STUDIES ON THE ASSEMBLY AND MORPHOLOGY OF

HUMAN T-CELL LEUKEMIA VIRUS TYPE 1

A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE

SCHOOL OF THE UNIVERSITY OF MINNESOTA

José Orlando Maldonado-Ortiz

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Advisor: Louis M. Mansky, PhD

August 2018

© José Maldonado, 2018 ACKNOWLEDGEMENTS

I would like to thank:

My wife, Lorie, whom since high school has been on my side, supporting every decision I have made, motivating me to continue regardless on the lengthiness of my studies, and for helping me stay sane during my graduate studies.

My family, whom have supported and motivated me during my 25+ years as a student; they are always there to assist me whenever I need them.

My friends, which in one-way or the other have provided support and friendship.

My advisor, Dr. Louis Mansky, for always being supportive and providing me with knowledge, freedom, and the necessary equipment to complete my projects. I am also thankful to Lou for encouraging me to extend my graduate studies for another year, which was essential for me to better train as an independent investigator.

My PhD dissertation committee members Drs. Paul Jardine, Peter Southern, and Wei Zhang, for their continued support, insightful discussions, and scientific and career advice.

All members of the Mansky laboratory for their continued support and scientific discussions.

My partial funding from the Institute for Molecular Virology Training Program (T32 AI083196) and my ongoing individual fellowship from the National Institute of Dental and Craniofacial Research (F30 DE22286).

i DEDICATION

I dedicate this thesis to my family and my wife, Lorie, for their constant support and unconditional love

ii ABSTRACT

The group-specific antigen (Gag) polyprotein is an essential retrovirus structural protein required for the assembly and release of virus particles. Present knowledge of

Gag biology has been limited to a few retroviruses. Furthermore, current understanding of the diversity in the nature of Gag structure and function in virus particle assembly is limited. Human T-cell leukemia virus type 1 (HTLV-1) is a deltaretrovirus that causes an adult T-cell leukemia/lymphoma (ATLL), HTLV-1-associated-myelopathy/tropical spastic paraparesis (HAM/TSP), and other neurotropic conditions. HTLV-1 has infected approximately 15 million individuals worldwide. A general knowledge gap exists regarding the details of HTLV-1 replication, including particle assembly. To address this, and to test the overarching hypothesis that HTLV-1 particle assembly is distinct from that of other retroviruses, this dissertation focused on investigating three key aspects of

HTLV-1 immature and mature particle morphology.

First, an analysis of the morphology and Gag stoichiometry of HTLV-1-like particles and authentic, mature HTLV-1 particles by using cryogenic transmission electron microscopy (cryo-TEM) and scanning transmission electron microscopy

(STEM) was conducted. HTLV-1-like particles mimicked the morphology of immature authentic HTLV-1 virions. Importantly, it was observed for the first time that the morphology of these virus-like particles (VLPs) has the unique local feature of a flat Gag lattice that does not follow the curvature of the viral membrane, resulting in an enlarged distance between the Gag lattice and the viral membrane. Measurement of the average size and mass of VLPs and authentic HTLV-1 particles suggested a consistent range of size and Gag copy numbers in these two groups of particles. The unique local flat Gag

iii lattice morphological feature observed suggests that HTLV-1 Gag could be arranged in a lattice structure that is distinct from that of other retroviruses characterized to date.

Second, the effects of Gag proteins labeled at the carboxy terminus with a fluorophore protein were analyzed for their influence on particle morphology. In particular, a HTLV-1 Gag expression construct with the yellow fluorescence protein

(YFP) fused to the carboxy-terminus was used as a surrogate for the HTLV-1 Gag-Pro to assess the effects of co-packaging of Gag and a Gag-YFP on virus-like particle morphology and particles were analyzed by cryo-TEM. STEM and fluorescence fluctuation spectroscopy (FFS) were also used to determine the Gag stoichiometry. Ratios of 3:1 (Gag:Gag-YFP) or greater were found to result in a particle morphology indistinguishable from that of VLPs produced with the untagged HTLV-1 Gag, i.e., a mean diameter of ~113 nm and a mass of 220 MDa as determined by cryo-TEM and

STEM, respectively. This information is useful for the quantitative analysis of Gag-Gag interactions that occur during virus particle assembly and in released immature particles.

Third, cryo-electron tomography (cryo-ET) was used to analyze mature HTLV-1 particle morphology. Particles produced from MT-2 cells were polymorphic, roughly spherical, and varied in size. Capsid cores, when present, were typically poorly defined polyhedral structures with at least one curved region contacting the inner face of the viral membrane. Most of the particles observed lacked a defined capsid core, which likely impacts HTLV-1 particle infectivity. Taken together, the findings of this dissertation provide new insights into the nature of immature and mature HTLV-1 assembly and morphology and provide foundational knowledge towards an advanced understanding of the HTLV particle assembly pathway.

iv

TABLE OF CONTENTS

PAGE

List of tables ix

List of figures x

List of publications xii

CHAPTER I: GENERAL INTRODUCTION 1

History and identification of human T-cell leukemia virus type 1 2

HTLV-1 clinical association 2

HTLV-1 head and neck clinical association 4

HTLV-1 genome organization 5

HTLV-1 infectious replication cycle 6

Attachment and fusion 6

Reverse transcription, nuclear transport and integration 7

Viral gene transcription 8

Post-transcriptional regulation 8

Viral protein translation 9

Gag trafficking 9

Gag and viral RNA trafficking 10

Budding and maturation 10

Gag-membrane interactions 11

Identification of Gag assembly sites along the inner leaflet of the plasma membrane 12

v Determinants of Gag-Gag interactions 13

Gag oligomerization 14

HTLV-1 transmission 15

Inter-host transmission 15

Cell-to-cell transmission 16

Virological synapses 17

Viral biofilms 18

Monoclonal expansion of HTLV-1 infected cells & leukemogenesis 18

Dissertation objectives 19

Figures 22

CHAPTER II: DISTINCT MORPHOLOGY OF HUMAN T-CELL LEUKEMIA VIRUS TYPE 1-LIKE PARTICLES 27

Introduction 28

Material and methods 30

Transfection and HTLV-1-like particle production 30

Gradient purification of authentic virus particles and VLPs 31

Cryo-TEM of HTLV-1-like particles and authentic virus particles 32

Determination of particle size 32

Determination of particle mass by STEM 33

Results 34

Analysis of the morphology of HTLV-1-like particles 34

Morphology of authentic HTLV-1 mature particles produced from MT-2 cells 35

STEM analyses of HTLV-1-like particles and authentic mature HTLV-1

vi Particles 36

Calculation of Gag stoichiometry in HTLV-1-like particles 37

Estimating Gag stoichiometry in authentic immature HTLV-1 particles by calculating Gag copy number in authentic mature HTLV-1 particles 38

Discussion 39

Figures 42

CHAPTER III: PERTURBATION OF HUMAN T-CELL LEUKEMIA VIRUS TYPE 1 PARTICLE MORPHOLOGY BY DIFFERENTIAL GAG CO-PACKAGING 48

Introduction 49

Materials and methods 52

Production and purification of HTLV-1-like particles 52

Cryo-TEM analysis of HTLV-1-like particles 53

Measurement of virus-like particle size 53

Fluorescence fluctuation spectroscopy, experimental setup and data analysis 54

Scanning transmission electron microscopy mass measurements of virus-like particles 56

Protein and RNA content of HTLV-1 virus-like particles 57

Results 58

Morphology of HTLV-1-like particles 58

Gag stoichiometry of HTLV-1-like particles determined by FFS 59

VLP diameter as determined by FFS 60

STEM analysis of HTLV-1-like particle mass 61

STEM determination of Gag stoichiometry in HTLV-1-like particles 62

Discussion 63

vii Conclusions 65

Figures 67

CHAPTER IV: ANALYSIS OF HUMAN T-CELL LEUKEMIA VIRUS TYPE 1 PARTICLES USING CRYO-ELECTRON TOMOGRAPHY 78

Text 79

Figures 88

CHAPTER V: DISSERTATION SUMMARY AND FINAL DISCUSSION 92

BIBLIOGRAPHY 101

APPENDIX I: DECLARATION OF CONTRIBUTIONS TO CO- AUTHORED PUBLICATIONS: DISTINCT PARTICLE MORPHOLOGIES REVEALED THROUGH COMPARATIVE PARALLEL ANALYSES OF RETROVIRUS-LIKE PARTICLES 129

APPENDIX II: DECLARATION OF CONTRIBUTIONS TO CO-AUTHORED PUBLICATIONS: DUAL ANTI-HIV MECHANISM OF CLOFARABINE 149

APPENDIX III: DECLARATION OF CONTRIBUTIONS TO CO- AUTHORED PUBLICATIONS: THE HIV-1 REVERSE TRANSCRIPTASE A62V MUTATION INFLUENCES REPLICATION FIDELITY AND VIRAL FITNESS IN THE CONTEXT OF MULTI-DRUG RESISTANCE MUTATIONS 168

APPENDIX IV: COPYRIGHT PERMISSIONS 183

viii LIST OF TABLES

CHAPTER II PAGE

Table 2-1 Summary of the mass determinations and the calculated Gag copy number per particle in human T-cell leukemia virus type 1(HTLV-1)- like particles and authentic HTLV-1 particles 47

CHAPTER III

Table 3-1 Summary of HTLV-1-like particle mass determination and calculation of Gag copy number per particle 76

CHAPTER V

Table 5-1 Summary of authentic HTLV-1 particle, HTLV-1-like particle, and HTLV-1-like particle expressing Gag/Gag-YFP at a 3 to 1 ratio particle size, mass determination and calculation of Gag copy number per particle 100

ix LIST OF FIGURES

CHAPTER I PAGE

Figure 1-1 HTLV-1 genome organization 22

Figure 1-2 HTLV-1 life cycle 23

Figure 1-3 Gag and retrovirus particle assembly 25

Figure 1-4 Schematic representation of HTLV-1 and HIV-1 Gag-membrane association 26

CHAPTER II

Figure 2-1 Analysis of the diameter and morphology of human T-cell leukemia virus type 1 (HTLV-1) virus-like particles (VLPs) by transmission electron microscopy (TEM) 42

Figure 2-2 Cryogenic transmission electron microscopy (Cryo-TEM) images of HTLV-1-like particles and comparison of Gag lattice between HTLV-1 and human immunodeficiency virus type 1 (HIV-1) 43

Figure 2-3 Analysis of the diameter of authentic mature HTLV-1 virus particles 44

Figure 2-4 Scanning transmission electron microscopy (STEM) analysis of HTLV-1-like particles 45

Figure 2-5 STEM analysis of authentic mature HTLV-1 virus particles 46

CHAPTER III

Figure 3-1 Calibration of YFP molecular brightness for measurements of Gag copy number by fluorescence fluctuation spectroscopy 67

Figure 3-2 Images of HTLV-1-like particles as determined by fluorescence microscopy and cryogenic transmission electron microscopy 68

Figure 3-3 Analysis of HTLV-1-like particle diameter by cryogenic transmission electron microscopy 69

Figure 3-4 Analysis of HTLV-1-like particles by fluorescence fluctuation spectroscopy 70

Figure 3-5 Determination of HTLV-1-like particle Gag copy number and particle diameter 72

x

Figure 3-6 Scanning transmission electron microscopy analysis of virus-like particles produced from expression of HTLV-1 Gag-YFP and Gag:Gag-YFP (3:1) 73

Figure 3-7 Relationship between virus-like particle size and Gag copy number to that of HTLV-1 Gag-YFP copy number 75

CHAPTER IV

Figure 4-1 Morphology of HTLV-1 particles as determined by cryo-ET 88

Figure 4-2 Tomographic slices of core-containing HTLV-1 particles 90

Figure 4-3 Turret structures on HTLV-1 particles 91

xi LIST OF PUBLICATIONS

1. Maldonado, J. O., Martin, J. L., Mueller, J. D., Zhang, W., & Mansky, L. M. (2014). New insights into retroviral Gag-Gag and Gag-membrane interactions. Frontiers in Microbiology, 5, 302. doi: 10.3389/fmicb.2014.00302

2. Cao, S., Maldonado, J. O., Grigsby, I. F., Mansky, L. M., & Zhang, W. (2015). Analysis of human T-cell leukemia virus type 1 particles by using cryo-electron tomography. Journal of Virology, 89(4), 2430-2435. doi: 10.1128/JVI.02358-14

3. Daly, M. B., Roth, M. E., Bonnac, L., Maldonado, J. O., Xie, J., Clouser, C. L., Mansky, L. M. (2016). Dual anti-HIV mechanism of clofarabine. Retrovirology, 13, 20. doi: 10.1186/s12977-016-0254-0

4. Maldonado, J. O., Cao, S., Zhang, W., & Mansky, L. M. (2016). Distinct Morphology of Human T-Cell Leukemia Virus Type 1-Like Particles. Viruses, 8(5). doi: 10.3390/v8050132

5. Martin, J. L., Cao, S., Maldonado, J. O., Zhang, W., & Mansky, L. M. (2016). Distinct Particle Morphologies Revealed through Comparative Parallel Analyses of Retrovirus-Like Particles. Journal of Virology, 90(18), 8074-8084. doi: 10.1128/JVI.00666-16

6. Martin, J. L., Maldonado, J. O., Mueller, J. D., Zhang, W., & Mansky, L. M. (2016). Molecular Studies of HTLV-1 Replication: An Update. Viruses, 8(2). doi: 10.3390/v8020031

7. Maldonado, J. O., Angert, I., Cao, S., Berk, S., Zhang, W., Mueller, J. D., & Mansky, L. M. (2017). Perturbation of Human T-Cell Leukemia Virus Type 1 Particle Morphology by Differential Gag Co-Packaging. Viruses, 9(7), 191. doi: 10.3390/v9070191

8. Maldonado, J. O., & Mansky, L. M. (2018). The HIV-1 Reverse Transcriptase A62V Mutation Influences Replication Fidelity and Viral Fitness in the Context of Multi-Drug-Resistant Mutations. Viruses, 10(7), 376. doi: 10.3390/v10070376

xii

CHAPTER I

GENERAL INTRODUCTION

Portions of the text presented in this chapter were previously published:

Maldonado, J. O., Martin, J. L., Mueller, J. D., Zhang, W., & Mansky, L. M. (2014). New insights into retroviral Gag-Gag and Gag-membrane interactions. Frontiers in Microbiology, 5, 302. doi: 10.3389/fmicb.2014.00302

Martin, J. L., Maldonado, J. O., Mueller, J. D., Zhang, W., & Mansky, L. M. (2016). Molecular Studies of HTLV-1 Replication: An Update. Viruses, 8(2). doi: 10.3390/v8020031

1 1.1 History and Identification of human T-cell leukemia virus type 1

Human T-cell leukemia virus type 1 (HTLV-1) was independently discovered in the late-1970s, early-1980s, by two research groups and identified as the etiological agent of adult T-cell leukemia/lymphoma (ATLL) [1, 2]. As the first human retrovirus discovered, research on HTLV-1 laid the foundational framework for subsequent studies of human immunodeficiency virus type 1 (HIV-1), infectious causes of cancer, and the molecular mechanisms of leukemogenesis [3].

Shortly after the discovery of HTLV-1, another human retrovirus was discovered

– human T-cell leukemia virus type 2 (HTLV-2) – which closely resembled HTLV-1 in genome structure and nucleotide sequence [4]. Unlike HTLV-1, HTLV-2 has not been convincingly associated with human pathology. Nevertheless, both HTLV-1 and -2 are included in worldwide prevalence estimates. Historically, it has been estimated that 15-20 million people are infected worldwide [5, 6]. A more recent study has estimated the number closer to 5-10 million, with most these individuals residing in Southwestern

Japan, Central Africa, Australia, and the Caribbean Basin [7]. A third and fourth type of

HTLV, human T-cell leukemia virus type 3 (HTLV-3) and human T-cell leukemia virus type 4 (HTLV-4), have been discovered in central Africa in the past decade; both are closely related to HTLV-1, and likely share similarities in replication, pathogenesis and transmission [8, 9].

1.2 HTLV-1 Clinical Associations

HTLV-1 is the etiological agent of ATLL as well as a variety of neurological pathologies, primarily HTLV-1-associated-myelopathy/tropical spastic paraparesis

2 (HAM/TSP) [10]. Both ATLL and HAM/TSP have a low incidence among HTLV-1 carriers. It is thought that approximately 2-6% of patients infected with HTLV-1 will acquire either pathology [11, 12]. The mechanisms of HTLV-1 pathogenesis were recently reviewed [13]. Briefly, ATLL generally presents after a long latency in patients infected during childhood. This is in contrast to HAM/TSP, which is associated with infection later in life [14]. ATLL is an aggressive malignancy of the peripheral T-cells and can be divided into four subtypes – acute, lymphomatous, chronic, or smoldering.

Patients with the acute form of ATLL have a prognosis of approximately 6 months – an estimate that has not significantly changed since the discovery of the disease, despite advances in treatments [15]. Current recommended therapies for ATLL include chemotherapy, monoclonal antibodies, allogeneic bone marrow transplants, and a combination of interferon-α (IFN- α) and azidothymidine (AZT) [16-19]. Interestingly, the mechanism of action of the combination of IFN- α and AZT appears to correlate with an induction of cell apoptosis by phosphorylation of p53 [20].

HAM/TSP is characterized by spasticity and weakness of the legs along with urinary disturbances [20]. The primary pathology of HAM/TSP is associated with HTLV-

1 infection in the spinal cord leading to inflammation. Unlike ATLL, which appears to have a complex and multi-faceted pathology, the incidence of HAM/TSP has been shown to correlate with HTLV-1 proviral loads as well as the site of proviral integration [21,

22]. Treatment of HAM/TSP is symptom-based and includes antispasmodic and anti- inflammatory medications [23].

3 1.3 HTLV-1 Head and Neck Clinical Association

HTLV-1 infected individuals may also develop Sjögren’s syndrome (SS)-like symptoms [24]. SS is a disorder of the immune system where an individual’s white blood cells attack and destroy salivary or lacrimal glands, or both, resulting in xerostomia or keratoconjunctivitis sicca, respectively. The most common complications of SS are yeast infections, dental cavities, and vision complications. This autoimmune disease is regularly seen together with other immune system disorders, such as scleroderma, lupus, and rheumatoid arthritis; however, SS is also seen alone. SS is estimated to affect over 30 million people worldwide, with about half of these individuals having only SS and the other half being diagnosed with SS associated with immune system disorders [25, 26].

Currently there is no cure for SS, however the use of artificial tears and saliva, nasal spray, increase of fluid intake as well as saliva flow stimulation may treat the symptoms.

SS is caused by a combination of genetic and environmental factors such as viral infections. Human viruses such as coxsackie virus, Epstein-Barr virus, hepatitis C virus, hepatitis D virus, HIV-1, and HTLV-1 are thought to play an important role in SS pathogenesis by triggering an autoimmune response, but a specific etiology has not been identified. [24, 27-30]. HTLV-1 seroprevalence rates in SS patients from HTLV-1 highly endemic regions are about 23% compared to only 3% of the individuals from non- endemic regions [31]. SS patients from HTLV-1 endemic regions have high prevalence of IgA antibodies to HTLV-1, which may be due to increased viral activity in the salivary glands [28, 31]. It is believed that HTLV-1 may play a role in SS pathogenesis by activating an immune response through anti-SS-related antigen A antibody production

[32]. Furthermore, HAM patients with a high prevalence of SS, may show a high number

4 of infiltrated peripheral blood mononuclear cells, mostly CD4+, CD45RO+, CD25-,

CD69, Bcl-2+, and memory T-cells, into salivary glands which may increase cytokine production, augmentation of cell proliferation, cell adhesion, inflammation, migration, and apoptosis [33-35]. It has been suggested that HTLV-1 is tropic for salivary and lachrymal glands ductal epithelium [36]. However, it is still not clear what the role of

HTLV-1 and SS is in HTLV-1 non-endemic areas, as well as the pathways used by

HTLV-1 in SS.

1.4 HTLV-1 Genome Organization

HTLV-1 viral particles contain two linear, positive sense, single-stranded RNA

(vRNA) copies of about 9 kb in size [37]. Figure 1-1 depicts the HTLV-1 genomic organization. The HTLV-1 proviral sequence contains 5’ and 3’ long terminal repeats

(LTRs) at both ends. The LTRs contain regulatory elements within the unique 3’ region

(U3), repeat region (R), and the unique 5’ region (U5). [37, 38] The 5’ end contains a primer-binding site, while the 3’ end contains a polypurine tract. Both of these cis-acting elements are important during reverse transcription of the vRNA into double-stranded

DNA (dsDNA). The integration of the dsDNA into the host cell chromosome forms the provirus. Like all retroviruses, HTLV-1 encodes structural (gag), enzymatic (pro-pol), and envelope (env) genes, which are essential for retroviral replication. [39-41] The Gag polyprotein contains matrix (MA), capsid (CA), and nucleocapsid (NC) domains. The pro gene encodes for the viral protease (PR), the pol gene encodes for reverse transcriptase

(RT) and for integrase (IN), while the env gene encodes the envelope (Env) glycoprotein, which contains the transmembrane (TM) and surface (SU) domains. [39, 40] HTLV-1

5 also encodes for the pX region in the 3’ portion of the genome that contains regulatory and accessory genes (tax, rex, pX-I, pX-II) as well as an antisense gene, which encodes a basic leucine zipper factor (hbz). The four open reading frames (ORFs) tax, rex, pX-I, pX-

II, as well as antisense hbz ORF, code for the Tax, Rex, p12, p27, p13, p30 and HBZ proteins, respectively. [42-45] These proteins play an important role in viral replication such as transcription initiation, post-transcriptional regulation and modification, infectivity, and persistence [13].

1.5 HTLV-1 Infectious Replication Cycle

1.5.1 Attachment and fusion

HTLV-1 primarily infects CD4+ T-cells, but has the potential to infect a wide variety of cells – including CD8+ T-cells, B-lymphocytes, endothelial cells, myeloid cells, fibroblasts, as well as other mammalian cells [46-50]. This wide variety of target cells is due in part to the ability of the SU protein of the HTLV-1 Env glycoprotein to interact with three widely distributed cellular surface receptors including the glucose transporter (GLUT1) [51], heparin sulfate proteoglycan (HSPG) [52], and the VEGF-165 receptor neuropilin-1 (NRP-1) [53]. Once HTLV-1 has attached to the cell, the membrane fusion process occurs by a series of proposed sequential events between SU and the target cell receptor proteins (Figure 1-2A-B) [53, 54]. Briefly, HTLV-1 Env initially interacts with HSPG, then with NRP-1, which results in the formation of a complex. Following this event, GLUT1 associates with the HSPG/NRP-1 complex to initiate the fusion process, through interactions with the HTLV-1 Env TM protein, allowing for the HTLV-

6 1 CA core containing the viral genome and viral proteins to be released into the cytoplasm of the permissive target cell (Figure 1-2B).

1.5.2 Reverse transcription, nuclear transport and integration

The HTLV-1 CA core contains 2 copies of the vRNA along with RT, IN, and the viral PR. Reverse transcription of HTLV-1 vRNA to dsDNA is thought to primarily occur after virus entry into target cell (Figure 1-2C) [55, 56]. For HIV-1, it is thought that reverse transcription is linked to intracellular uncoating of the CA core [57].

Additionally, HIV-1 RT and IN interactions have been shown to be necessary for production of early reverse transcription products [58]. Complementary studies with

HTLV-1 or other retroviruses have not been conducted to date. Recombination can occur during reverse transcription, and recent evidence from phylogenetic analyses strongly suggests that recombination has played an important role in the emergence of HTLV-1 in the human population approximately 4,000 years ago [59]. No studies, to date, have investigated the nature of HTLV-1 recombination. For example, recombination in HIV-1 occurs via template switching during reverse transcription [60].

The partially disassembled core containing the reverse transcription complex

(preintegration complex) is translocated to the nucleus (Figure 1-2D) where integration into the host cell chromosome occurs (Figure 1-2E) to form the provirus (Figure 1-2F).

Analysis of HTLV-1 integration sites indicates that there is a lack of preferred integration sites [61-66]. Extensive analysis of HTLV-1 integration sites have not resulted in the identification of hotspots [67, 68].

7 1.5.3 Viral gene transcription

The LTRs of the HTLV-1 provirus contain the necessary promoter and enhancer elements to initiate RNA transcription (Figure 1-2G), and synthesize viral RNAs that have a polyadenylation signal at the 3’end [1]. Tax, a non-structural protein, is a trans- acting activator of viral transcription, and potently activates viral transcription during the early phase of viral replication by recruiting multiple cellular transcription factors [69].

Three conserved 21-bp repeat elements, known as the Tax-responsive element 1 (TRE-1) bind the cyclic AMP response element binding protein (CREB) at the TRE-1 site via the amino-terminal domain (NTD) [70-76], while the carboxy-terminal domain (CTD) of

Tax is believed to promote the transcriptional initiation and RNA polymerase elongation by directly interacting with the TATA binding protein [5, 77]. The Tax-CREB promoter complex recruits the multifunctional cellular coactivators CREB binding protein (CBP), p300, and the p300/CBP-associated factor to the LTR [78-83].

1.5.4 Post-transcriptional regulation

Rex is a positive post-transcriptional regulator essential for splicing and transport of HTLV-1 RNAs (Figure 1-2H-I). Rex specifically interacts with the U3 and R regions of the HTLV-1 RNA known as the Rex-responsive element (RexRE). During the early stages of viral gene transcription, suboptimal levels of Rex are present [84], which results in the exclusive export of doubly spliced viral mRNAs encoding viral accessory genes

(i.e., tax, rex, p30II, p12, p13, and hbz) to the cytoplasm (Figure 1-2I) [85]. Once Rex accumulates in the nucleus, Rex reduces splicing of viral mRNA and the singly spliced

(env) and unspliced (gag-pro-pol) mRNAs are then exported from the nucleus to the

8 cytoplasm leading to the production of enzymatic and structural proteins (Figure 1-2J)

[85]. Rex binds to the RexRE through a highly basic RNA-binding NTD, while the CTD is required for protein oligomerization [86, 87]. Rex also contains an activation domain containing the nuclear export signal, which targets Rex to the nuclear pore complex in order for Rex to shuttle between the nucleus and cytoplasm [88, 89].

1.5.5 Viral protein translation

Once HTLV-1 mRNAs are exported to the cytoplasm, they are translated by the host translational machinery to produce the viral proteins [41]. Full-length vRNA is used for translation of the gag, pro and pol genes; this vRNA can also traffick to the plasma membrane (PM), where vRNA dimers are targeted to virus budding sites along with oligomerized Gag, Gag-Pro-Pol, and Env (Figure 1-2K-L) [90]. The doubly-spliced and unspliced mRNAs are translated by free ribosomes to express the structural and enzymatic proteins (i.e., Gag, protease, RT, integrase), respectively, while the singly spliced mRNA is translated by membrane-bound ribosomes to express Env [61].

1.5.6 Gag trafficking

Viral particle formation occurs after Gag translocation from the cytoplasm to the

PM to form immature virions (Figure 1-2K). HTLV-1 monomeric Gag exist in the cytoplasm and are detected at the PM shortly after the initiation of viral protein translation [91]. HIV-1 Gag interacts with many cellular proteins, including cytoskeleton- associated proteins, though their relationship to HIV-1 Gag trafficking is unclear [92].

9 HTLV-1-infected cells regulate cytoskeletal polarization [93]; how cell polarization influences Gag trafficking to the PM represents a current gap of knowledge in the field.

1.5.7 Gag and viral RNA trafficking

HTLV-1 Gag NC protein binds to HTLV-1 vRNA less tightly compared to that of other retroviral NC proteins, due in part to the anionic CTD of the HTLV-1 NC [94]. The

HTLV-1 MA has been reported to bind RNA, and it was found that HTLV-2 MA binds

RNA at much higher affinity than HTLV-2 NC. This is in direct contrast to HIV-1, where

NC binds to RNA more strongly than HIV-1 MA [95]. These recent findings highlight the importance of both the MA and NC domains in vRNA interactions that are likely critically important for vRNA genome recognition and vRNA packaging. How HTLV-1 vRNA translocates from the cytoplasm to virus budding sites along the PM is an open question in the field. A recent study with HIV-1 vRNA suggests that the vRNA diffuses through the cytoplasm to the membrane [96]. It is formally possible that HTLV-1 vRNA also diffuses through the cytoplasm to reach the membrane. However, it is also plausible that it could bind to Gag before reaching the membrane (Figure 1-2I-K).

1.5.8 Budding and maturation

Cellular factors required for particle release are recruited to virus budding sites, resulting in budding and subsequent release of immature virus particles (Figures 1-2L-M and 1-3) [97-99]. The reported Gag stoichiometry in each particle varies greatly from approximately ~750 to 5000 copies [100-103]. The viral PR cleaves the Gag and Pol polyproteins during and shortly after the release of immature virus particles (Figure 1-

10 2N) [41] – MA is closely associated with the PM, CA forms a capsid core that contains

RT, integrase and the NC-coated vRNA (Figure 1-3). Maturation of released immature particles occurs by the enzymatic activity of PR. Infectious, mature particles can infect permissive target cells (Figure 1-2N) [104].

1.6 Gag-Membrane Interactions

The Gag polyprotein must recognize an assembly site at the PM in order to initiate particle assembly. It has been shown that Gag targets and assembles at specific

PM microdomains known as lipid rafts, which are dense, ordered groups of tightly packed saturated lipids stabilized by cholesterol (Figure 1-4C). The molecular composition of these lipid rafts is different from that of the surrounding membrane [105-

107].

MA is responsible for the binding of Gag to the inner leaflet of the PM, which is mediated by the MA NTD that contains multiple membrane binding signals necessary for membrane interactions. A hydrophobic myristic acid moiety found in MA of most retroviruses, such as HIV-1 and HTLV-1, plays an important role in targeting and inserting Gag to the inner leaflet of the PM [108-112]. MA also contains a highly basic region, mostly arginines and lysines, that interacts electrostatically with the inner leaflet of the PM. Furthermore, it has been shown that HIV-1 Gag is flexible and can adopt a closed conformation, which brings the MA and NC terminal domains in close proximity, allowing these to interact with the anionic inner leaflet of the PM (Figure 1-4B) [113].

The inner leaflet of the PM is rich in acidic phospholipids, such as phosphatidylserine

11 (PS) and acidic phosphatidylinositol-4,5-bisphosphate [PI(4,5)P2], which are important for efficient membrane binding and targeting to the PM (Figure 1-4C) [114, 115].

1.7 Identification of Gag Assembly Sites along the Inner Leaflet of the Plasma

Membrane

The PM is composed of different transmembrane proteins and a wide variety of lipids. These include cholesterol and multiple phospholipids such as phosphatidylinositol phosphates (PIPs), phosphatidyl glycerol (PG), and PS; however, the most abundant lipids are phosphatidylcholine (PC) and phosphatidylethanolamine (PE). The inner leaflet of the PM is mostly composed of PE, PC, PS, and PI(4,5)P2, which makes it acidic and better suited for MA to interact with the PM [116, 117]. HIV-1 Gag uses each one of these lipids as a signal to recognize the PM. However, as one of the main lipids in the inner leaflet, it has been suggested that HIV-1 Gag primarily interacts with PI(4,5)P2 to target the PM [118]. Further studies have contradicted these findings and suggest that

PI(4,5)P2 is not the most important site-specific acidic signal in the PM for HIV-1 Gag.

HTLV-1 and other retroviral Gag proteins (e.g., RSV Gag) do not have a PI(4,5)P2 binding signal but interact strongly with other acidic phospholipids [119, 120].

Contradictory results show that RSV Gag does interact with PI(4,5)P2 at the PM [121].

These lipid rafts could explain how retroviral Gag is able to colocalize at the PM to assemble immature virions.

12 1.8 Determinants of Gag-Gag Interactions

Gag-vRNA, Gag-Gag and Gag-membrane interactions are all required for the assembly and budding of virus particles (Figure 1-2L) [122]. Gag forms higher order oligomers through Gag-Gag interactions primarily involving the CA domain and to some extent the NC domain [99, 123-127]. A recent study showed that the HTLV-1 CA NTD and not the CTD, is the primary determinant of Gag-Gag oligomerization, and play an important role in particle size, morphology, and biogenesis [128].

While CA is typically thought to be the primary site of Gag-Gag interactions, other Gag domains are necessary for stabilizing these interactions. NC has been shown to be an important factor in the formation of Gag-Gag interactions, likely due to its RNA- binding capacities. However, it has also been shown that the HIV-1 MA NTD is capable of binding to the vRNA through an electrostatic interaction [129]. vRNA may serve as a binding platform for Gag assembly, as it may promote Gag oligomerization and expose domains necessary for interactions with the PM (Figures 1-4A,B) [127, 130, 131].

Myristoylation appears to be another requirement for Gag-Gag multimerization, at least in the case of HIV-1 [132, 133]. Fluorescence resonance energy transfer (FRET) studies found that without a myristic acid moiety, there was a significant decrease in

Gag-Gag interactions. The simplest explanation for this phenomenon is that myristoylation concentrates Gag molecules at the PM, bringing the CA domains into contact and facilitating oligomerization.

13 1.9 Gag Oligomerization

Gag-Gag interactions have been primarily studied using HIV-1 as a model system. This system has been thought to be useful due to the structural similarities between retroviruses, specifically the MA domain. Initial studies used crosslinking to show that HIV-1 Gag mainly exists as a dimer in the cytoplasm and does not form trimers or hexamers until it reaches the PM [134]. This observation was recently confirmed using dual-color z-scan fluorescence fluctuation spectroscopy (dczFFS), which quantifies fluorescent proteins inside of living cells [91, 135].

It has typically been thought that other retroviruses, such as HTLV-1, trafficked to and assembled at the PM in a similar fashion to that of HIV-1. Using dcz-FFS together with total internal reflection fluorescence and conventional, epi-illumination imaging, it was recently reported that HIV-1 requires micromolar concentrations of Gag in order for it to target and associate with the PM (Figure 1-4B), while HTLV-1 only requires nanomolar (nM) concentrations of Gag to become associated with the PM (Figure 1-4A)

[135]. These results correlate with previous observations that HTLV-1 Gag-Gag interactions were absent in the cytoplasm [91]. This data also supports the hypothesis that

HTLV-1 reaches the PM as a monomer where it then forms higher order oligomers

(Figure 1-4A), which is in contrast to HIV-1 Gag-Gag interactions, which traffics as lower order oligomers in the cytoplasm prior to targeting the PM (Figure 1-4B) [136,

137]. Therefore, HTLV-1 Gag monomers must translocate to the same area at the inner leaflet of the PM to form Gag-Gag oligomers (Figure 1-4A).

14 1.10 HTLV-1 Transmission

1.10.1 Inter-host transmission

There are generally three modes of inter-host HTLV-1 transmission that have been described: 1) blood and blood products, 2) vertical or 3) sexual transmission [138], but the main mode of transmission is thought to be vertical, i.e. from mother-to-child through breastfeeding [139]. Mother-to-child transmission rates vary from 5% to 27% for children nursed by infected mothers and correlate with the duration of breastfeeding [140,

141]. While the precise nature of how infection occurs through the mucosal and epithelial barriers of the gastrointestinal tract, it is generally believed that infected lymphocytes in breast milk carry the virus into the gut [142]. Once in the gut, either cell-free virus or virus-infected cells must pass through the epithelium. A recent study demonstrated in vitro that cell-free HTLV-1 may cross the epithelial barrier via transcytosis before infecting subepithelial dendritic cells [143]. The precise mechanism of transcytosis for

HTLV-1 remains unclear. However, studies with HIV-1 have shown that transcytosis across vaginal epithelial cells occurs via the endocytic recycling pathway [144]. It is plausible that other mechanism(s) are involved in HTLV-1 infection across the gut epithelial barrier due to the low infectivity of cell-free virus. While cell-free HIV-1 is generally thought to be much more infectious than cell-free HTLV-1, it has been suggested that HIV-1-infected lymphocytes more efficiently infect target cells in the gut than cell-free virus – which could be explained by the formation of a viral synapse that induces transcytosis [145]. The role of the virological synapse (VS) in these transmission events has not been reported to date. Another open question is whether HTLV-1 infected

15 lymphocytes can transmigrate as a whole cell across the epithelial barrier and infect subepithelial immune cells.

Zoonotic transmission events of simian T-cell leukemia virus type 1 (STLV-1) to humans after contact with nonhuman primates through bites or bushmeat slaughtering continue to be reported, establishing the emergence of new HTLV-1 infections in humans

[59]. More than 8% of individuals bitten by nonhuman primates in Africa are infected with HTLV-1, and virus transmission cannot be attributed to mother-to-child transmission [146]. The strains of HTLV-1 found in infected individuals closely resembled the subtypes of STLV-1 commonly found in the primate species by which they were bitten [146, 147]. In fact, it is likely that the emergence of HTLV-3 and

HTLV-4 may be attributable to recent STLV zoonotic transmission events, as STLV-4 is known to be endemic in gorillas, and phylogenetic analyses have shown that HTLV-4 is not an ancient human virus but recently emerged in the human population [148]. While these findings highlight the potential ongoing role of nonhuman primates as virus reservoirs, they also highlight interest in the virus-host interactions that facilitate cross- species transmission as well as potential risks in transmission and emergence of more highly pathogenic types of HTLV.

1.10.2 Cell-to-Cell Transmission

In general, there are two distinct methods of virus transmission between cells: virus infection of cells in the absence of cell-to-cell contacts and virus infection involving cell-to-cell contacts. Most retroviruses can efficiently infect target cells in the absence of cell-to-cell contacts – in which the virus buds from an infected cell, diffuses away and

16 infects a new target cell. Experimentally, HTLV-1 is notorious for being poorly infectious in the absence of direct cell-to-cell transmission, and co-cultivation of permissive target cells with virus-producing cells is the most efficient means of virus transmission [149].

1.10.3 Virological Synapses

Immunofluorescence and confocal microscopy have been previously used to demonstrate that Gag and Env proteins are more evenly distributed in individual infected

T-cells, but once the cell encounters another cell, cell polarization occurs – impacting the localization of HTLV-1 Gag, Env and the vRNA towards the cell-cell junction. This cell- to-cell junction, termed the VS, shares many features with the previously described immunological synapse, such as ordered talin domains and microtubule organizing center

(MTOC) polarization [150]. Cryoelectron tomography (cryo-TM) studies of HTLV-1 associated VS structures suggest that there is no fusion of the cell membranes [151]. To the contrary, HTLV-1 transmission occurs via rapid budding and fusion of the HTLV-1 virus across the VS from the infected to uninfected cell.

In comparison of HTLV-1 cell-cell transmission to that of HIV-1 cell-cell transmission, HTLV-1 is more dependent upon cell contact for infectivity than HIV-1.

For example, co-culturing of virus-producing cells with permissive target cells significantly enhances HTLV-1 transmission by several thousand-fold, but only increases

HIV-1 transmission 10-100 hundred-fold [152, 153]. Disruption of actin and tubulin polymerization inhibits HTLV-1 spread to a greater extent than that of HIV-1 spread

[153]. Nonetheless, HIV-1 cell-cell transmission remains a relevant form of virus spread

17 in vivo. A recent study helps to highlight this by the observation that granulocytes, more specifically basophils, are capable of capturing HIV-1 on the cell surface and efficiently transfer HIV-1 to susceptible T cells through cell-to cell contact [154].

