A NOVEL HIV-1 GENOMIC RNA PACKAGING ELEMENT AND ITS ROLE IN INTERPLAY BETWEEN RNA PACKAGING AND GAG-POL RIBOSOMAL FRAMESHIFTING

Transcription

1 A NOVEL HIV-1 GENOMIC RNA PACKAGING ELEMENT AND ITS ROLE IN INTERPLAY BETWEEN RNA PACKAGING AND GAG-POL RIBOSOMAL FRAMESHIFTING by MASTOOREH CHAMANIAN Submitted in partial fulfillment of the requirements For the degree of Doctor of Philosophy Dissertation Adviser: Eric J. Arts, Ph.D. Department of Molecular Biology and Microbiology CASE WESTERN RESERVE UNIVERSITY May 2013

2 CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES We hereby approve the dissertation of Mastooreh Chamanian candidate for the degree*. David McDonald, Ph.D. (signed) (chair of the committee) Jonatha Gott, Ph.D. Amy Hise, MD. MPH. Eric J Arts, Ph.D. (dissertation advisor) January 30, 2013 (date) *We also certify that written approval has been obtained for any proprietary material contained therein.

3 TABLE OF CONTENTS LIST OF TABLES... 9 LIST OF FIGURES ACKNOWLEDGEMENTS LIST OF ABBREVIATIONS CHAPTER 1: INTRODUCTION HIV-1 and AIDS Retroviruses HIV-1 genome and proteins HIV-1 virion Early stage of HIV-1 lifecycle Entry Reverse transcription Nuclear Import/Integration Late stage of HIV-1 lifecycle Transcription and splicing of Viral RNA Nucleo-Cytoplasmic transport of viral RNA mrna translation Ribosomal frameshifting Genomic RNA dimerization and packaging Virion assembly, budding and maturation Viral RNA structure

4 1.7.1 Secondary structure of major packaging signal in 5 UTR region D structure of major packaging signal in 5 UTR region Interplay between RNA packaging and translation by RNA switch Secondary structure of ribosomal frameshift signal NMD pathway Lentivirus-based retroviral vectors HIV-1 derived lentiviral vectors Lentiviral Vector packaging system Design and improvement of lentiviral transfer vector Production of lentiviral vector CHAPTER 2: MATERIALS AND METHODS Yeast recombination/gap repair cloning system Plasmid construction Construction of prec_nfl_hiv-1 and pcmv_ctpl 3 truncated vectors Construction of a universal prec- HIV-1/URA3 shuttle vector Construction of prec-5 LTR plasmids using yeast recombination technique Site-directed mutagenesis Construction of the remaining plasmids Cell culture Transfection and infection

5 2.5 Reverse transcriptase activity assay RNA extraction and quantitative real-time RT-PCR (qrt-pcr) Western blot and ELISA analysis RNA synthesis and secondary RNA structure prediction analysis D structure prediction RNA Interference In vitro RNA dimerization Vector and lentivirus production In vitro lentivirus transduction Flow cytometry CHAPTER 3: Optimal HIV-1 genomic RNA packaging is dependent on a cis acting RNA sequence found at the 3 terminus of gag using a dual RNA complementation system Preface Abstract Introduction Results Development of the yeast-based HIV-1 cloning and virus production system Complementation between nfl_hiv-1 and cplt RNA Validation and optimization of the complementation system

6 Testing the infectivity of viruses produced from bipartite system Optimizing the complementing system by increasing the length of complementing vector (pcmv_cplt) Using complementing system for RNA packaging studies Generation of truncated prec-5 LTR vectors HIV-1 replication system involving bi-partite HIV-1 genomic RNA where only one contributes to HIV-1 coding sequence Locating new RNA elements in the HIV-1 coding sequence that are necessary for genomic RNA encapsidation The GRPE RNA element is necessary for infectivity via its effect on genomic RNA packaging Recombination between two sgrna during reverse transcription of de novo infection Discussion CHAPTER 4: The novel packaging enhancer element overlaps with ribosomal frameshift signal and mediates an interplay between RNA packaging and frameshifting Preface Abstract Introduction Results

7 4.4.1 HIV-1 genomic RNA encapsidation is reduced with GRPE deletion Positional dependence of GRPE for grna encapsidation SHAPE technology for probing RNA secondary structure of GRPE Effect of GRPE secondary structure on grna encapsidation, dimerization, and virus infectivity In vitro dimerization assay to determine the impact of GRPE on sgrna dimerization Exploring the relationship between HIV-1 mrna translation and grna packaging Discussion CHAPTER 5: The lentivirus transduction efficiency is enhanced by incorporation of GRPE element in the lentiviral vectors Preface Abstract Introduction Results Generation of lentiviral vectors including the GRPE and flanking sequences GRPE increases the packaging efficiency of lentivirus RNA

8 5.4.3 GRPE enhances the transduction efficiency of lentivirus via effects on grna packaging Discussion CHAPTER 6: GENERAL DISCUSSION AND FUTURE DIRECTIONS Identification of an additional packaging element within HIV-1 Coding region Interplay between grna packaging and mrna translation/frameshifting Application of GRPE in lentivirus system COPYRIGHT RELEASES BIBLIOGRAPHY

9 A Novel HIV-1 Genomic RNA Packaging Element and its Role in Interplay between RNA Packaging and Gag-Pol Ribosomal Frameshifting Abstract by MASTOOREH CHAMANIAN Retroviruses belong to a diverse family of RNA viruses that utilize the eukaryotic translational machinery for viral protein synthesis during infection of its target cells. In the cytoplasm, the unspliced version of retroviral RNA takes one of two pathways: 1) to serve as mrna template for Gag protein translation, or 2) to act as genomic RNA that is packaged into virions. Packaging of two genomic RNAs in the retroviral particle involves interaction between a cis-acting packaging signal and GagNC. Various studies have mapped the canonical packaging signal or Ψ of most retroviruses such as HIV-1 to the 5 untraslated region (UTR) containing four hairpin structures. However, most of these studies only focused on small fragments of the 5 UTR and therefore the evidences were not sufficient to consider the Ψ as the sole packaging element. A central goal of this dissertation was to identify additional packaging determinant(s) in the coding region of HIV-1 genome. The ability to orchestrate genomic RNA packaging and gene expression is central to production of infectious and morphologically correct retrovirus containing sufficient amount of genomic (g) RNA required for further recombination and integration. Aside from housing the Ψ, 5 UTR forms a complex RNA structure that interacts with various 7

10 host and viral proteins to mediate multiple steps of viral life cycle. Although the Ψ packaging signal is often described as a set of static, discrete RNA stem loops, there is considerable evidence that HIV-1 genome and its 5 UTR is capable of adopting multiple conformations that could regulate dimerization, translation and genome packaging. The ribosomal frameshifting step of translation is critical for efficient synthesis of Gag and Gag-pol proteins through shifting of ribosome in -1 nt direction at the ribosomal frameshift signal (RFS). After identification of a Genomic RNA Packaging Element (GRPE) that overlapped with RFS using a dual RNA complementation system, another important goal of this study was to determine the role of GRPE in coordination of translation at the level of frameshifting and RNA packaging which is an enduring question in the field of retroviral biology. The studies presented here reveal a new understanding of retroviral genomic RNA packaging by showing that RNA that was used for translation of Gag-pol by ribosomal frameshifting can escape the nonsense-mediated mrna decay (NMD) pathway and be selected as genomic RNA for encapsidation into the assembled viruses. These findings provide broad knowledge for optimizing the inhibitors of packaging step and also basic understanding of retroviral replication and host translational control system. It is also important for lentiviral gene therapy system since current lentiviruses lack the GRPE element and all lentiviruses utilize a similar -1 ribosomal slippage mechanism to express the Gag-Pol precursor proteins. The following brief overview of the retroviral replication cycle and the adaptation of retroviruses as vectors for gene transfer will serve as background for these studies and emphasize two post-transcriptional steps of the lifecycle: RNA packaging, and ribosomal frameshifting. 8

11 LIST OF TABLES TABLE 1: Retroviral Genera TABLE 2: HIV-1 Genome in prec-5 LTR Vectors Converting the HIV-1 HXB2 DNA Numbering with RNA Numbering TABLE 3: Primers and Probes Used for PCR, RT-PCR and q-rt-pcr

12 LIST OF FIGURES FIGURE 1: Schematic representation of simple and complex retrovirus genomes FIGURE 2: Organization of the HIV-1 mature viral particle FIGURE 3: Early stage of HIV Life cycle FIGURE 4: The steps involved in reverse transcription FIGURE 5: Late stage of HIV-1 life cycle associated with RNA packaging FIGURE 6: HIV mrna splicing FIGURE 7: Rev/CRM-1 pathway for nucleo-cytoplasmic exports of intron-containing. viral RNAs FIGURE 8: Sequence and secondary structure model of HIV-1 packaging signal in 5 UTR FIGURE 9: Interactions between DIS of two RNA molecules and initiation of grna dimerization FIGURE 10: Representative high-resolution 3D structures of RNA elements within HIV-1 5 -UTR determined by NMR or X-ray crystallography FIGURE 11: Model for translation and encapsidation regulation of HIV-1 unspliced RNAs FIGURE 12: Structure of the HIV-1 gag-pol ribosomal frameshift site (RFS) FIGURE 13: Translation termination and NMD pathway FIGURE 14: Schematic representation of lentiviral vector system FIGURE 15: Schematic depiction of precnflhiv-1 production FIGURE 16: Schematic depiction of pcmv_cplt production FIGURE17: Overview of high-throughput selective 2 -hydroxyl acylation analyzed by primer extension (hshape) FIGURE 18: Complementation system for chimeric virus production

13 FIGURE 19: Generation of complementing vectors containing various lengths of HIV-1 genome and comparison of their infectivity FIGURE 20: Cartoon showing bi-partite plasmid FIGURE 21: Complementation system used for packaging studies and infectious virus production FIGURE 22: Schematic of the reverse transcription following host cell entry with homo- and heterodiploid viruses FIGURE 23: The level of Gag precursor polyprotein measured in 293T cells transfected with truncated prec-5 LTR Plasmids FIGURE 24: Comparing the expected and actual infectious titers of virus derived from co-transfected 293T cells FIGURE 25: The levels of 5 LTR and nfl-3 LTR sgrna within the cell and cytoplasm following co-transfections FIGURE 26: Relative packaging of the HIV-1 sub-genomic RNA in virus derived from co-transfected cell FIGURE 27: RNA packaging efficiency was not influenced by deletion of a Gag AUG start codon in the 5 LTR sgrnas FIGURE 28: Effect of deleting the putative GRPE element on packaging of the 5 LTR sgrna and full-length grna FIGURE 29: Effect of deleting the region separating GRPE and Ψ on packaging of the 5 LTR sgrna FIGURE 30: SHAPE analysis of GRPE/RFS element in 5 LTR sgrna with deletions of the region separating GRPE and Ψ FIGURE 31: Prediction of GRPE 3D strcuture of in the 5 LTR-RT-mutant sgrnas FIGURE 32: The role of RNA secondary structure in the GRPE/RFS region on HIV-1 RNA packaging FIGURE 33: The levels of nfl-3 LTR sgrna measured by qrt-pcr FIGURE 34: In vitro dimerization of 5 LTR and nfl-3 LTR sgrna

14 FIGURE 35: The effect of erf1 knockdown on virus production and grna Encapsidation FIGURE 36: Lentiviral vector constructs FIGURE 37: Lentivirus-based gene delivery system FIGURE 38: Quantitative analysis of the RNA content of lentiviral particles released from transfected 293T producer cells FIGURE 39: Transduction efficiency of Lentiviruses increase by addition of GRPE FIGURE 40: Model for translation and packaging of the unspliced HIV-1 RNA

15 ACKNOWLEDGEMENTS I would like to thank my advisor Eric Arts for his valuable guidance and support throughout the duration of my PhD. I highly appreciate the freedom that was given to me to learn from my mistakes and become an independent researcher. I would like to thank all my committee members, David McDonald, Jonatha Gott, and Amy Hise. I m truly grateful for their time, encouragement and insightful wisdom throughout this process. A special thanks to David McDonald; his helpful insights and enthusiasm has always inspired me to do better. I would like to thank the current and past members of the Arts lab. A special thanks to Annette Ratcliff for being a great friend and a part of almost all of my memories in the lab. Her company made for a very exciting experience that I could not imagine undertaking without. To Yong Gao who helped and supervised me during the beginning of this project. A special thanks to Ken, Rick and Denis for all their support and suggestions. My interactions with these wonderful people have made science even more enjoyable. I would like to thank Katarzyna Purzycka and Stuart Le Grice for their fruitful collaboration on RNA structure experiments presented in this dissertation. To my parents, Farzaneh and Hashem, and my two brothers, for their unconditional love and support. My mom always encouraged me to follow my dreams. And also to my dear friend Faegheh, who was there for me through the hard times and happy times. Finally, I would like to thank Arash, my beloved husband and best friend for being supportive of my goals and dreams. Without his constant encouragement and loving support this work would not have been possible. And to my dog, Hugo for being a constant source of love and energy. 13

16 LIST OF ABBREVIATIONS Ψ: major encapsidation/packaging signal 5-FOA: + 5-fluoro-1,2,3,6-tetrahydro-2,6-dioxo-4-pyrimidine carboxylic acid 5 LTR: 5 long terminal repeat 5 UTR: 5 untranslated region AIDS: acquired immune deficiency syndrome BMH: branched multiple hairpin CMV: cytomegalovirus cppt: central polypurine tract CTS : central termination sequence DIP: defective interfering particles DIS: dimerization initiation site DMEM: dulbecco s modified eagle medium DNA: deoxyribonucleic acid env: HIV envelope gene erf1: release factor 1 FBS: fetal bovine serum GAPDH: glyceraldehyde 3-phosphate dehydrogenase GFP: green fluorescent protein gp120: glycoprotein 120 grna: genomic RNA GRPE: genomic RNA packaging enhancer 14

17 HIV-1: human immunodeficiency virus type 1 HSC: hematopoietic stem cell IU: infectious unit LDI: Long distance interaction LTR: long terminal repeat MLV: murine leukemia virus MND: MPSV promoter with negative control region deleted MOI: multiplicity of infection mrna: messenger RNA NC: nucleocapsid NMD: nonsense-mediated mrna decay nt: nucleotide ORF: open reading frame PABP1 poly(a)-binding protein PBS (1): phosphate buffered saline PBS (2): primer binding site PPT: polypurine tract RCL: replication competent lentivirus RFS: ribosomal frameshift signal RLU: relative light unit RNA: ribonucleic acid RNaseH: ribonuclease H 15

18 RRE: rev responsive element SD: splice donor SHAPE: selective 2' hydroxyl acylation analyzed by primer extension SL: stem loop U3: unique 3 region U5: unique 5 region URA3: yeast gene that encodes orotidine 5-phosphate decarboxylase TCID 50 : tissue culture infectious dose for 50% infectivity (TCID 50 ) VLP: viral like particles VSVG: vesicular stomatitis virus G protein 16

19 CHAPTER 1 INTRODUCTION 17

20 1.1 HIV-1 and AIDS The epidemic of acquired immune deficiency syndrome (AIDS) started in the early 1980s in the United States by observation of increasing reports of patients presenting with life-threatening opportunistic infections. By March 1981, at least eight cases of Kaposi s sarcoma (KS), a rare form of skin cancer had been reported amongst homosexual men in New York. Simultaneously, there was an increase, both in California and in New York, in the number of cases of an unusual type of pneumonia caused by the agent Pneumocystis carnii. Within a few years of these observations, the virus was isolated from patients by two independent research groups (18,91). It was identified as a novel retrovirus called human immunodeficiency viruses (HIV) causing acute immune compromise and susceptibility to rare opportunistic infections. Currently, an estimated 33 million people globally are infected with HIV-1 (176). Since the beginning of the epidemic, nearly three decades ago, HIV-1 infection has claimed more than 25 million lives worldwide. Therefore, HIV leads to higher mortality than any other infectious disease worldwide. HIV-1 specifically targets the human immune system via infection of CD4+ helper T cells, macrophages and dendritic cells leading to cell disruption and general weakness of the immune system. 1.2 Retroviruses HIV-1 is a member of the Lentivirinae (Lentivirus) genus of the enveloped RNA viruses, Retroviruses that belong to the large and diverse Retroviridae family. These RNA 18

21 viruses can cause various diseases, such as leukemia, tumors, demyelinating diseases, and AIDS, but also are used as vectors for human gene therapy. The retroviral virion contains two copies of a positive single-stranded RNA that is non-segmented, 7 12 kb in size, and includes the 5 cap and 3 Poly(A) inside the virion (52). Retroviruses replicate through a distinctive life cycle in which virally encoded reverse transcriptase and integrase enzymes are used to copy the genomic RNA into double-stranded DNA and incorporate it into the host chromosomal genome. Next, the integrated DNA termed the provirus is transcribed into RNA to be used for protein production and genome packaging. Retroviruses are classified as either simple or complex based on their genomic structure. Simple retroviruses such as murine leukemia virus (MLV) encode only gag, pol and env genes (Fig. 1A), while the complex retroviruses such as HIV-1 contain additional small regulatory and auxiliary genes involved in various functions important for viral replication (Fig. 1B). Retroviruses are further subdivided into seven genera by the International Committee on Taxonomy of Viruses. The alpharetroviruses, betaretroviruses, and gammaretroviruses genera comprise simple retroviruses, while the deltaretroviruses, epsilonretroviruses, lentiviruses, and spumaviruses comprise complex retroviruses. Some representative members of each genera are briefly summarized in Table 1. 19

