Biochemical and Structural Characterization of Proteins of The

Biochemical and Structural Characterization of

Proteins of the Herpesvirus Inner Tegument

A thesis

submitted by

Jared D. Pitts

In partial fulfillment of the requirements

for the degree of

Doctor of Philosophy

Molecular Biology and Microbiology

TUFTS UNIVERSITY

Sackler School of Graduate Biomedical Sciences

Date

May, 2014

ADVISER: Ekaterina Heldwein, Ph.D.

Abstract

The closely related alphaherpesviruses Herpes Simplex Virus 1 (HSV-1) and

Pseudorabiesvirus (PRV) are neurotropic viruses of humans and pigs respectively. The hallmark of HSV infection is the mucocutaneous lesions it causes during lytic infection.

PRV, the causative agent of Aujeszky’s disease in pigs and other livestock, often results in lethal infection. The mechanisms by which these viruses replicate and are released from host cells are not well defined. PRV and HSV-1 are two model viruses used to better characterize alphaherpesvirus replication.

In cells infected with herpesviruses, two capsid-associated or inner-tegument proteins,

UL37 and UL36 control cytosolic trafficking of capsids by as yet poorly understood mechanisms. In this work, we set out to further define the roles of UL36 and UL37 in viral replication though the biochemical and structural characterization of these proteins and their complexes. For the first time, we have purified and started biochemical characterization of the PRV UL37 protein and regions of the PRV and HSV-1 UL36 proteins in complex with either UL37 or the auxiliary capsid protein UL25.

Notably, we have determined the crystal structure of the N-terminal half of UL37 from

PRV. The structure, which is the first for any alphaherpesvirus inner tegument protein – reveals an elongated molecule of a complex architecture, rich in helical bundles. Through evolutionary trace analysis we were able to identify a novel functional region important for cell-cell spread. These results suggest a novel role for UL37 in intracellular trafficking that promotes spread of viral infection, expanding the repertoire of UL37

ii functions in intracellular virus trafficking. Unexpectedly, the UL37 N terminus shares a structural similarity with cellular multi-subunit tethering complexes (MTCs), which control vesicular trafficking in eukaryotic cells by tethering transport vesicles to their destination organelles. We attempted to determine if this structural similarity also resulted in a functional similarity by looking for host-protein interactions mediated by

UL37N. Although we currently have been unable to validate these putative interactions, results from our structural and mutagenesis studies support the notion that UL37 traffics capsids to cytoplasmic budding destinations and potentially further on to cell junctions for spread to nearby cells.

iii

Acknowledgements

First and foremost, I would like to thank my thesis adviser Professor Ekaterina Heldwein for her encouragement, her patience, her willingness to share her expertise, and her wonderful mentorship. Her continued guidance throughout my graduate school career has made a huge impact on me as a person and as a scientist. She always has a way to make me look at the positives when experiments aren’t going as planned, and without her mentorship this thesis would not have been possible.

I would like to thank the members of my thesis committee, Professor Joan Mecsas,

Professor Andy Camilli, and Professor Linc Sonenshien for their helpful discussions and guidance during the course of my PhD. I also would like to thank Professor Ralph Isberg for stepping in for my examination committee and Professor John Wills for taking time out of his schedule to agree to be my outside examiner.

I need give a huge thank you to all the past and present members of the Heldwein Lab, for making each day in the lab enjoyable. Sam Stampfer, Jessica Silverman, Claire

Metrick, Heidi Burke, and Henry Rogalin – you have been a wonderful group of graduate students to be in the lab with. Tirumal Chowdary, Elvira Vitu, Janna Bigalke, and

Rebecca Cooper, your guidance and advice over the years have been invaluable. I have learned a lot from you all about what it means to be a scientist, and promise to incorporate what you have all taught me as I move onto my own post-doc.

Thank you to the entire Molecular Microbiology Program and the Department of

Molecular Biology and Microbiology, I am always thankful to have chosen such an amazing and nurturing department to aide in my development as a scientist. To my rotation students Zach, Arti, Andrew, Heidi, and Matt thank you all for helping me realize what it means to be a mentor. Thanks to my classmates EmilyKate, Dennise, and

Talya, I couldn’t ask for a better group of people to come into graduate school with, and I truly enjoyed all of our years together.

I want to thank the roommates I have had in Boston. To Adam, it was great living with my best friend for so many years and I’m thankful for all of the memories we have from our time together in Boston. To Nick and Claudia, you have both been terrific roommates over the last year and I’ve been very fortunate to have such wonderful roommates as I’ve finish out my PhD.

Thanks to all of the Mighty Microbes softball players for all of the fun every summer, I truly enjoyed managing and playing on the team for so many years. To the Elephants basketball team, we went through a lot of season together, some good and some not so good, but we always played hard and made the best out of it. To the Sweep the Leg trivia team, there were few things I looked forward to more than our weekly trivia nights. Also

I need to say a huge thanks Andrew and Megan, for always being there to celebrate personal and professional accomplishments.

Finally, I would like to thank my family for their never ending support. To my parents,

Dave and Rhonda, who have always been there when I’ve needed them. They have a wonderful ability to keep me calm when I was getting anxious and are always willing to listen to any problems I may have had. To my sister Jamie, you have been such a great support to me over the years in Boston. For all the meals you have cooked for me and for all the time that we have spent just relaxing and enjoying each other’s company I am incredibly grateful. To my grandparents, I just want to thank you for all your love and support throughout the years. I am truly lucky to have such a wonderful and supportive family.

Table of Contents

Abstract ...... ii

Acknowledgements ...... iv

Table of Contents ...... vii

List of Tables ...... xii

List of Figures ...... xiii

List of Abbreviations ...... xvi

Chapter 1: Introduction ...... 1 1.1 Viral Assembly and egress...... 2 1.1.1 General overview of viral lifecycle ...... 2 1.1.2 Assembly and egress strategies of non-enveloped viruses ...... 3 1.1.3 Assembly and egress strategies of enveloped viruses ...... 4 1.1.3.1 Assembly and egress strategies of enveloped viruses which bud at the plasma membrane ...... 6 1.1.3.2 Assembly and egress strategies of enveloped viruses which bud at internal membrane sites ...... 9 1.2 Herpesvirus morphology and replication ...... 11 1.2.1 Introduction to Herpesviruses ...... 11 1.2.2 Herpesvirus morphology and tegument ...... 12 1.2.3 Overview of Tegument-Capsid Interaction and Early Tegumentation ...... 15 1.2.2 Current understanding of Herpesvirus replication ...... 18 1.3 Roles of the tegument proteins in Herpesvirus replication ...... 20 1.3.1 Targeting and uncoating ...... 20 1.3.2 Capsid assembly ...... 22 1.3.3 Nuclear Egress ...... 24 1.3.4 Herpesvirus trafficking and secondary envelopment ...... 27 1.3.5 UL36 and UL37 in pathogenesis ...... 29 1.4 How eukaryotic cells regulate vesicle trafficking ...... 30

vii

1.4.1 Vesicle Formation ...... 31 1.4.2 Vesicle Transit ...... 31 1.4.3 Vesicle Tethering and Fusion ...... 32 1.4.4 Vesicle Trafficking and Microbes ...... 36 1.5 Major unanswered questions in Herpesvirus Egress ...... 38

Chapter 2: Materials and Methods ...... 39 2.1 Cloning ...... 40 2.1.1 Cloning of UL36 constructs ...... 40 2.1.2 Cloning of UL37 constructs ...... 40 2.2 Protein expression and purification ...... 43 2.2.1 E. coli expression and purification ...... 43 2.2.1.1 UL25/UL36 purification ...... 44 2.2.1.2 UL37 purification ...... 45 2.2.1.3 Purification of Rab proteins, Sec22b, and LidA ...... 47 2.2.2 Insect cell expression and purification ...... 48 2.2.2.1 UL25, UL36, and UL37 purification ...... 49 2.2.2.2 gB purification ...... 50 2.3 Biochemical characterization ...... 50 2.3.1 Western Blot Analysis ...... 50 2.3.2 Electron Microscopy ...... 51 2.3.3 Co-Immunoprecipitation ...... 52 2.3.4 Circular Dichroism ...... 52 2.3.5 Thermofluor Assay ...... 53 2.3.6 Size-exclusion chromatography (SEC)-coupled multi-angle light scattering (MALS) ...... 54 2.3.7 Mass Spectrometry ...... 54 2.4 Crystallization ...... 54 2.4.1 UL37N Crystallization ...... 55 2.4.2 gB crystallization ...... 55 2.5 Obtaining Heavy atom derivatives of UL37N ...... 56 2.5.1 Selenomethionine incorporation into proteins ...... 56 2.5.2 Heavy Metal Incorporation into proteins ...... 57 2.6 Structure Determination ...... 57

viii

2.7 Structure analysis ...... 58 2.8 Accession numbers ...... 59 2.9 Identification and validation of putative host binding proteins of UL37N ...... 59 2.9.1 Pull-down with UL37N sepharose ...... 59 2.9.2 GST Pull-downs ...... 61 2.9.3 GFP-Trap Co-Immunoprecipitation ...... 62 2.9.4 Subcellular localization of GFP fusions and immunofluorescence ...... 63

Chapter 3: Biochemical Characterization of the UL37 protein of Pseudorabiesvirus ...... 65 3.1 Introduction ...... 66 3.2 Results ...... 67 3.2.1 UL37 is truncated upon recombinant expression ...... 67 3.2.2 Biochemical characterization of full-length UL37 from PRV ...... 68 3.2.3 Increasing yield of soluble full-length UL37 ...... 70 3.2.4 UL37N and UL37C terminal constructs ...... 74 3.2.4.1 Expression and purification ...... 74 3.2.4.2 Biochemical characterization of UL37N and UL37C ...... 77 3.2.4.3 Additional UL37C terminal constructs ...... 84 3.2.5 UL37/UL36 Interaction studies ...... 86 3.3 Discussion ...... 90

Chapter 4: The crystal structure of PRV UL37N supports its critical role in control of viral trafficking ...... 94 4.1 Introduction ...... 95 4.2 Results ...... 96 4.2.1 Crystallization trials of UL37 constructs ...... 96 4.2.2 Obtaining heavy metal derivatives for UL37N ...... 97 4.2.3 Analysis of the crystal structure of UL37N ...... 101 4.2.3.1 UL37N is a crystallographic dimer but a monomer in solution ...... 101 4.2.3.2 The architecture of UL37N ...... 106 4.2.4 Identification of Functional Regions of UL37N ...... 110 4.2.4.1 Electrostatics of UL37N ...... 111 4.2.4.2 Mapping of conserved residues onto UL37N structure ...... 111 4.2.4.2 ETA reveals three conserved surface patches within UL37N ...... 115

4.2.5 ETA mutants in Region 2 display defect in cell-to-cell spread and retrograde trafficking ...... 117 4.2.6 UL37N possesses structural similarity to eukaryotic proteins of known function ...... 122 4.2.6.1 Domain 3 of UL37N possesses a weak structural similarity to 14-3-3 proteins .. 122 4.2.6.2 UL37 shares a structural similarity with subunits of the CATCHR family of tethering complexes ...... 123 4.3 Discussion ...... 129

Chapter 5: Identification of PRV UL37N putative host protein binding partners. 135 5.1 Introduction ...... 136 5.2 Results ...... 137 5.2.1 Identification of putative host cell binding partners of UL37N ...... 137 5.2.1.1 Rab and SNARE proteins identified ...... 142 5.2.2 Rab and SNARE proteins do not directly bind UL37N...... 145 5.2.3 Rab proteins are not observed to interact with UL37N using cell based methods ...... 150 5.2.4 Localization of UL37 and UL37N GFP fusions ...... 152 5.3 Discussion ...... 154

Chapter 6: Characterizing the HSV-1 UL25/UL36 cbd Interaction ...... 159 6.1 Introduction ...... 160 6.2 Results ...... 161 6.2.1 Construct Rationale ...... 161 6.2.2 UL25/UL36cbd do not interact from E. coli expressed protein ...... 163 6.2.3 Insect Cell Expression and Co-purification of UL25/UL36cbd ...... 165 6.2.4 Increasing the solubility of the UL25/UL36cbd complex ...... 167 6.3 Discussion ...... 172

Chapter 7: Discussion and Future Directions ...... 174 7.1 Significance of work ...... 175 7.2 Future Directions ...... 177 7.2.1 Characterization of R2 mutants and additional structure based mutagenesis 177 7.2.2 Identification of UL37 host-cell binding proteins ...... 179 7.2.3 Structural Determination of UL37C ...... 181

7.2.4 Further investigations into tegument-tegument interactions and capsid- tegument interactions ...... 183

Appendix A: Using gB ecto-domain as an anchor for structural studies of HSV-1 gB-cytoplasmic tail ...... 185 A.1 Introduction ...... 186 A.2 Results ...... 188 A.2.1 Construct Rationale ...... 188 A.2.2 Obtaining purified gB protein containing the cytoplasmic tail ...... 191 A.2.3 Electron Microscopy ...... 193 A.2.4 Crystallization ...... 195 A.2.5 Analysis of Diffraction ...... 195 A.3 Conclusions ...... 197

Appendix B: Table of Expression Plasmids Used ...... 199

References ...... 203

List of Tables

Table # Title Page

Table 3-1: Thermofluor Melting of UL37 N- and C- terminal proteins 78

Table 4-1: UL37N Data collection and refinement statistics 101

Table 4-2: Z scores of Alignments with components of MTCs 127

Table 5-1: Potential UL37N binding partners 140

Table B1: Expression plasmids used in these studies 198

xii

List of Figures

Figure 1-1: Model of virus egress of non-enveloped and enveloped viruses ...... 5

Figure 1-2: Viral proteins that drive budding at the plasma membrane ...... 7

Figure 1-3: Schematic of Herpesvirus morphology ...... 13

Figure 1-4: Interactions of UL25/UL36/UL37 ...... 17

Figure 1-5: Model of Herpesvirus replication cycle ...... 19

Figure 1-6: Herpesvirus egress from infected cells ...... 25

Figure 1-7: Multi-subunit tethering complexes of eukaryotic cells ...... 33

Figure 1-8: Mechanism of action of Multi-subunit tethering complexes in membrane fusion in eukaryotic cells ...... 35

Figure 3-1: The PRV UL37 protein is truncated during E. coli expression ...... 69

Figure 3-2: Biochemical characterization of full-length PRV UL37 ...... 71

Figure 3-3: Full-length UL37 protein always leads to truncations and is highly insoluble73

Figure 3-4: UL37N and UL37C terminal constructs can be readily purified ...... 76

Figure 3-5: UL37N (1-496) is more stable over time than UL37N (1-476) ...... 80

Figure 3-6: UL37N and UL37C are predominantly helical ...... 82

Figure 3-7: Multi-Angle Light Scattering of UL37N and UL37C ...... 83

Figure 3-8: Purification of additional UL37C constructs ...... 85

Figure 3-9: E. coli expressed PRV UL37 and PRV UL36 (344-561) interact ...... 88

Figure 3-10: Purified UL37-UL36 complex is aggregated ...... 89

Figure 4-1: Optimization of UL37N crystallization ...... 98

Figure 4-2: Native gels of heavy atom soaks of UL37N ...... 100

Figure 4-3: Overlay of the two UL37N monomers in the asymmetric unit ...... 103

Figure 4-4: UL37N is a calcium-dependent dimer in crystals but not in solution...... 105

xiii

Figure 4-5: UL37N structure and domain organization ...... 109

Figure 4-6: Electrostatic surface potential map of UL37N ...... 112

Figure 4-7: Multiple sequence alignment of UL37 homologs ...... 114

Figure 4-8: Mapping of conserved residues on the surface UL37N ...... 116

Figure 4-9: Phylogenetic tree of UL37 from the Evolutionary Trace Analysis ...... 118

Figure 4-10: Mapping of conserved residues from ETA analysis and identification of potential functional regions...... 119

Figure 4-11: Propagation and spread of PRV encoding mutant forms of UL37 ...... 121

Figure 4-12: Overlay of UL37N and 14-3-3 protein Tau ...... 124

Figure 4-13: UL37N shares structural similarities with several subunits of cellular MTCs tethering complexes ...... 127

Figure 4-14: Model of UL37 function in viral egress ...... 134

Figure 5-1: UL37N pull-down of host-cell binding partners ...... 140

Figure 5-2: Rab5 and Rab6 pull-down with UL37N sepharose ...... 144

Figure 5-3: UL37N does not directly interact with Rab proteins ...... 147

Figure 5-4: UL37N does not directly interact with the SNARE protein Sec22b ...... 149

Figure 5-5: Rab proteins do not co-immunoprecipitate with UL37-GFP fusions ...... 151

Figure 5-6: UL37N-GFP and UL37-GFP localize to distinct loci but different loci ...... 153

Figure 5-7: UL37N-GFP does not co-localize with Golgi or early endosomal markers 155

Figure 6-1: Linear representations of the UL25 and UL36cbd proteins ...... 162

Figure 6-2: Bacterially expressed UL25 and UL36cbd do not interact ...... 164

Figure 6-3: UL36cbd is required to be expressed in eukaryotic cells for interaction with UL25 ...... 166

Figure 6-4: UL25/UL36cbd complex obtained from Sf9 cells is aggregated ...... 168

xiv

Figure 6-5: Additives aide in solubility but not prevention of aggregation of UL25/UL36 complex ...... 170

Figure 6-6: Extending UL36cbd domain boundaries does not minimize aggregation ...171

Figure A-1: Schematic representation of the gB constructs ...... 190

Figure A-2: Purification of gB constructs containing the cytoplasmic domain ...... 192

Figure A-3: Negative stain electron microscopy of gB constructs ...... 194

Figure A-4: gB ectodomain-cytodomain fusion crystals ...... 196

List of Abbreviations

293T Human embryonic kidney 293 cells

Amp Ampicillin

BSA Bovine Serum Albumin

CA Constitutively Active

CATCHR Complexes associated with tethering containing helical rods

CD Circular dichroism

Cm Chloramphenicol

CV Column Volume

CVSC Capsid vertex specific component

DAPI 4,6-diamidino-2-phenylindole

DMEM Dulbecco’s Modified Eagle medium

DN Dominant negative

DNA Deoxyribonucleic acid

DUB Deubiquitinase

EBV Epstein-Barr virus

EDTA Ethylenediaminetetraacetic acid

EE Early Endosome

EM Electron Microscopy

ER Endoplasmic reticulum

ETA Evolutionary Trace Analysis

FBS Fetal bovine serum

FITC Fluorescein isothiocyanate

xvi gB (D,etc.) Glycoprotein B (D, etc.)

GDI Guanine nucleotide displacement inhibitor

GEF Guanine nucleotide exchange factor

GFP Green Fluorescent Protein

GST Glutathione S-transferase

HCMV Human cytomegalovirus

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HIV Human immunodeficiency virus

HRP Horseradish peroxidase

HSV Herpes Simplex Virus

IMAC Immobilized metal affinity chromatography

IP Immunoprecipitation

IPTG Isopropyl-β-D-thiogalactopyranoside

Kan Kanamycin

KSHV Kaposi’s Sarcoma herpesvirus

LB Luria Broth

MALS Multi-Angle light scattering

MPR Membrane-proximal region

MTC Multi-subunit tethering complex

NES Nuclear export signal

Ni-NTA Nickel-nitriloacetic acid

NLS Nuclear localization signal

NUP Nucleoporin

P3 Third baculovirus passage

PBS Phosphate buffered saline

PDB Protein databank

PEG Polyethylene glycol

PEI Polyethylenimine

PIPES Piperazine-N,N’-bis(2-ethanesulfonic acid)

PK15 Pig kidney epithelial cells

PMSF Phenylmethylsulfonyl fluoride

PRV Pseudorabies virus

R1/2/3 Region 1, 2, or 3 rmsd Root-mean squared deviation

S200 Superdex 200 column

SAD Single wavelength anomalous dispersion

SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis

SEC Size exclusion chromatography

SeMet Selenomethionine

Sf9 Spodoptera frugiperda siRNA Small interfering ribonucleic acid

SNARE Soluble NSF Attachment protein Receptor

SOE PCR Splicing by overlap extension polymerase chain reaction

TBST 20 mM Tris-HCl(pH 7.5), 500 mM NaCl, and 0.05% Tween 20

TCEP Tris-(2-carboxyethyl) phosphine

TFF Tangential flow filtration

iii

TGN Trans-Golgi Network

TM Transmembrane domain

UL36cbd Capsid binding domain of the UL36 protein

UL37 Unique Long (gene 37)

UL37N UL37 N-terminal half (predominantly 1-496)

UL37C UL37 C-terminal half (predominantly 478-919)

US3 Unique Short (gene 3)

Vero African green monkey kidney epithelial cells

VP Viral protein

VSV Vesicular stomatitis virus

VZV Varicella zoster virus

WT Wild type

Chapter 1: Introduction

1.1 Viral Assembly and egress

1.1.1 General overview of viral lifecycle

Viruses come in many shapes and sizes, and are known to infect all forms of life, from bacteria and fungi all the way up to animals. Even among viruses that infect vertebrate animals there is an amazing amount of diversity. Despite their many differences, there are basic patterns of viral replication, which all animal viruses follow to produce progeny virions inside of infected cells. In the initial stage of viral infection the virus must attach to the host cell, often through interaction of viral proteins with specific receptors present on the cell membrane. The specificity of attachment is often associated with the tropism of the virus. After the virus is bound to the cell, the virus must enter or penetrate the host cell membrane. The viral penetration step is mediated predominantly by either endocytosis of the viral particle or through direct membrane fusion at the plasma membrane. Upon entry into the cell the virus then releases its genome from the viral capsid in a step known as viral uncoating.

Once the genome is released from the capsid the viral genes are transcribed and viral proteins expressed. In addition to protein production the viral genome is replicated. Thus, the virus has produced all of the components required to make progeny virions. The viral components next assemble, and the virus can undergo additional processes of maturation required to yield infectious particles. These infectious particles are then released into the extracellular environment either after lysis of the cell, through budding at the plasma membrane, or in some cases through utilization of the host secretory machinery. This

2 very basic model of replication cycle is true for nearly all viruses; however, viruses differ greatly in exactly how and where in the cell each of these stages is performed.

In order to successfully produce progeny virions, viruses must have strategies to assemble new virions from the genome, protein (i.e. capsids), and potentially membranes components. In addition, maturation steps are often required to ensure that viruses released into the extracellular environment can move on to infect other cells. If we can better understand the mechanisms used by viruses to successfully assemble, mature, and release progeny virions, then we can potentially design strategies to prevent these processes and in doing so prevent viral spread.

1.1.2 Assembly and egress strategies of non-enveloped viruses

All viruses (enveloped and non-enveloped) assemble viral capsids utilizing fairly similar mechanisms. The capsids are often assembled through self-assembly of the capsid components; however, in some cases such as adenoviruses the capsid proteins will form around scaffolding proteins which are removed as the virus undergoes maturation (124).

Assembly of non-enveloped viruses can take place in the nucleus, as is the case with most naked DNA viruses (i.e. the SV40 polyomavirus or adenoviruses), or in the cytoplasm as is observed with non-enveloped RNA viruses (i.e. picornaviruses). Regardless of where the assembly process takes place, the process of virion assembly for naked virus is fairly simple, typically requiring the assembly of only a few capsid proteins and possibly an accessory protein or two. Furthermore, as the naked viruses do not need to acquire an envelope, the assembled viruses are often fully infectious upon packaging of the viral

3 genome. In some cases, the virus requires protease cleavage of structural proteins (i.e. L3 proteolysis of precursor proteins in adenoviruses) (38).

Mature virions of non-enveloped viruses will then accumulate in the nucleus or cytoplasm of the infected cell, ready to be released into the extracellular environment to ultimately spread infection to other cells. The most common mechanism used by naked viruses for egress is through lysis of the host cell (Fig. 1-1) (47, 51). In general the mechanisms utilized by non-enveloped viruses to induce cell lysis are not well understood. In the case of poliovirus, the 3A non-structural protein has been characterized as a protein that induces a cytopathic effect eventually leading to lysis of the infected cell and release of the progeny virions (90).

Although cell lysis is the predominant egress mechanism used by non-enveloped viruses, there are examples in which naked viruses are released from cells in the absence of cell lysis. For example, the Simian Virus 40 is released from polarized epithelial cells at apical membranes (30) by being incorporated into secretory granules and using the cells secretory transport pathway to be released from the cell (11). Interestingly, the process that Simian Virus 40 utilizes for egress is reminiscent of the viral release strategies used by some enveloped viruses.

1.1.3 Assembly and egress strategies of enveloped viruses

The assembly and egress strategies utilized by enveloped viruses are more complicated than those of non-enveloped viruses, primarily due to the requirement of acquisition

Figure 1-1: Model of virus egress of non-enveloped and enveloped viruses. Schematic displaying the differences in virus release mechanisms used by non-enveloped and enveloped viruses. Naked viruses are most often released from infected cells through cell lysis. In contrast, enveloped viruses most often are released from cells using non-lytic strategies. One strategy of budding into the plasma membrane to acquire the envelope and be released is displayed. Alternatively, enveloped viruses can bud into intracellular membrane compartments and be released using the cells secretory pathway. Adapted from H. Garoff et al., Mol. Microbiology. Rev. 62:1171-1190, 1998 (51) with permission.

5 of the viral envelope. In most cases, capsid assembly occurs at the site of envelope acquisition (although Herpesviruses and Hepadnaviruses are notable exceptions to this).

Thus the processes of viral assembly and envelopment are coordinated, in some cases these events are also synchronized with viral release as the virus buds from the plasma membrane of the infected cell (Fig 1-1).

1.1.3.1 Assembly and egress strategies of enveloped viruses which bud at the plasma membrane

Several families of enveloped viruses assemble virions and acquire their envelope at the plasma membrane of the infected cell. Among these viruses are togaviruses, retroviruses, orthomyxoviruses, and rhabdoviruses. The major advantage of viruses that bud at the plasma membrane is the coordinated actions of capsid assembly with envelope acquisition and release into the extracellular environment. Although all of these viral families are utilizing a similar pathway to assemble and egress from the cell, the exact mechanisms used for each family are more divergent (Fig 1-2).

Among togaviruses, the Sindbis virus capsid assembles at plasma membrane sites, which are studded with the envelope glycoproteins. The core capsid proteins directly interact with the cytoplasmic domains of these glycoproteins (predominantly the E2 protein) thus initiating budding into the plasma membrane and release of the newly formed virus (75).

In contrast, retroviruses such as human immunodeficiency virus (HIV) and orthomyxoviruses such as influenza A virus mediate budding through the utilization of matrix proteins. In both of these virus families the matrix protein (Gag for HIV and M1

Figure 1-2: Viral proteins that drive budding at the plasma membrane. Viruses can bud into the plasma membrane using a variety of mechanisms. A subset of the known mechanisms of viral budding at the plasma membrane is depicted here. The budding process can be driven by interactions between envelope glycoproteins (red) and capsid

(blue) as observed in Sindbis virus. In influenza A virus and HIV the matrix proteins

(brown) act alone to mediate budding. Budding for the coronavirus mouse hepatitis virus is driven by membrane glycoproteins. Rhabdoviruses such as rabies virus, can bud into the plasma membrane with only matrix proteins however, this budding is aberrant in the absence of glycoproteins. Adapted from H. Garoff et al., Mol. Microbiology. Rev.

62:1171-1190, 1998 (51) with permission.

7 and M2 for Influenza A) the matrix proteins are recruited to plasma membrane sites containing envelope glycoproteins, and the matrix proteins mediate the budding and scission process either through the action of the proteins themselves in the case of M1 and M2 (134, 137), or through a concerted action of the budding mediated by the matrix protein and several cellular proteins recruited (ESCRT proteins) to the budding site to aide in scission of the virus from the cell membrane, in the case of retroviral Gag (21,

126). In both cases, although there are glycoproteins present at the site of budding, these proteins do not play an active role in the budding process. The mechanism of budding for the sets of matrix proteins appears to the due to the oligomerization of the proteins which induces curvature of the plasma membrane from which the virus can be fully enveloped and ultimately released from the cell (21, 144).

Budding at the plasma membrane can also be mediated exclusively by the membrane as observed in a coronavirus mouse hepatitis virus (167). Finally, in rhabdovirus, such as rabies virus, the matrix protein M in the presence of ribonucleoprotein can mediate budding at the plasma membrane (similar to HIV Gag). However, in the absence of the envelope glycoprotein G this budding is aberrant, often resulting in abortive budding of viruses (110). This implies that M protein is the primary driver of budding, but the G protein and ribonucleoprotein protein play roles in this process as well, ensuring proper virus budding and release.

1.1.3.2 Assembly and egress strategies of enveloped viruses which bud at internal membrane sites

Although undergoing envelopment at the plasma membrane has the major advantage of being coordinated with release from the cell, not all viruses obtain their envelope at this site. Several viruses obtain their envelopes from intracellular membrane compartments or organelles that possess their envelope glycoproteins. The cytoplasmic tails of these glycoproteins often mediate the budding into the lumen of these internal membrane sites.

For example, in Rift Valley Fever Virus (a bunyavirus), the cytoplasmic tails of the G(C) and G(N) glycoproteins regulate the localization of the glycoproteins within Golgi membranes. These cytoplasmic tails then interact with ribonucleoproteins bound to the viral genome, and this interaction enables genome packaging and budding of the virus into the lumen of the Golgi (23, 120). It is then believed that the viruses are released utilizing the secretory pathway of the host-cell; however, the exact mechanism of release of the enveloped bunyaviruses within the Golgi lumen is currently unclear.

Hepadnaviruses are another example of viruses that bud into internal membrane compartments. Hepadnaviruses, such as hepatitis B virus, bud into the ER-Golgi intermediate compartment or multi-vesicular bodies, which are studded with envelope glycoproteins (127). One major difference in the assembly and budding process for these viruses is that the capsid is assembled and packaged at cytoplasmic sites not associated with the intracellular membranes used for budding. The major determinant of budding into the intracellular vesicle is the concentration of the L envelope glycoprotein within the membrane compartment (97). At sufficiently high levels the viral capsid can interact

9 with the cytoplasmic domains of the L glycoprotein, initiating the budding into the lumen of the intracellular membrane and subsequent release of the virus, again presumably using the secretory pathway of the cell (127). Interestingly, Hepatitis B virus also releases large numbers of non-enveloped capsids from the cell using a currently unknown mechanism (117).

One of the most complex assembly and egress pathways observed in viruses is that of the poxviruses. The prototypical poxvirus, vaccinia virus induces formation of membrane

“crescents” derived from ER membrane, which is used to wrap and assemble the viral genome and all of structural viral proteins of the virus. These newly assembled viruses then mature within these membranes, in part through loss of a scaffolding protein D13, which induces a large morphological change in the virus (161). In most cases the mature virus is released through lysis of the infected cell. However, in several instances the mature virions acquire a second membrane of Golgi or early endosomal origin and then are released by fusing the outermost envelope layer with the plasma membrane, resulting in cell associated virus (containing one additional envelope layer) that can infect neighboring cells (153).

Another virus family that has a complex mechanism for assembly and egress is the herpesvirus family of viruses. This mechanism is unusual in its complexity due to the fact that herpesviruses assemble their capsids in the nucleus but obtain the viral envelope by budding into the lumen of intracellular membranes (of Golgi or Early Endosomal origin).

The process of Herpesvirus egress will be described in greater detail in the following section.

1.2 Herpesvirus morphology and replication

1.2.1 Introduction to Herpesviruses

The order herpesvirales is made up of a diverse number of enveloped, double-stranded

DNA viruses that have the ability to infect a wide array of animal hosts (40). Everything from oysters and fish up to birds and mammals are known to be susceptible to these viruses. Within this order is the herpesviridae family, which primarily infects birds and mammals, and is comprised of three subfamilies alphaherpesvirinae, betaherpesvirinae, and gammaherpesvirinae. Members of each of these subfamilies are known to have the ability to infect humans, including herpes simplex viruses types 1 and 2 (HSV-1 and

HSV-2) and varicella zoster virus (VZV) belonging to the alphaherpesviruses; human cytomegalovirus (HCMV) and the human herpesviruses type 6 and 7 (HHV6 and HHV7) belonging to the betaherpesviruses, and finally Epstein-Barr virus (EBV) and Kaposi

Sarcoma-associated herpesvirus (KSHV) belonging to the gammaherpesviruses. These viruses are quite ubiquitous throughout the population with seropositivity rates from 10% for KSHV (7, 46) up to 90% for EBV as estimated by the Center for Disease Control and

Prevention. HSV-1 and HSV-2 are also quite common among the adult population of the

United States with seropositivity rates of approximately 68% (154) and 20% (46) respectively. In lytic HSV infections, the viruses replicate in epithelial cells of the maxillofacial and genital regions, leading to the skin lesions for which the virus is best known. As with all hosts infected with herpesviruses, HSV infected individuals are subject to a life-long infection, as the virus latently persists in a subset of infected cells

(i.e. the peripheral sensory ganglia for HSV-1). The neurotropic nature of the alphaherpesviruses can potentially lead to more severe neurological complications, such as herpes simplex encephalitis, meningitis, ascending myelitis, and retinal necrosis (12).

The most devastating of these complications is neonatal herpes simplex encephalitis, an often fatal disease, which occurs when infected women transmit the virus to their newborn during childbirth (37). Recent publications also indicate correlations between

HSV infection and increased risks of HIV infection and Alzheimer’s disease (68, 69).

In addition to being a major human pathogen, alphaherpesviruses are also major veterinary pathogens. The best characterized of these animal herpesviruses is the causative agent of Aujeszky’s disease in pigs, the pseudorabies virus (PRV), which is closely related to VZV, HSV-1, and HSV-2. Although a live-attenuated vaccine for PRV is available for livestock pig populations, PRV is still a major burden in developing countries as well as to wild pig and boar populations. As PRV is closely related to the human alphaherpesviruses, it is often used as a model system for lytic alphaherpesvirus infections.

1.2.2 Herpesvirus morphology and tegument

Regardless of the animal that is infected or the disease outcome all herpesviridae have a common morphology (Fig. 1-3). The large 120-300 kb dsDNA genome is encased in the viral capsid, which is made up primarily of the major capsid protein VP5 in addition to several minor capsid proteins (triplex proteins UL18 and UL38, the portal protein UL6, and additional accessory proteins UL17 and UL25, among others) (113). The capsid is

Figure 1-3: Schematic of Herpesvirus morphology. (A) A schematic diagram of a mature herpesvirus virion in which each component is labelled. (B) EM images of the alphaherpesviruses HSV-1 and PRV in which each component of the herpesvirus virion can be visualized. Figure was adapted from H Guo et al., Protein Cell 11:987-998, 2010 (57) and T Mettenleiter et al., Virus Res. 143:222-234 (113), 2009 with permission.

13 surrounded by a proteinaceous layer between the capsid, termed the tegument, which is wrapped in the lipid bilayer of the viral envelope studded with the viral glycoproteins.

The tegument of the alphaherpesviruses HSV-1 and PRV both possess roughly 20 virally encoded proteins and several cellular proteins (78, 85, 102). The proteins found in the tegument can be divided into two distinct subsets, the inner tegument and the outer tegument. The composition of the inner tegument proteins appears to be tightly regulated as overexpression of these proteins in the cell does not result in an increase in the amount of the protein present in extracellular virions (108). Inner tegument proteins are termed so because they are the first tegument proteins to appear on the capsid after it is made and also the last present during entry. Furthermore, these proteins are resistant to removal from high salt washes of the viral capsids (129). Among the proteins that make up the inner tegument are US3, UL36, and UL37. All three of these proteins are found on the capsid at early time-points during viral assembly and at late-time points in viral entry.

US3 is thought to phosphorylate several of outer tegument proteins during entry, ultimately releasing them from the incoming capsid (115). The UL36 and UL37 proteins both have late expression kinetics expected for the structural proteins, and both are also phosphorylated by cellular kinases, though neither protein is removed from the capsid upon phosphorylation (4, 109, 147). The UL36 protein is likely to be the first tegument protein to attach to the capsid as it directly interacts with the auxiliary capsid protein

UL25 (34, 122). The UL36 protein also interacts with other inner tegument proteins

UL37 (81) and UL48 (169), thus acting a scaffold for the acquisition of additional tegument proteins. The outer tegument proteins appear to be much less regulated as far as

14 the amount of each protein present in mature virus (78). This seems reasonable given that the outer tegument proteins are believed to be predominantly acquired during secondary envelopment through tegument-tegument and tegument-glycoprotein interactions. Some outer tegument proteins consist of UL11, UL16, UL21, and UL47 among many others.