1.10.4 Viral Biofilms

In addition to transmission via the VS, HTLV-1 particles have been reported to have the ability to form a biofilm-like, carbohydrate-rich extracellular structure on the surface of cells. These structures are composed of collagen, agrin, tetherin, and galectin-3 and may function as a way to concentrate HTLV-1 particles in a single location to increase the likelihood of infection of a permissive target cell [155].

While it has been observed that dendritic cells may be infected by cell-free

HTLV-1 in vitro [54], a recent study analyzed the infectivity of chronically infected

C91PL cell culture supernatant as compared to purified biofilms in primary human monocyte-derived dendritic cells (MDDCs) and CD4+ T-cells. It was observed that while

MDDCs were more easily infected than T-cells , both cell types achieved significantly higher proviral loads when exposed to viral biofilms as opposed to supernatant-derived virus [156].

1.11 Monoclonal Expansion of HTLV-1 Infected Cells & Leukemogenesis

While HTLV-1 cell-cell transmission is likely a critical determinant of virus transmission from an infected individual to a susceptible individual, many studies have established that the primary route of replication for HTLV-1 in vivo is through mitotic division of host cells and subsequent propagation of the provirus by clonal expansion.

18 DNA analysis of HTLV-1 sequences from patients in geographically distinct locations shows very little genetic variation among HTLV-1 isolates [157]. This is likely because of a low evolutionary rate of 7.06 x 10-7 – 1.38 x 10-5 substitutions per site per year in the

LTR and env regions [158]. While the HTLV-1 reverse transcriptase has a reduced mutation rate when compared with HIV-1 reverse transcriptase (7 x 10-6 mutations per target base pair per replication cycle for HTLV-1 as compared to 3.4 x 10-5 for HIV-1), the fourfold difference is likely not sufficient alone to explain the relative difference in genetic diversity between the two viruses [159, 160]. Also, administration of reverse transcriptase inhibitors has been demonstrated to not reduce proviral loads, even when administered shortly after infection [161, 162].

The most compelling evidence in favor of a clonal expansion of HTLV-1 proviruses is the clonality of T-cells in infected individuals. HTLV-1 integrates randomly into the genome [163, 164]. When HTLV-1 proviral integration sites are amplified by

PCR, clonal populations of infected cells have been detected in both symptomatic and asymptomatic carriers [165, 166].

1.14 Dissertation Objectives

As the first human retrovirus discovered in the early 1980s, HTLV-1 has been studied extensively, yet there is still no treatment or vaccine for HTLV-1 infection.

Additionally, ATL and HAM/TSP treatments are symptom-based and do not directly treat the viral infection. Continued research on the molecular aspects of HTLV-1 replication will enhance opportunities for the discovery of potential antiretroviral targets that can be exploited for the development of effective therapeutic strategies.

19 The assembly of virus particles is a key aspect of viral replication that is still poorly understood at the molecular level. Retroviral assembly has been extensively investigated, though detailed information is lacking for many aspects of the process. This is true for HIV-1, and particularly true for HTLV-1. Recent observations have demonstrated differences in the form and concentration of Gag that is associated with translocation to the PM. While HIV-1 Gag dimers are the primary form of Gag that is thought to translocate to the PM at µM cytoplasmic concentrations (i.e., concentration dependent translocation), HTLV-1 Gag has been found to translocate to the PM as a monomer at nM cytoplasmic concentrations. These observations imply that fundamental differences exist in the association of different Gag proteins with the PM, including interactions with lipids. Such variances suggest that differences in the movement of Gag along the inner leaflet of the PM may also be distinct among different retroviruses. Taken together, these observations argue for the importance of comparative studies of retroviruses in order to provide the greatest insights into the diversity of strategies associated with the virus assembly pathway.

The goals of the studies presented in this dissertation were to design a model system to further study Gag functions in the virus particle assembly pathway, to better characterize the morphology of both immature and mature HTLV-1 virus particles, and to better understand the function of Gag in the assembly of HTLV-1 virus particles. In

Chapter II, the morphology of both immature, virus-like particles (VLPs) and mature

HTLV-1 virus particles was characterized by determining the number of Gag polyproteins that are incorporated into the virus particles. Chapter III reports the development of an YFP-tagged Gag VLP system that can be used with fluorescent

20 microscopy techniques to mechanistically study the assembly pathway of Gag in living cells. Finally, Chapter IV reports, for the first time, the CA core morphology of authentic

HTLV-1 particles in a natural hydrated state.

Deciphering Gag lattice and capsid core structure along with Gag copy number is critically important for understanding the mechanistic details and morphologic changes that occur during virus particle assembly and maturation. These studies should aid in the discovery of novel therapeutic targets to block particle assembly and maturation.

21

Figure 1-1. HTLV-1 genome organization. Diagram of the human T-cell leukemia virus type-1 (HTLV-1) proviral DNA and gene organization. The HTLV-1 proviral sequence contains the 5’ and 3’ long terminal repeats (LTRs) at both ends. The LTRs contain regulatory elements within the unique 3’ region (U3), repeat region (R), and the unique 5’ region (U5). This proviral DNA, like all retroviruses, encodes structural (gag), enzymatic (pro-pol), and envelope (env) genes. The Gag protein contains matrix (MA), capsid (CA), and nucleocapsid (NC) domains. The pro gene encodes the viral protease

(PR), the pol gene encodes reverse transcriptase (RT) and integrase (IN), while the env gene encodes the envelope glycoprotein, which contains the surface (SU) and transmembrane (TM) domains. The HTLV-1 provirus also contains a pX region in the 3’ portion of the genome that encodes the accessory genes (tax, rex, pX-I, pX-II) as well as a basic leucine zipper factor (hbz) minus strand RNA which encode the Tax, Rex, p12, p27, p13, p30 and the antisense synthesized HBZ protein, respectively.

22

Figure 1-2. HTLV-1 life cycle. The major steps in the life cycle of HTLV-1 are shown.

A mature, infectious HTLV-1 virion attaches and fuses to the target cell membrane through interaction with the target cell surface receptors GLUT1/HSPG/NRP-1 via the

HTLV-1 envelope surface and transmembrane domains of the envelope (Env) protein

(A). Following fusion, the viral core containing the viral genomic RNA (vRNA) is delivered into the cytoplasm (B), and during and/or following entry the vRNA genome undergoes reverse transcription to convert the vRNA into double stranded DNA (dsDNA)

(C). The dsDNA is then transported into the nucleus (D), and it is integrated into the host genome (E, F). The provirus is then transcribed by cellular RNA polymerase II (G) as 23 well as post-transcriptionally modified (H). Both full-length and spliced viral mRNAs are exported from the nucleus to the cytoplasm (I). The viral proteins are then translated by the host cell translation machinery (J) and the Gag, Gag-Pol and Env proteins are transported to the plasma membrane (PM) along with two copies of the vRNA genome

(K). These viral proteins and vRNAs assemble at a virus budding site along the PM to form an immature virus particle (L). The budding particle releases from the cell surface

(M) and undergoes a maturation process through the action of the viral protease, which cleaves the viral polyproteins to form an infectious, mature virus particle (N).

24

Figure 1-3. Gag and retrovirus particle assembly. A graphic depiction is shown in cross section of the assembly of a prototypic retrovirus particle, emphasizing the oligomerization of Gag along the inner leaflet of the plasma membrane, incorporation of two copies of the viral RNA, the budding of an immature virus particle, and the conversion of the immature virus particle to mature infectious virus particle that is catalyzed by the viral-encoded protease. Gag is shown as being composed of the matrix domain (red circle), the capsid domain (blue oval), and the nucleocapsid domain (purple circle). Two copies of the viral RNA (two orange lines inside the viral particle) are shown packaged into the virus particle.

25

Figure 1-4. Schematic representation of HTLV-1 and HIV-1 Gag-membrane association. (A) HTLV-1 Gag associates with the plasma membrane (PM) as a monomer and is found in the PM at nM concentrations. HTLV-1 Gag is shown as recruiting viral

RNA (vRNA) after its association with the inner leaflet of the PM (though monomeric

Gag may recruit the vRNA in the cytoplasm and are transported together to the PM). (B)

Concentration-dependent HIV-1 Gag dimerization and translocation to the PM. HIV-1

Gag must reach a critical cytoplasmic concentration (~0.5µM) in order for Gag-Gag dimers to form and subsequent Gag-membrane association to occur. The myristoyl moiety of the HIV-1 Gag is exposed, allowing for association with the PM of Gag dimers

(including Gag dimers associated with dimeric vRNA). Both HIV-1 Gag N- & C- terminal domains interact with the inner leaflet of the PM and RNA, but is not until HIV-

1 Gag interacts with the PM, other HIV-1 Gag molecules, and to NA that the protein becomes extended. (C) Expanded view of lipid raft that HIV-1 Gag associates with at the

PM. Shown are three key constituents of lipid rafts: cholesterol (yellow), phosphatidylserine (PS; blue), and phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2; red].

26

CHAPTER II

Distinct Morphology of Human T-Cell Leukemia Virus Type 1-Like Particles

The text presented in this chapter has been published:

Maldonado, J. O., Cao, S., Zhang, W., & Mansky, L. M. (2016). Distinct Morphology of Human T-Cell Leukemia Virus Type 1-Like Particles. Viruses, 8(5). doi: 10.3390/v8050132

27 2.1 Introduction

Approximately 10–20 million people are infected with human T-cell leukemia virus type 1 (HTLV-1) worldwide [7, 167]. HTLV-1 is a deltaretrovirus and is associated with adult T-cell leukemia/lymphoma, tropical spastic paraparesis, as well as HTLV-1- associated myelopathy [168, 169]. These diseases are prevalent in places highly endemic for HTLV-1 infection such as southwestern Japan, central Africa, South America and the

Caribbean. Despite the association of HTLV-1 with cancer and its significant impact on human health and well-being, the molecular mechanisms of viral replication, virus particle assembly and morphology remain poorly understood due to difficulties in propagating the virus in tissue culture.

Like other retroviruses, the assembly and budding of HTLV-1 particles is directed by the viral Gag polyprotein (recently reviewed by Maldonado et al. [170]). Briefly,

HTLV-1 Gag molecules translocate to the plasma membrane (PM) soon after the protein is synthesized [91]. A previous study with human immunodeficiency virus type 1 (HIV-

1) suggested that the viral RNA is recruited to the PM by Gag and serves as a platform to promote Gag-Gag interactions, allowing Gag to form higher order oligomers in immature particles [171]. Infectious virions are produced via a maturation process that occurs either concomitantly with or after budding of the immature virus. During virus maturation, the viral protease (PR) cleaves the Gag polyprotein into three structural proteins: matrix

(MA), which remains associated with the inner leaflet of the viral membrane; capsid

(CA), which organizes into a closed protein shell to package the genomic RNA; and nucleocapsid (NC) which is in complex with the viral genome.

28 The diameters of retrovirus particles are typically variable and commonly appear to form a normal distribution [100, 103, 172-177]. Calculations of the average Gag copy number per virus particle vary somewhat depending on the methods used for the measurement as well as on the type of retrovirus being analyzed. Scanning transmission electron microscopy (STEM) has previously been used successfully to determine the average Gag copy number per particle [100, 102, 175, 177-179]. This method estimates the mass of the whole virus particle. Since the majority of the virus particle mass is contributed by Gag, the mass of the entire particle has been used for calculating the Gag stoichiometry. Multiple studies have reported varying Gag copy numbers, ranging from approximately 750 to 5000, which coincide with varying virus particle size distributions

[101-103, 172, 175-177, 179]. To date, there are no reported studies on Gag stoichiometry that would be present in the immature precursors of authentic, mature

HTLV-1 particles. Determination of Gag stoichiometry is critical to understanding the mechanisms of HTLV-1 replication, for this information assists in the interpretation of

HTLV-1 particle structures, and helps in determining the copy number of other viral proteins in the virus particle (e.g., Pol).

In this study, a comparative analysis of HTLV-1-like particles and authentic, mature HTLV-1 particles was performed by cryogenic transmission electron microscopy

(cryo-TEM) and scanning transmission electron microscopy (STEM). These findings provide the first demonstration of the morphology of these virus-like particles (VLPs) having the unique feature of local flat Gag lattice regions that did not follow the curvature of the viral membrane and had an enlarged distance toward the membrane.

Morphological features similar to that observed with other retroviruses [122] include (1)

29 a Gag lattice with multiple discontinuities; (2) a string of bead-like densities at the inner leaflet that is associated with the Gag lattice; and (3) a Gag lattice resembling a railroad track. We also demonstrate that HTLV-1-like particles and authentic mature HTLV-1 particles possess a consistent size and Gag stoichiometry.

2.2 Materials and Methods

2.2.1 Transfection and HTLV-1-Like Particle Production

A codon-optimized HTLV-1 gag gene expression construct (pN3 HTLV-1 Gag,

Figure 2-1A) was created in a similar manner to that of a previously described construct in which the yellow fluorescence protein (YFP) was fused to the carboxy-terminus of

Gag (pEYFP-N3 HTLV-1 Gag) [180]. The new gag gene which does not have a YFP tag was synthesized with an optimal Kozak consensus sequence at the 5′ end of the gene:

GCCACCATGG (start codon in bold and underlined) (Figure 2-1A). In order to produce

VLPs, six 10 cm tissue culture dishes each containing 2.2 × 106 human embryonic kidney

293T cells in 6 mL of Dulbecco's Modified Eagle Medium (DMEM) supplemented with

10% Fetal Clone III were co-transfected with the pN3-HTLV-1 Gag expression construct along with an HTLV-1 envelope protein expression construct (ratio of 10:1) using

GeneJet (SignaGen, Gaithersburg, MD, USA) following the manufacturer’s instructions.

Twenty-four hours post-transfection, 2 mL of fresh media were added to each plate and incubated for additional 24 h at 37 °C in 5% CO2. To harvest VLPs, cell culture supernatants from transfected cells were centrifuged at 3000 × g for 5 min to remove large cellular debris and then filtered through a 0.2 µm filter. The samples were then concentrated and purified in the same manner as with authentic particles.

30

2.2.2 Gradient Purification of Authentic Virus Particles and VLPs

Authentic HTLV-1 particles were produced from MT-2 cells, a T-cell line chronically infected with HTLV-1, which was obtained from Dr. Douglas Richman through the NIH AIDS Reagent Program, Division of Acquired Immune Deficiency

Syndrome (AIDS), National Institute of Allergy and Infectious Diseases (NIAID),

National Institutes of Health (NIH) [181, 182]. MT-2 cells were grown in two T-75 flasks with up to 60 mL of Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% Fetal Clone III, for ~10 days. After the cells reached about 90% confluency, which was indicative by the formation of large cell clumps, virus particles were harvested and centrifuged at 3000 × g for 5 min to remove large cellular debris and then filtered through a 0.2 µm filter. The concentrated particles (i.e., authentic virus particles or VLPs) were then ultracentrifuged through an 8% OptiPrep (60% iodixanol in water with a density of 1.32 g/mL, (Sigma-Aldrich, St. Louis, MO, USA) cushion at

109,000 × g for 1.5 h in a 50.1 Ti rotor (Beckman, Brea, CA, USA) at 4 °C. The particle pellet was resuspended in 0.5 mL of 1× STE buffer (100 mM NaCl , 10 mM Tris-Cl, pH

7.4, 1 mM sodium chloride-Tris ethylenediaminetetraacetic acid (EDTA), and overlaid onto a 4 mL 10%–40% OptiPrep gradient and centrifuged to equilibrium in a SW55 Ti rotor (Beckman) at 250,000 × g for 3 h at 4 °C. The virus- or VLP-containing fraction, at about 20% OptiPrep, was removed from the gradient using a hypodermic needle. The collected virus particles were diluted 10 fold in 1 × STE and pelleted at 195,000 × g for 1 h in a SW55 Ti rotor at 4 °C. Following centrifugation, the pellet was re-suspended in

~15 µL of 1 × STE at 4 °C overnight and then analyzed by cryo-TEM or STEM. The

31 compound 2, 2′-dithiodipyridine (aldrithiol-2; AT-2) was used to inactivate authentic

HTLV-1 infectivity prior to cryo-TEM or STEM analysis as previously described [183].

2.2.3 Cryo-TEM of HTLV-1-Like Particles and Authentic Virus Particles

Virus and VLP samples were prepared for cryo-TEM as previously described

[180]. Briefly, 3 µL concentrated virus or VLP sample was applied to a glow-discharged c-flat holey carbon grid (Ted Pella, Redding, CA, USA) and then blotted with filter paper to remove the sample excess. The grid was then plunged frozen into liquid ethane [184] with a FEI MarkIII Vitrobot system (FEI Company, Hillsboro, OR, USA). The frozen grids were then transferred to a FEI TF30 field emission gun transmission electron microscope at liquid nitrogen temperature (FEI Company). Images were then recorded at a nominal magnification of 39,000_x and 59,000_x at low-dose (~30 electrons/Å2) and 1 to 5 µm under focus conditions using a Gatan 4 k by 4 k CCD camera (Gatan Inc.,

Pleasanton, CA, USA).

2.2.4 Determination of Particle Size

Cryo-TEM images were analyzed by using ImageJ software (Version 1.49c, NIH,

Bethesda, MD, USA). For each virus particle or VLP analyzed, two perpendicular diameters were used to calculate the average diameter [180]. Histograms of particle diameters were generated by using GraphPad Prism 6 software (Version 6.0c, GraphPad,

La Jolla, CA, USA).

32 2.2.5 Determination of Particle Mass by STEM

The mass of virus particles or VLPs was determined by quantitative dark-field

STEM, which was developed at the Brookhaven National Laboratory (BNL, Upton, NY,

USA) [185]. This method allows for the study of individual unstained virus particles with minimal radiation damage. The particle sample was first mixed with tobacco mosaic virus (TMV) particles that were used as an internal control. Then the mixture was applied onto a thin-carbon transmission electronic microscopy (TEM) grid, extensively washed, blotted and freeze-dried overnight. The TEM grid was imaged under a 40 keV electron beam at −150 °C. The grid was first scanned with a low dose electron beam and areas with clean background were used for the final scan. Prior to the final scanning, the electron beam was focused in a nearby area to minimize radiation damage to the specimen. The low temperature and low dose imaging technique (<500 electrons/nm2) was used to help to reduce mass loss (less than 1%) caused by electron radiation as well as to eliminate contamination from mobile hydrocarbons. Each point in the STEM image corresponds to an area of 0.625 nm2 over the specimen. The whole image corresponds to

512 by 512 nm in the specimen with the center of the points separated by 1 nm. A large- and a small-angle annular dark-field detector were used to digitally record the number of scattered electrons in each scanning point. The number of scattered electrons at any scanning point is proportional to the sample mass in that local region.

STEM images were analyzed by using the PCMass software developed by the

BNL STEM facility (Version 32, Brookhaven National Laboratory, Upton, New York,

USA). Each virus particle or VLP in the STEM micrograph was first masked by a density profile model (i.e., a sphere) in order to mimic the virus density profile. The diameter of

33 the sphere was based on the dimension of the measured particle. The mass of each virus particle was then calculated using the sum of the electron densities within the mask and a scale factor, which was determined using the image of TMV and its mass per unit length

(i.e., 13.1 kDa/Å) [185, 186]. The resulting histograms and graphs of particle mass distribution were generated by using GraphPad Prism 6 software (Version 6.0c,

GraphPad, La Jolla, CA, USA).

2.3 Results

2.3.1 Analysis of the Morphology of HTLV-1-Like Particles

HTLV-1-like particles produced using the HTLV-1 Gag-only expression construct (Figure 2-1A) were observed to be spherical in shape with a mean diameter of

110 ± 32 nm measured from 1172 particles (Figure 2-1B,C). This is in contrast to a previous study using a Gag-YFP expression construct in which a mean particle diameter of 71 ± 20 nm was determined by cryo-TEM [180]. The electron density adjacent to the inner viral membrane was interpreted as being the immature Gag lattice. All particles with this electron density pattern were counted as Gag-containing particles. Intriguingly, many local regions of Gag assembly were observed to exhibit flat electron density features that did not strictly follow the curvature of the membrane and showed enlarged distance toward the viral membrane (4 vs. 8 nm) (Figure 2-2A,B). About 20% of particles had this morphological feature. This unique structural feature has not been reported for other retrovirus immature particles. The flat Gag density feature observed with HTLV-1- like particles suggests that HTLV-1 Gag could be arranged in a lattice structure that is

34 distinct from that of other retroviruses characterized to date (i.e., HIV-1, Mason-Pfizer monkey virus (MPMV) and Rous sarcoma virus (RSV)) [176, 187, 188].

Other morphological features have commonalities with other retroviruses. First, the Gag densities in the cryo-TEM images were not always continuous. In particular, smaller VLPs were observed to have multiple discontinuities in the immature Gag lattice

(Figure 2-2A,B). The membrane regions that associate with organized Gag lattices appear to be wider and more pronounced. Between the Gag lattice and the viral membrane, a string of bead-like densities is sometimes observed (i.e., in ~10% of particles analyzed) lining along the inner leaflet of the viral membrane (Figure 2-2E,G). These density features are likely due to the association of MA with the inner membrane of the virus particle. The MA lattice has been observed in a large membrane-enclosed multi-core structure in supernatants of HIV-1-infected cells [189]. In the cryo-TEM images (Figure

2-2A-G), the arrangement of the Gag molecules within the lattice is similar but not identical to that in other retrovirus immature particles such as HIV-1. The cryo-TEM image (Figure 2-2H) and cut-away view of the HIV-1 Gag lattice assembly in a three- dimensional (3D) reconstruction map [190, 191] has two density layers: closer to the viral membrane is the CA protein layer showing an array of rod-like densities, while towards the center of the particle is the NC layer that has a continuous density. In contrast, the

HTLV-1-like particles consistently display a continuous density at the region closer to the membrane inner leaflet. Underneath the continuous density layer, closer to the center of the virus particle, is an array of densities that resembled a railroad track (Figure 2-2H).

2.3.2 Morphology of Authentic HTLV-1 Mature Particles Produced from MT-2 Cells

35 As a control and to confirm our previous studies with authentic HTLV-1 particles by cryo-electron tomography [192], the morphology of authentic HTLV-1 particles, harvested and purified from MT-2 cells [181, 182], was also studied by cryo-TEM

(Figure 2-3A). Mature virus particles were identified by either the presence of readily observable electron-dense cores, or by the presence of significant electron density within the particle. The lack of HTLV-1 protease inhibitors prevented the production of large numbers of authentic immature particles. Vesicles were identified by the absence of core structures or significant internal electron density. The particles were primarily spherical and heterogeneous in size. The particle diameter was determined by averaging the longest and shortest measurements of each particle. A total of 1074 authentic particles were measured and had a mean diameter of 113 ± 23 nm (Figure 2-3A,B). This measurement based on two-dimensional (2D) cryo-TEM images was in good agreement with our previous analyses using the cryo-electron tomography method [192].

A gallery of cryo-TEM images of authentic HTLV-1 particles (Figure 2-3C) revealed that the particles contained an unordered polyhedral-like capsid core structure, which is different in each particle regardless of particle size. The core size varied by particle, with some regions of the protein capsid of the cores following the curvature of the inner leaflet of the viral lipid bilayer, while other parts of the capsid appeared completely separated from the viral membrane.

2.3.3 STEM Analyses of HTLV-1-Like Particles and Authentic Mature HTLV-1 Particles

STEM analysis was used to determine the total molecular mass of HTLV-1 particles as previously described [185]. Representative dark-field electron micrographs of

36 HTLV-1-like particles and authentic mature HTLV-1 particles are shown in Figures 2-4A and 2-5A. Only isolated intact particles that were of the expected particle diameter range, as determined by cryo-TEM imaging, were used for mass measurements. Some smaller randomly distributed contaminants are visible in the background. Using the known mass of TMV as an internal control, we are able to obtain the average masses of HTLV-1

VLPs and authentic particles.

Both HTLV-1-like particles and authentic particles showed a wide distribution of mass diversity, which correlates with the wide particle size distribution (Figures 2-4B and

2-5B). The TMV-corrected masses of HTLV-1-like particles (Figure 2-4B) and HTLV-1 authentic particles (Figure 2-5B) were determined to be 174 ± 96 MDa and 204 ± 67

MDa, respectively (Table 2-1). The average masses were used for estimating the Gag copy numbers in HTLV-1-like particles and inferred immature precursors of the authentic

HTLV-1 particles. AT-2 was used to inactivate the particles. It is formally possible that

AT-2 treatment could affect particle morphology.

2.3.4 Calculation of Gag Stoichiometry in HTLV-1-Like Particles

The average mass of HTLV-1-like particles determined by STEM was used to estimate the average Gag copy number per virus particle. The viral RNA mass contribution from total particle mass was determined by extracting the RNA from particle lysates with RNA columns, using Roche’s High Pure Viral RNA Kit (Roche Diagnostics,

Indianapolis, IN, USA). The extracted viral RNA was quantified by determining the ultraviolet (UV) absorption at 260 nm and a conversion factor of 40 µg/mL × A260 optical density unit x dilution factor using a Beckman DU-65 spectrophotometer

37 (Beckman Coulter, Brea, CA, USA). The Thermo Scientific Pierce BCA Protein Assay

Kit (Thermo Fisher Scientific, Waltham, MA, USA) was used to estimate the protein content of the same sample used to determine the VLPs’ RNA content. The VLPs’

Gag/RNA mass ratio was determined to be 14.4:1, equivalent to about 4% of the averaged molecular mass of the VLPs measured by STEM (Table 2-1). Based upon the average size of the VLPs, which is 110 nm, and estimating the average thickness of the viral membrane to be 5 nm, an estimate of the number of lipid molecules in the virus envelope was made. Assuming that the distance between lipid molecules in the same leaflet is 0.85 nm [193], and the average molecular weight of lipids was 750, the mass of lipids in an averaged size particle was determined to be approximately 70 MDa (i.e.,

~40% of the total mass of an averaged sized particle) (Table 2-1).

Assuming that the mass of the Gag protein in HTLV-1-like particles is similar to that of other retroviruses, approximately 70%–90% of the total protein mass [102, 194], the mass contribution of Gag in an HTLV-1-like particle with an average size of 110 nm would be 70–87 MDa. Given the molecular weight for HTLV-1 Gag is ~53 kDa [39,

195], it was estimated that HTLV-1-like particles contain approximately 1300–1600 Gag polyproteins per VLP with a mass and diameter of 174 MDa and 110 nm, respectively

(Table 2-1).

2.3.5 Estimating Gag Stoichiometry in Authentic Immature HTLV-1 Particles by

Calculating Gag Copy Number in Authentic Mature HTLV-1 Particles

The same methodology used to calculate the Gag copy number in HTLV-1-like particles was used to estimate the Gag stoichiometry in authentic immature HTLV-1

38 particles by calculating the Gag copy number in authentic mature HTLV-1 particles.

Based upon the average size of the authentic mature HTLV-1 particles, which is 113 nm, we estimated the lipid mass to be approximately 80 MDa (Table 2-1). The viral RNA in authentic HTLV-1 particles was calculated by assuming that each authentic particle contains two copies of the 8.5 kb genomic RNA, plus tRNA and other small RNAs comprising approximately 30% of the genomic RNA by mass. The molecular weight of

RNA was estimated to be 7 MDa, equivalent to about 3.5% of the averaged molecular mass of the particle measured by STEM. Assuming that the mass of the Gag protein in retroviruses is about 70%–90% of the total protein, the Gag protein of authentic HTLV-1 particles would contribute 82–106 MDa to the total particle mass for a particle of an average size of 113 nm (Table 2-1). An authentic HTLV-1 particle contains three forms of the Gag polyprotein: Gag, Gag-Pro and Gag-Pro-Pol, with molecular weights of 53 kDa, 76 kDa and 180 kDa, respectively [39, 195]. Given that the estimated molar ratio of the Gag, Gag-Pro, and Gag-Pro-Pol is 100:10:1 based on in vitro translation of viral RNA

[196], it was calculated that the immature precursor of authentic HTLV-1 particles contains approximately 1500 to 1900 copies of Gag, which would result in a particle with a mass of 204 MDa and a size of 113 nm (Table 2-1) [193, 196].

2.4 Discussion

Although HTLV-1 was the first human retrovirus to be discovered [1, 197], the morphological details of HTLV-1 particles have been poorly characterized, including that of Gag stoichiometry. To combat the technical difficulties in working with HTLV-1 in cell culture, a HTLV-1 Gag-only expression model system was used to produce and

39 purify HTLV-1-like particles. A key technical advantage of this HTLV-1 Gag model system is that it is a highly robust system that results in highly efficient production of

VLPs from mammalian cells. In the absence of methodologies to efficiently produce authentic immature HTLV-1 particles, and given the absence of HTLV-1 PR inhibitors

[198-201], this construct was used as a surrogate to study immature particle morphology.

The electron density of the HTLV-1 Gag lattice appears more compact than what has been previously observed for Gag lattices from HIV-1, MPMV or RSV [172, 187,

202]. The most intriguing morphological feature of the HTLV-1 immature Gag lattice is that about 20% of the HTLV-1-like particles had regions that appeared to be flat and did not follow the curvature of the viral membrane in multiple regions. The maximum separation between these ‘flat’ regions and the viral membrane was approximately 8 nm.

This is the first time this observation has been made regarding the structure of an immature retroviral Gag lattice. One intriguing possibility is that this morphological feature is indicative of a more rigid lattice structure compared to that of other previously reported HIV-1 immature Gag lattice structures. The addition of a fluorophore tag on the carboxy terminus of the HTLV-1 Gag protein did affect the diameter of the VLPs (i.e., average diameter of 110 nm without tag versus 75 nm with tag) as well as the Gag-Gag interactions, given the distinct morphological differences in the presence and absence of the fluorophore tag [180]. This is in contrast to that observed with HIV-1 Gag, where particles produced from a Gag-YFP expression construct did not influence particle size

[101]. Taken together, these results imply distinct differences in the Gag assemblies in

HTLV-1 immature particles compared to that of other retroviruses, particularly HIV-1.

40 STEM analysis led to the observation that the HTLV-1 Gag copy number distribution per particle spanned a wide range for both HTLV-1-like particles and HTLV-

1 authentic particles (Figures 2-4B and 2-5B), which corresponded to the diverse particle size population (Figures 2-1C and 2-3B). HTLV-1-like particles and mature particles were found to have Gag copy numbers of 1300–1600 and 1500–1900 Gag molecules/particle, respectively, which is in the general range of Gag copy numbers observed for other retroviruses including MPMV and RSV [102, 179].

The observations made by this study emphasize both unique and common morphological features of the HTLV-1-like particles in comparison to other retrovirus immature VLPs. Future studies will include a detailed determination of the immature Gag lattice, which should provide important new insights into the unique aspects of HTLV-1 particle assembly, in particular, and new insights into retroviral assembly, in general.

41 Figure'1'

A.# B.# C.#

Figure 2-1. Analysis of the diameter and morphology of human T-cell leukemia

virus type 1 (HTLV-1) virus-like particles (VLPs) by transmission electron

microscopy (TEM). (A) HTLV-1-like particle expression construct. A codon-optimized

Gag expression construct (pN3 HTLV-1 Gag) with a Kozak sequence was used to

produce HTLV-1 VLPs; (B) Representative micrograph of HTLV-1-like particles of

different sizes and morphology; (C) Size distribution of HTLV-1-like VLPs.

42

Figure 2-2. Cryogenic transmission electron microscopy (Cryo-TEM) images of

HTLV-1-like particles and comparison of Gag lattice between HTLV-1 and human immunodeficiency virus type 1 (HIV-1). (A–G) Cryo-TEM images of HTLV-1-like particles. The white arrows indicate regions of the Gag lattice that appear flat in contrast to the curvature observed with the viral membrane. The black arrows show the membrane regions that are associated with Gag lattice and exhibit a string of bead-like densities in the inner membrane leaflet. The black arrowheads demark discontinuity of the Gag lattice. The black dash-lined box in D shows a region displayed in the top panel of H. The scale bar in G is applicable to the panels C–G; (H) Comparison of Gag lattice morphology between HTLV-1-like and HIV-1-like particles. The electron densities representing the Gag lattice structure are indicated by the left and right bracket, respectively.

43 Figure'3'

A.# B.# C.#

Figure 2-3. Analysis of the diameter of authentic mature HTLV-1 virus particles.

(A) Authentic mature HTLV-1 particles produced from MT-2 cells; (B) Size distribution

of authentic mature HTLV-1 particles ; (C) Magnified images of authentic mature

HTLV-1 particles showing irregular polyhedral-like core structures. The scale bars in A

and C are 100 nm.

44 Figure'4'

A.# B.# C.#

Figure 2-4. Scanning transmission electron microscopy (STEM) analysis of HTLV-

1-like particles. (A) A STEM micrograph of HTLV-1-like particles mixed with tobacco mosaic virus (TMV); (B) TMV-corrected mass measurement distribution in MDa of purified HTLV-1-like particles, which was determined based on the known TMV mass per unit length of 13.1 kDa/Å.

45 Figure'5'

A.# B.# C.#

Figure 2-5. STEM analysis of authentic mature HTLV-1 virus particles. (A) A

STEM micrograph of authentic HTLV-1 particles mixed with TMV. The region labeled as “clusters” represents closely associated viral particles and is excluded from the calculation; (B) The TMV-corrected measurement of mass distribution in MDa of purified authentic mature HTLV-1 particles was determined. The TMV-corrected particle mass determination was based on the known TMV mass per unit length of 13.1 kDa/Å.

46 Table 2-1. Summary of the mass determinations and the calculated Gag copy number per particle in human T-cell leukemia virus type 1 (HTLV-1)-like particles and authentic HTLV-1 particles.

HTLV-1 Particle Sample Measurement Virus-Like Authentic Particle Particle Average Diameter (nm) a 110 113 Average Particle Mass (MDa) b 174 204 Mass of RNA, RNA c 7 7 Lipid and Lipid d 70 80 Protein (MDa) Total protein e 97 118 Mass of Total Gag 70–87 82–106 Gag polyprotein f Molecules Gag 70–87 70–90 (MDa) Gag-Pro N/A 10–13 Gag-Pro-Pol N/A 2.5–3 Gag polyprotein copy number g 1300–1600 1500–1900

a As determined by cryogenic transmission electron microscopy (Cryo-TEM); b As determined by scanning transmission electron microscopy (STEM); c Mass contributed by RNA in virus-like particles was estimated experimentally as described in the Materials and Methods. The RNA mass contribution for authentic HTLV-1 particles was estimated based upon the genome size; d Mass contributed by lipids was estimated from average particle size and membrane thickness; e Mass of total protein was determined by subtraction of the RNA and lipid mass from the total particle mass as determined by

STEM; f Total Gag polyprotein was estimated based upon the assumption that Gag contributes ~70%–90% of the total protein mass; g The Gag polyprotein copy number represents the range of Gag copy number in a particle that has both average mass and dimensions.

47

CHAPTER III

Perturbation of Human T-Cell Leukemia Virus Type 1 Particle Morphology by

Differential Gag Co-Packaging

The text presented in this chapter has been published:

Maldonado, J. O., Angert, I., Cao, S., Berk, S., Zhang, W., Mueller, J. D., & Mansky, L. M. (2017). Perturbation of Human T-Cell Leukemia Virus Type 1 Particle Morphology by Differential Gag Co-Packaging. Viruses, 9(7), 191. doi: 10.3390/v9070191

48 3.1 Introduction

Human T-cell leukemia virus type 1 (HTLV-1) was the first human retrovirus identified, and is the etiological agent of adult T-cell leukemia /lymphoma (ATLL) and the neurological disorder HTLV-1-associated myelopathy/tropical spastic paraparesis

(HAM/TSP) [168, 169]. It is estimated that 15 million people are infected with HTLV-1 worldwide [7, 167]. Regions in the world with high prevalence levels include the

Caribbean basin, South America, Central Africa, southwestern Japan, southern Africa, and the Middle East [7]. About 5% of HTLV-1 infected individuals develop an aggressive form of ATLL [10, 12].

HTLV-1 is a deltaretrovirus and like other retroviruses, requires the expression of the Gag polyprotein – the main retroviral structural protein – to drive particle assembly and release [203]. Gag translocates from the point of translation to particle budding sites at the plasma membrane (PM) where Gag, the Gag-Pol (a Gag protein with the Pol protein fused to the carboxy-terminus of Gag), viral RNA, and the virus envelope (Env) protein assemble and result in virus particle production [171]. Aspects of this process have been suggested to involve motor proteins and the microtubule network [204], though this remains an open question.

Lipid rafts, i.e., tightly packed saturated lipids, are typically associated with virus budding sites (reviewed by Maldonado et al. [170]), and it is at the lipid-raft associated, virus budding sites that Gag is thought to form higher-ordered oligomers [105]. Gag has three structural domains: matrix (MA), capsid (CA), and nucleocapsid (NC) [41]. The N- terminal domain (NTD) of MA, which contains a hydrophobic myristic acid moiety, is responsible for the targeting and insertion of Gag into the PM [108, 109]. This interaction

49 is thought to stimulate Gag oligomerization [132] at the site of particle assembly primarily through CA interactions [123, 124], with the assistance of the viral RNA, as well as cellular proteins [130, 205-207]. Virus particle budding is initiated by recruitment of the ESCRT proteins [97, 98].

Budding and newly released particles are typically immature and are observed via thin section transmission electron microscopy as having an electron dense Gag layer along the inner leaflet of the viral membrane. The activation of the viral protease cleaves

Gag to release the mature MA, CA and NC proteins, which results in the conversion of an immature particle to a mature infectious particle. Following virus maturation, MA remains associated with the inner leaflet of the viral membrane, NC coats the viral RNA and this complex along with reverse transcriptase and integrase are encapsidated by the

CA core [41, 61].

We previously reported a tractable model system for HTLV-1 Gag expression and the production of HTLV-1-like particles using yellow fluorescent protein (YFP)-tagged

Gag [180]. These virus-like particles (VLPs) were small, about 73 nm in diameter, and contained about 500 copies of Gag. This is in contrast to studies of authentic HTLV-1 particles and VLPs produced from expression of an untagged HTLV-1 Gag, where particles were between 110 and 113 nm in diameter, and contained 1300-1900 copies of

Gag [208]. Taken together, these observations suggest that the YFP tag on the carboxy- terminus of Gag affects particle size and Gag stoichiometry. While there has been extensive use of labeled Gag proteins for studies of retroviral assembly, there is a paucity in the literature regarding their impact on Gag stoichiometry and particle morphology.

50 In this study, we sought to investigate whether HTLV-1-like particle morphology would be perturbed as a result of co-packaging of Gag and Gag fused to a fluorescent protein (i.e., YFP). Gag co-packaging is biologically relevant in the retrovirus assembly pathway as HTLV-1 and retrovirus particles naturally co-package Gag and Gag-Pol. To help address this, we have used cryogenic transmission electron microscopy (cryo-TEM), scanning transmission electron microscopy (STEM), and fluorescence fluctuation spectroscopy (FFS) to investigate co-packaging of Gag and Gag-YFP. We found that ratios of 3:1 (Gag:Gag-YFP) or greater resulted in a particle morphology indistinguishable from that of VLPs produced with the untagged HTLV-1 Gag – i.e., a mean diameter of ~113 nm, and a mass of 83 to 120 MDa as determined by cryo-TEM and STEM, respectively. Furthermore, STEM and FFS analyses indicated that HTLV-1

Gag-YFP was incorporated into VLPs in a predictable manner at the 3:1 ratio. Both

STEM and FFS analyses found that the Gag copy number in VLPs produced with a 3:1 ratio of Gag:Gag-YFP was in the range of 1500-2000 molecules per VLP. Thus, the

HTLV-1 Gag copy number and particle size distribution observed with a 3:1 ratio of

Gag:Gag-YFP was similar to that in authentic HTLV-1 particles [208], and within range of that observed with other retroviruses [100-103, 172, 173, 175-177, 179]. Taken together, the insights gained in these studies contribute new knowledge to how differential Gag co-packaging impacts virus-like particle morphology, which can have relevance to the assembly of authentic HTLV-1 particles by co-packaging of Gag and

Gag-Pol.