22 Figure 1. Schematic representation of simple and complex retrovirus genomes. (A) Representative simple retroviral genome. The genetic map of murine leukemia virus (MLV) contains three major coding regions, gag, pol, and env. The long terminal repeats (LTRs) include the unique 3 sequences (U3), repeat sequence (R), and a unique 5' region (U5). (B) Representative complex retroviral genome. In addition to structural proteins, HIV-1 expresses six additional proteins that regulate various facets of the virus lifecycle. The genes for regulatory and auxiliary proteins are tat, rev, vif, vpu, vpr, and nef, produced from singly and multiply spliced RNAs. The size of HIV-1 provirus is approximately 9.7 Kb. Both LTRs in proviral DNA have U3, R and U5 regions; while in genomic RNA U3 in 5 LTR and U5 in 3 LTR are removed after transcription of integrated proviral DNA. 20

23 Table 1. Retroviral Genera 21

24 1.3 HIV-1 Genome and Proteins The HIV-1 genome is comprised of nine genes from which six are unique to the HIV-1 virus and three (gag-pol-env) are shared by all retroviruses. The first protein encoded by the HIV-1 genome is the group-specific antigens (Gag) which is translated in the reading frame 1 and has an essential role in assembly of HIV-1 and other retroviruses. Gag is synthetized as a 55 kda polyprotein (Pr55) in the cytosol and transported to the cell membrane to coordinate viral particle assembly. In most cases, expression of Gag is enough for formation of viral like particles (VLPs) which are morphologically indistinguishable from infectious mature viruses. In order to study Gag localization, viral assembly and budding, fluorescent proteins are fused to Gag in different internal or C-terminal locations. If the fusion is done correctly, the resulted virions are able to function and assemble as efficiently as intact viruses. Alternatively, co-expression of unfused Gag proteins is required for formation of accurate particle morphology. Gag polyprotein precursor consists of four major domains as well as two spacers polypeptides: p1, p2 (Fig. 1B, orange) which are produced during maturation by virally encoded protease (PR) enzyme (85). Matrix (MA; p17) is the N-terminal domain which includes an N-terminally myristoylated globular head and a α-helical stalk. Matrix is the major structural protein in charge of directing Gag to the cell membranes. The next domain is capsid (CA; p24), which homo-oligomerizes in an ordered manner during assembly and is responsible for HIV-1 morphology. The nucleocapsid (NC; p7) domain allows viral RNA packaging by binding to the genomic RNA which also facilitates Gag 22

25 multimerization by simultaneous interaction of NC domain with an RNA molecule and other Gag proteins. The last C-terminal domain is called p6 which is less structured than other domains, but contains short peptide docking sites for the ESCRT and ESCRTassociated proteins necessary for virion envelope separation from the cell membrane during viral budding. The next protein, Pol is translated from ribosomal slippage into the -1 reading frame of the gag-pol transcript which was also used for Gag translation caused by frame shifting (124). pol gene is responsible for production of the enzymatic proteins including reverse transcriptase (RT), protease (PR) and integrase (IN) in order from 5 to 3 on the genome (Fig. 1B, blue). The reverse transcriptase enzyme is initially expressed as a 66 kda heterodimer consisting of p66 and p51 subunits, possessing the polymerase and ribonuclease H (RNase H) subdomains (149), and is subsequently proteolytically processed to form a 66-kDa and 51-kDa heterodimer by cleavage of the 15 kda carboxylterminus from the p66 subunit. RT is in charge of conversion of single stranded RNA genome into double stranded DNA after entry. Later, in the act of viral invasion of the host cell, the dsdna is integrated into DNA chromosome as a provirus by the 32 kda integrase enzyme. The Protease enzyme which is produced from an auto-catalytic cleavage process is required for the cleavage of immature Gag-Pol polyprotein precursors to the mature form (234). The env gene encodes the surface (SU or gp120) and transmembrane (TM or gp41) portions of the third major protein; envelope (Fig. 1B, red). In the endoplasmic reticulum, translation of a partially spliced bicistronic transcript which contains both 23

26 vpu and env ORFs produces the polyprotein precursor (Env, gp160) via the leaky scanning through vpu AUG (52). Following translation, the glycosylated Env precursor gp160 is transported to the Golgi complex to be cleaved into the gp41 and gp120 glycoproteins by host cell protease called Furin (28,57). In the next step, an envelope complex consisting of gp120 and gp41 is transported to the cell membrane where it can be incorporated into budding virions. The cell surface recruitment of Env complex is independent of the transportation of Gag and Gag/Pol precursor (87). The gp120 glycoprotein determines cell tropism by binding to the CD4 receptors and specific coreceptors on the cells. In addition to major proteins, alternative splicing of full-length mrna results in production of the regulatory and accessory proteins. The two regulatory proteins Tat and Rev and four accessory proteins Nef, Vif, Vpr and Vpu of HIV-1 are understood to different degrees (77,158,246). Among these proteins, the function of Tat and Rev are better characterized. Tat and Rev Proteins are each encoded from fully spliced transcripts containing two exons. Rev stands for the regulator of expression of virus and is a 19-kDa RNA binding protein (Fig. 1B) (79,173). The primary role of Rev is to aid in the nucleocytoplasmic transport of partially spliced and unspliced mrnas through interaction with a cis-acting element known as rev response element (RRE) [reviewed in (54,115,200)]. Rev binding also improves the stability and translational efficiency of RRE-containing RNAs. (for details see section 1.6.2). Tat (transcriptional transactivation protein) is generated from some RNA 24

27 transcripts early in cells which promote high level HIV mrna transcription is a positive feedback mechanism. Tat is responsible for viral mrna transcription by binding to the transactivation response region (TAR) hairpin located downstream of the transcription initiation site (23,206). Tat and TAR interaction results in recruitment of multiple cellular factors such as P-TEFb, cyclin T1 and CDK9 to the transcription complexes that promote processive polymerization by RNA polymerase II (11,161). Even though viral accessory proteins are called the non-essential accessory proteins, they are known to facilitate evasion from the intracellular innate and adaptive immune system and allow infection and functional replication in a subset of T- cell lines termed non-permissive cells (89). They include the Vif (viral infectivity factor), Nef (negative regulatory factor), Vpr (viral protein R), and Vpu (viral protein U) (Fig. 1B). Vif has been implicated as a viral infectivity promoter through inhibiting the encapsidation of apolipoprotein B mrna-editing catalytic polypeptide 3 (APOBEC3G or A3G), a relative of the activation-induced deaminase (AID) (225). APOBEC3G is an antiviral cellular protein which is encapsidated by budding viruses in the absence of Vif and converts cytidine (C) to uritidine (U) and subsequently causes hypermutation during reverse transcription in newly infected cells (25,107). It is not entirely clear, however, to what extent Vif is successful in neutralization of APOBEC3 proteins and protection from HIV-1 infectivity in vivo (157,225). Nef is a 27-kD myristoylated membrane-associated protein expressed early from nef gene that partially overlaps the 3 LTR. The fact that Nef is one of the first 25

28 proteins to be produced implies that it should be important for viral replication. Multiple functions have been reported for Nef, including: regulation of viral core s intracellular trafficking and helping virus to pass through the cortical barrier underlying the plasma membrane after fusion (39), internalization of CD4 from the infected cell surface of infected cells, as well as enhancement of viral infectivity. Downregulation of CD4 occurs through association of Nef with the cytoplasmic tail of CD4 and recruitment of AP2 (clathrin adaptor protein complex 2), internalization through clathrin coated pits, and subsequent transportation to the endosomes and lysosomes where Nef assists with degradation of CD4 (44,93). MHC-I and MHC-II as well as CD4, CD3 TCR complex and CD28-initiated co-stimulatory signal molecules are also shown to be down-regulated from the cell surface by Nef protein (26,215,238). The likely consequence of the limitation of antigen-specific signaling and cytotoxic T lymphocyte recognition in HIV-1-infected cells by Nef is to facilitate evasion from the host immune response. Vpu is an 81 amino acid dimeric integral membrane protein that is thought to be involved in enhancement of virus particle release and like Nef, degradation of newly synthesized CD4. Vpu is employed to co-localize with an interferon induced membrane protein termed tetherin and inhibit its activity. Tetherin functions to neutralize Vpu and causes retention of virions on the cell surface and therefore, block HIV-1 particle release (184). Vpu also induces the turnover of CD4 through direct interaction with the cytosolic end of CD4 that is associated with the newly synthetized gp160 precursor in the endoplasmic reticulum (ER) that would otherwise traffic to the cell surface for 26

29 incorporation into viral assembly (258). The vpr gene encodes a 14 kda protein which has several functions in viral life cycle. Vpr is incorporated into the viral particles to facilitate the nuclear targeting of the pre-integration complex (PIC) of the virus (202), influencing the mutation rate during reverse transcription (163), increasing the 3 polyadenylation activity (47), as well as inducing cell cycle arrest in the G2/M phase which contribute to the pathogenesis of HIV-1 (129). The protein-encoding regions are flanked by 5 and 3 LTRs, which consist of U3, R and U5. The LTRs harbor numerous cis-acting elements important for multiple steps of viral replication. The details of the signals in 5 LTRs are explained in and HIV-1 Virion A mature, infectious HIV-1 is formed through two steps: 1) budding of the immature, non-infectious particle and 2) maturation (86). The HIV-1 virion has an average size of nm in diameter which remains unchanged during maturation. The outside of the virus is surrounded by viral envelope proteins. Viral membrane is extended from the membrane of the host cells which produce the virus particle (40). HIV-1 packages two copies of a 9.2 kb single stranded RNA genome which is surrounded by nucleocapsid proteins encased in the cone-shaped viral core (or capsid) made from the protein p24. Three viral enzymes including reverse transcriptase and integrase and protease are also packaged in the virus (Fig. 2). 27

30 Figure 2. Organization of the HIV-1 mature viral particle. An HIV particle is around 110 nm in diameter and is surrounded by the viral envelope (or membrane). Multiple glycoproteins (gp120 and gp41) are projected from the envelope. Just under the envelope is a layer called the matrix, which is made from the protein p17. The double stranded RNA genome is surrounded by nucleocapsid proteins (NC) that are protected in the bullet-shaped viral core (or capsid) made from the protein p24. Three enzymes are also packaged in the virus, which are required for HIV replication. These enzymes are called reverse transcriptase (RT), integrase (IN), and protease (PR). 28

31 1.5 Early stage of HIV-1 Life cycle HIV-1 replication in infected cells is complexly linked to host cell biology and is broadly categorized into early and late events covering the series of finely organized steps (36,88,100). During early events of virus life cycle, HIV-1 binds to the host cell, enters the cell, and reverse transcribes its genome from RNA to DNA for integration of DNA into the host chromosome (187). Fig. 3 summarizes the early stage of HIV-1 replication. Late events start with transcription of integrated viral DNA by cellular polymerase and end with viral assembly and release Entry Early events begin when the viral envelope glycoprotein gp120 attaches to the cellular CD4 surface receptors and also one of the chemokine co-receptors, CCR5 or CXCR4 on the surface of the host cell. This interaction triggers a conformation change in the gp120 and exposure of the transmembrane (TM) glycoprotein gp41. Next, viral and cellular membranes are brought in close proximity which results in lipid fusion, and the viral core enters the cytoplasm (Fig. 3A) Reverse transcription Retroviruses have evolved a complex replication process that convert the genomic RNA into an expanded DNA transcript by addition of sequences at both ends required for controlling transcription and proviral DNA integration (Fig. 3B and 4). In the core, the viral enzyme reverse transcriptase (RT) copies the viral ssrna genome into a 29

32 dsdna copy which is transported by host cell proteins into the nucleus within the preintegration complex for integration into the host cell genome (112). Reverse transcription is initiated shortly after the virus enters the cytoplasm by binding of a host trna Lys,3 primer to the highly conserved 18-nucleotide primer binding site (PBS) located about 180 nts from the 5 -end of the RNA genome (Fig. 4A). In the case of HIV-1, trna Lys,3 is selectively packaged into the virion via interactions of the RT domain in pol with Gag and lysyl-trna synthetase (LysRS) (221). After annealing of the primer, RT reverse transcribes the U5 and R to generate minus-strand strong-stop DNA (Fig. 4B). Two enzymatic properties of RT, DNA polymerase activity and RNase H activity are essential for strand transfer during viral replication. The synthesis rate and accuracy of viral RT is lower than cellular polymerases (204), resulting in accumulation of mutations during reverse transcription which accounts for high diversity within HIV viral strains, especially in the presence of APOBEC3G (25,107). RNase H has both sequence specific and non-specific activities; both degrade RNA within an RNA/DNA hybrid (108). RNase H domain cleaves RNA sequences from the RNA/DNA complex in a sequence non-specific manner (Fig. 4C) (164). In order to complete synthesis of the minus strand DNA, the resulting (-) ssdna transfers to an equivalent R region in 3 LTR of the same or the other co-packaged genome (Fig. 4D) (242). Following DNA/RNA priming, the reverse transcription continues from the 3 LTR over the entire genome, ending at the PBS (Fig. 4E). RT RNase H makes a sequence-specific cleavage of the RNA genome, with the exception of two small conserved RNA fragments, which are resistant to RNase H cleavage because the fragments possess a unique polypurine tract (PPT) sequence (142). 30

33 The main PPT is termed 3 PPT and is located in the env region. The second PPT is central (c) PPT that is positioned upstream of the 3 PPT in the integrase domain of pol gene. Both cppts and 3 PPT can be used to initiate plus-strand DNA synthesis (Fig. 4F). However, only 3 PPT contains the att sequence which is required for efficient DNA integration; thus internally-initiated plus-strand DNA cannot integrate properly into the host DNA (243). Therefore, it is important for RNase H to pick the correct PPT sequences. The (+) strong-stop DNA is synthesized in a similar way to minus strand DNA from the 3 PPT to the 5 end until U3, R, U5, and the first 18 nt of the trna are produced (Fig. 4G). The plus strand DNA is then transferred to 3 end of the nascent minus strand DNA and complemented by the terminal PBS to complete the synthesis of full length double stranded DNA with duplicated long terminal repeat (LTR) ends (Fig. 4H). This dsdna product is ready for integration into the host DNA where it can be transcribed by cellular machinery Nuclear import/integration For transported to the cell nucleus (Fig. 3C), the newly synthesized DNA genome join a nucleoprotein structure known as the pre-integration complex (PIC) which also includes some viral proteins: IN, Vpr, RT, and MA (33). The mature HIV-1 PIC contains cellular proteins such as lens epithelium derived growth factor (LEDGF/p75) which has been initially shown to facilitate lentiviral integration through direct binding to IN within the PIC. It is suggested that this interaction protects IN from proteasomal degradation (152) and promotes its binding to the host cell chromatin (37). Other proteins such as 31

34 integrase interactor 1 (INI-1/hSNF5) which is the first identified IN co-factor, transportin- 3 (TNPO3), and transportin-sr2 (TRN-SR2) have also been proposed to join the PIC as nuclear import shuttles (33). Transport of the viral DNA between the cytoplasm and the nucleus occurs through the nuclear pore complexes (NPCs). PIC cannot be carried passively from the cytoplasm to the nucleus due to its size. To overcome this, lentiviruses adopted a strategy to actively transport the PIC through nuclear pore complexes in an ATP-dependent manner, resulting in their ability to infect non-dividing cells (32). MA and IN are believed to assist this process by recruitment of nuclear import machinery, such as importin α/β and importin 7 proteins via their functional nuclear localization sequences (NLSs) (236). However simple retroviruses, unlike lentiviruses, have to wait for cell division to occur for PIC to reach the chromosome. After PIC/DNA complex is localized in the nucleus, the PIC disassembles to permit the integration of the linear DNA. Although integration site selection is not fully understood, it is known that IN targets regions of chromatin DNA that are transcriptionally active. The process of integration (Fig. 3D) starts with integrase-mediated removal of two nucleotides from both 3 ends of ds DNA and generation of CA overhangs during an event called 3 -end processing. Concurrently, integrase cleaves the phosphodiester backbone of cellular DNA at the site of integration, resulting in generation of non-blunt 5 ends. The 3 recessed ends of the viral DNA bind with the 5 ends of the cellular DNA, while leaving gaps at the place of integration. In the last step, the gaps between the integrated viral DNA and cellular DNA are repaired and the ends are ligated by host cell enzymes to generate the provirus [for review see (248)]. 32