Some of these proteins have functions at early time-points of infection, like UL48, which is involved in transcribing the immediate-early genes of HSV-1 and PRV (48, 89).

However, complexes between UL11, UL16, and UL21 are also implicated in secondary envelopment (24, 158). Inner tegument proteins in general are also more highly conserved than outer tegument proteins, for example UL36 and UL37 are conserved throughout herpesviridae, while several of the outer tegument proteins including UL47 and UL49 are only present in alphaherpesviruses (33, 61, 78). There are exceptions to this however, as some outer tegument proteins (i.e. UL16 and UL11) are also conserved among all herpesviruses (101).

1.2.3 Overview of Tegument-Capsid Interaction and Early Tegumentation

UL36, a 3164 amino acid protein in HSV-1, is the largest protein encoded in the herpesvirus genome and is involved in several points of the viral lifecycle. Among the functions of the protein is a deubiquitinase (DUB) at the N terminus. Although the DUB domain is conserved among herpesviruses (142) and is important for neurovirulence in vivo, its activity is dispensable for infection in cell culture (13, 14). In addition to the

DUB domain, the UL36 protein is thought to be one of the major mediators in tegument acquisition, essentially acting as a scaffold for tegument proteins to bind. In line with this, the N-terminal region of HSV-1 UL36 mediates interactions with tegument proteins

UL48 (UL36 residues 260-596) and UL37 (UL36 residues 512-767) based upon Co-IP and yeast two hybrid studies (81, 114). UL36 also interacts with the auxiliary capsid protein UL25, thus bridging the capsid and tegument (34). This interaction is potentially mediated by two binding sites on UL36, the C-terminal 62 amino acids, referred to as the capsid binding domain (cbd) (34) and a second region, residues 2200-2353 (122).

The 580 amino acid UL25 protein possesses five-fold symmetry at the pentameric capsid vertices via an interaction of its N-terminal 40 residues with UL17, this complex making up the capsid vertex specific components (CVSC) (Figure 1-4B) (35, 163). UL25 is required for proper retention of DNA in the capsid, possibly signaling to the DNA packaging machinery that the capsid is full and the DNA ready to be cut (32, 160). The structure for amino acids 134-580 of UL25 is known, and this region was recently determined to interact with the UL31 protein of the nuclear egress complex in addition to the UL36 protein.

The interactions between the capsid and inner tegument proteins are conserved throughout herpesviruses. These are vital interactions, ensuring efficient herpesvirus replication and maturation through proper localization of the inner tegument proteins and several additional functions during replication. Two of the inner tegument proteins, UL36 and UL37, help carry out several essential functions throughout the viral replication and maturation process. Additionally, interaction between UL36 and UL37 as well as the interaction between UL36 and the capsid protein UL25 are critical for proper localization and function of the inner tegument proteins. Thus, interactions between these proteins

Figure 1-4: Interactions of UL25/UL36/UL37. (A) Linear diagrams from UL36, UL37, and UL25 of HSV-1. The black bars indicate the known binding regions for interacting partners of each protein. Also displayed are the known nuclear localization signal (NLS) and nuclear export signal (NES) and the deubiquitinase region (DUB in green). (B)

Density map of a wild-type HSV-1 penton surrounded by UL17/UL25 proteins, displaying the five copies of the CVSC surrounding the penton (C) A schematic of interactions between UL25, UL36, UL37, and VP16 displayed at only one vertex of the capsid for simplicity. (D) Schematic representation of the CVSC surrounding the pentameric vertex of an HSV-1 capsid. Panels B and D were adapted from Toropova et al.

J Virol. 85(15):7512-22 (2011) (164) with permission. Panel (C) was adapted from

Coller et al., J. Virol 81:11790-7, 2007 (34) with permission.

17 represent potential targets for novel therapeutics, and need to be better understood at a biochemical and structural level. Figure 1-4 provides an overview of the interactions mediated by the UL25, UL36, and UL37 proteins and displays a simplified schematic of how capsid-inner-tegument interactions are arranged.

1.2.2 Current understanding of Herpesvirus replication

Herpesviruses initially infect cells through fusion of their viral membranes with the membranes of host cells (Fig 1-5 number 2). Four viral glycoproteins (gD, gB, and the gH/gL heterodimer) are required to mediate the fusion event. Once the fusion pore is formed at the cellular envelope, the infecting viral capsid enters the cytoplasm and is trafficked to a nuclear pore via microtubules (Fig. 1-5 numbers 3-4). The DNA genome is then released into the nucleus, and the virus can start the processes of replication of its genome, transcription and translation of its proteins, and finally assembly of new capsids

(Fig. 1-5 numbers 5-7). Once the pro-capsids are assembled, the DNA is packaged into them through the herpesvirus portal protein (Fig. 1-5 number 8). Nascent nucleocapsids undergo egress through a three step Envelopment-De-envelopment-Re-envelopment process. In this process, nucleocapsids that contain complete genomes are trafficked to the nuclear membrane where they bud into the perinuclear space (envelopment) (Fig. 1-5 number 9). At this stage, the capsid contains an envelope complete with glycoproteins required for fusion (gB and gH/gL), which mediate the fusion of the enveloped perinuclear capsid with the outer nuclear membrane (de-envelopment), resulting in un- enveloped capsids releasing into the cytoplasm (Fig. 1-5 number 10). Once released into the cytoplasm, it is believed most of the inner tegument proteins are acquired. These

Figure 1-5: Model of Herpesvirus replication cycle. See text in section 1.2.2 for a description of the numbered steps. Adapted from T Mettenleiter et al., Virus Res.

143:222-234, 2009 (113) with permission.

19 inner-tegument proteins assist the capsid in trafficking to tubular vesicles of trans-Golgi network (TGN) or early endosomal (EE) (62) origin, which comprise the site of secondary envelopment (Fig. 1-5 number 11). The capsid then buds into the secondary envelopment vesicle, which provides the viral envelope (re-envelopment) (Fig.1-5 number 12). Once secondary envelopment is complete, the viral particle is surrounded by two membranes and utilizes the cells normal secretory pathway to be released into the environment (Figure 1-5 numbers 13-14).

1.3 Roles of the tegument proteins in Herpesvirus replication

Tegument proteins play a lot of major roles in the replication of herpesviruses from viral targeting to release. In this section, I will go into further detail of the functions of major tegument proteins in these processes with a primary focus on the roles played by the

UL36 and UL37 proteins in addition to the auxiliary capsid protein UL25.

1.3.1 Targeting and uncoating

After fusion of the viral membrane with the cellular membrane, the US3 kinase phosphorylates several outer tegument proteins such as UL48 and VP22, which is believed to result in the release of the phosphorylated proteins from the incoming capsid

(115). This release of outer tegument proteins leaves only a few remaining inner tegument proteins attached to the capsid, among these proteins are UL36 and UL37 (53).

Recent studies implicate both UL36 and UL37 in the process of targeting the viral capsid to the nucleus for uncoating. A PRV UL37-null virus is still able to traffic capsids to the nucleus during viral entry but does so with reduced kinetics (86). Additionally, the UL37

20 interaction with dystonin has been shown to be important for directed capsid movement toward the nucleus during entry (107). In the case of the UL36 protein, the recent interaction observed between this protein and components of the dynein motor complex, shed light onto roles for UL36 in retrograde trafficking towards the nucleus (179). Thus, both the UL36 and UL37 proteins appear to play major roles of the trafficking in incoming capsids to the nucleus of the infected cell, although in some cases UL36 appears to be playing a dominant role in this process (133).

Once the capsid arrives at the nuclear pore it appears a number of protein-protein interactions mediate the docking of the capsid and subsequent uncoating of the viral

DNA. Two of these protein-protein interactions are between viral proteins UL25 and

UL36 interacting with nucleoporins CAN/Nup214 (122) and Nup358 (36) respectively.

Both of these interactions were observed in live cells and abrogation of these interactions by either antibodies or siRNA treatment of the cells to decrease levels of the nucleoporins prevented capsids from docking at the nuclear pores, and thus release of DNA into the nucleus. The UL36 and UL25 proteins may both also play active roles in the delivery of

DNA into the nucleus. In the case of UL36, the protein may need to be cleaved leaving only a C-terminal 60 kDa region of the protein still attached to the capsid, which signals uncoating (76); however, in these experiments there is still so much full-length UL36 present in these fractions that it is difficult to interpret whether the cleaved product is important for genome uncoating or just a random event. With UL25, it has been observed that this protein is able to Co-IP with the portal protein (UL6) at early times of infection

(122), indicating that it may be playing an active role in the uncoating and release of the

21 genome into the nucleus. Additionally, HSV-1 UL25 point mutant E233K showed a reduced ability to uncoat at nuclear pore (128). However, none of these experiments have elucidated the mechanisms by which UL25 or UL36 may be required for DNA to be delivered into the nucleus, and further analysis of these processes are needed.

Upon release of herpesvirus DNA into the nucleus, the viral genes are expressed and the viral DNA is replicated. Additionally, proteins are translated in 3 phases (immediate early, early, and late). Initiation of immediate early gene expression is regulated by the

UL48 protein in HSV-1, through formation of a transcription activation complex with several cellular factors (159). The proteins produced by immediate early genes are involved in viral DNA replication and early gene expression. Ultimately, proteins translated from mRNA of early genes activate transcription and translation of late gene products, which consist predominantly of structural proteins, including many tegument proteins such as UL36 and UL37 (150). After synthesis of these late proteins the herpesvirus capsids can then start to be assembled in the nucleus of the infected cell.

1.3.2 Capsid assembly

Formation of the capsid begins with the dodecameric complex of the UL6 portal, ensuring only one portal is present per capsid (26). The major capsid proteins, VP5, and a several minor ones (triplex proteins UL18 and UL38) are next added to the portal and the capsid is assembled around a scaffolding protein (UL26.5), which is later degraded by the

UL26 protease to create the cavity for the DNA genome. Once the capsid is built, DNA is packaged into the empty capsid through the portal. During this process, UL25 is believed to become capsid associated, as DNA containing nucleocapsids contain more UL25 than

22 capsids containing scaffolding protein or empty capsids (32). Cryo-EM analysis suggests the CVSC (UL25 and UL17) attaches to the pentameric vertices of the capsid and possesses a five-fold symmetry around them (35, 65, 164). The crystal structure for a

UL25 fragment (residues 133-580) has been elucidated (15), but it is lacking the 40 N- terminal amino acids, which interact with UL17 (31). Despite the known crystal structure, the exact requirement for the UL25 protein during nucleocapsid assembly remains unclear, but it is believed to be required for stabilization of DNA containing capsid. This belief based on UL25 mutants that are able to package DNA, but do so aberrantly, often times not packaging entire HSV-1 genomes (160). This also may suggest the UL25 plays a role in signaling to the packaging machinery that a complete genome is enclosed, although it doesn’t likely play an active role in the genome encapsidation process.

Recently reported interactions with the UL31 protein have highlighted a role of the UL25 protein in delivery of DNA-filled capsids to the nuclear egress complex UL31/UL34 (176,

177).

There are studies which suggest that HSV-1 and/or PRV UL36 and UL37 may also be present on nuclear capsids, however this remains quite controversial. First fluorescently tagged UL36 and UL37 proteins are found in both the cytoplasmic and nuclear fractions during the course of infection (33, 109) . HSV-1 UL36 also contains a well conserved nuclear localization signal (NLS) at its N terminus (1, 2). This NLS maps to 425-444 region of UL36 and is required to complement UL36 null virus. Moreover, fluorescently tagged UL36 containing the NLS co-localizes with capsid proteins in the nucleus (2).

Meanwhile, HSV-2 UL37 has a consensus LXXXLXXLXLX nuclear export motif at

23 residues 263-273, which is conserved throughout alphaherpesviruses. Truncation of the first 400 amino acids of the N-terminus of UL37, deletion of nuclear export motif, or mutation of any of the leucines to Q/R greatly increases amount of UL37 retained in nucleus (172). The presence of these nuclear signals indicate that at least a fraction of the

UL36 and UL37 protein may need to be in the nucleus. Additionally, in a fractionation study, UL36 was observed with DNA-containing nucleocapsids while UL37 was observed on both empty and DNA containing nucleocapsids (18). It is unclear how UL37 would attach to empty capsids as UL37 localization to capsids requires the presence of

UL36. In contrast to this study several other groups using various techniques such as immuno-electron microscopy (54) in PRV and mass spectrometry (129) in HSV-1 have failed to detect either protein associated with nuclear capsids. Thus the location (nuclear or cytoplasmic) of the start of the viral tegumentation process remains controversial.

1.3.3 Nuclear Egress

In order for nucleocapsids to escape the nucleus they are required to traverse the nuclear membrane (Fig 1-6). To accomplish this, herpesviruses first bud into the perinuclear space where they obtain their initial envelope and this process is known to be mediated by the UL31/UL34 complex (130) in a manner reminiscent of matrix protein mediated budding of influenza A at the plasma membrane. The UL25 protein has been implicated in delivering DNA-filled capsids to the UL31/UL34 complex (176, 177); thus selecting for predominantly completed capsids to undergo nuclear egress. Once the UL31/UL34- mediated budding is complete the virion contains an envelope complete with glycoproteins required for fusion, but likely lacking most tegument proteins, with the

Figure 1-6: Herpesvirus egress from infected cells. The egress process from nuclear egress to release from the plasma membrane is displayed. See the text in sections 1.3.3 and 1.3.4 for a more in depth description of the processes. Adapted from Johnson D and

Baines., Nat. Rev. Micro. 5:382-94, 2011 (74) with permission.

25 exception of US3 (54). In the case of HSV-1, the US3 kinase is responsible for phosphorylation the cytoplasmic domain of gB and regulates its fusion activity with the outer nuclear membrane (173). This fusion event is required to release un-enveloped capsids into the cytoplasm (Fig 1-6).

There also exists a controversy on whether or not the inner tegument proteins play a role in this process. In most studies, herpesviruses lacking either UL36 or UL37 aggregate only in the cytoplasm and have no trouble escaping the nucleus (41, 42, 82, 104).

However, in some cases for both PRV UL36-null virus (104) and HSV-1 UL37-null virus

(41) capsids are also observed to have defects in nuclear egress. In both of these studies, major defects in the ability to undergo primary envelopment were observed, with up to

80% of nascent capsids being retained in the nucleus (41). In the case of UL37, there was a recent study looking at UL37 null-viruses of HSV-1 and PRV viruses in 2 different cell lines within the same lab and no defects were observed in nuclear egress for either virus (96). Thus, it is unlikely that UL37 is actually involved in the process of nuclear egress. However, it should be noted that the “null” virus in these studies were not clean deletions, but expressed small 50-100 amino acid fragments of the protein. A similar study needs to be done for UL36 to help answer the question of whether or not this protein is involved in the process of nuclear egress. The differences observed between labs may be simply due to the time-points used as every paper seems to use differing times post infection to observe the virus at this early stage in egress (additional differences may also have to do with cell-types being used).

1.3.4 Herpesvirus trafficking and secondary envelopment

One of the main roles of the inner tegument proteins UL36 and UL37 appears to be trafficking to the site of secondary envelopment. In the absence of these proteins HSV-1 and PRV viral capsids tend to aggregate in the cytoplasm (41, 42, 82, 104), and the release of infectious virions is either completely blocked, as seen in HSV-1, (41, 96) or is strongly impaired, as observed in PRV (82, 96). It was also observed that PRV viruses lacking either UL36 or UL37 had defects in directed movement (104, 133) to additional nuclei in syncytia (133) or to sites of secondary envelopment (104), with the defect in

UL36-null virus being more severe. This defect can be imitated using wild type-virus in the presence of nocodazole, suggesting that the capsids are being trafficked along the microtubule network (104). It is suggested that the inner-tegument proteins, likely one of

UL36 or UL37, is involved in binding of the cellular microtubule motors. Consistent with this are the recent interactions found for UL36 with dynein motor components (179) and

UL37 with dystonin (123). Additionally, capsids stripped of most outer tegument proteins and left predominantly with inner tegument proteins UL36, UL37, US3 kinase and

UL21(162) have the ability to bind several microtubule motors simultaneously such as kinesin-1, kinesin-2, dynein, and dynactin (129). How the capsid is able to traffic properly given that it can possess both microtubule motors that can undergo both anterograde and retrograde trafficking needs to be further investigated. It has been observed, however that capsid trafficking is not fluid in either direction and capsids will often stop and reverse directions for brief periods of time before resuming transport in the desired direction (129). Most recent evidence indicates that both the UL36 and UL37 proteins are playing critical roles in the trafficking of capsids to the site of secondary

27 envelopment (140). Although both of these proteins are important for this process, we currently do still not have a good understanding as to why both proteins are required to bind host-cell proteins involved in trafficking. Hopefully, future studies on these proteins in viral egress will provide answers to some of these remaining questions.

When the capsid arrives at the site of cytoplasmic envelopment, it accumulates the outer tegument proteins as well as the envelope glycoproteins during the secondary envelopment process. The acquisition of the rest of the tegument is thought to be due to a variety of protein-protein interactions. For example, the HSV-1 UL48 protein requires

UL36 to be incorporated into virions, and once bound to UL36, the UL48 protein appears to be partly required to bring the UL47 protein into the tegument as well (48, 83). The

UL37 protein of PRV is also involved in acquisition of the secondary envelope and tegument as evidenced by lack of gE and UL49 proteins in extracellular virions of UL37 mutant virus (82). The actual mechanistic process of secondary envelopment remains a mystery, but it is known to require several outer tegument proteins, such as UL11 and

UL16 and the cytoplasmic tails of glycoproteins, such as gE (10, 59, 111, 112, 158, 178).

Both UL36 and UL37 are required for HSV-1 to undergo secondary envelopment, and lack of either protein usually results in accumulation of capsid aggregates in the cytoplasm (41, 42). However, this appears to be predominantly due to the inability to traffic to the sites of secondary envelopment. As previous reports have shown that UL37 is not required for the formation of the so-called light or L-particles, which lack capsids and contain only enveloped tegument (82, 133), it seems not to be required for the

28 assembly of the outer tegument or for membrane deformation during cytoplasmic budding.

1.3.5 UL36 and UL37 in pathogenesis

The UL37 and UL36 proteins are known to play critical roles in viral replication in cell culture. However, many recent studies have also uncovered roles for these proteins in pathogenesis in infected animals. In UL36, the DUB domain has been shown to be a critical determinant in neurovirulence in infected mice (14, 95). Furthermore, the TRAF3 protein was recently identified as a substrate of the UL36 DUB domain, with the deubiquitination of this protein leading to an inhibition in production of beta interferon

(171). PRV UL37-null virus is still able to replicate to low levels in cell culture indicating that the PRV UL37 protein is not essential for the basic mechanisms of viral replication

(82, 96). In contrast, mice infected with UL37-null PRV virus do not exhibit any signs of neuronal infection, suggesting that the UL37 protein plays a major role in neuroinvasion and neurovirulence in vivo (79, 80) although it is possible this decrease in virulence may be solely due to a slightly reduced ability to replicate. Further, the HSV-1 and KSHV homologues of UL37 have both been implicated in regulation of the immune system during infection as HSV-1 UL37 activates the NFkB pathway early in infection through interaction with TRAF6 (99), and KSHV ORF63 inactivates the inflammasome response through interactions with NOD2, NLRP1, and NLRP3 (55). These experiments reveal additional roles of the inner tegument in promoting virulence through regulation of the immune system during infection.

1.4 How eukaryotic cells regulate vesicle trafficking

Enveloped viruses must interact with host cell membranes at either the plasma membrane or internal membrane sites to acquire their viral envelope. Given that viruses often co-opt host systems to aide in their replication (i.e. HIV Gag recruitment of ESCRT proteins to mediate scission of budding particles); it is of interest to know how eukaryotic cells mediate budding, trafficking, and fusion of their membranes. These systems could potentially be recruited and/or mimicked by viruses during replication.

Vesicular transport within cells is a very tightly regulated process, which is coordinated by a large number of proteins. Among the most critical proteins in these processes are

Rab proteins - small-GTPases, which are the master regulators involved in nearly all steps of vesicular trafficking, from vesicle formation and movement to vesicle uncoating, tethering and fusion (66). As GTPases, Rabs can be found in both the inactive GDP bound state and the active GTP bound state. Membrane bound Rabs are predominantly in the active GTP state, which allows these proteins to interact with effectors that aide in regulation of vesicular trafficking. In addition to Rab proteins, several protein complexes also are involved in vesicular trafficking. For example, the COP and Clathrin coat proteins play active roles in vesicle formation, while multisubunit tethering complexes

(MTCs) are involved in tethering transport vesicle to their target organelles and SNARE

(NSF attachment protein receptor) proteins, which is a critical interaction for fusion of vesicles with the target organelle. There are an abundance of interactions between these various classes of proteins, which assures that vesicles are being trafficked and fusing to

30 the proper organelles. In this section I will briefly go into the roles that these proteins play in mediating proper vesicular trafficking within eukaryotic cells.

1.4.1 Vesicle Formation

Intracellular membranes are often coated with COPI, COPII, clathrin, or other coat proteins, and the nascent vesicles are initially formed based upon recognition of proteins that mark the cargo for transport by components of the coat complexes. In clathrin and

COP mediated vesicular formation, coat proteins most often recognize members of a family of small GTPases, the Sar/Arf proteins, which mark membranes for cargo vesicle formation and trafficking. However, Rab proteins are also known to play roles in vesicle selection with other coat proteins. An example is the Rab7 dependent recruitment of retromer coat trimer Vps26-Vps29-Vps35 to endosomal membranes, initiating retromer vesicle formation from early endosomes membranes, which are destined to be trafficked to the trans-Golgi network (TGN) (136).

1.4.2 Vesicle Transit

Several Rabs can interact with proteins that are involved in trafficking along actin or microtubule networks. The Rab11 protein mediates a complex between the vesicle it is attached to with myosin Vb by recruitment of the myosin protein with the effector Rab11 family interacting protein 2 (Rab11-FIP2), once recruited the myosin vesicle is trafficked to the plasma membrane for recycling (58). Rab6 is observed to have a similar function of recruiting Rab6 containing vesicles to dynein motor complexes through interaction of its effector protein Bicaudal D1/D2 (56, 106). In addition to Rab proteins interacting with cellular motor proteins, multisubunit tethering complexes have also been shown to recruit motor proteins, such as the DSL1 complex recruitment of dynein to the

31 kinetochore during nuclear segregation (45, 146), or the recruitment of myosin motors to plasma membrane mediated by the exocyst complex (72). These types of interaction between Rab proteins or MTC with molecular motors make up a small subset of ways that vesicles are trafficked between sites of vesicle formation and destination organelles for tethering and ultimately fusion.

1.4.3 Vesicle Tethering and Fusion

Tethering of transport vesicles to their destination organelles is predominantly carried out by Multisubunit tethering complexes (MTCs). As the name implies MTCs are complexes comprised of between 3 and 10 proteins that can be divided into three classes, the

CATCHR (complexes associated with tethering containing helical rods) family, the

CORVET/HOPS family, and the TRAPP family (Fig. 1-7) (16). The CATCHR family of

MTCs is predominantly involved in the tethering and fusion of vesicles to organelles in the secretory pathway. The structures of the components of these CATCHR MTCs are remarkably similar, especially considering that they possess very little sequence similarity. The CORVET/HOPS family of MTCs is involved in fusion of vesicles to organelles of the endolysosomal pathway (predominantly early endosomes and multi- vesicular bodies). Finally, the TRAPP family MTCs actually a multi-subunit guanine exchange factor that acts upon several Rab proteins in addition to tethering vesicles via

COPI and COPII coat proteins to either the Golgi or the lysosome.

Figure 1-7: Multi-subunit tethering complexes of eukaryotic cells. An overview of the vesicular trafficking pathways mediates by MTCs. The MTCs are in orange and red at membrane organelles consistent with their localization in cells. The arrows pointing to the MTC are indicating the origin of transport vesicles in which the MTC recognizes for tethering and fusion with to target organelle. Adapted from C. Brocker et al., Curr. Biol.

20:943-952, 2010 (16) with permission.

MTC localize to various organelles of the secretory and endolysosomal pathways and tether vesicles of specific origins to their destination organelles, leading to fusion of the vesicle with the organelle (Fig 1-7). Localization of MTCs is mediated by the binding of specific components of the MTC most often to SNARE proteins on the organelle surface.

For example, the Dsl1 complex is known to interact with 3 t-SNARE proteins, which are present on the ER membrane, Sec20, Ufe1, and Use1 (156). The interaction of Dsl1 complex with these three proteins aides in localizing the Dsl1 complex to the ER and allows the complex to interact with incoming vesicles of Golgi origin destined for fusion with the ER.

Just as the localization of MTCs is dependent upon interactions of components of the

MTCs with specific (often t-SNAREs) on the organelle surface, the tethering of transport vesicles to the organelles is also dependent upon specific interactions. Incoming vesicles are decorated with specific subsets of coat proteins as well as Rabs and SNAREs. Only vesicles containing the correct proteins on the surface can be bound by the MTCs and tethered to the organelle. The two most common classes of proteins that MTCs bind to mediate this tethering are Rab and v-SNARE proteins (Fig 1-8). In the case of the Dsl1 complex discussed earlier, incoming vesicles containing the SNARE protein Sec22 can tether to ER through an interaction with the Dsl3 (or Sec39) protein of the Dsl1 complex.

In addition to the Sec22 interaction – the Dsl1 complex can also recognize the COPI coat proteins via the Dsl1 protein (5, 6), thus there are often several mechanisms by which the

MTCs can mediate tethering to the vesicle. The MTC also is believed to help assemble the SNARE complex, essentially priming the vesicle for fusion with the destination

Figure 1-8: Mechanism of action of Multi-subunit tethering complexes in membrane fusion in eukaryotic cells. Targeting and tethering of vesicles to organelle membranes is completed using coordinated interactions between MTC with SNARE proteins at the organelle and MTC interactions with Rab or SNARE proteins on the incoming transport vesicle. Ultimately the tethered vesicles will fuse with the target organelle after assembly of the SNARE complex. Adapted from C. Brocker et al., Curr. Biol. 20:943-952, 2010

(16) with permission.

35 organelle (156). All of the other MTCs (other than Dsl1) are known to be able to interact with at least one Rab protein, which can also mediate tethering of the vesicle (Fig 1-8).

An example of this is the interaction between Sec15 component of the Exocyst complex with Rab11 (174).

As with many of the effector interactions mediated by Rab proteins, the Rab11 protein needs to be in the active GTP bound state in order to bind Sec15 (174). This appears to be the case for most of the interactions between MTC components and Rab proteins. It is not unexpected as the Rab proteins in the active GTP bound state the predominant state that is membrane bound (as GDP bound inactive Rabs are often removed from the membranes via GDI (guanine nucleotide dissociation inhibitor) proteins). Similarly to the

Dsl1 complex aided fusion described previously, upon tethering of the vesicle to the organelle the rest of the MTC of the CATCHR family also aide in assembly of SNARE complexes ultimately leading to fusion of the two membranes.

1.4.4 Vesicle Trafficking and Microbes

Pathogens that replicate within host cells, such as the bacteria Legionella pneumophila and all viruses often utilize host pathways to manipulate the intracellular environment in ways that promote their own replication. Several of these intracellular pathogens are known to interact with components of vesicular trafficking to this cause. In the case of

Legionella pneumophila the bacterial secreted effector protein LidA, is known to interact with several host cell Rab GTPases (Rab1, Rab6A’, and Rab8A) (105, 148). The LidA interaction with Rab6A’ blocks hydrolysis of GTP to GDP, which was beneficial to the

36 intracellular growth of L pneumophila, as expression of constitutively active Rab6A’ effectively reduced the ability for L pneumophila to replicate (27). Although the exact advantage that is gained through this blocking is currently not understood, it is clear that the pathogen is using this interaction to promote its intracellular replication.

In addition to bacteria taking advantage of proteins involved in vesicular trafficking to promote their replication, there is an increasing number of studies finding similar mechanisms used by viruses during infection. For example the plant virus Cauliflower mosaic virus, utilizes Rab GTPase (ARA7) decorated vesicles to traffic its movement proteins (MPs) throughout the cell, and the proper trafficking of these proteins is critical for virus replication (22). Additional studies have found that the HCMV protein pp150 interacts with Bicaudal D1 and Rab6, and that these interactions are essential for proper intracellular localization of the pp150 protein and infectious virus production (67).

It can easily be imagined that interaction with other components of these pathways for example SNAREs or COP coat proteins or potential mimicry of them could greatly enhance the ability for intracellular pathogens to transit inside the cell. This trafficking and potential manipulation of these host processes could be quite beneficial, especially for viruses that may need to co-opt host systems or interact with vesicular components during envelopment and/or release of viral particles.

1.5 Major unanswered questions in Herpesvirus Egress

Genetic and virus based assays have provided a lot of information on the proteins involved in herpesvirus egress from infected cells. Interaction studies have additionally started to provide information about how viral proteins work in concert to mediate egress from the cell. Furthermore, the identification of interactions between viral and host proteins have highlighted how some of these protein utilize host processes to mediate trafficking of viral capsids. However, despite continually acquiring more information about which proteins are involved in the process of virus assembly and egress, we know surprisingly little about the mechanisms behind how these proteins function.

The inner tegument proteins UL36 and UL37 are conserved throughout the herpesviridae family. Although we are starting to get a clearer picture that these two proteins are involved in early events of tegument acquisition and viral capsid trafficking based upon interactions with host trafficking proteins, the exact roles that these proteins are playing remains largely unknown. To gain insight into the mechanisms by which these proteins enable trafficking and secondary envelopment, we sought to biochemically and structurally characterize both proteins and their interactions. As the UL36 protein is so large, we opted to investigate primarily the interaction domains with the UL25 and UL37 proteins. Structural and biochemical information about these proteins and their complexes would provide both valuable insight into how tegumentation and assembly of herpesviruses occurs, and also novel targets for the design of more effective therapeutics.

Chapter 2: Materials and Methods

2.1 Cloning

2.1.1 Cloning of UL36 constructs

The truncated construct for the HSV-1 UL36 cbd (50) was generated using the pGS3334 plasmid as a template with primers 5’-

CTAGGGATCCGGATCTTACGTGCAACGCACC

3’- CTAGAAGCTTTTAGCCCAGTAACATGCGCAC. The resulting product was then digested with BamHI and HindIII and inserted into pET24b resulting in the plasmid pJP50. For plasmids containing the PRV UL37 binding site of UL36 (344-561), several constructs were generated, the first using the endogenous PRV gene from the Becker strain as the template. PCR products were made using pGS1521 as a template and the primers 5’- CTAGCAATTGCTGCAGGCCCACGCCGACGAG

3’ – CTAGGCTAGCTTAGTTCGACTCGGCCGCGCC and were ligated into pET24b using the MfeI and NheI restriction enzymes, resulting in the plasmid pJP52. Additional constructs were made using a codon optimized gene made by GeneArt. pJP53 was made by cutting the codon optimized gene from the GeneArt plasmid with BamHI and HindIII and ligating it into pKH89 producing a Strep-SUMO tagged UL36(344-561) construct.

An additional construct, pJP54 was made by cutting pJP53 with BamHI and XhoI restriction enzymes and ligating the product into pGEX-6P-1 resulting in an N-terminal

GST-fusion to PRV UL36(344-561). All UL36 constructs were sequence verified prior to use.

2.1.2 Cloning of UL37 constructs

Plasmid pGS3610 (a gift from Greg Smith) encodes the PRV-Becker UL37 gene fused to an N-terminal His6-SUMO tandem tag. This was made by cutting the pETDuet-

SUMO vector (a derivative of pETDuet-1, and a gift from Dr. Thomas Schwartz) and the pGS1740 subclone of UL37 with BamHI and HindIII. The pJP10 construct (for Sf9 expression of UL37 was made through restriction digest of pFastBac1 and pGS3610 with

BamHI and HindIII, and ligation of the UL37 gene into the cut pFastBac1 vector. The pJP4 plasmid, which contains a His-SUMO-PreScission tag in frame with the BamHI restriction site of the multiple cloning site in a pET24b vector, was made through PCR of the His-SUMO-PreScission tag from pETDuet-SUMO using the primers 5'-

GGGAATTCCATATGGGCAGCAGCCATCACCATCA and 3'-

CTAGGGATCCGGGCCCCTGGAACAGAACTT (restriction sites are underlined). The

PCR product was subcloned into the pET24b using NdeI and BamHI restriction sites. The codon optimized PRV UL37 gene for E. coli expression was synthesized by GeneArt.

The His-SUMO tagged full-length codon optimized PRV UL37 (pJP14) was made by cutting the plasmid synthesized by GeneArt and the pJP4 plasmid with BamHI and

HindIII and ligating the UL37N gene product into the cut pJP4 plasmid.

For the UL37 N-terminal constructs (UL37N (1-476) and UL37N (1-496)) of codon- optimized PRV UL37, the N-terminal fragments were was amplified by PCR pJP14 plasmid using the forward primer 5'-CTAGGGATCCATGGAAGCACTGGTTCGTGC and the reverse primers 3'-CTAGAAGCTTCTAGGCTGCGCTGGTCGGTG and 3’-

CTAGAAGCTTTCAACGGCTCTGTGCACGAA (restriction sites are underlined). The

PCR products were then subcloned into pJP4 using the BamHI and HindIII restriction sites to yield plasmid pJP23 (UL37N496) and pJP27(UL37N476), respectively.

Plasmids containing genes for UL37C terminal protein were constructed identically as

UL37N terminal constructs. The primer sets used for amplification of UL37 C-terminal constructs were the reverse primer 3’ –

CTAGAAGCTTTCATTTGCTAACGGCATCAA and the forward primer 5’ –

CTAGGGATCCGGTCTGCGTGCAGATGGTGCC, 5’-

CTAGGGATCCAGCGCGGCCCTGGTTGATCT, and 5’-

CTAGGGATCCGATCTGGCAGCCGCAGCCGAT. These primer sets and subsequent subcloning into the JP4 plasmid cut with BamHI and HindIII resulted in the generation of pJP35 (UL37C(478-919)), pJP31 (UL37C(494-919)), and pJP57 (UL37C(499-919). Two additional constructs were made, which started at amino acid 499 were created with truncated C-termini stopping at amino acid 831 and 856. These constructs used the 499 forward primer listed above and reverse primers of 3’ -

CTAGAAGCTTTCATTATGCAGGCGGTGCCTGTTCTTC and 3’-

CTAGAAGCTTTCATTAACCATCACCGGCAACTTCATG generating plasmids pJP58 and pJP59, respectively.

Generation of both the V498D and V498G mutant constructs were done using splicing by overlap extension (SOE) PCR (60). For generation of the 5’ SOE PCR fragments from the pJP14 plasmid for mutagenesis of the V498 site, the forward primer used was 5’-

CTAGGGATCCATGGAAGCACTGGTTCGTGC while the reverse primers were 3’-

TGCCAGATCATCCAGGGCTGC (V498D) and 3’ –

TGCCAGATCACCCAGGGCTGC (V498G). The 3’ SOE PCR fragments were generated using the reverse primer 3’ –

CTAGAAGCTTTCATTTGCTAACGGCATCAA and the forward primers 5’-

GCAGCCCTGGATGATCTGGCA (V498D) and 5’- GCAGCCCTGGGTGATCTGGCA

(V498G). The resulting PCR products were subjected to an additional round of PCR creating a product containing the full-length UL37 gene with the desired mutation. This

PCR product was then sub-cloned into pJP4, resulting in the plasmids pJP55 (V498D) and pJP56 (V498G). The plasmid for mammalian expression of the UL37N496-GFP fusion protein was also generated using SOE PCR using the pGS1499 plasmid as a template. For generation of the 5’ SOE PCR fragment the forward primer that was used was 5’- GCATACGCGTACGTGAGCGAGCT and the reverse primer was 3’-

CTCGCCCTTGCTCACCGCCGCGCTCGTCGG. The 3’ SOE PCR was produced using the primers 5’- CCGACGAGCGCGGCGGTGAGCAAGGGCGAG and 3’ –

GCAGGCACGCGTCCACGGCGTC. Additional PCR of using the two fragments generated above resulted in a gene product, which spliced out the 3’ half of the UL37 gene, resulting in an frame fusion between UL37N and GFP, which was sub-cloned into pGS1499 (a gift from Greg Smith) using endogenous MluI restriction sites. All UL37 constructs were sequence verified prior to use.