51 3.2 Materials and Methods

3.2.1 Production and purification of HTLV-1-like particles

A plasmid construct expressing HTLV-1 Gag with YFP fused to the carboxy- terminus (i.e., pN3 HTLV-1 Gag-YFP) [180] or a construct expressing HTLV-1 Gag only [170] have been previously described and were used to produce VLPs, as previously described [208]. Briefly, 2.2x106 HEK 293T cells maintained in Dulbecco’s modified

Eagle’s medium (DMEM) supplemented with 10% Fetal Clone III were co-transfected with 6µg of pN3-HTLV-1 Gag-YFP or 6µg total of pN3-HTLV-1 Gag and pN3-HTLV-1

Gag-YFP (at ratios of 1:1, 2:1, 3:1, 4:1 and 5:1) along with an HTLV-1 envelope protein

(10:1 ratio of Gag to envelope plasmids) expression construct using GeneJet (SignaGen,

Gaithersburg, MD) following the manufacturer’s instructions. Twenty-four hours post- transfection, 2ml of fresh media was added to cells and incubated for another 24h. Cell culture supernatants of transfected cells were collected 48h post-transfection, clarified by centrifugation 3000 ´ g for 5 min, and filtered through a 0.2 µm filter.

VLPs were then subjected to ultracentrifugation through an 8% OptiPrep (Sigma-

Aldrich) cushion at 109,000 × g for 1.5h using a 50.1 Ti rotor (Beckman) at 4ºC. The

VLP pellet was resuspended in 0.5 ml of 1x STE buffer (10 mM Tris-CL, pH 7.4, 100 mM NaCl, 1 mM EDTA), and overlaid onto a 4 ml 10%-40% OptiPrep gradient and centrifuged to equilibrium in a SW55 Ti rotor (Beckman) at 250,000 × g for 3h at 4ºC.

The fraction containing the concentrated particles was removed from the gradient using a hypodermic needle, diluted 10 fold in 1xSTE buffer and pelleted at 195,000 × g for 1h in a SW55 Ti rotor at 4ºC through an 8% Optiprep cushion. Following centrifugation, the

VLP pellet was resuspended in ~15 µl of 1x STE overnight at 4ºC. Samples were then

52 used for cryo-TEM and STEM analyses. For FFS measurements, 293T cells were transfected and cell culture supernatant was collected in the manner described above.

Supernatant was filtered through a 0.2 µm filter and 200 µl was added to 8-well Nunc

Lab-Tek Chamber Slides (Thermo Fisher Scientific, Pittsburgh, PA). Slides were then sealed to prevent sample loss due to evaporation and used in FFS analyses.

3.2.2 Cryo-TEM analysis of HTLV-1-like particles

VLP samples were prepared for cryo-TEM analysis as previously described [180].

Briefly, concentrated VLP samples were incubated on a glow-discharged c-flat holey carbon grid (Ted Pella, Redding, CA), and blotted with filter paper to remove excess sample. Grids were quickly frozen into liquid ethane [184] with a FEI Vitrobot MarkIII system. A FEI TF30 field emission gun transmission electron microscope at liquid nitrogen temperature (FEI Company, Hillsboro, OR) was used to analyze frozen samples.

Imaging of samples was done at a nominal magnification of 59k at low-dose (~30 electrons/angstrom2) and 1 to 5 µm underfocus conditions using a Gatan 4k by 4k CCD camera (Gatan Inc, Pleasanton, CA). The fluorescence image and corresponding TEM picture was recorded using a FEI iCorr installed on a FEI Tecnai Spirit TEM.

3.2.3 Measurement of virus-like particle size

Two perpendicular diameter measurements were done for each VLP, as previously described [170, 180]. Cryo-TEM images were analyzed using ImageJ software

(NIH, Bethesda, MD). A histogram was generated for each VLP sample using GraphPad

53 Prism 6 software (GraphPad, La Jolla, CA) to determine the particle mean diameter and size distribution.

3.2.4 Fluorescence fluctuation spectroscopy, experimental setup and data analysis

VLPs were analyzed on an inverted microscope (AxioObserver, Zeiss,

Thornwood, NY) modified for two-photon FFS with an excitation wavelength of 1000 nm provided by a titanium-sapphire laser (MaiTai, Spectra Physics, Mountain View,

CA). FFS measurements were performed by focusing the laser inside the VLP solution and recording the fluorescence emission as the VLPs moved in and out of the observation volume via passive diffusion. Fluorescence emission was collected by a 63x water immersion objective (Zeiss, C-Apochromat, N.A.=1.2), separated from excitation light by a dichroic filter, and detected by a single-photon counting module. The photon counts were recorded with a sampling frequency of 20 kHz and stored for later analysis with routines written in IDL 8.3 (Research Systems, Boulder, CO).

FFS measurements recorded photon counts for ~30 minutes and contained approximately 700 to 1300 distinct VLP events. The photon counting histogram (PCH) was calculated for each data record and fit to determine the brightness of the VLP sample as previously described [101, 209]. The fit model includes background counts, two brightness species to describe VLP events, and accounts for deadtime and afterpulsing of the detector. The fit characterizes the VLPs by their effective brightness lVLP and particle occupation number N of the two brightness species. The normalized brightness b of the

VLP sample was defined by the ratio

b = ll VLP YFP (Equation 1)

54 which specifies the average number of Gag-YFP molecules per VLP. The reference brightness, lYFP, was obtained from a separate series of FFS measurements where the excitation power was systematically varied and the brightness of a purified YFP solution was determined for each excitation power. A representative PCH of purified YFP measured at 0.47 mW excitation power and its corresponding fit to a model with a single fluorescent species is excerpted from this calibration series and shown in Figure 3-1A.

This PCH fit recovered a brightness of 760 counts per second per molecule, an occupation number of 130 YFP and a goodness of fit (reduced chi-square) of 1.7. This occupation number corresponds to a concentration of ~1 µM, which is far below the reported Kd of YFP dimers (~100 µM) [210], ensuring that the brightness determined by this calibration represents the brightness of true YFP monomers. The monomeric nature of the enhanced green fluorescent protein (EGFP) and related fluorescent proteins for concentrations of 10 µM or less has been experimentally verified in several FFS experiments [211, 212].

The experimentally determined brightness values of the purified YFP solution were proportional to the squared excitation power as expected for two-photon microscopy (Figure 3-1B) [213]. Because each VLP typically contains several hundred

YFP labels, VLP samples must be measured at relatively low excitation power to avoid detector saturation. However, a direct measurement of the YFP monomeric brightness is technically difficult at these low powers [101]. Therefore, we used the linear fit of the brightness versus squared power to identify the reference brightness lYFP at the excitation power used for VLP measurements (dashed lines, Figure 3-1B). This procedure identified a YFP reference brightness of lYFP = 330/s. A second calibration method described in the

55 literature [101], which is based on the power-dependence of the fluorescence intensity, was applied as a consistency check and confirmed the value of the YFP reference brightness (data not shown).

t The diffusion time D of VLPs was determined from the autocorrelation function

(ACF) of the photon count record and converted to the diffusion coefficient D by

t = wD2 4 Do (Equation 2)

w with 0 representing the radial beam waist of the observation volume. The average diameter d of VLPs was obtained from the Stokes-Einstein relation

DkT= 3ph d B (Equation 3)

kT where h is the viscosity of water and B is the thermal energy at room temperature.

Measurements were again grouped by the transfection ratio, and the median diameter of

VLPs was calculated for each group. The standard deviation of the median diameter was determined by the bootstrap method.

3.2.5 Scanning transmission electron microscopy mass measurements of virus-like particles

The total mass of HTLV-1-like particles was determined (as previously shown

[208]) by using a custom-built quantitative dark-field STEM developed at the

Brookhaven National Laboratory (BNL, Upton, NY) [185]. Purified VLPs were mixed with tobacco mosaic virus (TMV) particles, which comprise rod-like structures of known linear mass-density (13.1 kDa/Å) and are commonly used as a standard for STEM mass measurements [185, 186, 214]. The sample was subsequently freeze-dried onto a thin-

56 carbon TEM grid and imaged on a low-temperature stage (-150ºC) by raster-scanning a

40-keV electron beam across the grid. The number of scattered electrons at each point within the sample was detected and recorded, yielding images with pixel intensity proportional to the mass-density at each location. The mass of individual VLPs was then determined by integrating the STEM image over each VLP and comparing against the observed mass-density of TMV particles in the same image. This comparison with TMV particles is conducted in order to ensure that mass perturbations that may be introduced during sample preparation and measurement are removed from the data collected on VLP mass [39,42,43]. For example, to account for TMV-mass changes, the raw mass values of the particles of interest is compared to the known TMV mass of 13.1 kDa/Å [100, 172].

Previous analyses have shown that TMV can gain up to 20% of mass during STEM analysis during sample preparation, but can be reduced via protocol modifications [215].

Exposure to the electron beam can also cause mass loss, and this can be reduced by using the low-dose technique [185, 215]. Contaminants (e.g., salts) can also contribute to a relative increase of TMV mass [216]. Following data collection, the PCMass software, developed by the BNL STEM facility, was used to analyze each virus particle’s total mass. GraphPad Prism 6 software (GraphPad, La Jolla, CA) was used to generate the resulting histograms and graphs of particle mass distributions.

3.2.6 Protein and RNA content of HTLV-1 virus-like particles

The total RNA content for each sample of VLPs was determined by extracting the

RNA from particle lysates with RNA columns, using Roche’s High Pure Viral RNA Kit

(Roche Diagnostics, Indianapolis, IN). The extracted RNA was detected by UV

57 absorption at 260 nm using a Beckman DU-65 spectrophotometer (Beckman Coulter,

Brea, CA). A Thermo Scientific Pierce BCA Protein Assay Kit (Thermo Fisher

Scientific, Waltham, MA) was used to estimate the protein content of the same sample used to determine VLP RNA content. A bovine serum albumin protein standard and measurement of UV absorption at 562 nm were used to establish an approximate absorbance-to-concentration curve and estimate the total protein concentration of VLPs samples.

3.3 Results

3.3.1 Morphology of HTLV-1-like particles

We recently reported that HTLV-1-like particles produced by expression of

HTLV-1 Gag have an electron dense ring of Gag along the inner leaflet of the viral membrane [208]. The average particle size (~110 nm) and estimated Gag copy number

(1300-1600) was found to be comparable to that of authentic HTLV-1 virions. In contrast, VLPs produced from expression of Gag-YFP were smaller (~73 nm), had a lower Gag copy number (~500), and lacked an electron dense ring indicative of a properly formed Gag lattice [180]. In order to study the impact of Gag-YFP on VLP morphology, we produced VLPs by transfecting HTLV-1 Gag and Gag-YFP at various ratios, ranging from 0:1 to 5:1 of Gag:Gag-YFP (Figure 3-2).

Correlative light and electron microscopy demonstrated that the fluorescence signal co-localizes with the distribution of YFP-labeled VLPs on a frozen-hydrated TEM grid (Figure 3-2A,B). Figure 3-2C-H shows the particle morphology of VLPs produced by transfecting Gag:Gag-YFP ratios of 0:1 to 5:1. There was no Gag lattice observed in

58 VLPs produced from Gag-YFP construct (Figure 3-2C) or Gag:Gag-YFP at a ratio of 1:1

(Figure 3-2D). VLPs produced by transfecting HTVL-1 Gag and Gag-YFP at 2:1 ratio

(Figure 3-2E) showed a stronger electron densities inside the viral membrane but missed clear lattice configuration. At higher Gag and Gag-YFP ratio (Figure 3-2F-H), the VLPs were found to be spherical and contain a lattice-like arrangement indistinguishable from that observed from VLPs produced by expression of untagged HTLV-1 Gag [208].

Figure 3-3 shows the distribution of particle sizes produced. Interestingly, the average size of the HTLV-1-like particles increased as the ratio of Gag to Gag-YFP increased

(i.e., 83 ± 21 nm to 113 ± 37 nm as illustrated in Figure 3-3). The particle size of

Gag:Gag-YFP at ratios of 2:1 to 5:1 (i.e., 110 ± 26 nm to 113 ± 37 nm as shown in Figure

3-3C-F), were comparable to those reported for VLPs produced from untagged HTLV-1

Gag [208].

3.3.2 Gag stoichiometry of HTLV-1-like particles determined by FFS

FFS was used to measure Gag stoichiometry and determine an average Gag copy number. The experiment records the photon counts of the fluorescence emitted as VLPs diffuse through the observation volume. A representative photon count trace of HTLV-1

Gag-YFP VLPs is shown in Figure 3-4A; each intensity spike corresponds to the passage of a single VLP. The photon counting histogram (PCH, Figure 3-4B) corresponds to the probability distribution function of all photon counts calculated from a 30-minute photon count trace of HTLV-1 Gag-YFP VLPs. The VLP signal and the background signal are identified by a fit of the PCH (solid line, Figure 3-4B). Background counts (dashed line) dominate the left side of the histogram, while VLP events lead to high photon counts

59 (right side of the histogram). The average brightness of the VLPs determined from the

PCH fit allows estimation of the average Gag-YFP stoichiometry of VLPs by comparison to the brightness of a YFP monomer. This PCH fit identified an average brightness of

120,000 counts per second per VLP. The Gag-YFP copy number of this sample was then estimated by the ratio of the VLP brightness to the YFP monomer brightness (see

Materials and Methods section) as 120,000/330 = 360. Repeated measurements (n = 9) of

HTLV-1 Gag-YFP VLP samples resulted in a copy number of 370 ± 50.

VLPs were then harvested and analyzed from cells transfected with HTLV-1 Gag and HTLV-1 Gag-YFP at a ratio of 1:1. Copy number determinations by FFS represent only the Gag-YFP population packaged into the VLPs. Therefore, we estimated the total

Gag copy number by using the assumption that the ratio of Gag to Gag-YFP within the

VLPs is equivalent to the transfection ratio, resulting in a total copy number of 730±130

Gag molecules per VLP. FFS measurements of VLP samples with 2:1 and 3:1 transfection ratios of Gag:Gag-YFP were then analyzed by PCH and the total Gag stoichiometry was inferred from the transfection ratio to be 990±130 and 1600±530 molecules of Gag per VLP, respectively. The FFS experiments identified an increase in the total Gag copy number as a function of the Gag:Gag-YFP ratio (Figure 3-5A), which correlated well with the increase in VLP size as determined by cryo-TEM (Figure 3-3).

3.3.3 VLP diameter as determined by FFS

In addition to PCH analysis, we also investigated the autocorrelation function

t (ACF) of the photon counts to estimate the size of VLPs by FFS. The diffusion time D

g t of a VLP sample was determined as the time when the ACF ( ) decayed to half its

60 original amplitude (Figure 3-4C). The diffusion time was first converted to a diffusion coefficient and subsequently to a particle diameter by using the Stokes-Einstein relation

(equation 3). This analysis was conducted on VLP samples produced from cells transfected with HTLV-1 Gag and Gag-YFP plasmid at ratios of 0:1, 1:1, 2:1, and 3:1, respectively. We determined the median diameter and its standard deviation from several measurements of VLP samples with the same ratio of labeled and unlabeled Gag. The median diameter of VLPs produced by expression in 293T cells of HTLV-1 Gag-YFP was 75 nm. Increasing the ratio of unlabeled to labeled Gag led to an increase in the particle diameter of ~110 nm for the 2:1 and 3:1 ratios of Gag:Gag-YFP as determined by cryo-TEM analysis (Figures 3-3 and 3-5B). The values determined by FFS are in good agreement with the particle size determined by cryo-TEM (Figures 3-3 and 3-5B).

3.3.4 STEM analysis of HTLV-1-like particle mass

STEM analysis was used to determine the total mass of HTLV-1-like particles as previously described [185], which allowed for the determination of the average Gag copy number per particle. Representative dark-field electron micrographs of VLPs produced by expression of HTLV-1 Gag-YFP and that of VLPs produced by co-expression of a 3:1 ratio of Gag:Gag-YFP are shown in Figures 3-6A and 3-6D, respectively. Only isolated and intact particles that were in the range of particle sizes observed by cryo-EM were used for VLP mass measurements. Small contaminants are visible in the background; however, these did not affect the mass measurements based on the measured mass of

TMV particles that were used as an internal control. The analyzed VLPs had a wide mass variability (Figures 3-6B and 3-6E), which correlates well with the VLP size range

61 observed by using cryo-TEM (Figure 3-3). We determined that there was no more than

4% increase or decrease in TMV mass (13.1 kDa Å-1) (Figures 3-6C and 3-6F). By assuming that similar mass changes exist with the HTLV-1-like particles, the TMV- corrected particle mass measurements for VLPs produced by HTLV-1 Gag-YFP (Figure

3-6C) and by HTLV-1 Gag:Gag-YFP at a 3:1 ratio (Figure 3-6F) were determined to be

80±53 MDa and 219±89 MDa, respectively.

3.3.5 STEM determination of Gag stoichiometry in HTLV-1-like particles

To confirm the FFS measurements of Gag stoichiometry, we also conducted measurements using STEM. The total mass measurements of VLPs produced by expression of HTLV-1 Gag-YFP or HTLV-1 Gag:Gag-YFP particles at a ratio of 3:1 were used to estimate the average Gag copy number per VLP. The Gag:RNA mass ratio was determined to be 19.8:1 and 18.3:1 for VLPs produced with HTLV-1 Gag:Gag-YFP at a 3:1 ratio and HTLV-1 Gag-YFP, respectively, which yielded an RNA content of approximately 3% of the total VLP mass (Table 3-1).

The average VLP diameter of 83 nm and 113 nm for VLPs produced by expression in 293T cells of HTLV-1 Gag-YFP and HTLV-1 Gag:Gag-YFP at a ratio of

3:1, respectively, and the 5nm thickness of the viral membrane were used to estimate the number of lipid molecules in the VLP membrane. The average molecular weight of membrane lipids was estimated to be 750 Da, and it was assumed that the distance between lipid molecules is 0.85 nm [193]. Given these lipid estimates, the mass of lipids in an averaged size particle was estimated to be 36 MDa and 80 MDa for VLPs produced with HTLV-1 Gag-YFP and HTLV-1 Gag:Gag-YFP at a 3:1 ratio, respectively (Table 3-

62 1). Based upon the assumption that the mass contribution of the Gag protein to the total

VLP protein mass is similar to that of other retroviruses (i.e., in the range of 70%-90%)

[102, 194], the mass of Gag-YFP in VLPs of average size and mass produced by expression of Gag-YFP was determined to be 29-38 MDa. Similarly, the mass of Gag and

Gag-YFP in VLPs of average size and mass produced by expression of HTLV-1

Gag:Gag-YFP at a 3:1 ratio was determined to be 93-120 MDa (Table 3-1). VLPs of average size and mass produced by expression of Gag-YFP were calculated to contain approximately 360 to 480 Gag copies per VLP (Table 3-1), while VLPs of average size and mass produced by Gag:Gag-YFP at a 3:1 ratio contained approximately 1600 to 2100 copies of Gag per VLP (Table 3-1). A comparison of the Gag copy number as determined by STEM analysis at HTLV-1 Gag:Gag-YFP expression plasmid ratios of 0:1 and 3:1

(Table 3-1) indicates that the STEM and FFS measurements are in good agreement, which provides an independent confirmation of the Gag copy number determination.

Taken together, elevation of the Gag ratio led to similar observations. In contrast, increasing the Gag-YFP ratio resulted in smaller VLPs, fewer Gag molecules per VLP, as well as an altered particle morphology (Figure 3-7).

3.4 Discussion

While HTLV-1 was the first human retrovirus discovered 35 years ago [1, 197], the details of viral replication – particularly the steps involved in virus particle assembly

– are poorly understood. HTLV-1 is notorious for being difficult to grow in cell culture, and the proviral sequences are prone to recombination. Given these technical limitations, knowledge of HTLV-1 replication has been lacking. We previously devised a tractable

63 model system for the robust expression of the HTLV-1 Gag protein in mammalian cells, which resulted in efficient production of HTLV-1-like particles that have a morphology comparable to that of immature HTLV-1 particles [208]. In this study, we sought to use this model system investigate whether HTLV-1-like particle morphology would be perturbed as a result of co-packaging of Gag and Gag fused to a fluorescent protein (i.e.,

YFP). Gag co-packaging is biologically relevant in the retrovirus assembly pathway as

HTLV-1 and retrovirus particles naturally co-package Gag and Gag-Pol (a Gag protein with the Pol protein fused to the carboxy-terminus of Gag), reviewed in [170, 217]. Thus, the insights gained in these studies have relevance to the assembly of authentic HTLV-1 particles.

HTLV-1-like particles were produced by transient transfection into 293T cells of a HTLV-1 Gag expression plasmid, a HTLV-1 Gag-YFP expression plasmid, or by co- transfection of these two plasmids at various ratios. VLP production was highly robust, due in part to the fact that the human-codon optimized versions of the Gag gene were used. We previously found that the HTLV-1 Gag-YFP construct can be used to produce

VLPs [91, 180]. Interestingly, in contrast to studies with HIV-1 [101], the YFP-tag was observed to affect the size and Gag stoichiometry of HTLV-1-like particles [192, 208].

FFS (Figure 3-5A) and STEM (Table 3-1) measurements of VLPs produced from

HTLV-1 Gag or a series of mixed transfection ratios of Gag:Gag-YFP showed that when the Gag-YFP was less than 25% of the total Gag population (3:1 ratio of Gag to Gag-

YFP), the average Gag copy number of 1600 to 2100 (Table 3-1), consistent with average

Gag copy numbers observed in VLPs produced by transfection of 293T cells with HTLV-

1 Gag alone [208]. The Gag copy number determinations from the STEM analysis relied

64 on the assumption that regardless of the particle size, the RNA and protein content as well as the Gag:lipid ratio by mass remained constant among the VLP sample population.

Given the diversity of retrovirus particle populations, these ratios are difficult to directly determine experimentally. Additionally, cryo-TEM analysis found that the VLPs produced from the co-packaging of HTLV-1 Gag:Gag-YFP, at a 3:1 ratio, had a morphology and size that closely resembled VLPs produced by transient expression of

HTLV-1 Gag in 293T cells [208]. The VLPs produced by co-packaging of Gag and Gag-

YFP were found to be spherical, had a Gag lattice that follows the inner leaflet of the virus membrane (Figure 3-2D), and were about 113nm in diameter (Figure 3-3D).

Furthermore, FFS analysis independently confirmed the VLP size and Gag stoichiometry measurements made by cryo-TEM and STEM. In particular, FFS analysis indicated that

HTLV-1 Gag-YFP was incorporated into VLPs in a predictable manner at the 3:1 ratio.

Based upon FFS, the VLPs were found to be approximately 110nm in diameter (Figure 3-

5B), and contained about 1600 copies of Gag (Figure 3-5A).

3.5 Conclusions

The findings of this study support the conclusion that biologically relevant Gag-

Gag interactions occur between Gag and Gag-YFP when they are at a ratio of 3:1

(Gag:Gag-YFP). Increasing the amount of Gag led to similar observations, while increasing the amount of Gag-YFP resulted in small VLPs, fewer Gag molecules per

VLP, and an altered particle morphology (Figure 3-7).

These findings are important for the analysis of Gag-YFP in living cells, particularly regarding the study of the processes involved in Gag oligomerization and

65 Gag lattice formation. Furthermore, these studies are important for understanding the natural interactions and co-packaging between Gag and Gag-Pol, which is critical to ensure that the viral enzymes are packaged into virus particles. Proper formation of the immature Gag lattice is likely critical in order to ensure efficient Gag processing by the viral protease and formation of mature, infectious virus particles. Therefore, the ratio of co-packaging Gag and Gag-Pol is critical in order to ensure proper Gag stoichiometry as well as immature particle morphology (and likely particle infectivity). These observations support previous reports of the importance of Gag to Gag-Pol ratios in retroviral assembly [218, 219]. Given the extensive use of labeled Gag proteins for studies of retroviral assembly, our observations provide important insights into the impact of Gag co-packaging on Gag stoichiometry and particle morphology.

66 A B

Figure 3-1. Calibration of YFP molecular brightness for measurements of Gag copy number by fluorescence fluctuation spectroscopy. (A) The photon counting histogram

(hexagon symbols) for YFP measured at 0.47 mW excitation power is plotted together with a single species fit (solid line) with a reduced chi-square of 1.7. (B) A graph of the molecular brightness of YFP versus excitation power squared is fit to a linear relation

(solid line). Individual brightness values are determined by FFS measurements of YFP at differing excitation powers. The reference brightness of YFP for the VLP experiment is indicated by the dashed line.

67

Figure 3-2. Images of HTLV-1-like particles as determined by fluorescence microscopy and cryogenic transmission electron microscopy. (A) Image recorded by using a FEI iCorr showing fluorescence signal, in green, of HTLV-1-like particles with

Gag:Gag-YFP ratio of 3:1 on a frozen TEM grid. The I and H labels indicate areas of the grid that only has vitrified ice of frozen buffers (I), and an empty hole on the lacey carbon film (H), respectively, as observed in both panels A and B. The scale bar is also applicable to both A and B. (B) The TEM image of the grid at the same location of A, showing distribution of frozen hydrated HTLV-1-like particles (P). (C-H) Representative cryo-TEM images of HTLV-1-like particles showing size and morphology VLPs produced by transfection of HTLV-1 Gag expression plasmid(s) at 0:1 (C), 1:1 (D), 2:1

(E), 3:1 (F), 4:1 (G) and 5:1 (H) ratio of HTLV-1 Gag to HTLV-1 Gag-YFP respectively. All HTLV-1-like particles contained an electron dense Gag lattice with gaps except for the fully labeled Gag (C), which is consistent with previous observation [180].

The scale bar shown is equivalent to 100 nm.

68

Figure 3-3. Analysis HTLV-1-like particle diameter by cryogenic transmission electron microscopy. Particle size distribution for VLPs produced by transfection of Gag expression plasmids into 293T cells at a HTLV-1 Gag to Gag-YFP ratio of 0:1 (A), 1:1

(B), 2:1 (C), 3:1 (D), 4:1 (E), and 5:1 (F).

69 A B

C

Figure 3-4. Analysis of HTLV-1-like particles by fluorescence fluctuation spectroscopy. (A) Shown is a two-minute excerpt from a 30-minute fluorescence intensity trace. Intensities are displayed as photon counts per second. (B) The experimental photon counting histogram of the fluorescence intensity trace determines the probability distribution of the photon counts. Photon counting histogram analysis determines particle brightness and occupation number by a fit of this probability distribution. (C) This fit results in a reduced chi-square of 1.7. The experimental

70 autocorrelation function (ACF) of the fluorescence intensity trace serves to identify the

t diffusion time D as indicated by the dashed lines.

71 A B

Figure 3-5. Determination of HTLV-1-like particle Gag copy number and particle diameter. (A) Average HTLV-1 Gag copy number and (B) VLP diameter as determined by fluorescence fluctuation spectroscopy and cryogenic transmission electron microscopy measurements.

72 A D

B E

C F

Figure 3-6. Scanning transmission electron microscopy analysis of virus-like particles produced from expression of HTLV-1 Gag-YFP and Gag:Gag-YFP (3:1).

(A) Micrograph from scanning transmission electron microscopy of HTLV-1 Gag-YFP particles mixed with TMV. (B) Total mass measurement distribution in MDa of HTLV-1

Gag-YFP particles. (C) Mass distribution per grid analyzed of HTLV-1 Gag-YFP

73 particles. (D) Micrograph from scanning transmission electron microscopy of HTLV-1

Gag:Gag-YFP (3:1) particles mixed with TMV. (E) Total mass measurement distribution in MDa of HTLV-1 Gag:Gag-YFP (3:1) particles. (F) Mass distribution per grid analyzed of HTLV-1 Gag:Gag-YFP (3:1) particles. Intact HTLV-1-like particles and TMV rods are identified by arrows for both scanning transmission electron microscopy micrographs.

Corrected mass of virus-like particles is indicated in bold, which was determined based upon the known TMV mass per unit length of 13.1 kDa Å-1. Standard deviation is indicated from 3 measurements.

74

Figure 3-7. Relationship between virus-like particle size and Gag copy number to that of HTLV-1 Gag-YFP copy number. A summary of the general findings of this study are indicated with increasing ratio of HTLV-1 Gag to HTLV-1 Gag-YFP, resulting in particle diameters and Gag copy numbers that are comparable to that of particles produced with only unlabeled HTLV-1 Gag.

75 Table 3-1. Summary of HTLV-1-like particle mass determination and calculation of

Gag copy number per particle.

HTLV-1 Like Particle

Measurement Gag-YFP Gag:Gag-YFP (3:1) Average Diameter (nm)a 83 113 Average Particle Mass (MDa)b 80 219

c

RNA 2.4 6.6 Lipidd 36 80 (MDa) Mass of e and Protein Protein and RNA, Lipid RNA, Lipid Total protein 42 133

f

Total Gag polyprotein 29 - 38 93 – 120 Gag - 70 – 90 (MDa)

Molecules Molecules Gag-YFP 29 - 38 23 – 30 Mass of Gag

Gag polyprotein copy numberg 360 - 480 1600 – 2100

a As determined by cryogenic transmission electron microscopy. b As determined by scanning transmission electron microscopy. c Mass contributed by RNA in virus-like particles was estimated experimentally as described in the Materials and Methods. d Mass contributed by lipids was estimated from average particle size and membrane thickness. e Mass of total protein was determined by subtraction of the RNA and lipid mass from the total particle mass as determined by scanning transmission electron microscopy. f Total Gag polyprotein was estimated based upon the assumption that Gag contributes ~

70-90% of the total protein mass.

76 g The Gag polyprotein copy number represents the range of Gag copy number in a particle that has both average mass and dimensions.

77

CHAPTER IV

ANALYSIS OF HUMAN T-CELL LEUKEMIA VIRUS TYPE 1 PARTICLES

USING CRYO-ELECTRON TOMOGRAPHY

The text presented in this chapter has been published:

Cao, S., Maldonado, J. O., Grigsby, I. F., Mansky, L. M., & Zhang, W. (2015). Analysis of human T-cell leukemia virus type 1 particles by using cryo-electron tomography. Journal of Virology, 89(4), 2430-2435. doi: 10.1128/JVI.02358-14

78 4.1 Text

Human T-cell leukemia virus type 1 (HTLV-1) is a human cancer-causing retrovirus that causes an adult T-cell leukemia and HTLV-1-associated myelopathy/tropical spastic paraparesis [168, 169]. It is generally accepted that cell-cell contact is indispensible for the infection of HTLV-1 [220]; however, cell-free infection through dendritic cells also contributes to viral transmission and pathogenesis [54].

HTLV-1 particle assembly and maturation involve the budding of immature virus particles from the plasma membrane (PM), followed by the triggering of virally mediated proteolysis and particle maturation [206, 221]. Similar to other retroviruses, the HTLV-1

Gag protein has three major functional domains: matrix (MA), capsid (CA), nucleocapsid

(NC). The MA binds directly to the inner leaflet of the cell membrane, which is critical for particle assembly; the CA protein is necessary for Gag-Gag interactions that form the

Gag lattice in immature particles and is also required in forming the viral capsid core in mature particles [170]; and the NC protein interacts and coats the viral RNA genome.

During virus maturation, Gag is cleaved by the viral protease, after which CA proteins reassemble into a capsid core that encapsulates the NC-RNA complexes and the viral replication enzymes.

Studies using cryo-electron microscopy (cryo-EM) and cryo-electron tomography

(cryo-ET) have demonstrated that retrovirus cores are highly polymorphic. The cores in human immunodeficiency virus type 1 (HIV-1) are predominately conical in shape [222,

223], whereas those in Rous sarcoma virus (RSV) are either irregularly polyhedral or spherical [174, 224]. Tubular cores have also been observed for both HIV-1 and RSV.

Despite the morphological diversity among retroviral cores, the fundamental principles

79 by which the CA proteins are organized are similar. For example, the amino-terminal domain (NTD) of CA mediates formation of hexameric or pentameric rings on the exterior of the capsid, while the carboxy-terminal domain (CTD) links the rings together through dimerization at the floor of the capsid. In addition to variation in the dihedral angles between neighboring CA hexamers [224, 225], closely packed hexamers are interspersed with CA pentamers to mediate larger changes in curvature [224, 226]. The correct formation of cores is likely related to their function in replication and the infectivity of the virus particles [227-229].

Despite its significant impact on human health, many of the details of its particle assembly and structure remain poorly understood. Previous structural analyses of HTLV-

1 assembly have been largely limited to observing HTLV-1-infected cells by using thin- section transmission electron microscopy (TEM). In those studies, HTLV-1 particles produced from chronically infected MT-2 cells [230] appeared to be assembly intermediates, with electron density inside particles forming either a half-donut shape

[180, 231] or complete donut-shape [230] along the inside of the viral membrane. Virus particles were observed that had a distinct capsid core or had electron densities evenly distributed in the interior of the particle in the absence of a core structure. In the current study, we purified HTLV-1 particles from MT-2 cells and employed cryo-EM and cryo-

ET methods to determine the morphology of HTLV-1 virions in a frozen-hydrated state.

A two-step procedure was used for the purification of HTLV-1 particles. First,

MT-2 cells were cultured in 50 ml of RPMI 1640 medium supplemented with 10% Fetal

Clone III for about 1 week or until the cells reached ~90% confluency (as determined by the formation of cell clusters). Two new flasks were then each inoculated with ~5% of

80 the suspended cells from the first expansion cycle and filled with 25 ml of culture medium supernatant from the previous expansion cycle and 25 ml of fresh medium. The cultures were then grown for an additional week, after which HTLV-1 particles were purified from the culture medium supernatant and concentrated using a procedure modified from that of Grigsby et al. [180, 231]. After removing large cellular debris by centrifugation and filtration, the sample was pelleted using an 8% OptiPrep (Sigma-

Aldrich) cushion and applied to a 10%-to-50% OptiPrep gradient for centrifugation. The virus-containing fraction was extracted, pelleted, and resuspended in STE buffer (10 mM

Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA). Previous immunoblot analyses of

HTLV-1 particles produced from MT-2 cells found both processed and non-processed forms of HTLV-1 Gag [231, 232]. HTLV-1 particles purified from MT-2 cells were found to be infectious but with an infectivity that was about 1,000-fold lower than that of

HIV-1. Prior to cryo-EM sample preparation, 2,2'-dithiodipyridine (AT-2) was used to inactivate infectious HTLV-1 particles. Previous studies with HIV-1 have indicated that

AT-2 covalently modifies the NC domain zinc fingers, which results in the blocking of reverse transcription [183].

For tomography data collection, the virus sample was mixed with 10-nm colloidal gold particles treated with bovine serum albumin solution (Sigma-Aldrich) and then vitrified on 200-mesh Quantifoil grids (Ted Pella, Inc.) in liquid ethane. An FEI F30 transmission electron microscope was used to record a single-axis tilt series from -60° to

+60° in 3° incremental steps. Data were collected automatically using SerialEM [233] on a Gatan Ultrascan 4000 charge-coupled device (CCD) camera at 39,000´ nominal magnification. The defocus level of the image at the 0° tilt angle in each tilt series was

81 determined to be ~13 µm based on the first zero of the contrast transfer function shown in

RobEM [234]. The tomograms, reconstructed in IMOD [235], were bin averaged so that the dimension of each voxel in the tomogram corresponded to 0.62 nm in the specimen.

The final reconstruction maps were denoised using nonlinear anisotropic diffusion implemented in IMOD. We used 30 tomographic reconstructions for further analysis.

The HTLV-1 virions embedded in vitreous ice were roughly spherical and vary in size (Figure 4-1). About 15% of the particles were found to contain an interior core-like structure that had a distinct ~5-nm-thick density layer corresponding to the viral capsid shell. These particles appeared to best represent mature HTLV-1 particles with complete cores. Other density-filled particles (~60%) were observed to contain evenly distributed electron densities inside the viral membrane but no density attributable to a capsid core.

While the nature of these HTLV-1 particles is not entirely clear, they likely represent particles that lack viral infectivity, as they do not possess a complete (mature) capsid core. Assembly defects may have contributed to the morphology of some or perhaps most of these particles, although further analyses are needed to better characterize these particles. Approximately 5% of the particles observed contained partially mature capsid cores in which the shell of the capsid was not closed, and the electron densities at the open region appear similar to those in the frequently observed particles that were density- filled. Single HTLV-1 particles that contained either conically shaped cores or double cores were rare, and particles with tubular cores were not observed. Densities that resembled small knob-like structures projecting away from the viral membrane were observed (Figure 4-2E). However, the limit of resolution in the tomographic data set did

82 not allow the determination of whether these structures represent the HTLV-1 envelope protein.

The average diameter of the mature and partially mature HTLV-1 virions, as measured from the outer margins of the viral membrane in 83 particles, was determined to be 113 ± 23 nm (mean ± standard deviation), which is smaller than the diameters of mature RSV (~125 nm) [174] and HIV-1 (~145 nm) [178, 222] particles. The HTLV-1 particles are larger than the virus-like particles previously produced from 293T cells by an HTLV-1 Gag expression construct with yellow fluorescence protein (YFP) fused to the carboxy-terminus [180], suggesting that YFP influenced the particle diameter. The capsid cores in HTLV-1 virions have a polyhedron-like structure. In the 2-dimensional

(2-D) slices of the core-containing particles (Figure 4-2), linear segments of the capsid shell were either jointed at a sharp angle or connected by a continuous density outline with a smooth curvature. In addition, we occasionally observed spherical cores (Figure 4-

1). The HTLV-1 cores tended to adjoin the boundary of the viral envelope, and there was at least one curved region or vertex in each capsid that touched the inner face of the viral membrane. This membrane-proximal region is situated ~13 nm away from the viral envelope, which is similar to the case in HIV-1, where a 12-nm space was observed between the viral membrane and the broad end of the core [172]. We observed that four particles had cores with regions of double-layered structure; that is, a second capsid sheet of the same thickness was stacked on top of the primary capsid shell (Figure 4-2D).

Capsids with double-layered regions have also been observed in HIV-1 [172, 222] and

RSV [174, 225]. In HTLV-1, the stacked shell sheets observed in the four particles appear to be flat (Figure 4-2D). The average electron density inside the complete cores of

83 HTLV-1 virions is very similar to that in the region between the capsid and viral membrane, suggesting that the packing densities of viral or cellular components are similar in these two compartments.