35 Figure 3. Early stage of HIV Life cycle. (A) Entry: Viral replication begins when the viral envelope glycoprotein gp120 H binds to CD4 and a chemokine co-receptor on the surface of the target cell. After binding to CD4, conformational changes in gp41 promote fusion of the viral membrane and target cell plasma membrane. Following viral core entry, the core is removed to release the viral RNA and proteins in the cytoplasm. (B) Reverse transcription: Single stranded genomic RNA is synthesized to a double stranded DNA in the cytoplasm of the target cell. (C) Nuclear import: When the reverse transcription is completed, pre-integration complex (PIC) including the provirus DNA, several viral proteins such as IN, Vpr, RT, and MA, together with cellular proteins are actively transported into the nucleus via nuclear pore complex (NPC). (D) Integration: Integrase (IN) catalyzes integration of viral DNA into cellular chromosomal DNA. Once integrated, the viral DNA is called proviral DNA. 33

36 Figure 4. The steps involving reverse transcription. (A) The trna Lys,3 (green) anneals to PBS hairpin of the ssrna genome. (B) (-) strand strong-stop DNA is generated by trna extension through the RNA in a 5 to 3 direction. (C) The 5 end of the genomic RNA is degraded by RNase H domain of RT. (D) (-) ss DNA are transferred to the 3 end of the genomic RNA, where the repeat region, R, at the 5 end of (-) ss DNA base pairs to the 3 R of genomic RNA. (E) DNA synthesis continues to complete DNA synthesis. (F) RNaseH degrades the genomic RNA except the polypurine tract, 3 PPT and/or cppt located in the RNA strand. (G) Synthesis of (+) strong-stop DNA proceeds from the 3 PPT 5 to 3 until a DNA copy of the PBS is generated. trna and PPT are removed by the action of RNaseH. (H) The second strand transfer occurs by base paring between complementary PBS sequences in both DNA strands, following by folding of the minus-strand DNA over the plus-strand DNA for binding of the U5, R, and U3 regions. Subsequently, each strand is used as template until a double stranded DNA containing both LTRs is completely synthesized. 34

37 1.6 Late stage of HIV-1 lifecycle The late stage of HIV-1 life cycle is also called the post-transcription phase since it starts by synthesis of viral RNA. Following transcription, the viral mrna is used for production of the structural, enzymatic, and regulatory proteins necessary for production of infectious virus (132). For the genomic RNA to follow the intracellular fate and reach its final destination which is the capsid of progeny virus, controlled assembly of multiple ribonucleoprotein (RNP) complexes are required. In general, the processing of retroviral RNAs is directed by interaction of several cis-acting elements with viral and cellular proteins. A summary of Late stage of HIV life cycle is shown in Fig Transcription and splicing of viral RNA Transcription of the integrated proviral DNA generates full length RNA genome (Fig. 5A) by a process similar to the rest of the host genomic DNA. Tat is the key transcription transactivator protein which interacts with the transactivation response element (TAR) cis-element in the 5 LTR, to facilitate the transcription function of cellular RNA polymerase II complex (RNAP II) (87). Unlike typical transcriptional transactivators which bind to a DNA, Tat protein binds to the bulge of an RNA stem-loop (80). Thus, the initial transcription of TAR is required for Tat binding, prior to HIV transcription. Tat-TAR interaction recruits the positive transcription elongation factor (PTEFb) which is a multicomponent kinase complex composing of cyclin T1 (CycT1) and cyclin dependent kinase 9 (CDK9) subunits. CDK9 phosphorylates the C-terminal domain of RNAP II, which in turn results in relieving the polymerase from negative elongation 35

38 factors and stimulation of transcriptional processivity. The presence of Tat dramatically increases (by more than two logs) basal transcriptional activity from the HIV LTR [for review see (15)]. Complex retroviruses have evolved sophisticated strategies to regulate the generation of multiple species of mrna (up to 30 products) from a single primary viral RNA. This function is controlled by employing a complex splicing pattern in which a portion of the pre-mrnas interacts with the cellular RNA splicing machinery. Three classes of viral mrnas are present in infected cells: 1) The unspliced, 9.2 kb RNA encoding Gag and Gag/Pol and also serving as genomic RNA, 2) The approx. 4 kb partially spliced RNAs expressing Vif, Vpr, Vpu and Env, and 3) The approx. 1.8 kb, multiply or fully spliced RNAs that encode Tat, Rev and Nef. The resultant mrnas are shown in Fig. 6. The production of this spectrum of viral RNAs is achieved through the combinatorial use of five 5' splice sites (SD1-5) and nine 3' splice sites (ss) (SA1-3, SA4a,b,c, SA5-SA7) (205) Nucleo-Cytoplasmic transport of viral RNA The export of eukaryotic cellular mrnas from the nucleus to the cytoplasm is a unique and fundamental step in post-transcriptional gene expression as it is tightly controlled by the rule that allows only fully spliced mrna to reach the cytoplasm. In contrast to cellular mrna, retroviral mrna are exceptions to this rule and can evade the restrictions by using viral regulatory protein (Rev) and a number of cellular proteins to transport the intron-containing viral transcripts from the nucleus (Fig. 5B) [reviewed in 36

39 (54,115,200)]. Fully spliced tat and rev mrnas are the first mrnas to be exported to the cytoplasm, leading to production of additional Tat protein and subsequent elevation of mrna transcription and Rev-mediated rescue of unspliced and partially spliced HIV-1 RNA species from the nucleus (182,216). Nucleo-cytoplasmic export is necessary for the cytoplasmic presence of viral RNAs responsible for genome packaging and protein translation. The Rev-RNA binding occurs at a complex stem-loop structure called Rev response element or RRE, located in the 3 end and within the intronic sequences of unspliced and partially spliced RNA (Fig. 6) (159,160). HIV-1 Rev has two functional domains at C and N termini. The C-terminus of Rev contains the leucine rich nuclear export signal (NES) which is recognized by Crm-1 upon multimerization of Rev monomers at the RRE (173). Crm-1 is a cellular nuclear export receptor that belongs to the Importin family (159). One feature of the Importin family is their regulation by the small GTPase Ran (178). Crm-1 classically is used for the export of 5SrRNA and proteins with NES (29,56). Nuclear export of viral RNA proceeds by interaction of Rev/Crm1 with RRE and other cellular proteins (Fig. 7A). The interaction of Crm-1 and Rev, initiates the delivery of RRE/Rev/Crm-1/RanGTP complex to the nuclear pore complex (NCP) and out of the nucleus with the assistance of nuclear pore proteins including Nup214 and Nup98, which bind to Crm-1 (Fig. 7A) (84,185). Following translocation of the RRE/Rev/Crm-1/Ran-GTP complex to the cytoplasm, Ranspecific GTPase-activating protein (RanGAP) hydrolyses the Ran-GTP to Ran-GDP in order to release the Crm1 from the complex (Fig. 7B) to be returned to the nucleus for additional Rev export. This action permits the detachment of RRE containing viral RNA 37

40 from the Rev in the cytoplasm. The other important Rev domain at amino terminus encompasses an arginine rich nuclear localization signal (NLS) which serves as a binding site for RRE and Importin-β to coordinate the import of Rev back to the nucleus (Fig. 7C) (239). This process of Rev Shuttle through the nuclear pores is facilitated by interaction of Ran-GDP to the Rev/Importin-β. In the nucleus, conversion of Ran-GDP to Ran-GTP by Ran-specific guanine nucleotide exchange factor (RanGEF) results in Rev dissociation from Importin-β (Fig. 6D). Importin-β is then exported to the cytoplasm for further Rev Interaction. In recent studies, it is also suggested that HIV-1 might contain other motifs in the genome in addition to RRE that mediate the nucleo-cytoplasmic transport. The reason for this hypothesis is the lack of the RRE-like RNA element in other retroviruses such as MMV and the possibility of sequences in proximity of a gag splice site (188,189) mrna translation Following export of multiple species of mrna into the cytoplasm, during the next post-transcriptional gene expression step, various regulatory and structural proteins are synthesized using mrna as template (Fig. 5C) (55). Eukaryotic translation includes three steps: initiation, elongation and termination. Translation initiation is considered as the rate-limiting step and is tightly controlled by the sequence and structure of RNA (14,100). Similar to cellular mrna, retroviral transcript contain a 7- methyl-guanosine 5 cap structure and 3 poly(a) tail. The 5 cap should be recognized by cytoplasmic eukaryotic initiation factors (eifs) to initiate the mrna translation at the first AUG start codon downstream of 5 UTR (102). Cytoplasmic initiation factor eif4e 38

41 which is part of eif4f cap-binding protein mediates 5' cap recognition and facilitates the recruitment of the complex to the 5' end (210). eif4f is a trimeric complex which is composed of other proteins including eif4e, the scaffold protein eif4g, and an ATPasedependent RNA helicase eif4a (139,172). Concurrently, 40S ribosomal subunit associates with eif2 ternary complex (eif2, GTP, Met-tRNA) and eif3 to form the 43S pre initiation complex. eif3 is the ribosomal dissociation factor and interacts with the central region of eif4g. eif4g acts as a platform which coordinates the connection of eif4e, eif4a and eif3. Interaction of the capped transcripts with eif4f recruits the charged 43S complex to scan the mrna in a 5'-3' direction until an initiator codon in appropriate Kozak consensus context is detected (14). These events trigger recruitment of the large 60S ribosomal subunit to join the initiator codon to form the 80S ribosome complex and elongation phase of translation ensues (31,105). Although cap-dependent translation initiation has been demonstrated as the main mechanism to initiate protein synthesis, translation can also rely on an alternative mechanism using the presence of highly structured internal ribosome entry sites (IRESs) in mrnas when cap-dependent translation is inhibited. Efficient ribosome scanning is possible in transcripts that contain a relatively short (<100 nt) and unstructured 5 UTR. Conversely, stably secondary structures placed between the 5' terminus of the RNA and the AUG initiation codon are refractory to ribosomal scanning and translation initiation (83,138). In this pathway, ribosomes interact in a cap-independent way and initiate translation at the next downstream AUG codon. IRESs are assumed to allow ribosomes to bypass the scanning of structured 5' UTRs. For example, it has been suggested that 39

42 IRES is also used for translation of Env in addition to the leaky scanning process which uses a Kozak sequence surrounding the vpu AUG start codon that promotes translation of the downstream env gene from bicistronic vpu/env (9). Combination of these two translation mechanisms maximizes the expression of Env Ribosomal frameshifting The 5 cap-dependent translation of most eukaryotic mrnas restricts protein synthesis to the first coding sequence on the mrna. However, some retroviruses display a variety of strategies such as ribosomal frameshifting to allow downstream open reading frames (ORFs) to be accessed. The coding region for the Gag and Gag-Pol proteins, translated from unspliced mrna, are in separate reading frames. A programmed single 1 frameshift has to occur to produce the Gag-pol precursor polyprotein. It is a process where a specific signal called Gag-Pol ribosomal frameshift signal (RFS) forces the ribosome to change reading frame from the 0 to the 1 frame during translation of Gag leading to production of the Gag-Pol (124,261). The ratio of Gag-pol to Gag molecule synthesis is approximately one to twenty (124,259). The ribosomal frameshift signal is explained in more detail in Genomic RNA dimerization and packaging All retroviruses, except foamy viruses, contain two copies of identical positivestrand unspliced, 5 capped, and 3 polyadenylated RNA genome (52). The diploid RNAs are provided during the late stage of the viral replication cycle, by a mechanism known 40

43 as RNA packaging or encapsidation that have been extensively studied [see reviews (24,59,103,192,208,219)] but still remained to be entirely understood. The process of HIV-1 RNA packaging is illustrated in Fig. 5. RNA packaging is a crucial step in the retrovirus lifecycle because viral particles that lack viral genome or contain a defective genome, function as defective interfering particles (DIPs) which are not infectious (24). Although a single genomic (g) RNA template can support reverse transcription and proviral DNA synthesis (119), genetic recombination and re-assortment of polymorphisms is a hallmark feature of the retrovirus and dependent on a diploid genome. Prior to packaging, intergenomic annealing initiates formation of loose noncovalent RNA dimer (Fig 5.D), which is then selected for encapsidation over the excess host cellular and viral spliced HIV-1 mrnas (Fig 5.E) (~99% of the total cellular RNA) (125,192). Extensive genetic and virological studies indicated that the stringent selectivity for packaging of grna is due to the recognition of a cis-acting packaging element (also referred to as Psi or Ψ) by the trans-acting nucleocapsid (NC) domain of Gag precursor protein via two zinc fingers. [reviewed in (24,60,208)]. In HIV-1, the Ψ signal is located in a ~120 nucleotides region downstream of the primer binding site (PBS) in the 5 UTR and extending into the 5 terminus of gag coding sequences (Fig. 8) (5,60,111,132,144,169). Ψ contains four stable hairpins referred to as stem loops 1-4 (SL1, SL2, SL3 and SL4). In particular, SL1 contains the primary dimerization initiation signal (DIS) which stimulates grna dimerization prior to packaging. Both SL2 (containing the splice donor site) and SL3 have high binding affinity for NC (6) but only residues in SL3 are recognized to form the independent principal 41

44 structure of Ψ signal. Several studies provided convincing evidence that deletion and mutations of SL3 which were designed to disrupt the base pairing with NC have been reported to decrease the genome packaging (51,111,168). High-affinity binding of both the splice donor site (SD) and SL3 with NC is mediated by interactions between the zinc fingers of NC and exposed guanosine residues in the GGUG tetraloop of SL2 and GGAG tetraloop of SL3 (69). However, deletion of SL1 and SL3 decreases but does not eliminate HIV-1 grna packaging (51), suggesting other HIV-1 RNA element(s) may play a role in encapsidation during viral assembly. SL4 is also a part of packaging signal which extends beyond the AUG start codon within gag. The major splice donor (SD) hairpin resides in the GGUG loop of the SL2 hairpin. Presence of the SD within the packaging signal may provide a potential mechanism for preferential selection of the full-length genomic RNAs over spliced RNA for packaging into the virions (169). This is due to the fact that SL3 is removed or eliminated upon splicing. The exposed loop of the DIS hairpin contains a hexamer palindromic sequence (GCGCGC) with an auto-complementary sequence (ACS) property (126,194). The GC-rich loops from two grna monomers interact by a Watson-Crick base pairing and form a loose unstable kissing loop intermediate (Fig. 9A) which promote RNA dimerization. The unstable kissing dimer recruits the NC domain of Gag resulting in rearrangement to form a thermodynamically more stable form called extended duplex (Fig. 9B). Since both DIS and Ψ elements overlap and both dimerization and packaging are promoted by NC, it is likely that these two processes are linked, or alternatively the RNA dimerization is a prerequisite step for RNA packaging (141,230). NC seems to mediate this process (Fig. 42

45 9C) through destabilization of the DIS hairpin and subsequent rearrangement of the base pairing between the palindromic sequences resulting in strand exchange between residues of the stems without disrupting the kissing loop (174). Formation of dimer form exposes the SL3 that was previously sequestered in monomer form, allowing for NC binding during RNA packaging. The essential elements for NC-RNA interaction have been shown to be two conserved Cys-X2-Cys-X4-His-X4-Cys (X=non-Cys/His amino acid) motifs that bind zinc with high affinity in vitro. (104) NMR studies determined the structure of HIV-1 NC with zinc fingers, and a representative structure is shown in Fig. 9C (60,235). In simple retroviruses such as moloney murine leukemia virus (MMLV), the grna packaging element does not appear to extend beyond a localized sequence at 5 end of the viral genome (97,162). However, in complex retroviruses like FIV, HIV-1 and HIV-2, grna packaging and dimerization maps to multiple sequences in 5 LTR, 3 LTR and the 5 end of gag gene, including the trans-activating-responsive (TAR) stem-loop, splice donor site and primer binding sequence (48,50,135,156,167,168,212,218) Virion assembly, budding, and maturation In retroviruses in the late stage of viral replication, the newly synthesized Gag protein precursors undergo myristoylation (86,95,208) and glycosylation (53,203) which mediates the multimerization of Gag and its assembly at the inner lipid raft site of the plasma membrane (PM) (Fig. 8F). This function promotes assembly of viral core 43

46 particles, which subsequently results in packaging of viral RNA (Fig. 8H). Gag is targeted and then attached to the plasma membrane by the myristoylated (myr) MA domain. The cell biology of HIV-1 Gag trafficking (Fig. 5F) is not well characterized compared to other aspects of viral life cycle. However, confocal microscopy and recent live cell imaging studies by Jouvenet et al (127,128) suggest that RNA containing the Ψ is captured in a perinuclear/centrosomal site to form a ribonucleoprotein complex following by their transport to the inner plasma membrane. In these studies, the genome RNA contained multiple stem loops that could bind to the bacteriophage MS2 coat protein, which could then be imaged in vivo upon GFP fused MS2 protein. Simultaneously, the HIV-1 Gag protein which was tagged with the mcherry fluorophore was visualized by fluorescence microscopy. They observed that grna molecules did not concentrate at specific sites at or near the PM in the absence of Gag, while were targeted at lipid rafts on PM in the presence of Gag. Virus release happens by budding, permitting the virus to obtain a host derived lipid envelope. The composition of the viral envelope and the plasma membrane are not equivalent, suggesting that virus assembly may occur in nonrandom subdomains of cellular membrane. A complex of cellular proteins named ESCRTs (endosomal sorting complex required for transport), which are originally the mediators of cargo protein sorting into the inner vesicles of endosomes, is required for virus budding (12). It has been shown that multiple component of ESCRT machinery interacts with the viral proteins during budding to promote virus release at the plasma membrane. 44