2.2 Protein expression and purification

2.2.1 E. coli expression and purification

All proteins were initially expressed and purified from T7 Express (New England

Biolabs) or Rosetta E. coli strains. Freshly transformed cells were incubated at 37C overnight in 5 mL Luria Broth (LB) starter culture supplemented with the required antibiotics. The starter culture was diluted into 1 L LB supplemented with antibiotics and

43 grown at 37C until the OD600 reached 0.8-1.0. At this point, the temperature was shifted to 16C and the cells were induced with 0.5 mM isopropyl-B-D-thiogalactopyranoside

(IPTG) and the proteins were incubated on a shaker for 16-20 hours. Cells were harvested by centrifugation at 12,000 g for 40 minutes, resuspended in 25 mL 20 mM Tris 8.0, 150 mM NaCl, 0.1% IGEPAL CA-630 (Sigma), 5% glycerol, 0.5 mM TCEP, and 1 tablet of

EDTA-free “Complete” protease inhibitor cocktail tablet (Roche), and flash frozen in liquid nitrogen and stored at -80C until ready for purification. To begin purification the resuspended pellets were thawed and then lysed by French Press. The insoluble fraction was removed by centrifugation of the whole cell lysate at 14,000 g for 30 min at 4C.

Soluble lysate was then loaded onto the appropriate affinity column and subsequently washed 3 times with 10 CV of wash buffer (20 mM Tris pH 8.0, 150 mM NaCl, 0.5 mM

TCEP). Proteins were eluted as recommended by manufacturers and the protein was concentrated and the buffer exchanged to wash buffer using Ultra-15 concentrators

(Millipore). Protein purity was assessed by Coomassie G-250 staining after SDS-PAGE.

As a polishing step the proteins were run over S200 size exclusion columns equilibrated with wash buffer and the peak fractions harvested.

2.2.1.1 UL25/UL36 purification

All UL25 and UL36 proteins were expressed in Rosetta cells as described above.

Attempts to co-purify both the UL25 and UL36 proteins were done by running lysates over either glutathione resin (for purification using the GST-UL25 protein) or Ni-NTA

(for purification based on the His-SUMO-UL36 protein). For glutathione purification after the lysates were loaded the columns was extensively washed 3 times with 10

44 column volumes of wash buffer (20 mM Tris pH 8.0, 150 mM NaCl, 0.5 mM TCEP) and eluted with 20 mM reduced glutathione in wash buffer. For Ni-NTA purification the 3 washes consisted of wash buffer with increasing amounts of imidazole (10 mM, 25 mM, and 40 mM) and the protein(s) were eluted using wash buffer containing 300 mM imidazole. Regardless of the column used the eluted proteins were immediately concentrated and the buffer exchanged into wash buffer using an Ultra-15 30-kDa-cutoff concentrator (Millipore). Protein purity and complex formation was evaluated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie G-250 staining and subsequent polishing steps on an S200 size exclusion column.

2.2.1.2 UL37 purification

All UL37 proteins were expressed in T7 express cells as described above. Briefly, all were induced with 0.5 mM IPTG at an optical density between 0.6-1.0 and incubated overnight at 16°C. Cells were harvested by centrifugation for 30 minutes at 4000 x g and the supernatant discarded. For the UL37N496 protein the cells were resuspended in 25 ml lysis buffer (20 mM PIPES (pH 7.0), 50 mM NaCl, 0.5% IGEPAL-630, 5% glycerol, and

0.5 mM TCEP and a protease inhibitor tablet (Roche)) and lysed using French press.

Lysates were next centrifuged at 15,000 x g for 30 minutes and the supernatant fractions collected. Soluble lysate was loaded onto a 5 mL Ni-Sepharose 6B FF column (GE

Healthcare). The column was subsequently washed with 10 column volumes (CV) of 20 mM PIPES pH 7.0, 50 mM NaCl, 0.5 mM TCEP (buffer A) containing increasing amounts of imidazole, 10 mM or 25 mM. Protein was eluted in buffer A containing 100 mM Imidazole. The eluate was immediately concentrated and the imidazole removed by

45 buffer exchange into buffer A in an Ultra-15 50-kDa-cutoff concentrator (Millipore).

Protein concentration was determined from 280-nm absorbance using a calculated extinction coefficient. GST-tagged-PreScission protease was added to the protein solution at a 1:50 protease: protein ratio, and the protein was cleaved overnight at 4 C to remove the His6-SUMO tag. The protease-protein solution was applied sequentially to glutathione sepharose 4B (GE Healthcare) and Ni-Sepharose 6B to remove the GST- tagged PreScission protease and the His6-SUMO tag, respectively. Cleaved protein was present in the unbound and wash fractions. UL37N was further purified by size-exclusion chromatography using a Superdex 200 column (GE Healthcare) and concentrated to 3.5-

4.0 mg/mL using an Ultra-15 30-kDa-cutoff concentrator (Millipore). Protein purity was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and

Coomassie G-250 staining. The final yield was 15-20 mg of pure protein per 1 L of E. coli culture. All UL37N protein samples used for crystallization and biochemical studies were stored in 20 mM PIPES pH 7.0, 50 mM NaCl, and 0.5 mM TCEP.

For all other UL37 proteins (full-length, mutants, and C-terminal domains) the purification scheme was nearly identical to that for UL37N. However for the UL37C terminal constructs, all proteins were purified in optimal buffers for UL37C protein, which contained 20 mM HEPES 7.5 and 500 mM NaCl. Subsequent storage buffers for

UL37C were 20 mM HEPES pH 7.5, 300 mM NaCl, and 0.5 mM TCEP. For full-length

UL37 constructs the buffers consisted of 20 mM Tris-HCl (pH 8.0) and 150 mM NaCl.

Other than these differing buffer conditions (used throughout purification) the overall scheme of purification is the same for all UL37 protein constructs. Final yields for

UL37C terminal constructs were often between 6-10 mg of pure protein per liter, while full-length proteins were substantially less at 0.2-0.4 mg/L.

For purification of the PreScission protease used to remove the His-SUMO tags a BL21 E. coli strain expressing GST-tagged PreScission protease was a gift from Peter Cherepanov

(London Research Institute, UK). Protein expression was induced with 0.5 mM IPTG at

30 ºC for 4 hours before the cells were harvested and lysed. The PreScission protease was purified over glutathione sepharose in a buffer containing 20 mM Tris pH 8.0, 200 mM

NaCl, and 1 mM TCEP. The column was washed 3 times with 10 column volumes (CVs) of the binding buffer and protein eluted from the column in binding buffer containing 5 mM reduced glutathione. The eluted protein was concentrated in a 30-kDa-cutoff concentrator (Millipore) and further purified over a Superdex 200 (S200) size exclusion column equilibrated with the binding buffer. The protein was concentrated to 1 mg/mL, flash frozen, and stored at -80 ºC.

2.2.1.3 Purification of Rab proteins, Sec22b, and LidA

Sec22b, LidA, and all of the Rab proteins were expressed as described above (Section

2.2.1) using the Rosetta strain of E. coli. Briefly, all were induced with 0.5 mM IPTG at an optical density between 0.6-1.0 and incubated overnight at 16°C. Cells were harvested by centrifugation for 30 minutes at 4000 x g and the supernatant discarded. The cells were resuspended in 25 ml lysis buffer containing (20 mM HEPES (pH 7.5), 100 mM

NaCl, 0.5% IGEPAL-630, 5% glycerol, and 0.5 mM TCEP) and lysed by French Press.

The lysates were then centrifuged at 15,000 x g for 30 minutes and the supernatant

47 fractions collected. As all of these proteins contained a GST tag, the lysates were loaded onto a glutathione sepharose column. The column was washed 3 times with 10 column volumes of wash buffer (20 mM HEPES (pH 7.5), 100 mM NaCl, 0.5 mM TCEP) and the protein eluted with wash buffer containing 10 mM reduced glutathione. All of the proteins were concentrated on Ultra-15 30 kDa cut-off concentrators (Millipore) immediately after elution from the columns. The buffers were exchanged to wash buffer and the proteins were subjected to gel filtration over an S200 column as a polishing step.

The peak fractions containing mono-disperse protein were pooled and concentrated to

2.5-4.0 mg/ml and the proteins stored at 4°C.

2.2.2 Insect cell expression and purification

All insect cell expression was performed in Spodoptera frugiperda (Sf9) cells, which were grown in suspension in SF-900 II SFM medium (Invitrogen) at 27°C. Recombinant baculoviruses were generated for all gB constructs using the Bac-to-Bac system

(Invitrogen). Sf9 cells were transfected using lipofectamine and the recombinant bacmid

DNA and the first passage (P1) virus stock was harvested and filtered through a 0.22 micron filter at 72 hours post-transfection. The virus underwent an additional two rounds of amplification and the third passage (P3) viral stocks of baculovirus were harvested, filtered, and stored at 4°C until ready to use for infection. For infection Sf9 cells were infected with 7.5 ml of baculovirus P3 stock per liter of insect cell culture.

2.2.2.1 UL25, UL36, and UL37 purification

All of these proteins were purified at 72 hours post infection with P3 baculovirus. The cells were harvested by centrifugation for 30 minutes at 4000 x g and the supernatant discarded. The cells were respuspended in 25 ml lysis buffer (20 mM Tris-HCl (pH 8.0),

150 mM NaCl, 0.5% IGEPAL-630, 5% glycerol, and 0.5 mM TCEP) and lysed by

French press. The lysates were then centrifuged at 15,000 x g for 30 minutes and the supernatant fractions collected.

The UL25 protein possesses a GST tag, and thus was loaded onto glutathione sepharose column. The column was washed 3 times with 10 column volumes of wash buffer (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.5 mM TCEP) and the protein eluted with wash buffer containing 10 mM reduced glutathione.

Both the UL36 and UL37 (HSV-1 and PRV) proteins contain and N-terminal His-SUMO tag. Therefore, these proteins were purified over Ni-NTA sepharose. The columns were washed with 3 times with 10 column volumes of wash buffers containing increasing amounts of imidazole (10 mM, 25 mM, and 40 mM imidazole). Proteins were eluted from the Ni-NTA column using wash buffer containing 300 mM imidazole. The complex of UL25 and UL36 was also purified using Ni-NTA following the protocol for purification of UL36 alone.

Each of these proteins was concentrated on Ultra-15 30 kDa cut-off concentrators

(Millipore) immediately after elution from the columns. The buffers were exchanged to

49 wash buffer and the proteins were subjected to gel filtration over an S200 column to assess protein mono-dispersity and purity. The fractions from mono-disperse protein peaks were pooled and concentrated to 2.0-6.0 mg/ml and the proteins stored at 4°C.

2.2.2.2 gB purification

For purification of the secreted gB proteins, the supernatant of the cultures were harvested 3 days after infection. The cells were removed through centrifugation for 30 minutes at 4000 x g and the cell supernatant was filtered through a 0.22 micron filer and subsequently applied to tangential flow filtration (TFF) with a 30 kDa cartridge. During

TFF the buffer was exchanged to 1 x PBS and the supernatant was loaded onto a DL16 column to purify gB by immunoaffinity chromatography. The column was then washed with 30-50 column volumes of 10 mM Tris-HCl (pH 8.0) and 500 mM NaCl and the gB protein eluted with 3 M KSCN. The protein was immediately concentrated with an Ultra-

15 50 kDa cut-off concentrator (Millipore) and the buffer exchanged to Tris-HCl (pH 8.0) and 150 mM NaCl to remove KCSN. As a polishing step the concentrated gB proteins were run over an S200 gel filtration column previously equilibrated with the Tris-NaCl buffer. The peak fractions were pooled and the proteins again concentrated to 3.0-4.0 mg/ml and stored at 4°C or flash frozen and stored at -80°C.

2.3 Biochemical characterization

2.3.1 Western Blot Analysis

All western blots were transferred using a Semi-Dry method from a polyacrylamide gel to a nitrocellulose membrane. The membranes were next incubated for 1 hour with 5 %

50 milk in 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, and 0.05% Tween 20 (TBST) for 1 hour at room temperature on a shaker. After 3 five minute washes with TBST the membranes were incubated with the primary antibody (1:2500 anti-His5-HRP conjugated antibody (Qiagen), 1:5000 anti-myc (Pierce), 1:2000 anti-UL25 (gift from Greg Smith),

1:5000, anti-GST (Santa Cruz), 1:2000 anti-GFP (Thermo Scientific), 1:500 anti-Rab (all variants) (Santa Cruz) or 1:1000 Streptactin-HRP (IBA)) in TBST for 1 hour at room temperature. After the incubation with the primary antibody the membranes were again washed 3 times with for 5 minutes with TBST. For blots using antibodies that were not

HRP conjugated, the membranes were incubated with either goat anti-rabbit or goat anti- mouse IgG HRP-conjugated pAbs (1:8,000) in TBST for 1 hour before washing away the secondary with 3 final washes with TBST for 5 minutes each. Following the final wash, all blots were developed using the Pierce western Pico chemiluminescence kit (Bio-Rad).

2.3.2 Electron Microscopy

Purified proteins were diluted into a buffer containing 20 mM Tris pH 8.0 and 150 mM

NaCl to a final concentration of between 10 - 20 μg/ml. Samples were incubated on formvar/carbon coated 200 mesh copper grids (SPI), which had been glow discharged at room temperature. Grids were next stained using a 0.75% solution of uranyl formate. The

EM images were obtained at the Harvard Medical School EM core facility using a Tecnai

G2 Spirit BioTWIN with an AMT 2k charge-coupled device camera microscope. All images were processed and lengths of proteins measured using Adobe photoshop.

2.3.3 Co-Immunoprecipitation

For all Co-IP studies 1-5 ml of bacterial or insect cell cultures expressing proteins of interest were harvested at times consistent with harvest times for protein purification (12-

16 hours for bacterial cultures and 3 days for insect cells). The cultures were centrifuged at 2000 x g for 15 minutes and then the pellets resuspended in 500 μl Co-IP buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 0.5% IGEPAL-630, 1 mM EDTA, 0.5 mM TCEP,

5% glycerol, and a protease inhibitor cocktail. Cells were lysed by 3 rounds of freeze- thaw between liquid nitrogen and a 42°C water bath and 2 subsequent rounds of sonication. The lysate were centrifuged at 20,000 x g and the soluble fraction was pre- cleared with 5 μl of Protein G beads (with no antibody bound) by incubating the bead- lysate solution for 1 hour at 4°C on a rotator. Next the lysate was mixed with 15 μl of

Protein G beads that had been incubated for 2-3 hours with 5-10 μg of antibody (anti-His, anti-GST, or anti-Strep) at 4°C. The lysate-protein G bead solution is then incubated overnight at 4°C on a rotator. The magnetic protein G beads were applied to a magnetic strip and the unbound fraction removed and the beads washed 3 times with 500 μl Co-IP buffer. The beads were then resuspended in 30 μl of SDS sample buffer and boiled at

95°C for 10 minutes. All samples were then loaded onto SDS-PAGE gels and further analyzed by western blot.

2.3.4 Circular Dichroism

Circular dichroism was used to determine if the E. coli proteins were well folded.

Proteins were diluted to 0.2-0.3 mg/mL into CD buffer (20 mM sodium phosphate buffer pH 8.0 and 100 mM NaF) and the far-UV spectrum for each protein was analyzed at

52 room temperature using a JASCO model 810 spectropolarimeter. Three scans were taken for each sample and the values were averaged. These scans were further processed by subtraction of background signal from blank samples and finally smoothed to minimize noise. The measurements are displayed as units of mean residual ellipticity (Θ).

Spectrums from 190-240 nm were analyzed using the K2D2 program

(http://www.ogic.ca/projects/k2d2/) for percentages of each secondary structure element.

2.3.5 Thermofluor Assay

Optimal buffer composition and NaCl concentration for the stability of UL37 proteins, was determined using the Thermofluor method (125). Protein was diluted to 0.15 mg/mL into the storage buffer and a fluorescent dye Sypro Orange (Invitrogen) was added at a

1:1000 dilution. 10 μl of the protein-dye solution was pipetted into each well of a 96-well

PCR microplate. Next, 10 μl of desired buffer (custom-made screen containing buffers with pH 4.5-10.5 and NaCl concentrations from 0-500 mM) were added to wells containing the protein-dye solution. The plate was sealed and centrifuged for 1 min at

500 x g at 25 °C. Samples were analyzed on a Roche LightCycler 480 qPCR machine using an excitation wavelength of 465 nm and detection of emission at 610 nm. The emission signal was analyzed from 25 °C to 95 °C at a continuous acquisition rate of 3 measurements per °C. Data were analyzed using the ThermoQ software program

(http://jshare.johnshopkins.edu/aherna19/thermoq/). Conditions that stabilized UL37N further increased its solubility.

2.3.6 Size-exclusion chromatography (SEC)-coupled multi-angle light scattering

(MALS)

For SEC-coupled MALS, the UL37N protein (2.0 mg/mL injected concentration) was subjected to SEC using Superdex 200 column (GE Healthcare) equilibrated overnight in

20 mM Tris pH 8.0, 200 mM NaCl, and 0.5 mM TCEP. The chromatography system connected in line to a DAWN Heleos II light-scattering instrument equipped with a 658 nm laser, and Optilab T-rEX interferometric refractometer (Wyatt Technology). Data were analyzed with ASTRA 6 software (Wyatt Technology), yielding the molar mass and mass distribution (polydispersity) of the sample. For normalization of the light scattering detectors and data quality control, monomeric BSA was used.

2.3.7 Mass Spectrometry

For mass spectrometry analysis, the UL37 protein was analyzed using sinapinic acid

(Agilent Technologies, US) as matrix. MS measurements were performed on a Voyager

DE-Pro MALDI-TOF Mass Spectrometer (Applied Biosystems).

2.4 Crystallization

Crystallization screening for all proteins were originally conducted using the vapor diffusion method in a sitting drop set-up in trays stored at room temperature. The crystallization trials were set up using the Phoenix robot (Art Robbins Instruments) with drop containing 0.2 μl protein and 0.2 μl well solution. Crystallization drops were viewed and analyzed for crystal formation/growth using a Zeiss Discovery V8 microscope at least every other day for one month, and then at least once a month for one year.

2.4.1 UL37N Crystallization

Crystals for UL37N were originally observed in a screening solution containing 20%

PEG 1000, 200 mM Ca(CH3COO)2, and 0.1 M imidazole, pH 8.0 within 3 days after set- up. Optimization of crystals were conducted using the vapor diffusion method at room temperature in hanging drops using 1 L protein and 1 L well solution containing 24-

26% PEG 1000, 0.3 M Ca(CH3COO)2, and 0.1 M imidazole, pH 8.0. Large plates formed in 3-8 days and were harvested 2-4 weeks later. For data collection, crystals were incubated in the solution identical to the well solution plus 10% glycerol for 30 seconds to 2 minutes prior to flash freezing in liquid N2. Heavy-atom derivative crystals were obtained by soaking native crystals in well solution containing 5 mM thimerosal (Na salt of ethylmercurithiosalicylic acid or C9H9HgNaO2S) for 12-16 hours. Derivative crystals were harvested and frozen using the protocol developed for the native crystals.

2.4.2 gB crystallization

Crystals of the gB730-L6-cytoplasmic domain proteins were observed in many conditions between 10 days and 1 month after set-up. All conditions in which crystals appears contained PEG (3000-8000), salt (NaCl, Calcium Acetate, etc.), and buffers with pHs between 5.5-8.5. Optimization of gB crystals were set-up as was done for UL37N crystals. However, many screens that initially provided crystals, failed to repeat upon optimization. Regardless several crystals were able to be harvested for data collection, the freezing protocols used were similar to those used for UL37N crystals, except the cryoprotectants used were either 20-27.5% glycerol or 20% xylitol.

2.5 Obtaining Heavy atom derivatives of UL37N

2.5.1 Selenomethionine incorporation into proteins

In order to express of UL37N proteins that possessed the heavy atom selenium we used a protocol modified from Walden et al. (170). In this protocol we incorporated

Selenomethionine into the UL37N protein after inhibition of methionine biosynthesis of

E. coli. Briefly, the T7 express strain of E. coli containing the pJP23 plasmid was incubated on a shaker at 37°C in 5 ml LB containing 50 μg/ml of kanamycin. The following day the cells were spun down at 2000 x g, the supernatant was removed and the pelleted cells were washed once with M9 minimal medium and spun again. The cells were resuspended in M9 medium and diluted into 1 L of M9 minimal media containing

0.5% glucose and 0.01% Biotin and Thiamine and 50 μg/ml of kanamycin. The culture was incubated on a shaker at 37°C until an optical density between 0.6-0.8 was reached.

At this OD 100 mg of phenylalanine, lysine and threonine were added to the culture in addition to 50 mg of leucine, isoleucine, valine, and Selenomethionine. The culture was allowed to incubate for another 15 minutes on a shaker after with the temperature was reduced to 16°C and the cells were induced with 0.5 mM IPTG and the cells were incubated overnight. Harvesting and purification of UL37N containing Selenomethionine was identical to native UL37N protein, except that all buffers contained additional reducing agent (2.5 mM TCEP). Selenium incorporation into UL37N was validated by

Mass Spectrometry and the protein was crystallized under identical conditions to native

UL37N.

2.5.2 Heavy Metal Incorporation into proteins

In order to screen for the ability of additional heavy atoms to bind UL37N, several 10 mM stock solutions for heavy atom compounds were prepared. Included in this preparation were several lead, potassium, nickel, and gold containing compounds.

Additionally, many mercury and platinum compounds contained in heavy atoms screens

(Hampton Research) were prepared as well. To test the binding capability of these compounds to UL37N, solutions containing 10 μl of heavy atom compound were mixed with 10 μl of purified UL37N protein for a final heavy atom concentration of 5 mM and the solutions were incubated at room temperature for 15 minutes. After the incubation, sample buffer was added to the solution and the samples were run on a native gel using a

MOPS running buffer. Gels were stained by Coomassie and the gels analyzed for gel shifts in the apparent molecular weight of UL37N.

Heavy atoms that produced gel shifts were incubated with crystals of UL37N by placing crystals in mother liquor containing 5 mM of the heavy atom compound. The crystals were incubated in the heavy atom compound for 12-16 hour and then harvested as was done for native crystals of UL37N.

2.6 Structure Determination

X-ray diffraction data were collected at 100 K at the X25 beamline at the National

Synchrotron Radiation Source. The data were processed using HKL2000 (119) and indexed in space group P21. Native data set was processed up to 2.0-Å resolution and a

SAD Hg data set was processed to 2.3 Å. All 12 heavy atom sites were found using

57 phenix.autosol, and the experimental density allowed the tracing of ~70% of the residues in phenix.autobuild. Additional residues were manually built using Coot (44). There are two UL37N molecules in the asymmetric unit.

Before refinement of the heavy atom model, 10% of the data was set aside for cross- validation. The model was refined against the SAD Hg data set to 2.3-Å resolution using phenix.refine. Next, test set flags were transferred to the native data set; additionally,

10% of the native data between 2.3 and 2.0 Å were set aside for cross-validation. After several cycles of refinement in phenix.refine (3) and rebuilding in Coot (44), the Rwork is

17.3% and Rfree is 22.0%. The final model contains all amino acids from 1-479 including

3 of the 4 linker residues left after protease cleavage of the N-terminal tag. The final model is missing residues 480-496 in both chains. Molprobity (39) was used to assess the stereochemical quality of all models. According to Molprobity, 99.0% of the residues lie in the most favored and 1%, in the additionally allowed regions of the Ramachandran plot. Final statistics are listed in Table 4-1.

2.7 Structure analysis

Sequence alignment was generated and analyzed using ClustalW (92) and ESPRIPT (52).

Interfaces were analyzed using PISA (87). Structural homology searches were performed using the DALI server (63) and the top hits were superposed onto the UL37N protein using Dalilite pairwise comparison tool. Evolutionary Trace Server was used for evolutionary trace analysis

(http://www.cryst.bioc.cam.ac.uk/~jiye/evoltrace/evoltrace.html). All structure figures were made in Pymol (http://www.pymol.org).

2.8 Accession numbers

Atomic coordinates and structure factors for the UL37N structure have been deposited to the RCSB Protein Data Bank under accession number 4K70.

2.9 Identification and validation of putative host binding proteins of UL37N

2.9.1 Pull-down with UL37N sepharose

Purified UL37N protein was bound to CNBr-Activated Sepharose 4B in order to pull- down putative host protein interacting partners of PRV UL37N. To attach UL37N to the sepharose, 1 g of sepharose powder (makes 3.5 ml of sepharose) was washed for 15 minutes with 1 mM HCl at room temperature. The HCl solution was removed and the sepharose was washed 2 times in coupling buffer (100 mM NaHCO3 pH 8.3 and 500 mM

NaCl) and subsequently resuspended in 5 ml of coupling buffer containing 15 mg of purified UL37N protein. The protein-sepharose mixture was allowed to incubate overnight at 4°C with gentle shaking. The following day the mixture was allowed to settle and to residual protein was removed by washing 2 times with 10 mL of the coupling buffer. Remaining active groups of the CNBr sepharose were blocked by adding

1 M ethanolamine pH 8.0 incubating the slurry for 2 hours at room temperature. The

UL37N bound sepharose was washed with 10 column volumes of Tris-HCl pH 8.0 prior

59 to use in pull-down. This exact procedure was followed in the absence of any UL37N protein, to provide a control for sepharose without any protein bound.

To conduct the pull-down experiment Vero and PK15 cells were grown at 37°C and 5%

CO2 in DMEM supplemented with 5% FBS and 1% Penicillin/Streptomycin solution. 10

T175 flasks for each cell type were harvested using 3 ml Typsin per flask and the respuspended cells were pooled and spun down at 600xg for 6 minutes. The cells were washed 2 times with 1 X PBS and then resuspended in 10 ml lysis buffer (20 mM Tris

8.0, 100 mM NaCl, 0.5 mM TCEP, 5% Glycerol, and 0.5% IGEPAL-630) and incubated on ice for 30 minutes. To ensure efficient lysis, the cells were briefly sonicated for 2 cycles and then centrifuged at 20,000xg for 30 minutes. The soluble lysates were then divided into 2 fractions, one to be run over sepharose alone, and the other to be run over

UL37N-sepharose. The lysates were incubated with the sepharose overnight at 4°C on a rotator. The unbound protein was allowed to flow through the resin and the column was washed 3 times with 20 column volumes of wash buffer (20 mM Tris 8.0, 100 mM NaCl,

0.5 mM TCEP). The bound protein was analyzed by adding 50 μl of SDS sample buffer to the sepharose and boiling the sample at 95C for 10 minutes. The samples run on SDS-

PAGE and the bound proteins visualized by Coomassie stain.

For identification of bound proteins, gel sections were sliced from the gel and sent to the

Taplin Mass Spectrometry Facility at Harvard Medical School. The proteins were identified though tryptic digest of samples followed by identification of peptides using a

Velos LTQ Linear ion trap, with Orbitraps from Thermo Electron for high throughput

60 identification of peptides. The proteins were identified by identification of peptides from proteomic databases for pigs (PK15 cells) and humans (Vero cells and PK15 cells).

Confidently identified proteins from the pull-down studies possess at least 3 unique peptides matches.

2.9.2 GST Pull-downs

A protocol for pull-downs using purified GST-tagged Rab proteins has been modified from Langemeyer et. al. (91). This modified protocol was used for all wild-type Rab proteins. Briefly, 0.05 mg of purified GST-Rab proteins was incubated with glutathione-

Sepharose 4B (GE) in 1 ml of NE100 buffer (20 mM HEPES, pH 7.5, 100 mM NaCl, 10 mM EDTA, and 0.1% Triton X-100) for 1 hour at 4°C. The beads were centrifuged for 1 minute at 2000 x g at room temperature and the supernatant removed from the pellet. The beads were washed 3 times in 500 μl NE100 buffer, with centrifugation steps as above between each wash. After the last wash the beads were resuspended in 200 NL100 buffer

(NE100 buffer, replacing 10 mM EDTA with 10 mM MgCl2) and 20 μl of GDP or

GMppNp, and 0.1 mg of purified UL37N or LidA (positive control for interaction) and incubated on a rotator at 4°C for 1 hour. Beads were again washed 3 times in 500 μl

NL100 buffer with 1 minute spins at 2000 x g, removing supernatant after each wash.

After the last wash supernatant was removed the beads were resuspended in 30 μl SDS sample buffer and boiled at 95°C for 10 minutes. Samples were analyzed by Coomassie staining after running on a 4-15% SDS-PAGE gel.

The pull-down protocol using constitutively active and dominant negative GST-Rab proteins and GST-Sec22b has several modifications to the above protocol. Briefly, the

GST-tagged Rab or Sec22b proteins were incubated initially with the glutathione sepharose beads in the NL100 buffer (instead of NE100 buffer). Additionally, we did not add GDP or GMppNp to the solution when the UL37N or LidA proteins were added, as the mutant Rabs are already locked into the nucleotide bound forms desired. The rest of the protocol for Sec22b and mutant Rab proteins is identical to that for the wild-type Rab proteins.

2.9.3 GFP-Trap Co-Immunoprecipitation

5 x 105 Vero or 293T cells were seeded into 6 well plates and incubated overnight at

37°C and 5% CO2 to allow for the cells to adhere. Cells were next transfected with 2 mg of plasmid (eGFP, pJP60, or pGS1499) and 12 μg PEI. After incubation at 37°C and 5%

CO2 the transfected cells were washed 2 times in PBS and harvested by resuspension in

200 μl lysis buffer (20 mM HEPES 7.5, 100 mM NaCl, 0.5% IGEPAL CA-630 (Sigma), and 0.5 mM TCEP). 5 μl of HALT protease inhibitor cocktail (Thermo) was added to the resuspended cells and the cells were placed on ice for 30 minutes for lysis, with extensive pipetting every 10 minutes. Cells were then briefly sonicated for 2 cycles to ensure efficient lysis and spun down at 20,000 x g for 10 minutes at 4°C. The supernatant was harvested and diluted to 500 μl with wash buffer (lysis buffer without IGEPAL-630).

During centrifugation, 25 μl of GFP-Trap_M (per reaction) beads (ChromoTek) were washed with wash buffer three times, using the magnetic strip to pellet the beads and the supernatant was removed. After equilibration, the lysate was added to the beads and the mixture was incubated on a rotator at 4°C for 2 hours. The lysate-bead solution was then applied to the magnetic strip and the unbound fraction removed. The GFP-Trap beads

62 were subjected to 3 washes with 500 μl wash buffer and finally resuspended in 40 μl SDS sample buffer and boiled at 95°C for 10 minutes. Samples were run on SDS-PAGE on 4-

15% gels and western analysis was performed.

2.9.4 Subcellular localization of GFP fusions and immunofluorescence

1 x 105 Vero cells were seeded onto coverslips in a 24 well plate and incubated overnight at 37°C and 5% CO2 to allow for cell adherence. Cells were then transfected with 0.5 μg of plasmid (peGFP, pJP60, or pGS1499) and 3 μl of PEI. After a second 24 hour incubation at 37°C and 5% CO2 , the transfected cells were fixed with 3.7% formaldehyde in 1 X PBS for 20 minutes at room temperature. The fixative was removed and the cells on the coverslips were washed with 1 x PBS. The fixed cells were next permeablized with 0.1% Triton-X100 in 1 x PBS for 10 minutes. The detergent was removed and the cells were washed 2 times with 1 x PBS and stained with 1 μg/ml 1,6- diphenyl-1,3,5-hexatriene (DAPI) for 5 minutes. Coverslips were washed 2 final times with 1 x PBS and mounted onto glass slides using 8 μl of ProLong Gold Antifade reagent

(Invitrogen) and then sealed with nail polish. The mounted coverslips were allowed to cure overnight in the dark at room temperature prior to imaging. GFP localization was visualized using a 100X 1.4 NA lens on a Zeiss IM200 fluorescent microscope.

For co-localization studies of GFP fusions with Golgi or Early Endosomal structures immunofluorescence studies were conducted. These studies were conducted exactly as described above up to the permeablization step. After the incubation with Triton X100, the cells were blocked for 30 minutes using 4% goat serum in 1X PBS (blocking buffer).

The cells were next incubated overnight at 4°C with primary antibodies for markers of the Golgi, α-GM130 at 1:500, or early endosome ,α-EEA1 at 1:500 (gifts from Ralph

Isberg), in blocking buffer. The following day coverslips were washed 3 times in blocking buffer and incubated with 1:500 goat anti-mouse Alexa Fluor 594 (Invitrogen) in blocking buffer for 1 hour at 37°C. Coverslips were washed 3 times in blocking buffer and nuclei were stained with DAPI and cells mounted and visualized as described previously.

Chapter 3: Biochemical Characterization of the UL37 protein of

Pseudorabiesvirus

3.1 Introduction

UL37 is a large structural tegument protein that is conserved throughout the

Herpesviridae family. Several studies have previously shown the requirement for UL37 in efficient egress of herpesviruses. In UL37-null viruses, un-enveloped capsids aggregate in the cytosol of infected cells (41, 42, 82, 96, 133, 140) and the release of infectious virions is either completely blocked, as observed in HSV-1, (41, 96) or is strongly impaired, as seen in PRV (82, 96). Consistent with these findings, the UL37 protein of HSV-1 has shown to play in important role in capsid trafficking in the cytoplasm of the infected cells both during entry and egress (107, 123). Another protein, the UL36 protein of HSV-1 and PRV, has also been shown to play an important role in capsid trafficking (34, 94, 104) as it interacts directly with components of dynein (179).

The UL36 protein is known to interact with the C-terminal half UL37 protein in several herpesviruses (17); thus, it may be possible that UL36 and UL37 are working in coordination (with each other and potentially additional proteins i.e UL21) to traffic capsids throughout the cytoplasm. The interaction of the C-terminal half of UL37 with

UL36 is essential for correct addition and localization of UL37 with capsid structures. An interaction with the host-protein dystonin was also mapped to C-terminal half of UL37, and this interaction was shown to be essential for directed capsid movement in the cytosol (123). Furthermore, in PRV, UL37 has been implicated for the correct addition of tegument proteins, which may play a role during secondary envelopment (82). All of these previous studies have highlighted the importance of the UL37 protein in capsid trafficking and potentially secondary envelopment; however, the exact role(s) of the

UL37 protein in these processes remains largely unknown. Therefore, additional studies

66 using alternative approaches are required to acquire a better understanding of the functions of the UL37 protein during viral replication.

Previous work on UL37 has primarily focused on genetic studies of the protein in the context of viral infection. Thus, there is currently little information about its biochemical properties and molecular characteristics. In this study, we sought to investigate the properties of purified UL37 protein of pseudorabiesvirus. It was hoped that these studies would provide fundamental information needed for subsequent structural studies of UL37.

Additionally, we sought to purify and biochemically characterize the UL37 protein in complex with its interaction domain from the large tegument protein UL36. Ultimately, we hope to be able to better define the role(s) of UL37 in herpesvirus replication.

3.2 Results

3.2.1 UL37 is truncated upon recombinant expression

Initially, we expressed the full-length PRV UL37 protein with an N-terminal His6-SUMO tag in E. coli (Fig. 3-1A). During expression, this protein underwent spontaneous proteolysis, which generated a fragment containing the His6-SUMO tag and the N terminus of UL37 that was purified along with full-length molecules using metal affinity chromatography (Fig. 3-1B). This N-terminal cleavage product was highly soluble in comparison to the full-length UL37 protein, which was predominantly observed in the insoluble fraction (Fig 3-1B). Due to the decreased solubility of the full-length protein, the major species present in purified fractions was the UL37N terminal cleavage product.

