S. Bioconjug. Chem. 2006, 17, 1265.

109

(93) Polizzotti, B. D.; Kiick, K. L. Biomacromolecules 2006, 7, 483.

(94) Rai, P.; Padala, C.; Poon, V.; Saraph, A.; Basha, S.; Kate, S.; Tao, K.; Mogridge, J.; Kane, R. S. Nat. Biotechnol. 2006, 24, 582.

(95) Polizzotti, B. D.; Maheshwari, R.; Vinkenborg, J.; Kiick, K. L. Macromolecules 2007, 40, 7103.

(96) Rai, P. R.; Saraph, A.; Ashton, R.; Poon, V.; Mogridge, J.; Kane, R. S. Angew. Chem. Int. Ed. Engl. 2007, 46, 2207.

(97) Carlson, C. B.; Mowery, P.; Owen, R. M.; Dykhuizen, E. C.; Kiessling, L. L. ACS Chem. Biol. 2007, 2, 119.

(98) Ohlson, S. Drug Discov. Today 2008, 13, 433.

(99) Lutz, J.-F.; Börner, H. G. Prog. Polym. Sci. 2008, 33, 1.

(100) Hartmann, L.; Börner, H. G. Adv. Mater. 2009, 21, 3425.

(101) Hawker, C. J.; Wooley, K. L. Science 2005, 309, 1200.

(102) Börner, H. G. Macromol. Rapid Commun. 2011, 32, 115.

(103) Dehn, S.; Chapman, R.; Jolliffe, K.; Perrier, S. Polym. Rev. 2011, 51, 214.

(104) Canalle, L. A.; Löwik, D. W. P. M.; van Hest, J. C. M. Chem. Soc. Rev. 2010, 39, 329.

(105) Gauthier, M. A.; Klok, H.-A. Chem. Commun. (Camb). 2008, 2591.

(106) Klok, H.-A. Macromolecules 2009, 42, 7990.

(107) Le Droumaguet, B.; Nicolas, J. Polym. Chem. 2010, 1, 563.

(108) Gauthier, M. A.; Gibson, M. I.; Klok, H.-A. Angew. Chem. Int. Ed. Engl. 2009, 48, 48.

(109) Yanjarappa, M. J.; Gujraty, K. V; Joshi, A.; Saraph, A.; Kane, R. S. Biomacromolecules 2006, 7, 1665.

(110) Gujraty, K. V; Yanjarappa, M. J.; Saraph, A.; Joshi, A.; Mogridge, J.; Kane, R. S. J. Polym. Sci. A. Polym. Chem. 2008, 46, 7246.

(111) Pieters, R. J. Org. Biomol. Chem. 2009, 7, 2013.

110

(112) Top, A.; Kiick, K. L. Adv. Drug Deliv. Rev. 2010, 62, 1530.

(113) Duncan, R. Nat. Rev. Drug Discov. 2003, 2, 347.

(114) Twaites, B.; de las Heras Alarcón, C.; Alexander, C. J. Mater. Chem. 2005, 15, 441.

(115) Kiick, K. L. Science 2007, 317, 1182.

(116) Liu, S.; Maheshwari, R.; Kiick, K. L. Macromolecules 2009, 42, 3.

(117) Gordon, E. J.; Gestwicki, J. E.; Strong, L. E.; Kiessling, L. L. Chem. Biol. 2000, 7, 9.

(118) Kopecek, J.; Kopecková, P.; Minko, T.; Lu, Z. Eur. J. Pharm. Biopharm. 2000, 50, 61.

(119) Kopecek, J.; Kopecková, P. Adv. Drug Deliv. Rev. 2010, 62, 122.

(120) Li, C. Adv. Drug Deliv. Rev. 2002, 54, 695.

(121) Totani, K.; Kubota, T.; Kuroda, T.; Murata, T.; Hidari, K. I.-P. J.; Suzuki, T.; Suzuki, Y.; Kobayashi, K.; Ashida, H.; Yamamoto, K.; Usui, T. Glycobiology 2003, 13, 315.