47 During or shortly after the nascent virions are released, the virus undergoes proteolytic maturation (Fig. 5I) in which the viral protease cleaves the immature Gag-Pol polyprotein to produce mature MA, CA, and NC proteins, which rearrange to form the mature infectious virion. 45

48 46

49 Figure 5. HIV-1 late stage of HIV life cycle associated with RNA packaging. Schematic description of HIV-1 RNA packaging during late stage of viral life cycle. (A) Transcription: Transcription of the integrated viral genome generates full length RNA genome that include the Ψ. Part of the RNA remains full length, while the other portion is spliced into fully and partially spliced RNAs. (B) RNA export: All the spliced and full length viral RNAs are exported from the nucleus to the cytoplasm for further processes. (C) Translation: Following export, full length RNA plays a dual role as either a genomic RNA for packaging into progeny virus or as the mrna template for ribosomal synthesis of Gag and Gag-Pol (by a frameshift: not shown) polyproteins. (D) RNA dimerization: Prior to RNA encapsidation, two monomeric viral genomic RNAs are fated to join at a dimer linkage site to form dimer via NC domain of Gag. (E) Dimerization of grna results in exposure of packaging signal which is sequestered in monomer form and subsequent recognition and binding of small number of NC. If the cell is infected with a heterodimer virus, then association of grna dimers in the cytoplasm leads to a random assortment of homodimeric and heterodimeric grnas. (F) Dimer grnas/gag complex is transported to the plasma membrane lipid raft mediated by the myristoylated MA domain of Gag. (G) RNA encapsidation: Dimer grnas are packaged within the capsid core and incorporated into the assembling virions. (H) Assembly: During virus assembly of the nascent virus, the packaged grnas become condensed in the core while binding to NC. (I) Maturation: The resultant virus like particles (VLPs) are released from the membrane; during budding, they undergo a process termed maturation, whereby polyproteins are cleaved by protease into mature proteins, resulting in the formation of fully infectious viruses. Abbreviations: MA: Matrix, CA: Capsid, NC: Nucelocapsid. 47

50 Figure 6. HIV mrna splicing. (A) Elements present in the provirus. The cis elements TAR, DIS, SD, Ψ, cppt, CTS, SA, RRE and PPT are indicated. (B) Viral mrnas are generated from provirus transcription and RNA splicing. This process results in production of three groups of viral mrna with 9.2 kb, 4 kb and 1.8 kb size range. Rectangles represent expected final protein products that result from transcription, RNA splicing and protease cleavage of protein precursors. Ribosomal frameshift site is located at the 3 end of gag. Abbreviations: LTR, long terminal repeat; U3, 3 unique element; R, repeat element; U5, 5 unique element; TAR, transactivation response element; DIS, dimerization signal; SD, splice donor site; SA, splice acceptor site; Ψ, packaging signal; cppt, central polypurine tract; CTS, central termination sequence; RRE, Rev response element; PPT, 3' polypurine tract; polya, polyadenylation signal; UTR, untranslated region; cap, terminal 7-methylguanosine;; AAAAAA, polya tail. 48

51 Figure 7. Rev/CRM-1 pathway for nucleo-cytoplasmic exports of intron-containing viral RNAs. (A) Interactions of nuclear pore proteins (Nup214 and Nup98) with Crm-1 carry the RRE/Rev/Crm-1/Ran-GTP complex through the nuclear pores complexes (NPCs). (B) After the RRE/Rev/Crm-1/Ran-GTP complex is transported to the cytoplasm, Ran-GTP is converted to Ran-GDP and dissociates from the complex. Crm1 re-imports to the nucleus to assist with more Rev export, while RRE dissociates from the Rev protein. (C) Binding of Importin-β the Rev initiates its return into the nucleus. Ran-GDP facilitates the entrance of Rev/Importin-β through the nuclear pores. (D) Finally, conversion of Ran-GDP to Ran-GTP in the nucleus, mediates detachment of Rev from Importin-β. Importin-β is then exported to the cytoplasm for more Rev Import. Binding of Rev to other viral RNAs recycles the pathway. 49

52 Figure 8. Sequence and secondary structure model of HIV-1 packaging signal in 5 UTR. The main stem-loops important for virus replication are represented. These structural elements are TAR, poly(a), U5-PBS hairpins as well as major packaging signal (shown in the box) containing the SL1 or DIS, SL2 or SD, the packaging core signal or SL3 and SL4 (the gag start codon). Nucleotides and numbering correspond to the HIV-1 HXB2 sequence. Adapted with permission from (219). Abbreviation: TAR: DIS: Dimerization initiation site, SD: major splice donor 50

53 Figure 9. Interactions between DIS of two RNA molecules and initiation of grna dimerization. (A) The palindromic GC-rich motifs in two DIS hairpins base pair to form a kissing loop complex and initiate RNA dimerization. (B) Binding of NC domain of Gag register-shifts that occur during dimer maturation and formation of extended complex which is more stable. Nucleotides numbering correspond to the HIV-1 HXB2 sequence. (C) Representative NMR structure of the HIV-1 NC domain of Gag protein. The zinc fingers are indicated as grey backbones. Cys and His are yellow and blue ligands. The zinc atoms are cyan blue and the flexible tails and linkers are red. Adapted with permission from (60,219) 51

54 1.7 Viral RNA Structure Secondary structure of major packaging signal in 5 UTR region Over the last two decades, numerous studies were performed to identify the RNA sequence and structures responsible for genome selection and packaging. Most of the findings are consistent with the primary packaging determinants that are located in the 5 untranslated region (UTR). As a result, most structural studies focused on this highly conserved region of retroviral RNA genome. HIV-1 5 UTR compromised of a collection of cis-acting highly ordered structures recognized by RNAs or proteins that mediate a variety of steps in the viral replication cycle, including reverse transcription, polyadenylation, splicing, nuclear export, RNA packaging, dimerization, and translation (19,22,181,245). The secondary structure of the HIV-1 5 UTR has been extensively studied and explored by biochemical, mutational and phylogenetic approaches in vitro (19,22,49,61,109,110,122,130,190). Using in situ chemical modification assay, this region was also investigated in the infected cells and in viral particles during viral assembly for the first time which showed very similar structure as in vitro (190). A few years later, a paper was published on the RNA secondary structure of the whole genome of NL4-3 HIV-1 by Dr. Weeks group (255). Different models for the architecture of secondary structure of these stem loop are proposed by a collection of site directed mutagenesis, chemical and enzymatic accessibility assays, phylogenetic studies, and free-energy calculations. Fig. 8 presents a simple model for secondary structure of 5 UTR. 52

55 D structure of major packaging signal in 5 UTR region In addition to the secondary structure models, more recent NMR studies by Dr. Summers lab have described the tertiary structure in conserved regions of the 5 UTR which is shown in Fig. 10 (154,155). The 5 end of capped HIV-1 RNA starts with the R region which is roughly 100-nt sequence. R includes TAR hairpin which is important for Tat-mediated transcription activation. In general, it has been difficult to investigate the role of TAR in other steps of replication by mutation analysis due to the negative effect that the mutations have had on transcription. For example in one study, disrupting the structure of TAR by mutation was shown to reduce genome packaging (64). In contrast, more recent studies indicated that that effect resulted from transcription reduction (50), and that in fact, TAR deletion does not affect RNA packaging (66). The highresolution three-dimensional (3D) solution NMR structure of TAR has been determined in the presence of TAT protein (Fig. 10A). TAR is followed immediately by the second hairpin, poly(a) which contains the AAUAAA polyadenylation signal. Poly(A) also exist in the 3 LTR region which mediates the polyadenylation of mrna. The function of 5 poly(a) is not understood, however, a mutational study has suggested a genome packaging role for Poly(A) (50). Later it was suggested that this effect was mainly caused by structural disruption of major packaging signal which is adjacent to this hairpin (65). The next stem loop contains the PBS which functions as a binding site for trna binding for reverse transcription initiation as described in Chemical and enzymatic probing methods have proposed slightly different secondary structural models for 53

56 residues of the PBS (1,2,61,62,109,256). Some of the structures suggest presence of an apical hairpin in an A-rich loop while some do not support this. The structure of the kissing complex in DIS GC-rich palindromic loops of RNA dimer has been determined by NMR and crystallography which shows a perfect coaxial alignment with the stems (Fig. 10B). As described before, the isolated splice donor (SD) hairpin has high binding affinity for NC. The 3D structure for the isolated SD hairpin in the presence of NC has been determined by NMR methods (Fig. 10C) (7). Deletion and mutations of SL3 which were designed to disrupt the base pairing with NC have been reported to largely abolish the genome packaging. The structure of SL3 which in known as the core packaging element was proved to be the mediator of its binding to NC since the packaging efficiency was restored by compensatory mutations which repaired the base pairing of the stem loop (51). The 3D structure of SL3 and NC complex is shown in Fig. 10D. AUG gag initiation codon has been suggested to adopt numerous secondary structures depending on the experiment (154,168). The 3D structure of AUG is solved using NMR method which is shown in Fig. 10E. Some mutagenesis studies and freeenergy predict a hairpin structure for AUG (109,167), while some other experiments suggest an unstable AGAG stem for the structure of this region (133) Interplay between RNA packaging and translation by RNA switch. Recent studies revealed that a conformation switch in the native 5'UTR can control the interplay between RNA packaging and translation. In one conformation, dimer-promoting residues and NC binding sites are sequestered and cap-dependent 54

57 mrna translation is promoted, while in the other one, these sites are exposed and instead packaging is stimulated (154). Some groups studied the regulation of different viral replication steps by conformational switch within the 5 UTR. They proposed a slightly different base-pairing pattern for each conformation, but they all agreed that this riboswitch is attributable to control of packaging versus translation. Abbick, Huthoff and Berkhout demonstrated that HIV-1 5 UTR can adopt either the branched multiple-hairpin (BMH) or the long distance interaction (LDI) alternative conformations (Fig. 11) (2,22,122). These two conformers differ in the presentation of the DIS hairpin motif and consequently in their ability to dimerize in vitro through longrange RNA RNA interactions between the AUG and residues in U5 that connect the poly(a) and PBS stem loop. Therefore, this LDI-BMH riboswitch has been shown to regulate the dimerization and translation. After binding of nucleocapsid protein to the RNA packaging signal, ribosome scanning would be arrested and efficient capdependent translation of the viral RNA inhibited (11,156,156). The SL3 and DIS hairpins are folded in BMH conformer (Fig. 11B) resulting in promotion of dimerization, NC binding and subsequently RNA packaging (Fig. 11E,F). On the other hand, these domains are base paired in LDI conformation (Fig. 11C) which is favorable for translation (Fig. 11D) (122). Efficient cap-dependent translation of the unspliced RNA results in disruption of the specific RNA structure in 5 UTR including the RNA packaging signal. Another group examined secondary and tertiary structural elements within the 5' 744 nucleotides of the HIV-1 genome, using combination of bioinformatics, enzymatic probing, phylogenetic and intra-molecular UV-cross linking strategies. They provided 55

58 supporting evidence for U5:AUG base paring (61). However, subsequent NMR studies of the stem loops in 5 UTR by Lu et al. (155) revealed a comparable model that is somewhat different, but is still supportive of previous BHM/LDL conformations. Taken together, these scenarios imply that the cellular translation machinery and viral assembly complexes compete for the unspliced viral RNAs in the cytoplasm. However as an outcome of this dissertation, we propose a model suggesting a novel mechanism for the interplay between translation and packaging. In our model, which will be described later in Chapter 6, genome-length RNA is first used for translation of Gag and Gag-pol. If the mrna is used for translation of Gag, it might result in formation of a complex involving release factor 1 (erf1) near the RFS after termination which initiates the NMD pathway via accumulation of other proteins. Otherwise, if the ribosome shift -1 nt at RFS to express Gag-pol (at <5% rate), it clears the RFS, prevents binding of Upf1-dependent decay complex, and promotes termination of translation while inhibiting RNA degradation. Eventually, the free RFS in the RNA that was used for Gag-pol translation can participate in long-range RNA-RNA or RNA-NC interactions which facilitate the genomic RNA encapsidation. 56

59 Figure 10. Representative 3D structures of relatively small RNA elements within HIV-1 5 UTR determined by NMR or X-ray crystallography (P,O atoms colored red; C, N atoms gray and blue, respectively, except as noted). Shown are (A) the NMR structure of the TAR hairpin bound to a cyclic peptide mimic of the Tat protein (brown) (67), (B) crystal structure of the DIS:DIS kissing complex (C atoms of the two RNA molecules shown in green and gray) (78), (C) the NMR structure of the SD hairpin bound to NC (backbone displayed as a ribbon; basic side chains displayed in blue color) (8), (D) the NMR structure of the Ψ hairpin bound to NC (in ribbon format) (69), (E) the NMR structure of the AUG hairpin (refined with a force field that included sodium ions, colored brown) (133), (F) the NMR structure of the NC protein (ribbon format, with basic residues shown in blue) bound to a U5 oligoribonucleotide (232). Adapted with permission from (155). Abbreviations: PBS, primer-binding site; SD, splice-donor site. 57

60 Figure 11. Model for translation and encapsidation regulation of HIV-1 unspliced RNAs. Unspliced transcript is transported to the cytoplasm where it can be used as grna or mrna. (A) The long-distance interaction (LDI) secondary structure allows the transcript to function as mrna for translation of Gag and Gag-Pol (C). (B) The branched multiple hairpin (BMH) secondary structure is proposed to initiate dimerization and packaging of grna by exposing the DIS (red) and Ψ (blue) core packaging signal for Gag recognition and interaction (E). (F) The putative Conformational switch between two secondary structures of 5 UTR is proposed to regulate HIV-1 grna packaging and mrna translation processes. The regulatory motifs are shown in different colors. The BHM and LDI secondary structures are adapted from a model proposed by Paillart,J.C et al (190,192). 58

61 1.7.4 Secondary structure of ribosomal frameshift signal Although the event of -1 ribosomal frameshift during translation which resulted in expression of the HIV-1 Gag-Pol polyprotein was shown experimentally almost 20 years ago (124), the structure of the programmed frameshifting signal (RFS) and its role in virus replication have remained partially unsolved. The ribosomal frameshift signal (RFS) includes two essential elements: a ribosomal stimulatory hairpin where the ribosome may be stalled and an upstream slippery polyu sequence, which facilitates and regulates ribosomal pausing and -1 nt shifting (30,74,76). For HIV- the polyu sequence of slippery hairpin is AAUUUUUU (255) or UUUUUUA (according to earlier models). Recently the structure of RFS has been characterized by Watts et al. (255) and suggested a type C three-way junction structure for RFS (143) in which a base stem (P1 in Fig. 12A) is connected with a stem loop containing the polyu slippery sequence (shown as P2 in Fig. 12) and ribosomal stimulatory hairpin (P3 in Fig. 12A). 59

62 Figure 12. Structure of the HIV-1 gag-pol ribosomal frameshift site (RFS). (A) SHAPE analysis of the HIV-1 RFS using pnl4-3 strain. Nucleotides reactivities are color-coded as shown in a scale in Fig. 30. (B) Sequence and structural conservation for the type C three-way junction model across 37 HIV-1 group M reference sequences. Adapted with permission from (255). 60

63 1.8 NMD pathway The nonsense-mediated mrna decay (NMD) is a RNA surveillance mechanism that ensures fast degradation of aberrant mrnas containing truncated ORFs due to premature translation termination codons (PTCs) (43). In addition to degrading PTC containing mrnas caused by a nonsense mutation, frameshift, aberrant alternative splicing, or rearrangement of genomic DNA, this pathway also regulates the stability of 1-10% of wild-type transcripts (170). The model for canonical termination suggests that stop codon is recognized by the eukaryotic release factors, erf1 and erf3, which results in interaction between PABP and erf3. These events promote speedy polypeptide release, disassembly of the ribosomal subunits, and reinitiation of the ribosome at the start codon (Fig. 13A). Close proximity of authentic stop codons with the poly(a) tail is thought to facilitate interactions between erf3 and PABP that positively contribute to peptide release. Transcripts preferentially targeted by NMD include those with PTCs encoded by alternatively spliced exons, introns downstream of the termination codon (TC), long 3 UTRs, or upstream open reading frames [reviewed in (186)]. Degradation of aberrant mrnas by NMD can affect the progression of inherited genetic disorders and many forms of cancer. Up-frameshift protein 1 (UPF1) is evolutionarily conserved RNA helicase that identifies premature stop codons or begins to specify the downstream events of NMD. Upf1 has been proposed to bridge the interaction between additional factors thought to mediate its association with mrna targets and regulate a cycle of Upf1 phosphorylation and dephosphorylation required for establishment of translational repression and recruitment of RNA decay enzymes. The exon-junction 61