However, the full-length protein could be separated from the N-terminal truncation product though size exclusion chromatography on an S200 column (Fig. 3-1C). During this process, additional full-length His6-SUMO-UL37 protein was lost to the void fraction, as the protein appears to aggregate heavily. Despite the high amounts of aggregated protein, mono-disperse full-length UL37 protein was present in peak 2 (Fig 3-1C) while the truncated UL37 N-terminal fragment eluted slightly later (peak 3). Based upon the elution volume in the S200 column, it appeared that the His6-SUMO-UL37 protein was a monomer in solution, with an apparent molecular weight of approximately 120 kDa. The final yield for the His6-SUMO-UL37 protein from a 1 liter E. coli culture was quite low, approximately 200 μg/L. However, even with this low level of protein we were able to start biochemical characterization of the UL37 protein from PRV.

3.2.2 Biochemical characterization of full-length UL37 from PRV

Due to the low levels of pure protein that we are able to acquire, we conducted the biochemical characterization of the UL37 protein with the His6-SUMO tag still attached.

We first wanted to determine whether the His6-SUMO-UL37 protein was folded when expressed in E. coli. Based upon circular dichroism (CD), we were able to determine that

UL37 is predominantly helical in nature, consistent with secondary structure predictions.

Although the observed helicity of 38% was well under the predicted helicity of 59%, the predominantly helical nature of the CD trace provides reasonable assurance that the

UL37 protein was folded.

Figure 3-1: The PRV UL37 protein is truncated during E. coli expression. (A) A schematic representation of the His6-SUMO tagged full-length UL37 protein. (B)

Coomassie stained gel of UL37 protein purified over Ni-NTA resin indicated the presence of full-length protein along with a ~70 kDa band. Anti-His western analysis of insoluble (I) and soluble (S) fractions of cell lysates expressing UL37 revealed the 70 kDa protein is an N-terminally truncated form of UL37. (C) Gel filtration of Ni-NTA purified UL37 protein was able to separate the monodisperse full-length UL37 protein

(peak 2) from aggregated protein (peak 1) and the N-terminal truncated protein (peak 3).

In addition to determining the secondary structure of the His6-SUMO-UL37 protein, we wanted to visualize to protein though negative-stain electron microscopy. Using this technique we have for the first time visualized the UL37 protein of PRV. The protein is a globular and elongated protein with approximate dimensions of 18 nm in length x 6 nm in width (Fig 3-2B). The His6-SUMO tag is still attached to this protein, so the UL37 protein alone likely has somewhat smaller dimensions – however subsequent attempts to visualize the UL37 protein without the His6-SUMO tag were unsuccessful. Taken together, the results from the CD analysis and electron microscopy indicate that the full- length UL37 protein expressed and purified from E. coli appears to be folded and additional studies towards structure and function can be pursued with this protein if we can determine methods to increase the yield of purified protein.

3.2.3 Increasing yield of soluble full-length UL37

We decided to take several approaches in our attempts to obtain larger yields of pure full- length UL37 protein. In the first approach, we sought to determine whether UL37 protein codon-optimized for E. coli expression would limit the proteolysis or potentially produce additional soluble full-length UL37 protein. The yield of protein was slightly increased

(up to ~300 μg/L) when using codon optimized UL37; however, the amount of N- terminal product was also increased (Fig 3-3A) and became increasingly difficult to separate from full-length protein. Additionally, the use of various additives during lysis failed to improve the insolubility issues observed with full-length UL37 protein (Fig. 3-

3D). Our next approach was to determine whether disruption of the cleavage site would

Figure 3-2: Biochemical characterization of full-length PRV UL37. (A) Circular dichroism of His6-SUMO-UL37 reveals a characteristic alpha helical trace for the full- length protein with an observed helical content of 38.2%. Helical content was estimated by the K2D2 program http://www.ogic.ca/projects/k2d2/ using the spectrum in the range of 190-240 nm. (B) Negative stain electron microscopy of His6-SUMO-UL37 reveals a globular and elongated molecule approximately 18 nm in length x 6 nm wide.

Measurements of the protein were made using the tape measure tool in Adobe photoshop.

71 prevent proteolysis of the full-length UL37 protein. Through our initial mass spectrometry analysis, we determined the cut site to be near residue 476 based upon peptide coverage or residue 496 of PRV UL37 based upon the mass-centroid of the truncated N-terminal protein. We further were able to determine the exact location of proteolysis through N-terminal sequencing of a 55 kDa truncation product present when we expressed His-SUMO UL37 C-terminal constructs of UL37 (See Section 3.2.4) to valine 498. We changed this residue to either glycine (V498G) or aspartate (V498D) and purified the proteins using metal affinity chromatography. In both cases, we still observed the truncated UL37 N-terminal protein, but in reduced levels. However, both substitutions appeared to destabilize the protein, as additional truncations and impurities appeared that were not observed with WT protein (Fig. 3-3A).

We next aimed to change the expression system used for PRV UL37, as previous western analysis of HSV-1 and PRV in mammalian cells infected with virus or transfected with

UL37 has not given any indication of proteolysis of UL37 (96). Instead of mammalian cell expression, where protein yields are quite small, we expressed the UL37 protein in

Sf9 insect cells. However, we again were able to observe the N-terminal truncation product of UL37. Additionally, yields of purified full-length His6-SUMO-UL37 from Sf9 cells were only about 100 μg/L, and thus, expression in insect cells did little to assist in increasing protein yields.

One final attempt to acquire full-length UL37 protein was to express the protein of the

UL37 homologue from HSV-1 in E. coli. This yielded only ~50 μg of protein/L, much

Figure 3-3: Full-length UL37 protein always leads to truncations and is highly insoluble.

(A) Coomassie stained gels of purified codon optimized UL37 proteins (wild type, and proteolytic site mutants V498D and V498G) all possess N-terminal cleavage products,

V498D and V498G also appear to further destabilize UL37. (B) anti-His western analysis of Sf9 cells infected with two baculoviruses containing the PRV UL37 gene. The N- terminal cleavage product is still observed in eukaryotic expressed protein. (C)

Coomassie stained gel of purified HSV-1 UL37 protein reveals a reduction in the N- terminal truncation, but also a large reduction in purified protein yield. (D) anti-His western analysis indicating that additives present during lysis of E. coli expressing UL37 fail to solubilize much additional full-length protein.

73 less than observed with PRV UL37. Additionally, although there appeared to be a reduction in the amount of truncated protein present for the HSV-1 UL37 homologue, the

N-terminal truncation was still observed. Thus, the proteolytic event appears to occur (at least at low levels) regardless of alteration of the proteolysis site, the homologue being expressed, and the expression system being used. Furthermore, in the best case scenario, the yield of pure full-length His6-SUMO-UL37 protein was ~300 μg/L, a very low yield that makes subsequent biochemical characterization and crystallization trials difficult.

3.2.4 UL37N and UL37C terminal constructs

The full-length PRV UL37 protein has proven to be a difficult protein to work with, given the issues of spontaneous proteolysis into an N-terminal cleavage product, in addition to poor protein solubility during lysis and a tendency to aggregate. In contrast, the N-terminal truncation product of PRV UL37 appears to be highly soluble and does not appear to have major problems with aggregation (Fig 3-1B and C). Therefore, we sought to characterize the N- and C- terminal regions of PRV UL37 separately.

3.2.4.1 Expression and purification

To investigate the PRV UL37 N- and C- terminal regions separately, four new expression plasmids were constructed. Each protein was expressed using boundaries identified from the original mass spectrometry analysis. For the N-terminal UL37 constructs, we expressed two proteins with N-terminal His6-SUMO tags that had termination codons after residue 476 or 496. C-terminal UL37 constructs also possessed an N-terminal His6-

SUMO tag and started at either reside 478 or 494 and expressed residues to the C- terminal residue 919 (Figure 3-4A).

All of the constructs were purified using simple metal affinity chromatography on a Ni-

NTA column. The His6-SUMO tags were then cleaved off through treatment with

PreScission protease, and the untagged proteins were further purified from the cleaved

His6-SUMO-tag and the GST-tagged PreScission protease by serial passage over Ni-NTA and glutathione columns. Lastly, the proteins were subjected to size exclusion chromatography as a final polishing step resulting in pure untagged UL37 N- and C- terminal proteins (Figure 3-4B). The amount of purified protein for N- and C- terminal constructs were dramatically higher than those obtained for full-length protein with yields of 16-20 mg/L for UL37N proteins and 6-8 mg/L for UL37C proteins.

The UL37N-terminal proteins appeared to be readily soluble and had very little tendency to aggregate. In contrast, the UL37C -terminal proteins both had a tendency to precipitate out of solution roughly 5-10 days after purification. Additionally, the C-terminal half was proteolytically cleaved in a fashion similar to full-length UL37 and prior to treatment with PreScission protease treatment. There was a doublet present in the UL37C terminal proteins that we believe is a direct result of this proteolysis. Thus, the C-terminal half of

UL37 appears to be the main cause of aggregation and solubility issues for the full-length

PRV UL37 protein. Additionally, the UL37C terminal constructs appear to possess the site of proteolytic cleavage, which through N-terminal sequencing of the truncated ~55 kDa fragment we determined to be at Valine 498. Regardless of the short-falls for UL37C terminal proteins, for both the UL37 N- and C- terminal proteins we had sufficient quantities for biochemical characterization and crystallization screening.

Figure 3-4: UL37N and UL37C terminal constructs can be readily purified. (A)

Schematic representations of the two UL37N terminal constructs and the two UL37C terminal constructs. Molecular weights of expressed proteins with tags are displayed. (B)

Coomassie gels of purified UL37N and UL37C terminal proteins after removal of the 14 kDa His6-SUMO tags. All of the proteins are purified with high yields, but the UL37C constructs run as a doublet on SDS PAGE.

3.2.4.2 Biochemical characterization of UL37N and UL37C

The Thermofluor assay (125) is a technique used to measure the thermal stability of proteins. It has been previously shown that conditions that enhance the thermal stability of proteins can often aide in the ability of the protein to crystallize (71). This assay uses a fluorescent dye, Sypro Orange, which displays an increase in fluorescence upon protein unfolding due to a quantum effect as the dye concentrates as it binds hydrophobic regions of the protein as they are being exposed. When the derivative of the increase in fluorescence is plotted, the apex of the derivative peak corresponds to the melting temperature of the protein.

We used the Thermofluor assay to assess the thermal stability of each of the UL37 N- and

C-terminal proteins across a wide range of buffer conditions (primarily varying pH, NaCl concentration). This assay indicated that in nearly all buffer conditions tested, the longer

UL37 N- and C- terminal constructs (UL37N (496) and UL37C (478)) were more stable than the shorter proteins for each terminal half. Given the increased thermal stability of these proteins, we opted to focus our efforts of further biochemical characterization and crystallization on these longer constructs (herein called UL37N for UL37N (496) and

UL37C for UL37C (478), unless otherwise specified). Furthermore, we were able to use this assay to determine optimal buffer conditions for further purification and storage of the UL37 N- and C-terminal proteins. The UL37N protein was found to have the highest thermal stability in PIPES (pH 7.0) or Na-Malonate (pH 7.0) buffers, while the UL37C protein was found to be the most stable in HEPES buffer (pH 7.5) (Table 3-1). All of the buffer conditions (1-16 in Table 3-1) contained 150 mM NaCl. Interestingly, the effect of

NaCl concentration on thermal stability was opposite between the N- and C-terminal proteins. The UL37N protein was the most thermally stable in very low concentrations of

NaCl, while the UL37C proteins displayed the highest thermal stability in more saline conditions (Table 3-1). Increasing levels of glycerol was also found to increase the thermal stability of both proteins. Thus, all subsequent purification and storage for the

UL37N protein was conducted in its optimal buffer of 50 mM PIPES pH 7.0, 50 mM

NaCl, and 0.5 mM TCEP. (Glycerol was left out as it can sometime impede crystallization.) The UL37C protein was subsequently purified and stored in its optimal buffer of 50 mM HEPES (pH 7.5), 500 mM NaCl, and 0.5 mM TCEP.

In addition to the reduced thermal stability of the short UL37N protein (UL37N (476)), we determined through CD analysis that the protein gradually loses helicity over time.

Over the course of one month, the observed helical content of the UL37N476 protein was reduced from 56%, to 40% at 2 weeks and finally to 28% after a month of storage at 4°C

(Fig. 3-5). In contrast, the longer UL37N construct was found to hold a constant helical content of 42% regardless of how long the protein had been stored (Fig. 3-5).

We also performed CD analysis on the UL37C terminal construct to ensure that the protein was properly folded. It was determined to actually be more helical when compared to the UL37N construct with helical contents of 67.2% and 56.3%, respectively (Fig. 3-6). It is important to note that in this experiment the helical content of

UL37N (496) was found to be higher than previously observed. This difference is likely due to errors in estimation of protein concentration between different batches of protein,

Table 3-1: Thermofluor Melting of UL37 N- and C- terminal proteins Buffer Screen Condition Protein Melting Temperature (°C) UL37N496 UL37N476 UL37C494 UL37C478 1 Na-Acetate 5.0 49.6 49.6 38.91 41.28 2 Na-Citrate 5.5 57.38 53.22 38.91 41.28 3 MES 6.0 55.88 52.24 38.91 41.28 4 Na Cacodylate 6.5 59.71 57.38 38.91 44.98 5 Bis-Tris 6.5 58.06 55.06 40.26 42.64 6 Imidazole 7.0 57.57 54.58 41.9 42.96 7 PIPES 7.0 59.54 56.9 38.91 44.98 8 Na-Malonate 7.0 59.54 56.9 38.91 45.32 9 MOPS 7.5 57.91 56.37 41.9 45.64 10 HEPES 7.5 58.21 56.55 42.27 45.98 11 Tris 8.0 57.19 55.88 43.61 45.64 12 Bicine 9.0 57.57 53.89 41.61 45.32 13 Glycine 9.0 57.91 53.58 42.92 45.32 14 CHES 9.5 52.41 49.6 40.9 41.66 15 Ethanolamine 9.5 56.19 52.41 41.61 44.63 16 CAPS 10.0 54.39 51.93 38.91 41.66 17 0 NaCl, HEPES 7.5 59.37 57.57 38.91 43.96 18 50 NaCl, HEPES 7.5 59.05 57.57 41.9 44.63 19 150 NaCl, HEPES 7.5 58.38 56.9 38.91 45.98 20 300 NaCl, HEPES 7.5 57.74 56.19 42.6 46.32 21 500 NaCl, HEPES 7.5 57.19 55.88 44.59 46.99 150 NaCl 5% Glycerol, HEPES 22 7.5 59.54 57.74 42.92 47.32 150 NaCl 10% Glycerol, HEPES 23 7.5 60.55 58.89 44.2 48.64 500 NaCl 5% Glycerol, HEPES 24 7.5 58.38 56.9 44.59 48.31 *Melting temperatures in red indicate several of the most thermally stable conditions for each protein.

Figure 3-5: UL37N (1-496) is more stable over time than UL37N (1-476) as measured using circular dichroism. The UL37N (1-476) protein is loses 50% of its helicity over the course of one month, while the UL37N (1-496) protein retains its helicity. Helical content was estimated by the K2D2 program http://www.ogic.ca/projects/k2d2/ using the spectrum in the range of 190-240 nm.

80 as the protein concentration contributes to the mean residual ellipticity value for each measure. Nevertheless, in both the UL37N and UL37C proteins were determined to be predominantly helical in nature.

In addition to the thermal stability and the helical content of the UL37N and UL37C proteins, we sought to determine the oligomerization state of the proteins in solution. It has been previously shown that UL37 of HSV-1 self-associates, and this interaction is predominantly dependent upon the C-terminal half of UL37 (17). To determine whether the UL37N and UL37C proteins were monomeric or part of higher order oligomers, we performed Multi-Angle Light Scattering (MALS) on the two proteins. Consistent with the findings of Bucks et al. (17), we found that UL37N is monomeric with a determined experimental molecular weight of 52.2 kDa, and no self-association is observed for this region of the protein. Furthermore, UL37C, which was observed to elute earlier on an

S200 column was determined to have a molecular weight of 92.5 kDa (Fig. 3-7), indicating that the UL37C protein is a dimer in solution. To our knowledge, this is the first time that the precise oligomerization status for any region of the UL37 protein has been determined.

Figure 3-6: UL37N and UL37C are predominantly helical. Circular dichroism revealed that both UL37N (Blue) and UL37C (Red) display characteristic alpha helical traces. The observed helical content of both UL37N and UL37C are consistent with secondary structure predictions. Helical content was estimated by the K2D2 program http://www.ogic.ca/projects/k2d2/ using the spectrum in the range of 190-240 nm.

Figure 3-7: Multi-Angle Light Scattering of UL37N and UL37C. UL37N is in blue and

UL37C in red. The thin solid lines are the UV traces of the proteins being run over an

S200 column, while the dotted lines are the measure of light scattering of the protein in the peaks. Molecular weights are determined by the flatter region of the light scattering present near the center of the peak. The experimentally determined molecular weight for

UL37N is 52.2 kDa, which is consistent with a monomer. In contrast, the determined molecular weight for UL37C is 92.5 kDa, which is the approximate weight of a dimer of

UL37C.

3.2.4.3 Additional UL37C terminal constructs

In contrast to the readily soluble UL37N protein, both of the UL37C terminal constructs exhibited several issues. First, the proteins were still truncated, presumably at the same site of proteolysis for the UL37 full-length protein. Purified UL37C ran as a doublet on

SDS-PAGE gels, likely as a result of this proteolytic cleavage. Furthermore, the UL37C terminal protein precipitated out of solution in 5-10 days post-purification, even when stored in optimal buffer conditions. To attempt to resolve some of these issues, we made a UL37C terminal construct, which started at residue 499 (Fig. 4-8A). Although the longer UL37C construct was found to be more thermodynamically stable, we hypothesized that the ability to remove the site of proteolysis from UL37C expressed proteins may be beneficial. However, upon purification using metal affinity chromatography of this protein the doublet was still observed (Fig. 4-8B), indicating that there may be addition cleavage events occurring on the C-terminal end of the protein.

In additional attempts to obtain UL37C terminal protein that possessed only one species of protein, we constructed two UL37 C-terminal expression plasmids that have premature stop codons at residues 831 or 856 (Fig 3-8A). These two sites were chosen as they are predicted to be in unstructured loops between predicted helices of the C-terminal half of

UL37. Both of these UL37 C-terminal proteins were purified by metal affinity chromatography and run on SDS PAGE. Neither protein appeared as a doublet, indicating that there is likely a cleavage event occurring on the C-terminal end of UL37 – downstream of residue 856. Unfortunately, for both of the proteins, the yields were

Figure 3-8: Purification of additional UL37C constructs. (A) Schematic representations of the smaller UL37C terminal constructs starting at residue 499 and terminating at residues 919, 856, and 831 respectively. (B) Coomassie stained gels of Ni-NTA purified protein. A doublet is still observed in UL37C (499-919) indicating there is likely a cleavage event at the C-terminal end of the protein. Both UL37C (499-856) and (499-

831) run as a single band indicating the C-terminal proteolysis site is after residue 856, however both proteins display drastically reduced yields of protein.

85 dramatically decreased, from 6-8 mg/L for other UL37 C-terminal constructs to ~1 mg/L for the UL37C (499-856) protein and ~200 mg/L for the shorter UL37C(499-831) protein.

Additionally, these protein preps seemed much dirtier than other preps of UL37 C- terminal constructs, indicating that the proteins are likely disordered and possibly prone to aggregation. Therefore, the best UL37C terminal protein construct is still the longest

UL37C protein (478-919).

3.2.5 UL37/UL36 Interaction studies

In many cases, proteins that exhibit issues in solubility or tendencies to aggregate can be stabilized though the use of ligands or binding partners. In the case of the UL37 protein, previous studies have shown the importance of the UL36-UL37 interaction in localizing

UL37 to capsid structures (49). Additionally, in co-IP studies (17), co-expression of

HSV-1 UL37 with the UL37 binding region of HSV-1 UL36 resulted in a reduction of self-association between UL37 proteins. The interaction site in this study was further mapped to the C-terminal half of HSV-1 UL37 (17). Given these studies and the solubility issues observed with the purified PRV UL37 full-length and UL37C, we hypothesized that co-expression of UL37 or UL37C with the UL37 binding region of

UL36 may prevent aggregation and proteolysis of UL37. Therefore, we constructed two expression plasmids of the UL37-binding region of PRV UL36 (residues 344-561) with either an N-terminal Strep-SUMO tag or GST tag. All results shown herein are for the

Strep-SUMO tag protein, but consistent results were obtained with GST-UL36 (344-561).

We first wanted to validate that we could observe the interaction between UL37 and

UL36 (344-561) using proteins expressed in E. coli. Through co-IP experiments, we were able to determine that both the full-length His6-SUMO-UL37 protein and the His6-

SUMO-UL37C terminal protein were able to immunoprecipitate Strep-SUMO tagged

UL36 (344-561) (Fig. 3-9). Furthermore, UL36 was not observed in the IP fraction when not co-expressed with UL37. These results suggested that we could likely use these constructs to obtain complexes of UL37 and UL36.

Using metal affinity chromatography, we were able to co-purify His6-SUMO-UL37C and

Strep-SUMO-UL36 (344-561) when the two proteins were co-expressed in E. coli (Fig 3-

10A and B). This is significant because this is the first time the complex has been purified or observed in the absence of other viral or mammalian cell proteins.

Unfortunately, based upon size exclusion chromatography, the region of UL36 (344-561) that was expressed caused the UL37C-UL36 complex to aggregate, as all complex is observed in the void volume (Fig. 3-10C). Similar results were obtained for the purification of full-length UL37 with UL36 (344-561). Furthermore, purification of UL36

(344-561) using Strep resin and subsequent analysis of the protein by gel filtration revealed that UL36 alone, with the current boundaries, is comletely aggregated. Hence, further characterization of the complex will require the discovery of a UL36 fragment that retains the ability to interact with UL37, but is also mono-disperse in solution.

Figure 3-9: E. coli expressed PRV UL37 and PRV UL36 (344-561) interact. Anti-strep and anti-His western blots of co-Immunoprecipitation experiments using anti-His antibody. UL36 (344-561) co-immunoprecipitated with both the full-length UL37 protein and the UL37C terminal protein, consistent with previous results (17). UL36 did not immunoprecipitate in the absence of UL37.

Figure 3-10: Purified UL37-UL36 complex is aggregated. (A) Schematic representations of the His-SUMO-UL37C terminal protein and the Strep-SUMO-UL36 (344-561) protein used in purification. (B) Coomassie stained gel of UL37C and UL37 (344-561) after purification over Ni-NTA resin. The presence of Strep-SUMO-UL36 indicates that the complex can be purified from E. coli co-expressing both proteins. (C) Gel filtration chromatogram and Coomassie stained gel of UL37-UL36 (344-561) complex. The two peaks observed in the chromatogram are aggregated UL36-UL37 complex (in the void) and dimeric UL37C (in the smaller second peak), indicating UL36 induces aggregation of

UL37-UL36 complex.

3.3 Discussion

In these studies we (for the first time) were able to purify and perform preliminary biochemical studies on the full-length UL37 protein of PRV. When tagged with a His6-

SUMO tag, the protein was predominantly alpha-helical, consistent with secondary structure predictions. Furthermore, we were able to visualize this protein using negative- stain EM imaging, revealing an elongated protein with approximate dimensions of 18 nm x 6 nm.

Unfortunately, the UL37 protein underwent spontaneous proteolysis, and this could not be fully prevented regardless of mutagenesis, expression system, or the use of alternative homologues. In light of this cleavage event and the low yields associated with purification of full-length UL37 protein, we pursued the N- and C-terminal regions of

UL37. Both were found to be predominantly helical in nature, consistent with secondary structure predictions and CD analysis of the full-length UL37 protein. Interestingly, both the N- and C- terminal domain were found to be more helical in content than the full- length protein. However, this may be attributed to errors associated with determining helical content, as we observed variations between experiments on different protein preps of up to 16%.

The optimal buffers for both UL37N and UL37C were determined using the Thermofluor assay. Both the UL37N and UL37C proteins were found to be the most stable in buffers of near neutral pH (7.0 and 7.5 respectively). However, the effect of salt on the thermal stability was found to be opposite for the two halves of UL37. While UL37N became less

90 stable with increasing concentrations of NaCl, the UL37C protein actually increased its stability. This is somewhat surprising as the two halves of UL37 have very similar amino acid compositions. The reasoning for this differing response to NaCl concentration is not clear at this time, but it seems reasonable that the residue distribution of the surface of

UL37N and UL37C proteins may be quite different.

Through the use of Multi-Angle Light Scattering, we were able to assess the oligomerization status of both the UL37N and UL37C proteins. In agreement with previous work with HSV-1 UL37 (17), we observed that the UL37N protein did not self- associate in solution and possessed a molecular weight of 52 kDa consistent with a monomer. In contrast, the UL37C protein was determined to have a molecular weight of

93 kDa, consistent with UL37C being a dimer in solution. The putative self-association domain in HSV-1 is present in the C-terminal half of the protein (17) and here for the first time report that this “self-association” is a dimerization of the C-terminal half.

Interestingly, based upon the elution volume of the full-length UL37 protein on an S200 column, the apparent molecular weight of this protein is 120 kDa – or a monomer of the full-length protein. This indicates that although the UL37 protein contains a self- association (or dimerization) domain in its C-terminal portion – no dimer was observed for full-length protein. It is possible, that the N-terminal half of UL37 is somehow interfering with dimerization of the C-terminal half, although it is currently unclear if there is any interaction between the N- and C- terminal halves of the protein. The only

“self-association” observed for full length UL37 is aggregated protein in the void. It cannot be ruled out that this protein is not truly aggregated, but rather a larger ordered

91 oligomer (dimer of dimers, pentamer, etc.). This seems unlikely as these larger ordered oligomers were not observed with UL37C.

Through biochemical characterizations it is clear that the UL37N and UL37C terminal halves are quite distinct from one another – in oligomerization, thermal stability, and tolerance to NaCl. It will be interesting to continue to characterize these two domains and ultimately determine their structures. Furthermore, it will be interesting to determine whether these two halves function independently in the full-length UL37 protein or if they work together to mediate a function.

The UL37 protein is also known to interact with the largest tegument protein, UL36. We sought to co-express and ultimately purify the PRV UL37 protein (full-length or UL37C) with the binding region of PRV UL36 (344-561). We hypothesized that the presence of the UL36 binding partner may further stabilize the UL37 proteins, as the C-terminal protein is highly prone to aggregation shortly after purification. We were able to observe interaction by Co-IP of both full-length UL37 and UL37C with its binding region of the

UL36 protein (residues 344-561). Additionally, for the first time, we were able to purify complex between these two proteins. Unfortunately, the complex was completely aggregated due to the instability of the UL36 protein. Further characterization of the

UL37-UL36 interaction is interesting on many levels, as this interaction is critical for the localization of UL37 with capsid structure and downstream addition of tegument proteins

(49). However, we first need to better define the boundaries of the UL37-binding region

92 in UL36 to find a protein domain that will prevent aggregation of the UL36-UL37 complex.

Taken together, in these studies we for the first time were able to characterize the UL37 protein of PRV. Additional biochemical characterization on the N- and C- terminal halves revealed two regions that possessed very differing biochemical qualities. Also for the first time we were able to purify a complex between the UL37 protein and its binding region in UL36. Hopefully, future work will be able to further characterize the UL37 protein alone and in complex with UL36, revealing new information about the potential roles of UL37 in herpesvirus replication, and how it interacts with other tegument proteins.

Chapter 4: The crystal structure of PRV UL37N supports its critical

role in control of viral trafficking

The data presented in this chapter are included in the following manuscript recently accepted for publication.

Pitts, J.D., Klabis J., Richards, A.L., Smith, G.A. and Heldwein E.E. 2014. “Crystal structure of the herpesvirus inner tegument protein UL37 supports its essential role in control of viral trafficking.” Journal of Virology. Epub ahead of print.

Figure 4-11 was obtained by experiments conducted by Jenifer Klabis and Alexsia

Richards from the lab of Professor Greg Smith.

4.1 Introduction

The UL37 proteins from HSV-1 and PRV have been shown to play critical roles in viral replication of these viruses. UL37-null viruses are unable to efficiently be released from infected cells, highlighting an essential role for UL37 during viral egress. Previous studies have found that HSV-1 UL37 and dystonin are required for directed capsid transport in infected cells (123). This result sheds light on a possible role of UL37 during herpesvirus egress; however, the precise mechanism(s) remains unclear.

In the previous chapter, we have shown the ability to purify and biochemically characterize the full-length UL37 protein from pseudorabiesvirus as well as truncated N- and C-terminal regions of the protein. In addition to these studies, we hoped to be able to determine the structure of the UL37 protein from PRV. A crystal structure of UL37 would assist in identifying functional regions of the protein. Furthermore, identification of any proteins that share structural similarity to UL37 may provide additional insight to the critical role(s) of UL37 during viral infection.

In this chapter, we report the crystal structure of UL37N and the insight that this structure has provided in identifying functional regions of UL37. Additionally, the structure of

UL37N does show structural similarity to proteins of known structure and we analyze this similarity and what it may mean for potential functions of the UL37 protein of PRV.

4.2 Results

4.2.1 Crystallization trials of UL37 constructs

The full-length UL37 protein, the UL37 C-terminal protein (478-919), and the UL37N terminal protein (1-496) were all screened in crystallization trials. Both the UL37 full- length protein and UL37C failed to provide any crystals that could be further optimized.

The UL37N protein, however, yielded crystals in 2 conditions. In 3-4 days, thin plate-like crystals were observed in a condition containing 20% PEG 1000, 200 mM calcium acetate, and 100 mM imidazole (pH 8.0). Additionally, smaller crystals appeared in a similar condition, 20% PEG 1000, 200 mM MgCl2, and cacodylate (pH 6.5) in 2-3 weeks.

As the crystals obtained from the first condition appeared much faster and provided larger crystals, I predominantly focused on the optimization of this condition to provide larger and thicker crystals that could be used for x-ray diffraction.

To repeat crystal growth in the optimization process, I first needed to increase the PEG

100 concentration to 25%, as 20% PEG 100 did not provide crystals when increasing the drop size from 0.2 μl (protein) + 0.2 μl (well solution) during screening to 1 μl +1 μl for optimization. Early in the optimization process for UL37N crystals, it became apparent that the presence of calcium acetate was essential for nucleation and growth, as calcium acetate concentrations below 100 mM failed to provide any crystals. Furthermore, increasing the concentration of calcium acetate to 300 mM provided more single plate- like crystals as opposed to the stacks of plates of observed at 100 mM and 200 mM concentrations (Fig 4-1). Other calcium containing compounds tested (i.e. calcium chloride, calcium carbonate) failed to yield crystals. The plates acquired from these 300

96 mM calcium acetate set-ups were still too small for diffraction, thus we attempted to optimize the PEG concentration to produce larger crystals. By lowering the concentration of PEG 1000 to 23% and increasing the drop size to 2 μl +2 μl, we were able to increase the thickness and size of the crystals to be adequate for x-ray diffraction (Fig 4-1). These crystals were harvested and frozen in a cryoprotectant containing glycerol in mother liquor and sent to the synchrotron for analysis. Excitingly, the crystals obtained from these optimized conditions diffracted up to 2.0 angstrom resolution, providing a native data set that could be used for structural determination of UL37N.

4.2.2 Obtaining heavy metal derivatives for UL37N

With a good native data set, the major remaining hurdle before the structure of UL37N could be solved was determination of the phases. Phases are determined through the diffraction of crystals containing heavy atom derivatives. When crystals containing heavy atoms are diffracted at the absorption edge for that heavy atom, an anomalous signal should be observed. From this anomalous signal, the phases for the crystals can be estimated and the structure of the protein of interest solved. I took two approaches in my efforts to obtain crystals containing heavy atom derivatives of UL37N.

The first approach was to incorporate Selenomethionine (SeMet) into the protein during expression. To do this, I inhibited methionine biosynthesis through the addition of several amino acids and a source of methionine (Selenomethionine) to E. coli in M9 medium.

Figure 4-1: Optimization of UL37N crystallization. UL37N crystal nucleation and growth is dependent upon the presence of calcium acetate. Lower concentrations of calcium acetate results in heavy clustering of crystals, in the presence of 300 mM calcium acetate however small single plates are observed. Decreasing the concentration of PEG

1000 and increasing the drop size resulted in thicker and larger crystals, which could be harvested and tested for diffraction. Crystals grown in these conditions diffracted up to

2.0 angstrom resolution.

The UL37N protein containing SeMet was purified in the same manner as with native protein, with the exception of using 2.5 mM TCEP instead of 0.5 mM, ensuring that the

Selenomethionine remained reduced. Using this purification scheme, we were able to acquire sufficient amounts of SeMet UL37N (~ 2 mg/liter) to crystallize the protein under the same conditions as native UL37N. To ensure SeMet had been incorporated, we sent the protein to be analyzed by Mass Spectrometry and found that 80% had SeMet incorporated into all 7 methionine sites of UL37N while the remaining 20% of UL37N had SeMet incorporated in 6 of 7 sites. The crystals containing SeMet diffracted well when tested at the synchrotron (up to 2.5 angstroms); however, they contained very low anomalous signal, and it was not possible to acquire the needed phase information from these crystals.

The second approach to obtaining heavy atom derivatives for acquiring phases was to perform heavy atom soaks of native UL37N crystals. Before we started using crystals for heavy atoms soaks, we tested which heavy atoms would be the most likely to incorporate with the protein. To do this, we screened 48 different heavy atoms for their ability to bind

UL37N by incubating the protein in 5 mM solutions containing heavy atom for 15 minutes at room temperature. I then ran all proteins on a native gel and looked for shifts in mobility (Figure 4-2), which would be indicative of heavy atom binding the protein.

This high throughput assay revealed several heavy atoms (mostly containing mercury or platinum) that shifted the mobility of the UL37N protein, suggesting they would likely be good candidates for soaking crystals.

Figure 4-2: Native gels of heavy atom soaks of UL37N. UL37N protein was incubated with 5 mM solutions of various heavy atoms for 15 minutes and then analyzed by running on a native gel. Shifts in mobility on the gel indicate that the heavy atom is interacting with the protein. The starred heavy atoms were those used for soaking crystals of UL37N to obtain heavy atom derivative crystals. Hg5 (ethylmercurithiosalicylic acid) was the heavy atom used for structural determination of UL37N.

100

Three to four native UL37N crystals were soaked with each of the eight heavy atoms compounds (six mercuries and two platinums) found to change mobility on the native gel

(starred compounds in Figure 4-2). Crystals were soaked in mother liquor containing

5mM of the heavy atom for 12-16 hours and were then harvested, frozen, and tested for diffraction at the synchrotron. Several of the heavy atom derivative crystals diffracted to between 3.0 to 4.5 angstroms several with clear anomalous signal. However, one heavy atom derivative crystal containing ethylmercurithiosalicylic acid (Thimerosal) diffracted to 2.05 angstroms with high anomalous signal to 2.3 angstrom resolution. From this one crystal we were able to determine the phases and solve the structure of the UL37N protein.

4.2.3 Analysis of the crystal structure of UL37N

The crystal structure of UL37N was determined using single anomalous dispersion with the Thimerosal heavy atom derivative. The phases were then transferred and extended, and the structure refined against the 2.0-Å native data set of UL37N (Table 4-1).

4.2.3.1 UL37N is a crystallographic dimer but a monomer in solution

There are two monomers of UL37N in the asymmetric unit, and the final model of both monomers includes residues 1-479 plus residual linker residues PGS on the N-terminus, which were left after proteolysis with PreScission protease. Each monomer of the asymmetric unit adopts a very similar conformation (Figure 4-3), with the rmsd = 0.4 when overlaying the 482 common Cs (63).

101

Table 4-1: UL37N Data collection and refinement statistics Native Thimerosal Data collection Space group P21 P21 Cell dimensions a, b, c (Å) 51.67, 156.589, 67.381 51.526, 156.303, 66.343 , ,  () 90, 91.33, 90 90, 91.78, 90 Resolution (Å) 43.12-2.00 (2.07-2.00) 48.91-2.05 (2.12-2.05)

Rsym or Rmerge 0.086 (0.516) 0.097 (0.280) I/σI 20.32 (2.74) 13.87 (2.18) Completeness (%) 89.4 (49.5) 85.1 (35.31) Redundancy 6.3 (4.4) 3.9 (2.1)

Refinement Resolution (Å) 43.12-2.00 No. reflections (free set) 64,342 (2,347)

Rwork / Rfree 17.30/22.01 No. atoms 7983 Protein 7342 Ligand/ion 45 Water 596 B-factors 35.05 Protein 35.03 Ligand/ion 42.8 Water 37.5 R.m.s. deviations Bond lengths (Å) 0.007 Bond angles () 0.96

*Values in parentheses are for highest-resolution shell.