(122) Zeng, X.; Murata, T.; Kawagishi, H.; Usui, T.; Kobayashi, K. Biosci. Biotechnol. Biochem. 1998, 62, 1171.

(123) Sunderland, C. J.; Steiert, M.; Talmadge, J. E.; Derfus, A. M.; Ã, S. E. B. Drug Dev. Res. 2006, 67, 70.

(124) Natarajan, A.; Xiong, C.-Y.; Albrecht, H.; DeNardo, G. L.; DeNardo, S. J. Bioconjug. Chem. 2005, 16, 113.

(125) Kurfürst, M. M. Anal. Biochem. 1992, 200, 244.

(126) Kastenholz, B. Protein Pept. Lett. 2006, 13, 503.

(127) Ly, M.; Wang, Z.; Laremore, T. N.; Zhang, F.; Zhong, W.; Pu, D.; Zagorevski, D. V; Dordick, J. S.; Linhardt, R. J. Anal. Bioanal. Chem. 2011, 399, 737.

(128) Seelert, H.; Krause, F. Electrophoresis 2008, 29, 2617.

(129) Joshi, A.; Kate, S.; Poon, V.; Mondal, D.; Boggara, M. B.; Saraph, A.; Martin, J. T.; McAlpine, R.; Day, R.; Garcia, A. E.; Mogridge, J.; Kane, R. S. Biomacromolecules 2011, 12, 791.

(130) Young, J. A. T.; Collier, R. J. Annu. Rev. Biochem. 2007, 76, 243.

111

(131) Moayeri, M.; Robinson, T. M.; Leppla, S. H.; Karginov, V. A. Antimicrob. Agents Chemother. 2008, 52, 2239.

(132) Maynard, J. A.; Maassen, C. B. M.; Leppla, S. H.; Brasky, K.; Patterson, J. L.; Iverson, B. L.; Georgiou, G. Nat. Biotechnol. 2002, 20, 597.

(133) Numa, M. M. D.; Lee, L. V; Hsu, C.-C.; Bower, K. E.; Wong, C.-H. Chembiochem 2005, 6, 1002.

(134) Panchal, R. G.; Hermone, A. R.; Nguyen, T. L.; Wong, T. Y.; Schwarzenbacher, R.; Schmidt, J.; Lane, D.; McGrath, C.; Turk, B. E.; Burnett, J.; Aman, M. J.; Little, S.; Sausville, E. A.; Zaharevitz, D. W.; Cantley, L. C.; Liddington, R. C.; Gussio, R.; Bavari, S. Nat. Struct. Mol. Biol. 2004, 11, 67.

(135) Sellman, B. R.; Mourez, M.; Collier, R. J. Science 2001, 292, 695.

(136) Ragle, B. E.; Karginov, V. A.; Bubeck Wardenburg, J. Antimicrob. Agents Chemother. 2010, 54, 298.

(137) Fan, E.; Merritt, E. A.; Verlinde, C. L.; Hol, W. G. Curr. Opin. Struct. Biol. 2000, 10, 680.

(138) Szejtli, J. Chem. Rev. 1998, 98, 1743.

(139) Gu, L. Q.; Braha, O.; Conlan, S.; Cheley, S.; Bayley, H. Nature 1999, 398, 686.

(140) Liao, K.-C.; Mogridge, J. Infect. Immun. 2009, 77, 4455.

(141) Kolb, H. C.; Finn, M.; Sharpless, K. B. Angew. Chemie Int. Ed. 2001, 40, 2004.

(142) Rostovtsev, V. V; Green, L. G.; Fokin, V. V; Sharpless, K. B. Angew. Chem. Int. Ed. Engl. 2002, 41, 2596.

(143) Tornøe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67, 3057.

(144) Lutz, J.-F. Angew. Chem. Int. Ed. Engl. 2007, 46, 1018.

(145) Gujraty, K.; Sadacharan, S.; Frost, M.; Poon, V.; Kane, R. S.; Mogridge, J. Mol. Pharm. 2005, 2, 367.

(146) Kane, R. S. Langmuir 2010, 26, 8636.