64 complex (EJC), a multiprotein assembly deposited at exon-exon junctions in the process of splicing, directly interacts with the Upf1 and the terminating ribosome (Fig. 13B). The competition between Upf1 and cytoplasmic poly(a) binding protein 1 (PABPC1) for binding to the translation release factors erf1 and erf3 has been proposed to be a crucial factor in the decision to decay diverse transcripts. Upf1 binding to release factors at the terminating ribosome stimulates phosphorylation of Upf1 by the SMG-1 kinase, translational repression, and recruitment of decay factors. Conversely, binding of PABPC1 to release factors is proposed to preserve transcript stability and translational competence (227). An additional model (the 3 UTR model) proposes that Upf1 directly connects with the 3 UTR of an mrna, which would cause a PTC to be associated with a longer 3 UTR sequence than a true termination codon. This would provide a larger target for Upf1 binding. Also, when there is a long 3 UTR, the poly(a) tail is less proximal to the stop codon. Some studies have suggested that when is a stop codon is closer to the poly(a) tail it is less likely to be recognized as premature, while other studies argue that the role of poly(a) tail in NMD is minimal (Fig. 13B). 62

65 Figure 13. Translation termination and NMD pathway (227). (A) Canonical termination model. The model postulates that canonical translation termination involves the recognition of a stop codon by the eukaryotic release factors, erf1 and erf3. In the next step, interaction between PABP and erf3 promotes speedy polypeptide release, disassembly of the ribosomal subunits and reinitiation of the ribosome at the start codon. (B) Nonsense-mediated decay is activated by the presence of an exon junction complex (EJC) downstream of a premature stop codon or alternatively a long 3 UTR. In the EJC model, which is more common (particularly in lower eukaryotes), communication between the termination factors and the EJC is efficiently linked by Upf1 in association with the release factors. In the long 3 UTR model, a larger binding platform for Upf1 binding is provided, which initiate the NMD pathway. Gray rectangle, ORF; thick black line, 5 and 3 UTRs. In the 3 UTR model, UPF proteins may associate with a prematurely terminating ribosome because essential interactions between poly(a)-binding protein (PABPC1) and eukaryotic release factor 3 (erf3) have been disrupted due to the lack of their proximity. Adapted with permission from (227). 63

66 1.9 Lentivirus-based retroviral vectors One of the most promising applications of retroviruses is their modification to function as vectors for the delivery of corrective human genes for gene therapy due to their unique characteristic of long latency after infection and integration into the host chromosome. The ability of retroviruses to function as vehicles for therapeutic gene delivery into mammalian cells was established in the early 1980 s when It was discovered that mutated murine and avian retroviral RNAs that lack viral protein coding regions can be packaged if the missing viral proteins were provided in trans (99,226,240). These studies showed that basically a coordinate translation of Gag protein is not necessary to target RNA for packaging in the retroviruses. Based on these findings, in the initial design of retroviral plasmids, the segregated retroviral plasmids provided packaging and translation functions separately. One plasmid can yield the helper RNA to be used as mrna template for translation, while the other plasmid produces vector RNA to serve as genomic RNA for packaging. The same property was also observed in the lentiviruses such as HIV-1 and HIV-2 indicating that a coordinate Gag protein synthesis is not absolutely essential for packaging of lentivector RNA (79,107,136). In general, retroviral vector refers to a modified virus in which the structural genes are replaced with one or more of the strategic marker gene or therapeutic gene while keeping the intact cis-acting elements essential for gene expression and replication (118). Most retroviral vectors encompass the major packaging signal (Ψ+), but lack some or all of the viral protein coding sequences which makes them replication- 64

67 incompetent. To support a single cycle of replication of the vector RNA, the essential structural genes that comprise the virion are encoded in trans from a helper cell. Upon co-transfection of a vector plasmid along with a helper plasmid, only the Ψ+ vector RNA is packaged into helper virus particle because the helper RNA lacks the major packaging signal (Ψ) and functions exclusively as mrna template for translation of viral Gag, Polymerase, and Envelope proteins. In the new generation of lentiviral vectors the env gene is expressed from a separate plasmid called envelope vector (Fig. 14B). In a major improvement in the development of lentivurses, the retroviral envelope protein is substituted with the G protein of vesicular stomatitis virus (VSVG). Use of the VSVG envelope 1) broadens the range of targets by facilitating transduction of numerous cell types, 2) mediates lentiviral vector entry in the acidified endosome pathway, which suppress the requirements for some viral accessory proteins such as Nef protein (3), and 3) improves pseudotyped viral particle stability during ultracentrifugation, which allows for further concentration of the viruses to high titer (35). In some cases, the rev gene has been positioned in a separate cassette, thereby adding one more plasmid to the package HIV-1 derived lentiviral vectors The HIV-1 genome is a common backbone of gene transfer vector and was first considered as a lentivector in 1990s. However, because of the pathogenicity of HIV-1, genomes of other lentiviruses such as HIV-2, FIV, equine infectious anemia virus (EIAV) and simian immunodeficiency virus (SIV) have also been used to derive lentiviral 65

68 vectors. Lentiviral vectors are currently used in delivery of reporter or therapeutic genes in vitro, ex vivo and in vivo studies. An example is the successful delivery of green fluorescent protein (GFP) to non-dividing or terminally differentiated cells such as neurons, macrophages, hematopoietic stem cells, retinal photoreceptors, and muscle and liver cells, cell types for which previous lentivirus transduction methods could not be used Lentiviral vector packaging system Lentiviral systems have evolved since their first generation into safer and more efficient systems (Fig. 14A). The currently used packaging system is the third generation in which the trans acting elements are put on three separate plasmids to avoid the generation of replication competent lentiviruses (RCL). Early versions of HIV-based vectors contained only HIV cis elements; however, later version acquired additional heterologous elements through some modifications. The plasmid that provides the structural proteins is termed packaging vector, while the gene of interest is put in the transfer vector. The Env pseudotyping glycoprotein is expressed in trans from the envelope plasmid (Fig. 14). The first generation packaging plasmid expressed the entire HIV Gag, Pol, Tat and Rev as well as all of the viral accessory proteins (Fig. 14A). The structural proteins Gag and Gag-Pol are viral particle components, Tat transactivates transcription from LTR and Rev binds RRE to stimulate export of vector RNA from nucleus to cytoplasm. The Ψ and PBS signals were removed to ensure that this plasmid is only used to expresses viral 66

69 necessary proteins and enzymes without being packaged. Moreover, the viral LTR promoter was replaced with CMV promoter and poly(a) tail was also added to the 3 end. In the second-generation packaging system, with more understanding of viral genes for infectivity and virulence, four accessory genes (vif, vpr, vpu and nef) were also removed without affecting viral titer and infectivity (Fig. 14A) (96). The third generation which is the currently used packaging system has deletion in U3 region of the 3 LTR, which results in deleted U3 in both 5 and 3 LTRs in the proviral DNA following reverse transcription. These vectors are referred to as selfinactivating (SIN) vectors and their utilization prevents interference with the promoter/enhancer elements of the host genome following integration and thus mitigate the potential for insertional mutagenesis in transduced cells (177). Additionally, the regulatory gene, rev is placed on the transfer plasmid to increase bio-safety, and tat is removed (Fig. 14A) because the transfer vector harbors a constitutively active promoter instead of the enhancer/promoter sequence in U3 region of the 5 LTR (75) Design and improvement of lentiviral transfer vector In HIV-based lentiviral vector, the transgene expression cassette including internal promoter and transgene sequence is flanked by HIV cis acting sequences (Fig. 14C). LTR regions contain viral promoter and polyadenylation signal. Substitution of the U3 in the 5 LTR with a potent internal promoter rendered the independence from Tat protein which has been removed from the 3 rd generation of packaging vector as was 67

70 explained above (266). To avoid the possible genetic recombination with wild-type lentiviruses, the transfer vector only includes the minimum viral genome. The RNA from transfer vector is the only RNA that carries the Ψ and thus is preferentially incorporated into virions and then integrated into targeted cell. The RNA from the transfer plasmid also contains the PBS, splice donor, RRE, splice acceptor, and polypurine tract (PPT) (Fig. 14C). Recent versions of transfer vectors carry additional cis-acting elements to enhance efficacy of HIV-1 based lentiviral vectors. The elements that are routinely incorporated in lentiviral vectors include central polypurine tract (cppt) at the central termination sequence (CTS) located centrally in the pol gene and woodchuck hepatitis virus posttranscriptional regulatory element (WPRE). Interaction of the cppt-cts cis elements with the host cell nuclear import machinery and nuclear localization signals facilitates the generation of the central DNA flap determined to be essential for active preintegration complexes (PIC) at the step that involves nuclear entry (70,263). The addition of WPRE to the 3 side of transgene enhances mrna transcript stability and level of post-transcriptional transgene expression (265) due to efficient RNA processing. As described above, the U3 region of the 5 LTR was replaced with the CMV or RSV promoter, resulting in Tat-independent transcription (75). One of the major improvements that have been made is deletion of U3 region resulting in the selfinactivating (SIN) lentiviral vector with an inactive LTR during transgene expression. The deleted portion of the U3 region incorporated the TATA box and the transcription factors Sp1 and NF-κB binding sites. Even though the deletion does not reduce viral titer, it is advantageous for safety because it minimizes RCL generation further by 68

71 reducing common viral genome with wild-type HIV-1. It also decreases the possibility of host gene activation around the insertional site Production of lentiviral vector Triple transfection of the human embryonic kidney 293T cells with these three different plasmids including the packaging plasmid, the envelope plasmid, and the transfer vector plasmid yield the lentiviral particles carrying vector RNA. In order to avoid the production of empty viral particles, the plasmid ratio of packaging plasmid vs transfer vector plasmid vs envelope plasmid is 3:3: hours post transfection; the progeny lentiviral particles are harvested, filtered, and used for transduction of target cells. The transduction of cells with the lentivirus, results in a single cycle of vector RNA replication. Subsequently, the integrated vector provirus functions exclusively as mrna for synthesis of vector-encoded protein by the host cell machinery. Importantly, the vector does not express any viral proteins and infectious virions are not produced and do not spread to other target cells. 69

72 Figure 14. Schematic representation of lentiviral vector system. (A) Three generations of lentiviral packaging vectors. (B) Pseudotyping envelope vectors. (C) Improved lentiviral gene transfer vector. Abbreviations: Ψ: packaging signal; SD: splice donor; RRE: Rev responsive element; cppt: central polypurine tract; WPRE: woodchuck hepatitis virus post-transcriptional regulatory element. 70

73 CHAPTER 2 MATERIALS AND METHODS 71

74 2.1 Yeast recombination/gap repair cloning system. The Yeast recombination technique works when a plasmid with a double stranded gap is transformed into Saccharomyces Cerevisiae yeast cell along with an insert containing the regions of homology flanking the gap in the plasmid. The yeast performs homologous recombination in order to integrate the insert into the plasmid. The yeast prec vectors that were used in this study are derived from pcdna3.1 (Invitrogen) with addition of yeast centromere sequence (CEN6), the autonomously replicating sequence (ARSH4), and the β-isopropylmalate dehydrogenase (LEU2) for maintenance of this plasmid in media lacking leucine as described previously (74). Transformations were performed using the lithium acetate method. Briefly, yeast was grown in the YPD media overnight. The overnight culture was used to inoculate a new culture grown approximately 4-6 hours to remain in log-growth phase. This yeast was pelleted and resuspended in fresh LiAc (100 mm ph 7.5) and TE (10 mm Tris-Cl ph 7.5 and 1 mm EDTA ph 8.0). Approximately 1 ug of Insert/fragment mixture was added to competent cells along with 50 μg of single stranded salmon sperm carrier DNA (BD Biosciences/Clontech, Palo Alto, CA) and sterile polyethylene glycol (50%)/TE (10 mm Tris-Cl, 1 mm EDTA)/LiAc (100 mm). This mixture was incubated at 30 C on a shaker for 30 minutes. Yeast was then heat shocked at 42 C for 15 minutes and plated on agar plates containing the appropriate selection for 4-5 days at 30ºC. Colonies were grown in 2mL of CMM-Leu media, lysed with μm glass beads to extract crude DNA and transformed into electrocompetent Top10 cells (Invitrogen). Bacterial colonies were screened for the correct insertion by colony PCR using primers and sequencing. 72

75 Orotidine-5'-phosphate decarboxylase (URA3) gene is used as selective marker for insertion of HIV genome. Colonies containing plasmids in which URA3 is properly replaced with desired HIV genome are selected on minimal media plates lacking leucine but containing 5 FOA (5-fluoro-2-deoxyuridine monophosphate), i.e. toxic to yeast carrying a functional URA3 gene (27). This technology has been designed to allow cloning any gene in the HIV-1 backbone. To insert HIV-1 PCR products and replace the URA3 gene, the PCR product along with the prec/ura3 vector was transformed into the S cerevisiae yeast strain. Yeast colonies containing the recombined gene in prec vectors were then selected on plates containing CMM-Leu +5 FOA. FOA is converted to the toxic substrate 5-fluorouracil by the URA3. FOA-resistant yeast cells are then lysed to purify prec vectors. The success of this cloning procedure (replacement of the toxic URA3 gene with the gene of interest) is between 85-95%. It is important to note that the system has been fully tested. 2.2 Plasmid Construction Construction of prec_nfl_hiv-1, pcmv_ctpl, and 3 truncated vectors. The prec_nfl_hiv-1 and pcmv-ctpl vectors were previously generated in out lab (74). For prec_nfl_hiv-1 (also termed prec-nfl-3 LTR in this dissertation) construction, first, env was replaced with URA3 in the precenv plasmid using yeast homologous recombination. URA3 was amplified from the prs316 (228) vector with primers that also contained regions of homology to the first half of the NL4-3 genome and to precenv (Fig. 15B, step 1). Thus, the recombination of URA3 into precenv also inserts 40 nt of 73

76 homology to the first half of the NL4-3 genome to allow yeast homologous recombination of NL4-3 into precδenv/ura3 in the subsequent step (Fig. 15B, step 1, 2). The first half of the NL4-3 genome was PCR amplified from pnl4-3, a plasmid containing the full-length genome, and recombined into the prec/ura3 plasmid (Fig. 15B, step 2). Following insertion of URA3 into prec, the first half of NL4-3 was then inserted in place of URA3 immediately behind the CMV promoter under 5-FOA selection. Next, URA3 was inserted with regions of homology to allow for the recombination of the second half of the genome. Last, the second half of the pnl4-3 HIV-1 genome was inserted (Fig. 15B, steps 3, 4). The prec_nfl_hiv-1 plasmid was then digested with restriction endonucleases to confirm the correct insertion of the genome, as well as being sequenced in the sites of recombination to ensure that no insertions/deletions were introduced during the recombination process (Fig. 15C. and data not shown). In this plasmid, the transcription started from the CMV promoter is the first nucleotide of the primer binding sequence (PBS) in the HIV-1 genome. In order to generate the pcmv_cptl complementing vector, the cytomegalovirus (CMV) sequence was PCR-amplified from pcdna3.1zeo/cat (Invitrogen). This amplicon was utilized as a primer to amplify the R, U5, and gag regions from pnl4 3; the PCR product was then cloned into the pcr XL TOPO vector (Invitrogen). The TOPO vector was then subjected to digestion by MluI and BstXI to generate a CMV promoter-driven R, U5 and gag fragment. The resulting fragment was cloned back into the pcdna3.1zeo/cat backbone to generate the pcmv_cplt vector (Fig. 16). 74

77 In order to generate the 3 end truncated plasmids, a plasmid (prec_full length_hiv-1) containing the whole HIV-1 genome was truncated from 3 terminus by digestion with Xho I in one end and other enzymes to cut in different location to create shorter plasmids. When the plasmid is digested by only Xho I enzyme, 2 sites are cut at nt 8902 in HIV-1 genome and in the prec plasmid downstream of the HIV-1 genome which results in the production of prec-5 LTR/XhoI. Xho was used for digestion along with Sbf I, Bsu 36I, Nhe I, Bbvc and just by itself it cut fragments with the length of 2339, 5501, 6796, 7426 and 8433 bp respectively. 75

78 Figure 15. Schematic depiction of prec_nfl_hiv-1 production. (A) The HIV-1 genome has two regions of homology between the 5 LTR and 3 LTR which results in removing of the 8.5 kb intervening HIV-1 sequence during yeast homologous recombination of the full-length genome. (B) Stepwise cloning, outlining the introduction of near full-length HIV-1 pnl4-3 genome into the precenv shuttle vector. The first step demonstrates insertion of URA3 gene harboring 40 nt homologous region at both ends into a digested precenv. In step 2, precδenv/ura3 is digested with NotI, as well as a PCR product of the first half of pnl4-3 (starting from the PBS to the middle of the pol gene) being inserted via yeast homologous recombination. Next, URA3 is once again amplified from prs316 and inserted into precδenv/nfl, which is digested with XhoI (step 3). Finally, the second half of pnl4-3 is amplified and cloned into precδenv/nfl1/ura3 (step 4). (C) In order to verify the correct cloning, the resultant prec_nfl_hiv-1 vector was digested by restriction mapping with HindIII, PstI, and SacI, along with the pnl4-3 and the final pcmv_cplt plasmids. Adapted with permission from (74). 76