The average size of the particles with evenly distributed electron densities, as measured from the central sections of 156 particles in the tomograms, was determined to be 111 ± 20 nm. For these particles, we did not observe density features attributable to the capsid protein shell, as seen in the mature HTLV-1 virion. We also did not see density attributable to the immature Gag-Gag lattice structure, as seen in HIV-1 mutant samples subjected to partial Gag cleavage [236] or in samples treated with maturation inhibitor bevirimat [237]. Furthermore, we observed particles with incomplete formation of the capsid shell, which suggests that the Gag proteins in these particles may not have been properly assembled when budding out of the cell or that the CA protein failed to reorganize into a complete capsid core. It has been reported that MT-2 cells contain one complete provirus and seven defective proviral sequences [238]. A 3.4-kb RNA transcript of the defective proviruses expresses a myristylated truncated Gag protein that is composed of MA, a truncated CA, a short pX region, and two long terminal repeats

[239]. The 3.4-kb RNA transcript and the truncated Gag proteins have previously been found packaged into virus particles produced from MT-2 cells [232]. Since the truncated

Gag has only about 30% CA, it is likely that after protease cleavage, the truncated CA does not form proper capsid cores. The assembly of defective Gag may contribute to particles with the evenly distributed internal density or those particles with partially closed capsid cores. A long-held belief about HTLV-1 particles that is in contrast with data for the HIV-1 and RSV is that cell-free particles are poorly infectious [220]. For

84 example, it has been previously reported that a number of T cell lines [240], B-cell lines

[241], and nonlymphoid cells [242] could be infected with cell-free HTLV-1 particles produced from MT-2 cells but only at very low levels. We found that most of the virus particles analyzed from MT-2 cells had evenly distributed electron densities or partially mature particles instead of mature capsid cores. The low frequency of HTLV-1 particles with mature cores, along with the encapsulation of defective RNA transcripts, could help to explain the low infectivity associated with cell-free HTLV-1 particles produced from

MT-2 cells in vitro.

Intriguingly, we observed in about 10% of the particles analyzed (both mature particles and those with evenly distributed electron densities) a turret structure associated with the viral membrane (Figure 4-3A,B). In some instances, this organization was observed between two particles. In the cross section, these turret structures are rectangular or trapezoidal in shape. The wider end of the structure is 30-60 nm in length and proximal to the associated viral membrane, while the height of the structure is 10-40 nm. We detected no correlation between the size of the turret structure and the size of the virus particle. The boundary of the turret has an electron density as strong as that of the viral envelope. In the tomograms, the turret structures have a filled interior with a density greater than that corresponding to the region inside a viral core. A 2-D cryo-EM image of the turret revealed juxtaposed density stacks with ~3.5 nm in between each density stack

(Figure 4-3C), which is similar to the distance between the cholesterol layers in the cryo-

EM structure of human low-density lipoprotein [243]. The turret structure does not resemble the previously reported in vitro-assembled endosomal-sorting complex required for transport III (ESCRT-III), which forms spiral [244, 245] or a hollow cylindrical [244]

85 structures. A turret structure such as that observed here with HTLV-1 particles from MT-

2 cells has not been previously reported in other HTLV-1 particle analyses with TEM, nor did we observe such a structure in our previous cryo-EM study with HTLV-like particles produced from 293T cells using an HTLV-1 Gag-only expression construct.

Also, such a turret structure has not been observed with other retroviruses. It is formally possible that the appearance of the turret structure is cell-type dependent. The molecular composition, origin, and biological relevance of these turret structures in HTLV-1 replication are currently not known but represent a clear focus for further investigations.

This study has found that the CA cores of HTLV-1 particles are neither conical, as are those in HIV-1 [222], nor polyhedral, as are the angular cores in RSV [174].

Instead, the cores of HTLV-1 were found to be generally polyhedral in shape, with regions of smooth surface. It is possible that hexameric CA proteins are located in the flat facets, such as the double-layered regions of the cores, and that pentameric CAs are located at the corners of irregular polyhedrons. Both CA pentamers and hexamers tilted out of the plane [224, 225] may contribute to the curvature of the capsid shell, which has a continuously curving surface. The malleability of HTLV-1 particle cores may reflect the flexibility of the linkers that tether the NTD and CTD of the CA proteins. Indeed, the solution structure of the full-length CA protein from HTLV-1 indicates that these two domains behave independently, without a preference for a particular relative orientation

[246], which suggests that the NTD and CTD of HTLV-1 CA molecules are able to dynamically adopt relative orientations within the assembled capsid core. Higher resolution structures of the HTLV-1 Gag-lattice in the immature particle and CA lattice

86 in the mature particle will help to establish the molecular organization of immature and mature HTLV-1 particles and provide insight into HTLV-1 assembly and maturation.

87

Figure 4-1. Morphology of HTLV-1 particles as determined by cryo-ET. (A) A section in a tomogram showing authentic HTLV-1 particles produced and purified from

MT-2 cells. White arrows point to particles with evenly distributed electron densities that do not have discernable cores. Black arrows point to virus particles with mature cores.

White arrowheads point to particles that have partially complete cores. Particle 3 has a capsid core with roughly spherical morphology. (B) Surface-shaded models of the four particles indicated by number in panel A, showing the viral envelope (cyan), capsid shell

(red), genome and/or mixture of proteins and nucleic acids derived from the host cell

88 (beige), and densities between the capsid and viral membrane (beige). The image was created using Chimera software [247].

89

Figure 4-2. Tomographic slices of core-containing HTLV-1 particles. (A-B) Small and large particles containing complete (mature) cores; note that there are curling edges in the capsid shell adjoining the viral membrane. (C) Particles with incomplete cores. (D)

Particles with partially double-layered cores. (E) Particles with prominent spike structures (arrowheads) on the viral envelope. (F) Particles with external turret structure

(arrows) on the viral envelope. The scale bar is 100 nm.

90

Figure 4-3. Turret structures on HTLV-1 particles. (A-B) Tomographic slices highlighting four particles with external structures (black arrow-heads). (C) One particle with a turret structure on its surface. (D) Enlarged view of the turret structure showing layers of strong density separated by gaps of ~3.5 nm.

91

CHAPTER V

DISSERTATION SUMMARY AND FINAL DISCUSSION

92 The research presented in this dissertation has led to new insights into the nature of HTLV-1 particle morphology, as well as to define experimental parameters for the study of HTLV-1 particle assembly. The morphology of authentic HTLV-1 virions was investigated, for the first time, in a frozen hydrated state, as well as that of HTLV-1-like particles. Due to the unstable nature of the HTLV-1 proviral genome, the difficulties to replicate the virus in tissue culture, as well as the lack of a good HTLV-1 protease inhibitor, an HTLV-1-like particle system was used to study the morphology of immature virus particles as shown in Chapter II. Additionally, HTLV-1 Gag stoichiometry was also determined, for the first time, in both authentic particles and HTLV-1-like particles.

These studies found that VLPs and authentic particles had comparable particle shapes and sizes. The HTLV-1-like particles were found to be spherical in shape with a size of

~110nm. These particles also had a thick layer of electron density adjacent to the inner leaflet of the viral membrane that corresponds to the Gag lattice of immature retroviral particles. This, discontinuous, railroad track-looking lattice structure, for the most part, followed the curvature of the virus membrane, except for certain flat regions where the

Gag lattice did not follow the curvature of the membrane. These characteristics have not previously been described for any other immature retroviral particles. Likewise, authentic particles were primarily spherical and measured ~113nm.

STEM analysis was performed to determine the total particle mass of both VLPs and authentic particles, which in turn allows for the calculation of Gag copy number.

These results indicate that HTLV-1-like particles contain 1300-1900 molecules of Gag, while authentic particles were found to contain 1500-1900 copies (Table 5-1). These values are in the range of those observed with other retroviruses. The studies presented in

93 Chapter II show that VLPs and authentic particles have comparable particle diameters as well as Gag copy numbers. These studies also demonstrated that HTLV-1 particles possess a consistent size and Gag stoichiometry (Table 5-1), as well as to confirm that the

Gag-only VLP model system is appropriate surrogate for morphological studies of immature HTLV-1 particles.

In Chapter III, the effect of a HTLV-1 Gag-YFP fusion protein on particle morphology via copackaging with HTLV-1 Gag was studied. Other fluorophores such as

ReAsH, green fluorescent protein, and mCherry were tried with and without a flexible or rigid protein linker between the fluorophore and the Gag protein to produce VLPs, however, none of them resulted in reproducible production of VLPs (data not shown).

The rationale for this study was two-fold. First, the Gag-fluorophore protein has been used extensively for many studies of HTLV-1 Gag trafficking and particle assembly studies. Second, the copackaging of Gag and Gag-YFP serves as a surrogate for Gag and

Gag-Pol copackaging. Gag and Gag-Pol copackaging is critical for infectious virus spread. Previous attempts to produce VLPs using authentic HTLV-1 Gag and Gag-Pol were unsuccessful, the yield of particle production was too low for cryo-TEM analysis

(data not shown). The studies in Chapter III describe the results from multiple biophysical techniques showing that Gag/Gag-YFP ratios of 3 to 1 or greater result in the production of VLPs that were morphologically indistinguishable to that of Gag-only

VLPs as shown in Chapter II. In particular, cryo-TEM analysis revealed that particles produced with Gag/Gag-YFP ratios of 3 to 1 or greater were spherical and commonly possessed a Gag lattice with a non-continuous railroad rail-like morphology. Both cryo-

TEM and FFS analysis showed that the size of the particles increases to levels similar to

94 those of wt VLPs as the ratio of Gag-YFP to Gag increased. FFS analysis demonstrated that the Gag copy number increased as particle size increased. Particles produced from cells expressing Gag/Gag-YFP proteins at a 3 to 1 ratio were comparable to Gag-only

VLPs in both morphology (with a diameter of ~113nm) and Gag copy number (~1700)

(Table 5-1). STEM analysis, demonstrated that the Gag copy number in particles produced from cells expressing Gag/Gag-YFP (3:1) contained 1600-2100 Gag molecules per particle with a mean size of 113nm (Table 5-1). These results were found to be in general agreement with the FFS measurements, this is particularly important for morphological studies that do not use a fluorophore. Together, these results show that authentic HTLV-1 particles, HTLV-1-like particles, and HTLV-1-like particles expressing Gag/Gag-YFP at a 3 to 1 ratio have similar particle size and Gag copy number

(Table 5-1). Similar results have been reported with other retroviruses [100-103], however, no studies to date have performed a multi-faceted comparative study as reported in this dissertation.

Cryo-TEM/ET and authentic HTLV-1 particles collected from MT-2 cells were used to complete the studies presented in Chapter IV. Authentic HTLV-1 virions produced by MT-2 cells were found to be polymorphic and roughly spherical, with about

15% of them containing a complete polyhedron-like CA core. About 5% of the authentic particles contained a partially mature CA core, while the rest of them lacked a discernable complete CA core. The particles lacking complete cores could help explain why cell-free HTLV-1 virions have relatively low infectivity. However, further studies need to be done to understand the prevalence of particles with incomplete particle cores.

It was also found that about 10% of the authentic particles had a turret structure

95 associated with the viral membrane, which size did not correlate with that of the particle size. It was also noted that the internal content of this structure is similar to that observed in the cholesterol layers of the human low-density lipoprotein. This turret structure has not been associated with other retroviruses, which suggests that it could be a cell- dependent phenomenon.

Considering that the MT-2 cell line harbors truncated HTLV-1 proviruses and expresses aberrant forms of the Gag protein, a recent study from our research group analyzed the morphology of authentic HTLV-1 virions produced by the chronically infected SP cell line [248]. The SP cell line appears to be a god system to study HTLV-1 virion morphology because it harbor a relatively low number of HTLV-1 proviruses as compared to other HTLV-1 chronically infected cell lines. Their results show that only the full-length Gag protein was incorporated into virus particles. Cryo-TEM further confirmed that authentic HTLV-1 virions contains a combination of complete cores, incomplete cores, and particles without distinct electron densities that would correlate with the CA region. The particles collected from the SP cells contained generally polygonal CA cores, measured, on average, 115 nm in diameter, and presented a Turret- like structure associated with the particles at similar rates as reported in this dissertation.

These results demonstrate that the Turret-like structure associated with authentic HTLV-1 particles is not a cell dependent phenomenon. Combined, these studies provide evidence that the high number of particles containing incomplete CA cores could correlate with the known poor infectivity of HTLV-1 virions.

Further studies are needed to understand the composition, origin, and biological relevance of the Turret-like structures in the replication of HTLV-1. Mass spectrometry

96 could be conducted to determine the composition of these structures by analyzing purified HTLV-1 virions and compare these with that of VLPs. Further structural analysis could be performed with nuclear magnetic resonance analysis. Fluorescence microscopy could also be used to study composition and location of the Turret-like structures in purified HTLV-1 virions and at the surface of infected living cells. Ideally, lipid specific fluorescent molecular probes should be used and the labeled virions analyzed by correlative cryo-TEM fluorescence microscopy. With the current technology, it might not be possible to determine if the Turret-like structure associated with authentic HTLV-1 particles play a role in the infectivity of the cell-free virions.

Correlative cryo-TEM fluorescence microscopy will be used to better understand the virus particle assembly events that occur at the plasma membrane. Aditionally, higher resolution cryo-ET analysis should be performed to determine the arrangements of CA molecules that form a mature CA core. The studies presented in this dissertation showed that, unlike other retroviral immature particles reported to date, HTLV-1-like particles contain a Gag lattice with flat regions that does not follow the viral membrane. However, the origin, and implications of these “rigid” structures in the formation of a mature CA core are still to be determined. The curvature of the HIV-1 immature Gag lattice is achieved by the incorporation of defects into the lattice, while an hexameric lattice is more organized and flat [249]. The oligomeric HTLV-1 Gag arrangement, as well as the density at which they are packed in these flat regions, should be analyzed by high- resolution cryo-ET and subtomogram averaging. In future studies the Gag domains responsible for Gag-Gag interactions should be determined to elucidate whether these

97 interactions are important for the appearance of the flat regions observed in the Gag lattice structure of these particles.

The HTLV-1 NC vRNA chaperon activity was highly improved when the HTLV-

1 anionic CTD of NC was truncated [94]. Preliminary studies using an HTLV-1 Gag expression construct containing a truncated NC acidic CTD have suggested that this acidic CTD in NC may play important roles in both Gag stoichiometry as well as particle morphology (data not shown). A comparative study between this truncated HTLV-1 Gag and the bovine leukemia virus (BLV) Gag (which is naturally missing this acidic CTD in

NC) could be informative. In this proposed project, HTLV-1 and BLV Gag only constructs, with and without a truncation of the CTD of NC, will be used to produce

VLPs. The same set of constructs will be tagged with a fluorescent protein for the analysis of Gag and VLPs in living cells using fluorescence microscopy analyses. These experiments are important to better understand the role of the NC CTD in Gag-Gag and

Gag-PM interactions, VLP formation, release, and morphology.

There are currently no effective anti-HTLV-1 drugs for treating HTLV-1 infections. It is our goal to identify drug targets and design novel, small molecule inhibitors that are less prone to drug resistance development. Our research group is interested in developing drugs that specifically target HTLV-1 particle assembly, budding, release or maturation. Current efforts in identifying HIV-1 assembly small molecule inhibitors were recently reviewed [250], similar efforts will be employed to identify HTLV-1 assembly inhibitors. Our research group, in collaboration with the

University of Minnesota’s Center for Drug Design, will develop small molecule inhibitors targeting HTLV-1 MA-PM, CA-CA, and NC-RNA interactions. The work

98 presented in this dissertation as well as the proposed experiments, are essential to better understand the steps necessary to effectively produce infectious HTLV-1 virions.

99 Table 5-1. Summary of authentic HTLV-1 particle, HTLV-1-like particle, and

HTLV-1-like particle expressing Gag/Gag-YFP at a 3 to 1 ratio particle size, mass

determination and calculation of Gag copy number per particle.

HTLV-1 Particle Sample

Measurement Virus-Like Particle Gag:Gag-YFP (3:1) Authentic Average Diameter (nm) 110 113 113 Average Particle Mass (MDa) 174 219 204

RNA 7 6.6 7 Mass of RNA, Lipid and Protein (MDa) Lipid 70 80 80 Total protein 97 133 118 Total Gag polyprotein 70 - 87 93 - 120 82 - 106 Gag 70 - 87 70 - 90 70 - 90 Mass of Gag Molecules (MDa) Gag-Pro - - 10 - 13 Gag-Pro-Pol - - 2.5 - 3

Gag-YFP - 23 - 30 - Gag polyprotein copy number 1300 - 1600 1600 - 2100 1500 - 1900

100 BIBLIOGRAPHY

1. Poiesz, B.J., et al., Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-cell lymphoma. Proc Natl Acad Sci U S A, 1980. 77(12): p. 7415-9.

2. Yoshida, M., I. Miyoshi, and Y. Hinuma, Isolation and characterization of retrovirus from cell lines of human adult T-cell leukemia and its implication in the disease. Proc Natl Acad Sci U S A, 1982. 79(6): p. 2031-5.

3. Gallo, R.C., History of the discoveries of the first human retroviruses: HTLV-1 and HTLV-2. Oncogene, 2005. 24(39): p. 5926-30.

4. Kalyanaraman, V.S., et al., A new subtype of human T-cell leukemia virus (HTLV- II) associated with a T-cell variant of hairy cell leukemia. Science, 1982. 218(4572): p. 571-3.

5. Caron, C., et al., Functional and biochemical interaction of the HTLV-I Tax1 transactivator with TBP. EMBO J, 1993. 12(11): p. 4269-78.

6. Proietti, F.A., et al., Global epidemiology of HTLV-I infection and associated diseases. Oncogene, 2005. 24(39): p. 6058-68.

7. Gessain, A. and O. Cassar, Epidemiological Aspects and World Distribution of HTLV-1 Infection. Front Microbiol, 2012. 3: p. 388.

8. Calattini, S., et al., Discovery of a new human T-cell lymphotropic virus (HTLV-3) in Central Africa. Retrovirology, 2005. 2: p. 30.

9. Wolfe, N.D., et al., Emergence of unique primate T-lymphotropic viruses among central African bushmeat hunters. Proc Natl Acad Sci U S A, 2005. 102(22): p. 7994-9.

10. Yoshida, M., et al., Monoclonal integration of human T-cell leukemia provirus in all primary tumors of adult T-cell leukemia suggests causative role of human T- cell leukemia virus in the disease. Proc Natl Acad Sci U S A, 1984. 81(8): p. 2534-7.

11. Murphy, E.L., et al., Sexual transmission of human T-lymphotropic virus type I (HTLV-I). Ann Intern Med, 1989. 111(7): p. 555-60.

12. Kaplan, J.E., et al., The risk of development of HTLV-I-associated myelopathy/tropical spastic paraparesis among persons infected with HTLV-I. J Acquir Immune Defic Syndr, 1990. 3(11): p. 1096-101.

13. Bangham, C.R.M., Human T Cell Leukemia Virus Type 1: Persistence and Pathogenesis. Annu Rev Immunol, 2017.

101 14. Hisada, M., et al., Persistent paradox of natural history of human T lymphotropic virus type I: parallel analyses of Japanese and Jamaican carriers. J Infect Dis, 2004. 190(9): p. 1605-9.

15. Tsukasaki, K., et al., Definition, prognostic factors, treatment, and response criteria of adult T-cell leukemia-lymphoma: a proposal from an international consensus meeting. J Clin Oncol, 2009. 27(3): p. 453-9.

16. Tanosaki, R., et al., Allogeneic hematopoietic stem cell transplantation using reduced-intensity conditioning for adult T cell leukemia/lymphoma: impact of antithymocyte globulin on clinical outcome. Biol Blood Marrow Transplant, 2008. 14(6): p. 702-8.

17. Tsukasaki, K., et al., VCAP-AMP-VECP compared with biweekly CHOP for adult T-cell leukemia-lymphoma: Japan Clinical Oncology Group Study JCOG9801. J Clin Oncol, 2007. 25(34): p. 5458-64.

18. Yamamoto, K., et al., Phase I study of KW-0761, a defucosylated humanized anti- CCR4 antibody, in relapsed patients with adult T-cell leukemia-lymphoma and peripheral T-cell lymphoma. J Clin Oncol, 2010. 28(9): p. 1591-8.

19. Bazarbachi, A., et al., Meta-analysis on the use of zidovudine and interferon-alfa in adult T-cell leukemia/lymphoma showing improved survival in the leukemic subtypes. J Clin Oncol, 2010. 28(27): p. 4177-83.

20. Kinpara, S., et al., Interferon-alpha (IFN-alpha) suppresses HTLV-1 gene expression and cell cycling, while IFN-alpha combined with zidovudine induces p53 signaling and apoptosis in HTLV-1-infected cells. Retrovirology, 2013. 10: p. 52.

21. Meekings, K.N., et al., HTLV-1 integration into transcriptionally active genomic regions is associated with proviral expression and with HAM/TSP. PLoS Pathog, 2008. 4(3): p. doi:10.1371/journal.ppat.1000027.

22. Nagai, M., et al., Analysis of HTLV-I proviral load in 202 HAM/TSP patients and 243 asymptomatic HTLV-I carriers: high proviral load strongly predisposes to HAM/TSP. J Neurovirol, 1998. 4(6): p. 586-93.

23. Oh, U. and S. Jacobson, Treatment of HTLV-I-associated myelopathy/tropical spastic paraparesis: toward rational targeted therapy. Neurol Clin, 2008. 26(3): p. 781-97, ix-x.

24. Kawakami, A. and K. Eguchi, Role of HTLV-I infection in the pathogenesis of Sjogren's syndrome and rheumatoid arthritis. Mod Rheumatol, 2001. 11(2): p. 87-90.

25. Binard, A., et al., Epidemiology of Sjogren's syndrome: where are we now? Clin Exp Rheumatol, 2007. 25(1): p. 1-4. 102 26. Alani, H., et al., Systematic review and meta-analysis of the epidemiology of polyautoimmunity in Sjogren's syndrome (secondary Sjogren's syndrome) focusing on autoimmune rheumatic diseases. Scand J Rheumatol, 2017: p. 1-14.

27. Talal, N., et al., Detection of serum antibodies to retroviral proteins in patients with primary Sjogren's syndrome (autoimmune exocrinopathy). Arthritis Rheum, 1990. 33(6): p. 774-81.

28. Eguchi, K., et al., Primary Sjogren's syndrome with antibodies to HTLV-I: clinical and laboratory features. Ann Rheum Dis, 1992. 51(6): p. 769-76.

29. Sipsas, N.V., M.N. Gamaletsou, and H.M. Moutsopoulos, Is Sjogren's syndrome a retroviral disease? Arthritis Res Ther, 2011. 13(2): p. 212.

30. Weller, M.L., et al., Hepatitis Delta Virus Detected in Salivary Glands of Sjogren's Syndrome Patients and Recapitulates a Sjogren's Syndrome-Like Phenotype in Vivo. Pathog Immun, 2016. 1(1): p. 12-40.

31. Terada, K., et al., Prevalence of serum and salivary antibodies to HTLV-1 in Sjogren's syndrome. Lancet, 1994. 344(8930): p. 1116-9.

32. Hida, A., et al., Association of human T lymphotropic virus type I with Sjogren syndrome. Ann Rheum Dis, 2010. 69(11): p. 2056-7.

33. Nakamura, H., et al., High prevalence of Sjogren's syndrome in patients with HTLV-I associated myelopathy. Ann Rheum Dis, 1997. 56(3): p. 167-72.

34. Triantafyllopoulou, A. and H.M. Moutsopoulos, Autoimmunity and coxsackievirus infection in primary Sjogren's syndrome. Ann N Y Acad Sci, 2005. 1050: p. 389-96.

35. Nakamura, H. and A. Kawakami, What is the evidence for Sjogren's syndrome being triggered by viral infection? Subplot: infections that cause clinical features of Sjogren's syndrome. Curr Opin Rheumatol, 2016. 28(4): p. 390-7.

36. Nakamura, H., et al., Direct Infection of Primary Salivary Gland Epithelial Cells by Human T Lymphotropic Virus Type I in Patients With Sjogren's Syndrome. Arthritis Rheumatol, 2015. 67(4): p. 1096-106.

37. Boxus, M. and L. Willems, Mechanisms of HTLV-1 persistence and transformation. Br J Cancer, 2009. 101(9): p. 1497-501.

38. Chou, K.S., et al., Nucleotide sequence analysis of a full-length human T-cell leukemia virus type I from adult T-cell leukemia cells: a prematurely terminated PX open reading frame II. Int J Cancer, 1995. 60(5): p. 701-6.

103 39. Nam, S.H., et al., Processing of gag precursor polyprotein of human T-cell leukemia virus type I by virus-encoded protease. J Virol, 1988. 62(10): p. 3718- 28.

40. Lee, T.H., et al., Human T-cell leukemia virus-associated membrane antigens: identity of the major antigens recognized after virus infection. Proc Natl Acad Sci U S A, 1984. 81(12): p. 3856-60.

41. Le Blanc, I., et al., HTLV-1 structural proteins. Virus Res, 2001. 78(1-2): p. 5-16.

42. Bartoe, J.T., et al., Functional role of pX open reading frame II of human T- lymphotropic virus type 1 in maintenance of viral loads in vivo. J Virol, 2000. 74(3): p. 1094-100.

43. Collins, N.D., et al., Selective ablation of human T-cell lymphotropic virus type 1 p12I reduces viral infectivity in vivo. Blood, 1998. 91(12): p. 4701-7.

44. Silverman, L.R., et al., Human T-cell lymphotropic virus type 1 open reading frame II-encoded p30II is required for in vivo replication: evidence of in vivo reversion. J Virol, 2004. 78(8): p. 3837-45.

45. Zhao, T., The Role of HBZ in HTLV-1-Induced Oncogenesis. Viruses, 2016. 8(2).

46. Yamamoto, N., et al., Unique cell lines harbouring both Epstein-Barr virus and adult T-cell leukaemia virus, established from leukaemia patients. Nature, 1982. 299(5881): p. 367-9.

47. Ho, D.D., T.R. Rota, and M.S. Hirsch, Infection of human endothelial cells by human T-lymphotropic virus type I. Proc Natl Acad Sci U S A, 1984. 81(23): p. 7588-90.

48. Longo, D.L., et al., Isolation of HTLV-transformed B-lymphocyte clone from a patient with HTLV-associated adult T-cell leukaemia. Nature, 1984. 310(5977): p. 505-6.

49. Yoshikura, H., et al., Isolation of HTLV derived from Japanese adult T-cell leukemia patients in human diploid fibroblast strain IMR90 and the biological characters of the infected cells. Int J Cancer, 1984. 33(6): p. 745-9.

50. Koyanagi, Y., et al., In vivo infection of human T-cell leukemia virus type I in non-T cells. Virology, 1993. 196(1): p. 25-33.

51. Manel, N., et al., The ubiquitous glucose transporter GLUT-1 is a receptor for HTLV. Cell, 2003. 115(4): p. 449-59.

52. Jones, K.S., et al., Heparan sulfate proteoglycans mediate attachment and entry of human T-cell leukemia virus type 1 virions into CD4+ T cells. J Virol, 2005. 79(20): p. 12692-702. 104 53. Ghez, D., et al., Neuropilin-1 is involved in human T-cell lymphotropic virus type 1 entry. J Virol, 2006. 80(14): p. 6844-54.

54. Jones, K.S., et al., Cell-free HTLV-1 infects dendritic cells leading to transmission and transformation of CD4(+) T cells. Nat Med, 2008. 14(4): p. 429-36.

55. Temin, H.M. and S. Mizutani, RNA-dependent DNA polymerase in virions of Rous sarcoma virus. Nature, 1970. 226(5252): p. 1211-3.

56. Baltimore, D., RNA-dependent DNA polymerase in virions of RNA tumour viruses. Nature, 1970. 226(5252): p. 1209-11.

57. Hulme, A.E., O. Perez, and T.J. Hope, Complementary assays reveal a relationship between HIV-1 uncoating and reverse transcription. Proc Natl Acad Sci U S A, 2011. 108(24): p. 9975-80.

58. Tekeste, S.S., et al., Interaction Between Reverse Transcriptase and Integrase is Required for Reverse Transcription During HIV-1 Replication. J Virol, 2015.

59. Desrames, A., et al., Northern African strains of human T-lymphotropic virus type 1 arose from a recombination event. J Virol, 2014. 88(17): p. 9782-8.

60. Onafuwa-Nuga, A. and A. Telesnitsky, The remarkable frequency of human immunodeficiency virus type 1 genetic recombination. Microbiol Mol Biol Rev, 2009. 73(3): p. 451-80, Table of Contents.

61. Coffin, J.M., S.H. Hughes, and H.E. Varmus, The Interactions of Retroviruses and their Hosts, in Retroviruses, J.M. Coffin, S.H. Hughes, and H.E. Varmus, Editors. 1997: Cold Spring Harbor (NY).

62. Kitamura, Y., Y.M. Lee, and J.M. Coffin, Nonrandom integration of retroviral DNA in vitro: effect of CpG methylation. Proc Natl Acad Sci U S A, 1992. 89(12): p. 5532-6.

63. Doi, K., et al., Preferential selection of human T-cell leukemia virus type I provirus integration sites in leukemic versus carrier states. Blood, 2005. 106(3): p. 1048-53.

64. Holman, A.G. and J.M. Coffin, Symmetrical base preferences surrounding HIV-1, avian sarcoma/leukosis virus, and murine leukemia virus integration sites. Proc Natl Acad Sci U S A, 2005. 102(17): p. 6103-7.

65. Kang, Y., et al., Integration site choice of a feline immunodeficiency virus vector. J Virol, 2006. 80(17): p. 8820-3.

66. Derse, D., et al., Human T-cell leukemia virus type 1 integration target sites in the human genome: comparison with those of other retroviruses. J Virol, 2007. 81(12): p. 6731-41. 105 67. Gillet, N.A., et al., Strongyloidiasis and infective dermatitis alter human T lymphotropic virus-1 clonality in vivo. PLoS Pathog, 2013. 9(4): p. doi:10.1371/journal.ppat.1003263.

68. Cook, L.B., et al., The role of HTLV-1 clonality, proviral structure, and genomic integration site in adult T-cell leukemia/lymphoma. Blood, 2014. 123(25): p. 3925-31.

69. Kashanchi, F. and J.N. Brady, Transcriptional and post-transcriptional gene regulation of HTLV-1. Oncogene, 2005. 24(39): p. 5938-51.

70. Jeang, K.T., et al., Characterization of cellular factors that interact with the human T-cell leukemia virus type I p40x-responsive 21-base-pair sequence. J Virol, 1988. 62(12): p. 4499-509.

71. Baranger, A.M., et al., Mechanism of DNA-binding enhancement by the human T- cell leukaemia virus transactivator Tax. Nature, 1995. 376(6541): p. 606-8.

72. Adya, N. and C.Z. Giam, Distinct regions in human T-cell lymphotropic virus type I tax mediate interactions with activator protein CREB and basal transcription factors. J Virol, 1995. 69(3): p. 1834-41.

73. Adya, N., et al., Expansion of CREB's DNA recognition specificity by Tax results from interaction with Ala-Ala-Arg at positions 282-284 near the conserved DNA- binding domain of CREB. Proc Natl Acad Sci U S A, 1994. 91(12): p. 5642-6.

74. Goren, I., et al., The amino terminus of Tax is required for interaction with the cyclic AMP response element binding protein. J Virol, 1995. 69(9): p. 5806-11.

75. Tie, F., et al., Interaction of the human T-lymphotropic virus type 1 Tax dimer with CREB and the viral 21-base-pair repeat. J Virol, 1996. 70(12): p. 8368-74.

76. Zhao, L.J. and C.Z. Giam, Human T-cell lymphotropic virus type I (HTLV-I) transcriptional activator, Tax, enhances CREB binding to HTLV-I 21-base-pair repeats by protein-protein interaction. Proc Natl Acad Sci U S A, 1992. 89(15): p. 7070-4.

77. Ching, Y.P., et al., Specific TATAA and bZIP requirements suggest that HTLV-I Tax has transcriptional activity subsequent to the assembly of an initiation complex. Retrovirology, 2004. 1: p. doi:10.1186/1742-4690-1-18.

78. Seiki, M., et al., Direct evidence that p40x of human T-cell leukemia virus type I is a trans-acting transcriptional activator. EMBO J, 1986. 5(3): p. 561-5.

79. Giebler, H.A., et al., Anchoring of CREB binding protein to the human T-cell leukemia virus type 1 promoter: a molecular mechanism of Tax transactivation. Mol Cell Biol, 1997. 17(9): p. 5156-64.

106 80. Kwok, R.P., et al., Control of cAMP-regulated enhancers by the viral transactivator Tax through CREB and the co-activator CBP. Nature, 1996. 380(6575): p. 642-6.

81. Jiang, H., et al., PCAF interacts with tax and stimulates tax transactivation in a histone acetyltransferase-independent manner. Mol Cell Biol, 1999. 19(12): p. 8136-45.

82. Harrod, R., et al., p300 and p300/cAMP-responsive element-binding protein associated factor interact with human T-cell lymphotropic virus type-1 Tax in a multi-histone acetyltransferase/activator-enhancer complex. J Biol Chem, 2000. 275(16): p. 11852-7.

83. Bex, F., et al., Differential transcriptional activation by human T-cell leukemia virus type 1 Tax mutants is mediated by distinct interactions with CREB binding protein and p300. Mol Cell Biol, 1998. 18(4): p. 2392-405.

84. Green, P.L. and I.S. Chen, Regulation of human T cell leukemia virus expression. FASEB J, 1990. 4(2): p. 169-75.

85. Hidaka, M., et al., Post-transcriptional regulator (rex) of HTLV-1 initiates expression of viral structural proteins but suppresses expression of regulatory proteins. EMBO J, 1988. 7(2): p. 519-23.

86. Bogerd, H.P., et al., The type I human T-cell leukemia virus (HTLV-I) Rex trans- activator binds directly to the HTLV-I Rex and the type 1 human immunodeficiency virus Rev RNA response elements. Proc Natl Acad Sci U S A, 1991. 88(13): p. 5704-8.

87. Nosaka, T., et al., Nucleolar targeting signal of human T-cell leukemia virus type I rex-encoded protein is essential for cytoplasmic accumulation of unspliced viral mRNA. Proc Natl Acad Sci U S A, 1989. 86(24): p. 9798-802.

88. Rehberger, S., et al., The activation domain of a hormone inducible HTLV-1 Rex protein determines colocalization with the nuclear pore. Exp Cell Res, 1997. 233(2): p. 363-71.

89. Palmeri, D. and M.H. Malim, The human T-cell leukemia virus type 1 posttranscriptional trans-activator Rex contains a nuclear export signal. J Virol, 1996. 70(9): p. 6442-5.

90. Butsch, M. and K. Boris-Lawrie, Destiny of unspliced retroviral RNA: ribosome and/or virion? J Virol, 2002. 76(7): p. 3089-94.

91. Fogarty, K.H., et al., Characterization of cytoplasmic Gag-gag interactions by dual-color z-scan fluorescence fluctuation spectroscopy. Biophys J, 2011. 100(6): p. 1587-95.

107 92. Ritchie, C., et al., Analysis of HIV-1 Gag protein interactions via biotin ligase tagging. J Virol, 2015. 89(7): p. 3988-4001.

93. Barnard, A.L., et al., Engagement of specific T-cell surface molecules regulates cytoskeletal polarization in HTLV-1-infected lymphocytes. Blood, 2005. 106(3): p. 988-95.

94. Qualley, D.F., et al., C-terminal domain modulates the nucleic acid chaperone activity of human T-cell leukemia virus type 1 nucleocapsid protein via an electrostatic mechanism. J Biol Chem, 2010. 285(1): p. 295-307.

95. Sun, M., et al., Retrovirus-specific differences in matrix and nucleocapsid protein-nucleic acid interactions: implications for genomic RNA packaging. J Virol, 2014. 88(2): p. 1271-80.

96. Chen, J., et al., Cytoplasmic HIV-1 RNA is mainly transported by diffusion in the presence or absence of Gag protein. Proc Natl Acad Sci U S A, 2014. 111(48): p. E5205-13.

97. Demirov, D.G. and E.O. Freed, Retrovirus budding. Virus Res, 2004. 106(2): p. 87-102.

98. Morita, E. and W.I. Sundquist, Retrovirus budding. Annu Rev Cell Dev Biol, 2004. 20: p. 395-425.

99. Lingappa, J.R., et al., How HIV-1 Gag assembles in cells: Putting together pieces of the puzzle. Virus Res, 2014. 193: p. 89-107.

100. Briggs, J.A., et al., The stoichiometry of Gag protein in HIV-1. Nat Struct Mol Biol, 2004. 11(7): p. 672-5.

101. Chen, Y., et al., Fluorescence fluctuation spectroscopy on viral-like particles reveals variable gag stoichiometry. Biophys J, 2009. 96(5): p. 1961-9.

102. Vogt, V.M. and M.N. Simon, Mass determination of rous sarcoma virus virions by scanning transmission electron microscopy. J Virol, 1999. 73(8): p. 7050-5.

103. Yeager, M., et al., Supramolecular organization of immature and mature murine leukemia virus revealed by electron cryo-microscopy: implications for retroviral assembly mechanisms. Proc Natl Acad Sci U S A, 1998. 95(13): p. 7299-304.

104. Konvalinka, J., H.G. Krausslich, and B. Muller, Retroviral proteases and their roles in virion maturation. Virology, 2015. 479-480: p. 403-17.

105. Sonnino, S. and A. Prinetti, Membrane domains and the "lipid raft" concept. Curr Med Chem, 2013. 20(1): p. 4-21.

108 106. Lingwood, D., et al., Lipid rafts as functional heterogeneity in cell membranes. Biochem Soc Trans, 2009. 37(Pt 5): p. 955-60.

107. Ono, A., HIV-1 Assembly at the Plasma Membrane: Gag Trafficking and Localization. Future Virol, 2009. 4(3): p. 241-257.

108. Ootsuyama, Y., et al., Myristylation of gag protein in human T-cell leukemia virus type-I and type-II. Jpn J Cancer Res, 1985. 76(12): p. 1132-5.

109. Resh, M.D., Fatty acylation of proteins: new insights into membrane targeting of myristoylated and palmitoylated proteins. Biochim Biophys Acta, 1999. 1451(1): p. 1-16.

110. Zhou, W. and M.D. Resh, Differential membrane binding of the human immunodeficiency virus type 1 matrix protein. J Virol, 1996. 70(12): p. 8540-8.

111. Resh, M.D., A myristoyl switch regulates membrane binding of HIV-1 Gag. Proc Natl Acad Sci U S A, 2004. 101(2): p. 417-8.

112. Tang, C., et al., Entropic switch regulates myristate exposure in the HIV-1 matrix protein. Proc Natl Acad Sci U S A, 2004. 101(2): p. 517-22.

113. Datta, S.A., et al., HIV-1 Gag extension: conformational changes require simultaneous interaction with membrane and nucleic acid. J Mol Biol, 2011. 406(2): p. 205-14.

114. Murray, P.S., et al., Retroviral matrix domains share electrostatic homology: models for membrane binding function throughout the viral life cycle. Structure, 2005. 13(10): p. 1521-31.

115. Chukkapalli, V., et al., Interaction between the human immunodeficiency virus type 1 Gag matrix domain and phosphatidylinositol-(4,5)-bisphosphate is essential for efficient gag membrane binding. J Virol, 2008. 82(5): p. 2405-17.

116. van Meer, G., D.R. Voelker, and G.W. Feigenson, Membrane lipids: where they are and how they behave. Nat Rev Mol Cell Biol, 2008. 9(2): p. 112-24.

117. Chukkapalli, V. and A. Ono, Molecular determinants that regulate plasma membrane association of HIV-1 Gag. J Mol Biol, 2011. 410(4): p. 512-24.

118. Ono, A., et al., Phosphatidylinositol (4,5) bisphosphate regulates HIV-1 Gag targeting to the plasma membrane. Proc Natl Acad Sci U S A, 2004. 101(41): p. 14889-94.

119. Chan, J., R.A. Dick, and V.M. Vogt, Rous sarcoma virus gag has no specific requirement for phosphatidylinositol-(4,5)-bisphosphate for plasma membrane association in vivo or for liposome interaction in vitro. J Virol, 2011. 85(20): p. 10851-60. 109 120. Inlora, J., et al., Gag localization and virus-like particle release mediated by the matrix domain of human T-lymphotropic virus type 1 Gag are less dependent on phosphatidylinositol-(4,5)-bisphosphate than those mediated by the matrix domain of HIV-1 Gag. J Virol, 2011. 85(8): p. 3802-10.