(147) Kramer, R. H.; Karpen, J. W. Nature 1998, 395, 710.

112

(148) Krishnamurthy, V. M.; Semetey, V.; Bracher, P. J.; Shen, N.; Whitesides, G. M. J. Am. Chem. Soc. 2007, 129, 1312.

(149) Knoll, D.; Hermans, J. J. Biol. Chem. 1983, 258, 5710.

(150) Diestler, D.; Knapp, E. Phys. Rev. Lett. 2008, 100, 1.

(151) Centers for Disease Control and Prevention. Early estimates of seasonal influenza vaccine effectiveness--United States, January 2013; MMWR 2013; Vol. 62; U.S. Department of Health and Human Services: Atlanta, GA, 2013.

(152) Osterholm, M. T.; Kelley, N. S.; Sommer, A.; Belongia, E. A. Lancet Infect. Dis. 2012, 12, 36.

(153) Cohen, C.; White, J. M.; Savage, E. J.; Glynn, J. R.; Choi, Y.; Andrews, N.; Brown, D.; Ramsay, M. E. Emerg. Infect. Dis. 2007, 13, 12.

(154) Centers for Disease Control and Prevention. Prevention and control of seasonal influenza with vaccines. Recommendations of the Advisory Committee on Immunization Practices--United States, 2013-2014; MMWR 2013; Vol. 62; U.S. Department of Health and Human Services: Atlanta, GA, 2013.

(155) Laursen, N. S.; Wilson, I. A. Antiviral Res. 2013, 98, 476.

(156) Lee, P. S.; Yoshida, R.; Ekiert, D. C.; Sakai, N.; Suzuki, Y.; Takada, A.; Wilson, I. A. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 17040.

(157) Ekiert, D. C.; Bhabha, G.; Elsliger, M.; Friesen, R. H. E.; Jongeneelen, M.; Throsby, M.; Goudsmit, J.; Wilson, I. A. Science 2009, 324, 246.

(158) Sui, J.; Hwang, W. C.; Perez, S.; Wei, G.; Aird, D.; Chen, L.; Santelli, E.; Stec, B.; Cadwell, G.; Ali, M.; Wan, H.; Murakami, A.; Yammanuru, A.; Han, T.; Cox, N. J.; Bankston, L. A.; Donis, R. O.; Liddington, R. C.; Marasco, W. A. Nat. Struct. Mol. Biol. 2009, 16, 265.

(159) Throsby, M.; van den Brink, E.; Jongeneelen, M.; Poon, L. L. M.; Alard, P.; Cornelissen, L.; Bakker, A.; Cox, F.; van Deventer, E.; Guan, Y.; Cinatl, J.; ter Meulen, J.; Lasters, I.; Carsetti, R.; Peiris, M.; de Kruif, J.; Goudsmit, J. PLoS One 2008, 3, e3942.

(160) Whitehead, T. A.; Chevalier, A.; Song, Y.; Dreyfus, C.; Fleishman, S. J.; De Mattos, C.; Myers, C. A.; Kamisetty, H.; Blair, P.; Wilson, I. A.; Baker, D. Nat. Biotechnol. 2012, 30, 543.

113

(161) Fleishman, S. J.; Whitehead, T. A.; Ekiert, D. C.; Dreyfus, C.; Corn, J. E.; Strauch, E.-M.; Wilson, I. A.; Baker, D. Science 2011, 332, 816.

(162) Einhauer, A.; Jungbauer, A. J. Biochem. Biophys. Methods 2001, 49, 455.

(163) Wall, S. T.; Saha, K.; Ashton, R. S.; Kam, K. R.; Schaffer, D. V; Healy, K. E. Bioconjug. Chem. 2008, 19, 806.

(164) Conway, A.; Vazin, T.; Spelke, D. P.; Rode, N. A.; Healy, K. E.; Kane, R. S.; Schaffer, D. V. Nat. Nanotechnol. 2013.

(165) Hermanson, G. Bioconjugate Techniques; 2nd ed.; Academic Press: Boston, MA, 2008.