79 Figure 16. Schematic depiction of pcmv_cplt production. (A) The CMV promoter was amplified from pcdna 3.1zeo/CAT and then used as a primer for amplification of the R- U5-gag region of pnl4 3. (B) In the next step, both the amplicon and a plasmid named pcdna3.1 zeo/cat were digested with MluI and BstXI. (C) The digested amplicon containing NL4-3 R-U5-gag region was inserted into pcdna3.1zeo/cat to generate the pcmv_cplt complementing vector. Adapted with permission from (74). 77

80 2.2.2 Construction of a universal prec- HIV-1/URA3 shuttle vector. In order to construct the vector for universal HIV-1 cloning, in the beginning, URA3 selection marker was inserted into prec_nfl_hiv-1 (74) to replace the entire nfl HIV-1 genome. This Plasmid (prec- HIV-1/URA3) was used a general shuttle vector which allowed us to introduce any HIV-1 gene of interest containing approximately 40 nucleotides (nt) of homology with end junctions of URA3 in the plasmid. The resultant plasmids were used as RNA expressing prec-5 LTR plasmids for packaging studies Construction of prec-5 LTR plasmids using yeast recombination technique. Most of the plasmids in this study were prepared utilizing yeast homologous recombination. For generation of prec-5 LTR plasmids with incremental HIV-1 sequence, different lengths of the HIV-1 genome starting from the first nucleotide of R were PCR amplified from NL4-3 lab strain and cloned into prec- HIV-1/URA3. Every fragment to be inserted contained 40 nt of homology to the prec- HIV-1/URA3 to promote yeast recombination. Deletions ( nt and nt) were introduced into prec-5 LTR- RT and prec-5 LTR-RT by PCR amplification of two overlapping fragments which contained the deletions. The amplicons were recombined into prec- 5 LTR /URA3 vector (in which the HIV-1 genome downstream of nt 1091 was replaced with URA3). 78

81 2.2.4 Site-directed mutagenesis. The prec-5 LTR-2845-mu (#1-5), mutations were generated by site directed mutagenesis using Quick Change II XL Site-Directed Mutagenesis Kit as per the manufacturer s direction Construction of the remaining plasmids. Plasmid pnl4-3 GRPE was generated by a deletion between nucleotides1341 and 3717 using ligation bacterial cloning. A helper virus construct (pcmv R8.91) was used in pnl4-3 (WT and GRPE) transfection to compensate the lack of Gag and Gag-pol proteins production from pnl4-3 GRPE. Plasmid HIV Gag-iGFP is a replication competent molecular clone that has been generated by Hubner et al. (120). It is derived from NL4-3 and contains a Gaginterdomain green fluorescent protein (igfp) fused in MA-CA domains of Gag. The list of main prec-5 LTR vectors with HXB2 DNA and RNA numbering system is shown in Table 2. 79

82 Table 2. HIV-1 genome in prec-5 LTR vectors converting the HIV-1 HXB2 DNA numbering with RNA numbering. DNA numbering starts with the 5 end of the U3 region whereas RNA numbering starts with 5 end of the R region. Plasmid name DNA HXB2# RNA HXB2# prec-5 LTR-MA 1280 ( )nt (1-824)nt prec-5 LTR-CA 1415 ( )nt (1-959)nt prec-5 LTR-CA 1878 ( )nt (1-1422)nt prec-5 LTR-NC 2085 ( )nt (1-1629)nt prec-5 LTR-p ( )nt (1-1836)nt prec-5 LTR-PR 2549 ( )nt (1-2093)nt prec-5 LTR-RT 2845 ( )nt (1-2389)nt prec-5 LTR-Nef 8902 ( )nt (1-8446)nt prec-5 LTR-RT ( )nt ( )nt prec-5 LTR-RT ( )nt ( )nt prec-5 LTR-p6 868 ( )nt ( )nt prec-5 LTR-p6 716 ( )nt ( )nt prec-5 LTR-p6 706 ( )nt ( )nt prec-5 LTR-p6 996 ( )nt ( )nt prec-5 LTR-RT-mu (1-5) ( )nt (1-2389)nt prec-nfl-3 LTR ( )nt ( )nt 80

83 2.3 Cell culture. The human embryonic kidney 293T cells were was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, provided by Dr. Andrew Rice (101). The cells were cultured in Dulbecco s modified Eagle s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100μg/mL penicillin/streptomycin. U87.CD4.CXCR4 human glioblastoma cells were obtained through the AIDS Research and Reference Reagent Program and were maintained in DMEM supplemented with 15% FBS, 100μg/mL penicillin/streptomycin, 1μg/mL puromycin and 300 μg/ml G Transfection and infection. 293T cells were transfected with FuGENE6 (Roche) by the manufacturer s protocol. Briefly, T cells were grown in 100 mm dish for 24 h and were transfected with DNA mixtures containing 3x10 11 copies of prec-5 LTR_HIV-1 and prec-3 LTR_HIV-1 plasmids (~3 µg of each plasmid). The cells were washed at 8 h to remove unincorporated plasmid/liposome mixtures. Cells and cell-free supernatant were harvested 48 h post-transfection. Virus particle production in supernatant was monitored using a radiolabelled RT or p24 antigen capture kit. Virus was cleared of cellular debris by centrifugation at 1500 g for 10 min and filtered through a 0.45mm pore size membrane (Millipore). One quarter of the supernatant was utilized to measure RT activity and virus protein levels by Western blot and p24 antigen capture assays. Another quarter was used for infection, while RNA was extracted with the MagMax (Ambion) kit from the remaining half and then used to determine the level of grna encapsidation 81

84 Serial dilutions of the supernatant and following RT assays were performed for each virus mutant to compare with the RT activity relative to infectious titers of serially diluted wild type NL4-3 viruses. RT activity correlated with the infectious titer as measured by the Reed Muench technique to determine the tissue culture infectious dose for 50% infectivity (TCID 50 ) (92). As such, a virtual, expected infectious titer can be derived for the mutant viruses using the RT activity and infectious titer of a known wild type virus (NL4-3). To calculate actual infectious titers, mutant or wild type virus particle numbers were equalized based on RT activity (derived from serial dilution analyses), serially diluted (10-fold from undiluted; 8 dilutions), and then used to infect the U87.CD4.CXCR4 cells (96 well plate). Virus production was monitored over 10 days in the U87.CD4.CXCR4 cell cultures. The actual TCID 50 for U87.CD4.CXCR4 cells was then calculated using Reed-Muench technique (same as above). Titers were expressed as infectious units per milliliter. 2.5 Reverse transcriptase activity assay. Reverse transcriptase activity is a measure of virus production in supernatant and is used to quantify the level of transfection and infection. Supernatant (10μL) was collected and incubated for 2 hr at 37ºC with buffer (25μL) containing nucleotides datp, dgtp, dctp and [α32p]-labeled dttp. A volume of 10μL was spotted on DEAE filtermats, dried, and washed five times for five minutes each with 1X saline-sodium citrate buffer and twice with 85% ethanol. Filtermats were dried and radioactivity measured either by a Packard beta-counter to quantify counts per 82

85 minute (cpm) or filtermats were exposed to phosphoimaging screen for 2 hr at room temperature and densitometry quantified by Imager FX. 2.6 RNA extraction and quantitative real-time RT-PCR (qrt-pcr). Virion-associated RNA was extracted from 50 μl of transfection supernatant using MagMAX RNA Isolation kit (Ambion, Austin, TX) and was eluted in 50 µl. Whole cellular RNA was also extracted from transfected cell pellets using RNA RNeasy kit (Qiagen) according to the manufacturer's instructions. Cytoplasmic and nuclear RNA isolations from transfected cells were performed using the PARIS Kit (Ambion) according to manufacturer protocols. All isolated RNAs were subjected to a DNase digestion using a Turbo DNA-free kit (Ambion) and treated with RNaseOUT RNase inhibitor (Invitrogen). Quality and quantity of viral RNA from cell-free supernatant, total cellular RNA, as well as cytoplasmic and nuclear fraction was determined by absorbance at 260 nm/280 nm and electrophoresis on formaldehyde agarose gels. The proper fractionation of cytoplasmic and nuclear RNA was tested by RT-PCR amplifying the spliced and unspliced actin respectively. The OTR580, OTR581 and OTR582 primers were employed as described (98). The resulting products were photographed on 2% agarose gels stained with ethidium bromide. The amount of total RNA extracted from virions were measured using RiboGreen RNA Quantitation Kit (Invitrogen) and input of RNA for RT-PCR was normalized based on the concentration. Next, cell and virion-associated RNAs were reverse-transcribed using specific antisense primer Gag CCCCGCTTAATACTGACGCTCTCGC-3 for 5 LTR-cDNA 83

86 and oligo dt primer (Invitrogen) for 3 LTR cdna using Superscript III transcriptase (Invitrogen). In details, 5ul of RNA was added to 40 pmol of primer and 10 mm dntps and subjected to the following cycles; 65 C for 5 min and 4 C hold. Next, 5X first strand buffer (Invitrogen), 0.1 M dtt (Invitrogen) and 1 ul of enzyme were added to each reaction and cycled for 55 C for 60 min, 70 C for 15min and 4 C hold. All PCR primers were synthesized by Invitrogen, and Taqman fluorescent-tagged probes were synthesized by Applied Biosystems. The list of the primers and probes which were used for PCR, RT-PCR and q-rt-pcr is shown is table 3. Each 15 µl reaction contained 0.5 µl of each forward- and reverse-specific primer (25 µm), 0.75 µl of probes (5 µm), 7.5 µl of 2X Universal Taqman PCR Master Mix, 3.75 µl of RNase-free water, and 2 µl of template cdna. The reactions were carried out under the following conditions: 50 C for 2 min, 95 C for 10 min, 40 cycles of 95 C for 15 s and 60 C for 1 min. The ABI PRISM 7000 sequence detection system (Applied Biosystems) was used for qrt-pcr amplifications. The U5 and U3 target sequences were specifically amplified using the primer pairs and probes described in Fig. 4A. The 18S ribosomal RNA was amplified as internal control for cellular RNA with the forward primer human R18.seq-948F and reverse primer human R18.seq-1014R, and detected with probe R18. Absolute RNA copy numbers were calculated by using 10-fold serial dilutions of plasmid DNA with known copy numbers containing the target sequence (ranging from 10 4 to 10 9 copies). For 18s ribosomal RNA standard, quantuserial dilution of mrna Universal 18S Internal Standard (Ambion) was used. Lack of plasmid DNA contamination from the DNA transfection of cells was routinely confirmed by a control excluding reverse transcriptase (RT). Agarose 84

87 gel analysis was used to verify that each primer pair produced single amplicons, and the identities of the PCR products were verified by cloning into TOPO plasmid followed by DNA sequencing. It should be noted that for all RNA analysis in this study, after RNA extraction from cells and supernatants and prior to RT-PCR reactions, all the RNAs were treated with DNase and subsequently amplified. The absence of any detectable level of plasmid DNA contamination was confirmed by lack of amplification in control reactions without RT enzyme, followed by quantitative PCR amplification. Having confirmed this, RNA preparations were reverse transcribed to make the cdna. Another control experiment was also performed in order to evaluate the cellular RNA background level by running a real-time PCR with RNA samples extracted from supernatant of mock-transfection. These background signals were removed from the RNA copy numbers measured in HIV- 1 positive samples and the RNA data were normalized for viral particle production as quantitated by Reverse transcription assay. 85

88 Table 3. Primers and probes used for PCR, RT-PCR and q-rt-pcr. Primer/probe name Sequence Description Gag (-) CCCCGCTTAATACTGACGCTCTCGC 5 LTR sgrna q-rt-pcr Primer U5 (+) CGTCTGTTGTGTGACTCTGGTAACT 5 LTR sgrna q-rt-pcr Primer U5 (-) CTGCTAGAGATTTTCCACACTGACTAA 5 LTR sgrna q-rt-pcr Probe U5 6FAM-AGATCCCTCAGACCCT-MGBNFQ 5 LTR sgrna q-rt-pcr Primer U3 (+) GAGAGAGAAGTGTTAGAGTGGAGGTT nfl-3 LTR sgrna q-rt-pcr Primer U3 (-) CTCTCGGGCCACGTGATG nfl-3 LTR sgrna q-rt-pcr Probe U3 6FAM-ACAGCCGCCTAGCAT-MGBNFQ nfl-3 LTR sgrna q-rt-pcr Primer R18 GACGGTATCTGATC 18S ribosomal RNA RT-PCR Primer R18.seq- 948 (+) Primer R18.seq (-) CGCCGCTAGAGGTGAAATTC CATTCTTGGCAAATGCTTTCG 18S ribosomal RNA q-rt-pcr 18S ribosomal RNA q-rt-pcr probe R18 6FAM- ACCGGCGCAAGACGGACCAGA-TAMRA 18S ribosomal RNA q-rt-pcr Primer OTR580 (+) TGAGCTGCGTGTGGCTCC Spliced actin mrna PCR OTR581 ( ) (GGCATGGGGGAGGGCATACC 3 Spliced/unspliced actin mrna PCR OTR582(+) CCAGTGGCTTCCCCAGTG 3 Unspliced actin mrna PCR 86

89 2.7 Western blot and ELISA analysis. 293T cells from a 10 cm 2 dish transfection were lysed with lysis buffer (cell signaling) and heated at 95 C for 10 min. Virus-containing supernatants were collected from transfected 293T cells. Virus was pelleted at 32,000 x g for 1 h at 4 C then resuspended in 80 μl of SDS lysis buffer. Samples were loaded and electrophoresed on 10% SDS PAGE gels (bio rad). Protein was transferred onto nitrocellulose (Whatman, Dassel, Germany) at 100 cvolts for 1 h. Nitrocellulose membranes were blocked in PBST-5% dry milk for 1 h and probed with mouse anti-p24 (#24-4-AIDS Research and Reference Reagent Program) (1:100,000 dilution), Goat anti- Actin (C11-Santa Cruz Biotechnology, Inc) (1:5,000 dilution), or Rabbit Anti-eRF1 (Ambion#ab31799). Membranes were then labeled with either a goat anti mouse HRPlabeled secondary antibody or a rabbit anti-goat HRP-labeled secondary antibody. Blots were developed using the ECL-plus chemiluminescence kit (GE healthcare). Gag-P24 ELISA was also performed on viruses from transfection in order to quantify the virus production. 2.8 RNA synthesis RNA secondary structure prediction analysis. High-throughput selective 2' hydroxyl acylation analyzed by primer extension (hshape) was utilized in this study for prediction of RNA secondary structure with single nucleotide resolution (255). This approach employs a novel chemical probing technology, reverse transcription, capillary electrophoresis and secondary structure prediction software to determine the structures of RNAs from several hundred to several thousand nucleotides at single nucleotide resolution. 87

90 RNAs were prepared by in vitro transcription using the MEGAscript T3 kit (Ambion) from a PCR product (with T3 extended primer) derived from NL4-3 following the manufacturer s protocol. Transcripts were purified using MEGAclear kit (Ambion) and the size and integrity was checked by denaturing formaldehyde agarose gel electrophoresis. 20 pmol of RNA were heated at 95 C for 3 min in 20 μl of renaturing buffer [10 mm Tris HCl (ph 7.5), 100 mm KCl, and 0.1 mm EDTA] and slowly cooled to 4 C. Subsequently, 97 μl of water and 29 μl of 5x folding buffer [200 mm Tris HCl ph 7.5, 650 mm KCl, 2.5 mm EDTA, 25 mm MgCl 2 and 40U RNase inhibitor] were added and the RNA was incubated at 37 C for 15 minutes. Alternatively, RNA was refolded for 1h at 37 C (255) in 50 mm Tris HCl ph 8, 200mM KCl, 5mM MgCl 2.The mixture was divided into equal parts, treated with 7.3 μl of 20 mm 1M7 in anhydrous DMSO (+) or DMSO alone (-) (Fig. 17A) and the reaction was allowed to proceed at 37 C for 5 minutes (Fig. 17B). The RNA was precipitated and resuspended in 15 μl of 1mM Tris ph 8, 0.1 mm EDTA. Fluorescently labeled primer (1μl) was added to the 2.5pmols of RNA [Cy5 (+) and Cy5.5 (-); 8μM] and 12 μl of primer-template solutions were incubated at 85 C for 1 min, 60 C for 5 min and 35 C for 5 min. RNA was reverse transcribed at 50 C for 45 min with 100 U RT (Invitrogen superscript III), 1 RT buffer (Invitrogen), 5 mm DTT and 0.5 mm dntps (Fig. 17C). After adding 1ul of 4N NaOH and heating 3 min at 95 C to hydrolyze the RNA template, samples were neutralized with an equivalent amount of HCl. Sequencing ladders were prepared using the Thermo Sequenase Cycle Sequencing kit (USB) according to the manufacturer s protocol. Primers labeled with WellRed D2 and LicorIR-800 dyes were used. Samples and sequencing ladders were precipitated, washed 88