121. Nadaraia-Hoke, S., et al., Alterations in the MA and NC domains modulate phosphoinositide-dependent plasma membrane localization of the Rous sarcoma virus Gag protein. J Virol, 2013. 87(6): p. 3609-15.

122. Zhang, W., et al., Morphology and ultrastructure of retrovirus particles. AIMS Biophys, 2015. 2(3): p. 343-369.

123. Gamble, T.R., et al., Structure of the carboxyl-terminal dimerization domain of the HIV-1 capsid protein. Science, 1997. 278(5339): p. 849-53.

124. Rayne, F., et al., The NH2-terminal domain of the human T-cell leukemia virus type 1 capsid protein is involved in particle formation. J Virol, 2001. 75(11): p. 5277-87.

125. Ganser-Pornillos, B.K., et al., Assembly properties of the human immunodeficiency virus type 1 CA protein. J Virol, 2004. 78(5): p. 2545-52.

126. Ako-Adjei, D., M.C. Johnson, and V.M. Vogt, The retroviral capsid domain dictates virion size, morphology, and coassembly of gag into virus-like particles. J Virol, 2005. 79(21): p. 13463-72.

127. Hogue, I.B., A. Hoppe, and A. Ono, Quantitative fluorescence resonance energy transfer microscopy analysis of the human immunodeficiency virus type 1 Gag- Gag interaction: relative contributions of the CA and NC domains and membrane binding. J Virol, 2009. 83(14): p. 7322-36.

128. Martin, J.L., et al., Disparate Contributions of Human Retrovirus Capsid Subdomains to Gag-Gag Oligomerization, Virus Morphology, and Particle Biogenesis. J Virol, 2017. 91(14).

129. Purohit, P., et al., Sequence-specific interaction between HIV-1 matrix protein and viral genomic RNA revealed by in vitro genetic selection. RNA, 2001. 7(4): p. 576-84.

130. Khorchid, A., et al., Role of RNA in facilitating Gag/Gag-Pol interaction. J Virol, 2002. 76(8): p. 4131-7.

131. Rein, A., et al., Diverse interactions of retroviral Gag proteins with RNAs. Trends Biochem Sci, 2011. 36(7): p. 373-80.

132. Li, H., et al., Myristoylation is required for human immunodeficiency virus type 1 Gag-Gag multimerization in mammalian cells. J Virol, 2007. 81(23): p. 12899- 910. 110 133. O'Carroll, I.P., et al., Functional redundancy in HIV-1 viral particle assembly. J Virol, 2012. 86(23): p. 12991-6.

134. Kutluay, S.B. and P.D. Bieniasz, Analysis of the initiating events in HIV-1 particle assembly and genome packaging. PLoS Pathog, 2010. 6(11): p. e1001200.

135. Fogarty, K.H., et al., Interrelationship between Cytoplasmic Retroviral Gag Concentration and Gag-Membrane Association. J Mol Biol, 2013.

136. Lindwasser, O.W. and M.D. Resh, Multimerization of human immunodeficiency virus type 1 Gag promotes its localization to barges, raft-like membrane microdomains. J Virol, 2001. 75(17): p. 7913-24.

137. Perez-Caballero, D., et al., Human immunodeficiency virus type 1 matrix inhibits and confers cooperativity on gag precursor-membrane interactions. J Virol, 2004. 78(17): p. 9560-3.

138. Goncalves, D.U., et al., Epidemiology, treatment, and prevention of human T-cell leukemia virus type 1-associated diseases. Clin Microbiol Rev, 2010. 23(3): p. 577-89.

139. Wiktor, S.Z., et al., Mother-to-child transmission of human T-cell lymphotropic virus type I (HTLV-I) in Jamaica: association with antibodies to envelope glycoprotein (gp46) epitopes. J Acquir Immune Defic Syndr, 1993. 6(10): p. 1162-7.

140. Takahashi, K., et al., Inhibitory effect of maternal antibody on mother-to-child transmission of human T-lymphotropic virus type I. The Mother-to-Child Transmission Study Group. Int J Cancer, 1991. 49(5): p. 673-7.

141. Nyambi, P.N., et al., Mother-to-child transmission of human T-cell lymphotropic virus types I and II (HTLV-I/II) in Gabon: a prospective follow-up of 4 years. J Acquir Immune Defic Syndr Hum Retrovirol, 1996. 12(2): p. 187-92.

142. Li, H.C., et al., Provirus load in breast milk and risk of mother-to-child transmission of human T lymphotropic virus type I. J Infect Dis, 2004. 190(7): p. 1275-8.

143. Martin-Latil, S., et al., Transcytosis of HTLV-1 across a tight human epithelial barrier and infection of subepithelial dendritic cells. Blood, 2012. 120(3): p. 572- 80.

144. Kinlock, B.L., et al., Transcytosis of HIV-1 through vaginal epithelial cells is dependent on trafficking to the endocytic recycling pathway. PLoS One, 2014. 9(5): p. e96760.

111 145. Alfsen, A., et al., HIV-1-infected blood mononuclear cells form an integrin- and agrin-dependent viral synapse to induce efficient HIV-1 transcytosis across epithelial cell monolayer. Mol Biol Cell, 2005. 16(9): p. 4267-79.

146. Filippone, C., et al., A Severe Bite From a Nonhuman Primate Is a Major Risk Factor for HTLV-1 Infection in Hunters From Central Africa. Clin Infect Dis, 2015. 60(11): p. 1667-76.

147. Kazanji, M., et al., Origin of HTLV-1 in hunters of nonhuman primates in Central Africa. J Infect Dis, 2015. 211(3): p. 361-5.

148. LeBreton, M., et al., A gorilla reservoir for human T-lymphotropic virus type 4. Emerg Microbes Infect, 2014. 3(1): p. e7.

149. Fan, N., et al., Infection of peripheral blood mononuclear cells and cell lines by cell-free human T-cell lymphoma/leukemia virus type I. J Clin Microbiol, 1992. 30(4): p. 905-10.

150. Igakura, T., et al., Spread of HTLV-I between lymphocytes by virus-induced polarization of the cytoskeleton. Science, 2003. 299(5613): p. 1713-6.

151. Majorovits, E., et al., Human T-lymphotropic virus-1 visualized at the virological synapse by electron tomography. PLoS One, 2008. 3(5): p. e2251.

152. Dimitrov, D.S., et al., Quantitation of human immunodeficiency virus type 1 infection kinetics. J Virol, 1993. 67(4): p. 2182-90.

153. Mazurov, D., et al., Quantitative comparison of HTLV-1 and HIV-1 cell-to-cell infection with new replication dependent vectors. PLoS Pathog, 2010. 6(2): p. doi:10.1371/journal.ppat.1000788.

154. Jiang, A.P., et al., Human Blood-Circulating Basophils Capture HIV-1 and Mediate Viral trans-Infection of CD4+ T Cells. J Virol, 2015. 89(15): p. 8050-62.

155. Pais-Correia, A.M., et al., Biofilm-like extracellular viral assemblies mediate HTLV-1 cell-to-cell transmission at virological synapses. Nat Med, 2010. 16(1): p. 83-9.

156. Alais, S., R. Mahieux, and H. Dutartre, Viral Source-Independent High Susceptibility of Dendritic Cells to Human T-Cell Leukemia Virus Type 1 Infection Compared to That of T Lymphocytes. J Virol, 2015. 89(20): p. 10580-90.

157. Gessain, A., R.C. Gallo, and G. Franchini, Low degree of human T-cell leukemia/lymphoma virus type I genetic drift in vivo as a means of monitoring viral transmission and movement of ancient human populations. J Virol, 1992. 66(4): p. 2288-95.

112 158. Van Dooren, S., et al., The low evolutionary rate of human T-cell lymphotropic virus type-1 confirmed by analysis of vertical transmission chains. Mol Biol Evol, 2004. 21(3): p. 603-11.

159. Mansky, L.M., In vivo analysis of human T-cell leukemia virus type 1 reverse transcription accuracy. J Virol, 2000. 74(20): p. 9525-31.

160. Mansky, L.M. and H.M. Temin, Lower in vivo mutation rate of human immunodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase. J Virol, 1995. 69(8): p. 5087-94.

161. Taylor, G.P., et al., Zidovudine plus lamivudine in Human T-Lymphotropic Virus type-I-associated myelopathy: a randomised trial. Retrovirology, 2006. 3: p. 63.

162. Miyazato, P., et al., De novo human T-cell leukemia virus type 1 infection of human lymphocytes in NOD-SCID, common gamma-chain knockout mice. J Virol, 2006. 80(21): p. 10683-91.

163. Seiki, M., et al., Nonspecific integration of the HTLV provirus genome into adult T-cell leukaemia cells. Nature, 1984. 309(5969): p. 640-2.

164. Ohshima, K., et al., Random integration of HTLV-1 provirus: increasing chromosomal instability. Cancer Lett, 1998. 132(1-2): p. 203-12.

165. Wattel, E., et al., Clonal expansion of human T-cell leukemia virus type I-infected cells in asymptomatic and symptomatic carriers without malignancy. J Virol, 1995. 69(5): p. 2863-8.

166. Etoh, K., et al., Persistent clonal proliferation of human T-lymphotropic virus type I-infected cells in vivo. Cancer Res, 1997. 57(21): p. 4862-7.

167. de The, G. and R. Bomford, An HTLV-I vaccine: why, how, for whom? AIDS Res Hum Retroviruses, 1993. 9(5): p. 381-6.

168. Gessain, A., et al., Antibodies to human T-lymphotropic virus type-I in patients with tropical spastic paraparesis. Lancet, 1985. 2(8452): p. 407-10.

169. Osame, M., et al., HTLV-I associated myelopathy, a new clinical entity. Lancet, 1986. 1(8488): p. 1031-2.

170. Maldonado, J., et al., New Insights Into Retroviral Gag-Gag and Gag-Membrane Interactions. Frontiers in Microbiology, 2014. 5(302): p. doi:10.3389/fmicb.2014.00302.

171. Jouvenet, N., S.M. Simon, and P.D. Bieniasz, Imaging the interaction of HIV-1 genomes and Gag during assembly of individual viral particles. Proc Natl Acad Sci U S A, 2009. 106(45): p. 19114-9.

113 172. Briggs, J.A., et al., Cryo-electron microscopy reveals conserved and divergent features of gag packing in immature particles of Rous sarcoma virus and human immunodeficiency virus. J Mol Biol, 2006. 355(1): p. 157-68.

173. Briggs, J.A., et al., Cryoelectron microscopy of mouse mammary tumor virus. J Virol, 2004. 78(5): p. 2606-8.

174. Butan, C., et al., RSV capsid polymorphism correlates with polymerization efficiency and envelope glycoprotein content: implications that nucleation controls morphogenesis. J Mol Biol, 2008. 376(4): p. 1168-81.

175. Carlson, L.A., et al., Three-dimensional analysis of budding sites and released virus suggests a revised model for HIV-1 morphogenesis. Cell Host Microbe, 2008. 4(6): p. 592-9.

176. Fuller, S.D., et al., Cryo-electron microscopy reveals ordered domains in the immature HIV-1 particle. Curr Biol, 1997. 7(10): p. 729-38.

177. Yu, F., et al., Characterization of Rous sarcoma virus Gag particles assembled in vitro. J Virol, 2001. 75(6): p. 2753-64.

178. Briggs, J.A., et al., The mechanism of HIV-1 core assembly: insights from three- dimensional reconstructions of authentic virions. Structure, 2006. 14(1): p. 15-20.

179. Parker, S.D., J.S. Wall, and E. Hunter, Analysis of Mason-Pfizer monkey virus Gag particles by scanning transmission electron microscopy. J Virol, 2001. 75(19): p. 9543-8.

180. Grigsby, I.F., et al., Biophysical analysis of HTLV-1 particles reveals novel insights into particle morphology and Gag stochiometry. Retrovirology, 2010. 7: p. 75.

181. Haertle, T., et al., Metabolism and anti-human immunodeficiency virus-1 activity of 2-halo-2',3'-dideoxyadenosine derivatives. J Biol Chem, 1988. 263(12): p. 5870-5.

182. Harada, S., Y. Koyanagi, and N. Yamamoto, Infection of HTLV-III/LAV in HTLV- I-carrying cells MT-2 and MT-4 and application in a plaque assay. Science, 1985. 229(4713): p. 563-6.

183. Rossio, J.L., et al., Inactivation of human immunodeficiency virus type 1 infectivity with preservation of conformational and functional integrity of virion surface proteins. J Virol, 1998. 72(10): p. 7992-8001.

184. Baker, T.S., N.H. Olson, and S.D. Fuller, Adding the third dimension to virus life cycles: three-dimensional reconstruction of icosahedral viruses from cryo- electron micrographs. Microbiol Mol Biol Rev, 1999. 63(4): p. 862-922.

114 185. Wall, J.S., J.F. Hainfeld, and M.N. Simon, Scanning transmission electron microscopy of nuclear structures. Methods Cell Biol, 1998. 53: p. 139-64.

186. Namba, K. and G. Stubbs, Structure of tobacco mosaic virus at 3.6 A resolution: implications for assembly. Science, 1986. 231(4744): p. 1401-6.

187. Bharat, T.A., et al., Structure of the immature retroviral capsid at 8 A resolution by cryo-electron microscopy. Nature, 2012. 487(7407): p. 385-9.

188. de Marco, A., et al., Conserved and variable features of Gag structure and arrangement in immature retrovirus particles. J Virol, 2010. 84(22): p. 11729-36.

189. Frank, G.A., et al., Maturation of the HIV-1 core by a non-diffusional phase transition. Nat Commun, 2015. 6: p. 5854.

190. Briggs, J.A., et al., Structure and assembly of immature HIV. Proceedings of the National Academy of Sciences of the United States of America, 2009. 106(27): p. 11090-5.

191. Wright, E.R., et al., Electron cryotomography of immature HIV-1 virions reveals the structure of the CA and SP1 Gag shells. The EMBO journal, 2007. 26(8): p. 2218-26.

192. Cao, S., et al., Analysis of human T-cell leukemia virus type 1 particles by using cryo-electron tomography. J Virol, 2015. 89(4): p. 2430-5.

193. Cockburn, J.J., et al., Membrane structure and interactions with protein and DNA in bacteriophage PRD1. Nature, 2004. 432(7013): p. 122-5.

194. Fleissner, E., Chromatographic separation and antigenic analysis of proteins of the oncornaviruses. I. Avian leukemia-sarcoma viruses. J Virol, 1971. 8(5): p. 778-85.

195. Nam, S.H., et al., Characterization of ribosomal frameshifting for expression of pol gene products of human T-cell leukemia virus type I. J Virol, 1993. 67(1): p. 196-203.

196. Mador, N., A. Panet, and A. Honigman, Translation of gag, pro, and pol gene products of human T-cell leukemia virus type 2. J Virol, 1989. 63(5): p. 2400-4.

197. Poiesz, B.J., et al., Isolation of a new type C retrovirus (HTLV) in primary uncultured cells of a patient with Sezary T-cell leukaemia. Nature, 1981. 294(5838): p. 268-71.

198. Ding, Y.S., D.H. Rich, and R.A. Ikeda, Substrates and inhibitors of human T-cell leukemia virus type I protease. Biochemistry, 1998. 37(50): p. 17514-8.

115 199. Louis, J.M., S. Oroszlan, and J. Tozser, Stabilization from autoproteolysis and kinetic characterization of the human T-cell leukemia virus type 1 proteinase. J Biol Chem, 1999. 274(10): p. 6660-6.

200. Pettit, S.C., et al., HIV type 1 protease inhibitors fail to inhibit HTLV-I Gag processing in infected cells. AIDS Res Hum Retroviruses, 1998. 14(11): p. 1007- 14.

201. Tozser, J. and I.T. Weber, The protease of human T-cell leukemia virus type-1 is a potential therapeutic target. Curr Pharm Des, 2007. 13(12): p. 1285-94.

202. Schur, F.K., et al., The Structure of Immature Virus-Like Rous Sarcoma Virus Gag Particles Reveals a Structural Role for the p10 Domain in Assembly. J Virol, 2015. 89(20): p. 10294-302.

203. Wills, J.W. and R.C. Craven, Form, function, and use of retroviral gag proteins. AIDS, 1991. 5(6): p. 639-54.

204. Hirokawa, N. and Y. Noda, Intracellular transport and kinesin superfamily proteins, KIFs: structure, function, and dynamics. Physiol Rev, 2008. 88(3): p. 1089-118.

205. Burniston, M.T., et al., Human immunodeficiency virus type 1 Gag polyprotein multimerization requires the nucleocapsid domain and RNA and is promoted by the capsid-dimer interface and the basic region of matrix protein. J Virol, 1999. 73(10): p. 8527-40.

206. Fogarty, K.H., et al., New insights into HTLV-1 particle structure, assembly, and Gag-Gag interactions in living cells. Viruses, 2011. 3(6): p. 770-93.

207. O'Carroll, I.P., et al., Elements in HIV-1 Gag contributing to virus particle assembly. Virus Res, 2013. 171(2): p. 341-5.

208. Maldonado, J.O., et al., Distinct Morphology of Human T-Cell Leukemia Virus Type 1-Like Particles. Viruses, 2016. 8(5).

209. Johnson, J., Y. Chen, and J.D. Mueller, Characterization of brightness and stoichiometry of bright particles by flow-fluorescence fluctuation spectroscopy. Biophys J, 2010. 99(9): p. 3084-92.

210. Zacharias, D.A., et al., Partitioning of lipid-modified monomeric GFPs into membrane microdomains of live cells. Science, 2002. 296(5569): p. 913-6.

211. Chen, Y., L.N. Wei, and J.D. Muller, Probing protein oligomerization in living cells with fluorescence fluctuation spectroscopy. Proc Natl Acad Sci U S A, 2003. 100(26): p. 15492-7.

116 212. Chen, Y., et al., Observing protein interactions and their stoichiometry in living cells by brightness analysis of fluorescence fluctuation experiments. Methods Enzymol, 2010. 472: p. 345-63.

213. Chirico, G., et al., Molecular heterogeneity of O-acetylserine sulfhydrylase by two-photon excited fluorescence fluctuation spectroscopy. Biophys J, 2001. 80(4): p. 1973-85.

214. Freeman, R. and K.R. Leonard, Comparative mass measurement of biological macromolecules by scanning transmission electron microscopy. J Microsc, 1981. 122(Pt 3): p. 275-86.

215. Shirley A. Müller, K.N.G., Roland Bürki, Robert Häring, Andreas Engel, Factors influencing the precision of quantitative scanning transmission electron microscopy. Ultramicroscopy, 1992. 46(1-4): p. 317-334.

216. Thomas, D., et al., Mass analysis of biological macromolecular complexes by STEM. Biol Cell, 1994. 80(2-3): p. 181-92.

217. Martin, J.L., et al., Molecular Studies of HTLV-1 Replication: An Update. Viruses, 2016. 8(2).

218. Shehu-Xhilaga, M., S.M. Crowe, and J. Mak, Maintenance of the Gag/Gag-Pol ratio is important for human immunodeficiency virus type 1 RNA dimerization and viral infectivity. J Virol, 2001. 75(4): p. 1834-41.

219. Johnson, S.F., et al., Determinants of Moloney murine leukemia virus Gag-Pol and genomic RNA proportions. J Virol, 2014. 88(13): p. 7267-75.

220. Pique, C. and K.S. Jones, Pathways of cell-cell transmission of HTLV-1. Front Microbiol, 2012. 3: p. 378.

221. Sundquist, W.I. and H.G. Krausslich, HIV-1 assembly, budding, and maturation. Cold Spring Harb Perspect Med, 2012. 2(7): p. a006924.

222. Briggs, J.A., et al., Structural organization of authentic, mature HIV-1 virions and cores. Embo J, 2003. 22(7): p. 1707-15.

223. Yu, Z., et al., Unclosed HIV-1 capsids suggest a curled sheet model of assembly. J Mol Biol, 2012. 425(1): p. 112-23.

224. Cardone, G., et al., Visualization of a missing link in retrovirus capsid assembly. Nature, 2009. 457(7230): p. 694-8.

225. Bailey, G.D., et al., Proton-linked dimerization of a retroviral capsid protein initiates capsid assembly. Structure, 2009. 17(5): p. 737-48.

117 226. Zhao, G., et al., Mature HIV-1 capsid structure by cryo-electron microscopy and all-atom molecular dynamics. Nature, 2013. 497(7451): p. 643-6.

227. Katoh, I., et al., Murine leukemia virus maturation: protease region required for conversion from "immature" to "mature" core form and for virus infectivity. Virology, 1985. 145(2): p. 280-92.

228. Scheifele, L.Z., et al., Overlapping roles of the Rous sarcoma virus Gag p10 domain in nuclear export and virion core morphology. J Virol, 2007. 81(19): p. 10718-28.

229. Tang, S., et al., Human immunodeficiency virus type 1 N-terminal capsid mutants that exhibit aberrant core morphology and are blocked in initiation of reverse transcription in infected cells. J Virol, 2001. 75(19): p. 9357-66.

230. Miyoshi, I., et al., Type C virus particles in a cord T-cell line derived by co- cultivating normal human cord leukocytes and human leukaemic T cells. Nature, 1981. 294(5843): p. 770-1.

231. Wang, H., N.J. Machesky, and L.M. Mansky, Both the PPPY and PTAP motifs are involved in human T-cell leukemia virus type 1 particle release. J Virol, 2004. 78(3): p. 1503-12.

232. Morozov, V.A. and R.A. Weiss, Two types of HTLV-1 particles are released from MT-2 cells. Virology, 1999. 255(2): p. 279-84.

233. Mastronarde, D.N., Automated electron microscope tomography using robust prediction of specimen movements. J Struct Biol, 2005. 152(1): p. 36-51.

234. http://cryoem.ucsd.edu/programs.shtm.

235. Kremer, J.R., D.N. Mastronarde, and J.R. McIntosh, Computer visualization of three-dimensional image data using IMOD. J Struct Biol, 1996. 116(1): p. 71-6.

236. Carlson, L.A., et al., Cryo electron tomography of native HIV-1 budding sites. PLoS Pathog, 2010. 6(11): p. e1001173.

237. Keller, P.W., et al., HIV-1 maturation inhibitor bevirimat stabilizes the immature Gag lattice. J Virol, 2011. 85(4): p. 1420-8.

238. Kobayashi, N., et al., Genomic structure of HTLV (human T-cell leukemia virus): detection of defective genome and its amplification in MT-2 cells. EMBO J, 1984. 3(6): p. 1339-43.

239. Iino, T., et al., Structural analysis of p28 adult T-cell leukaemia-associated antigen. J Gen Virol, 1986. 67 ( Pt 7): p. 1373-9.

118 240. Kuroda, M.J., et al., Inefficient transmission of HTLV-I to MOLT-4 cells by cell- free virus and cocultivation. Intervirology, 1992. 34(4): p. 202-12.

241. Agadjanyan, M.G., et al., Identification of an 80-kilodalton membrane glycoprotein important for human T-cell leukemia virus type I and type II syncytium formation and infection. J Virol, 1994. 68(1): p. 485-93.

242. Clapham, P., et al., Productive infection and cell-free transmission of human T- cell leukemia virus in a nonlymphoid cell line. Science, 1983. 222(4628): p. 1125- 7.

243. Ren, G., et al., Model of human low-density lipoprotein and bound receptor based on cryoEM. Proc Natl Acad Sci U S A, 2010. 107(3): p. 1059-64.

244. Dobro, M.J., et al., Electron cryotomography of ESCRT assemblies and dividing Sulfolobus cells suggests that spiraling filaments are involved in membrane scission. Mol Biol Cell, 2013. 24(15): p. 2319-27.

245. Shen, Q.T., et al., Structural analysis and modeling reveals new mechanisms governing ESCRT-III spiral filament assembly. J Cell Biol, 2014. 206(6): p. 763- 77.

246. Khorasanizadeh, S., R. Campos-Olivas, and M.F. Summers, Solution structure of the capsid protein from the human T-cell leukemia virus type-I. J Mol Biol, 1999. 291(2): p. 491-505.

247. Pettersen, E.F., et al., UCSF Chimera--a visualization system for exploratory research and analysis. J Comput Chem, 2004. 25(13): p. 1605-12.

248. Meissner, M.E., et al., Polymorphic Nature of Human T-Cell Leukemia Virus Type 1 Particle Cores as Revealed through Characterization of a Chronically Infected Cell Line. J Virol, 2017. 91(16).

249. Briggs, J.A., et al., Structure and assembly of immature HIV. Proc Natl Acad Sci U S A, 2009. 106(27): p. 11090-5.

250. Spearman, P., HIV-1 Gag as an Antiviral Target: Development of Assembly and Maturation Inhibitors. Curr Top Med Chem, 2016. 16(10): p. 1154-66.

251. Gheysen, D., et al., Assembly and release of HIV-1 precursor Pr55gag virus-like particles from recombinant baculovirus-infected insect cells. Cell, 1989. 59(1): p. 103-12.

252. Wilk, T., et al., Organization of immature human immunodeficiency virus type 1. J Virol, 2001. 75(2): p. 759-71.

253. Pietschmann, T., et al., Foamy virus capsids require the cognate envelope protein for particle export. J Virol, 1999. 73(4): p. 2613-21. 119 254. Martin-Serrano, J., T. Zang, and P.D. Bieniasz, Role of ESCRT-I in retroviral budding. J Virol, 2003. 77(8): p. 4794-804.

255. Goldstone, D.C., et al., A unique spumavirus Gag N-terminal domain with functional properties of orthoretroviral matrix and capsid. PLoS Pathog, 2013. 9(5): p. e1003376.

256. Cruceanu, M., et al., Nucleic acid binding and chaperone properties of HIV-1 Gag and nucleocapsid proteins. Nucleic Acids Res, 2006. 34(2): p. 593-605.

257. Zhang, Y., et al., Analysis of the assembly function of the human immunodeficiency virus type 1 gag protein nucleocapsid domain. J Virol, 1998. 72(3): p. 1782-9.

258. Flugel, R.M. and K.I. Pfrepper, Proteolytic processing of foamy virus Gag and Pol proteins. Curr Top Microbiol Immunol, 2003. 277: p. 63-88.

259. Gao, W., et al., UpGene: Application of a web-based DNA codon optimization algorithm. Biotechnol Prog, 2004. 20(2): p. 443-8.

260. Kozak, M., Compilation and analysis of sequences upstream from the translational start site in eukaryotic mRNAs. Nucleic Acids Res, 1984. 12(2): p. 857-72.

261. Kozak, M., An analysis of 5'-noncoding sequences from 699 vertebrate messenger RNAs. Nucleic Acids Res, 1987. 15(20): p. 8125-48.

262. Trobridge, G., et al., Improved foamy virus vectors with minimal viral sequences. Mol Ther, 2002. 6(3): p. 321-8.

263. Golovkina, T.V., J.P. Dudley, and S.R. Ross, B and T cells are required for mouse mammary tumor virus spread within the mammary gland. J Immunol, 1998. 161(5): p. 2375-82.

264. Jorgenson, R.L., V.M. Vogt, and M.C. Johnson, Foreign glycoproteins can be actively recruited to virus assembly sites during pseudotyping. J Virol, 2009. 83(9): p. 4060-7.

265. Delamarre, L., et al., A novel human T-leukemia virus type 1 cell-to-cell transmission assay permits definition of SU glycoprotein amino acids important for infectivity. J Virol, 1997. 71(1): p. 259-66.

266. Landau, N.R. and D.R. Littman, Packaging system for rapid production of murine leukemia virus vectors with variable tropism. J Virol, 1992. 66(8): p. 5110-3.

267. Zabransky, A., et al., Premature processing of mouse mammary tumor virus Gag polyprotein impairs intracellular capsid assembly. Virology, 2009. 384(1): p. 33- 7. 120 268. Smith, G.H. and N.A. Wivel, Intracytoplasmic A particles: mouse mammary tumor virus nucleoprotein cores? J Virol, 1973. 11(4): p. 575-84.

269. Stansell, E., et al., Basic residues in the Mason-Pfizer monkey virus gag matrix domain regulate intracellular trafficking and capsid-membrane interactions. J Virol, 2007. 81(17): p. 8977-88.

270. Grigorov, B., et al., Assembly of infectious HIV-1 in human epithelial and T- lymphoblastic cell lines. J Mol Biol, 2006. 359(4): p. 848-62.

271. Johnson, M.C., H.M. Scobie, and V.M. Vogt, PR domain of rous sarcoma virus Gag causes an assembly/budding defect in insect cells. J Virol, 2001. 75(9): p. 4407-12.

272. Mortuza, G.B., et al., Structure of B-MLV capsid amino-terminal domain reveals key features of viral tropism, gag assembly and core formation. J Mol Biol, 2008. 376(5): p. 1493-508.

273. Kyere, S.K., P.R. Joseph, and M.F. Summers, The p12 domain is unstructured in a murine leukemia virus p12-CA(N) Gag construct. PLoS One, 2008. 3(4): p. e1902.

274. Nandhagopal, N., et al., Dimeric rous sarcoma virus capsid protein structure relevant to immature Gag assembly. J Mol Biol, 2004. 335(1): p. 275-82.

275. Rovnak, J., J.W. Casey, and S.L. Quackenbush, Intracellular targeting of walleye dermal sarcoma virus Orf A (rv-cyclin). Virology, 2001. 280(1): p. 31-40.

276. Garbitt-Hirst, R., S.P. Kenney, and L.J. Parent, Genetic evidence for a connection between Rous sarcoma virus gag nuclear trafficking and genomic RNA packaging. J Virol, 2009. 83(13): p. 6790-7.

277. Schur, F.K., et al., Structure of the immature HIV-1 capsid in intact virus particles at 8.8 A resolution. Nature, 2015. 517(7535): p. 505-8.

278. Schur, F.K., et al., Determination of protein structure at 8.5A resolution using cryo-electron tomography and sub-tomogram averaging. J Struct Biol, 2013. 184(3): p. 394-400.

279. Menendez-Arias, L., et al., Purification of immature cores of mouse mammary tumor virus and immunolocalization of protein domains. J Virol, 1992. 66(9): p. 5615-20.

280. Hamann, M.V., et al., The cooperative function of arginine residues in the Prototype Foamy Virus Gag C-terminus mediates viral and cellular RNA encapsidation. Retrovirology, 2014. 11: p. 87.

121 281. Boyko, V., et al., Coassembly and complementation of Gag proteins from HIV-1 and HIV-2, two distinct human pathogens. Mol Cell, 2006. 23(2): p. 281-7.

282. Goepfert, P.A., et al., An endoplasmic reticulum retrieval signal partitions human foamy virus maturation to intracytoplasmic membranes. J Virol, 1999. 73(9): p. 7210-7.

283. Reichard, P., Interactions between deoxyribonucleotide and DNA synthesis. 1988, Ann. Rev. Biochem. p. 349-74.

284. Guarino, E., I. Salguero, and S.E. Kearsey, Cellular regulation of ribonucleotide reductase in eukaryotes. Semin Cell Dev Biol, 2014. 30: p. 97-103.

285. Håkansson, P., A. Hofer, and L. Thelander, Regulation of mammalian ribonucleotide reduction and dNTP pools after DNA damage and in resting cells. J Biol Chem, 2006. 281(12): p. 7834-41.

286. Gandhi, V.V. and D.C. Samuels, A review comparing deoxyribonucleoside triphosphate (dNTP) concentrations in the mitochondrial and cytoplasmic compartments of normal and transformed cells. Nucleosides, nucleotides & nucleic acids, 2011. 30(5): p. 317-39.

287. Traut, T.W., Physiological concentrations of purines and pyrimidines. Molecular and cellular biochemistry, 1994. 140(1): p. 1-22.

288. Niida, H., et al., Mechanisms of dNTP supply that play an essential role in maintaining genome integrity in eukaryotic cells. Cancer Sci, 2010. 101(12): p. 2505-9.

289. Björklund, S., et al., S-phase-specific expression of mammalian ribonucleotide reductase R1 and R2 subunit mRNAs. Biochemistry, 1990. 29(23): p. 5452-8.

290. Diamond, T.L., et al., Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse transcriptase. J Biol Chem, 2004. 279(49): p. 51545- 53.

291. Skasko, M., T.L. Diamond, and B. Kim, Mechanistic variations among reverse transcriptases of simian immunodeficiency virus variants isolated from African green monkeys. Biochemistry, 2009. 48(23): p. 5389-95.

292. Amie, S.M., E. Noble, and B. Kim, Intracellular nucleotide levels and the control of retroviral infections. Virology, 2013. 436(2): p. 247-54.

293. Ayinde, D., N. Casartelli, and O. Schwartz, Restricting HIV the SAMHD1 way: through nucleotide starvation. Nat Rev Microbiol, 2012. 10(10): p. 675-80.

294. St Gelais, C. and L. Wu, SAMHD1: a new insight into HIV-1 restriction in myeloid cells. Retrovirology, 2011. 8: p. 55. 122 295. Lahouassa, H., et al., SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting the intracellular pool of deoxynucleoside triphosphates. Nat Immunol, 2012. 13(3): p. 223-8.

296. Kim, B., et al., Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral DNA synthesis kinetics in human primary monocyte- derived macrophages. J Biol Chem, 2012. 287(26): p. 21570-4.

297. Goujon, C., et al., Characterization of simian immunodeficiency virus SIVSM/human immunodeficiency virus type 2 Vpx function in human myeloid cells. J Virol, 2008. 82(24): p. 12335-45.

298. Hrecka, K., et al., Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1 protein. Nature, 2011. 474(7353): p. 658-61.

299. Allouch, A., et al., p21-mediated RNR2 repression restricts HIV-1 replication in macrophages by inhibiting dNTP biosynthesis pathway. Proc Natl Acad Sci U S A, 2013. 110(42): p. E3997-4006.

300. Yamauchi, T., et al., Cytarabine-resistant leukemia cells are moderately sensitive to clofarabine in vitro. Anticancer Res, 2014. 34(4): p. 1657-62.

301. Aye, Y. and J. Stubbe, Clofarabine 5'-di and -triphosphates inhibit human ribonucleotide reductase by altering the quaternary structure of its large subunit. Proc Natl Acad Sci U S A, 2011. 108(24): p. 9815-20.

302. Aye, Y., et al., Clofarabine targets the large subunit (α) of human ribonucleotide reductase in live cells by assembly into persistent hexamers. Chem Biol, 2012. 19(7): p. 799-805.

303. Xie, K.C. and W. Plunkett, Deoxynucleotide pool depletion and sustained inhibition of ribonucleotide reductase and DNA synthesis after treatment of human lymphoblastoid cells with 2-chloro-9-(2-deoxy-2-fluoro-beta-D- arabinofuranosyl) adenine. Cancer Res, 1996. 56(13): p. 3030-7.

304. Parker, W.B., et al., Effects of 2-chloro-9-(2-deoxy-2-fluoro-beta-D- arabinofuranosyl)adenine on K562 cellular metabolism and the inhibition of human ribonucleotide reductase and DNA polymerases by its 5'-triphosphate. Cancer Res, 1991. 51(9): p. 2386-94.

305. Beach, L.B., et al., Novel inhibitors of human immunodeficiency virus type 2 infectivity. J Gen Virol, 2014. 95(Pt 12): p. 2778-83.

306. Rawson, J.M., et al., 5,6-Dihydro-5-aza-2'-deoxycytidine potentiates the anti- HIV-1 activity of ribonucleotide reductase inhibitors. Bioorg Med Chem, 2013. 21(22): p. 7222-8.

123 307. Arnold, L.H., et al., A continuous enzyme-coupled assay for triphosphohydrolase activity of HIV-1 restriction factor SAMHD1. Antimicrob Agents Chemother, 2015. 59(1): p. 186-92.

308. Joe, A.K., et al., Resveratrol induces growth inhibition, S-phase arrest, apoptosis, and changes in biomarker expression in several human cancer cell lines. Clin Cancer Res, 2002. 8(3): p. 893-903.

309. Bianchi, V., E. Pontis, and P. Reichard, Changes of deoxyribonucleoside triphosphate pools induced by hydroxyurea and their relation to DNA synthesis. J Biol Chem, 1986. 261(34): p. 16037-42.

310. Kohler, J.J. and W. Lewis, A brief overview of mechanisms of mitochondrial toxicity from NRTIs. Environ Mol Mutagen, 2007. 48(3-4): p. 166-72.

311. Kopp, E.B., et al., Steady state kinetics and inhibition of HIV-1 reverse transcriptase by a non-nucleoside dipyridodiazepinone, BI-RG-587, using a heteropolymeric template. Nucleic Acids Res, 1991. 19(11): p. 3035-9.

312. Tyrsted, G., Effect of hydroxyurea and 5-fluorodeoxyuridine on deoxyribonucleoside triphosphate pools early in phytohemagglutinin-stimulated human lymphocytes. Biochem Pharmacol, 1982. 31(19): p. 3107-13.

313. Gao, W.Y., et al., Low levels of deoxynucleotides in peripheral blood lymphocytes: a strategy to inhibit human immunodeficiency virus type 1 replication. Proc Natl Acad Sci U S A, 1993. 90(19): p. 8925-8.

314. Gao, W.Y., D.G. Johns, and H. Mitsuya, Anti-human immunodeficiency virus type 1 activity of hydroxyurea in combination with 2',3'-dideoxynucleosides. Mol Pharmacol, 1994. 46(4): p. 767-72.

315. Preston, B.D., B.J. Poiesz, and L.A. Loeb, Fidelity of HIV-1 reverse transcriptase. Science, 1988. 242(4882): p. 1168-71.

316. Roberts, J.D., K. Bebenek, and T.A. Kunkel, The accuracy of reverse transcriptase from HIV-1. Science, 1988. 242(4882): p. 1171-3.

317. Amie, S.M., et al., Anti-HIV host factor SAMHD1 regulates viral sensitivity to nucleoside reverse transcriptase inhibitors via modulation of cellular deoxyribonucleoside triphosphate (dNTP) levels. J Biol Chem, 2013. 288(28): p. 20683-91.

318. Lisziewicz, J., et al., Hydroxyurea in the treatment of HIV infection: clinical efficacy and safety concerns. Drug Saf, 2003. 26(9): p. 605-24.

319. Fozza, C., The role of Clofarabine in the treatment of adults with acute myeloid leukemia. Crit Rev Oncol Hematol, 2015. 93(3): p. 237-45.

124 320. Jacoby, M.A., et al., Phase I study of oral clofarabine consolidation in adults aged 60 and older with acute myeloid leukemia. Am J Hematol, 2014. 89(5): p. 487-92.

321. Harrington, R.D. and A.P. Geballe, Cofactor requirement for human immunodeficiency virus type 1 entry into a CD4-expressing human cell line. J Virol, 1993. 67(10): p. 5939-47.

322. Vodicka, M.A., et al., Indicator cell lines for detection of primary strains of human and simian immunodeficiency viruses. Virology, 1997. 233(1): p. 193-8.

323. Rawson, J.M., et al., HIV-1 and HIV-2 exhibit similar mutation frequencies and spectra in the absence of G-to-A hypermutation. Retrovirology, 2015. 12: p. 60.

324. Chugh, P., et al., Akt inhibitors as an HIV-1 infected macrophage-specific anti- viral therapy. Retrovirology, 2008. 5: p. 11.