102

Figure 4-3: Overlay of the two UL37N monomers in the asymmetric unit. The two monomers (A in red and B in cyan) align very closely with an rmsd of 0.4 angstroms when overlaid using DaliLite (64).

103

The two UL37N monomers in the asymmetric unit form an X-shaped dimer (Figure 4-

4A) that buries 1734.8 Å2 of surface area as determined using the Pisa interface tool (88).

Four calcium ions are coordinated at the interface of dimerization as two symmetry- related sets each consisting of two calcium ions. Each set is coordinated by carboxyl oxygens from the side chains of Asp79, Asp81, and Glu82 of one monomer and carboxyl oxygens from the side chains of Asp382 and Asp383 plus the carbonyl oxygen of Trp379 of the other monomer and two water molecules (Figure 4-4B). As a result, one calcium ion is hexahedrally coordinated while the second is pentahedrally coordinated.

Despite forming a dimer in crystals, UL37N is a monomer in solution. Crystal formation required the presence of at least 100 mM Ca(CH3COO)2, and the best crystals were obtained in the presence of 300 mM Ca(CH3COO)2. However, in solution the UL37N protein remained monomeric even in the presence of 200 mM CaCl2, judging by its elution volume during size-exclusion chromatography (Figure 4-4C). We conclude that the dimerization of UL37N observed in crystals is likely induced by the high protein concentrations and presence of calcium in the crystallization conditions. The coordination of four calcium ions at the crystallographic dimer interface does aide in explaining the importance of calcium ions in mediating crystal contacts. In the absence of calcium, the buried interface would have been smaller, 1504.0 Å2 instead of 1734.8 Å2 (88).

Additionally, the interface between the two monomers was predicted to be not biologically significant, based upon an algorithm used by the PISA interface tool, which analyzes interfaces of both biologically significant and crystallographic contacts. Taken

104

Figure 4-4: UL37N is a calcium-dependent dimer in crystals but not in solution. (A) Two

UL37N monomers in the asymmetric unit are shown in grey and beige. Calcium ions at the dimer interface are shown as blue spheres. (B) A close-up view of the calcium- binding site at the dimer interface. The color scheme is the same as in (A). Residues coordinating calcium ions are shown as sticks while waters involved in coordinating calcium are shown as maroon spheres. Distances between calcium and coordinating atoms are shown as dashed lines and labeled. (C) An overlay of size-exclusion chromatograms of UL37N in the presence or absence of 0.2 M CaCl2.

105 together, these analyses indicate that a single monomer of the UL37N protein represents the biologically relevant state.

4.2.3.2 The architecture of UL37N

UL37N is an elongated molecule with dimensions of 99 x 42 x 26 Å composed of 24  helices and 6 310 helices arranged into a series of helical bundles (Fig. 1C). The structure can be divided into three domains: domain I, residues 1-184 and 432-479; domain II, residues 185-295; and domain III, residues 296-431 (Fig. 4-5A, C).

Domain I is formed by two non-contiguous segments of the polypeptide chain, residues

1-184 and 432-479 (Fig. 4-5C). Residue 479 is the last resolved residue; no electron density was observed for residues 480-496, as these residues are likely disordered.

Domain I consists of five helical hairpins with the “down-up” topology (Fig. 4-5C), which are formed by 12  helices (1-10 and 23-24) and three 310 helices 1-3.

Hairpins 1-4 make up the N-terminal region of domain I (I-N), while hairpin 5 makes up the non-contiguous C-terminal region of domain I (I-C) (Figure 4-5A).

Linker residues GS, which precede the start methionine, form the N-terminus of helix 1.

Hairpins 1 through 3 form a helical stack (Fig. 4-5C) as the three hairpins layer on top of one another. Hairpin 1 consists of two short antiparallel helices, while hairpin 2 consists of two longer kinked helices, and hairpin 3 has two short “up” helices followed by a loop and a single extended “down” helix. Helix 1 is contained within a long loop that connects hairpins 2 and 3. The last two helices, 23 and 24, form hairpin 5 (Fig. 4-5C).

Only the upper portion of the hairpin 5 helices interact with hairpins 1-3, an arrangement

106 that results in a large U-shaped groove within domain I. Hairpin 4, formed by helix 10 running antiparallel to helices 9 and 2, forms a “plug” in the U-shaped groove in domain I. The small helix 3 forms the tip of this “plug” (Fig 4-5C). At the opposite end of the “plug”, a solitary helix 8 resides at the tip of a long extension which interacts with domain II. The Dali structural homology search (63) revealed that domain I bears structural resemblance to helical bundle domains of several subunits of multi-subunit tethering complexes

Within domain one of the crystal structure, we were also able to better understand why the shorter UL37N (1-476) construct possessed a lower thermal stability and progressively lost secondary structure over time (Fig. 3-5). The reason for this is that the conserved residue W477 plays a key role not only in the stability of domain I but also the entire UL37N. As W477 helps anchor the hairpin 4 “plug” in domain I-N through van der

Waals interactions with several hydrophobic residues and a hydrogen bond with the carboxyl of D169 (Fig. 4-5B). Additionally, W477 mediates several van der Waals interactions with hydrophobic residues in domain III, thus aiding in the stability of the whole protein. Therefore, it is easy to envision the shorter construct, which lacks this conserved residue, would be less stable as it is lacking these inter-domain contacts in the core of the protein.

Domain II, residues 185-295, consists of helices 11-14 and two 310 helices 4-5 (Fig.

4-5C). Helices 11-13 form a helical bundle, in which the last turn of helix 12 adopts a -helix conformation. The putative conserved nuclear export motif, residues 263-273 of

107

108

Figure 4-5: UL37N structure and domain organization. (A) Crystal structure of a UL37N monomer is shown in two orientations related by a 180-degree rotation around the vertical axis. A linear diagram indicating domain boundaries of UL37N is also represented. (B) A close-up view of residue W477 and its surroundings. Domains are colored as in (A). (C) UL37N domains are shown individually. The color scheme is the same as in (A) and (B) except the layers of domain 1 are highlighted by different shades of green and the putative NES region in domain II is shown in teal. Orientations were chosen to show all secondary structure elements. Topology diagrams displaying individual domain organization of helices are also shown.

109

HSV-2 (172), maps to the buried helix 12 (Fig. 4-5C) and is unlikely to be functional.

Two long loops at the bottom of domain II are extended from the molecule, yet remain well structured (Fig. 4-5C) and adopt similar conformations in the two NCS molecules

(Fig 4-3). Helix 14 appears to buttress both loops. Domain II does not have any structural homologs according to DALI (63).

Domain III, residues 296-431, is composed of helices 15-22 and one 310 helix 6 (Fig.

1E). This domain is also a helical bundle, with the central helix 19 surrounded by the other six helices. This central helix maintains the structural integrity of domain III and is highly conserved among alphaherpesviruses. Domain III has a weak structural similarity to several 14-3-3 family proteins (63).

4.2.4 Identification of Functional Regions of UL37N

With the structure of UL37N solved, we wanted to attempt to identify functional regions of the protein. Functional regions of proteins are often highly conserved across homologous proteins, thus the mapping of conserved residues onto the structure of

UL37N may highlight regions of potential functional importance. Additionally, the mapping of electrostatics onto the surface of proteins can reveal highly charged electro- positive or electro-negative patches. These patches are also often sites of functional importance as they mediate protein-protein interactions. In this section we seek to determine functional regions of UL37 through use of mapping electrostatics and conserved residues onto the structure of UL37N.

110

4.2.4.1 Electrostatics of UL37N

Electrostatics maps of the UL37N protein were generated using the CHARMM server

(http://www.charmm-gui.org/?doc=input/pbeqsolver). This mapping revealed very few charged patches on the surface of UL37N as the charges were mostly distributed approximately equally throughout the protein. The one region that did appear to be more highly charged was the central groove, which contains residues important for calcium binding and additional residues involved in the crystallographic dimerization interface of

UL37N (Figure 4-6). Although UL37N is not observed to dimerize in solution, given the highly charged nature of this calcium binding region, this area may participate in intramolecular contacts with the C-terminal half of UL37 in the full-length protein.

However, beyond the calcium binding region, the electrostatics mapping did not provide much information about additional putative functional regions.

4.2.4.2 Mapping of conserved residues onto UL37N structure

To analyze sequence conservation within UL37N, we generated a sequence alignment of

15 UL37 homologs from alphaherpesviruses – the subfamily of herpesviruses that includes HSV and PRV. From the sequence alignment, we identified 35 strictly- conserved residues (Fig. 4-7) and mapped them onto the structure of UL37N. Most of these residues are located within the hydrophobic core of the protein and are likely predominantly required for maintaining the structural integrity of the UL37N. However,

11 of these residues are surface-exposed (Figure 4-8) and are the logical choices for mutational analysis, as surface-exposed conserved residues often participate in protein/protein interactions. Unfortunately, these surface conserved residues were spread

111

Figure 4-6: Electrostatic surface potential map of UL37N generated using CHARMM

(http://www.charmm-gui.org/?doc=input/pbeqsolver). Charged electro-negative patches involved in binding calcium in crystals are circled.

112

113

Figure 4-7: Multiple sequence alignment of UL37 homologs from 15 alphaherpesviruses.

Only the alignment of residues corresponding to residues 1-479 of PRV UL37 is shown.

Helices are indicated by rectangular boxes, labeled, and colored by domain, similar to Fig.

4-5. Identical residues are boxed in red, and those exposed on the surface of UL37N are marked by purple asterisks. Residues involved in binding calcium are marked by green circles. All residues specified for mutagenesis are highlighted by triangles (R1 - blue, R2

- pink, R3 - orange) using the same color scheme as in Fig. 4-10B.

114 throughout the surface of the protein; thus, mapping of these residues onto UL37N did not reveal clusters that would help pinpoint regions of potential functional importance

(Fig. 4-8).

4.2.4.2 ETA reveals three conserved surface patches within UL37N

To locate potentially important functional sites on the surface of UL37N, we took the approach of mapping of conservation by evolutionary trace analysis (ETA) (98). ETA uses the sequence alignment of homologous proteins to generate a phylogenetic tree, which is then broken up into partitions with more closely related sequences grouped into classes. Within each partition, consensus sequences are generated for each set of proteins within a class. Each position within the sequence alignment is designated as conserved, class-specific, or neutral. Conserved residues are the same in all consensus sequences, but class-specific residues are common only within closely related subgroups and differ among more divergent subgroups. Positions lacking consensus among at least one subgroup are considered to be neutral. Clustering of conserved and class-specific residues onto the protein surface may indicate regions of potential functional importance (98).

This method has previously been used to detect functional sites in a number of proteins

(25, 155).

Evolutionary trace analysis was performed on the 15 alphaherpesviruses previously used in generation of the sequence alignment (Fig 4-7) and conservation mapping (Fig 4-8).

ETA of these sequences generated the phylogenetic tree (Fig 4-9), and though mapping of each partition of the analysis we determined that partition 5 provided a sufficient

115

Figure 4-8: Mapping of conserved residues on the surface UL37N. (A) UL37N structure is shown in surface representation. Two orientations related by a 180-degree rotation around the vertical axis are shown. Strictly conserved surface residues are in purple.

116 number of class-specific and conserved residues while minimizing the background associated with later partitions. The conserved and class specific residues were mapped onto UL37N structure (Figure 4-10A), revealing several surface-exposed clusters that may be important for function of UL37. To identify the most likely candidates for functional regions, we sought surface-exposed clusters that possess both conserved and class specific residues. ETA analysis revealed three surface exposed clusters, referred to as regions 1 through 3, which contained both conserved and class-specific residues that projected into solution (Fig. 4-10B). The 3 regions we chose to pursue contained the following residues: region 1 (R1), V249/R254/R285/D287/H311; region 2 (R2),

H421/H425/Q324/D362/R365; and region 3 (R3), K203/P204/D239/E240/D295.

4.2.5 ETA mutants in Region 2 display defect in cell-to-cell spread and retrograde trafficking

To probe the functional roles of surface clusters identified in ETA, we altered multiple residues within each cluster (Fig. 4-10B). Additionally, the calcium-binding region was changed to test whether the potential calcium-mediated dimerization of UL37 is important for function of the protein. Within each region, five of the residues identified by ETA were changed to either eliminate a bulky side chain or, in two cases, to replace a small side chain with a bulky one. Four mutants of UL37 were generated: region 1 (R1),

V249R/R254A/R285A/D287A/H311A; region 2 (R2),

H421A/H425A/Q324A/D362A/R365A; region 3 (R3),

K203A/P204Q/D239A/E240A/D295A; and calcium-binding region (Ca):

117

Figure 4-9: Phylogenetic tree of UL37 from the Evolutionary Trace Analysis (ETA) using UL37 homologs from 15 alphaherpesviruses. ETA analysis was carried out at partition 5 to avoid high background levels associated with later partitions.

118

Figure 4-10: Mapping of conserved residues from ETA analysis and identification of potential functional regions (A) ETA class-conserved and class specific residues are shown in green and red, respectively. (B) Mutated residues in region 1 (R1), region 2

(R2), region 3 (R3), and calcium-binding site (Ca) are shown in blue, pink, orange, and red, respectively.

119

D79A/D81A/E82A/D382A/D383A.

All mutations were introduced into the PRV strain Becker background, and each mutant virus was observed to propagate to wild-type titers. To further investigate these mutants, viral replication and spread were measured in a single-step growth and plaque formation assays, respectively. The Ca, R1, and R3 mutants did not display any reduction in plaque size or viral propagation. In contrast, the R2 mutant plaques were restricted to about half the diameter of virus encoding wild-type UL37 (Fig. 4-11A). A defect in plaque formation can be the result of a defect in cell-cell spread or be due to delays in propagation kinetics. To address this question, the rate of cell-associated virus production and virus release into the supernatant were measured. The amount of virus released into the supernatant by the UL37 R2 mutant virus was not significantly different from WT virus at any of the time points evaluated (Fig. 4-11B), suggesting that the R2 mutations cause a defect in cell-cell spread. Furthermore, this defect in cell-to-cell spread is not due to the UL37 protein being misfolded as western analysis reveals that the UL37 protein is incorporated into extracellular virions. We hypothesize that the R2 cluster serves as a potential binding site for as of yet unidentified cellular or viral protein(s) important for

UL37 function in virus trafficking that is essential in cell-cell spread.

In line with this hypothesis, additional experiments investigating viral trafficking along the stalks of cultured neurons, revealed a complete abrogation of viral retrograde trafficking for the R2 mutant virus (Alexsia Richards and Greg Smith, unpublished data).

120

Figure 4-11: Propagation and spread of PRV encoding mutant forms of UL37. (A)

Relative plaque diameters of mutant viruses: Ca, R1, R2, and R3. Plaque diameters are compiled from three individual experiments per virus and are plotted as a percentage of the average wild-type plaque diameter determined in parallel. Mean diameters are indicated by a horizontal bar with error bars representing standard deviation. (B) Single- step growth curves comparing propagation of PRV-GS4284 (UL37 WT) and PRV-

GS5604 (UL37 R2 mutant). Virus was harvested from cell media (dashed lines) and cells

(solid lines) and titers were measured by plaque assay. (C) Western blot analysis of UL37 protein incorporation into wild-type (WT) and R2 mutant (R2) extracellular virions.

Virions were probed with antibodies raised against UL37 or VP5, with the latter serving as loading control. The relative incorporation of the UL37 R2 mutant protein was calculated as a percentage relative to the wild-type protein. A representative blot is shown from four independent experiments.

121

This observation is consistent with UL37 playing a significant role in the trafficking of herpesvirus capsids in the cell.

4.2.6 UL37N possesses structural similarity to eukaryotic proteins of known function

The UL37 proteins have no notable sequence homology to any viral or host proteins. In addition to the identification of putative functional regions of UL37N through mapping of conserved residues and ETA analysis, we searched for structural similarities between

UL37N and proteins of known function. Proteins that possess similar structures often also have similar functions. Thus, we used the DALI (64) structural homology search program on the full UL37N protein, as well as on individual domains of UL37N to try to identify structural similarities that could provide major insight into the function. The DALI search did reveal two classes of proteins that shared structural similarity to domains of UL37.

The strongest structural similarity is observed between domain I of UL37N with a specific family of multi-subunit tethering complexes (MTCs). The second observed structural homolog is the 14-3-3 class of proteins, which possesses a weak structural similarity to domain III of UL37N. Domain II did not possess any structural similarity to any proteins.

4.2.6.1 Domain 3 of UL37N possesses a weak structural similarity to 14-3-3 proteins

There are five helices (α15, 18, 19, 20, and 22) of domain III, which can be superposed onto five helices of a 14-3-3 protein Tau (Fig 4-12). The Z-scores for these aligned

122 residues are in line with Z-scores for other structural homologues of UL37N (i.e. MTCs), with values between 3.6 and 5.0. A few caveats to this structural similarity is that is involves only half of the 14-3-3 monomer, and it is known that the region missing is important for protein-protein interactions of 14-3-3 proteins. Additionally, the putative dimerization interface that would be required for 14-3-3-like proteins to function is blocked by domain II (Fig 4-12).

14-3-3 proteins are known to bind small molecules, and these ligands often activate the protein, which gives it the ability to regulate various cellular functions (depending on the

14-3-3 protein). Several cellular processes have been implicated to be regulated by 14-3-

3 proteins, for example, the coordination of microtubules, regulation of apoptosis, and ER export (50, 118, 180). It is evident that herpesviruses would likely benefit through regulation of these processes (especially coordination of microtubules). However, based upon the fact that important functional regions of 14-3-3 proteins are missing from this aligned region and that the putative dimerization interface of the 14-3-3 like region of

UL37N would be blocked by domain II, it seems unlikely that this structural similarity is functionally significant.

4.2.6.2 UL37 shares a structural similarity with subunits of the CATCHR family of tethering complexes

In addition to the structural similarity with 14-3-3 proteins the DALI search (63) revealed that domain I shares structural similarity with several subunits of the CATCHR

(complexes associated with tethering containing helical rods) family of eukaryotic multi-

123

Figure 4-12: Overlay of UL37N and 14-3-3 protein Tau. The UL37N protein is shown in grey, while the 14-3-3 protein Tau dimer is shown in shades of green. The aligned helices

(cyan for UL37N and light green for 14-3-3) are at the dimerization interface of 14-3-3 protein. This putative dimerization interface in UL37N is blocked by domain II, therefore

UL37N is unlikely to function similarly to 14-3-3 proteins.

124 subunit tethering complexes (MTCs) (16, 70). In intracellular trafficking pathways,

MTCs tether vesicles to their target organelles both to bring the vesicles closer to their target membranes and to help ensure the delivery of the vesicle to the correct target organelle (16). There are several subunits of the four CATCHR MTCs (the Dsl1 complex, the exocyst, the GARP complex, and the COG complex) that share strong structural similarities despite low sequence identity (43, 70, 131, 165), which points to their common evolutionary origin and mechanistic similarities. Their structures consist of one to five helical bundle domains of similar folds (43, 131, 165). UL37N shares the highest structural similarity with several subunits of the Dsl1 complex and the exocyst, with Dali

Z scores of 4.2-5.4 for alignments with full-length UL37N (Fig. 4-13A and Table 2), while similarity to other MTC subunits is less pronounced. Although the similarity scores for the top “hits” are modest, they are comparable to scores for some of the more distantly related MTC subunits. The structural similarity to MTC subunits is particularly remarkable because the sequence identity is under 10% (63).

Residues 1-136 of domain I resemble the helical bundles of MTCs the most and typically align with the domain C of MTCs (70). A wire representation showing the structural similarity of MTCs with the first 136 residues of UL37N is show in Figure 4-

13B. However, the structural similarity between UL37N and MTCs extends beyond domain I and includes domain II (Fig. 5). The domain I+II module resembles domains

C+D+E of MTC subunits and has an overall J shape (Fig. 4-13A) found in some tethering subunits as the result of an additional domain E that follows domains C and D. Although the tip of domain II of UL37N only remotely resembles domain E of MTC subunits, the folds of domains E are highly divergent even among MTC subunits themselves (Fig. 4-

125

126

Figure 4-13: UL37N shares structural similarities with several subunits of cellular MTCs.

(A) UL37N is shown side by side with Sec6 and Exo70 of the exocyst complex and Dsl1 and Tip20 of the Dsl1 complex. Structure alignments were carried out using Dali (64), with the Z scores of 5.0, 4.4, 4.0, and 5.4, respectively. Domain I of UL37N aligns with domain C of these proteins (green). The overall J shape of domains I and II of UL37N is reminiscent of these MTC components. (B) Overlays of Sec6, Dsl1, and Exo70 onto

UL37N. (C) UL37N possesses a salt bridge similar to one observed in Cog4. The salt bridge between residues D216 and R260 in domain II between putative D and E subdomains of UL37N is strictly conserved among alphaherpesviruses and likely plays a structural role. This salt bridge resembles the inter-domain salt bridge between residues

R729 and E764 of Cog4 (132).

127

Table 4-2: Alignments of UL37N with components of MTCs Protein Z-score rmsd aligned residues %id TIP20 5.4 10.1 186 4 Sec6 5.0 3.9 145 7 Exo70 4.4 15.7 182 6 Dsl1 4.2 11.1 159 6 Cog4 3.6 4.7 135 10 Sec15 3.2 3.4 72 4 Exo84 2.5 4.1 75 4

Alignments of UL37N (helical bundle, residues 1-136) with components of MTCs Protein Z-score rmsd aligned residues %id Exo70 5.8 3.3 123 9 Dsl1 5.7 3.8 98 7 Exo84 5.4 4.0 102 9 Sec6 5.1 3.7 111 8 Sec15 4.4 3.5 97 4 Cog4 3.8 4.6 107 10 Tip20 3.8 3.8 92 14 All alignments were carried out using DALI (64). Either the entire UL37N or only residues 1-136 were used in the Dali search.

128

13). Of note, Cog4 has a salt bridge between domains D and E that involves a conserved arginine (131). UL37N also has a salt bridge between D216 and R260 in putative subdomains II-D and II-E, respectively (Fig 4-13C) that is conserved among alphaherpesviruses (Fig. 4-7).

Despite the noticeable similarity, the UL37N structure differs from those of MTC subunits in two aspects. First, instead of multiple helical bundles of similar topology

(152), it only has one helical bundle of the topology similar to those found in MTC subunits. Second, unlike in MTC subunits, where domains D or E are C-terminal, the polypeptide chain in UL37N continues into domain III and a hairpin in domain I. Thus, the structural resemblance of domains I and II of UL37N to the MTCs may be the result of convergent evolution.

4.3 Discussion

UL37 is one of the largest and most highly conserved herpesvirus tegument proteins yet little is known about its role during the herpesvirus infectious cycle. In this study, we report the crystal structure of the N terminal half of PRV UL37, UL37N, which revealed an elongated molecule rich in helical bundles that shares structural similarity to eukaryotic multi-subunit tethering complexes (MTCs). This is the first structure of an inner tegument protein from an alphaherpesvirus and only the third structure available for any herpesvirus tegument protein, in addition to the structures of the transcription activator UL48 (VP16) from HSV-1 (100) and the deubiquitinase domain of the UL36 homologue of murine cytomegalovirus (143). The structure provided a three-dimensional

129 framework for investigating the function of UL37 N terminus and a first glimpse into the architecture of the inner tegument of alphaherpesviruses. Using evolutionary trace analysis, we identified several conserved surface regions and explored their importance using mutagenesis. We found that one of these, the R2 cluster, is dispensable for virus propagation in non-differentiated cells yet important for viral cell-cell spread, which suggests that UL37 is necessary for intracellular virus trafficking to cell junctions.

Additionally, this R2 mutant cluster was found to be essential for retrograde trafficking in the stalks of cultured neurons, further confirming the importance of UL37 in intracellular capsid trafficking. While UL37 has been previously implicated as an effector of both capsid microtubule transport and secondary envelopment (82, 107, 121, 123, 129, 140), these new insights into UL37N structure and function indicate UL37 also coordinates viral spread between cells. The UL37 protein is critical for efficient propagation of both

PRV and HSV-1, however, in PRV UL37-null virus small amount of residual propagation is still able to occur (resulting in a several log decrease in infectious virus production); in contrast, HSV-1 propagation in the absence of UL37 is totally abrogated

(41, 82). With this in mind, it is interesting to speculate if the R2 mutant region would have more severe defects in HSV-1 infection. However, whether R2 functionality is conserved in herpesviruses other than PRV requires additional testing.

Although the mechanism by which UL37 participates in viral trafficking to cell junctions is unknown, we speculate that the R2 cluster is a potential binding site for a viral or a cellular co-factor that promotes viral transmission between cells. The large tegument protein UL36, the best-characterized binding partner of UL37, is essential for UL37

130 localization to capsids (30, 81, 140, 169). Additionally, the role of UL36 in viral cytosolic trafficking is well documented (104, 140, 149). However, UL36 is unlikely to interact with the R2 cluster, as it has been previously shown to interact with the C- terminal end of UL37 (17), and disruption of the interaction between UL36-UL37 results in a more severe defect in viral replication (49) than was observed with the R2 mutant virus. HSV-1 UL37 has also been shown to interact with capsid proteins UL35 and UL38 and the outer tegument protein UL46 in yeast two-hybrid screens (93), but these interactions have not yet been further validated. Furthermore, interactions between UL37 and capsid proteins may not be biologically relevant as UL37 recruitment to capsids appears to be solely dependent upon UL36 (49, 83, 140). In addition to viral proteins,

UL37 also associates with host protein dystonin/BPAG1 (68), a protein involved in microtubule-based transport (139). Although dystonin has been shown to be necessary for efficient capsid trafficking during entry and egress (68, 107), the dystonin-binding region also has been determined to reside within the C terminus of UL37 (68). Thus, UL37 likely binds additional, yet unidentified cellular or viral protein partners using its N- terminus.

While the R2 cluster attests to the importance of UL37 for cell-cell spread and expands the repertoire of UL37 functions in intracellular virus trafficking, other conserved clusters identified on the surface of UL37N may play important functional roles despite not displaying defects in our assays. Additional studies should monitor the phenotypes associated with R1, R2 and R3 cluster mutants in animal infection models, which often show more severe defects. As PRV UL37-null virus shows some ability to replicate in

131 cell culture models, but is required for neurovirulence in mice (79). By contrast, the calcium-binding site is unlikely to play any functional roles as the acidic residues chelating calcium ions are not well conserved among alphaherpesviruses (Fig. 4-7).

It is interesting to note that the defect in cell-to-cell spread observed in R2 in domain III was found to be outside of the domain I, which possesses the structural similarity to

MTCs. This may imply that the UL37 protein has multiple functions during herpesvirus egress. This is highly likely as it has already been shown that in addition to aiding in capsid trafficking, that UL37 homologues plays roles in modulation of the immune system, by either suppressing the inflammasome response to KSHV infection (55) and activation of the NFkB pathway at early times in HSV-1 infection (99).

The structural similarity to eukaryotic multi-subunit tethering complexes (MTCs) as it may further suggest that UL37 is actively involved in the secondary envelopment, by tethering capsids to cytoplasmic vesicles. MTCs control vesicular trafficking by tethering incoming vesicles to their target organelles and helping ensure the correct “match” between the vesicle and its destination (16). These “matches” are determined by protein- protein interactions of the MTC complexes, most often with SNARE and Rab proteins, which decorate the surfaces of intracellular organelles and vesicles. Functions of some

MTCs may even extend beyond tethering vesicles, e.g., recruitment of dynein to kinetochores during chromosome segregation by the mammalian homolog of yeast Dsl1 complex (145) or interaction of the exocyst with microtubules during vesicle transport .

The four MTC subunits with the highest structural similarity to UL37N are components

132 of either the Dsl1 complex (TIP20 and Dsl1) or the exocyst complex (Sec6 and Exo70).

Both the Dsl1 complex and the exocyst coordinate tethering of Golgi-derived vesicles to the endoplasmic reticulum or the plasma membrane, respectively. The notable structural resemblance between UL37N and MTC subunits suggests that UL37 could be the first viral MTC mimic. Structural determination of the UL37 C-terminal half could assist in further reinforcing this structural similarity, as secondary structure predictions indicate that the C-terminal half of UL37 is also predominantly helical in nature. Additionally, it needs to be determined if this structural similarity extends to a functional similarity, perhaps by mediating similar protein-protein interactions (i.e. interactions of UL37N with

Rab or SNARE proteins), such finding would strongly reinforce the role for UL37 in tethering capsids to secondary envelopment vesicles. A model which combines all of these potential functions for UL37 is shown in Figure 4-14.

Considering that the herpesvirus capsids undergo cytoplasmic envelopment at tubular vesicles derived from either the TGN or early endosomes (62, 113, 122), our data imply that herpesviruses may have co-opted MTC functionality of UL37 to bring capsids to target membranes, perhaps, even using UL37 to specify the cytoplasmic budding destination for the capsids. This idea is additionally supported by the observation that in cells infected with HSV-1 UL37-null virus, cytosolic capsids do not co-localize with vesicular membranes (96, 121). Although the precise mechanism by which UL37 participates in virus trafficking remains enigmatic, its similarity to subunits of cellular

MTCs and the identification of the function R2 regions, will hopefully guide future work towards defining more precisely the roles of UL37 in herpesvirus replication.

133

Figure 4-14: Model of UL37 function in viral egress. Green stars indicate UL37 in potential roles of virus egress. Previous studies have implicated UL37 to be important for trafficking and secondary envelopment of herpesvirus capsids. Results of R2 mutants with cultured neurons further implicate the involvement of UL37 in capsid trafficking.

Meanwhile, the structural similarity to MTCs suggests that UL37 may function to tether capsids to secondary envelopment vesicles, thus specifying the sites of secondary envelopment. Additionally, the reduction in plaque size observed for R2 mutants reveal that UL37N may be playing an additional role in trafficking mature capsids to cell junctions, thus promoting efficient cell-to-cell spread.

134

Chapter 5: Identification of PRV UL37N putative host protein binding

partners

135

5.1 Introduction

Based upon the observed structural similarity between PRV UL37N and several different components of multisubunit tethering complexes, we proposed that UL37 may have some shared functions with these proteins. We have hypothesized that UL37 may be functioning to direct capsids to sites of secondary envelopment, and upon reaching these sites aide in tethering the viral capsid to the tubular vesicle, in a manner similar to vesicle tethering mediated by MTC’s. The types of proteins that MTC’s use to aide in tethering vesicles to their destination organelles are Rab proteins, SNARE proteins, and in the case of the DSL1 complex, COP coatamer proteins. Additionally the novel functional R2 region has been shown to be important for cell-to-cell spread via a currently unknown mechanism. It is possible that this region binds to an unidentified host or viral protein to mediate this function.

There have been several studies that have identified host binding partners of UL37 proteins from various herpesviruses (55, 77, 99). Some of these interaction have roles in immune regulation, for example the interactions between TRAF6 with UL37 from HSV-

1 (99) (leading to NFkB activation) and NLRP1, NLRP3, and NOD2 with ORF63 (UL37 homologue) of KSHV (55) (leading to suppression of the inflammasome). The only larger scale proteomic study to date looking to identify host proteins that interact with

UL37 was done using a yeast-two hybrid approach using the HSV-1 UL37 protein as bait

(77). This study identified 10 proteins as putative interactions of HSV-1 UL37; among them are the kinase TAOK1, the dead box protein DDX5, the GAP protein GARNL1, and beclin1. However, only the TAOK1 (thousand and one kinase) has been validated by

136 additional co-IP experiments. Moreover, with all of these studies, the host proteins which were identified are not in line with the classes of proteins we expect to see interacting with MTC like proteins or to be involved in cell-to-cell spread.

For this study, we sought to identify additional host proteins that may be interacting with

PRV UL37N using a combinatorial pull-down approach with subsequent identification of proteins using mass spectrometry. Identification of proteins that also are known to interact with MTC’s may allow us speculate that UL37 and MTC’s share not only a structural similarity, but a functional one as well. This would indicate a likely role for

UL37 in specifying and tethering capsids to the tubular vesicles used for secondary envelopment.

5.2 Results

5.2.1 Identification of putative host cell binding partners of UL37N

To identify potential interactions between PRV UL37N and host proteins, we performed pull-down experiments in which mammalian cell lysates from either pig kidney (PK15) or green monkey kidney (Vero) cells were passed over cyanogen-bromide (CNBr) activated sepharose with the purified UL37N protein attached. Both of the cell lines being used are permissive for PRV infection. The use of two cell lines should help in determining conserved interactions between PRV and the host cell that are used for viral replication in a non-host specific manner. The pull-down approach is advantageous for two additional reasons. First, we are using PRV UL37N protein that does not contain any

137 tags, thus negating any non-specific interactions with tags associated with GST or His-tag pull-down experiments. Additionally, we have two controls which should further remove non-specific interactions. In the first control, we passed lysate over sepharose without any

UL37N bound, while in the second control we passed lysis buffer alone (no lysate) over

UL37N-sepharose. Proteins from the experimental lysate as well as both controls that are retained on the column were run on a gel and identified by mass-spectrometry analysis.

Using this approach, we found hundreds of proteins for both cell types, while only a handful of protein bands were observed in the sepharose controls without bound UL37N

(Fig. 5-1). To identify these bands, gel sections were excised from the Coomassie stained gel and sent for analysis by mass spectrometry. In addition to proteins bound by UL37N sepharose, samples of both controls were also analyzed to identify any potential contaminants obtained during purification and handling of gels as well as proteins that bind non-specifically to sepharose. Peptides were identified using databases from the pig proteome (for PK15 cells) or the human proteome (for Vero and PK15 cells). Proteins from the samples were considered confidently identified when 3 or more unique peptides from that protein were found. Proteins identified based upon 1 or 2 peptides were not considered confidently identified and were not further considered as potential UL37N binding partners.

Mass spectrometry performed on the pulled-down fractions confidently identified 312 proteins from Vero cells and 132 from PK15 cells, none of which were identified in either of the controls. A likely reason for the much smaller number of proteins identified

138 from PK15 cells is the limitations in the current pig proteome database. Many proteins in the pig database are currently partial segments that likely are missing regions of the expressed protein and will produce fewer confidently identified hits. Despite this, there were 26 proteins that were confidently identified from pull downs in both cell lines.

These 26 proteins listed in Table 5-1 represent the most likely host protein binding partners of UL37N.

Excitingly, included in this list of 26 proteins are several proteins involved in vesicular transport. Small GTPase proteins Rab5 and Rab14 were identified in both cell types; additionally, the Rab6 protein was identified confidently from PK15 cells, but only had 2 peptides identified from Vero cells. Nevertheless, I believe the possible interaction with

Rab6 is an interesting one to pursue further. In addition to the Rab proteins identified, the

Sec22b SNARE protein was also pulled-down. This is exciting as it falls in line with

UL37N mimicking MTC function. A couple additional proteins of interest that were identified were the COP coatamer beta, another protein involved in vesicular transport, and the heavy chain of dynein, which may provide further evidence for UL37 being important for intracellular trafficking of capsids. We are currently unsure of how many of the other identified potential interacting partners may function with the UL37 protein. For example, several mitochondrial proteins were pulled-down in both cell lines. Although there is some evidence for disruption and trafficking of mitochondria during herpesvirus infection (84), UL37 has never been observed to play a role in this phenomenon.

139

Figure 5-1: UL37N pull-down of host-cell binding partners. Coomassie stained gel of

UL37N pull downs from PK15 and Vero cell lysates reveal hundreds of proteins, which are pulled down in the presence of UL37N. Sepharose alone controls show only small subsets of these proteins bind to sepharose non-specifically. The Coomassie stained gel was sliced into sections and the proteins identified through Mass Spectrometry analysis.