(166) Protein Calculator v3.4. http://protcalc.sourceforge.net/ (accessed Mar 20, 2014).

(167) Rathnayaka, T.; Tawa, M.; Nakamura, T.; Sohya, S.; Kuwajima, K.; Yohda, M.; Kuroda, Y. Biochim. Biophys. Acta 2011, 1814, 1775.

(168) Studier, F. W. Protein Expr. Purif. 2005, 41, 207.

(169) Han, J. C.; Han, G. Y. Anal. Biochem. 1994, 220, 5.

(170) Getz, E. B.; Xiao, M.; Chakrabarty, T.; Cooke, R.; Selvin, P. R. Anal. Biochem. 1999, 273, 73.

(171) Hansen, R. E.; Winther, J. R. Anal. Biochem. 2009, 394, 147.

(172) Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. M. J. Org. Chem. 1991, 56, 2648.

(173) Bagiyan, G.; Koroleva, I. Russ. Chem. Bull. 2003, 52, 1135.

(174) Ochs, C. J.; Such, G. K.; Stadler, B.; Caruso, F. Biomacromolecules 2008, 9, 3389.

(175) Jastrzabek, K. G.; Subiros-Funosas, R.; Albericio, F.; Kolesinska, B.; Kaminski, Z. J. J. Org. Chem. 2011, 76, 4506.

(176) Kamiński, Z. J.; Kolesińska, B.; Kolesińska, J.; Sabatino, G.; Chelli, M.; Rovero, P.; Błaszczyk, M.; Główka, M. L.; Papini, A. M. J. Am. Chem. Soc. 2005, 127, 16912.

(177) Khan, M. N. J. Pharm. Sci. 1984, 73, 1767.

(178) Baldwin, A. D.; Kiick, K. L. Bioconjug. Chem. 2011, 22, 1946.

114

(179) Baldwin, A. D.; Kiick, K. L. Polym. Chem. 2013, 4, 133.

(180) Kalia, J.; Raines, R. T. Bioorg. Med. Chem. Lett. 2007, 17, 6286.

(181) Lin, D.; Saleh, S.; Liebler, D. C. Chem. Res. Toxicol. 2008, 21, 2361.

(182) Shen, B.-Q.; Xu, K.; Liu, L.; Raab, H.; Bhakta, S.; Kenrick, M.; Parsons-Reponte, K. L.; Tien, J.; Yu, S.-F.; Mai, E.; Li, D.; Tibbitts, J.; Baudys, J.; Saad, O. M.; Scales, S. J.; McDonald, P. J.; Hass, P. E.; Eigenbrot, C.; Nguyen, T.; Solis, W. A.; Fuji, R. N.; Flagella, K. M.; Patel, D.; Spencer, S. D.; Khawli, L. A.; Ebens, A.; Wong, W. L.; Vandlen, R.; Kaur, S.; Sliwkowski, M. X.; Scheller, R. H.; Polakis, P.; Junutula, J. R. Nat. Biotechnol. 2012, 30, 184.

(183) Manual for the laboratory diagnosis and virological surveillance of influenza; WHO Press: Geneva, Switzerland, 2011.

(184) Klimov, A.; Balish, A.; Veguilla, V.; Sun, H.; Schiffer, J.; Lu, X.; Katz, J. M.; Hancock, K. Influenza Virus Titration, Antigenic Characterization, and Serological Methods for Antibody Detection; Kawaoka, Y.; Neumann, G., Eds.; Methods in Molecular Biology; Humana Press: Totowa, NJ, 2012; Vol. 865, pp. 25–51.

(185) Patke, S.; Boggara, M.; Maheshwari, R.; Srivastava, S. K.; Arha, M.; Douaisi, M.; Martin, J. T.; Harvey, I. B.; Brier, M.; Rosen, T.; Mogridge, J.; Kane, R. S. Angew. Chemie [Online Early Access]. DOI: 10.1002/ange.201400870. Published Online: Apr 6, 2014. http://onlinelibrary.wiley.com.libproxy.rpi.edu/doi/10.1002/ange.201400870/full (accessed May 7, 2014).