325. Boussif, O., et al., A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine. Proc Natl Acad Sci U S A, 1995. 92(16): p. 7297-301.

326. Clouser, C.L., S.E. Patterson, and L.M. Mansky, Exploiting drug repositioning for discovery of a novel HIV combination therapy. J Virol, 2010. 84(18): p. 9301-9.

327. Sirivolu, V.R., et al., Clicking 3'-azidothymidine into novel potent inhibitors of human immunodeficiency virus. J Med Chem, 2013. 56(21): p. 8765-80.

328. Kim, B., Genetic selection in Escherichia coli for active human immunodeficiency virus reverse transcriptase mutants. Methods, 1997. 12(4): p. 318-24.

329. Weiss, K.K., et al., A role for dNTP binding of human immunodeficiency virus type 1 reverse transcriptase in viral mutagenesis. Biochemistry, 2004. 43(15): p. 4490-500.

330. Kennedy, E.M., et al., Ribonucleoside triphosphates as substrate of human immunodeficiency virus type 1 reverse transcriptase in human macrophages. J Biol Chem, 2010. 285(50): p. 39380-91.

331. Chen, L.F., J. Hoy, and S.R. Lewin, Ten years of highly active antiretroviral therapy for HIV infection. Med J Aust, 2007. 186(3): p. 146-51.

332. Sluis-Cremer, N., D. Arion, and M.A. Parniak, Molecular mechanisms of HIV-1 resistance to nucleoside reverse transcriptase inhibitors (NRTIs). Cell Mol Life Sci, 2000. 57(10): p. 1408-22.

333. De Clercq, E., Non-nucleoside reverse transcriptase inhibitors (NNRTIs): past, present, and future. Chem Biodivers, 2004. 1(1): p. 44-64.

125 334. Detels, R., et al., Effectiveness of potent antiretroviral therapy on time to AIDS and death in men with known HIV infection duration. Multicenter AIDS Cohort Study Investigators. JAMA, 1998. 280(17): p. 1497-503.

335. Murphy, E.L., et al., Highly active antiretroviral therapy decreases mortality and morbidity in patients with advanced HIV disease. Ann Intern Med, 2001. 135(1): p. 17-26.

336. Chen, R., et al., Human immunodeficiency virus mutagenesis during antiviral therapy: impact of drug-resistant reverse transcriptase and nucleoside and nonnucleoside reverse transcriptase inhibitors on human immunodeficiency virus type 1 mutation frequencies. J Virol, 2005. 79(18): p. 12045-57.

337. Coffin, J., HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science, 1995. 267: p. 483-489.

338. Domingo, E., et al., Basic concepts in RNA virus evolution. Faseb J, 1996. 10(8): p. 859-64.

339. Clavel, F., E. Race, and F. Mammano, HIV drug resistance and viral fitness. Adv Pharmacol, 2000. 49: p. 41-66.

340. Blankson, J.N., D. Persaud, and R.F. Siliciano, The challenge of viral reservoirs in HIV-1 infection. Annu Rev Med, 2002. 53: p. 557-93.

341. Persaud, D., et al., Latency in human immunodeficiency virus type 1 infection: no easy answers. J Virol, 2003. 77(3): p. 1659-65.

342. Shirasaka, T., et al., Emergence of human immunodeficiency virus type 1 variants with resistance to multiple dideoxynucleosides in patients receiving therapy with dideoxynucleosides. Proc Natl Acad Sci U S A, 1995. 92: p. 2398-2402.

343. Shafer, R.W., et al., Combination therapy with zidovudine and didanosine selects for drug-resistant human immunodeficiency virus type 1 strains with unique patterns of pol gene mutations. J Infect Dis, 1994. 169(4): p. 722-9.

344. Matamoros, T., et al., Molecular determinants of multi-nucleoside analogue resistance in HIV-1 reverse transcriptases containing a dipeptide insertion in the fingers subdomain: effect of mutations D67N and T215Y on removal of thymidine nucleotide analogues from blocked DNA primers. J Biol Chem, 2004. 279(23): p. 24569-77.

345. Menendez-Arias, L., T. Matamoros, and C.E. Cases-Gonzalez, Insertions and deletions in HIV-1 reverse transcriptase: consequences for drug resistance and viral fitness. Curr Pharm Des, 2006. 12(15): p. 1811-25.

346. Cases-Gonzalez, C.E., et al., Mutational patterns associated with the 69 insertion complex in multi-drug-resistant HIV-1 reverse transcriptase that confer increased 126 excision activity and high-level resistance to zidovudine. J Mol Biol, 2007. 365(2): p. 298-309.

347. Mas, A., et al., Role of a dipeptide insertion between codons 69 and 70 of HIV-1 reverse transcriptase in the mechanism of AZT resistance. Embo J, 2000. 19(21): p. 5752-61.

348. Iversen, A.K., et al., Multidrug-resistant human immunodeficiency virus type 1 strains resulting from combination antiretroviral therapy. J Virol, 1996. 70(2): p. 1086-90.

349. Scherrer, A.U., et al., Predictors for the emergence of the 2 multi- nucleoside/nucleotide resistance mutations 69 insertion and Q151M and their impact on clinical outcome in the Swiss HIV cohort study. J Infect Dis, 2011. 203(6): p. 791-7.

350. Henry, M., et al., Vertical transmission of multidrug-resistant Q151M human immunodeficiency virus type 1 strains. Pediatr Infect Dis J, 2008. 27(3): p. 278- 80.

351. Dapp, M.J., R.H. Heineman, and L.M. Mansky, Interrelationship between HIV-1 fitness and mutation rate. J Mol Biol, 2013. 425(1): p. 41-53.

352. Rawson, J.M., C.L. Clouser, and L.M. Mansky, Rapid Determination of HIV-1 Mutant Frequencies and Mutation Spectra Using an mCherry/EGFP Dual- Reporter Viral Vector. Methods Mol Biol, 2016. 1354: p. 71-88.

353. Anastassopoulou, C.G., et al., Escape of HIV-1 from a small molecule CCR5 inhibitor is not associated with a fitness loss. PLoS Pathog, 2007. 3(6): p. e79.

354. Gervaix, A., et al., A new reporter cell line to monitor HIV infection and drug susceptibility in vitro. Proc Natl Acad Sci U S A, 1997. 94(9): p. 4653-8.

355. Dapp, M.J., et al., Discovery of novel ribonucleoside analogs with activity against human immunodeficiency virus type 1. J Virol, 2014. 88(1): p. 354-63.

356. Liu, Y., et al., A sensitive real-time PCR based assay to estimate the impact of amino acid substitutions on the competitive replication fitness of human immunodeficiency virus type 1 in cell culture. J Virol Methods, 2013. 189(1): p. 157-66.

357. Zhang, H., et al., Novel single-cell-level phenotypic assay for residual drug susceptibility and reduced replication capacity of drug-resistant human immunodeficiency virus type 1. J Virol, 2004. 78(4): p. 1718-29.

358. Maeda, Y., D.J. Venzon, and H. Mitsuya, Altered drug sensitivity, fitness, and evolution of human immunodeficiency virus type 1 with pol gene mutations

127 conferring multi-dideoxynucleoside resistance. J Infect Dis, 1998. 177(5): p. 1207-13.

359. Kosalaraksa, P., et al., Comparative fitness of multi-dideoxynucleoside-resistant human immunodeficiency virus type 1 (HIV-1) in an In vitro competitive HIV-1 replication assay. J Virol, 1999. 73(7): p. 5356-63.

360. Ueno, T., T. Shirasaka, and H. Mitsuya, Enzymatic characterization of human immunodeficiency virus type 1 reverse transcriptase resistant to multiple 2',3'- dideoxynucleoside 5'-triphosphates. J Biol Chem, 1995. 270(40): p. 23605-11.

361. Quinones-Mateu, M.E., et al., Insertions in the reverse transcriptase increase both drug resistance and viral fitness in a human immunodeficiency virus type 1 isolate harboring the multi-nucleoside reverse transcriptase inhibitor resistance 69 insertion complex mutation. J Virol, 2002. 76(20): p. 10546-52.

362. Huang, H., et al., Structure of a covalently trapped catalytic complex of HIV-1 reverse transcriptase: implications for drug resistance. Science, 1998. 282(5394): p. 1669-75.

363. Sarafianos, S.G., et al., Structure and function of HIV-1 reverse transcriptase: molecular mechanisms of polymerization and inhibition. J Mol Biol, 2009. 385(3): p. 693-713.

364. Helen M. Berman, J.W., Zukang Feng, Gary Gilliland, T. N. Bhat, Helge Weissig, Ilya N. Shindyalov, Philip E. Bourne, The Protein Data Bank. Nucleic Acids Research, 2000. 28(1): p. 235-242.

128

APPENDIX I

DECLARATION OF CONTRIBUTIONS TO CO-AUTHORED PUBLICATION: DISTINCT PARTICLE MORPHOLOGIES REVEALED THROUGH COMPARATIVE PARALLEL ANALYSES OF RETROVIRUS-LIKE PARTICLES

I am a co-author on the following publication:

Martin, J. L., Cao, S., Maldonado, J. O., Zhang, W., & Mansky, L. M. (2016). Distinct particle morphologies revealed through comparative parallel analyses of retrovirus-like particles. Journal of virology, 90(18), 8074-8084.

My contributions to this publication are as follows: • Produced and collected HTLV-1 biological samples • Partial analysis of HTLV-1 particle morphology • Generated and edited a draft of the manuscript

129 Introduction The Gag polyprotein is the primary retroviral structural protein responsible for orchestrating retrovirus assembly. Human immunodeficiency virus type 1 (HIV-1) Gag has arguably been the most extensively studied Gag polyprotein to date [2, 100, 172, 187, 251-254]; however, distinct differences in the biology of retroviral Gag proteins have emphasized the importance of comparative analyses in order to gain further insights [122]. The Gag polyprotein is functionally conserved across the two retroviral subfamilies (i.e. Orthoretrovirinae and Spumaretrovirinae). While the six genera in the Orthoretrovirinae subfamily (alpha-, beta-, delta-, epsilon-, gamma-, and lentiviruses) and single genus in the Spumaretrovirinae subfamily (spumaretrovirus) all encode for Gag and have similar genomic organizations, distinct differences among the Gag proteins impact their replication and assembly processes. Specifically, while the Gag polyprotein serves the same function for all seven genera, analysis of Gag tertiary structures reveal differences between the two subfamilies. Analysis of the Gag N-terminal domain of human foamy virus (HFV, spumaretrovirus) has revealed a structure that is unrelated to orthoretroviral Gag proteins, containing many β-sheet structures whereas orthoretroviral Gags contain α-helical structures [255]. The Gag polyprotein of the Orthoretrovirinae consists of three structurally distinct but functionally overlapping structural domains. The N-terminal domain of Gag (matrix [MA]) facilitates both Gag-membrane and Gag-RNA interactions [95]. The capsid (CA) domain is primarily responsible for Gag-Gag interactions, and the C- terminal region encodes for the nucleocapsid (NC) domain, which is particularly crucial for retroviral RNA packaging and stabilizing CA-CA interactions [256, 257]. Other domains are encoded by some orthoretroviral Gag proteins. For example, lentiviruses such as HIV-1 encode for the spacer peptide 1 (SP1) and spacer peptide 2 (SP2), which flank the NC domain, and the p6 domain, which is located at the C-terminus of HIV-1 Gag. During or at completion of virus budding, the viral protease is activated, cleaving the Gag polyprotein into mature viral proteins. MA remains associated with the viral membrane, and CA forms a capsid core structure that encapsulates the NC protein bound to the viral RNA along with reverse transcriptase and integrase. Unlike the orthoretroviral Gag proteins, spumaretroviral Gag proteins do not undergo extensive proteolytic processing and typically remain in a polyprotein form during the virus assembly process [258]. In this study, we sought to devise a parallel comparative analysis of Gag proteins and the resultant Gag-based VLPs from representatives of the various retroviral genera. The goal of this comparative analysis was to determine whether there were differences in the general subcellular distribution of the various Gag proteins and whether variations existed regarding the general morphology of VLPs produced in a parallel manner. Differences were observed among both the Orthoretrovirinae and the representative member of the Spumaretrovirinae. Key observations include the nuclear localization of epsilonretrovirus Gag, deltaretrovirus-like particles with flat regions of electron density representative of the immature Gag lattice, and HIV-2 lentivirus-like particles with a narrow range of particle size and consistent electron density below the viral membrane, suggesting a tightly packed Gag lattice. These results support the argument for how

130 comparative analyses can be critically important for gaining new insights into the retroviral assembly pathway.

Materials and Methods

Plasmids and cell lines. Codon-optimized gag genes of human immunodeficiency virus type 1 (HIV-1, lentiretrovirus), human immunodeficiency virus type 2 (HIV-2, lentiretrovirus), murine leukemia virus (MLV, gammaretrovirus), mouse mammary tumor virus (MMTV, betaretrovirus), human foamy virus (HFV, spumaretrovirus), and walleye dermal sarcoma virus (WDSV, epsilonretrovirus) were designed using the UpGene program [259] and synthesized by Genscript Co. (Piscataway, NJ). Gag genes were based on the following sequences: HIV-1 (bases 1-500 from GenBank accession # NP_057850.1), HIV-2 (bases 1103-2668 from GenBank accession # M30502.1), MMTV (bases 313-2088 from GenBank accession # AF033807.1), MLV (bases 357-1973 from GenBank accession # AF033811.1), HFV (bases 1869-2815 from GenBank accession # Y07725.1), and WDSV (bases 800-2548 in GenBenk accession # AF033822.1). Each gag gene expression cassette was created to contain a 5’ Kozak sequence (GCACCATG, start codon in bold) [260, 261]. HindIII and BamHI restriction sites were engineered at the 5’ and 3’ ends of the genes, respectively, for subcloning. Each Gag expression plasmid was created by cloning the gag gene into the peYFP-N3 vector, creating a carboxy-terminal Gag-eYFP fusion, or by cloning into the pN3 vector, where the eYFP tag was removed by deletion of the BamHI-NotI restriction fragment. The human T-cell leukemia virus type 1 (HTLV-1, deltaretrovirus) codon optimized Gag-eYFP plasmid has been previously described [180]. A codon-optimized Rous sarcoma virus (RSV, alpharetrovirus) gagpro was created (bases 381-2485 from GenBank accession # AF033808.1) using the same general strategy as that of the gag expression constructs in order to create peYFP-N3-RSV gagpro and pN3-RSV gagpro. The RSV pro sequence was removed from pN3-RSV gagpro by engineering a stop codon at position 2110 using primers 5’- CCTCCGGCCGTGTCCTAAGCGATGACCATGG-3’ and 5’- CCATGGTCATCGCTTAGGACACGGCCGGAGG-3’ (stop codon in bold). To create peYFP-N3-RSV gag, the following dinucleotide sequence was synthesized as a geneblock (Integrated DNA Technologies, Coralville, IA) containing a 5’ flanking BglII site and a 3’ flanking BamHI site: sequence 5’- CGGGATGGGGCATAACGCTAAGCAGTGCCGAAAGCGAGACGGGAACCAGGG ACAGCGCCCTGGCAGGGGGCTGTCCTCCGGCCCATGGCCGGGTCCCGAGCCT CCGGCCGTGTCCGGATCCATCGCCACCATGGTGAGC-3’. The restriction sites were used to clone the sequence into the N3-eYFP vector, removing the pro sequence. A codon-optimized WDSV env gene (bases 5974-9651 in GenBank accession # AF033822.1) was constructed as described above for the gag gene expression constructs. The HFV env plasmid, pCiES [262], was a kind gift from Maxine Linial. The MMTV env expression plasmid, Q61 [263], was a kind gift from Jackie Dudley and Susan Ross. The RSV env expression plasmid was kindly provided by Marc Johnson [264]. The HTLV-1 env expression construct, CMV-ENV, was kindly provided by Kathryn Jones and Marie- Christine Dokhelar [265]. The MLV env expression construct, SV-A-MLV-env has been previously described [266]. HeLa cells and HEK293T cells were purchased from ATCC

131 (Manassas, VA) and maintained in Dulbecco’s modified Eagles medium (DMEM) supplemented with 10% FetalClone III (FC3; GE Healthcare Lifesciences, Logan, UT).

Confocal microscopy. HeLa cells were grown in six-well plates with 1.5 mm glass slides coated with poly-L-lysine. The cells were transiently transfected with peYFP-N3-gag plasmids and pN3-gag plasmids at a 1:4 weight ratio using Genjet according to the manufacturer’s instructions (SignaGen, Gaithersburg, MD). The homologous env expression construct for each retroviral Gag (e.g. HIV-2 Env and HIV-2 Gag) was co- transfected into the HeLa cells at a plasmid weight ratio of 1:10 for the Env and Gag expression plasmids for RSV, HTLV-1, MLV, and HIV-2. The MMTV and WDSV Env and Gag expression plasmids were co-transfected at a weight ratio of 1:1, respectively. HFV was co-transfected at a Env to Gag expression plasmid weight ratio of 1:16. Sixteen to 24 hours post-transfection, cells were washed with PBS and fixed with 4% paraformaldehyde for 30 minutes. Cells were then washed with PBS containing 0.05% Triton X-100 (Sigma-Aldrich, St. Louis, MN). The actin proteins were stained using ActinRed™ 555 ReadyProbes® Reagent (Life Technologies, Gaithersburg, MD) according to the manufacturer’s inctructions. Coverslips were mounted on slides and preserved with ProLong® Diamond Antifade Mountant with DAPI (Life Technologies). Subcellular localization of Gag-eYFP was determined using a Zeiss LSM 700 confocal laser scanning microscope with a Plan-Apochromat 63x/1.40 NA Oil objective at 1.2 zoom. A range of WDSV Env and Gag expression plasmid co-transfection ratios were tested, and at least 50 cells were imaged for each transfection ratio. The percentage of cells containing YFP Gag outside of the nucleus were determined by qualitative evaluation of the images. Each replicate was normalized to the percentage of cells containing cytoplasmic Gag in the transfection of no Env and fold increase was calculated for 1:5 and 1:10 Env:Gag transfections.

Tranmission electron microscopy of 293T cells producing virus-like particles. Transfection of 293T cells with pN3-gag plasmids was done at a plasmid weight ratio of 10:1 for the Gag and Env expression plasmids, respectively, unless otherwise noted previously. Forty-eight hours post-transfection, cells were harvested and washed two times with 0.1 M sodium cacodylate. Cells were pelleted, fixed with 2.5% glutaraldehyde for 40 minutes, and were then washed three times with 0.1 M sodium cacodylate. Samples were then post-fixed with 1% OsO4 for 30 minutes. Cells were washed three times before they were dehydrated in a grade series of ethanol. Samples were then embedded in EMbed 812 resin. Ultrathin sections (65 nm) were stained with uranyl acetate and lead citrate. Samples were examined with a JEOL 1200EX II transmission electron microscope.

Cryo-electron microscopy of virus-like particles. Co-transfection of 293T cells with the pN3-gag expression constructs and their homologous Env expression plasmids was done using GenJet at a weight ratio of 10:1, respectively, unless otherwise noted previously. Forty-eight hours post-transfection, the cell culture supernatant was harvested and centrifuged at 3,000 x g for 5 minutes to remove large cellular debris. The supernatant was then passed through a 0.2 µm filter to remove any other debris or vesicles. VLPs were concentrated by ultra-centrifugation in a 50.2 Ti rotor (Beckman) at

132 35,000 x g for 90 min through an 8% OptiPrep™ (Sigma-Aldrich) cushion. The VLP pellets were resuspended in 200 uL 1X STE (10 mM Tris-Cl, pH 7.4, 100 mM NaCl, 1 mM EDTA). The concentrated VLPs were then centrifuged through a 10-30% OptiPrep™ gradient in a SW55 Ti rotor (Beckman) at 45,000 x g for 3 hours. The visible VLP band was extracted and pelleted in 1X STE at 35,000 x g for 1 hour using a SW55 Ti rotor. The pellet was resuspended in 10 ul 1X STE and frozen at -80 °C. The sample was thawed on ice prior to analysis by cryo-TEM. VLP samples were prepared for cryo-TEM as previously described [180]. Briefly, 3µl of concentrated VLP sample was applied to a glow-discharged c-flat holey carbon grid (Ted Pella, Redding, CA) and then blotted with filter papers to remove the sample excess. The grid was then plunged frozen into liquid ethane [184] with a FEI MarkIII Vitrobot system. The frozen grids were then transferred to a FEI TF30 field emission gun transmission electron microscope at liquid nitrogen temperature (FEI Company, Hillsboro, OR). Images were then recorded at a nominal magnification of 39,000 x and 59,000 x at low-dose (~30 electrons/Å2) and 1 to 5µm underfocus conditions using a Gatan 4k by 4k CCD camera (Gatan Inc., Pleasanton, CA).

Image analysis of virus-like particles. Cryo-TEM images of VLPs were used for all analyses. VLP diameter was measured using ImageJ software (NIH, Bethesda, MD). For each VLP, two perpendicular diameters were measured and used to calculate an average diameter. The histogram was generated using GraphPad Prism 5 software (GraphPad, La Jolla, CA). Radial density profiles were calculated using the Radial Profile Extended applet in ImageJ. The ImageJ circle tool was used to encompass a VLP of average size for each genus tested, with the center of the circle laying in the center of the particle. The average image intensity for each radial point was obtained from running the applet. This was repeated for three VLPs of the same size and shape. The background was subtracted, and both the average and standard deviation were normalized to 1 The radial density profiles for all three VLPs measured were aligned based on the electron dense peak for the lipid bilayer. The averaged radial density profiles were plotted using GraphPad Prism 5 software.

Results

Gag subcellular localization in cells. In order to investigate the subcellular distribution of the eight different retroviral Gag proteins, peYFP-N3-Gag expression constructs were engineered to express a Gag-YFP fusion for each of the retroviral genera. The tagged Gags were co-transfected at a 1:4 ratio with untagged Gags to maintain untagged Gag distribution phenotypes. Transiently transfected HeLa cells were analyzed using confocal microscopy 16 to 24 hours post-transfection. The cell perimeter was visualized using actin staining. All images shown are optical sections of a representative cell at the widest section of the cell. More than 50 cells were qualitatively analyzed before representative cells were selected for imaging. While some variability existed based on hours post- transfection and Gag expression level, Gag distribution within the cells remained fairly consistent within genera. RSV had puncta that appeared in the cytoplasm and around the plasma membrane in addition to a low level of diffuse fluorescence found in the nucleus (Figure 1A). HTLV-1, MLV, HIV-1 and HIV-2 all have puncta appearing both near the

133 cell membrane and throughout the cytoplasm (Figure 1C-F). Unlike the other retroviral Gag proteins analyzed, both MMTV and HFV were primarily found in the cytoplasm (Figure 1B and G), with very few puncta appearing around the membrane. This was expected for MMTV and HFV, as these Gag proteins are known to traffic to intracellular membranes to initiate particle assembly [253, 267]. HFV was also found on the apical side of the nucleus (Figure 1G, arrow). Cells transiently transfected with peYFP-N3-WDSV Gag primarily had fluorescence localized in or around the nucleus (Figure 1F). The subcellular localization of WDSV Gag has not been previously described, and the primarily nuclear localization appears novel among the other retroviral genera tested here. Diffuse eYFP-WDSV Gag was also observed to be closely associated with the nucleus but not overlapping with DAPI stained regions (Figures 1G and 2B, arrow). In the absence of WDSV Env expression, less than 15% of transfected cells were observed to have eYFP-WDSV Gag puncta in the cytoplasm. Co-expression of WDSV Gag and Env at 1:10 and 1:5 ratios, respectively, increased the proportion of cells expressing cytoplasmic Gag puncta (Figure 2B and 2C); however, the overall percentage of cells with Gag in the cytoplasm was still relatively low compared to other retroviral genera. This indicates that the WDSV Env influenced Gag localization in cells.

Thin-section transmission electron microscopy. Thin section TEM of 293T cells was performed to visualize VLPs produced by cells transiently transfected with a Gag and corresponding Env expression constructs. Forty-eight hours post-transfection, cells were fixed and stained in preparation for thin-section TEM. Electron densities, consistent with the immature Gag lattice, were observed in cells transfected with RSV, MLV, MMTV, HTLV-1, HIV-1, and HIV-2 Gags (Figure 3A-F). No evidence of particle assembly was observed in or around cells transfected with WDSV Gag, presumably due to the nuclear localization of this Gag protein. As expected, electron dense MMTV-like particles were observed primarily intracellularly and were associated with large, electron lucent vesicles (Figure 3A). This has been previously observed for viruses found in mouse mammary tumor cells and more recently demonstrated for Mason-Pfizer monkey viruses produced from a full-length MPMV genome [268, 269]. The electron densities consistent with the Gag lattice were approximately 90 nm in diameter and were abundant in transfected cells. Few electron densities resembling the MMTV Gag lattice were observed near the plasma membrane. Transfection with HIV-1, HIV-2, RSV, MLV, and HTLV-1 Gag resulted in VLP production at the plasma membrane (Figure 3B-F). HIV-2 VLPs were found budding at both the cell membrane and into intracellular vesicles, which has been observed before in 293T cells transfected with HIV-1 Gag [270]. Observations of cells transfected with HIV-1, HTLV-1, MLV, and RSV Gags correlated with previous observations for these viral proteins [231, 271, 272]. Unlike the other orthoretroviruses, VLPs produced by HFV Gag were not observed to be associated with a distinct electron density (Figure 3G), but particles with diffuse electron density were observed that were approximately 80-90 nm in diameter. In addition, these VLPs contained small electron dense regions around them that were consistent with HFV Env protein spikes (Figure 3G, arrow).

134 Cryo-TEM of Gag-based VLPs. Cryo-TEM was done to analyze VLPs produced from 293T cells that had been transiently transfected with Gag and Env expression constructs for each of the representative retroviruses under study. Forty-eight hours post- transfection, the VLPs were purified on an Opti-Prep gradient and visualized using cryo- TEM. VLPs were not readily produced by transfection of 293T cells with MMTV Gag or WDSV Gag, so no particle analysis was performed. RSV, HTLV-1, MLV, HIV-1, HIV- 2, and HFV VLPs were produced at sufficient levels to allow for morphological analyses (Figure 4A-F). With the exception of VLPs produced by HFV Gag (Figure 4E), all observed VLPs were spherical with electron density below the VLP membrane. In addition, VLPs produced from all Gag proteins resulted in a distribution of particle sizes, with a general range from 81-200 nm in diameter (Figure 5A-F). VLPs produced from RSV Gag were found to be 128 ± 13 nm in diameter (Figure 5A), which is similar to previously reported values [172]. The majority of RSV VLPs had electron density consistent with an immature Gag lattice that followed the curvature of the inner viral membrane, but there were some instances of particles where there were discontinuities in electron density or where the density did not strictly follow the curvature of the inner viral membrane (Figure 4A, arrow). Similar to RSV-like particles, the VLPs produced by HTLV-1 Gag expression also had regions where there were discontinuities in electron density below the viral membrane (Figure 4B). MLV-like particles also contained discontinuities in electron density beneath the viral membrane (Figure 4C, arrow), with an average particle diameter of 124 nm. Intriguingly, there appeared to be two distinct particle subpopulations among the MLV-like particles. The majority (~80%) of these particles were observed in the range of 70-140nm. A minority (~20%) of the particles were observed in the range of 160-210 nm. The particle morphology for both subpopulations was comparable (Figure 4C). Of the five orthoretrovirus VLPs observed, VLPs produced by HTLV-1 Gag were the smallest, with an average size of 116 ± 28 nm in diameter (Figure 5B). HTLV-1-like particles had regions of electron density which resembled the Gag lattice that were flat and did not follow the curvature of the lipid bilayer (Figure 4B, arrow) [208]. HIV-2-like particle were found to have an average particle diameter of 154 ± 20 nm in diameter (Figure 5E), which was significantly larger than the observed diameter of HIV-1-like particles (141 ± 35.1 nm, t-test, P = 0.0002). The diameter of HIV-1-like particles falls within the range of previous reports of the diameter of HIV-1 or HIV-1-like particles [101]. Compared to the other retrovirus-like particles in this study, the HIV-2- like particles had a narrow size distribution and a uniform particle morphology. In particular, these VLPs were observed to have a very organized electron density beneath the viral membrane. This organized density is suggestive of a tightly packed immature Gag lattice that closely follows the inner viral membrane. HFV-like particles had a unique particle morphology that was clearly distinguishable from the other retrovirus-like particles analyzed in this parallel comparative study. First, there was no distinguishable electron density found within the particles that would be consistent with an immature Gag lattice. The vast majority of particles observed were studded with electron dense projections (16 ± 2 nm) that were morphologically consistent with HFV Env proteins. In addition, more than 50% of the visible particle surface was decorated with these electron dense projections. It is interesting to note that the production of HFV-like particles was from a 1:16 co-

135 transfection of the HFV Env expression plasmid (pCiEs) to that of the HFV Gag expression plasmid (pN3-HFV-gag), respectively, which is a lower level of Env plasmid as compared to the other Gags. The HFV-like particles were also distinct in their relatively small size, with an average diameter of 82.5 ± 30 nm excluding the length of the Env projections (Figure 5F).

Radial density profile analysis. Particles characterized by cryo-TEM were further analyzed to determine their radial density profiles. For each retrovirus-like particle, 3 representative particles that were within 2 nm of each other in particle diameter were analyzed using the Radial Profile Extended applet in ImageJ. The integrated intensity units were corrected for background, and the average and standard deviation were normalized to a value of 1. The integrated intensity units for each of the three VLPs were then averaged and plotted with the standard error of the mean (Figure 6A-E). A distinct electron dense peak, which is representative of the lipid bilayer, was seen in all radial density profiles (indicated by the blue box) (Figure 6A-F). It is generally assumed that the MA domain for orthoretroviral Gags remains associated with the viral membrane and therefore has no distinct peak. Distinct electron density peaks were observed that are consistent with the CA NTD and CTD for the VLPs of RSV, HTLV-1, MLV, HIV-1, and HIV-2. In contrast, no electron density peaks consistent with the CA NTD or CTD were observed for HFV-like particles. Analysis of MLV- and RSV-like particles revealed a large space in electron density between the viral membrane and the presumed CA subunit, 12 and 18 nm respectively. This space in electron density could be due to a long flexible linker connecting the MA domain and the CA NTD (p12 in MLV and p10 in RSV) (Table 1) [273, 274]. In addition, the MLV- and RSV-like particles also have a relatively large space between the electron density that likely represent the inner and outer CA subdomains. While there were no distinct peaks visible that would be consistent for a CA NTD or CTD for HFV-like particles, which may be due to the lack of a typical orthoretroviral Gag conformation, there was a distinct density from 49 nm to 52 nm that would be consistent with the HFV Env protein on the surface of particles.

Discussion To our knowledge, this is the first comprehensive comparative study of retrovirus- like particle morphology performed in parallel using a Gag expression construct along with the corresponding retroviral Env expression plasmid (e.g., HIV-2 Gag expression plasmid and HIV-2 Env expression plasmid). In particular, we selected at least one representative virus from each of the retroviral genera to analyze the differences in Gag subcellular localization as well as in general particle morphology and size. A practical advantage of this experimental system is that it helps to limit variables related to biological differences in virus replication among the different retroviruses. While authentic viral RNA for each virus was not available for packaging into particles, the system is standardized in that the same cellular RNA is available for all RNA packaging, limiting cell type variability as well. Through the use of this system, we observed both distinct differences and similarities in the characteristics of the retrovirus-like particles produced, and the distribution of Gag in cells.

136 In this study, WDSV (epsilonretrovirus) Gag appeared to localize in the nucleus, and perhaps near the nucleus, which is a new observation. There have been no previous analyses of WDSV Gag localization or protein structure, but based on sequence homology, WDSV Gag has similarities to other orthoretrovirus Gag proteins; however, the distinct subcellular localization distinguishes this Gag. We also observed that the WDSV Env appears to influence WDSV Gag localization, as co-expression of both Gag and Env resulted in the cytoplasmic appearance of WDSV Gag, though the majority remained in the nucleus regardless of Env expression. When investigated by thin-section EM, there was no evidence of intracellular assembly of particles. The WDSV accessory protein, Orf A (rv-cyclin), has also been observed to localize in the nucleus and may interact with the Gag protein there [275]. Further studies are needed to better understand the nature of the nuclear localization of WDSV Gag, whether WDSV Gag nuclear localization may facilitate interaction with RNA as it does for RSV [276], and the general requirements for virus particle release from cells. Many previous studies have evaluated immature particle morphology of MLV (gamma-), HIV-1 (lenti-), RSV (alpha-), and MPMV (beta-) [202, 277] [172] [187, 278]. In contrast, limited information is available regarding MMTV (beta-), HTLV-1 (delta-), HIV-2 (lenti-), HFV (spuma-), and WDSV (epsilon-) immature or mature particle morphology. For example, MMTV particles were previously purified from murine cells and visualized by thin section microscopy [279], and authentic HTLV-1, HFV, and MLV have been visualized by cryo-TEM or cryo-electron tomography [103, 192, 280]. Our study provides, to our knowledge, the first information regarding HIV-2-like particle morphology, and no reports are available on WDSV particle morphology. Based on the structural similarities among most Gags, HIV-1 Gag trafficking and localization is often used as a reference model for many different retroviruses. Our findings here indicate that HIV-1 Gag packaging may not be a good model for many retroviral Gag studies, as we found a wide range of particle sizes and morphologies in our analysis. This is highlighted by our analysis of HIV-2 Gag-based VLPs, which is a lentivirus closely related to HIV-1. HIV-2 Gag has a similar subcellular distribution to what is seen for HIV-1 based on confocal analysis, but it produces particles that are distinct in size and morphology, indicating differences between particle assembly and/or the nature of the immature Gag lattice. Intriguingly, HIV-1 and HIV-2 Gags can coassemble together into the same VLPs, indicating that the Gags still share many similarities [281]. Our observations indicate that orthoretrovirus VLPs share some morphological features, such as a spherical shape, a distribution of diameter sizes, and distinct inner and outer CA ring densities. This is not particularly surprising given that retroviral CA structures are markedly similar even with little amino acid sequence homology. Despite these similarities, all of the orthoretrovirus VLPs observed here have distinguishing morphological features. These differences may be in part due to distinguishing features of the Gags outside of the CA domain, such as the linker domains that connect the RSV and MLV MA and CA, which can lead to a large distance from the lipid bilayer to the outer CA ring. We observed that this may be different for HIV-2, however, as the electron density remained closely associated with the viral membrane, possibly due to a smaller linker between HIV-2 MA and CA.

137 HFV VLPs were markedly different from the orthoretrovirus VLPs. This could be due to the nature of HFV particle assembly, which occurs intracellularly at the endoplasmic reticulum and requires expression of the homologous HFV Env protein [253, 282]. The requirement of Env protein expression for HFV-like particle assembly perhaps explains the high occurrence of spike-like protrusions on the surface of HFV-like particles. Furthermore, the lack of electron density that would correspond to an immature Gag lattice may be due to an unusual structure and folding of HFV Gag, which helps to distinguish spumaretroviral Gag proteins from orthoretroviral Gag proteins. Taken together, these parallel comparative analyses have identified distinct features that exist among retrovirus-like particles. In particular, differences were observed with WDSV, HTLV-1 and HIV-2. These studies emphasize the differences that exist among retroviruses in regards to particle assembly/morphology, and the importance of their study in order to gain deeper insights into retroviral assembly.

138

139 Figure 1. Cellular localization of transient Gag-YFP expression in HeLa cells. HeLa cells were transiently cotransfected with 4:1 weight ratios of untagged Gag to YFP- tagged Gag in addition to homologous envelope plasmids as described in Materials and Methods. (A) Rous sarcoma virus (RSV); (B) mouse mammary tumor virus (MMTV); (C) human T-cell leukemia virus type 1 (HTLV-1); (D) murine leukemia virus (MLV); (E) HIV-1; (F) HIV-2; (G) human foamy virus (HFV); (H) walleye dermal sarcoma virus (WDSV). Optical sections close to the bottom of the cells were standardly used for collecting representative images of Gag-YFP localization. Nuclei were identified with DAPI stain (blue), and actin filaments were identified with the ActinRed 555 ReadyProbes reagent (red). Gag-YFP expression was identified by green fluorescence. Scale bar, 10 µm.

140

Figure 2. Differential walleye dermal sarcoma virus (WDSV) Gag-YFP localization by WDSV Env expression. Differential walleye dermal sarcoma virus (WDSV) Gag- YFP localization by WDSV Env expression. HeLa cells were transiently transfected as described in Materials and Methods with 1:4 weight ratios of WDSV Gag-YFP and untagged WDSV Gag in addition to WDSV Env expression plasmids at ratios of either 1:10 or 1:5. Cells were fixed, DAPI stained, and stained with ActinRed 16 h posttransfection. Images collected by confocal microscopy were used to evaluate cells for subcellular localization of WDSV Gag-YFP. (A) Representative examples of HeLa cells with no WDSV Gag-YPF puncta observed in the cytosol from transfections with no Env.

141 (B) Representative examples of HeLa cells with WDSV Gag puncta observed in the cytosol from 1:10 or 1:5 cotransfections. Scale bar, 10 µm. (C) Fold increase in percentage of cells with Gag in the cytoplasm in different Env cotransfection ratios based on analysis of 50 cells (n = 3).

142

Figure 3. Transmission electron microscopy of 293T cells transiently transfected with retroviral Gag expression constructs. 293T cells were transiently transfected as described in Materials and Methods with a Gag expression construct along with the homologous Env expression construct. Cells were prepared and analyzed by transmission electron microscopy as described in Materials and Methods. Shown are virus-like particle productions from 293T cells of mouse mammary tumor virus (MMTV) Gag (A), HIV-2 Gag (B), HIV-1 Gag (C), Rous sarcoma virus (RSV) Gag (D), murine leukemia virus Gag (E), human T-cell leukemia virus type 1 (HTLV-1) Gag (F), human foamy virus (HFV) Gag (G). Scale bar, 200 nm.

143

144 Figure 4. Cryo-electron microscopy of retrovirus-like particles produced by transient expression of Gag in 293T cells. 293T cells were transiently transfected with selected retroviral Gag expression plasmids and homologous Env expression constructs (ratio of 10:1 for RSV, HTLV-1, MLV, HIV-1, and HIV-2; ratio of 16:1 for HFV) as described in Materials and Methods. Virus-like particles were collected and purified from the cell culture supernatant. (A) Rous sarcoma virus (RSV); (B) human T-cell leukemia virus type 1 (HTLV-1); (C) murine leukemia virus (MLV); (D) HIV-1; (E) HIV-2; (F) human foamy virus (HFV). Scale bar, 50 nm.

145

Figure 5. Analysis of the distribution of retrovirus-like particle diameters. Using images collected by cryo-TEM, retrovirus-like particle diameters were measured in two perpendicular directions and averaged. A minimum of two independently prepared particle preparations were used for the particle diameter analyses. (A) Rous sarcoma virus (RSV); (B) human T-cell leukemia virus type 1 (HTLV-1); (C) murine leukemia virus; (D) HIV-1; (E) HIV-2; (F) human foamy virus (HFV).

146

Figure 6. Radial density profile analysis of retrovirus-like particles. The solid black line indicates the average retrovirus-like particle density measured. The dashed red line indicates the standard error of the mean. The radial density profiles shown were averaged from three particles of median size. (A) Rous sarcoma virus (RSV); (B) human T-cell leukemia virus type 1 (HTLV-1); (C) murine leukemia virus (MLV); (D) HIV-1; (E) HIV-2; (F) human foamy virus (HFV). As points of reference, the blue horizontal box indicates the location of the lipid bilayer density, the purple box indicates the location of CA densities, and the orange box indicates the Env density.