140

Table 5-1: Potential UL37N binding partners Protein MW Localization Function Sarcoplasmic/Endoplasmic reticulum calcium ATPase 2 114 kDa ER/SR Calcium transport cell junction Rap1B 20.8 kDa PM maintenance Cytoskeleton-associated protein 4 66 kDa ER ER integrity Signal recognition particle receptor subunit ER protein targeting alpha 69.8 kDa ER GTPase ER protein Sec61 subunit beta 10 kDa ER translocation Calcium-binding mitochondrial carrier protein Aralar1 74.7 kDa Mito glu-asp transport Calcium-binding mitochondrial carrier protein Aralar2 74.1 kDa Mito glu-asp transport Dolichyl-diphosphooligosaccaride STT3A 80.5 kDa ER glycosyltransferase Guanine nucleotide binding protein subunit Int alpha-13 44 kDa Membranes G-protein Int integral membrane Extended synaptotagmin 1 122 kDa Membranes protein Int integral membrane Extended synaptotagmin 2 102 kDa Membranes protein Transmembrane and coiled coil domains integral membrane protein 1 72 kDa PM protein Trifunctional Enzyme subunit alpha, mitochondrial 83 kDa Mito lipid metabolism macropinocytosis RhoG 21.3 kDa PM induction Annexin A2 38.6 kDa ECM membrane binding Voltage-dependent anion selective channel protein 1 30.7 kDa Mito membrane channel Voltage-dependent anion selective channel protein 2 31.5 kDa Mito membrane channel Mitochondrial inner membrane protein mitochondrial isoform 1 83 kDa Mito transport Armadillo repeat containing protein 10 37.5 kDa ER p53 regulator peroxisomal Peroxisomal membrane protein 11B 28 kDa Peroxisome biogenesis plasminogen Transmembrane protein c9orf46 17.2 kDa PM receptor Cytoplasmic Dynein Heavy chain 1 532 kDa cytoskeleton retrograde transport Rab5 23.6 kDa EE vesicle transport Rab14 23.9 kDa Golgi/EE Vesicle transport Coatamer subunit beta 107 kDa Golgi Vesicle transport Sec22b 24.5 kDa Golgi Vesicle transport

Rab6 23.6 kDa Golgi Vesicle Transport

141

5.2.1.1 Rab and SNARE proteins identified

The pull-down of three Rab proteins and one SNARE protein was exciting, as these types of host-protein interactions with UL37N could provide additional support for UL37 mimicking MTC functionality during herpesvirus infection. Furthermore, the Rabs identified localize to either the early endosome (Rab5 and Rab14) or the trans-Golgi network (Rab6). These two intracellular membranes are thought to be the most likely sites for secondary envelopment of alphaherpesviruses (62, 82, 121). Furthermore, several proteomic studies have been conducted on herpesvirus virions (85, 102, 168) and have identified Rab proteins as being present in extracellular virions. In PRV, both Rab6 and Rab14 have been found; for HSV-1, Rab5 and Rab6 were observed; even in more distant viruses such as the gammaherpesvirus murid herpesvirus 4, both Rab6 and Rab14 were found to be incorporated into extracellular virions (168).

Additional studies have investigated the effects of siRNA knockdowns of Rab proteins on

HSV-1 replication. In the case of both Rab 5 and Rab6 knockdown, over a one log decrease in infectious viral titer was observed (73). Rab14 also displayed over a 0.5 log reduction in infectious virus titer and was among the top 10 Rab proteins that aide in efficient HSV-1 replication (73). Therefore, it appears that these Rab proteins are playing active roles in herpesvirus replication in infected cells, and it is possible that a part of this role is interaction with the UL37 protein. In line with Rabs playing active roles in herpesvirus replication, the Rab6 protein has been shown to play a role in morphogenesis of human cytomegalovirus though an interaction with the tegument protein pp150 and

Bicaudal D1 (67). Also, Rab5 and Rab11 have been implicated to be involved in

142 secondary envelopment of HSV-1 (62), although the exact roles of Rab proteins in these processes are not well defined.

To further confirm that the Rab proteins were being pulled down by UL37N, we repeated the assay three additional times. In these subsequent experiments, we also performed western analysis probing for pull-down of Rab proteins 5 and 6. In two out of the three pull-downs we were able to observe that both Rab5 and Rab6 pulled-down when passed over UL37N Sepharose but not when passed over sepharose alone (Fig 5-2). There was one replicate that failed to identify Rab5 and Rab6 pull-down with UL37N although the proteins were observed in the loading control. The pull-down of Rab5 and Rab6, did seem to be specific for certain Rabs, as a Rab that localizes to the late endosome (Rab7) was not retained on UL37N sepharose in any of the experimental replicates.

In addition to the Rab proteins, the one SNARE protein, Sec22b, is also a very interesting candidate for a host-binding partner of UL37N. This SNARE protein mediates trafficking and fusion of vesicles between the Golgi and Endoplasmic Reticulum (8), and thus may also be present at sites of secondary envelopment. However, none of the proteomic studies to date have identified Sec22b as a component of extracellular virions (85, 102).

As the Rab and SNARE proteins are proteins that were expected to be bound if UL37N was a functional mimic of MTCs we decided to further attempt to validate these interactions.

143

Figure 5-2: Rab5 and Rab6 pull-down with UL37N sepharose. Western blots revealed that both Rab5 and Rab6, but not Rab7 pulled down when lysates were passed over

UL37N sepharose columns. No Rab proteins were observed in the bound fractions of lysates passed over sepharose alone. The western shown is a representative of 3 replicate experiments.

144

5.2.2 Rab and SNARE proteins do not directly bind UL37N

We first sought to determine whether the potential interaction between PRV UL37N and the Rab and SNARE proteins were direct. To test this we chose to perform a GST pull- down assay using purified proteins (Rabs and UL37N). Expression plasmids for GST- tagged wild-type Rab proteins for 5 of the 7 subtypes of Rab5, Rab6, and Rab14 (gifts from David Lambright, Ralph Isberg, and Matthias Machner) were obtained, and the Rab proteins purified using a glutathione sepharose column followed by gel filtration.

Additionally two expression constructs for constitutively-active mutants (Rab5A Q79L and Rab6A’ Q72L) and a dominant-negative mutant (Rab6A’ T27N) (gifts from David

Lambright and Ralph Isberg) were obtained, and the proteins purified, exactly as for wild-type.

To ensure that we could identify Rab-effector interactions using this pull-down approach with purified proteins, we also obtained a positive control for interaction. The LidA protein of Legionella pneumophila is a known effector of Rab6, Rab6A’, and Rab14 (27,

116, 148). An E. coli expression construct for a GST-tagged LidA protein was acquired

(a gift from Matthias Machner) and LidA was purified using glutathione sepharose resin and the GST-tag removed through proteolytic cleavage with PreScission protease.

Consistent with the literature (27, 148), we were able to detect LidA in the bound fraction for the GMppNp loaded wild type Rab proteins Rab6A, Rab6A’, and Rab14 (Figure 5-

3A). In addition LidA was able to interact with the constitutively active Rab6A’, but not the dominant negative Rab6A’. No interaction was observed between GMppNp loaded

145

Rab5A or Rab5C and LidA, which further indicates that there is specificity in Rab- effector interaction.

One caveat with this control is that when the Rab6A, Rab6A’, Rab14 were loaded with

GDP to mimic the inactive state, we still observed binding of LidA. This however was not totally unexpected as it has been shown that LidA binds to Rab6A-GDP with Kd of approximately 5 nM. This affinity is nearly 200 fold weaker than the affinity for Rab6A-

GMppNp with a Kd of 0.03 nM (148); however, it indicates that this assay is not sensitive enough to observe the differences in binding between LidA with the inactive and active states of Rab proteins. The fact that LidA binds Rab6A’ in the GDP state but not to the dominant negative Rab6A’ indicates that there may be additional conformational changes brought on by this point mutation, and the “inactive“ GDP loaded state alone is not enough to abrogate binding of LidA. Thus, there is potential that the Rab proteins are not in the truly active conformation when GMppNp is added. However, we are able to detect direct Rab-effector interactions using this protocol and chose to examine potential interactions between these Rab proteins and PRV UL37N.

Using the same protocol as described above for the LidA-Rab pull-down, I probed direct interactions between UL37N and Rab proteins. Unfortunately, all of the Rab proteins tested were unable to pull-down UL37N, as all UL37N protein was detected in the unbound fraction regardless of whether the Rab proteins were GMppNp bound wild-type or GTP bound constitutively active mutants (for Rab5C and Rab6A’) (Figure 5-3B).

146

Figure 5-3: UL37N does not directly interact with Rab proteins. (A) Coomassie stained gels of GMppNp loaded GST-Rabs pull-down of LidA on glutathione resin. Rab6A,

Rab6A’ (WT and CA), and Rab14 all are able to pull down LidA, while Rab5A, Rab5C, and the DN Rab6A’ are unable to pull-down LidA – indicating the specificity of interaction. (B) Coomassie stained gels of GMppNp loaded GST-Rabs pull-down of

UL37N. The UL37N protein fails to pull-down with any purified Rab protein, indicating that no direct interaction is observed between Rab proteins and UL37N from E. coli expressed protein. Ub = Unbound fraction; W = final wash fraction; B= protein bound to glutathione resin after pull-down; In = input protein (LidA or UL37N) added to glutathione resin with bound GST-Rab.

147

As expected, all GDP bound Rab proteins also failed to pull-down UL37N. This result indicates that there is no direct interaction between UL37N and Rab proteins purified from E. coli lysates. This result may be explained by the requirement of a co-factor or a post-translational modification, which can only be acquired in mammalian cells for one or both of these proteins. It has been previously shown that UL37N is phosphorylated in transfected cells (4), a modification that would not be present in E. coli purified protein.

Another explanation may be that the GST-tag on the Rab proteins is interfering with interaction; however, this is unlikely as GMppNp loaded Rab6A, which had the GST-tag removed, did not co-elute with UL37N protein on a gel-filtration column. Furthermore, several other proteins have been identified as effectors using GST-tagged Rab proteins, indicating that GST-tags do not normally interfere with Rab-effector interactions.

In addition to testing the Rab proteins identified in the pull-down from cell-lysates, I acquired a Sec22b E. coli expression plasmid (a gift from Jonathan Goldberg). Sec22b is a SNARE protein involved in Golgi-ER and ER-Golgi trafficking. SNARE proteins are the second major class of proteins that MTCs use to tether transport vesicles to destination organelles. Therefore, I wanted to determine if this protein directly interacted with UL37N. The expressed protein possesses an N-terminal His-SUMO tag and a C-

Terminal GST-tag. His-SUMO-Sec22b-GST was expressed in Rosetta cells and purified using metal affinity chromatography and the protein further purified by gel filtration. The purified protein was then used in the pull-down assay for purified proteins using the same protocol as for the constitutively active Rab proteins. Similar to the Rab proteins tested

148

Figure 5-4: UL37N does not directly interact with the SNARE protein Sec22b.

Coomassie stained gel of glutathione sepharose pull-down assay using purified Sec22b with a C-terminal GST-tag and UL37N. UL37N does not pull-down in the presence of

Sec22b indicating that there is no direct interaction between these two proteins when expression from E. coli. In = input UL37N protein; Ub = unbound fraction; W1 = 1st wash fraction; W3 = final wash fraction; B = protein bound to glutathione resin after pull- down.

149 in the pull-down assay, UL37N failed to be retained by Sec22b on the glutathione resin

(Fig. 5-4). This again indicates that there is likely no interaction between UL37N and

Sec22b with proteins purified from E. coli.

5.2.3 Rab proteins are not observed to interact with UL37N using cell based methods

To further probe the putative interaction of UL37N and Rab proteins, we decided to use cell-based methods such as co-immunoprecipitation and co-localization. These methods should aide in reconciling whether or not a co-factor or post-translational modification is required for interaction, which would be absent in proteins purified from E. coli. For the cell-based approach of investigating UL37 and UL37N Rab interactions and localization,

I used UL37 and UL37N protein with C-terminal fusions to GFP (UL37-GFP and

UL37N-GFP respectively).

Co-IPs were performed using both Vero cells and 293T cells transfected either with GFP alone, UL37N-GFP, or UL37FL-GFP. The cells were transfected using PEI and were incubated overnight at 37°C. Cells were harvested and lysed by sonication and the immunoprecipitation was performed using the GFP-Trap nanobody system (ChromoTek).

Both Rab proteins and GFP tagged proteins were found in the lysates of both transfected cells types (although expression in 293T cells was much higher than Vero cells).

However, even though the GFP proteins were immunoprecipitated with the GFP-Trap system for both Vero and 293T cells, no Rab proteins were observed in the Co-IP with the UL37-GFP or UL37N-GFP proteins with either cell type (Figure 5-5). This result

150

Figure 5-5: Rab proteins do not co-immunoprecipitate with UL37-GFP fusions. Western blots (using anti-GFP and anti-Rab antibodies) of immunoprecipitation experiments using the GFP-Trap system fail to detect Rab protein in the IP fractions, despite it being detectable in the loaded lysates. This result likely indicates that there is no interaction between UL37 and Rab proteins.

151 implies that UL37N-GFP and UL37-GFP do not interact with Rab proteins. Taken together with the pull-down results (Fig 5-3) using purified proteins, it does not appear that UL37 and Rab proteins interact either directly or as part of a larger complex.

5.2.4 Localization of UL37 and UL37N GFP fusions

In addition to the cell-based interaction method of co-IP, we also sought to determine if we could visualize the localization of UL37 proteins fused with GFP in transfected Vero cells. Vero cells were transfected as was done for co-IPs, except the cells were seeded onto cover-slips. The cells were incubated for 24 hours at 37°C at which point the cells were fixed and permeablized and stained with DAPI to visualize the cell nuclei. The coverslips were then mounted, and the cells visualized by fluorescence microscopy. As expected, GFP alone displayed a diffuse localization throughout the transfected cells

(Figure 5-6). Interestingly, in about 40% of the UL37FL-GFP expressing cells the protein was observed to be localized to a region near the nucleus of the cell. It is possible that this protein is localizing to the ER; however, co-localization with ER markers is required to verify this. In contrast to the localization observed in UL37FL-GFP, the UL37N-GFP construct did not localize to this perinuclear region. Instead there is diffuse staining throughout the cell with several distinct foci of increased UL37N-GFP intensity observed throughout cytoplasm and nuclei of the cells (Figure 5-6). The difference in localization between the full-length and N-terminal UL37N fusion suggests that the C-terminal half of

UL37 is playing a dominant role in the localization of UL37.

152

Figure 5-6: UL37N-GFP and UL37-GFP localize to distinct loci but different loci. GFP when expressed alone localizes throughout the cytoplasm and nucleus of the transfected cell. When fused to full-length-UL37 the GFP signal localizes to perinuclear regions – possibly consistent with ER localization. In contrast, when GFP is fused to UL37N the

GFP signal again is observed in both the nucleus and cytoplasm, with several distinct foci visible in the cytoplasm.

153

Co-localization studies of the UL37N-GFP fusion were conducted using antibodies against the Golgi (GM130) and early endosome (EEA1). These are the sites consistent with localization of Rab5 and Rab14 (early endosome) and Rab6 (trans-Golgi).

Visualization of the merged images reveal that UL37N-GFP does not co-localize with either one of these markers (Fig. 5-7), and thus is not localizing to either site where

UL37N could interact with the Rab proteins identified in the initial pull-down experiment. This further implies that the UL37 protein does not interact with Rab proteins. The localization of UL37N-GFP is still unknown, and it will be interesting to look for co-localization with additional markers to determine where UL37N-GFP is localizing in transfected cells.

5.3 Discussion

Structure based mutagenesis of the PRV UL37 protein revealed a novel functional region

(R2) that is deficient in cell-to-cell spread during infection. We have hypothesized that this region may be a binding site for an unknown cellular or viral protein, and this interaction helps promote cell-to-cell spread. Additionally, domain I the UL37N protein bears a structural similarity to several components of MTCs. This similarity led us to hypothesize that UL37 may also be acting as a tethering protein between viral capsids and vesicles of secondary envelopment. MTCs often most often interact with Rab and

SNARE proteins during vesicular transport, to mediate tethering between the transport vesicle and the destination organelle. If this structural similarity of domain 1 of UL37N extends to a functional similarity, it may be expected that UL37N may also interact with

154

Figure 5-7: UL37N-GFP does not co-localize with Golgi or early endosomal markers.

UL37N-GFP transfected Vero cells were further analyzed for co-localization of GFP signal with markers of the Golgi (GM130) or early endosome (EEA1). The right panel displays the merged images with UL37N-GFP in green and Golgi or early endosome in red. The foci containing intense GFP signal to not co-localize with either marker, suggesting that UL37N-GFP is not localizing to either the Golgi or early endosome in transfected cells.

155

proteins involved in vesicular trafficking (Rabs and SNAREs). These finding highlight two regions in which UL37 may be involved in protein-protein interaction with host-cell proteins.

In this study we attempted to identify host-proteins that potentially interact with UL37N using a pull-down approach using purified UL37N covalently attached to sepharose with a downstream identification of bound proteins by mass spectrometry. Excitingly, using this approach we identified three Rab proteins and one SNARE protein, which fell in line with our hypothesis that UL37N is a functional mimic of multi-subunit tethering factors.

However, subsequent attempts to validate these interactions though pull-down of purified proteins, Co-Immunoprecipitation, and localization failed to detect interaction between these proteins. These results likely indicate that these proteins were non-specifically pulled down in the initial pull-down experiment from lysates of Vero and PK15 cells.

With the inability to validate 3 (and Rab6) out of the 26 proteins identified in this initial screen, it is easy to question how valid the remaining 23 putative host-protein interactions with UL37N. However, in the remaining 23 there are still a couple proteins of interest.

The one remaining protein that falls in line with the hypothesis that UL37 may aide in tethering capsids to secondary envelopment vesicles in a similar fashion to MTCs is the

COP coatamer subunit B. This protein is of interest, as it has been shown that the DSL1 protein of the DSL1 complex also interacts with components of the COP coatamer (5,

156). Additionally, the heavy chain of dynein was pulled down, which may be of interest as UL37 is also known to important for vesicular trafficking of capsids along

156 microtubules (86, 140). However, it is likely that this interaction would require an adapter protein (intermediate or small chain of dynein) as very few proteins bind directly to the heavy chain of molecular motor proteins. Two additional proteins that I find interesting in this pull-down are the Rap1B protein and Annexin A2. The Rap1B protein is a small-

GTPase, similar to Rab proteins, that is important for maintenance of cell junctions. Thus, we could speculate that interaction of UL37N with this protein may provide rationale for the defects in cell-to-cell spread observed with the R2 mutants. Finally, the Annexin A2 protein is interesting as it has been shown to be incorporated into extracellular virions of nearly every herpesvirus for which proteomic characterization has been done for (85, 102,

168). This protein is involved in multiple cellular functions, including sorting of endosomes and the coordination of exocytosis (166), the putative interaction with

Annexin A2 with UL37 thus could be important for secondary envelopment and subsequent cell-to-cell spread.

Before investigating additional interactions identified in this pull-down, it may be beneficial to conduct additional proteomics screens looking for protein-protein interactions with PRV UL37N. One main issue with this current study is that we only were monitoring for interaction with one protein (UL37N). If we expanded the study to include several other herpesvirus proteins of interest (i.e. UL25, UL36, UL37C) we may be able to remove false-positive interactions that pull-down with CNBr sepharose that contains any protein covalently bound.

157

The difference in localization of UL37-GFP and UL37N-GFP is another interesting finding in this section. The localization observed for UL37-GFP is in contrast with, a couple of previous studies, which found a more diffuse localization for transfected UL37

(175). There is one other study however, in which the UL37 protein was found to be localized to perinuclear regions during HSV-2 virus infection (172). Additionally, the localization of UL37N-GFP is quite different from the full-length protein, with diffuse signal throughout the cell, with several foci of intense GFP signal. The nature of these foci is still not determined, but it is clear from co-localization studies that the UL37N-

GFP protein is not localizing with early endosomal or Golgi structures. Additional experiments using a larger panel of molecular markers may aide in the identification of the foci that UL37N is localizing to.

In this study we have not currently been able to validate any UL37 interactions with host- proteins identified from the pull-downs. Thus, at this time we cannot determine if the structural similarity of UL37N to components of MTCs extends to protein functionality.

This does not take away from the structural similarity that is observed between these two proteins, but currently we are unable to say conclusively that UL37N is a mimic of MTC structure for purposes of utilizing the tethering function to transport vesicles without additional experimentation.

158

Chapter 6: Characterizing the HSV-1 UL25/UL36 cbd Interaction

159

6.1 Introduction

The interaction network between capsid proteins and tegument proteins are vital to ensuring proper viral assembly and egress in Herpesviruses. Among the most critical of these capsid-tegument interactions is the association between the auxiliary capsid protein

UL25 and the large tegument protein UL36 (34, 122). In addition to the interaction with

UL25, the UL36 protein of HSV-1 also has been shown to interact with several additional tegument proteins, such as VP16 (83) and UL37 (17, 114). Therefore, UL36 is believed to act as a scaffold for the addition of tegument proteins during the egress process, ensuring proper addition of these proteins to capsid structures. Moreover, both UL36 and

UL37 both have been shown to be required for directed capsid trafficking during viral egress (94, 103, 104, 121, 123, 179). These roles highlight the requirement for proper localization of the UL36 (and in turn additional tegument proteins) with capsid structures during egress.

At 3164 amino acids, the UL36 protein is the largest protein encoded in the HSV genome.

However, only the C-terminal 61 amino acids of UL36 are required to interact with the

UL25 protein – this small 61 amino acid domain is termed the UL36 capsid binding domain (cbd)(34). The goal of the present study was to purify and determine the crystal structure of the complex between UL25 and the UL36cbd. This structure (which would be the first between a capsid and tegument protein) would provide information that would enhance our understanding of the interactions between capsid and tegument proteins in herpesviruses. This knowledge will illuminate new approaches that can be utilized to disrupt this complex, ultimately preventing herpesvirus egress and spread.

160

6.2 Results

6.2.1 Construct Rationale

All expression plasmids for this project were provided by the lab of Dr. Greg Smith

(Northwestern University). The genes were cloned into a pET-Duet-1 vector, which possessed one of two UL36 genes, or one of four UL25 genes. The pET-Duet-1 vector allowed for both UL25 and UL36cbd proteins to be expressed simultaneously off of one plasmid in E. coli after induction with IPTG.

For expression of HSV-1 UL36cbd, we chose to pursue two differentially tagged variants of the protein. The first contained only an N-terminal His6-tag, while the second

UL36cbd protein possessed a larger N-terminal tag containing both a His6-tag as well as a

SUMO tag (Fig. 6-1). Both proteins contained a PreScission protease site allowing for removal of all tags. The SUMO was added to aide in potential solubility issues, found by recombinant expression of UL36cbd.

Four different constructs were made for recombinant expression of UL25. Two of the constructs contained the full-length UL25, while the other two were N-terminally truncated and started at amino acid 45. The rationale for starting two of the constructs at amino acid 45 is that HSV-1 UL25 has been previously purified from E. coli using an N- terminal truncated protein in addition to an N-terminal GST and S-Tag (15). Therefore, we also included an N-terminal GST-tags that could be removed by treatment with

PreScission protease for two of the proteins, while the other two remained untagged (Fig

6-1).

161

Figure 6-1: Linear representations of the UL25 and UL36cbd proteins as well as the expected Molecular Weights for each protein construct used for expression. Expression in

E. coli for each of the proteins was evaluated in addition to whether the expressed protein was soluble after cell lysis.

162

6.2.2 UL25/UL36cbd do not interact from E. coli expressed protein

Each dual-expression plasmid was transformed into the Rosetta pLysS expression strain of E. coli and expression of soluble protein was assessed. In the case of UL36cbd, only protein that contained both the His6 and SUMO tag was able to be expressed in the soluble fraction. Untagged UL25 protein was either not able to be expressed recombinantly (45-580) or was found to be completely insoluble (1-580). However, both of the GST-tagged UL25 proteins were found to be expressed and soluble (Fig 6-1), thus all further experiments were done with His6-SUMO-tagged UL36cbd and GST-tagged

UL25 protein (all data shown is for full-length UL25, but all experiments were also done with GST UL25(45-580)).

Despite the ability to express soluble protein for both UL25 and UL36cbd, we were unable to observe complexes between these two proteins. Lysates that expressed both proteins were unable to pull-down complex using glutathione resin, resulting instead in only purified GST-tagged UL25 (Fig. 6-2A). In the reciprocal pull-down experiment using Ni-NTA UL36cbd was highly enriched, and although a small band was visible for

UL25, subsequent gel-filtration runs separated UL25 from UL36cbd. Additionally, His6-

SUMO-UL36cbd failed to co-immunoprecipitate with GST-UL25 (Fig. 6-2B), providing further evidence that complex between UL25 and UL36cbd could not be obtained when the proteins are bacterially expressed.

It is unlikely that the tags are interfering in the interaction, as GST-tagged UL25 can co- immunoprecipitate with His6-SUMO-UL36 in mammalian cells (personal

163

Figure 6-2: Bacterially expressed UL25 and UL36cbd do not interact. A) Glutathione purification of E. coli expressing both GST-UL25(1-580) and His6-SUMO-UL36cbd. All of the His6-SUMO-UL36 is observed in the unbound and wash fractions and the elution fraction only contains GST-UL25. B) Co-immunoprecipitation using anti-GST antibody also reveals the presence of only GST-UL25 in the bound fraction. These assays indicate that UL25 and UL36cbd do not interact when expressed in E. coli.

164 communication with Greg Smith). Therefore, it is possible that one or both of these proteins need to be expressed in eukaryotic cells to interact, either due to the inability to properly fold in bacteria, or the requirement of post-translational modifications for interaction.

6.2.3 Insect Cell Expression and Co-purification of UL25/UL36cbd

Due to the inability to observe complex formation between UL25 and UL36cbd in bacterially expressed cells, we turned to an insect cell expression system. Bacmids were made for both GST-tagged UL25 proteins (1-580 and 45-580) and for His6-SUMO

UL36cbd from pFastBac1 plasmids provided by Greg Smith’s lab. We were able to confirm that all of the proteins were expressed and soluble in Sf9 cells and wanted to determine if they could interact.

To assess the requirement to express one of both UL25 and UL36cbd in insect cells we designed a co-immunoprecipitation experiment in which we expressed GST-tagged UL25 individually in either Sf9 cells or E. coli. Cells were harvested and mixed with Sf9 cells or E. coli cells, which expressed the His6-SUMO-UL36cbd. The cells were lysed together and immunoprecipitation was done using an anti-GST antibody. This experiment indicated that UL36cbd is required to be expressed in eukaryotic cells, as only His6-

SUMO-UL36cbd expressed from Sf9 cells was able to co-immunoprecipitate with GST-

UL25 (Fig 6-3). Conversely, GST-UL25 expressed in bacteria can still interact with His6-

SUMO-UL36 (from Sf9 cells) indicating that bacterially expressed UL25 is still functional, although the extent of interaction does appear to be less.

165

Figure 6-3: UL36cbd is required to be expressed in eukaryotic cells for interaction with

UL25. Lysates containing E. coli or Sf9 cells expressing GST-UL25 were mixed with E. coli or Sf9 cell lysates containing His6-SUMO-UL36. Co-IP using anti-GST antibody to immunoprecipitate GST-UL25 is able to Co-IP UL36 expressed in insect cells but not

UL36 expressed in E. coli.

166

Upon closer inspection of His6-SUMO-UL36, it appears that the requirement for eukaryotic expression is due to a cleavage event during bacterial expression. E. coli expressed His6-SUMO-UL36 appears to run slightly above the 15 kDa band, while Sf9 expressed protein runs slightly below the 27 kDa band. Notably, there is also a faint band below the 27 kDa band for bacterial expressed protein, likely low levels of His6-SUMO-

UL36 prior to cleavage. This cleavage event is not observed in Sf9 expressed UL36cbd and this is likely the reason why this protein is able to interact with GST UL25.

With the ability to detect interaction between GST-UL25 and His6-SUMO-UL36 by co- immunoprecipitation we next wanted to obtain purified complex for crystallization and biochemical studies. We were able to obtain GST-UL25/His6-SUMO-UL36 complex when Sf9 cell lysates expressing both proteins were purified over Ni-NTA resin (Fig. 6-

4A). However, the complex that was obtained was found to be heavily aggregated as evidenced by its elution in the void volume on an S200 column (Fig. 6-4B). This aggregation is due to the His6-SUMO-UL36cbd as GST-tagged UL25 does not aggregate when purified independently, while purified His6-SUMO-UL36cbd does.

6.2.4 Increasing the solubility of the UL25/UL36cbd complex

In order to continue with crystallization trials and biochemical characterization, we first need to be able to find methods to produce a mono-disperse protein complex. We sought to determine additives that may increase the solubility of UL25 and UL36cbd protein during lysis, as the use of these additives throughout lysis and purification may not only

167

Figure 6-4: UL25/UL36cbd complex obtained from Sf9 cells is aggregated. A) The complex between GST-UL25 and His6-SUMO-UL36cbd can be purified from Sf9 cells using metal affinity chromatography. B) GST-UL25/His6-SUMO-UL36cbd complex run on size exclusion chromatography reveals that the complex is aggregated and present only in the void peak of an S200 column.

168

aide in the solubilization of aggregates, but also provide more pure protein complex. Of the more than 20 additives that were screened, only 3 provided increases in soluble protein after lysis – 10% glycerol, 1M L-arginine, and 0.5M NDSB-256 (Fig. 6-5A).

Each of these 3 additives solubilizes more GST-UL25 as well as His6-SUMO-UL36cbd when compared to not having additives present (Fig. 6-5A). The current purification scheme already includes 5% glycerol, and increasing to 10% likely will not make a significant difference. Unfortunately, L-arginine, the additive that solubilizes the most complex, is not compatible with Ni-NTA resin, thus can’t be used during purification.

Thus, NDSB-256 is the only additive that can be used with our purification scheme which hasn’t previously been attempted. Although the overall protein yield obtained when NDSB-256 was present was slightly higher, the complex is still aggregated, and there is no improvement in mono-dispersity (Fig 6-5B).

Another method to try to obtain unaggregated His6-SUMO-UL36cbd is to redefine the boundaries used for the protein being expressed. Previous studies show that only the C- terminal 61 amino acids of UL36 are required for the UL25 interaction; however we decided to extend the N-terminus of UL36cbd by 10, 20, or 30 amino acids (Fig 6-6A) to determine if extending the boundary of UL36cbd may prevent aggregation. Each of these new constructs was expressed and purified from Sf9 cells as was done for the original

His6-SUMO-UL36cbd construct. Although all of these proteins were able to express soluble protein, all of the protein produced was again aggregated as observed by gel filtration (Fig 6-6B), indicating that extending the boundaries of UL36 does not decrease protein aggregation. This was not entirely surprising, as the secondary structure of the C-

169

Figure 6-5: Additives aide in solubility but not prevention of aggregation of UL25/UL36 complex. A) Representative western analysis of UL25 and UL36cbd expressing cells that were lysed in buffer conditions containing various additives. 10% glycerol, 1M L- arginine, and 0.5M NDSB-256 all show increases in solubilization of UL25 and UL36cbd during lysis. B) A 3mL S200 column gel filtration trace of UL25/UL36cbd complex purified throughout in Tris-NaCl buffer with (red) or without (blue) 0.5 M NDSB-256.

The complex is primarily in the aggregate void fraction regardless of the presence of

NDSB-256.

170

Figure 6-6: Extending UL36cbd domain boundaries does not minimize aggregation. A)

Sequence alignment of the UL36cbd with secondary structure predictions from jpred

(http://www.compbio.dundee.ac.uk/www-jpred/) indicated by green boxes for alpha helices and arrows for beta strands. Starting amino acids for all UL36cbd proteins are marked by blue triangles. B) Linear representations of the new UL36cbd constructs expressed and purified. C) S200 gel filtration analysis indicating that regardless of the N- terminal start site, all of the UL36cbd proteins being expressed are aggregated.

171 terminal half is predicted to be predominantly disordered outside of the terminal 61 amino acids. One last construct, which expressed only 50 amino acids was cloned and expressed in E. coli. This shorter construct removed two arginine residues that may have been the cause of proteolysis observed during bacterial expression. However, brief western analysis of this construct revealed that this shorter construct was still being digested into the smaller 15 kDa fragment. Removal of additional residues from the N- terminal half of the UL36cbd has not been attempted as it likely would start to interfere with complex formation with UL25.

6.3 Discussion

Consistent with previous studies (34, 122, 141) we were able to observe an interaction between the auxiliary capsid protein UL25 and the capsid binding domain of UL36 of

HSV-1 using co-immunoprecipitation. Furthermore, in these studies we for the first time have successfully purified complex of the UL25 and UL36cbd. However, the purified complex is highly aggregated due to the His-SUMO-UL36cbd protein, which was found to aggregate when it expressed and purified individually. The use of additives during purification and altering of the domain boundaries had little effect on decreasing the tendency of UL36cbd, and thus on the complex, to aggregate. Unfortunately, due to the inability to acquire mono-disperse protein complex, further biochemical and crystallization trials were not able to be conducted.

The biggest obstacle in obtaining mono-disperse HSV-1 UL25/UL36cbd complex, has been determining how to prevent aggregation or the UL36cbd. Future work with the

172

UL25/UL36cbd complex should begin with this goal in mind. The first approach would be to attempt to make more finite changes in the boundaries for expression of UL36cbd.

As evidenced by the UL37 protein, sometimes changes as small as one amino acid can provide a lot of additional stability in the protein. Additional approaches could start to look at the expression and purification of UL25/UL36cbd complexes from other herpesvirus proteins. We could also investigate the interaction of UL36 and UL25 of the second putative UL25 binding site (2200-2353) (122), although this interaction has not been investigated in viruses other than HSV-1. The complex of these two proteins has been observed in other alpha-herpesviruses, such as PRV (34). It is likely as both UL25 and UL36 are conserved throughout herpesviruses that this interaction is present in other viruses as well. Making and testing UL25/UL36cbd protein constructs from these additional viruses, may provide the mono-disperse protein required for subsequent structural and biochemical studies.

173

Chapter 7: Discussion and Future Directions

174

7.1 Significance of work

The overall goal of this work was to take an alternative biochemical and structural approach to better understand the proteins of the inner tegument of herpesviruses. The most notable achievement in these studies was the determination of the crystal structure of the N-terminal half of UL37. This structure revealed an unexpected similarity to multi- subunit tethering complexes (MTCs), which lead us to hypothesize that UL37 may be involved in tethering the viral capsid to secondary envelopment vesicles. In line with this hypothesis, pull-down approaches identified several Rab proteins as putative binding partners of UL37, although these interactions were not able to be validated in subsequent experiments. The inability to conclusively identify a host binding protein for UL37, at this time does not take away from the structural similarity observed with MTCs.

However, it is currently unclear whether this similarity between UL37N and MTCs will extend beyond structure to function.

The UL37N structure has also aided our ability to identify regions of the protein that are important for function. Through evolutionary trace analysis we were able to identify a novel functional region (R2) of PRV UL37. Although mutations in this region do not result in severe defects in viral replication kinetics, there is a drastic reduction in the ability of this virus to undergo cell-to-cell spread. Furthermore, the R2 mutants show a complete abrogation of retrograde trafficking in cultured neurons. These results add to a rapidly growing body of work, highlighting a role for UL37 in virus trafficking within the cytoplasm of infected cells. As the R2 region was identified through evolutionary trace

175 analysis, it will be interesting to test whether mutants in this R2 region of UL37 homologues exhibit similar phenotypes.

In addition to the structural studies of UL37N we have been able to biochemically characterize the full-length PRV UL37 and the PRV UL37 C-terminal half. Although both of these proteins were quite difficult to work with, we were able to determine that the UL37 protein is an elongated molecule with a predominantly alpha-helical secondary structure. Furthermore, for the first time we were able to determine the oligomerization state of the UL37 self-association domain present on the C-terminal half. Although

UL37C was determined to be a dimer in solution, the full-length UL37 protein appeared as a monomer – indicating that the N-terminal half of UL37 may play a regulatory role in the observed UL37 self-association. However, additional experiments are required to test this.

Finally, in these studies we were able to purify two complexes of the UL36 protein.

Using pull-downs we, for the first time, were able to observe complexes of PRV UL36-

UL37 and HSV-1 UL36-UL25 in the absence of other viral or host-cell proteins. These findings signify that although the complexes as they are currently expressed are aggregated, the interactions observed between UL36 with UL25 and UL37 are direct.