(186) Agrawal, L.; Lu, X.; Qingwen, J.; VanHorn-Ali, Z.; Nicolescu, I. V.; McDermott, D. H.; Murphy, P. M.; Alkhatib, G. J. Virol. 2004, 78, 2277.

(187) Signoret, N.; Pelchen-Matthews, A.; Mack, M.; Proudfoot, A. E.; Marsh, M. J. Cell Biol. 2000, 151, 1281.

(188) Muniz-Medina, V. M.; Jones, S.; Maglich, J. M.; Galardi, C.; Hollingsworth, R. E.; Kazmierski, W. M.; Ferris, R. G.; Edelstein, M. P.; Chiswell, K. E.; Kenakin, T. P. Mol. Pharmacol. 2009, 75, 490.

(189) Foster, T. J. Nat. Rev. Microbiol. 2005, 3, 948.

(190) Galy, R.; Bergeret, F.; Keller, D.; Mourey, L.; Prévost, G.; Maveyraud, L. Acta Crystallogr. Sect. F. Struct. Biol. Cryst. Commun. 2012, 68, 663.

115

(191) Joubert, O.; Viero, G.; Keller, D.; Martinez, E.; Colin, D. A.; Monteil, H.; Mourey, L.; Dalla Serra, M.; Prévost, G. Biochem. J. 2006, 396, 381.

(192) Viero, G.; Cunaccia, R.; Prévost, G.; Werner, S.; Monteil, H.; Keller, D.; Joubert, O.; Menestrina, G.; Dalla Serra, M. Biochem. J. 2006, 394, 217.

(193) Jayasinghe, L.; Bayley, H. Protein Sci. 2005, 14, 2550.

(194) Yamashita, K.; Kawai, Y.; Tanaka, Y.; Hirano, N.; Kaneko, J.; Tomita, N.; Ohta, M.; Kamio, Y.; Yao, M.; Tanaka, I. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 17314.

(195) Gravet, A.; Colin, D. A.; Keller, D.; Girardot, R.; Monteil, H.; Prévost, G.; Giradot, R. FEBS Lett. 1998, 436, 202.

(196) Alonzo, F.; Kozhaya, L.; Rawlings, S. A.; Reyes-Robles, T.; DuMont, A. L.; Myszka, D. G.; Landau, N. R.; Unutmaz, D.; Torres, V. J. Nature 2013, 493, 51.

(197) Reyes-Robles, T.; Alonzo, F.; Kozhaya, L.; Lacy, D. B.; Unutmaz, D.; Torres, V. J. Cell Host Microbe 2013, 14, 453.

(198) Lee, B.; Sharron, M.; Blanpain, C.; Doranz, B. J.; Vakili, J.; Setoh, P.; Berg, E.; Liu, G.; Guy, H. R.; Durell, S. R.; Parmentier, M.; Chang, C. N.; Price, K.; Tsang, M.; Doms, R. W. J. Biol. Chem. 1999, 274, 9617.

(199) Westby, M.; Smith-Burchnell, C.; Mori, J.; Lewis, M.; Mosley, M.; Stockdale, M.; Dorr, P.; Ciaramella, G.; Perros, M. J. Virol. 2007, 81, 2359.

(200) Xiao, Z.; Shangguan, D.; Cao, Z.; Fang, X.; Tan, W. Chemistry 2008, 14, 1769.

(201) Zhang, Y.; Chen, Y.; Han, D.; Ocsoy, I.; Tan, W. Bioanalysis 2010, 2, 907.

(202) Hicke, B. J.; Marion, C.; Chang, Y. F.; Gould, T.; Lynott, C. K.; Parma, D.; Schmidt, P. G.; Warren, S. J. Biol. Chem. 2001, 276, 48644.

(203) Shangguan, D.; Li, Y.; Tang, Z.; Cao, Z. C.; Chen, H. W.; Mallikaratchy, P.; Sefah, K.; Yang, C. J.; Tan, W. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 11838.