147

Table 1. Summary of the dimensions of virus-like particles and Gag lattice structures based upon cryo-TEM.

148

APPENDIX II

DECLARATION OF CONTRIBUTIONS TO CO-AUTHORED PUBLICATION: DUAL ANTI-HIV MECHANISM OF CLOFARABINE.

I am a co-author on the following publication:

Daly, M. B., Roth, M. E., Bonnac, L., Maldonado, J. O., Xie, J., Clouser, C. L., Mansky, L. M. (2016). Dual anti-HIV mechanism of clofarabine. Retrovirology, 13, 20. doi: 10.1186/s12977-016-0254-0 .

My contributions to this publication are as follows: • Produced and collected HIV-1 RT mutant biological samples used in the study • Contributed text for samples generated and edited a draft of the manuscript

149 Background Deoxyribonucleotide triphosphates (dNTPs) are essential for the genomic DNA replication of all organisms. In mammalian cells, two pathways supply the cell with dNTPs: the de novo synthesis pathway and the salvage pathway. Ribonucleotide reductase (RNR) is the rate-limiting enzyme in de novo dNTP synthesis and acts by reducing ribonucleotides to deoxyribonucleotides. Nucleotide levels and cell cycle status tightly regulate the expression and activity of RNR [283, 284]. Prior to S phase, RNR activity greatly increases ensuring a sufficient supply of dNTPs for DNA replication. However, cells that are in nondividing or resting states display restricted RNR activity, which results in a low dNTP environment [285, 286]. In contrast, transformed/cancer cells, which are rapidly dividing with uncontrolled cell cycles, have significantly higher dNTP levels compared to normal dividing cells, and elevated dNTP levels are considered a biochemical marker for cancer cells [286, 287]. Various intracellular pathogens that synthesize DNA, including human immunodeficiency virus type 1 (HIV-1), use cellular dNTPs for their genome replication. Two primary target cells of HIV-1 are activated CD4+ T cells and macrophages. Activated CD4+ T cells are actively dividing cells that contain abundant dNTPs (1-5 µM), synthesized by the highly expressed RNR, to support their genomic replication [288- 290]. On the other hand, macrophages are terminally differentiated and nondividing. Due in part to their extremely low RNR expression they have substantially lower dNTPs (20- 50 nM) than dividing cells [289-291]. HIV-1 replication kinetics is much slower and restricted in macrophages than in activated CD4+ T cells [290, 292]. We previously reported that the low dNTP pools found in nondividing macrophages kinetically delays HIV-1 reverse transcription, suggesting that limited dNTPs serve as a restriction mechanism against HIV-1 in nondividing cells [290]. Recently, SAM domain and HD domain containing protein 1 (SAMHD1) was identified as a potent myeloid-specific host restriction factor of HIV-1 that depletes cellular dNTPs [293], which subsequently suppresses HIV-1 replication in nondividing cell types such as macrophages [294-296]. Interestingly, unlike HIV-1, HIV-2 and many SIVs encode a viral accessory protein, viral protein X (Vpx), that counteracts SAMHD1 and promotes retroviral replication in macrophages by elevating cellular dNTPs [295, 297, 298]. In addition to SAMHD1, other dNTP pool modulators have been shown to effect HIV reverse transcription. Most notably, cyclin-dependent kinase p21 inhibits HIV replication by repressing the expression of an alternative RNR subunit called RNR2 [299]. Overall, these data indicate that sufficient dNTP levels are necessary for viral reverse transcription and depleting dNTPs restricts viral replication. Clofarabine is an FDA approved RNR inhibitor (RNRI) used in the treatment of acute lymphoblastic leukemia. As a purine nucleoside analog (Figure 1A), clofarabine is transported into cells by nucleoside transporters and is phosphorylated by host enzymes to the active forms of the drug, clofarabine di- and triphosphate [300]. Both clofarabine di- and triphosphate inhibit RNR by binding the catalytic and the allosteric regulatory sites inducing hexamerization of the large subunit of RNR and subsequently preventing formation of the active enzyme. Due to the necessity of RNR in de novo dNTP synthesis, this inhibition causes a reduction in endogenous dNTPs, which in turn can inhibit DNA synthesis due to limited substrate availability [301, 302]. Additionally, clofarabine triphosphate (clofarabine-TP) is incorporated as an adenosine analog by DNA

150 polymerase-α and –ε and can induce chain termination. [303, 304]. Other DNA polymerases can also incorporate clofarabine albeit at a much-reduced rate. The mechanism for chain termination of clofarabine is not entirely clear; as clofarabine has a 3’ OH, it is not an obligate chain terminator like azidothymidine (AZT). It has been observed that incorporation of clofarabine-TP reduces the rate at which the next nucleotide will be incorporated and incorporation of two consecutive clofarabine-TPs makes it extremely unlikely that the DNA chain will be further elongated. The current hypothesis for this inhibition of extension is that the 2’ fluorine atom may affect the reactivity of the 3’ OH and/or the quaternary structure of the DNA such that the polymerase fails to efficiently make the next bond. We have previously demonstrated that clofarabine has anti-HIV activity in cell culture using transformed cells lines [305], however these studies did not examine the mechanism of action, or the anti-HIV activity and toxicity in the primary target cells of HIV: activated CD4+ T cells and macrophages. Our biochemical and cell culture data indicate that clofarabine effectively blocks HIV-1 replication by two distinct mechanisms: 1) reduction of cellular dNTPs and 2) direct incorporation by and inhibition of HIV-1 reverse transcriptase, with very limited toxicity in macrophages.

Results and Discussion Anti-HIV-1 activity and cytotoxicity of clofarabine in human primary target cells Clofarabine, a purine nucleoside analog (Figure 1A) and RNRI, is an FDA approved anticancer compound that we have recently shown to have antiretroviral potency in a transformed cell line [305]. Here, we examined the anti-HIV-1 activity of clofarabine in the primary human cells that are targets of HIV-1: activated CD4+ T cells and terminally differentiated monocyte derived macrophages (MDMs). Cells were isolated from five healthy donors, pretreated with varying concentrations of clofarabine for 8 hours, and then infected with HIV-1 pseudotyped with vesicular stomatitis virus G protein (VSV-G). The construct used expresses the full-length HIV-1 genome, with a frameshift in env and two fluorescent protein genes, mCherry and enhanced GFP (EGFP), replacing a portion of rev and nef [306]. Cells were analyzed with flow cytometry at 5 days (MDM) or 3 days (T cells) after the addition of virus, and infected cells were determined by EGFP expression. Macrophages, as expected, showed a more restricted HIV infection than the CD4+ T cells; however, similar infectivity was achieved by using much larger volumes of virus (Supplemental Figure 1A). As shown in Figures 1B and 1C (blue lines), clofarabine caused a concentration-dependent decrease in HIV-1 infection in both cells types, with half maximal inhibitory concentration (IC50) values of 21.6 nM (95% confidence interval (95% CI): 17.4 – 25.8 nM) in macrophages and 60.3 nM (95% CI: 24.1 – 96.5 nM) in activated CD4+ T cells. This three-fold increase in potency in macrophages compared to T cells is surprisingly minor – in the low dNTP environment of macrophages, we expected that the ratio of clofarabine-DP and -TP to dADP and dATP, respectively, would be much higher than that found in T cells, and therefore considerably more potent. However, this analysis is complicated by the fact that clofarabine transport into cells and conversion to the di- and tri-phosphate forms is also dependent on the cellular enzymes responsible for dNTP formation, all of which may be expressed at lower levels in macrophages than in T cells, where the need for dNTPs is much higher. In addition,

151 clofarabine-TP has recently been identified as a substrate for SAMHD1, which is highly expressed in macrophages but not T cells [307]. We also determined the cytotoxicity of clofarabine in activated CD4+ T cells and macrophages (red lines in Figure 1B and 1C) using the XTT assay, and found that macrophages are far more resistant to clofarabine-induced toxicity than activated CD4+ T cells, with CC50 values of 6.8 µM (95% CI: 3.2 – 9.4 µM) and 854 nM (95% CI: 713 – 996 nM), respectively. Additional toxicity assays, including analysis of membrane integrity and cell size, were performed and supported this result (Supplemental Figures 1B-E). This eight–fold difference in cytotoxicity indicates that macrophages are significantly more resistant to the toxic effects of clofarabine. This difference in toxicity could be due to the greater need for dNTPs in dividing cells such as activated CD4+ T cells, where nucleotide deprivation causes cell cycle arrest in S phase, often accompanied by cell death [308, 309]. Another possible explanation is that the degradation of clofarabine-TP by SAMHD1 in macrophages limits the impacts of mitochondrial toxicity, which is common for nucleotide reverse transcriptase inhibitors (NRTIs) [310]. Despite the fact that clofarabine-TP can be degraded by SAMHD1, clofarabine remains very potent in macrophages (IC50=20.3 nM) and has limited cytotoxicity in this cell type. The selectivity index (SI, CC50/IC50) for clofarabine in macrophages is 314.8, 22-fold greater than the SI in activated CD4+ T cells (Figure 1D), suggesting that clofarabine is a highly selective inhibitor of HIV-1 specifically in macrophages.

Effect of clofarabine on cellular dNTP levels and HIV-1 DNA synthesis We previously reported that the dNTP concentration in activated CD4+ T cells (1- 5 m M) is above the Km value of HIV-1 RT (100-200 nM) [290, 311]. On the other hand, macrophages harbor very limited dNTPs (50 nM) with concentrations that are below the Km value of HIV-1 RT, suggesting that the low dNTP levels kinetically delay HIV-1 reverse transcription in macrophages [290]. Clofarabine is a known RNR inhibitor (RNRI) that can deplete endogenous dNTPs in transformed cell lines [303, 304]. Here, we wanted to investigate the extent to which clofarabine depleted endogenous dNTPs in activated CD4+ T cells and macrophages. Specifically, we were interested in whether clofarabine would be effective in the extremely low dNTP environment in macrophages where RNR is not robustly expressed. We pretreated the two cells types with two different clofarabine concentrations, 10 nM (below IC50 values for both cell types, Figure 1B & C) or 300 nM (above IC50, but below CC50 values for both cell types), and methanol extracted the cellular dNTPs 8 hours later. To measure the cellular dNTP levels we utilized a single nucleotide primer extension assay that we previously developed [290]. Briefly, this assay uses a 5’ P32 radiolabelled 23-mer primer (P) annealed to one of four distinct 24-mer templates (T). The single nucleotide overhang on the 24-mer template (A, C, G or T) determines the dNTP to be measured. The template/primer was incubated with extracted cellular dNTPs and purified HIV-1 RT. The increase in radiolabelled 24-mer product indicates that the dNTP specific for the template has been incorporated. In three donors, clofarabine induced at least 50% reduction of dATP, dCTP and dGTP at 300 nM in activated CD4+ T cells (Figure 2A) and macrophages (Figure 2B). As with many RNRIs, dATP levels were the most affected by clofarabine-induced inhibition of RNR. This is possibly due to inefficient dATP synthesis via the salvage pathway, which does not involve RNR and is

152 not affected by RNRIs [312-314]. Consistent with other reports, TTP levels were the least affected in both cell types possibly due to synthesis of TTP from dCMP and dUMP via the salvage pathway [303, 309]. One caveat of this assay is that clofarabine-TP is present in the cells and could potentially be incorporated by HIV-1 RT. As there is only a single nucleotide overhang on the template, any inhibition of RT extension caused by clofarabine-TP incorporation would have no effect; therefore, the only predictable effect of clofarabine-TP incorporation would be an increase in the amount of dATP calculated. Despite this potential problem, we saw a strong depletion of dATP in clofarabine-treated cells. We also measured the effect of clofarabine on dNTP levels in MAGI cells (a transformed cell line) using a mass spectrometry-based assay, which would be unaffected by clofarabine triphosphate, and saw a similar dNTP depletion profile to that seen in activated CD4+ T cells (Supplemental Figure 2A). As shown in Figure 2C, the dNTP concentrations in untreated macrophages (M) are already well below the Km of HIV-1 RT, and the clofarabine-induced dNTP reduction (M+ clof) would be expected to further delay HIV-1 reverse transcription and inhibit viral infection (Figure 1C). However, the dATP concentration in activated CD4+ T cells (T) and in clofarabine-treated T cells (T + clof) remains above the Km value of HIV-1 RT, suggesting that the clofarabine-induced dNTP reduction in T cells should not significantly affect HIV-1 reverse transcription kinetics. Despite this, we do see a reduction in pseudovirus infection in CD4+ T cells at 300 nM clofarabine treatment (Figure 1B). One possible explanation is that the toxicity associated with clofarabine treatment may be responsible for the decrease in infection; however, we see potent anti- HIV-1 activity at levels that are not toxic, making this unlikely. Another possibility is it that clofarabine may be acting through a mechanism not directly related to its RNR inhibition. Next, we confirmed that the observed clofarabine-induced HIV inhibition (Fig 1B & 1C) is due to the inhibition of HIV-1 DNA synthesis by employing quantitative PCR to measure reverse transcription product. Clofarabine treatment led to a decrease in viral DNA synthesis that correlated with a decrease in infection (Figure 2D). These results are similar to those seen with azidothymidine (AZT), a known chain terminator. In contrast, the integrase inhibitor raltegravir (Ralt) decreased infectivity by approximately 80%, but this loss did not correlate to a reduction in viral DNA synthesis. These results indicate that clofarabine reduces reverse transcriptase-mediated viral DNA synthesis through, at least in part, its inhibition of RNR, which deplete endogenous dNTPs.

Incorporation of Clofarabine-TP by Reverse Transcriptase Our activated CD4+ T cell data indicated that the clofarabine-induced dNTP depletion would still provide a kinetically favorable environment for reverse transcription. It has been reported that clofarabine triphosphate (clofarabine-TP) is efficiently incorporated into DNA by DNA polymerase α and ε as a dATP analog [303, 304]. Therefore we tested whether HIV-1 RT can directly incorporate clofarabine-TP into DNA. For this test, clofarabine-TP and purified HIV-1 RT were incubated with a 5’ 32P- labeled 23-mer DNA primer (P) annealed to a 24-mer DNA template containing a single T overhang, which allows RT to incorporate a single clofarabine-TP. As shown in Figure 3A, a 24-mer extended product (E) was observed in the presence of clofarabine triphosphate at concentrations as low as 50 nM and the amount of 24-mer produced

153 increased with increasing concentrations of clofarabine-TP. These results demonstrate that clofarabine triphosphate is a substrate for HIV-1 reverse transcriptase. HIV-1 RT is known for its high error rate and low fidelity [315, 316], raising the possibility that clofarabine could act as a general purine nucleoside analog. Therefore, we also tested whether clofarabine could be incorporated as a dGTP analog by using a “C” overhang template/primer. In this assay, we did not observe any clofarabine-TP incorporation by HIV-1 RT (data not shown), indicating that clofarabine-TP is incorporated only as a dATP analog, not as a nonspecific purine analog. Next, we investigated whether the incorporation of clofarabine-TP by HIV-1 RT inhibits primer extension. For this experiment, reactions contained 5’ 32P-labeled 17-mer primer (P) annealed to a 38-mer RNA template, dNTPs (at either the concentrations found in activated T cells, 5 m M, or macrophages, 50 nM), increasing concentrations of clofarabine-TP, and excess HIV-1 RT. As shown in Figure 3B, when the primer was extended by HIV-1 RT at the T cell dNTP concentration (5 m M), pausing of DNA synthesis (* in Figure 3B) was observed only at the highest dose of clofarabine-TP (125 nM). However, at the macrophage dNTP concentration (50 nM), the RT pausing was observed at much lower clofarabine-TP concentrations, and the full-length extension was completely inhibited at the highest clofarabine-TP concentration (125 nM). As expected, the first stall site was observed across from a U site in the template (* in Figure 3B) indicating that clofarabine-TP was incorporated as an adenosine analog. More interestingly, the first stall site induced inefficient subsequent dNTP incorporation ( ] in Figure 3B), causing RT pausing and eventually blocking the full length DNA synthesis. These data indicate that HIV-1 RT can directly incorporate clofarabine-TP and that its incorporation inhibits processive viral DNA synthesis, particularly at the low dNTP environment found in non-dividing macrophages. There are numerous mutations in RT that render HIV-1 resistant to NRTIs, many of which confer multi-drug resistance. To determine whether these mutations would also be confer resistance to clofarabine, we treated MAGI cells with increasing amounts of clofarabine for 2 hours prior to infection with pseudotyped HIV-1 vectors containing a panel of known RT mutations. As shown in Figure 3C, clofarabine had similar potency against the NRTI-resistant RT mutants as against wild type (NL4-3 MIG). This could indicate that the primary mechanism of HIV-1 inhibition is RNR inhibition rather than incorporation by RT. However, many mutations in RT that confer resistance to individual NRTIs have no effect on different NRTIs, and some can in fact increase susceptibility to certain drugs. Selection experiments to develop a clofarabine-resistant strain are planned to further explore this question. In a final attempt to separate the two distinct mechanisms of clofarabine, reduction of dNTPs and direct inhibition of DNA synthesis via incorporation, we biochemically compared the DNA synthesis efficiency of HIV-1 RT at the cellular dNTP concentrations described in Figure 2C. For this, we extended a 5’ 32P-labeled 17-mer primer (P) annealed to a 38-mer RNA template with an equal amount of purified HIV-1 RT protein at three different dNTP concentrations: that found in T cells (T, 5 m M), in macrophages (M, 50 nM) and in macrophages that had been treated with 300 nM clofarabine (M/C, Figure 3D). This experiment was conducted in the absence of clofarabine so that we could observe the effect of clofarabine induced dNTP reduction on DNA synthesis in the absence of any other clofarabine-induced effects. As shown in

154 Figure 3D, HIV-1 RT generated the fully extended product (F) at T cell dNTP concentrations, whereas HIV-1 RT displayed limited primer extension with clear kinetic pausing (*) in the macrophage simulated dNTP concentrations. This RT-pausing was further enhanced at the dNTP concentrations observed in clofarabine treated macrophages, which ranged from 10-40 nM (Figure 3D). These data recapitulate previous data indicating that the dNTP pools in macrophages have a restrictive effect on reverse transcription [290, 296] and support our hypothesis that a further reduction in dNTPs will lead to a corresponding decrease in HIV-1 viral DNA synthesis. From these biochemical analysis, we suggest that clofarabine induced dNTP depletion in macrophages reduces reverse transcription (Figure 3D) while also increasing the likelihood of direct incorporation of clofarabine-TP during reverse transcription, which in turn can cause inhibition of processive DNA synthesis (Figure 3B). As summarized in Figure 4, we report that clofarabine, which is converted to clofarabine-DP and clofarabine-TP in cells, can inhibit HIV-1 reverse transcription by two distinct mechanisms: 1) reduction of cellular dNTPs through the inhibition of RNR (by clofarabine-DP and clofarabine-TP) and 2) direct incorporation of clofarabine-TP by HIV-1 RT and inhibition of viral DNA synthesis. Importantly, the inhibitory impacts of clofarabine imposed by these two mechanisms becomes more effective in non-dividing macrophages because this viral target cell type is less sensitive to the toxic effects of clofarabine. In addition, macrophages maintain limited dNTP pools, in part due to the dNTPase activity of host protein SAMHD1. Interestingly, many nucleoside analogs are not substrates for SAMHD1, which increases nucleoside reverse transcriptase inhibitor (NRTI) efficacy in macrophages [317]; however, it was recently published that clofarabine triphosphate is degraded by SAMHD1 in vitro [307]. Despite this, clofarabine was still able to reduce dNTPs and restrict viral replication efficiently in macrophages. Our previous publication reported that clofarabine is 3-fold more effective against HIV-2 than HIV-1 [Insert Ref Beach et al]. This result is surprising in light of our current report that clofarabine exerts its antiviral effect at least in part through reducing dNTP levels: the HIV-2 protein Vpx counteracts SAMHD1 and promotes retroviral replication in macrophages by elevating cellular dNTPs, and therefore would be expected to be more resistant to clofarabine than HIV-1, which does not contain Vpx. However, the previous work on HIV-2 was performed in U373-MAGI-CXCR4CEM cells, a transformed cell line derived from a human glioblastoma, which have abundant dNTPs and are highly permissive to HIV-1 infection. This indicates that the difference seen between HIV-1 and HIV-2 inhibition by clofarabine in the previous report is likely not Vpx or dNTP related but more likely related to differences in RT fidelity. Comparing the effect of clofarabine on HIV-1 and HIV-2 in primary human macrophages would be interesting, but we believe it would be difficult to explain the outcome mechanistically due to the SAMHD1- mediated hydrolysis of clofarabine-TP (31). Degradation of SAMHD1 by Vpx would lead to an increase in both dNTP and clofarabine-TP levels and could either increase or decrease the clofarabine TP:dATP ratio, and therefore the potency of clofarabine. A dual RNRI / HIV-1 RT inhibitor that is resistant to SAMHD1 hydrolysis would be expected to have significantly higher efficacy against HIV-1 than does clofarabine, and would be a much better tool to study the effects of Vpx on RNR inhibition. Any suggestion of using an RNRI as a clinical treatment for HIV-1 runs into the specter of hydroxyurea, an RNRI which failed clinically as an anti-HIV drug due

155 primarily to its accentuation of NRTI toxicity [318]. Given the self-potentiation inherent to clofarabine’s dual mechanisms of action, potentiation of toxicity would be a serious risk; however, there are significant differences between the two RNRIs. Hydroxyurea and clofarabine differ greatly in structure and in the mechanism of RNR inhibition. Clofarabine has been used as a second line cancer treatment since its FDA approval in 2004 and recent trials have had positive results using it as a first line cancer treatment that can be administered orally [319, 320]. Any use of clofarabine as an HIV drug would require further drug interaction studies to establish the safety of clofarabine both alone and with commonly used combination therapies. The current study was designed to study the mechanism of action and toxicity of clofarabine as a model for dual RNR / HIV-1 RT inhibitors in primary human target cells. We propose that clofarabine be used as a base molecule for further drug design studies. A derivative that shows resistance to SAMHD1 hydrolysis while maintaining the activity and low toxicity of clofarabine in macrophages would be of great interest. This would increase drug efficacy by reducing the natural dNTP competition and ensure that drug concentrations remain high for maximum efficacy. In conclusion, our experimental observations of cellular dNTP level reduction by clofarabine (below Km value of HIV-1 RT) and more effective incorporation of clofarabine-TP at these lowered dNTP concentrations support that clofarabine can be a model agent for dual function anti-HIV agents.

Methods

Cell lines, plasmids, and chemicals The 293T cell line was obtained from the American Type Culture Collection (Manassas, VA) while U373-MAGI-CXCR4CEM (MAGI) cells were obtained from Dr. Michael Emerman through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH [321, 322]. The HIV plasmid pNL4-3 MIG has been previously described [306]. RT resistant mutants were created as previously described [insert Dapp et al 2013], and subcloned into pNL4-3 MIG. The VSV-G envelope plasmid, pHCMV- VSVG, was provided by J. Burns (University of California, San Diego). Clofarabine was purchased from Carbosynth (Berkshire, UK). The following reagents were obtained from the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH: Raltegravir (Cat # 11680) from Merck & Company, Inc. and Zidovudine.

Cell Culture 293T and MAGI cells were maintained as previously described [323]. Primary human monocytes and CD4+ T cells were isolated from peripheral blood buffy coats by positive selection using CD14 or CD4 beads (Miltenyi Biotec, San Diego, CA), as previously described [324], and maintained in RPMI medium with 10% FCS and penicillin/streptomycin. CD14+ monocytes were matured to macrophages using 5 ng/mL human GM-CSF for 7 days (Miltenyi Biotec). CD4+ T cells were activated with 5ug/ml phytohemagglutinin (Sigma-Aldrich) and 20 units/ml IL2 (Miltenyi Biotec) for 1 day, then 20 units/ml IL2 alone for 5 days.

156 Production of viral stocks 293T cells were co-transfected with pNL4-3 MIG and pHCMV-VSVG using linear polyethylenimine from Polysciences, Inc. (Warrington, PA) as previously described [325]. Virus was harvested at 48 and 72 hours post-transfection and used immediately or frozen at −80°C.

Drug Treatments and Infections MAGI cells were infected as described previously [323]. Cells were treated with clofarabine for 2 h prior to infection with viral supernatant. Cells were harvested for analysis 48-72 h after infection and analyzed with a BD Biosciences LSRII flow cytometer. Vehicle treated cells had 15-35% infection. Primary cells were treated with clofarabine for 8 hours, washed twice with PBS and then infected with viral supernatant. Monocyte derived macrophages (MDM) were collected 5 days post-infection and activated CD4+ T cells were collected 3 days post-infection for analysis via flow cytometry. Data analysis was performed using FloJo and Prism 6. IC50 and CC50 were determined using GraphPad Prism nonlinear fit analyses; log (inhibitor) vs. response- variable slope.

Cell Proliferation and Cytotoxicity MDMs and T cells were treated with clofarabine for 8 hours, washed twice with PBS and maintained in media 5 or 3 days, respectively (same treatment as infection protocol). The XTT assay from ATCC was used as per the manufacturer’s protocol. The optimized assay incubation time for both cell types was 5 hours.

Real-time qPCR of RT products Real time qPCR was performed essentially as previously described [326]. MAGI cells (150,000 per well) were plated on a six-well plate. Twenty-four hours later, cells were treated with the indicated drug for 2 h prior to infection. As a control, an aliquot of virus was heat inactivated for 30 min at 95°C. Eighteen hours after infection, cells were harvested with trypsin; half were re-plated for the analysis of infection, and DNA was isolated from remaining cells using HighPure PCR Template Preparation Kit (Roche, Basel, Switzerland). Quantitative PCR (qPCR) mixtures contained 4 µl of eluted DNA using iTaq Universal SYBR Green Supermix according to the manufacturer's suggestions (BioRad, Hercules, CA). Primers for 18S rRNA were used to normalize sample-to-sample variation. The primers used to detect late reverse transcription (RT) products were 5′TGTGTGCCCGTCTGTTGTGT (forward) and 5′GAGTCCTGCGTCGAGAGAGC (reverse). The primers used to detects 18S rRNA were 5′GTAACCCGTTGAACCCCATT (forward) and 5′CCATCCAATCGGTAGTAGGG (reverse). The conditions for amplification were 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s.

Synthesis of clofarabine triphosphate Clofarabine triphosphate was synthesized as previously described [327] with minor modifications. In separate round-bottom flasks, clofarabine (0.1 g, 1.0 eq.) and tributylammonium pyrophosphate (0.361 g, 2.0 eq) were dried under high vacuum for 1 h

157 at ambient temperature. Throughout the entire experiment, the reaction was maintained under an argon atmosphere. Clofarabine (0.1 g, 1.0 eq.) was dissolved in anhydrous dioxane (0.9 mL, 3 mL/mmol) and anhydrous pyridine (0.3 mL, 1 mL/mmol). The flask was cooled with an ice bath. A solution of 2-chloro-4H-1,3,2- benzodioxaphosphorin-4- one (80 mg, 1.2 eq.) in anhydrous dioxane (0.4 mL, 1mL/mmol) was added and the mixture was then stirred at room temperature for 15 min. A solution of tributylammonium pyrophosphate (361 mg, 2.0 eq.) in anhydrous DMF (0.7 mL, 1 mL/mmol) was added to the mixture, prior to quick addition of tributylamine (0.4 mL, 5.0 equiv). The mixture was stirred for 15 min at room temperature. A solution of iodine (1% solution in pyridine/ water, 9:1) was then added dropwise until the permanent brown color of iodine persisted, and the mixture was stirred for 20 min. The excess of iodine was quenched with a 5% aqueous solution of Na2S2O3. The reaction mixture was evaporated to dryness under vacuum, dissolved in 25% ammonia solution, and stirred for 1 h at room temperature. The reaction was monitored by TLC (Rf 0.1 isopropanol/ aqueous ammonia = 1/1 and MS). The reaction mixture was concentrated under reduced pressure and resulting crude product was purified on a short pad of silica gel with a gradient of isopropanol/aqueous ammonia 8/2 to 1/1 affording the desired 5’-triphosphate ammonium salt, 21% yield (75% pure with diphosphate as an impurity, determined by 31P NMR). The synthesized nucleoside 5’-triphosphate was confirmed by 1H NMR, 31P NMR, and HR-MS analyses consistent with published analytical data for clofarabine 5’- triphosphate [301].

HIV-1 RT Purification HIV-1 (NL4-3) RT homodimer with a hexahistidine tag was expressed using an overexpression system in in BL21 E.coli [328, 329], and purified using Ni2+ chelation chromatography as described previously [290, 329]. dNTP quantification dNTPs from primary cells were extracted and quantified using the protocol previously described by Diamond et al [290]. MAGI cells were treated with 100 nM or 300 nM clofarabine for 8 hours, then harvested, counted, resuspended in 60% methanol, and stored at -20 °C for 18 hours. Samples were then vortexed, heated to 95 °C for 3 minutes, and centrifuged at 16,000xg for 5 minutes. Supernatant was transferred to new tubes and dried in an Eppendorf Vacufuge. Samples were then stored at -80 °C. To perform LC-MS/MS analysis, dried extracts were reconstituted in water at a concentration of 1 million cells per 100 µL. For each reconstituted sample, a 50 µL aliquot of resuspension was added to an Eppendorf tube containing 50 µL of internal standard (10 µM 5-iodo-dCTP in water), and then diluted with 100 µL water. The samples were then centrifuged at 14,000 rpm for 5 min at 4 °C. LC-MS/MS was used to determine levels of TTP, dGTP, dCTP and dATP in the supernatants following a previous published method [Reference: Cohen S et al., Journal of Chromatography B, 2009, 877: 3831-3840] with minor modifications. The LC-MS/MS system consists of an AB Sciex QTrap 5500 mass spectrometer and an Agilent 1260 Infinity HPLC. The chromatographic separation of analytes was achieved using a Thermo Scientific Hypercarb column (100 × 3 mm, 5 µm). The two eluents were: (A) 0.5 % diethylamine in water, pH adjusted to 10 with acetic acid; and (B) 50% acetonitrile in water. The mobile

158 phase was delivered at a flow rate of 0.5 mL/min using stepwise gradients of A and B: 0– 20 min, 0-25% B (v/v); 20-28 min, 25-50% B (v/v); 28-28.5 min, 50-95% B (v/v); 28.5- 30.5 min, 95-95% B (v/v); 30.5-31 min, 95-0% B, (v/v); 31-39 min, 0-0% B (v/v). Only eluate from 10-30 min was diverted into the mass spectrometer for analysis. MS/MS detection of the analytes was conducted using an ESI ion source with MRM detection in negative mode. The curtain gas was set at 20 psi. The ionspray voltage was set at -4500 V, and the temperature at 650 °C. The nebulizer gas (GS1) and turbo gas (GS2) were both set at 45 psi.

In vitro clofarabine triphosphate incorporation assay This single nucleotide extension assay was modified from a previously described assay [329]. A 5’ 32P-labeled DNA 18-mer (5’-GTCCCTCTTCGGGCGCCA-3’) was annealed to a 19-mer DNA template (3’-CAGGGAGAAGCCCGCGGTG-5’) at a 1:2 ratio. Extension from an 18-mer to 19-mer indicates that clofarabine triphosphate has been incorporated by HIV-1 reverse transcriptase. 20 µl reactions contained 200 fmol template/primer, 2 µl clofarabine triphosphate at concentrations indicated or 50 µM of dNTPs for the positive control, 4 µl of purified RT (HIV-1 NL4-3), 25mM Tris–HCl, pH 8.0, 2 mM dithiothreitol, 100 mM KCl, 5 mM MgCl2, and 10 µM oligo(dT). Reactions were incubated at 37° C for 5 min and then quenched with 10 µL of 40 mM EDTA and 99% (vol/vol) formamide at 95 °C for 2 min. The reactions were resolved on a 20% urea- PAGE gel (American Bio Sequel NE reagent) and imaged using Pharos FX molecular imager (Biorad).

Primer Extension Assay An HIV-1 RT primer extension assay was performed as previously described with slight modifications [330]. A 5’ 32P-labeled 17-mer DNA primer (5’- CGCGCCGAATTCCCGCT-3’) was annealed to a 40-mer RNA template (5’- AAGCUUGGCUGCAGAAUAUUGCUAGCGGGAAUUCGGCGCG-3’) in the presence of 100 mM NaCl, 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA. 20 µl Reaction mixtures contained 10 nM template-primer, 4 µl of purified HIV-1 RT, 5 µM or 50 nM of all four dNTPs (ThermoScientific), 12.5 mM Tris-HCl (pH 7.5), 12.5 mM NaCl, 2.5 mM MgCl2 and 20 µM oligo(dT). Reactions were initiated upon addition of RT and incubated at 37° C for 5 min. Reactions were terminated with 10 µl of 40 mM EDTA, 99% formamide and the products were resolved on a 14% urea-PAGE gel (American Bio Sequel NE reagent) and imaged using Pharos FX molecular imager (Biorad) and analyzed using ImageLab software.

159

Figure 1. Anti-HIV-1 activity of clofarabine in primary human activated CD4+ T cells and monocyte derived macrophages. a The structure of clofarabine. b Clofarabine inhibition (blue lines) and cytotoxicity (red lines) on activated CD4+ T cells of 5 healthy donors. Cells were treated with increasing concentrations of clofarabine for 8 h, washed with PBS, and then infected with pseudotyped HIV-1, (inhibition) or cultured with media (cytotoxicity). Analysis was conducted at 72 h post-infection via flow cytometry (inhibition) or XTT assay (cytotoxicity). The IC50 is 60.3 nM with a 95 % confidence interval (95 % CI) of 24.1–96.5 nM; the CC50 is 854.2 nM with a 95 % CI of 712.6– 995.8 nM. c The clofarabine inhibition (blue lines) and cytotoxicity curves (red lines) for monocyte-derived macrophages of 5 healthy donors. Macrophages were treated as described for T cells except analysis was at 5 days post infection. IC50 = 21.6 nM (95 % CI 17.4–25.8 nM): CC50 = 6.8 µM (95 % CI 3.2–9.4 µM). d Selectivity Index (SI) difference between activated CD4+ T cells and macrophages. SI values were determined by dividing the average CC50 of five donors by the average IC50 of five donors.

160

Figure 2. Clofarabine induced depletion of cellular dNTPs and inhibition of reverse transcription. Effect of clofarabine on cellular dNTP levels in primary activated CD4+ T cells (a) and macrophages (b) from three healthy donors. Activated CD4+ T cells and macrophages were treated with the indicated concentration of clofarabine for 8 h, washed with PBS, and dNTPs were extracted for analysis. dNTP levels are expressed as a percentage of the vehicle control (DMSO). Each value represents an individual donor with mean ± SD indicated. *p < .05 compared to vehicle control (multiple t test with Holm-Sidak post hoc test). c Cellular dATP concentrations of untreated and clofarabine (300 nM) treated T cells and macrophages from three healthy donors. Measured dATP levels were converted to cellular dATP concentrations as previously described [8], and are expressed as mean ± SD). Blue barindicates the Km range of HIV-1 RT [8, 31]. T: untreated T cells, T + clof: T cells + 300 nM clofarabine, M: untreated macrophages, M + clof: macrophages + 300 nM clofarabine. d Clofarabine effect on HIV-1 proviral DNA synthesis. MAGI cells were incubated with DMSO (NI, HI, ND), 200 nM AZT, 300 nM clofarabine (Clof), or 50 nM raltegravir (Ralt) for 2 h prior to infection with pseudotyped HIV-1. Infection (black bars) was determined at 48 h post-infection, and reverse transcription (grey bars) efficiency was determined at 18 h post-infection. Data represent mean ± SD from three independent experiments and are expressed as percentage of vehicle control (DMSO, 100 %). NI no infection, HI heat inactivated virus, ND no drug.

161

Figure 3. Biochemical examination of the dual mechanism of clofarabine. a Direct clofarabine-TP incorporation by HIV-1 RT. A 5′ 32P-labeled 23-mer DNA primer (P) annealed to a 24-mer DNA template with a single T overhang was incubated with HIV-1 RT and increasing concentrations of clofarabine-triphosphate (Clof-TP). E, Extended product; +, 50 µM dATP positive control; −, no dATP control. b Effect of clofarabine-TP incorporation on DNA synthesis. A 5′ 32P-labeled 17-mer DNA primer (P) annealed to a 38-mer RNA template was extended by HIV-1 RT with fixed dNTP concentrations found in either activated T cells (T cell, 5 µM) or monocyte-derived macrophages (MDM, 50 nM) with increasing concentrations of clofarabine-TP (two-fold dilutions starting at 125 µM). F, Fully extended product; +, 50 µM dNTP positive control; −, no dNTP control. Blue asterisks (*) indicates the clofarabine-TP incorporation site followed by kinetic pauses (]). c Clofarabine inhibition of NRTI-resistant RT mutants. MAGI cells were treated with increasing concentrations of clofarabine for 2 h prior to infection with Vsvg-pseudotyped HIV-1 containing mutations in RT. Flow cytometry analysis for infected cells was conducted at 48–72 h post-infection. IC50 values and 95 % confidence intervals are shown. NL4-3 MIG: wild-type HIV-1 RT, Q151: A62V, V75I, F77L, 162 F116Y and Q151M, T69: M41L, A62V, T69S, K70R, T215Y and serine–serine insertion between 69 and 70. d Biochemical simulation of HIV-1 RT activity at dNTP concentrations found in cells with and without clofarabine treatment. A 5′ 32P-labeled 17- mer DNA primer (P) annealed to a 38-mer RNA template (shown in box) was extended using an equal amount of purified HIV-1 RT protein with dNTP concentrations found in T cells (T, 5 µM), macrophages (M, 50 nM) or macrophages treated with 300 nM clofarabine (M/C, 10 nM dATP, 28 nM dCTP, 28 nM dGTP, 40 nM TTP). Blue stars (*) indicate kinetic pause sites, F, fully extended 38 bp product; −, no dNTP control.

163

Figure 4. Model for the anti-HIV-1 dual action mechanisms of clofarabine in macrophages. Clofarabine di- and triphosphate inhibit RNR to reduce dNTP levels, leading to the kinetic suppression of HIV-1 reverse transcription (blue). Clofarabine triphosphate is also directly incorporated by HIV-1 RT, inhibiting extension (red).

164

Figure S1. Infectivity and toxicity in T cells and macrophages. (A) Activated CD4+ T cells and macrophages were infected with pseudotyped HIV-1. In order to achieve similar levels of infection more pseudovirus was used to infect macrophages (5X more virus used). Individual donor infectivity is shown with technical triplicates shown as standard error (mean with standard deviation). (B) Cell viability in acti- vated CD4+ T cells. Cells were treated with varying amounts of clofarabine for 8 h, washed in PBS and maintained in media for 72 h. Cell viability was determined by exclusion of trypan blue indicating membrane integrity by the Vi-cell counter. (C) Cell size of activated CD4+ T cells treated with clofarabine. Cells were treated with varying amounts of clofarabine for 8 h, washed in PBS and maintained in media for 72 h. Cells were counted and cell size was determined using the Vi-Cell counter. (D) Cell viability in macrophages. Cells were treated with varying amounts of clofarabine for 8 h, washed in PBS and maintained in

165 media for 5 days. Cell viability was determined by exclusion of trypan blue indicating membrane integrity by the Vi-cell counter. (E) Cell size of activated CD4+ T cells treated with clofarabine. Cells were treated with varying amounts of clofarabine for 8 h, washed in PBS and maintained in media for 5 days. Cells were counted and cell size was determined using the Vi-Cell counter.