91 twice with 70% EtOH, dried and resuspended in deionized formamide. Primer extension products were analyzed on a Beckman CEQ8000 Genetic Analysis System (Fig. 17D). Electropherograms were processed using the ShapeFinder software (Fig. 17E) (249). 1M7 reactivity at each nucleotide position was normalized as described (256). Briefly, peak area difference for each nucleotide position was divided by the average of the highest 8% of peak area differences calculated after excluding outliers. Outliers were defined as values greater than the 3rd quartile plus 1.5 the interquartile difference (Fig. 17F). Resulting data was introduced into RNAstructure software as pseudo-energy constraints. Secondary structures were visualized using PseudoViewer web application ( M-fold web server ( was used to predict secondary structures of RNA mutants. 89

92 Figure 17. Overview of high-throughput selective 2 -hydroxyl acylation analyzed by primer extension (hshape). Adapted with permission from (249). 90

93 2.9 3D structure prediction. The concept of the RNA 3D structure system is founded on translation machine system which is comparable to that employed in computational linguistics (23). In summary, the secondary structure of 5 LTR-2845-wt and mu1-5 RNAs were introduced to the RNA FRABASE database (webserver). The translation engine operates on the RNA FRABASE database tailored to the dictionary relating the RNA secondary structure in the dot-bracket notation and tertiary structure elements using different dictionaries (databases). First, user-defined RNA secondary structure was fragmented to the elements. Related tertiary structure elements were then found in the RNA FRABASE dictionary. The tertiary structure elements were superimposed with reference to common canonical base pairs, and merged to give an initial RNA 3D structure. This structure was subjected to the refinement using energy minimization in torsion angle space and Cartesian coordinated to generate the final, high-quality RNA 3D model RNA interference. A sirna pool including three sirnas against erf1 (Ambion), s4838, s4839, and s4840 was employed to transfect 293T cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer s protocol. The final concentration of total sirna was 50 nm. 24 hours after sirna transfection, 293T cells were co-transfected with total 0.6 g of prec-5 LTR and prec-nfl-3 LR vector plasmids using FuGENE6 (Roche). Seventy-two hours after transfection, the cells and supernatant were collected for further experiments. More detailed procedure was previously described (137). For fluorescence microscopy, a 6 well plate which contained 293T cell attached to three 91

94 coverslips (place at the bottom of each well), were transfected as explained above. 72 hours post transfection; the cells were fixed with 4% electron microscopy (EM) grade formaldehyde (Polysciences) in PBS for 20 minutes at room temperature. Coverslips were rinsed in PBS and were mounted on glass slides using Gel Mount (Biomedia) for surface accessibility assays. A Deltavision RT epifluorescent microscope (Applied Precision, Inc.) was used to image slides. Images were captured in z-series and deconvolved using Softworx software (Applied Precision, Inc.) In vitro RNA dimerization. 5 LTR RNAs starting with nt 456 (5 LTR-MA 1280, 5 LTR- RT 2845, 5 LTR- RT-mu5, 5 LTR- RT , 5 LTR- RT , and 5 LTR-p6 868) and a shorter version of 3 LTR RNA starting with nt 634 and extending up to 2292 nt (3 LTR-P ) were produced by in vitro transcription as described above. Transcripts were purified using MEGAclear kit (Ambion), precipitated using ammonium acetate as directed, and resuspended in water. For experiments using homodimer RNAs, 200 nm of RNA fragments were diluted in 8 L of RNAse-free water (Ambion). For experiments using heterodimer RNAs, 100 nm of each diluted RNAs in 8 L of water were added to 10 nm of the described labeled RNA. Samples were denatured at 88 C for 2 min and then incubated on ice for 2 min. 5 dimerization buffer [50 mm sodium cacodylate, ph 7.5, 300 mm KCl, and 5 mm MgCl 2 (229)] was added and the mixture was incubated at 37 C for 30 min. For denaturation curves, different aliquots of the dimer RNAs were incubated at the indicated temperatures for 10 min and then placed back on ice until electrophoresis. 6 loading buffer (40% glycerol, 1 dimerization buffer, bromophenol 92

95 blue, xylene cyanol) was added to each sample, and separated on TBE-agarose gels containing ethidium bromide at 5 V/cm, 4 C (0.8% Agarose 3:1 high resolution blend [Amresco], 89 mm Tris [Amresco], 89 mm Boric acid [Sigma-Aldrich], 2 mm EDTA [Fisher]). Homodimer images were captured on AlphaImager EP (Alpha Innotech) and bands were analyzed using AlphaView (Alpha Innotech) software. Melting temperature (T m ) was defined as the temperature at which the amount of dimeric RNA was reduced by 50%, as compared with its value at 37 C. Labeled gels were fixed in 10% trichloroacetic acid for 10 min, dried for 2 h under vacuum at room temperature then 2 h under vacuum at 60 C Vector and lentivirus production. plv-mnd-palag was a kind gift of Dr. Yuan Lin and Dr. Stanton Gerson (150), containing egfp, luciferase and MGMTP140K genes under the control of MND promoter, which enhances transgene expression in hematopoietic cells in a SIN lentiviral backbone. Three fragments from nt , nt , and nt of HIV-1 were amplified from pnl4-3 plasmid using primers that contained SbfI and ClaI enzyme sequences. Plasmid was digested with SbfI and ClaI cutting two sites separated by 45 nucleotides, located between Ψ and RRE. SbfI and ClaI were also used to digest the GRPE PCR fragment, and then the fragments were ligated to plv-mnd-palag, to generate plasmids: plv-mnd-palag +( ) -GRPE, plvmnd-palag+( ) +GRPE, and plv-mnd-palag+( ) +GRPE. After cloning, plasmid was digested and sequenced to ensure the correct insertion and orientation of the fragments. Lentiviruses named PALAG -GRPE, PALAG +( ) - 93

96 GRPE, PALAG+( ) +GRPE, and PALAG+( ) +GRPE were generated by triple transient transfection in 293T cells as described in (150). In brief, component plasmids included pmd.g (VSVG envelop plasmid), pcmv R8.91 (packaging plasmid), and plv-mnd-palag were added to 293T cells at a ratio of 1:3:3 with Fugene6 (Roche) in Opti-MEM (Gibco, Paisley, UK), and 12 hours later, Opti-MEM was replaced with fresh DMEM (Mediatech, Manassas, VA). For each 10 cm 2 dish containing T cells, a total of 9 µg DNA plasmids were used. After another 36 hour-incubation, lentiviral supernatant was collected and filtered with 0.45 µm millipore steriflip filter (Millipore, Billerica, MA). The lentiviruses were concentrated by centrifuge for 2 hrs at g using 20% sucrose cushion. 1ml of the supernatant at the bottom of the tubes was collected and centriguded again at g for 1 hr. The pellet was dissolved in 1 ml of PBS. Virus titer was measured by RT assay on serially diluted supernatant to calculate TCID50 explained in In vitro lentvirus transduction. 293T cells were plated in two 96-well plates in triplicate (1x104 cells/well) for luciferase and GFP gene expression experiments. Lentiviral vectors LV-mnd-PALAG (+/-) GRPE supernatant from transfections were serially diluted with highest concentration at MOI of 1 and were added to each well in the presence of 8 µg/ml polybrene (Sigma-Aldrich, Milwaukee, WI). After 72 hours, supernatant from one plate was removed and cells were lysed using 50 μl Glo lysis buffer (Promega, Madison, WI). Cells were lysed for 20 minutes at room temperature, transferred to white polystyrene 96 well plates, and read with a PerkinElmer VICTOR 94

97 plate reader using injectors to introduce 50μL Bright Glo reagent per well (Promega, Madison, WI). Luciferase activity was measured as relative light units (RLU). Cell in the other were measured for GFP expression by flow cytometry Flow cytometry. FACS analysis of GFP expression was performed on 293T cells. Cells were pelleted at 2000 rpm for 5 min, washed with PBS, and washed in FACS staining buffer (PBS, 5% FBS, 1% BSA, 0.1% sodium azide). Cells were again pelleted (2000 rpm x 5 min) and resuspended in FACS staining buffer for GFP detection. Cells were diluted in an additional 150 μl FACS staining 96 buffer and pelleted again (2000 rpm x 5 min) and resuspended in 400 μl FACS staining buffer for analysis. Cells were analyzed on a FACScalibur flow cytometer (Beckton Dickinson). 95

98 CHAPTER 3 OPTIMAL HIV-1 GENOMIC RNA PACKAGING IS DEPENDENT ON A CIS ACTING RNA SEQUENCE FOUND AT THE 3 TERMINUS OF GAG USING A DUAL RNA COMPLEMENTATION SYSTEM Mastooreh Chamanian 1,2, Dawn M. Dudley 1,2,Kenneth R. Henry 2,Richard M. Gibson 2, Janice Ha 2, Yong Gao 1,2, Eric J. Arts 1,2 1 Department of Molecular Biology and Microbiology 2 Division of Infectious Diseases, Department of Medicine Case Western Reserve University, Cleveland, OH, 44106, USA Figure 18 is adapted from Dudley et al. (A novel yeast-based recombination method to clone and propagate diverse HIV-1 isolates. Biotechniques [(74)]) Figures 21, 22, 23, 24, 25, 26, and 27 are adapted with modification from Chamanian et al (A cis-acting element in retroviral genomic RNA links Gag-Pol ribosomal frameshifting to selective viral RNA encapsidation, Cell Host and Microbe, Feb 2013 (42)). 96

99 3.1 Preface Several years ago, our laboratory adopted a yeast-based cloning strategy for HIV- 1 to circumvent intricacies associated with HIV-1 heterogeneity and difficulties regarding conventional ligation cloning. After establishing this technique for HIV-1 genome cloning, additional studies were performed to optimize the system. These studies demonstrated that extending the HIV-1 genome to the coding sequence in the complementary vector, (which was used in this system as the template for reverse transcription and therefore needed to be packaged in the virus for infectivity) resulted in production of more infectious viruses. This observation suggested the existence of potential element(s) in the coding region of HIV-1 involved in a process that affects the packaging of viral RNAs. After my involvement in this project, we aimed to explore this hypothesis by using the unique recombination technology as well as a dual transfection system that allowed us to examine the existence of a packaging signal in a nontranslated HIV-1 RNA genome. The first part of this chapter is based on previous work on yeast recombination technique and dual complementing transfection system for virus production. The second part is about identification of additional packaging signals within the HVI-1 genome. All the work pertaining to the development of the yeast recombination system was done by Dawn Dudley, Yong Gao, Kenneth Henry and Richard Gibson. The original idea for optimizing the complementing system by increasing the length of complementing vector was proposed by Yong Gao. Comparison of the infectivity of 97

100 viruses made from short and long complementing vectors was performed by Yong Gao and Kenneth Nelson. The prec_nfl_hiv-1 (prec-nfl-3 LTR) vector was produced by Dawn Dudley. The prec HIV-1/URA3 and prec-5 LTR vectors for packaging study were produced by the author. Transfection assays, virus harvesting and reverse transcriptions were set up by the author and Janice Ha. The Author performed all other experiments and analyses on RNA packaging. 98

101 3.2 Abstract In order to study the role of a particular HIV-1 gene for various experiments, the gene or mutation should be cloned into a common HIV-1 backbone. Conventional ligation cloning technique can be applied in many organisms; but is challenging in HIV-1 due to the extreme heterogeneity in many of the viral gene products, leaving few conserved sites for ligation cloning strategies. Previously in our lab, a yeast homologous recombination/gap repair technique was developed, aiming to facilitate the cloning of diverse genes into an HIV-1 backbone and production of viruses in a mammalian packaging cell line using a replication bi-partite system. Up to now, this technique has been used to study multiple aspects of the primary virus including replication, fitness, and drug sensitivity to antiviral drugs. Two plasmids were co-transfected for virus production; one plasmid contained the near full length of HIV genome for insertion purposes, while a short complementing plasmid (expressing a short RNA containing R- U5-PBS-MA) was used for initiation of reverse transcription. However, the resulting virus from this system was minimally infectious compared to WT NL4-3 virus (<0.01% infectivity). During optimization of the infectivity, the length of the short complementing RNA was extended to a near full-length HIV-1 RNA genome, which caused a significant increase in infectivity. Packaging of two copies of genomic (g) RNA in the budding retroviral particle involves interaction between the RNA packaging signal (Ψ) in the 5 untranslated region (UTR) and nucleocapsid (NC) domain of Gag. Few studies have explored the coding 99

102 region for possible determinants of genomic RNA packaging. Knowing the fact that packaging of the short RNA is required for production of virus with wild type infectivity from the dual-transfection system, we hypothesized that the reduced infectivity was attributable to poor packaging of the short RNA, despite the fact that it harbored Ψ. We explored this possibility by utilizing truncation analysis of HIV-1 RNA, employing a modified version of dual complementing transfection system. Using RNAs with various truncated lengths of HIV-1 genome, revealed that high levels of grna packaging were dependent on a cis acting RNA element (termed the Genomic RNA Packaging Enhancer or GRPE element) found within the p1-p6 domain of the gag coding sequence. 100

103 3.3 Introduction Cloning specific genes from the primary virus is often necessary for several reasons, including the inability to isolate a primary virus as well as the importance of assessing the function of a single gene in the context of a whole genome. There have been multiple methods to clone a gene from a primary isolate into a common viral backbone in order to study its role in different steps of the viral lifecycle and virus-host interaction (21,113,134). The traditional cloning technique involving bacterial ligation requires unique and conserved restriction sites for cloning. Using these traditional cloning methods, it has been difficult to build pseudotyped viruses, especially in genetically heterogeneous viruses with diverse sequences like HIV-1, due to the absence of conserved restriction enzyme sites between isolates (165). One of the techniques to study the effect of specific HIV-1 genes involves the production of pseudotyped viruses via trans-gene expression from molecular clones on two or more separate plasmids (131). Usually, these techniques result in a virus that has the ability to infect cells but is unable to functionally replicate due to the absence of fulllength genomes. Pseudotyped viruses, although sufficient for some studies, are not suitable when studying most aspects of viral replication that involves multiple rounds of infection. Thus, production of replication-competent viruses is necessary for this task. The initial attempt at establishing a technique to create replication-competent chimeric viruses in our laboratory was by Marozsan et al (165), and was based on cloning an HIV-1 gene into a shuttle vector using yeast homologous recombination. This 101

104 shuttle vector had the ability to replicate in bacterial, yeast, and mammalian cells. The gene of interest could be digested out from this shuttle vector using restriction sites and be ligated into an HIV-1 backbone (pnl4-3) (74,165). Later, the yeast-based cloning/virus production system was enhanced in a way that any HIV-1 gene or other coding region from any HIV-1 subtype could be inserted into a common HIV-1 backbone without need of restriction digestion. The development of the second generation system was performed by Dudley et al. (74) in which the pnl4-3 vector was subdivided into two vectors: prec_nfl_hiv-1 (later called prec-nfl-3 LTR) and pcmv_cptl. This method has since been improved and is being used in our lab for whole genome recombination, production of multivalent env-based vaccine development, and evaluation of drug resistance. Retroviral RNA encapsidation is a highly ordered and precise process by which two full-length genomic (g) RNAs are incorporated into assembling virions. Unspliced HIV-1 RNA constitutes less than 1% of the total mrna in the cytoplasm of infected cell (24) yet is preferentially incorporated into new virus particles. Prior to packaging, intergenomic annealing initiates formation of loose non-covalent dimers of unspliced HIV-1 RNA, which is then selected by nucleocapsid (NCp7) protein for encapsidation (125). This selectivity of the loose grna dimer over the excess host cellular and viral spliced HIV-1 mrnas is due to the recognition of a cis-acting packaging element located in the 5 untranslated region by two zinc fingers of the NC (24,60,145,191,208). Prior studies have described specific RNA sequences and secondary structures adjacent to the canonical packaging elements (Ψ) with modest but significant effects on 102

105 grna encapsidation (220). As described in and shown in Fig. 8, Ψ element of HIV-1 comprises four stem-loops SL1-4 located downstream of the PBS and extends into the 5 end of gag (5,144). In particular, SL1 contains the dimerization initiation site (DIS), which forms the kissing loop for grna dimerization (208,230). SL3, a GGAG hairpin, is highly conserved among different strains of HIV-1 and is known as the core packaging signal. However, deletion of SL1 and SL3 has been associated with a defect but not elimination of HIV-1 grna packaging (51), suggesting importance of other RNA element(s) for the encapsidation process. In complex retroviruses like HIV-1, grna packaging and dimerization maps to multiple sequences in 5 LTR, 3 LTR, and the 5 end of gag gene, including the trans-activating-responsive (TAR) stem-loop and primer binding sequence (PBS) (48,50,167,168,212,218). Despite attempts to define the complete HIV-1 grna packaging signal, the relative role of each sequence has remained debatable and few studies have examined the role of sequences within the HIV-1 coding sequence. In this study, the HIV-1 g/mrna was split into two subgenomic (sg) RNA species: one containing the entire HIV-1 genome and coding sequence but lacking the R and U5 (5 LTR-nfl sgrna), and the other lacking the R and U3 (nfl-3 LTR sgrna). In this system, HIV-1 particles were assembled with the full complement and correct stoichmetry of viral proteins expressed from the nfl-3 LTR sgrna. Both HIV-1 sgrnas must be encapsidated for subsequent de novo infection and virus propagation. Since the 5 LTRnfl sgrna only acts as a template for reverse transcription, mutations/deletions in this RNA can be evaluated for effects on packaging independent of the underlying coding sequence. Significantly, the nfl-3 LTR sgrna acts as internal control for accurate, relative 103