Taken together these experiments show the first biochemical and structural characterization of purified proteins and protein complexes of the inner tegument proteins of alphaherpesviruses. The structural studies have opened new potential avenues for

176 further study of UL37 through the finding of structural similarities to proteins of known function as well as structure guided mutagenesis. Our findings in this work are consistent with previous studies highlighting the importance of these proteins, especially UL37, in herpesvirus cytoplasmic capsid trafficking and secondary envelopment.

7.2 Future Directions

7.2.1 Characterization of R2 mutants and additional structure based mutagenesis

The identification of the novel functional R2 region of PRV UL37N is important on many levels. First it provides a new tool to be able to study the function of UL37 in viral replication and cell-to-cell spread. Previous studies have characterized UL37 protein mainly through the use of null-mutations, which drastically reduces or totally abrogate infectious virus production. The R2 mutant appears to replicate well in infected cells, however, defects have been observed in cell-to-cell spread and retrograde trafficking in cultured neurons. These studies suggest that the UL37 protein is functioning to enhance cell-to-cell spread possibly through a mechanism related to viral trafficking (to cell junctions?). This R2 mutant virus will provide a great tool for studying the role of UL37 protein in viral replication, and further characterization of the mutant in the context of infection in cell culture and in animal models is needed to help elucidate these roles.

UL37 null-virus of HSV-1 displays a more severe phenotype than UL37 null-virus from

PRV. In HSV-1 lacking UL37 there is a complete loss of infectious virus production in comparison to the 2-3 log reduction in infectious titer observed for PRV UL37 null-virus.

177

It would be interesting to make mutants in the R2 region of HSV-1 and determine if the functionality of this region is conserved in other homologues. It can be posited that mutations in the R2 region of UL37 in HSV-1 would exhibit a more severe defect as the protein is absolutely essential for HSV-1 replication. Regardless of the extent of the defect in HSV-1 however, the observation of defects with mutants in this region would further highlight the importance of the R2 region not only in PRV replication and cell-to- cell spread, but across other alphaherpesviruses as well.

Through evolutionary trace analysis we have identified two additional evolutionarily conserved surface exposed regions (R1 and R3), which may be of interest to further characterize (Fig. 4-10). Although no defects were observed for either of the mutants in any of the cell culture based assays tested, it may be interesting to investigate these mutant viruses in the context of an animal infection. UL37-null viruses have shown the ability to replicate at low levels in cell culture, however the same null-viruses are totally unable to establish infection in neurons after intranasal inoculation in mice (79). Thus, more severe defects can be observed in vivo and it is possible that future in vivo characterization of the R1 and R3 regions may highlight the importance of these regions in neurovirulence or neuroinvasion.

Evolutionary trace analysis of the UL37N structure did reveal additional regions that possessed some conservation. Although we did not highlight these regions for mutagenesis in the first set of experiments it would be very interesting to revisit additional conserved patches identified by the ETA analysis and design mutants to test

178 for functionality in these regions. In addition to regions identified by ETA mapping our analysis of MTCs also revealed the similarity in the salt bridge between the D and E domains of Cog4 and the putative D and E domains of UL37N (Fig 4-13C). Mutational analysis of this conserved salt bridge between residues D216 and R260 of UL37N would be interesting as it may further highlight similarities between UL37N and multi-subunit tethering factors. Although extra attention to this mutation would be required to ensure that any observed phenotypes are not due solely to a complete mis-folding of the protein as a result of mutation of a conserved salt bridge that may also be of importance for the structural integrity of the protein.

Overall, we hope that the structure of UL37N provides many new opportunities for further probing of UL37 function through structure-guided mutagenesis.

7.2.2 Identification of UL37 host-cell binding proteins

Domain I (which bears structural similarity to MTCs) and Region 2 represent two regions in UL37N that may mediate important protein-protein interactions. In our studies we attempted to identify host proteins that interact with PRV U37N, potentially through one of these two identified regions. This analysis revealed 26 putative binding partners, several of which fell in line with our model of UL37 in tethering capsids to secondary envelopment vesicles. Our hopes with these studies was that using untagged UL37 protein to pull-down interacting proteins from cellular lysates would minimize false- positives often associated with affinity tags. Furthermore, using an approach in which we used two different cell lines, we hoped to identify protein-protein interactions that were

179 conserved regardless cell-type. However, we were unable to confirm interactions with 3 of the 26 proteins identified (we were also unable to confirm the putative Rab6 interaction). Thus, it is clear additional approaches may need to be taken to identify host protein interactions of UL37.

In future studies towards identification of host-cell binding partners, it will likely be very advantageous to perform the analysis in parallel with several other proteins. Our lab is interested in the biochemical and structural characterization of several other herpesvirus proteins and screening for binding partners of all of these proteins in parallel would likely aide in decreasing false-positives associated with host-proteins that may be retained on sepharose columns in the presence of any protein. Therefore, an approach that utilizes tagged proteins (i.e. His6-SUMO or GST) and performs pull-downs of UL37 or other proteins (UL31, UL34, UL16, gB, etc.) from transfected cells may provide more confidence in observed interactions.

A slight alternative to this would be to perform the pull-downs using several homologues of PRV UL37 or UL37N. Host-proteins identified using multiple homologues of UL37 would likely provide fewer hits, but the identified proteins should be more likely to be confirmed with additional validation techniques. Using full-length protein (especially of

HSV-1 UL37) would have an added advantage, as there are several known binding partners of HSV-1, which predominantly interact with C-terminal half of the protein. If we are able to identify known binding partners of UL37 we could have more confidence in our ability to discover additional novel interaction partners.

180

7.2.3 Structural Determination of UL37C

The structure of the UL37N protein of PRV has provided a lot of insight into the potential roles for UL37 in herpesvirus replication, through both the structural similarities observed with proteins on known function and the characterization of novel function regions. A structure for the C-terminal half of UL37 would provide a lot of additional information about the PRV UL37 protein. It will be fascinating to see if the structure of the UL37C terminal half further supports the structural similarity to MTC proteins observed with UL37N. Additionally, in is known that the UL37 protein interactions with

UL36 through its C-terminal half (17)and it would be very interesting to observe how conserved this region is on the surface of UL37C. In addition to interacting with UL36, the C-terminal half of UL37 in HSV-1 has also been implicated in the binding of dystonin (123). Although it is not currently known if PRV UL37 also interacts with dystonin, it would be interesting to see if there are conserved regions within this putative binding domain in which the UL37 protein could interact with dystonin. For both of these interactions the structure could likely be used in further characterization of these binding regions and highlight the residues in these regions that may be used to mediate these interactions.

Furthermore, we have shown that the UL37C protein is a dimer in solution. It will be very interesting to observe this dimerization interface in the crystal structure and design mutants to disrupt this interaction. These experiments should shed light on whether self- association of UL37 through UL37C is important for function.

181

The UL37C protein does present many issues in its current state as the protein is highly prone to aggregation and falls out of solution in as little as 5 days. Additionally, the

UL37C protein appears to be sensitive to proteolysis at both ends of the protein – one proteolytic cleavage occurs at residue 498 and a second which appears to be C-terminal to residue 856. Future work on this domain should start with identification of boundaries in which a protein can be expressed that is not proteolyzed during expression and further is readily soluble in solution. Future work could try to identify a core region of UL37C that is soluble and mono-disperse. The two C-terminally truncated UL37C (499-831) and

UL37C (499-856) appear to represent a decent starting point as these proteins do not appear as a dimer on SDS-PAGE gels indicating that we have likely removed the C- terminal truncation product. Further modifications of the boundaries and both the N- and

C-terminal ends of the proteins will hopefully reveal a protein that can be used for structural determination of this half of UL37.

It is possible that the use of additional herpesvirus homologues of UL37 or UL37C may aide in the production of protein that can be used for structural determination. We have previously attempted to purify full-length UL37 from HSV-1 with extremely low yields.

However, this gene could be codon optimized as was done for PRV UL37. We could then attempt to purify and crystallize the full-length or UL37C from HSV-1. As UL37 is conserved among all herpesviridae there are many different homologues of the protein that could be pursued including proteins more distant herpesviruses (i.e. HCMV, EBV,

KSHV etc.). The pursuit of structure through the use of multiple homologues often times greatly increases the chances of successful structural determination.

182

7.2.4 Further investigations into tegument-tegument interactions and capsid- tegument interactions

The interactions mediated by the large tegument protein UL36 between the capsid protein

UL25 and the UL37 protein are thought to be two of the most critical interaction for tegument acquisition during herpesvirus egress. In the current studies we have purified both the PRV UL36-UL37 complex and the HSV-1 UL25-UL36 complex, although in both cases the regions of UL36 currently being expressed is forcing the complexes to aggregate. For the UL36 capsid binding domain we have attempted to increase and decrease the domain boundaries in steps of 10 residues, yet all of these proteins containing altered boundaries persisted to aggregate.

For both regions of UL36 further manipulation of the domain boundaries should be attempted. I have seen with the UL37N (1-476) that sometimes 1 amino acid can drastically affect the biochemical properties of a protein (i.e. protein stability). Thus more discrete alterations to the domain boundaries, especially for the 61 amino acid UL36cbd, may aide in production of mono-disperse soluble complex that can be used for subsequent crystallization and biochemical studies. For the UL37 binding domain of

UL36 it is possible that we can express a much smaller region and still get binding to

UL37. As yeast-two-hybrid studies investigating PRV UL36-UL37 interaction identified several UL36 boundaries that could mediate binding. They found that UL36 peptides spanning from residue 214-398 and 312-478 (81) could both interact with UL37. Thus, although not tested directly, this may indicate that the minimal essential binding region of

UL36 interaction with UL37 is 312-398. Furthermore, in HSV-1 the binding site begins

183 at residue 512 (17), which is residue 344 of PRV UL36. Thus, further reduction of the

UL36 boundaries from 344-398 may be sufficient for UL37 binding and possibly for mono-disperse protein, although the ability to produce mono-disperse protein is very difficult to predict.

Similarly to UL37C, we have only approached these complexes thus far with proteins from one virus, PRV for UL36-UL37 and HSV-1 for UL25-UL36. It may be advantageous to attempt to acquire these protein complexes using several other homologues. Both of these interactions are believed to be well conserved among herpesviruses and to my knowledge these complexes have been observed in every virus in which these interactions have been investigated. Thus, the use of homologous protein complex may provide protein complexes that are mono-disperse and can be used for further biochemical and structural characterization.

184

Appendix A: Using gB ecto-domain as an anchor for structural studies

of HSV-1 gB-cytoplasmic tail

185

A.1 Introduction

To initiate infection, viruses must first enter the cell. This is a difficult task for viruses because the virus must traverse the impenetrable cellular membrane before the capsid can be released into the cell. Enveloped viruses utilize a process in which they fuse their viral envelope with the membrane of the host cell, a mechanism known as viral membrane fusion. This result of this fusion event is the formation of a pore, through which the viral capsid can traverse, ultimately leading to infection of the host cell.

The process of viral fusion is energetically very costly. The energy to mediate this is provided by a series of conformational changes by the fusion protein. Initially the fusion protein is in the pre-fusion state, until some trigger (i.e. receptor binding or low pH) induces a conformational change. This initial conformational change is to an extended conformation, which exposes fusion peptides or loops (short hydrophobic regions that can perturb cellular membrane) and insert them into the host membrane. This extended intermediate is extremely unstable, which drives a second conformational change in which the protein starts to collapse into a more stable form. During this second conformational change the fusion protein essentially “holds onto” both the viral and cellular membranes and pulls them together. At this point the outer leaflets of the membranes begin to mix, an event known as hemi-fusion. The fusion protein continues to pull the membranes closer as it continues to alter its conformation, and the inner leaflets of the membrane also begin to mix. Finally, the conformation changes will stop as the fusion protein refolds into its stable post-fusion form and the fusion pore is formed.

186

As with other enveloped viruses, Herpes Simplex Viruses enter cells using this process of viral fusion. The viral envelope of HSV-1 contains 12 glycoproteins, four of which (gD, gB, and the gH/gL complex) are required for fusion (19) The use of four glycoproteins to carry out fusion makes the HSV fusion mechanism more complex than those of other well studied enveloped viruses, as most of these viruses carry out fusion using only one or two proteins. There is evidence suggesting that HSV may carry out fusion through the sequential recruitment of surface glycoproteins. Initially, the receptor binding protein gD binds to one of its cellular receptors. This binding event is thought to trigger or activate the action of the gH/gL complex. The function of the gH/gL complex after activation remains poorly understood, but there is evidence to suggest that it interacts with the fusion protein gB (20, 28) leading to the conformational changes in gB that drive viral fusion.

The crystal structures for the ectodomains of each of the 4 glycoproteins required for

Herpes Simplex Virus Fusion have been elucidated. gD has been solved with and without bound receptor, consistent with it being the receptor binding protein during HSV entry.

The gH/gL heterodimer does not have and structural similarity to proteins of known function, thus further indicating that this complex is an activator protein required specifically for Herpesvirus entry. Finally, the structure of the gB ectodomain reveals a trimeric protein that is structurally related to the post-fusion structure of Vesicular

Stomatitis Virus (135). This similarity indicates that the gB structure that has been solved is also likely in the post-fusion conformation, while the pre-fusion confirmation still remains unknown.

187

In addition to the ectodomain of gB, the protein contains three regions not included in the crystal structure – the hydrophobic membrane proximal region, the transmembrane domain, and the cytoplasmic domain (Figure 6.1). When this 109 amino acid cytoplasmic domain is removed, the resulting recombinant virus is unable to undergo fusion, suggesting the cytoplasmic domain plays a significant role in the fusion process. Further studies on the cytoplasmic domain have revealed more specific regions that are required for fusion, as truncation to amino acid 851 also results in a protein that in unable to mediate viral fusion. Interestingly, smaller truncations in which the protein is prematurely terminated at amino acid 868 or 874 results in a deregulation of viral fusion, resulting in hyper-fusogenic (or syncytial) virus (9). Additionally there are many point mutations in the cytoplasmic domain that also cause a de-regulation of viral fusion(138). These phenotypes suggest that regions of the cytoplasmic domain negatively regulate the fusion event by a yet unknown mechanism. Although this process is poorly understood, the phenotypes observed with the point and truncation mutants are likely due to conformational changes within the cytoplasmic domain. Knowing the structures of the wild-type cytoplasmic domain as well as null and hyper-syncytial mutants may allow a correlation between conformational changes in this domain and the resulting phenotypes.

A.2 Results

A.2.1 Construct Rationale

Previous attempts to determine the structure of the cytoplasmic domain of gB, in which the cytoplasmic domain was expressed by itself, proved to be unsuccessful. We decided that it may be beneficial to link the cytoplasmic domain of gB to the ectodomain of the

188 protein, as this domain had been previously crystalized and the structure determined.

Thus, we would be trying to utilize the ectodomain as an anchor to aide in crystal formation, and ultimately structural determination of the cytoplasmic region.

We first made to constructs, which linked the ectodomain of gB (up to amino acid 730) to the full-length cytoplasmic domain (796-904) containing a C-terminal His-tag using either 6 or 10 amino acid linkers (Fig A-1). The C-terminal His-tag was added to aide in purification of the cytoplasmic tail and ectodomain together in the case that the protein was cleaved over time. We found that this tag was not very useful however as the ectodomain of gB non-specifically bound to Nickel resin, thus addition constructs did not contain the His-tag.

Several additional constructs were designed in which the cytoplasmic domain was linked to the ectodomain (gB730) using the shorter 6 amino acid linker. Two of these constructs possessed either the full-length cytoplasmic domain (796-904) or the truncated hyper- syncytial mutant domain (796-868). In addition cytoplasmic domains beginning at amino acid 801 (instead of 796) were engineered to remove a potentially proteolytically active

Arginine residue at residue 800 and potentially stabilize the ability for the ectodomain and cytoplasmic domains to remain linked. For constructs starting at 801, we designed constructs to express full-length (801-904), hyper-syncytial (801-868), and fusion-null

(801-851) truncation mutants of the cytoplasmic domain (Figure A-1)

189

Figure A-1: Schematic representation of the gB constructs used for these studies. The top panel shows a comparison between the full-length gB protein and the ectodomain- cytoplasmic domain fusion proteins. The domains are separated by color showing the crystal structure of the ectodomain, the membrane proximal region (lime green), the transmembrane domain (purple), and the cytoplasmic domain (cyan). The lower panels reveal a more basic schematic for all of the protein constructs used. The signal sequence

(SS) and mellitin (Mel) sequences (are not included in the final protein as these sequences are removed during processing in the Endoplasmic reticulum. The cytoplasmic domains begin at either residue 796 or 801, and terminate at residues 906(WT),

868(syncytial), or 851 (fusion-null).

190

Having several different constructs expressing the gB ectodomain fused to the cytoplasmic domain, provided several different targets with which we could pursue structural studies of this domain. Additionally, if we can obtain crystal structures of both wild-type and mutant cytoplasmic domains, it may provide valuable insight into how the cytoplasmic domain of gB is involved in regulating viral fusion.

A.2.2 Obtaining purified gB protein containing the cytoplasmic tail

Each of the protein constructs shown in Figure A.1 was cloned into the pFastBac1 plasmid, transformed into DH10Bac E. coli, which transposed each of the cloned constructs from the plasmid into a bacmid. Bacmids were then purified from the E. coli and transfected into Sf9 cells. After 3 rounds of viral amplification, the recombinant baculovirus was harvested and used to infect Sf9 cells for large scale expression. The proteins could be purified from the Sf9 cell cultures 3 days post-infection.

All of the protein constructs were purified using a DL16 antibody column that recognizes the trimeric form of the gB ectodomain (Fig A-2a). After elution of the proteins from the antibody column, the proteins were loaded onto a Superdex 200 gel filtration column, as a polishing step to remove any other protein contaminants from the gB-ectodomain- cytodomain fusion (Fig. A-2b). Nearly all of the constructs produced decent yields of between 1.5-2 mg of protein per liter of culture. A notable exception to this is the truncated null-mutant gB730-L6-cyto(801-851) in which the yields of pure protein were drastically reduced to 0.2-0.3 mg/L, likely due to a large increase in aggregation of this protein. One other notable observation, is that the gB730-L10-cyto(796-906)-His protein

191

Figure A-2: Purification of gB constructs containing the cytoplasmic domain. A) DL16 purification of gB730-L6-cyto(796-904) as a representative of how pure the proteins are coming off of the antibody column. B) Gel filtration of the gB730-L6-cyto(796-904) is used as a polishing step to remove additional impurities and separate the mono-disperse gB protein from aggregated protein found in the void. C) Pure protein for each of the constructs prior to crystallization and further analysis. The gB730-L10-cyto(796-904) protein shown in this panel is nearly 2 weeks old and was used to display how degraded protein appears after the cytoplasmic domain is removed.

192

which contains the longer linker is much more sensitive to proteolysis and much of the protein starts to lose its cytoplasmic tail in as little as 7 days (Fig A-2c).

A.2.3 Electron Microscopy

Using SDS-gel electrophoresis we were able to determine that the cytoplasmic domain was likely present in the gB proteins after purification, due to the increase in the apparent molecular weight on the gel. In addition to this we wanted to be able to visually confirm that the cytoplasmic domain was present in these proteins using electron microscopy. We performed negative stain transmission electron microscopy on each of the proteins within

2-5 days post-purification. The proteins were visualized using 0.75% Uranyl Formate stain and the sizes of each protein were measured using the tape measure tool in photoshop.

We were able to confirm with all of the gB730-L-cyto constructs (except gB730-L6- cyto(801-851)) that the proteins were longer than the gB ectodomain alone. Between 30-

50 individual protein molecules were measured for each construct and the fusion constructs were between 21.6-21.9 nm in length, versus 17.1 nm for the ectodomain of gB without the cytoplasmic domain (Figure A-3 shows a representative panel of the gB proteins possessing the cytodomain). This additional density observed for each of the constructs is likely due to the presence of the cytoplasmic domain. Additionally the fact that we can observe this domain indicates that the cytoplasmic domain is globular and at least somewhat folded. With this additional confirmation about the presence of the

193

Figure A-3: Negative stain electron microscopy of gB constructs. The first three sets of panels show gB730-L10-cyto(796-906)-His, gB730-L6-cyto(796-906)-His, and gB730-

L10-cyto(796-906), respectively. Each of these proteins contains additional density at the end of the molecule (red arrows). This additional density, the cytoplasmic domain, makes the fusion constructs approximately 4.5 nm longer than the ectodomain of gB alone.

194 cytoplasmic domain, we decided to move forward to crystallization trials with these proteins.

A.2.4 Crystallization

Each of the constructs produced was screened for the crystals using approximately 600 different crystallization conditions. Surprisingly, I was able to obtain crystals with nearly all of the constructs, with crystals often appearing between 7 days to 4 weeks after set-up.

The crystallization conditions for all of these constructs were consistent, containing some form of Polyethylene Glycol (3000-8000), a salt (NaCl, calcium acetate, etc.) and a buffer with a pH between 5.5-8.5. The resulting crystals take on many different morphologies from thin plates, to hexagonal crystals, to smaller needle like crystals (Fig

A-4). The one exception, was once again the gB730-L6-cyto(801-856) fusion-null construct, which provided no crystals in any of the conditions tested.

Many of the crystallization conditions proved to be difficult to reproduce, however I was able to obtain several large diffraction quality crystals from both the initial screens used as well as some crystals from conditions that did repeat on optimization plates. In all, I froze 68 crystals (at least 3 from each construct) using either glycerol (20%-27.5%) or xylitol (20%) in the mother liquor as the cryoprotectant.

A.2.5 Analysis of Diffraction

Of the 68 crystals that were sent to the synchrotron to test diffraction, we were able to obtain useful diffraction data from 10 of them (from 5 different constructs), which

195

Figure A-4: gB ectodomain-cytodomain fusion crystals. A representation of the different types of crystals acquired from the gB fusion proteins. All of the crystals shown were acquired from crystallization conditions containing PEG (3000-8000), salt (NaCl, calcium acetate, etc.) and various buffers with pHs between 5.5-8.5.

196 diffracted anywhere between 6.6 - 3.3 angstrom resolution. Unfortunately, all of the crystals that diffracted possessed the same space group (P3) and cell dimension 117 x 117 x 159 angstroms. This happens to be the exact same space group and cell dimensions often observed for diffraction of the gB ectodomain alone (157). This indicates that the cytoplasmic domain is likely proteolytically cleaved off of the protein prior to crystallization – and the crystals that are formed are made up of the gB ectodomain alone.

A.3 Conclusions

We were able to acquire several promising early results as evidence by gel electrophoresis and electron microscopy indicating that the cytoplasmic domain was present in these each of these constructs. However, upon crystallization of these proteins, none of the crystals that were obtained still possessed the cytoplasmic domain, as evidenced by the space group and cell parameters that we observed from the diffraction patterns. This is likely due to the fact that these fusion constructs remain sensitive to degradation over time at or near the linker that fuses the two domains together. I have been able to observe this degradation by gel electrophoresis on protein that is 1-2 months old for the 6 amino acid linker constructs, and as little as 1 week for the 10 amino acid linked construct. Furthermore, the additional truncation of the N-terminal 5 amino acid, which removes the arginine residue at amino acid 800, do not seem to provide any additional benefit, as these proteins seems to be degraded at nearly the same rate as constructs starting at residue 796.

More recent evidence from our lab has also shown that the cytoplasmic domain, adopts a

197 more helical conformation in the presence of lipid and other membrane mimetics (29,

151). These observations indicate that in the absence of lipids that the cytoplasmic domain is likely not fully folded. Thus, testing these constructs in these constructs in the presence of lipids or lipid mimetics may yield more promising results, however, it may not overcome the degradation issues observed with these protein constructs. Thus, current studies towards the gB cytoplasmic domain have focused on reconstitution of full-length gB protein in membranes, as well as the use of membrane mimetics on the cytoplasmic domain of gB alone.

198

Appendix B: Table of Expression Plasmids Used

Table B1: Expression plasmids used in these studies Plasmid Protein(s) Expressed Expression From PRV UL37 pGS3610(pDS) His-SUMO-PP-UL37(FL) Bacterial (Amp) Greg Smith pGS1499 (pEGFP- C1) UL37(FL)-GFP Mammalian (Kan) Greg Smith pGS2905 (pCAGGS) Myc-UL37(FL) Mammalian (Amp) Greg Smith His-SUMO-PP-UL37 pJP10 (pFB) (FL) Insect (Amp) Heldwein Lab His-SUMO-PP- UL37 pJP14 (pET-DUET) (FL) (CO) Bacterial (Kan) Heldwein Lab His-SUMO-PP- UL37 pJP15 (pET-DUET) (FL) (CO) Bacterial (Amp) Heldwein Lab MBP-TEV-UL37- pJP16(pMal) Strep(CO-FL) Bacterial (Amp) Heldwein Lab MBP-TEV-His- Thrombin-UL37- pJP17(pMal) Strep(CO-FL) Bacterial (Amp) Heldwein Lab His-MBP-TEV-UL37- pJP19(p24) Strep(CO-FL) Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37N pJP23 (pET24) (496) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37N pJP27 (pET24) (476) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37C pJP31 (pET24) (494) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37C pJP35 (pET24) (478) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP- pJP55 (pET24) UL37FL(V498G) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP- pJP56 (pET24) UL37FL(V478D) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37C pJP57 (pET24) (499) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37C pJP58 (pET24) (499-831) CO Bacterial (Kan) Heldwein Lab His-SUMO-PP-UL37C pJP59 (pET24) (499-856) CO Bacterial (Kan) Heldwein Lab pJP60 (pEGFP-1) UL37N(496)-GFP Mammalian (Kan) Heldwein Lab

PRV UL36 pGS4351 (pEGFP- C1) GFP-UL36 (Becker) (FL) Mammalian (Kan) Greg Smith

199 pGS1521 (pEGFP- C1) GFP-UL36 (Becker) (FL) Mammalian (Kan) Greg Smith pGS3384 (pEGFP- GFP-UL36 (Becker) (R1- C1) R6 -missing cbd) Mammalian (Kan) Greg Smith pGS4287 (pEGFP- GFP-UL36 (Becker) (R1- C1) R5) Mammalian (Kan) Greg Smith pGS3296 (pEGFP- GFP-UL36 (Becker) (R1- C1) R5 - NLS mutated) Mammalian (Kan) Greg Smith GFP-UL36 (Becker) (N- pGS1722 (pEGFP- terminal third to BglII C1) site) Mammalian (Kan) Greg Smith pJP51 (pET-DUET- Strep-SUMO-UL36(344- SS) 561) Bacterial (Amp) Heldwein Lab Strep-SUMO-UL36(344- pJP52 (pET24-SS) 561) Bacterial (Kan) Heldwein Lab Strep-SUMO-UL36(344- pJP53 (pET24-SS) 561) CO Bacterial (Kan) Heldwein Lab pJP54 (pGEX) GST-UL36(344-561) CO Bacterial (Amp) Heldwein Lab

HSV-1 UL37 His-SUMO-PP- pGS3966(pDS) HSV1UL37(FL) Bacterial (Amp) Greg Smith pGS4474(pCAGGS) Myc-UL37 (FL) Mammalian (Amp) Greg Smith

HSV-1 UL36 pGS3624 (pEGFP- C1) UL36-GFP (Strain 17) FL Mammalian (Kan) Greg Smith pGS3185(pEGFP- C1) GFP-UL36 (KOS)(FL) Mammalian (Kan) Greg Smith His6-SUMO-UL36cbd pGS3911 (pFB) (61AA) Insect (Amp) Greg Smith His6-SUMO-UL36cbd pGS4774 (pFB) (71AA) Insect (Amp) Greg Smith His6-SUMO-UL36cbd pGS4807 (pFB) (81AA) Insect (Amp) Greg Smith His6-SUMO-UL36cbd pGS4819 (pFB) (91AA) Insect (Amp) Greg Smith His6-SUMO-UL36cbd pJP51 (pET-Duet) (50 AA) Bacterial (Amp) Heldwein Lab

HSV-1 UL36/37 HS-UL37 (FL) in MCS1 + SS-UL36 (512-784) in pGS4374 (pDuetSS) MCS2 Bacterial (Amp) Greg Smith

HSV-1 UL25 GS3910(pFB) GST-PP-UL25(1-580) Insect (Amp) Greg Smith

200

HSV-1 UL25/36cbd pGS3053 (pET- UL25(1-580) + His- DUET) UL36cbd Bacterial (Amp) Greg Smith pGS3058 (pET- UL25(45-580) + His- DUET) UL36cbd Bacterial (Amp) Greg Smith UL25(1-580) + His- pGS3068 (pET- SUMO-UL36cbd (OOF DUET) don’t use) Bacterial (Amp) Greg Smith UL25(45-580) + His- pGS3123 (pet- SUMO-UL36cbd (OOF DUET) don’t use) Bacterial (Amp) Greg Smith pGS3318 (pET- UL25(1-580) + His- DUET) SUMO-UL36cbd Bacterial (Amp) Greg Smith pGS3319 (pET- UL25(45-580) + His- DUET) SUMO-UL36cbd Bacterial (Amp) Greg Smith pGS3284 (pET- GST-UL25(1-580) + His- DUET) UL36cbd Bacterial (Amp) Greg Smith pGS3349 (pET- GST-UL25(1-580) + His- DUET) -SUMO-UL36cbd Bacterial (Amp) Greg Smith pGS3334 (pET- GST-UL25(45-580) + DUET) His-UL36cbd Bacterial (Amp) Greg Smith pGS3336 (pET- GST-UL25(45-580) + DUET) His--SUMO-UL36cbd Bacterial (Amp) Greg Smith

Rabs/SNAREs/LidA pGEX6P1-Rab5A GST-Rab5A Bacterial (Amp) David Lambright pGEX6P1- Rab5A(CA) GST-Rab5A (CA) Bacterial (Amp) David Lambright pGEX6P1-Rab5C GST-Rab5C Bacterial (Amp) David Lambright pGEX6P1-Rab6A GST-Rab6A Bacterial (Amp) David Lambright pGEX6P1-Rab6A' GST-Rab6A' Bacterial (Amp) Ralph Isberg pGEX6P1-Rab6A' (CA) GST-Rab6A' (CA) Bacterial (Amp) Mathias Machner pGEX6P1-Rab6A' (DN) GST-Rab6A' (DN) Bacterial (Amp) Ralph Isberg pGEX6P1-Rab14 GST-Rab14 Bacterial (Amp) David Lambright pGEX6P1-LidA GST-LidA (Philadelphia) Bacterial (Amp) Mathias Machner pET24B-Sec22b His-SUMO-Sec22b-GST Bacterial (Kan) Jonathan Goldberg

HSV-1 gB gB730-L6-cyto (796- pKH12 904)- His6 Insect (Amp) Heldwein Lab gB730-L10-cyto (796- pKH13 904)- His6 Insect (Amp) Heldwein Lab pKH39 gB730-L6-cyto (796-904) Insect (Amp) Heldwein Lab pKH40 gB730-L6-cyto (796-868) Insect (Amp) Heldwein Lab

201 pKH41 gB730-L6-cyto (801-904) Insect (Amp) Heldwein Lab pKH42 gB730-L6-cyto (801-868) Insect (Amp) Heldwein Lab pKH43 gB730-L6-cyto (801-851) Insect (Amp) Heldwein Lab

202

References

1. Abaitua, F., M. Hollinshead, M. Bolstad, C. M. Crump, and P. O'Hare. 2012. A Nuclear localization signal in herpesvirus protein VP1-2 is essential for infection via capsid routing to the nuclear pore. J Virol 86:8998-9014. 2. Abaitua, F., and P. O'Hare. 2008. Identification of a highly conserved, functional nuclear localization signal within the N-terminal region of herpes simplex virus type 1 VP1-2 tegument protein. J Virol 82:5234-5244. 3. Adams, P. D., R. W. Grosse-Kunstleve, L. W. Hung, T. R. Ioerger, A. J. McCoy, N. W. Moriarty, R. J. Read, J. C. Sacchettini, N. K. Sauter, and T. C. Terwilliger. 2002. PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr D Biol Crystallogr 58:1948-1954. 4. Albright, A. G., and F. J. Jenkins. 1993. The herpes simplex virus UL37 protein is phosphorylated in infected cells. J Virol 67:4842-4847. 5. Andag, U., T. Neumann, and H. D. Schmitt. 2001. The coatomer-interacting protein Dsl1p is required for Golgi-to-endoplasmic reticulum retrieval in yeast. J Biol Chem 276:39150-39160. 6. Andag, U., and H. D. Schmitt. 2003. Dsl1p, an essential component of the Golgi-endoplasmic reticulum retrieval system in yeast, uses the same sequence motif to interact with different subunits of the COPI vesicle coat. J Biol Chem 278:51722-51734. 7. Antman, K., and Y. Chang. 2000. Kaposi's sarcoma. The New England journal of medicine 342:1027-1038. 8. Aoki, T., M. Kojima, K. Tani, and M. Tagaya. 2008. Sec22b-dependent assembly of endoplasmic reticulum Q-SNARE proteins. Biochem J 410:93-100. 9. Baghian, A., L. Huang, S. Newman, S. Jayachandra, and K. G. Kousoulas. 1993. Truncation of the carboxy-terminal 28 amino acids of glycoprotein B specified by herpes simplex virus type 1 mutant amb1511-7 causes extensive cell fusion. J Virol 67:2396-2401. 10. Baird, N. L., J. L. Starkey, D. J. Hughes, and J. W. Wills. 2010. Myristylation and palmitylation of HSV-1 UL11 are not essential for its function. Virology 397:80-88. 11. Bendayan, M., D. Gingras, E. Ziv, and Y. S. Haviv. 2008. Low-voltage transmission electron microscopy reveals SV40 viral particles within secretory granules in pancreatic cells. Microscopy research and technique 71:659-662. 12. Berger, J. R., and S. Houff. 2008. Neurological complications of herpes simplex virus type 2 infection. Archives of neurology 65:596-600. 13. Bottcher, S., H. Granzow, C. Maresch, B. Mohl, B. G. Klupp, and T. C. Mettenleiter. 2007. Identification of functional domains within the essential large tegument protein pUL36 of pseudorabies virus. J Virol 81:13403-13411. 14. Bottcher, S., C. Maresch, H. Granzow, B. G. Klupp, J. P. Teifke, and T. C. Mettenleiter. 2008. Mutagenesis of the active-site cysteine in the ubiquitin- specific protease contained in large tegument protein pUL36 of pseudorabies virus impairs viral replication in vitro and neuroinvasion in vivo. J Virol 82:6009- 6016.

203

15. Bowman, B. R., R. L. Welschhans, H. Jayaram, N. D. Stow, V. G. Preston, and F. A. Quiocho. 2006. Structural characterization of the UL25 DNA- packaging protein from herpes simplex virus type 1. J Virol 80:2309-2317. 16. Brocker, C., S. Engelbrecht-Vandre, and C. Ungermann. 2010. Multisubunit tethering complexes and their role in membrane fusion. Curr Biol 20:R943-952. 17. Bucks, M. A., M. A. Murphy, K. J. O'Regan, and R. J. Courtney. 2011. Identification of interaction domains within the UL37 tegument protein of herpes simplex virus type 1. Virology 416:42-53. 18. Bucks, M. A., K. J. O'Regan, M. A. Murphy, J. W. Wills, and R. J. Courtney. 2007. Herpes simplex virus type 1 tegument proteins VP1/2 and UL37 are associated with intranuclear capsids. Virology 361:316-324. 19. Cai, W. H., B. Gu, and S. Person. 1988. Role of glycoprotein B of herpes simplex virus type 1 in viral entry and cell fusion. J Virol 62:2596-2604. 20. Cairns, T. M., J. C. Whitbeck, H. Lou, E. E. Heldwein, T. K. Chowdary, R. J. Eisenberg, and G. H. Cohen. 2011. Capturing the herpes simplex virus core fusion complex (gB-gH/gL) in an acidic environment. Journal of virology 85:6175-6184. 21. Carlson, L. A., and J. H. Hurley. 2012. In vitro reconstitution of the ordered assembly of the endosomal sorting complex required for transport at membrane- bound HIV-1 Gag clusters. Proc Natl Acad Sci U S A 109. 22. Carluccio, A. V., S. Zicca, and L. Stavolone. 2014. Hitching a ride on vesicles: Cauliflower mosaic virus movement protein trafficking in the endomembrane system. Plant physiology. 23. Carnec, X., M. Ermonval, F. Kreher, M. Flamand, and M. Bouloy. 2014. Role of the cytosolic tails of Rift Valley fever virus envelope glycoproteins in viral morphogenesis. Virology 448:1-14. 24. Chadha, P., J. Han, J. L. Starkey, and J. W. Wills. 2012. Regulated interaction of tegument proteins UL16 and UL11 from herpes simplex virus. J Virol 86:11886-11898. 25. Chakravarty, S., A. M. Hutson, M. K. Estes, and B. V. Prasad. 2005. Evolutionary trace residues in noroviruses: importance in receptor binding, antigenicity, virion assembly, and strain diversity. J Virol 79:554-568. 26. Chang, J. T., M. F. Schmid, F. J. Rixon, and W. Chiu. 2007. Electron cryotomography reveals the portal in the herpesvirus capsid. J Virol 81:2065- 2068. 27. Chen, Y., and M. P. Machner. 2013. Targeting of the small GTPase Rab6A' by the Legionella pneumophila effector LidA. Infect Immun 81:2226-2235. 28. Chowdary, T. K., T. M. Cairns, D. Atanasiu, G. H. Cohen, R. J. Eisenberg, and E. E. Heldwein. 2010. Crystal structure of the conserved herpesvirus fusion regulator complex gH-gL. Nat Struct Mol Biol 17:882-888. 29. Chowdary, T. K., and E. E. Heldwein. 2010. Syncytial phenotype of C- terminally truncated herpes simplex virus type 1 gB is associated with diminished membrane interactions. Journal of virology 84:4923-4935. 30. Clayson ET, B. L., and Compans RW. 1989. Release of simian virus 40 virions from epithelial cells is polarized and occurs without cell lysis. J Virol 63.