166

Figure S2. Clofarabine induced depletion of cellular dNTPs in MAGI cells. MAGI cells were treated with 100 nM (~IC50) and 300 nM (~IC90) clofarabine for eight hours, washed with PBS, and dNTPs were methanol extracted and analyzed by LC–MS/MS. Data shown repre- sents mean ± SD of three independent experiments, and is expressed as a percentage of vehicle control. * = p < .05 compared to vehicle control (multiple t test with Holm-Sidak post hoc test).

167

APPENDIX III

DECLARATION OF CONTRIBUTIONS TO FIRST AUTHORED PUBLICATION: THE HIV-1 REVERSE TRANSCRIPTASE A62V MUTATION INFLUENCES REPLICATION FIDELITY AND VIRAL FITNESS IN THE CONTEXT OF MULTI-DRUG RESISTANCE MUTATIONS

I am the first author on the following publication:

Maldonado, J. O., & Mansky, L. M. (2018). The HIV-1 Reverse Transcriptase A62V Mutation Influences Replication Fidelity and Viral Fitness in the Context of Multi-Drug- Resistant Mutations. Viruses, 10(7), 376.

My contributions to this publication are as follows: • Assisted in the design of the study • Performed the experiments • Analyzed the data • Drafted the manuscript

168 1. Introduction Since the discovery of the human immunodeficiency virus type 1 (HIV-1), the etiological agent of acquired immunodeficiency syndrome (AIDS), in the early 1980s, approximately 30 million individuals have died of AIDS, and approximately 36 million are currently infected worldwide. Current treatment for HIV-1 infection consists of a combination drug therapy (highly-active antiretroviral therapy, HAART) [331]. HAART typically consists of nucleoside reverse transcriptase inhibitors (NRTIs) [332], non- nucleoside reverse transcriptase inhibitors (NNRTIs) [333], and protease inhibitors (PIs) or integrase inhibitors which target key steps of the viral life cycle. HAART dramatically reduced the rate of HIV-1 and AIDS-related morbidity and mortality [334, 335]. A key drawback of drug therapy is antiretroviral drug resistance, which is associated with the acquisition of drug-resistant mutations [336]. The high mutation rate of HIV-1 (i.e., 3.4 × 10−5 mutations per target bp per replication cycle) [160] arguably contributes to the evolution of drug resistance. This high mutation rate can, in principle, lead to the generation of viral genomes, each day, which possess every possible mutation within an infected individual, thus allowing for the selection of drug-resistant mutations from this population [337]. Selection of virus variants with drug-resistant mutations that confer greater fitness for replication emerge during selective drug pressure. Fitness is defined as a parameter describing the capacity of an organism to adapt and replicate in a given environment (reviewed in Reference [338]). Two general types of mutations associated with drug-resistant phenotypes are (1) primary mutations that confer direct drug resistance, and( 2) secondary mutations that emerge during continued drug-selective pressure [339]. The former types of mutations usually correlate with a decrease in viral fitness, while the latter may have no discernable phenotype, but could include adaptive mutations that improve fitness. Latently infected cells can harbor drug-resistant viruses [340, 341], which are an obstacle to antiretroviral therapy. Understanding HIV-1 population dynamics in the context of drug-resistant mutations is essential to better predict viral disease progression to AIDS, durable antiretroviral drug regimens, and vaccine development. HIV-1 reverse transcriptase (RT), which is the enzyme responsible for converting the single-stranded viral RNA genome into double-stranded DNA, was previously observed to acquire the A62V amino acid substitution, which is known to be associated with multi- drug resistance but is not a resistance-conferring mutation [342, 343]. In particular, A62V is normally seen in different mutational arrangements, located mostly on the flexible β3– β4 loop region of the fingers sub-domain of HIV-1 RT, including the multi- dideoxynucleoside resistant (MDR) Q151M complex (i.e., A62V, V75I, F77L, F116Y, and Q151M) [342, 343] and the T69SSS insertion complex, which has a serine–serine insertion between the amino acid positions 69 and 70 (i.e., M41L, A62V, T69SSS, K70R, and T215Y) [344-346]. Mutational insertions or deletions in the β3–β4 loop region may confer multi-drug resistance [347]. The Q151M complex and the T69SSS insertion complex confer resistance to most of the NRTIs currently used for treatment including didanosine, zalcitabine, stavudine, and zidovudine (AZT), while the Q151M complex additionally confers resistance to lamivudine and abacavir [346-348]. These MDR complexes may lead to higher mortality rates, and can be transmitted from mother to child [349, 350]. An initial study reported that the A62V mutation alone increases HIV-1 mutant frequencies, and causes a minor decrease in virus fitness [351].

169 In this study, we sought to determine the role of A62V in HIV-1 viral mutagenesis and fitness in the context of clinically relevant drug-resistant mutations. To do this, we first used a single-cycle assay to assess the mutant frequency of the HIV-1 MDR complex mutants (i.e., the Q151M complex and the T69SSS insertion complex) in the presence or absence of the A62V substitution. The inclusion of the A62V substitution increased the observed lower mutant frequencies with each of the HIV-1 MDR complex mutants. We also introduced these mutations into an HIV-1 NL4-3 infectious molecular clone, and assessed its impact on viral fitness. Notably, the inclusion of A62V in the context of various MDR mutation complexes improved viral fitness in the presence of AZT. These observations together implicate an adaptive role for A62V in enhancing virus persistence during drug-selective pressure.

2. Materials and Methods

2.1. HIV-1 Vectors and Cell Lines The HIV-1 dual-reporter vector, pNL4-3 MIG, which expresses mCherry and the enhanced green fluorescent protein (EGFP), was previously described [306, 352]. This HIV-1 vector was co-transfected with a vesicular stomatitis virus G protein (VSV-G) envelope expression plasmid (HCMV-G; kindly provided by J. Burns, University of California, San Diego, CA, USA), into 293T cells to produce an infectious vector virus for use in a single-cycle replication assay (Figure 1A). HIV-1 NL4-3 molecular clones with polymorphisms in the vif gene (a kind gift from E. Arts, Case Western Reserve University, Cleveland, OH, USA) were used in virus fitness assays as previously described [351, 353]. Briefly, the HIV-1 RT mutants analyzed in this study were inserted into HIV-1 pol, with the 2100–5983 base region from HIV-1 NL4-3 sub-cloned into pCR2.1-TOPO® (Invitrogen, Carlsbad, CA, USA). Site-directed mutagenesis (QuikChange II Site-Directed Mutagenesis; Stratagene, Santa Clara, CA, USA) was performed to introduce point mutations; the region was sequenced to confirm proper introduction of mutations, and was then cloned back into the pNL4-3 MIG vector, using SbfI (2844) and AgeI (3486) restriction enzyme sites (New England Biolabs, Ipswich, MA, USA), or into the NL4-3 molecular clone using MscI restriction enzyme sites (2683 and 4545). All clones were sequence-confirmed for orientation and presence of desired mutations. Human embryonic kidney (HEK 293T) cells (American Type Culture Collection, Manassas, VA, USA) were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Cellgro, Manassas, VA, USA) plus 10% FetalClone III (FC3; Hyclone, Thermo Scientific, Waltham, MA, USA) and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA). U373-MAGI-CXCR4 cells were obtained from Michael Emerman through the National Institutes of Health (NIH) AIDS Reagent Program, Division of AIDS, NIAID, NIH [322]. U373-MAGI cells were maintained as were the HEK 293T cells, but also included the addition of 1.0 µg/mL puromycin, 0.1 mg/mL hygromycin B, and 0.2 mg/mL neomycin. The CEM-EGFP cell line was obtained from the AIDS Research and Reference Reagent Program, contributed by J. Corbeil [354]. The CEM cell line was a kind gift from Michael Malim. The CEM and CEM-EGFP cell lines were maintained in Roswell Park Memorial Institute (RPMI) medium (Gibco, Life Technologies Invitrogen, Grand Island, NY, USA) plus 10% FC3.

170 2.2. Virus Production and Titer Assay Vector viruses from the NL4-3 MIG vector and infectious viruses from the NL4-3 molecular clone were produced via transient transfection of HEK 293T cells as previously described [351] (Figure 1B). Briefly, the polyethylenimine (PEI) method [325] was used to transfect DNA into 2 × 106 293T cells with 10 µg of vector virus/proviral plasmid DNA, 1 µg of HCMV-G envelope expression plasmid DNA, and 33 µL of 1 mg/mL PEI. The medium was replaced 18 h post-transfection, and the cell culture supernatants were collected 48 h post-transfection, before being filtered through a 0.2-µm filter. A tissue culture infectious dose (50%; TCID50) end-point dilution assay was used to determine infectious units (IU) of virus per milliliter of cell culture supernatant. Each cell culture supernatant (for wild-type (wt) and mutant viruses) was serially diluted 10-fold; 100 µL of diluted supernatant was added to 5 × 104 CEM-EGFP indicator cells in 250 µL of total volume, before being plated in a 96-well plate, n = 6. The media was replaced every 48 h, and, at day 10, the number of EGFP-positive wells was determined using fluorescence microscopy. The number of EGFP-positive wells was multiplied by 1/6 and summed with 0.5 to determine the TCID50. The TCID50 divided by 100 µL was defined as the equivalent of the IU in each milliliter of supernatant. For example, an MOI of 0.005 could then be computed by calculating the amount of supernatant with 50 IU (MOI = IU/number of cells; 0.005 × 10,000 cells = 50 IU). The titer of virus stocks was determined using U373-MAGI cells prior to drug treatment experiments as previously described [306]. Briefly, a 12-well plate was used to plate 62,500 cells/well the day before infection. The media was replaced 18 h after plating the cells, and varying amounts of virus ranging from 1 to 50 µL were added. The media was replaced 24 h post-infection, and cells were collected 72 h post-infection. The cells were then analyzed by flow cytometry as previously described [355]. Briefly, the infected target cells were washed in phosphate-buffered saline (PBS) and resuspended in 200 µL of 2% FC3-PBS. Expression of mCherry and EGFP was analyzed using a BD LSR II flow cytometer (BD Biosciences, San Jose, CA, USA). Gates were selected based on a forward scatter channel and a side scatter channel with a minimum of 10,000 gated cells per sample. The fluorescent reporter proteins were excited with a blue 488-nm laser and a 561-nm laser, respectively. Flow cytometry data were analyzed using the FlowJo (v.9.2) software (Ashland, OR, USA). Virus infectivity was determined by adding all positive quadrants of mCherry-positive (mCherry+) and EGFP-positive (EGFP+) cells, and was set relative to wt for each experimental replicate.

2.3. Mutant Frequency Analysis by Flow Cytometry Vector viruses were used to infect 5 × 104 CEM cells via spinoculation for 2 h at 1200× g. Viral stocks were then titered using CEM cells to maintain a transduction efficiency below 20–30% to limit the likelihood of co-infection. Experiments were conducted independently three to five times with six biological replicates. The cells were prepared and analyzed by flow cytometry as described above. Mutant frequencies were calculated by dividing the sum of the number of cells in the single-positive populations (i.e., mCherry+, EGFP- and mCherry-, EGFP+) by the total number of infected cells. The mutant frequencies were then set relative to wt for each experimental replicate.

171 2.4. Dual-Competition Assay Infections using RT variants were done in the presence of the isogenic wt NL4-3 clone. In each head-to-head competition assay, the wt and mutant viruses could be independently quantified using a qPCR assay based on specific polymorphisms (i.e., 11 synonymous mutations) in the vif gene. Dual infections of 5 × 105 CEM-EGFP cells were done with a 1:1 ratio of wt:mutant virus at an MOI of 0.005. The cultures were maintained for 10 days, adding fresh culture media every two days. Each pairwise competition was done independently three times with four biological replicates per competition.

2.5. TaqMan Duplex qPCR Assay Analysis of dual-competition experiments was done using a modified duplex qPCR assay as previously described [353]. Briefly, infected cells were collected on day 10 of the competition assay, and then resuspended in 200 µL of PBS, before the total genomic DNA was extracted using the ZymoBead™ Genomic DNA Kit (Zymo Research, Irvine, CA, USA), and eluted into a total volume of 35 µL. Next, 5 µL of extracted DNA was subjected to a brief PCR amplification reaction in a 50-µL total-volume reaction with an outer primer pair—i.e., Vif Out+ (5′–GCA AAG CTC CTC TGG AAA GGT GAA GGG–3′) and Vif Out− (5′–CTT CCA CTC CTG CCC AAG TAT CCC–3′) primers to amplify the HIV-1 vif gene. Reactions were performed with Platinum PCR Supermix (Invitrogen) under the following conditions: one cycle at 94 °C for 2 min; 10 cycles at 94 °C for 30 s, 55 °C for 30 s, and 68 °C for 45 s; and one cycle at 68 °C for 5 min. The reaction was then purified with a GenElute PCR Clean-Up Kit (Sigma), and was eluted to 35 µL. A single probe was used in the TaqMan qPCR assay to differentiate between the two vif polymorphisms (i.e., vifA and vifB) as previously described [356]. One common- sense primer sequence was used for both vifA and vifB variants (vifAB; 5′–GGT CTG CAT ACA GGA GAA AGA GAC T–3′), and vifA- and vifB-specific antisense primers were used to differentially quantify the presence of each sequence. The vifA antisense primer used was 5′–AGG GTC TAC TTG TGT GCT ATA TCT CTT TT–3′, and the vifB antisense primer was 5′–AGG AAG CTT GCA ATA TCT AGC GTT AGC A–3′. The single probe (5′–6-FAM–CCT CCA TTC TAT GGA GAC TCC CTG ACC–BHQ1– 3′), which was used for both vifA and vifB, was labeled with fluorescein (FAM) and Black Hole Quencher. All amplifications were done using a BioRad CFX96 Touch Real- Time PCR Detection System (BioRad, Hercules, CA, USA). Primers were optimized, and were added at a final concentration of 375 nM each, while the probe was added at a concentration of 250 nM. Next, 1 µL of PCR product was added to 10 µL of 2× iTaq™ Universal Probes Supermix (BioRad), and was adjusted to a 20-µL total reaction volume. Reaction conditions were as follows: 95 °C for 10 min, 95 °C for 15 s, and 53 °C for 1 min for 40 cycles. All reactions were performed in triplicate, including a 9-log range in the plasmid DNA template (i.e., 5 × 109 to 5 × 101 copies), as well as a no-template negative control. The PCR efficiency (E) of each standard curve (vifA or vifB) was calculated based on the curve slope, E = (10 − 1/s − 1) × 100%, and the standard curves were confirmed to be within 5%, while the R2 of each standard curve was ≥99% to maximize sample data confidence. Differences in relative amounts of each clone after appropriate viral growth were used to identify differences in virus fitness. 172 2.6. Calculation of Viral Fitness Fitness differences (WD) were calculated for each wt versus mutant competition assay. The complementary DNA (cDNA) copy numbers of each of the four biological replicates used in the experiment were used to calculate the relative viral fitness (d) using the Viral Growth Rate Calculator web tool (https://indra.mullins.microbiol.washington.edu/vgrc/) [356]. Data from the three experimental replicates were compiled to determine fitness differences.

2.7. Statistical Analyses All statistical analyses were done using the GraphPad Prism version 6.0 software (La Jolla, CA, USA). A one-way ANOVA statistical analysis was performed on the raw data prior to normalization, in order to determine differences between the wt and mutants for both mutant frequency and viral fitness. A Tukey’s multiple-comparison post-test was also performed, in order to compare differences between the wt and each mutant. Linear regression was used to identify R2 values.

3. Results

3.1. Mutant Frequency Analysis of HIV-1 RT Variants A single-cycle vector assay (Figure 1A) was used to analyze differences in mutant frequency (Figure 2). In this assay, all four MDR mutants (i.e., Q151M complex with or without A62V, and the T69SSS insertion complex with or without A62V) were found to have a lower mutant frequency relative to that of the wt virus in the absence of AZT (Figure 2A). Interestingly, the MDR mutants without the A62V mutation had the lowest observed mutant frequencies (i.e., roughly half that of HIV-1 wt), which was significantly different from those of MDR complex mutants, including the A62V mutation. The observed virus mutant frequency of the A62V mutant alone was the highest (1.25-fold) of all mutants tested relative to the wt virus, as anticipated based upon previous observations [351]. The changes in mutant frequency of the MDR mutants were then tested in the presence of AZT (Figure 2B). The mutant frequency of the MDR mutants (except for the T69SSS complex without A62V, which was significantly lower) was restored to wt levels under the pressure of AZT. The mutant frequency of virus variants harboring the A62V alone was further increased (i.e., 1.9-fold) in the presence of AZT relative to the wt virus in the absence of AZT. Of note, wt HIV-1 replication in the presence of AZT resulted in a 1.26-fold increase in mutant frequency.

3.2. Replication Capacity Analysis of HIV-1 RT Variants The single-cycle assay was used with normalized virus stocks to measure the expression of EGFP in infected cells to determine wt or mutant HIV-1 replication capacity (Figure 3), or drug susceptibility (Figure 4) in the absence or presence of AZT. The number of infected cells, as determined by flow cytometry, directly represented virus replication [357]. Two of the MDR complex mutants—the Q151M complex without A62V, and the T69SSS complex without A62V—each had an improved replication capacity compared to that of the wt virus, with the T69SSS complex without A62V mutant having the highest increase in replication capacity (Figure 3). In contrast, the two

173 MDR complexes with A62V, as well as the A62V mutant, had a reduced replication capacity compared to that of wt virus (Figure 3). The A62V mutant was observed to have the lowest replication capacity of the viruses analyzed.

3.3. Drug Susceptibility of HIV-1 RT Variants to AZT AZT drug susceptibility among the HIV-1 RT variants compared to the wt was analyzed by determining the IC50 values. The A62V mutant was observed to have the highest susceptibility to AZT with a 1.76-fold decrease in virus replication relative to that of the wt virus in the absence of AZT. The two MDR complex mutants without A62V were observed to have the highest level of drug susceptibility, with the T69SSS insertion complex having the highest fold increase (Figure 4). The Q151M complex and T69SSS insertion complex mutants with A62V had an AZT drug-susceptibility phenotype comparable to that of the wt virus.

3.4. Fitness Impact of HIV-1 RT Mutations To assess the replication fitness of HIV-1 harboring the MDR RT mutants, a real- time TaqMan PCR assay was used to monitor the mutant and parental viruses as previously described [356]. This method allowed for the net growth rate of each virus in competition assays to be determined. The net growth rate difference (d) per day (which was used to determine the fitness cost between the mutant and its parental virus) was calculated in the absence (Figure 5A) or presence (Figure 5B) of AZT. In the absence of AZT, the A62V mutant alone had significantly reduced viral replication fitness, while the T69SSS insertion complex with A62V had a non-statistically significant decrease in fitness (Figure 5A). The fitness differences of the Q151M complex and T69SSS insertion complex mutants without A62V were not significant. In the presence of AZT, all MDR mutant complexes, except for the T69SSS insertion complex without A62V, had a statistically significant increase in virus fitness (Figure 5B). Predictably, the A62V mutant alone had reduced fitness.

4. Discussion While the A62V amino acid substitution in HIV-1 RT is known to be associated with multi-drug resistance, it is not a resistance-conferring mutation, and its appearance remains an open question in the field. To investigate this, we tested the hypothesis that A62V provides a selective advantage to the virus in the context of multi-drug resistance by influencing replication fidelity and fitness. In particular, we used parallel analyses to look at the relationship between HIV-1 fitness and mutagenesis in the presence or absence of the RT A62V amino acid substitution. We first found that the A62V mutation alone could significantly increase viral mutant frequencies (Figure 2) [351], while negatively impacting replication capacity (Figures 3 and 4) and viral fitness (Figure 5) in the absence or presence of AZT. Both the Q151M complex and the T69SSS insertion complex share the A62V secondary drug-resistance-associated point mutation, which is located close to the active polymerization site of HIV-1 RT (Figures 6A,B, respectively). We observed that MDR mutants without the A62V mutation had the lowest mutant frequencies, suggesting that these RT variants have higher fidelity (Figure 2A). In the presence of AZT (Figure 2B), the mutant frequency of all the MDR viruses, except that of the T69SSS insertion complex without A62V, were restored to that of wt HIV-1 in the 174 absence of AZT. These observations support the conclusion that the A62V mutation plays an important role in RT fidelity by increasing mutant frequency (Figure 2), where mutant frequency is higher in the context of MDR complexes in the absence of AZT (Figure 2A), but highest in virus with the A62V mutation alone in the presence of AZT (Figure 2B). How A62V influences replication is currently unclear, and will be a topic of future analyses. The T69SSS insertion complex without A62V was observed to possess the highest replication capacity. In contrast, the two MDR variant viruses with A62V, along with the virus harboring the A62V mutant alone, had a lower replication capacity relative to wt HIV-1 (Figure 3). Under AZT-selective pressure, viruses harboring either the Q151M complex or the T69SSS insertion complex without A62V had the lowest level of drug susceptibility. In contrast, viruses harboring the Q151M complex or the T69SSS insertion complex in the context of A62V had replication efficiencies comparable to that of wt HIV-1 in the absence of a drug when under AZT drug-selective pressure (Figure 4). Taken together, these data support the conclusion that both MDR variant viruses (i.e., harboring the Q151M complex and the T69SSS insertion complex mutations), in the context of the A62V amino acid substitution, had mutant frequencies and replication capacities, when under AZT selective pressure, comparable to that of the wt virus in the absence of drug-selective pressure. These observations implicate an interrelationship between HIV-1 fitness and mutation rate [351]. All multi-drug-resistant viruses were observed to have a statistically significant increase in viral fitness under AZT-selective pressure, except viruses harboring the T69SSS insertion complex without A62V (Figure 5B). As anticipated, viruses harboring the A62V mutation alone had the most significant reduction in replication capacity. These observations, together with the observations made of the relative susceptibility to AZT (Figure 4), indicate that the MDR viruses without A62V analyzed in our study have reduced AZT drug susceptibility at the expense of replication capacity. This was most notable with the viruses harboring the T69SSS insertion complex in the absence of the A62V mutation. As previously reported [358, 359], we observed that the fitness of the Q151M complex, in the presence or absence of the A62V mutation, was not significantly different than that of the wt virus (Figure 5A). Together, these observations implicate A62V as an adaptive mutation arising via drug pressure. The replicative capacity of viruses harboring the MDR Q151M complex, in the presence or absence of the A62V mutation, was observed to be comparable under AZT drug pressure (Figure 5B). However, viruses harboring the Q151M complex without A62V had a higher level of drug resistance than that of viruses possessing the Q151M complex with A62V (Figure 4), which agrees with previous reports [342, 358, 360]. Taken together, these observations support the conclusion that the T69SSS insertion complex does not confer a fitness advantage in the absence of AZT (Figure 5A), but does improve fitness in the presence of AZT (Figure 5B). This conclusion agrees with previous observations regarding the increased fitness associated with the T69SSS insertion complex during drug-selective pressure [361]. In summary, the observations made in this study provide the first demonstration that A62V is an important adaptive mutation in multi-drug-resistant viruses that impacts the interplay of replication fidelity, virus fitness and drug susceptibility. These data argue in support of the importance of adaptive mutations that can “piggyback” along with drug-

175 resistant mutations to improve overall viral fitness during drug-selective pressure. The observations of this study complement previous observations where MDR viruses were observed to be more fit than the wt virus in the absence of a drug [359]. Differences observed between the findings presented in this study regarding particular phenotypes compared to those previously reported are likely due to biological differences that can exist among virus isolates and cell types used for the analyses of virus mutant frequency and viral fitness. These studies predict that, in general, viral mutation rate and fitness can be influenced by adaptive mutations that arise during drug-selective pressure.

176

Figure 1. Assays for measuring human immunodeficiency virus type 1 (HIV-1) mutant frequencies and replication capacity. (A) Mutant frequency assay. 293T cells were co-transfected with an HIV-1 vector containing the dual-reporter mCherry and enhanced GFP (EGFP) cassette (pNL4-3 MIG) and a vesicular stomatitis virus G (VSV-G) envelope expression plasmid (pHCMV-G) in order to produce pseudotyped vector virus. Cell culture supernatants were collected and used to infect permissive CEM or U373-MAGI cells. Determination of virus mutant frequencies and replication capacities under zidovudine (AZT)-selective pressure were done using U373-MAGI cells. Relative infectivity and mutant frequencies were determined by flow cytometry. (B) Dual competition assay. Mutations in HIV-1 reverse transcriptase (RT) were introduced in an infectious molecular clone (pNL4-3), and were used to produce infectious virus following transfection of proviral DNA into 293T cells. Mutant and wild-type (wt) viruses were used to co-infect 5 × 105 permissive CEM-EGFP target cells at a 1:1 ratio, using a multiplicity of infection (MOI) of 0.005. Infected target cells were maintained for 10 days by replenishing with fresh media every other day. Cells were then collected, genomic DNA extracted, and relative amounts of viral nucleic acid quantified by duplex qPCR. Abbreviations: IRES, internal ribosome entry site; pol, HIV-1 gene consisting of protease, RT, and integrase.

177 *** ns *P value < 0.05 *** (no AZT) wt HIV-1 Q151M w/o A62V ****P value < 0.0001 *** **P value < 0.01 Q151M w/o A62V n = 3 Q151M w/ A62V ***P value < 0.001 *** Q151M w/ A62V T69SSS w/o A62V n = 4 T69SSS w/o A62V * *** T69SSS w/ A62V T69SSS w/ A62V HIV-1 RT Mutants RT HIV-1

*** Mutants RT HIV-1 A62V ** **** A62V **** wt HIV-1 wt HIV-1

0.50 0.75 1.00 1.25 1.50 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 Mutant Frequency Mutant Frequency (Fold Change Relative to wt HIV-1) (Fold Change Relative to [no AZT] wt HIV-1) (A) (B) Figure 2. Mutant frequency analysis of multi-drug-resistant HIV-1 mutants in the presence or absence of A62V. Mutant frequency analysis was performed using the pNL4-3 MIG vector by infecting permissive target cells prior to analyzing for expression of a pair of marker genes (mCherry and EGFP) by flow cytometry. Mutant frequencies were calculated by dividing the sum of the number of cells in the single-positive populations (i.e., mCherry+, EGFP- and mCherry-, EGFP+) by the total number of infected cells. The mutant frequencies were then set relative to the no-AZT wt virus for each experimental replicate. The A62V mutation in HIV-1 RT was analyzed alone or in the presence or absence of the multi-dideoxynucleoside resistant (MDR) Q151M complex (i.e., A62V, V75I, F77L, F116Y, and Q151M) and the T69SSS insertion complex (i.e., M41L, A62V, T69SSS, K70R, and T215Y). The mutant frequency of the MDR mutants, in the absence of AZT (A) or in the presence of AZT (B). The standard error of the mean (SEM) is represented by error bars. * p-value <0.05; ** p-value <0.01; *** p-value <0.001; **** p-value <0.0001. Lack of statistical significance is denoted by ‘ns’.

178

*** Q151M w/o A62V *

Q151M w/ A62V *** *** T69SSS w/o A62V ***

T69SSS w/ A62V ** *P value < 0.05

HIV-1 RT Mutants RT HIV-1 A62V *** **P value < 0.01 ***P value < 0.001 n=3 wt HIV-1

0.50 0.75 1.00 1.25 1.50 1.75 Viral Titer (Fold Change Relative to wt HIV-1) Figure 3. Analysis of replication capacity of multi-drug-resistant HIV-1 mutants in the presence or absence of A62V. Normalized virus stocks were used in the single-cycle assay to determine the replication capacity of multi- dideoxynucleoside resistant (MDR) HIV-1 vector viruses, the Q151M complex (i.e., A62V, V75I, F77L, F116Y, and Q151M) and the T69SSS insertion complex (i.e., M41L, A62V, T69SSS, K70R, and T215Y), in the presence or absence of A62V. Replication capacity was determined via the detection of GFP expression by flow cytometry. All replication capacity values were set relative to wt HIV-1 for each experimental replicate. The standard error of the mean (SEM) is indicated by the error bars. * p-value <0.05; ** p-value <0.01; *** p-value <0.001.

179 ****

(no AZT) wt HIV-1 ****P value < 0.0001 n=3 **** Q151M w/o A62V

Q151M w/ A62V **** T69SSS w/o A62V

T69SSS w/ A62V HIV-1 RT Mutants RT HIV-1 **** A62V **** **** wt HIV-1

0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 Viral Titer (Fold Change Relative to [no AZT] wt HIV-1) Figure 4. Analysis of relative drug susceptibility of multi-drug-resistant HIV-1 mutants in the presence or absence of A62V. Normalized virus stocks were used in the single-cycle assay to determine the relative drug susceptibility of the multi-dideoxynucleoside resistant (MDR) HIV-1 Q151M mutant complex (i.e., A62V, V75I, F77L, F116Y, and Q151M), and the T69SSS insertion complex (i.e., M41L, A62V, T69SSS, K70R, and T215Y) in the presence or absence of A62V under AZT-selective pressure. Virus titers were determined by flow cytometry. All drug susceptibility values were set relative to no-AZT wt HIV-1 for each experimental replicate. The standard error of the mean (SEM) is indicated by the error bars. **** p-value <0.0001.

180 1.0 0.2

0.5 * * * (d, per day) per (d, (d, per day) per (d, 0.0 -0.3 VifA * VifA -g -g VifB VifB g g -0.5 *

-0.8 -1.0

A62V A62V wt VifB

T69SSS w/ A62V Q151M w/ A62V T69SSS w/ A62V Q151M w/ A62V T69SSS w/o A62V Q151M w/o A62V T69SSS w/o A62V Q151M w/o A62V (A) (B) Figure 5. Dual-competition virus fitness assay of multi-drug-resistant HIV-1 mutants in the presence or absence of A62V. The net growth rate difference (d) of multi-dideoxynucleoside resistant (MDR) HIV-1 mutants (with or without A62V), in the absence (A) or presence (B) of AZT, was determined using the difference between the growth rate of wt HIV-1 (vifB) and mutant (vifA), that is, gVifB − gVifA. To assess the replication fitness, a real-time TaqMan PCR assay was used to monitor and differentiate between the mutant and parental virus using polymorphisms in the vif gene. All growth difference values (i.e., panels A,B) were set relative to no-AZT wt HIV-1 (vifB), shown in panel (A). The polymorphisms in the vif gene had a small fitness impact, as shown in panel (A); this effect was taken into consideration for all calculations. The wt HIV-1 (vifB) was not directly included in panel (B) due to inability to replicate it in the presence of AZT. Reported are the means and 95% confidence intervals (CIs) from triplicate competition experiments. The asterisk (*) indicates a significant difference (p-value <0.05, calculated using methods explained elsewhere [356]) relative to the corresponding vif gene polymorphisms (vifA, mutant virus; vifB, parental virus).

181

(A) (B) Figure 6. Location of resistance-conferring amino acid residues in multi-drug- resistant HIV-1 reverse transcriptase. A ribbon structure of the covalently trapped catalytic complex of HIV-1 RT [362] with the multi-dideoxynucleoside resistant (MDR) Q151M complex amino acid residues shown in a close-up view (A). The MDR T69SSS insertion complex amino acid residues are shown in a close-up view (B). A portion of the HIV-1 RT p66 subunit is shown with color- coding of subdomains: fingers (blue) and palm (red). The template and primer DNA strands are shown in light gray. While the p66 subunit contains the DNA- binding groove and the polymerization active site, the non-catalytic p51 subunit is not shown [363]. Image obtained from the “Research Collaboratory for Structural Bioinformatics Protein Data Bank” [364] (Protein Data Bank identifier: 1RTD).

182

APPENDIX IV

COPYRIGHT PERMISSIONS

183 Authorization of: Maldonado, J. O., Martin, J. L., Mueller, J. D., Zhang, W., & Mansky, L. M. (2014). New insights into retroviral Gag–Gag and Gag–membrane interactions. Frontiers in Microbiology, 5, 302.

Title: New insights into retroviral Gag–Gag and Gag–membrane interactions Author: José O. Maldonado, Jessica L. Martin, Jochen D. Mueller, Wei Zhang, & Louis M. Mansky Publication: Frontiers in Microbiology Publisher: Frontiers Date: Jun 24, 2014 Copyright © 2014, American Society for Microbiology

Permissions Request

Under the Frontiers Conditions for Website Use and the Frontiers General Conditions for Authors, authors of articles published in Frontiers journals retain copyright on their articles, except for any third-party images and other materials added by Frontiers, which are subject to copyright of their respective owners. Authors are therefore free to disseminate and re-publish their articles, subject to any requirements of third-party copyright owners and subject to the original publication being fully cited. Visitors may also download and forward articles subject to the citation requirements and subject to any fees Frontiers may charge for downloading licenses. The ability to copy, download, forward or otherwise distribute any materials is always subject to any copyright notices displayed. Copyright notices must be displayed prominently and may not be obliterated, deleted or hidden, totally or partially. A charge may be made for some facilities (such as downloading of e-magazines), where stated.

184 Authorization of: Cao, S., Maldonado, J. O., Grigsby, I. F., Mansky, L. M., & Zhang, W. (2015). Analysis of Human T-Cell Leukemia Virus Type 1 Particles by Using Cryo- Electron Tomography. Journal of virology, 89(4), 2430-2435.

Title: Analysis of human T-cell leukemia virus type 1 particles using cryo-electron tomography Author: Sheng Cao, José O. Maldonado, Iwen F. Grigsby, Louis Mansky, & Wei Zhang Publication: Journal of Virology Publisher: American Society for Microbiology Date: Feb 15, 2015 Copyright © 2014, American Society for Microbiology

Permissions Request

Authors in ASM journals retain the right to republish discrete portions of his/her article in any other publication (including print, CD-ROM, and other electronic formats) of which he or she is author or editor, provided that proper credit is given to the original ASM publication. ASM authors also retain the right to reuse the full article in his/her dissertation or thesis. For a full list of author rights, please see: http://journals.asm.org/site/misc/ASM_Author_Statement.xhtml

185 Authorization of: Martin, J. L., Maldonado, J. O., Mueller, J. D., Zhang, W., & Mansky, L. M. (2016). Molecular studies of HTLV-1 replication: An update. Viruses, 8(2), 31.

Title: Molecular studies of HTLV-1 replication: An update Author: Jessica Martin, José O. Maldonado, Joachim Mueller, Wei Zhang, & Louis M. Mansky Publication: Viruses Publisher: MDPI Date: Jan 27, 2016 Copyright © 2016, Creative Commons Attribution License

Permissions Request

Until 2008, most articles published by MDPI contained the note: "© year by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes". During 2008, MDPI journals started to publish articles under the Creative Commons Attribution License and are now using the latest version of the CC BY license, which grants authors the most extensive rights. All articles published by MDPI before and during 2008 should now be considered as having been released under the post-2008 Creative Commons Attribution License.

186 Authorization of: Daly, M. B., Roth, M. E., Bonnac, L., Maldonado, J. O., Xie, J., Clouser, C. L., Patterson, S. E., Kim, B., & Mansky, L. M. (2016). Dual anti-HIV mechanism of clofarabine. Retrovirology, 13(1), 20.

Title: Dual anti-HIV mechanism of clofarabine Author: Michele B. Daly, Megan E. Roth, Laurent Bonnac, José O. Maldonado, Jiashu Xie, Christine L. Clouser, Steven E. Patterson, Baek Kim, & Louis M. Mansky Publication: Retrovirology Publisher: Springer Open Date: Mar 24, 2016 Copyright © 2016, Creative Commons Attribution License

Permissions Request

As part of our copyright and license agreement, open access articles may be reproduced without formal permission or payment of permission fees. As a courtesy, however, anyone wishing to reproduce large quantities of an open access article (250+) should inform the copyright holder and we suggest a contribution in support of open access publication.

187 Authorization of: Maldonado, J. O., Cao, S., Zhang, W., & Mansky, L. M. (2016). Distinct morphology of human T-cell leukemia virus type 1-like particles. Viruses, 8(5), 132.

Title: Distinct morphology of human T-cell leukemia virus type 1-like particles Author: José O. Maldonado, Sheng Cao, Wei Zhang, & Louis M. Mansky Publication: Viruses Publisher: MDPI Date: May 11, 2016 Copyright © 2016, Creative Commons Attribution License

Permissions Request

Until 2008, most articles published by MDPI contained the note: "© year by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes". During 2008, MDPI journals started to publish articles under the Creative Commons Attribution License and are now using the latest version of the CC BY license, which grants authors the most extensive rights. All articles published by MDPI before and during 2008 should now be considered as having been released under the post-2008 Creative Commons Attribution License.

188 Authorization of: Martin, J. L., Cao, S., Maldonado, J. O., Zhang, W., & Mansky, L. M. (2016). Distinct particle morphologies revealed through comparative parallel analyses of retrovirus-like particles. Journal of virology, 90(18), 8074-8084.

Title: Distinct particle morphologies revealed through comparative parallel analyses of retrovirus-like particles Author: Jessica Martin, Sheng Cao, José O. Maldonado, Wei Zhang, & Louis M. Mansky Publication: Journal of Virology Publisher: American Society for Microbiology Date: Aug 26, 2016 Copyright © 2016, American Society for Microbiology

Permissions Request

Authors in ASM journals retain the right to republish discrete portions of his/her article in any other publication (including print, CD-ROM, and other electronic formats) of which he or she is author or editor, provided that proper credit is given to the original ASM publication. ASM authors also retain the right to reuse the full article in his/her dissertation or thesis. For a full list of author rights, please see: http://journals.asm.org/site/misc/ASM_Author_Statement.xhtml

189 Authorization of: Maldonado, J. O., Angert, I., Cao, S., Berk, S., Zhang, W., Mueller, J. D., & Mansky, L. M. (2017). Perturbation of Human T-Cell Leukemia Virus Type 1 Particle Morphology by Differential Gag Co-Packaging. Viruses, 9(7), 191.

Title: Perturbation of Human T-Cell Leukemia Virus Type 1 Particle Morphology by Differential Gag Co-Packaging Author: José O. Maldonado, Isaac Angert, Sheng Cao, Serkan Berk, Wei Zhang, Joachin Mueller, & Louis M. Mansky Publication: Viruses Publisher: MDPI Date: Jul 19, 2017 Copyright © 2017, Creative Commons Attribution License

Permissions Request

Until 2008, most articles published by MDPI contained the note: "© year by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes". During 2008, MDPI journals started to publish articles under the Creative Commons Attribution License and are now using the latest version of the CC BY license, which grants authors the most extensive rights. All articles published by MDPI before and during 2008 should now be considered as having been released under the post-2008 Creative Commons Attribution License.

190 Authorization of: Maldonado, J. O., & Mansky, L. M. (2018). The HIV-1 Reverse Transcriptase A62V Mutation Influences Replication Fidelity and Viral Fitness in the Context of Multi-Drug-Resistant Mutations. Viruses, 10(7), 376.

Title: The HIV-1 Reverse Transcriptase A62V Mutation Influences Replication Fidelity and Viral Fitness in the Context of Multi-Drug-Resistant Mutations Author: José O. Maldonado & Louis M. Mansky Publication: Viruses Publisher: MDPI Date: Jul 19, 2018 Copyright © 2018, Creative Commons Attribution License

Permissions Request

Until 2008, most articles published by MDPI contained the note: "© year by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes". During 2008, MDPI journals started to publish articles under the Creative Commons Attribution License and are now using the latest version of the CC BY license, which grants authors the most extensive rights. All articles published by MDPI before and during 2008 should now be considered as having been released under the post-2008 Creative Commons Attribution License.

191