(204) Yethiraj, A. J. Phys. Chem. B 2009, 113, 1539.

(205) Kolesnikov, A. V.; Kozyr’, A. V.; Shemyakin, I. G. Mol. Genet. Microbiol. Virol. 2012, 27, 49.

(206) Lauridsen, L. H.; Veedu, R. N. Nucleic Acid Ther. 2012, 22, 371.

116

(207) Ellington, A. D.; Szostak, J. W. Nature 1990, 346, 818.

(208) Tuerk, C.; Gold, L. Science 1990, 249, 505.

(209) Mendonsa, S. D.; Bowser, M. T. J. Am. Chem. Soc. 2004, 126, 20.

(210) Cerchia, L.; Ducongé, F.; Pestourie, C.; Boulay, J.; Aissouni, Y.; Gombert, K.; Tavitian, B.; de Franciscis, V.; Libri, D. PLoS Biol. 2005, 3, e123.

(211) Lou, X.; Qian, J.; Xiao, Y.; Viel, L.; Gerdon, A. E.; Lagally, E. T.; Atzberger, P.; Tarasow, T. M.; Heeger, A. J.; Soh, H. T. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 2989.

(212) Merino, E. J.; Wilkinson, K. A.; Coughlan, J. L.; Weeks, K. M. J. Am. Chem. Soc. 2005, 127, 4223.

(213) Wilkinson, K. A.; Merino, E. J.; Weeks, K. M. Nat. Protoc. 2006, 1, 1610.

(214) Hendrix, M.; Priestley, E. S.; Joyce, G. F.; Wong, C. H. J. Am. Chem. Soc. 1997, 119, 3641.

(215) Wu, C.; Eder, P. S.; Gopalan, V.; Behrman, E. J. Bioconjug. Chem. 2001, 12, 842.

(216) Wecker, M.; Smith, D.; Gold, L. RNA 1996, 2, 982.

(217) Seelig, B.; Jäschke, A. Bioconjug. Chem. 1999, 10, 371.

(218) Nardese, V.; Longhi, R.; Polo, S.; Sironi, F.; Arcelloni, C.; Paroni, R.; DeSantis, C.; Sarmientos, P.; Rizzi, M.; Bolognesi, M.; Pavone, V.; Lusso, P. Nat. Struct. Biol. 2001, 8, 611.

(219) Nishiyama, Y.; Murakami, T.; Shikama, S.; Kurita, K.; Yamamoto, N. Bioorg. Med. Chem. 2002, 10, 4113.

(220) Ramnarine, E.; DeVico, A. L.; Vigil-Cruz, S. C. Lett. Pept. Sci. 2003, 10, 637.

(221) Secchi, M.; Longhi, R.; Vassena, L.; Sironi, F.; Grzesiek, S.; Lusso, P.; Vangelista, L. Chem. Biol. 2012, 19, 1579.

(222) Lusso, P.; Vangelista, L.; Cimbro, R.; Secchi, M.; Sironi, F.; Longhi, R.; Faiella, M.; Maglio, O.; Pavone, V. FASEB J. 2011, 25, 1230.

(223) Vangelista, L.; Secchi, M.; Lusso, P. Vaccine 2008, 26, 3008.

117

(224) Vangelista, L.; Longhi, R.; Sironi, F.; Pavone, V.; Lusso, P. Biochem. Biophys. Res. Commun. 2006, 351, 664.

(225) Duma, L.; Häussinger, D.; Rogowski, M.; Lusso, P.; Grzesiek, S. J. Mol. Biol. 2007, 365, 1063.

(226) Issafras, H.; Angers, S.; Bulenger, S.; Blanpain, C.; Parmentier, M.; Labbé-Jullié, C.; Bouvier, M.; Marullo, S. J. Biol. Chem. 2002, 277, 34666.

(227) Bennett, L. D.; Fox, J. M.; Signoret, N. Immunology 2011, 134, 246.

(228) Lin, Y.-L.; Mettling, C.; Portales, P.; Reynes, J.; Clot, J.; Corbeau, P. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 15590.