204

31. Cockrell, S. K., J. B. Huffman, K. Toropova, J. F. Conway, and F. L. Homa. 2011. Residues of the UL25 protein of herpes simplex virus that are required for its stable interaction with capsids. J Virol 85:4875-4887. 32. Cockrell, S. K., M. E. Sanchez, A. Erazo, and F. L. Homa. 2009. Role of the UL25 protein in herpes simplex virus DNA encapsidation. J Virol 83:47-57. 33. Cohrs, R. J., J. Wischer, C. Essman, and D. H. Gilden. 2002. Characterization of varicella-zoster virus gene 21 and 29 proteins in infected cells. J Virol 76:7228-7238. 34. Coller, K. E., J. I. Lee, A. Ueda, and G. A. Smith. 2007. The capsid and tegument of the alphaherpesviruses are linked by an interaction between the UL25 and VP1/2 proteins. J Virol 81:11790-11797. 35. Conway, J. F., S. K. Cockrell, A. M. Copeland, W. W. Newcomb, J. C. Brown, and F. L. Homa. 2010. Labeling and localization of the herpes simplex virus capsid protein UL25 and its interaction with the two triplexes closest to the penton. J Mol Biol 397:575-586. 36. Copeland, A. M., W. W. Newcomb, and J. C. Brown. 2009. Herpes simplex virus replication: roles of viral proteins and nucleoporins in capsid-nucleus attachment. J Virol 83:1660-1668. 37. Corey, L., and H. H. Handsfield. 2000. Genital herpes and public health: addressing a global problem. Jama 283:791-794. 38. Cotten M, a. W. J. 1995. The Adenovirus Protease is required for Virus Entry into Host Cells. Virology 213:494-505. 39. Davis, I. W., A. Leaver-Fay, V. B. Chen, J. N. Block, G. J. Kapral, X. Wang, L. W. Murray, W. B. Arendall, 3rd, J. Snoeyink, J. S. Richardson, and D. C. Richardson. 2007. MolProbity: all-atom contacts and structure validation for proteins and nucleic acids. Nucleic Acids Res 35:W375-383. 40. Davison, A. J. 2010. Herpesvirus systematics. Vet Microbiol 143:52-69. 41. Desai, P., G. L. Sexton, J. M. McCaffery, and S. Person. 2001. A null mutation in the gene encoding the herpes simplex virus type 1 UL37 polypeptide abrogates virus maturation. J Virol 75:10259-10271. 42. Desai, P. J. 2000. A null mutation in the UL36 gene of herpes simplex virus type 1 results in accumulation of unenveloped DNA-filled capsids in the cytoplasm of infected cells. J Virol 74:11608-11618. 43. Dong, G., A. H. Hutagalung, C. Fu, P. Novick, and K. M. Reinisch. 2005. The structures of exocyst subunit Exo70p and the Exo84p C-terminal domains reveal a common motif. Nature structural & molecular biology 12:1094-1100. 44. Emsley, P., and K. Cowtan. 2004. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60:2126-2132. 45. Famulski, J. K., L. Vos, X. Sun, and G. Chan. 2008. Stable hZW10 kinetochore residency, mediated by hZwint-1 interaction, is essential for the mitotic checkpoint. J Cell Biol 180:507-520. 46. Fatahzadeh, M., and R. A. Schwartz. 2007. Human herpes simplex virus infections: epidemiology, pathogenesis, symptomatology, diagnosis, and management. Journal of the American Academy of Dermatology 57:737-763; quiz 764-736.

205

47. Flint, S. J., L. W. Enquist, V. R. Racaniello, and A. M. Skalka. 2000. Principles of Virology: Molecular Biology, Pathogenesis, and Control. ASM Press, Washington, DC. 48. Fuchs, W., H. Granzow, B. G. Klupp, M. Kopp, and T. C. Mettenleiter. 2002. The UL48 tegument protein of pseudorabies virus is critical for intracytoplasmic assembly of infectious virions. J Virol 76:6729-6742. 49. Fuchs, W., B. G. Klupp, H. Granzow, and T. C. Mettenleiter. 2004. Essential function of the pseudorabies virus UL36 gene product is independent of its interaction with the UL37 protein. J Virol 78:11879-11889. 50. Gardino, A. K., and M. B. Yaffe. 2011. 14-3-3 proteins as signaling integration points for cell cycle control and apoptosis. Seminars in cell & developmental biology 22:688-695. 51. Garoff, H., Hewson R., and Opstelten DJ 1998. Virus Maturation by Budding. Microbiology and Molecular Biology Reviews 62. 52. Gouet, P., E. Courcelle, D. I. Stuart, and F. Metoz. 1999. ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics 15:305-308. 53. Granzow, H., B. G. Klupp, and T. C. Mettenleiter. 2005. Entry of pseudorabies virus: an immunogold-labeling study. J Virol 79:3200-3205. 54. Granzow, H., B. G. Klupp, and T. C. Mettenleiter. 2004. The pseudorabies virus US3 protein is a component of primary and of mature virions. J Virol 78:1314-1323. 55. Gregory, S. M., B. K. Davis, J. A. West, D. J. Taxman, S. Matsuzawa, J. C. Reed, J. P. Ting, and B. Damania. 2011. Discovery of a viral NLR homolog that inhibits the inflammasome. Science 331:330-334. 56. Grigoriev, I., D. Splinter, N. Keijzer, P. S. Wulf, J. Demmers, T. Ohtsuka, M. Modesti, I. V. Maly, F. Grosveld, C. C. Hoogenraad, and A. Akhmanova. 2007. Rab6 regulates transport and targeting of exocytotic carriers. Dev Cell 13:305-314. 57. Guo, H., S. Shen, L. Wang, and H. Deng. 2010. Role of tegument proteins in herpesvirus assembly and egress. Protein & cell 1:987-998. 58. Hales, C. M., J. P. Vaerman, and J. R. Goldenring. 2002. Rab11 family interacting protein 2 associates with Myosin Vb and regulates plasma membrane recycling. J Biol Chem 277:50415-50421. 59. Han, J., P. Chadha, D. G. Meckes, Jr., N. L. Baird, and J. W. Wills. 2011. Interaction and interdependent packaging of tegument protein UL11 and glycoprotein e of herpes simplex virus. Journal of virology 85:9437-9446. 60. Heckman, K. L., and L. R. Pease. 2007. Gene splicing and mutagenesis by PCR-driven overlap extension. Nat Protoc 2:924-932. 61. Helferich, D., J. Veits, T. C. Mettenleiter, and W. Fuchs. 2007. Identification of transcripts and protein products of the UL31, UL37, UL46, UL47, UL48, UL49 and US4 gene homologues of avian infectious laryngotracheitis virus. J Gen Virol 88:719-731. 62. Hollinshead, M., H. L. Johns, C. L. Sayers, C. Gonzalez-Lopez, G. L. Smith, and G. Elliott. 2012. Endocytic tubules regulated by Rab GTPases 5 and 11 are used for envelopment of herpes simplex virus. EMBO J 31:4204-4220.

206

63. Holm, L., and P. Rosenstrom. 2010. Dali server: conservation mapping in 3D. Nucleic acids research 38:W545-549. 64. Holm, L., and C. Sander. 1995. Dali: a network tool for protein structure comparison. Trends Biochem Sci 20:478-480. 65. Homa, F. L., J. B. Huffman, K. Toropova, H. R. Lopez, A. M. Makhov, and J. F. Conway. 2013. Structure of the pseudorabies virus capsid: comparison with herpes simplex virus type 1 and differential binding of essential minor proteins. J Mol Biol 425:3415-3428. 66. Hutagalung, A. H., and P. J. Novick. 2011. Role of Rab GTPases in membrane traffic and cell physiology. Physiological reviews 91:119-149. 67. Indran, S. V., and W. J. Britt. 2011. A role for the small GTPase Rab6 in assembly of human cytomegalovirus. J Virol 85:5213-5219. 68. Itzhaki, R. F., and M. A. Wozniak. 2008. Alzheimer's disease-like changes in herpes simplex virus type 1 infected cells: the case for antiviral therapy. Rejuvenation research 11:319-320. 69. Itzhaki, R. F., and M. A. Wozniak. 2008. Herpes simplex virus type 1 in Alzheimer's disease: the enemy within. J Alzheimers Dis 13:393-405. 70. Jackson, L. P., D. Kummel, K. M. Reinisch, and D. J. Owen. 2012. Structures and mechanisms of vesicle coat components and multisubunit tethering complexes. Curr Opin Cell Biol 24:475-483. 71. Jancarik, J., R. Pufan, C. Hong, S. H. Kim, and R. Kim. 2004. Optimum solubility (OS) screening: an efficient method to optimize buffer conditions for homogeneity and crystallization of proteins. Acta Crystallogr D Biol Crystallogr 60:1670-1673. 72. Jin, Y., A. Sultana, P. Gandhi, E. Franklin, S. Hamamoto, A. R. Khan, M. Munson, R. Schekman, and L. S. Weisman. 2011. Myosin V transports secretory vesicles via a Rab GTPase cascade and interaction with the exocyst complex. Dev Cell 21:1156-1170. 73. Johns, H. L., C. Gonzalez-Lopez, C. L. Sayers, M. Hollinshead, and G. Elliott. 2014. Rab6 dependent post-Golgi trafficking of HSV1 envelope proteins to sites of virus envelopment. Traffic 15:157-178. 74. Johnson, D. C., and J. D. Baines. 2011. Herpesviruses remodel host membranes for virus egress. Nature reviews. Microbiology 9:382-394. 75. Jose, J., L. Przybyla, T. J. Edwards, R. Perera, J. W. Burgner, 2nd, and R. J. Kuhn. 2012. Interactions of the cytoplasmic domain of Sindbis virus E2 with nucleocapsid cores promote alphavirus budding. J Virol 86:2585-2599. 76. Jovasevic, V., L. Liang, and B. Roizman. 2008. Proteolytic cleavage of VP1-2 is required for release of herpes simplex virus 1 DNA into the nucleus. J Virol 82:3311-3319. 77. Kelly, B. J., E. Diefenbach, C. Fraefel, and R. J. Diefenbach. 2012. Identification of host cell proteins which interact with herpes simplex virus type 1 tegument protein pUL37. Biochem Biophys Res Commun 417:961-965. 78. Kelly, B. J., C. Fraefel, A. L. Cunningham, and R. J. Diefenbach. 2009. Functional roles of the tegument proteins of herpes simplex virus type 1. Virus Res 145:173-186.

207

79. Klopfleisch, R., B. G. Klupp, W. Fuchs, M. Kopp, J. P. Teifke, and T. C. Mettenleiter. 2006. Influence of pseudorabies virus proteins on neuroinvasion and neurovirulence in mice. J Virol 80:5571-5576. 80. Klopfleisch, R., J. P. Teifke, W. Fuchs, M. Kopp, B. G. Klupp, and T. C. Mettenleiter. 2004. Influence of tegument proteins of pseudorabies virus on neuroinvasion and transneuronal spread in the nervous system of adult mice after intranasal inoculation. J Virol 78:2956-2966. 81. Klupp, B. G., W. Fuchs, H. Granzow, R. Nixdorf, and T. C. Mettenleiter. 2002. Pseudorabies virus UL36 tegument protein physically interacts with the UL37 protein. J Virol 76:3065-3071. 82. Klupp, B. G., H. Granzow, E. Mundt, and T. C. Mettenleiter. 2001. Pseudorabies virus UL37 gene product is involved in secondary envelopment. J Virol 75:8927-8936. 83. Ko, D. H., A. L. Cunningham, and R. J. Diefenbach. 2010. The major determinant for addition of tegument protein pUL48 (VP16) to capsids in herpes simplex virus type 1 is the presence of the major tegument protein pUL36 (VP1/2). J Virol 84:1397-1405. 84. Kramer, T., and L. W. Enquist. 2012. Alphaherpesvirus infection disrupts mitochondrial transport in neurons. Cell Host Microbe 11:504-514. 85. Kramer, T., T. M. Greco, L. W. Enquist, and I. M. Cristea. 2011. Proteomic characterization of pseudorabies virus extracellular virions. J Virol 85:6427-6441. 86. Krautwald, M., W. Fuchs, B. G. Klupp, and T. C. Mettenleiter. 2009. Translocation of incoming pseudorabies virus capsids to the cell nucleus is delayed in the absence of tegument protein pUL37. J Virol 83:3389-3396. 87. Krissinel, E., and K. Henrick. 2007. Inference of macromolecular assemblies from crystalline state. Journal of molecular biology 372:774-797. 88. Krissinel, E., and K. Henrick. 2007. Inference of macromolecular assemblies from crystalline state. J Mol Biol 372:774-797. 89. Lam, Q., C. A. Smibert, K. E. Koop, C. Lavery, J. P. Capone, S. P. Weinheimer, and J. R. Smiley. 1996. Herpes simplex virus VP16 rescues viral mRNA from destruction by the virion host shutoff function. EMBO J 15:2575- 2581. 90. Lama J, S. M., and Carrasco L. 1998. Genetic analysis of poliovirus protein 3A: characterization of a non-cytopathic mutant virus defective in killing Vero cells. J Gen Virol 79. 91. Langemeyer, L., and F. A. Barr. 2012. Analysis of rab GTPases. Current protocols in cell biology / editorial board, Juan S. Bonifacino ... [et al.] Chapter 15:Unit 15 18. 92. Larkin, M. A., G. Blackshields, N. P. Brown, R. Chenna, P. A. McGettigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson, and D. G. Higgins. 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23:2947-2948. 93. Lee, J. H., V. Vittone, E. Diefenbach, A. L. Cunningham, and R. J. Diefenbach. 2008. Identification of structural protein-protein interactions of herpes simplex virus type 1. Virology 378:347-354.

208

94. Lee, J. I., G. W. Luxton, and G. A. Smith. 2006. Identification of an essential domain in the herpesvirus VP1/2 tegument protein: the carboxy terminus directs incorporation into capsid assemblons. J Virol 80:12086-12094. 95. Lee, J. I., P. J. Sollars, S. B. Baver, G. E. Pickard, M. Leelawong, and G. A. Smith. 2009. A herpesvirus encoded deubiquitinase is a novel neuroinvasive determinant. PLoS pathogens 5:e1000387. 96. Leege, T., H. Granzow, W. Fuchs, B. G. Klupp, and T. C. Mettenleiter. 2009. Phenotypic similarities and differences between UL37-deleted pseudorabies virus and herpes simplex virus type 1. J Gen Virol 90:1560-1568. 97. Lenhoff, R. J., and J. Summers. 1994. Coordinate regulation of replication and virus assembly by the large envelope protein of an avian hepadnavirus. J Virol 68:4565-4571. 98. Lichtarge, O., H. R. Bourne, and F. E. Cohen. 1996. An evolutionary trace method defines binding surfaces common to protein families. J Mol Biol 257:342- 358. 99. Liu, X., K. Fitzgerald, E. Kurt-Jones, R. Finberg, and D. M. Knipe. 2008. Herpesvirus tegument protein activates NF-kappaB signaling through the TRAF6 adaptor protein. Proc Natl Acad Sci U S A 105:11335-11339. 100. Liu, Y., W. Gong, C. C. Huang, W. Herr, and X. Cheng. 1999. Crystal structure of the conserved core of the herpes simplex virus transcriptional regulatory protein VP16. Genes Dev 13:1692-1703. 101. Loomis, J. S., R. J. Courtney, and J. W. Wills. 2003. Binding partners for the UL11 tegument protein of herpes simplex virus type 1. Journal of virology 77:11417-11424. 102. Loret, S., G. Guay, and R. Lippe. 2008. Comprehensive characterization of extracellular herpes simplex virus type 1 virions. J Virol 82:8605-8618. 103. Luxton, G. W., S. Haverlock, K. E. Coller, S. E. Antinone, A. Pincetic, and G. A. Smith. 2005. Targeting of herpesvirus capsid transport in axons is coupled to association with specific sets of tegument proteins. Proc Natl Acad Sci U S A 102:5832-5837. 104. Luxton, G. W., J. I. Lee, S. Haverlock-Moyns, J. M. Schober, and G. A. Smith. 2006. The pseudorabies virus VP1/2 tegument protein is required for intracellular capsid transport. J Virol 80:201-209. 105. Machner, M. P., and R. R. Isberg. 2006. Targeting of host Rab GTPase function by the intravacuolar pathogen Legionella pneumophila. Dev Cell 11:47-56. 106. Matanis, T., A. Akhmanova, P. Wulf, E. Del Nery, T. Weide, T. Stepanova, N. Galjart, F. Grosveld, B. Goud, C. I. De Zeeuw, A. Barnekow, and C. C. Hoogenraad. 2002. Bicaudal-D regulates COPI-independent Golgi-ER transport by recruiting the dynein-dynactin motor complex. Nat Cell Biol 4:986-992. 107. McElwee, M., F. Beilstein, M. Labetoulle, F. J. Rixon, and D. Pasdeloup. 2013. Dystonin/BPAG1 promotes plus-end-directed transport of herpes simplex virus 1 capsids on microtubules during entry. J Virol 87:11008-11018. 108. McLauchlan, J. 1997. The abundance of the herpes simplex virus type 1 UL37 tegument protein in virus particles is closely controlled. J Gen Virol 78 ( Pt 1):189-194.

209

109. McNabb, D. S., and R. J. Courtney. 1992. Characterization of the large tegument protein (ICP1/2) of herpes simplex virus type 1. Virology 190:221-232. 110. Mebatsion, T., M. Konig, and C. K.K. 1996. Budding of Rabies virus particles in the absence of the spike glycoprotein. Cell 84. 111. Meckes, D. G., Jr., J. A. Marsh, and J. W. Wills. 2010. Complex mechanisms for the packaging of the UL16 tegument protein into herpes simplex virus. Virology 398:208-213. 112. Meckes, D. G., Jr., and J. W. Wills. 2007. Dynamic interactions of the UL16 tegument protein with the capsid of herpes simplex virus. Journal of virology 81:13028-13036. 113. Mettenleiter, T. C., B. G. Klupp, and H. Granzow. 2009. Herpesvirus assembly: an update. Virus Res 143:222-234. 114. Mijatov, B., A. L. Cunningham, and R. J. Diefenbach. 2007. Residues F593 and E596 of HSV-1 tegument protein pUL36 (VP1/2) mediate binding of tegument protein pUL37. Virology 368:26-31. 115. Morrison, E. E., Y. F. Wang, and D. M. Meredith. 1998. Phosphorylation of structural components promotes dissociation of the herpes simplex virus type 1 tegument. J Virol 72:7108-7114. 116. Neunuebel, M. R., S. Mohammadi, M. Jarnik, and M. P. Machner. 2012. Legionella pneumophila LidA affects nucleotide binding and activity of the host GTPase Rab1. J Bacteriol 194:1389-1400. 117. Ning, X., D. Nguyen, L. Mentzer, C. Adams, H. Lee, R. Ashley, S. Hafenstein, and J. Hu. 2011. Secretion of genome-free hepatitis B virus--single strand blocking model for virion morphogenesis of para-retrovirus. PLoS Pathog 7:e1002255. 118. Nufer, O., and H. P. Hauri. 2003. ER export: call 14-3-3. Curr Biol 13:R391- 393. 119. Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode, p. 307-326. In C. C. W. Jr and S. R. M. (ed.), Methods in Enzymology, vol. 276. Academic Press, New York, NY. 120. Overby, A. K., V. L. Popov, R. F. Pettersson, and E. P. Neve. 2007. The cytoplasmic tails of Uukuniemi Virus (Bunyaviridae) G(N) and G(C) glycoproteins are important for intracellular targeting and the budding of virus- like particles. J Virol 81:11381-11391. 121. Pasdeloup, D., F. Beilstein, A. P. Roberts, M. McElwee, D. McNab, and F. J. Rixon. 2010. Inner tegument protein pUL37 of herpes simplex virus type 1 is involved in directing capsids to the trans-Golgi network for envelopment. J Gen Virol 91:2145-2151. 122. Pasdeloup, D., D. Blondel, A. L. Isidro, and F. J. Rixon. 2009. Herpesvirus capsid association with the nuclear pore complex and viral DNA release involve the nucleoporin CAN/Nup214 and the capsid protein pUL25. J Virol 83:6610- 6623. 123. Pasdeloup, D., M. McElwee, F. Beilstein, M. Labetoulle, and F. J. Rixon. 2013. Herpesvirus tegument protein pUL37 interacts with dystonin/BPAG1 to promote capsid transport on microtubules during egress. J Virol 87:2857-2867.

210

124. Perez-Berna, A. J., W. F. Mangel, W. J. McGrath, V. Graziano, J. Flint, and C. San Martin. 2014. Processing of the l1 52/55k protein by the adenovirus protease: a new substrate and new insights into virion maturation. J Virol 88:1513-1524. 125. Phillips, K., and A. H. de la Pena. 2011. The combined use of the Thermofluor assay and ThermoQ analytical software for the determination of protein stability and buffer optimization as an aid in protein crystallization. Current protocols in molecular biology / edited by Frederick M. Ausubel ... [et al.] Chapter 10:Unit10 28. 126. Pincetic, A., G. Medina, C. Carter, and J. Leis. 2008. Avian sarcoma virus and human immunodeficiency virus, type 1 use different subsets of ESCRT proteins to facilitate the budding process. J Biol Chem 283:29822-29830. 127. Prange, R. 2012. Host factors involved in hepatitis B virus maturation, assembly, and egress. Medical microbiology and immunology 201:449-461. 128. Preston, V. G., J. Murray, C. M. Preston, I. M. McDougall, and N. D. Stow. 2008. The UL25 gene product of herpes simplex virus type 1 is involved in uncoating of the viral genome. J Virol 82:6654-6666. 129. Radtke, K., D. Kieneke, A. Wolfstein, K. Michael, W. Steffen, T. Scholz, A. Karger, and B. Sodeik. 2010. Plus- and minus-end directed microtubule motors bind simultaneously to herpes simplex virus capsids using different inner tegument structures. PLoS Pathog 6:e1000991. 130. Reynolds, A. E., B. J. Ryckman, J. D. Baines, Y. Zhou, L. Liang, and R. J. Roller. 2001. U(L)31 and U(L)34 proteins of herpes simplex virus type 1 form a complex that accumulates at the nuclear rim and is required for envelopment of nucleocapsids. J Virol 75:8803-8817. 131. Richardson, B. C., R. D. Smith, D. Ungar, A. Nakamura, P. D. Jeffrey, V. V. Lupashin, and F. M. Hughson. 2009. Structural basis for a human glycosylation disorder caused by mutation of the COG4 gene. Proceedings of the National Academy of Sciences of the United States of America 106:13329-13334. 132. Richardson, B. C., R. D. Smith, D. Ungar, A. Nakamura, P. D. Jeffrey, V. V. Lupashin, and F. M. Hughson. 2009. Structural basis for a human glycosylation disorder caused by mutation of the COG4 gene. Proc Natl Acad Sci U S A 106:13329-13334. 133. Roberts, A. P., F. Abaitua, P. O'Hare, D. McNab, F. J. Rixon, and D. Pasdeloup. 2009. Differing roles of inner tegument proteins pUL36 and pUL37 during entry of herpes simplex virus type 1. J Virol 83:105-116. 134. Roberts, L., R. Lamb, and R. W. Compans. 1998. The M1 and M2 proteins of Influenza A virus are important determinants in filamentous particle formation. Virology 240. 135. Roche, S., S. Bressanelli, F. A. Rey, and Y. Gaudin. 2006. Crystal structure of the low-pH form of the vesicular stomatitis virus glycoprotein G. Science 313:187-191. 136. Rojas, R., T. van Vlijmen, G. A. Mardones, Y. Prabhu, A. L. Rojas, S. Mohammed, A. J. Heck, G. Raposo, P. van der Sluijs, and J. S. Bonifacino. 2008. Regulation of retromer recruitment to endosomes by sequential action of Rab5 and Rab7. J Cell Biol 183:513-526.

211

137. Rossman, J. S., X. Jing, G. P. Leser, and R. A. Lamb. 2010. Influenza virus M2 protein mediates ESCRT-independent membrane scission. Cell 142:902-913. 138. Ruel, N., A. Zago, and P. G. Spear. 2006. Alanine substitution of conserved residues in the cytoplasmic tail of herpes simplex virus gB can enhance or abolish cell fusion activity and viral entry. Virology 346:229-237. 139. Ryan, S. D., K. Bhanot, A. Ferrier, Y. De Repentigny, A. Chu, A. Blais, and R. Kothary. 2012. Microtubule stability, Golgi organization, and transport flux require dystonin-a2-MAP1B interaction. J Cell Biol 196:727-742. 140. Sandbaumhuter, M., K. Dohner, J. Schipke, A. Binz, A. Pohlmann, B. Sodeik, and R. Bauerfeind. 2013. Cytosolic herpes simplex virus capsids not only require binding inner tegument protein pUL36 but also pUL37 for active transport prior to secondary envelopment. Cell Microbiol 15:248-269. 141. Schipke, J., A. Pohlmann, R. Diestel, A. Binz, K. Rudolph, C. H. Nagel, R. Bauerfeind, and B. Sodeik. 2012. The C terminus of the large tegument protein pUL36 contains multiple capsid binding sites that function differently during assembly and cell entry of herpes simplex virus. J Virol 86:3682-3700. 142. Schlieker, C., G. A. Korbel, L. M. Kattenhorn, and H. L. Ploegh. 2005. A deubiquitinating activity is conserved in the large tegument protein of the herpesviridae. J Virol 79:15582-15585. 143. Schlieker, C., W. A. Weihofen, E. Frijns, L. M. Kattenhorn, R. Gaudet, and H. L. Ploegh. 2007. Structure of a herpesvirus-encoded cysteine protease reveals a unique class of deubiquitinating enzymes. Mol Cell 25:677-687. 144. Schmidt, N. W., A. Mishra, J. Wang, W. F. DeGrado, and G. C. Wong. 2013. Influenza virus A M2 protein generates negative Gaussian membrane curvature necessary for budding and scission. J Am Chem Soc 135:13710-13719. 145. Schmitt, H. D. 2010. Dsl1p/Zw10: common mechanisms behind tethering vesicles and microtubules. Trends Cell Biol 20:257-268. 146. Schmitt, H. D. 2010. Dsl1p/Zw10: common mechanisms behind tethering vesicles and microtubules. Trends in cell biology 20:257-268. 147. Schmitz, J. B., A. G. Albright, P. R. Kinchington, and F. J. Jenkins. 1995. The UL37 protein of herpes simplex virus type 1 is associated with the tegument of purified virions. Virology 206:1055-1065. 148. Schoebel, S., A. L. Cichy, R. S. Goody, and A. Itzen. 2011. Protein LidA from Legionella is a Rab GTPase supereffector. Proc Natl Acad Sci U S A 108:17945- 17950. 149. Shanda, S. K., and D. W. Wilson. 2008. UL36p is required for efficient transport of membrane-associated herpes simplex virus type 1 along microtubules. J Virol 82:7388-7394. 150. Shelton, L. S., A. G. Albright, W. T. Ruyechan, and F. J. Jenkins. 1994. Retention of the herpes simplex virus type 1 (HSV-1) UL37 protein on single- stranded DNA columns requires the HSV-1 ICP8 protein. J Virol 68:521-525. 151. Silverman, J. L., N. G. Greene, D. S. King, and E. E. Heldwein. 2012. Membrane Requirement for Folding of the Herpes Simplex Virus 1 gB Cytodomain Suggests a Unique Mechanism of Fusion Regulation. Journal of virology 86:8171-8184.

212

152. Sivaram, M. V., M. L. Furgason, D. N. Brewer, and M. Munson. 2006. The structure of the exocyst subunit Sec6p defines a conserved architecture with diverse roles. Nature structural & molecular biology 13:555-556. 153. Smith, G. L., and A. Vanderplasschen. 1998. Extracellular enveloped vaccinia virus. Entry, egress, and evasion. Adv Exp Med Biol 440:395-414. 154. Smith, J. S., and N. J. Robinson. 2002. Age-specific prevalence of infection with herpes simplex virus types 2 and 1: a global review. The Journal of infectious diseases 186 Suppl 1:S3-28. 155. Sowa, M. E., W. He, K. C. Slep, M. A. Kercher, O. Lichtarge, and T. G. Wensel. 2001. Prediction and confirmation of a site critical for effector regulation of RGS domain activity. Nat Struct Biol 8:234-237. 156. Spang, A. 2012. The DSL1 complex: the smallest but not the least CATCHR. Traffic 13:908-913. 157. Stampfer, S. D., H. Lou, G. H. Cohen, R. J. Eisenberg, and E. E. Heldwein. 2010. Structural basis of local, pH-dependent conformational changes in glycoprotein B from herpes simplex virus type 1. Journal of virology 84:12924- 12933. 158. Starkey, J. L., J. Han, P. Chadha, J. A. Marsh, and J. W. Wills. 2014. Elucidation of the block to herpes simplex virus egress in the absence of tegument protein UL16 reveals a novel interaction with VP22. J Virol 88:110-119. 159. Stern, S., M. Tanaka, and W. Herr. 1989. The oct-1 homeodomain directs formation of a multiprotein-DNA complex with the HSV transactivator Vp16. Nature 341:624-. 160. Stow, N. D. 2001. Packaging of genomic and amplicon DNA by the herpes simplex virus type 1 UL25-null mutant KUL25NS. J Virol 75:10755-10765. 161. Szajner, P., A. S. Weisberg, J. Lebowitz, J. Heuser, and B. Moss. 2005. External scaffold of spherical immature poxvirus particles is made of protein trimers, forming a honeycomb lattice. J Cell Biol 170:971-981. 162. Takakuwa, H., F. Goshima, T. Koshizuka, T. Murata, T. Daikoku, and Y. Nishiyama. 2001. Herpes simplex virus encodes a virion-associated protein which promotes long cellular processes in over-expressing cells. Genes Cells 6:955-966. 163. Thurlow, J. K., M. Murphy, N. D. Stow, and V. G. Preston. 2006. Herpes simplex virus type 1 DNA-packaging protein UL17 is required for efficient binding of UL25 to capsids. J Virol 80:2118-2126. 164. Toropova, K., J. B. Huffman, F. L. Homa, and J. F. Conway. 2011. The herpes simplex virus 1 UL17 protein is the second constituent of the capsid vertex-specific component required for DNA packaging and retention. J Virol 85:7513-7522. 165. Tripathi, A., Y. Ren, P. D. Jeffrey, and F. M. Hughson. 2009. Structural characterization of Tip20p and Dsl1p, subunits of the Dsl1p vesicle tethering complex. Nature structural & molecular biology 16:114-123. 166. Umbrecht-Jenck, E., V. Demais, V. Calco, Y. Bailly, M. F. Bader, and S. Chasserot-Golaz. 2010. S100A10-mediated translocation of annexin-A2 to SNARE proteins in adrenergic chromaffin cells undergoing exocytosis. Traffic 11:958-971.

213

167. Vennema, H., G. J. Godeke, J. W. Rossen, W. F. Voorhout, M. C. Horzinek, D. J. Opstelten, and P. J. Rottier. 1996. Nucleocapsid-independent assembly of coronavirus-like particles by co-expression of viral envelope protein genes. EMBO J 15. 168. Vidick, S., B. Leroy, L. Palmeira, B. Machiels, J. Mast, S. Francois, R. Wattiez, A. Vanderplasschen, and L. Gillet. 2013. Proteomic characterization of murid herpesvirus 4 extracellular virions. PLoS ONE 8:e83842. 169. Vittone, V., E. Diefenbach, D. Triffett, M. W. Douglas, A. L. Cunningham, and R. J. Diefenbach. 2005. Determination of interactions between tegument proteins of herpes simplex virus type 1. J Virol 79:9566-9571. 170. Walden, H. 2010. Selenium incorporation using recombinant techniques. Acta Crystallogr D Biol Crystallogr 66:352-357. 171. Wang, S., K. Wang, J. Li, and C. Zheng. 2013. Herpes simplex virus 1 ubiquitin-specific protease UL36 inhibits beta interferon production by deubiquitinating TRAF3. J Virol 87:11851-11860. 172. Watanabe, D., Y. Ushijima, F. Goshima, H. Takakuwa, Y. Tomita, and Y. Nishiyama. 2000. Identification of nuclear export signal in UL37 protein of herpes simplex virus type 2. Biochem Biophys Res Commun 276:1248-1254. 173. Wisner, T. W., C. C. Wright, A. Kato, Y. Kawaguchi, F. Mou, J. D. Baines, R. J. Roller, and D. C. Johnson. 2009. Herpesvirus gB-induced fusion between the virion envelope and outer nuclear membrane during virus egress is regulated by the viral US3 kinase. J Virol 83:3115-3126. 174. Wu, S., S. Q. Mehta, F. Pichaud, H. J. Bellen, and F. A. Quiocho. 2005. Sec15 interacts with Rab11 via a novel domain and affects Rab11 localization in vivo. Nat Struct Mol Biol 12:879-885. 175. Xing, J., S. Wang, Y. Li, H. Guo, L. Zhao, W. Pan, F. Lin, H. Zhu, L. Wang, M. Li, L. Wang, and C. Zheng. 2011. Characterization of the subcellular localization of herpes simplex virus type 1 proteins in living cells. Medical microbiology and immunology 200:61-68. 176. Yang, K., and J. D. Baines. 2011. Selection of HSV capsids for envelopment involves interaction between capsid surface components pUL31, pUL17, and pUL25. Proc Natl Acad Sci U S A 108:14276-14281. 177. Yang, K., E. Wills, H. Y. Lim, Z. H. Zhou, and J. D. Baines. 2014. Association of Herpes Simplex Virus pUL31 with Capsid Vertices and Components of the Capsid Vertex Specific Complex. J Virol. 178. Yeh, P. C., J. Han, P. Chadha, D. G. Meckes, Jr., M. D. Ward, O. J. Semmes, and J. W. Wills. 2011. Direct and specific binding of the UL16 tegument protein of herpes simplex virus to the cytoplasmic tail of glycoprotein E. Journal of virology 85:9425-9436. 179. Zaichick, S. V., K. P. Bohannon, A. Hughes, P. J. Sollars, G. E. Pickard, and G. A. Smith. 2013. The herpesvirus VP1/2 protein is an effector of dynein- mediated capsid transport and neuroinvasion. Cell Host Microbe 13:193-203. 180. Zhou, Q., Y. S. Kee, C. C. Poirier, C. Jelinek, J. Osborne, S. Divi, A. Surcel, M. E. Will, U. S. Eggert, A. Muller-Taubenberger, P. A. Iglesias, R. J. Cotter, and D. N. Robinson. 2010. 14-3-3 coordinates microtubules, Rac, and myosin II to control cell mechanics and cytokinesis. Curr Biol 20:1881-1889.

214

215