(229) Yang, J.; Gitlin, I.; Krishnamurthy, V. M.; Vazquez, J. A.; Costello, C. E.; Whitesides, G. M. J. Am. Chem. Soc. 2003, 125, 12392.

(230) Davis, N. E.; Karfeld-Sulzer, L. S.; Ding, S.; Barron, A. E. Biomacromolecules 2009, 10, 1125.

(231) Wang, Y.; Kiick, K. L. J. Am. Chem. Soc. 2005, 127, 16392.

(232) Liu, S.; Kiick, K. L. Macromolecules 2008, 41, 764.

(233) Kitov, P. I.; Mulvey, G. L.; Griener, T. P.; Lipinski, T.; Solomon, D.; Paszkiewicz, E.; Jacobson, J. M.; Sadowska, J. M.; Suzuki, M.; Yamamura, K.-I.; Armstrong, G. D.; Bundle, D. R. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 16837.

(234) Cui, L.; Kitov, P. I.; Completo, G. C.; Paulson, J. C.; Bundle, D. R. Bioconjug. Chem. 2011, 22, 546.

118

APPENDIX A – List of abbreviations

AEM Aminoethyl-maleimide

ANTXR1 Anthrax toxin receptor 1, also referred to by the acronyms ATR, for

“anthrax toxin receptor,” and TEM8, for “tumor endothelial marker 8”

ANTXR2 Anthrax toxin receptor 2, also referred to by the acronym CMG2, for

human capillary morphogenesis protein 2

CCL5 Chemokine (C-C motif) ligand 5, also referred to by the acronym RANTES,

for “regulated on activation, normal T cell expressed and secreted”

CCR5 C-C chemokine receptor type 5

CDC Center for Disease Control

DHA dehydroascorbic acid, an oxidizing agent

DMEM Dulbecco’s modified Eagle’s medium

DMTMM 4-(4,6-Dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride

DTT dithiothreitol, a reducing agent

EC50 Half maximal effective concentration

EDC 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide, a water-soluble reagent

for activating carboxylates for attack by amines to create amide bond

linkages

EDTA Ethylenediaminetetraacetic acid

EdTx Anthrax edema toxin (edema factor + protective antigen)

EF Anthrax edema factor

Fab Fragment, antigen binding. The antigen-binding antibody fragment

composed of one constant and one variable domain of each of

the heavy and the light chain

FBS Fetal bovine serum

HA Hemagglutinin

HBS HEPES-buffered saline

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HIV Human immunodeficiency virus

119

HOBt 1-hydroxybenzotriazole, a peptide-coupling reagent used to create

activated acids, similar to NHS, but more hydrophobic

HyA Hyaluronic acid, a polyanionic, unsulfated glycosaminoglycan

IC50 Half maximal inhibitory concentration

LAIV Live attenuated influenza vaccine

LeTx Anthrax lethal toxin (lethal factor + protective antigen)

LF Anthrax lethal factor

MBS MES-buffered saline

MDCK Madin-Darby Canine-Kidney cells

ME β-mercaptoethanol, a reducing agent

MES 2-(N-morpholino)ethanesulfonic acid

MWCO Molecular weight cut-off

NA Neuraminidase

NHS N-Hydroxysuccinimide, a reagent often used with EDC for creating an

activated acid (an ester with a good leaving group) to create amide bonds

PA Anthrax protective antigen

PA63 63 kDa fragment of anthrax protective antigen

[PA63]7 Toroid-shaped, self-assembled heptamer of PA63

PCR Polymerase chain reaction

PDB Protein data bank

PEG Poly(ethylene glycol)

PGA Poly(L-glutamic acid)

RMS Root-mean square

RNP Ribonucleoprotein complex of influenza virus, composed of viral RNA

coated with nucleoprotein and the RNA-dependent RNA polymerase

TCEP tris(2-carboxyethyl)phosphine, a reducing agent

TCID50 50% tissue culture infective dose (the dose that will infect 50% of

inoculated tissue cultures)

THP tris(hydroxypropyl)phosphine, a reducing agent