106 measures of efficient grna encapsidation in the heterodiploid virus particle. Using this system, efficient grna encapsidation maps to two general regions. The first of these comprises the Ψ canonical RNA packaging elements in the 5 UTR, while the second, is located ~1200 nt downstream (i.e., at the 3 end of the gag coding sequence). This second element is termed the genomic RNA packaging enhancer element or GRPE element. 104

107 3.4 Results Development of the yeast-based HIV-1 cloning and virus production system. To address issues with HIV-1 high genetic diversity and lack of conserved restriction enzyme sites, a shuttle vector (prec_nfl_hiv-1) was generated for insertion of any gene of interest, using yeast homologous recombination. If a plasmid that contains both LTR regions at 3 and 5 ends of the HIV-1 genome is used for homologous recombination method of cloning, the intervening 8.5-kb coding region will be cut out of the plasmid due to homology between the two LTRs (Fig. 15A). To overcome this problem, the 5 LTR ( RU5) was deleted in the proviral DNA vector prec_nfl_hiv-1 (near full length). Additionally, a complementing vector was also generated to provide the missing region (R-U5) that is required for initiation of reverse transcription process. In order to prepare the prec_nfl_hiv-1, the near-full length or nfl genome of NL4-3 laboratory strain, which lacks the 5 LTR region, was inserted into a previously generated precenv using yeast homologous recombination, as explained and showed in section 2.2 and Fig. 15B (74). precenv contains the orotidine-5 -phosphate decarboxylase gene (URA3) for selection of homologous recombination, the yeast centromere sequence (CEN6) and autonomously replicating sequence (ARSH4) to maintain an episomal plasmid in yeast, and the beta-isopropylmalate dehydrogenase (LEU2) for maintenance of the plasmid in yeast in the absence of leucine. This vector also contains an ampicillin-resistance marker for selection in bacterial cells as well as a zeocin marker for maintenance in mammalian cells as an expression vector. 105

108 Next, a short RNA expression complementing vector (pcmv_cplt) containing R- U5-PBS-MA regions was produced to prepare the 5 LTR during transfection of both plasmids (Fig. 16). This plasmid expresses the NL4-3 5 LTR and portion of MA immediately following the cytomegalovirus (CMV) promoter, as described in 2.2 and showed in Fig. 16. RNAs expressed from both vectors contain the region that encompasses the Ψ signal Complementation between nfl_hiv-1 RNA and cplt RNA. The initiation of reverse transcription in HIV-1 requires cellular trna Lys 3 as a primer, which is packaged into the virus during its assembly. In this complementing system, trna Lys 3 is annealed to the PBS just downstream of U5 region in the cplt RNA and is used to prime the reverse transcriptase-catalyzed synthesis of minus-strand strong stop DNA, the first step in reverse transcription (Fig. 4). The (-) ss DNA then jumps to the R-U3 region of the nfl_hiv-1 RNA to complete minus-strand DNA synthesis, priming of (+) strand DNA synthesis, the second template switch, and finally, completing the double-stranded DNA genome. The first template switch involving the (-) ss DNA in our system is analogous to the intrastrand template switching model (197). In the follow up of the study by Dudley et al.(74), the prec_nfl_hiv-1 plasmid containing a gene of interest was transfected into 293T packaging cell line along with the pcmv_cplt plasmid to produce a long nfl_hiv-1 RNA and a short complementing cptl RNA [referred to as cplt (to denote complementing ) RNA], respectively (Fig. 18A). It was expected that both RNAs would be able to be packaged efficiently into budding virus 106

109 particles when all the viral proteins are translated from the nfl_hiv-1 RNA, since the two RNA fragments harbor the major packaging signal (Ψ). If both RNAs are competently incorporated, the ratio of heterodimeric virus to homodimeric virus should be directly related to the packaging efficiency of the short cplt (x) versus long nfl (y) HIV-1 RNA fragments (based on Hardy-Weinberg equilibrium X 2 +Y 2 +2XY). This means if both RNAs are encapsidated at equal efficiencies, half of the virus particles will be heterodiploid. Only the heterodiploid viruses that contain both the nfl_hiv-1 RNA and the cptl RNA are potentially infectious. This is because only the heterodimer viruses can complete the reverse transcription process following de novo entry into a host cell during infection, while the homodiploid viruses will infect but cannot complete reverse transcription due to lack of the U3-R or R-U5 sequences, and therefore die (depicted in Fig. 22). Hence, the generation of the fully replication-competent virus containing a full-length genome, in turn, is dependent on the validity of the intrastrand template switching model for retroviral reverse transcription (10,197,247), as seen in Fig. 18B Validation and optimization of the complementation system. To test the ability of the complementation system to produce a fully functional HIV-1 virus, the supernatants and cells were harvested to examine the production of RNA and proteins from prec_nfl_hiv-1 and/or pcmv_cplt vectors, 48 hours post transcription (Fig. 18). Quantitative reverse transcription PCR (q-rt-pcr) was utilized to quantify the levels of viral cplt RNA and nfl_hiv-1 RNA in the cells and supernatants (Fig. 18C). The RNAs were reverse-transcribed with specific primers and the cplt RNA was detected using primers 107

110 specific for detection of a cplt-specific tag sequence, while the nfl_hiv-1 RNA was detected by primers specific for the pol region. Fig. 18C shows that no 5 LTR (tag) RNA was detected in the supernatant and cells when the cells were only transfected with prec-nfl-hiv-1 or pnl4-3 (used as a positive control) and only RNA from the pol gene can be amplified above the background (set to 1x10 4 copies/µl DNA based on the mock transfection data). Similarly, when the pcmv_cplt was transfected without prec-nfl-hiv- 1, only tag was amplified out of the supernatants and cells, indicating specificity of the q- RT-PCR. When different ratios of pcmv_cplt and prec_nfl_hiv-1 were used for transfection, the difference in the levels of RNA expressed in the cells and packaged into virus particles was insignificant. This could be the result of using very high levels of plasmids in the transfection, resulting maximal levels of transcript expressions (even from the plasmid with the lowest concentration). Overall, both the cplt and nfl_hiv-1 RNAs were shown to be relatively available in the cells and supernatant, indicating that RNA was expressed off both plasmids and was also packaged into virions isolated from the transfection supernatants. In the next step, the production of p24 protein derived from prec_nfl_hiv-1 in the cells and virions were confirmed by western blots (Fig. 18D). Levels of p24 protein isolated from transfection cells reflected differences in the amount of prec_nfl_hiv-1 that was used in transfection. Levels of p24 protein isolated from virions in the supernatant derived from pnl4-3 (positive control) and prec-nfl-hiv-1 were similar, while p24 levels were slightly lower in cells transfected with prec-nfl-hiv-1 than with pnl4-3. Finally, reverse transcription (RT) activity was measured in viruses containing 108

111 supernatant of these transfections (Fig. 18E) as an indicator for virus maturation, as RT is relatively inactive in the absence of protease processing and in immature virus particles (22). The results revealed that again the RT from prec_nfl_hiv-1/pcmv_cplt and pnl4 3 were at the same level. Transfections with the lowest amount of prec_nfl_hiv-1 (0.1 μg) showed reduced RT activity in virus particles (Fig. 18E) showing a threshold of viral structural proteins produced from this plasmid required to yield infectious virus from the 293T packaging cell line [also shown by lack of virus produced by transfection of pcmv_cplt alone (Fig. 18C-E)] Testing the infectivity of viruses produced from bipartite system. In order to test the infectivity of viruses with the bipartite genome, the supernatant from 293T cells transfected with prec_nfl_hiv-1 and +/- pcmv_cplt and pnl4 3 were used to infect U87.CD4.CXCR4 cells, a human glioma cell line used for virus propagation (Fig. 18B). Upon de novo infection, q-rt-pcr was performed on RNAs isolated from infected cells and virus containing supernatant, indicating that both RNAs were available in cells and assembled viruses (Fig. 18D). However, the levels of cplt RNA found in the U87.CD4.CXCR4 cells and supernatants, which were most likely the remnants of the initial virus entry rather than new RNA transcripts, were much lower in comparison to nfl_hiv-1 RNA. Also, RT activities in the supernatant of U87.CD4.CXCR4 cells infected with viruses derived from dual transfection (prec_nfl_hiv-1 and pcmv_cplt) as well as pnl4-3 transfection (Fig. 18H) were above 200 cpm, which is indicative of a replicationcompetent virus. In contrast, as expected, viruses containing only the nfl_hiv-1 or cplt 109

112 genomic RNA were not infectious with RT activity levels below 200 cpm (Fig. 18 H, first set of bars). Gag precursor proteins and mature Gag p24 were again assessed by Western blot in cells and supernatant, respectively, 7 days post-infection (Fig. 18G) to verify production of intact HIV-1 proteins from this system. P24 protein was detected in all cell populations and viral supernatant from infections with viruses containing both the nfl_hiv-1 and cplt RNAs. (Fig. 18G). Gag precursor proteins were not produced in the cells infected with viruses harboring nfl_hiv-1 RNA or cplt RNA separately, and therefore, no p24 was incorporated into progeny viruses. This indicated that both the nfl_hiv-1 and cplt RNAs should be available for reverse transcription to occur and for the subsequent genomic DNA integration and transcription of viral proteins, including Gag-p

113 111

114 Figure 18. Complementation system for chimeric virus production. 293T cell were cotransfected with prec _nfl_hiv-1 and pcmv_cplt DNA vectors (A), and the assembled viruses were used to infect U87.CD4.CXCR4 cells (B). q-rt- PCR was used to measure cplt and nfl_hiv-1 RNA levels in cells transfected with different ratios of plasmid DNA and their virus-containing supernatant (C), or RNA obtained from U87.CD4.CXCR4 cells infected with 293T supernatants derived from different ratios of transfection of each vector and subsequent virus-containing supernatant (F). Primer and probe set that was used to target cplt RNA and nfl_hiv-1 RNA bind to the region in gag and pol sequences, respectively. Copy number per microliter is based on the real-time PCR amplification of serially diluted cplt and nfl_hiv-1 ( copies) cdna standards; copies under are considered background based on weak signals obtained from the mock-transfected cells. Cell and viral lysates from the 293T transfections (D) or U87.CD4.CXCR4 infections (G) were also subjected to western blot analyses; the text box below panels C and F also pertains to the lanes in D and G, respectively. HIV-1 Gag proteins were detected with a p24 antibody which detects Gag p55 precursor protein in the cell as well as p24- containing cleavage products in virus. To control for input of cell lysates, β-actin (D) or GAPDH (G) was probed using the appropriate antibodies. (E, H) Reverse transcriptase (RT) activity (y-axis is units 10 2 cpm/ml) was measured in cell-free supernatants 3 days post transfection of 293T cells (E) or 3, 5, 7, and 9 days following infection of U87.CD4.CXCR4 cells (H). (H) As positive control, RT levels of pnl4 3-infected cells were quantified on days 3, 4, and 5. Background was set at approximately cpm/ml based on the RT activity from mock transfections/infections. Adapted with permission from (74). 112

115 Optimizing the complementing system by increasing the length of complementing vector (pcmv_cplt). As the results in showed, co-transfection of prec_nfl_hiv-1 and pcmv_cplt at the appropriate ratio (1:1 to 1:3 ratios) resulted in the production of viruses capable of infecting target cells during the subsequent round of infection when the virus contains a heterodiploid genome (Fig. 18H). The infectivity of viruses resulting from this system was initially expected to be in the range of 2 fold less than pnl4-3, considering half of the viruses derived from transfection being heterodiploid. But the resulting virus was minimally infectious compared to WT NL4-3 virus (<0.01% infectivity) to the point that western blots and reverse transcription assays of the pnl4-3 control had to be performed in a separate experiment. This was because the pnl4-3 virus replicated quickly and destroyed the cells by day 7, where peak infection was seen for the dual transfection system (Fig. 18H). Thus, cells and viral supernatants were collected at day 5 for pnl4-3 virus and at day 7 for dual transfected viruses to preserve intact cells from which to extract RNA and protein. The lack of high infectivity continued to be observed when different HIV-1 genes (gag, PR-RT, RT, and env) from different primary isolates were cloned into prec_nfl_hiv-1 vector to produce chimeric viruses from co-transfection with pcmv_cplt (data not shown). One of the possible explanations for the lack of optimal infectivity in heterodiploid viruses resulting from this system was thought to be related to the lower packaging efficiency of the complementing vector (pcmv_cplt) which included a short genome. To assess this possibility, the length of complementing vector (pcmv_cplt) was extended incrementally to generate plasmids ending at various nucleotides in the 113

116 genome (Fig. 19). In order to generate these vectors, a plasmid (prec_full lenght_hiv-1) containing the whole HIV-1 genome was truncated from 3 terminus utilizing restriction enzyme digestion and ligation methods (Fig. 19A). The resulting plasmids were prec- 5 LTR/SbfI, prec-5 LTR/Bsu 36I, prec-5 LTR/NheI, prec-5 LTR/Bbvc, and prec- 5 LTR/XhoI from shortest to longest. The plasmids are named based on the enzyme that was used for digestion. These plasmids were then used as the complementing vector to transfect the 293T cells along with prec_nfl_hiv-1. Supernatant of transfection was harvested after two days and was used to infect U87.CD4.CXCR4 cells. RT activity was monitored from day 2 to day 11 post infection (Fig. 19B). We observed that the extension of HIV-1 RNA genome in the pcmv_cptl complementing vector (originally ended in the beginning of gag coding region) to a longer length beyond gag coding region (2845 nt) in prec-5 LTR/SbfI significantly increased the virus infectivity (Fig. 19B). Any further extension of the HIV-1 RNA genome did not result in increased infectivity in the relatively longer plasmids. Based on these findings and the result from Fig. 18D, showing lower packaging efficiency of cplt RNA compared to nfl_hiv-1 RNA in the infected cells, we suspected that there was an element in the gag coding region that corresponded to a step in viral replication including RNA packaging. 114

117 Figure 19. Generation of complementing vectors containing various lengths of HIV-1 genome and comparison of their infectivity. (A) prec_full-lenght_hiv-1 was digested by various enzymes to delete the HIV-1 genome from 3 end. An Xho I enzyme cuts the plasmid in two locations, one at nt 8902 in HIV-1 genome and the other one in the prec plasmid downstream of the HIV-1 genome. When Xho was used for digestion, along with Sbf I, Bsu 36I, Nhe I, Bbvc and just by itself, it cut fragments with the length of 2339, 5501, 6796, 7426, and 8433 bp respectively. When the cut sites were bluntended by klenow enzyme, the ends were ligated to produce prec-5 LTR/SbfI, prec- 5 LTR/Bsu 36I, prec-5 LTR/Nhe I, prec-5 LTR/Bbvc, and prec-5 LTR/XhoI truncated plasmids. (B) Increasing the genome length of complementing vector pcmv_hiv-1 from the original sequence, which ended in the beginning of gag coding region, to longer lengths in prec_5 LTR/SbfI with HIV-1 genome ended in 2845 nt and (contained the whole gag and part of pol gene), significantly increased the infectivity in day 3 to day 6 post infection. Adding the genome length in the longer plasmids--prec-5 LTR/Bsu 36I, prec-5 LTR/Nhe I, prec-5 LTR/Bbvc, and prec-5 LTR/XhoI--doesn t boost the infectivity from prec_5 LTR/Sbf. 115

118 3.4.2 Using yeast recombination/complementing system for RNA packaging studies. In the rest of this chapter, we will discuss how we modified the plasmids used in the bi-partite complementing system in order to investigate the existence of an RNA packaging element within the HIV-1 coding region. To start, we decided to construct more truncated vectors similar to what was done before, but with more focus on truncation in the gag region. It is worth noting that using the restriction enzyme digestion and bacterial legation method makes it impossible to truncate the plasmid of interest in specific regions (i.e., the end of sub-domains of the gag gene). This is because the truncations have to be chosen based on the availability of the enzymes that cut the plasmid once. As a result, we decided to construct a universal backbone vector, which allowed us to clone any region of interest without being limited by incomplete numbers of enzymes for prec full-length plasmids, which contain 16 kilo-bases (details of construction in 2.2.2). This universal plasmid, which was termed prec- HIV-1/URA3, is a yeast-based HIV-1 cloning plasmid that contains the factors (e.g., CEN6, ARSH, Leu2) necessary for yeast recombination. This plasmid can replicate in yeast, bacteria, and mammalian systems and can be used for any further cloning of HIV-1 genomes as well as in the bi-partite system for investigating the packaging efficiency of RNAs produced from these plasmids Generation of truncated prec-5 LTR vectors. By using yeast homologous recombination, multiple fragments containing incremental lengths of the HIV-1 genome were amplified from pnl4-3 clone, were cloned into the prec- HIV- 116

119 1/URA3 background, and replaced URA3 to produce truncated constructs (Fig. 20). From the various plasmids that were produced, eight were selected for this study. Depending on where in the coding region and at what nucleotide the HIV-1 genome ended, these plasmids were termed as prec-5 LTR-MA 1208, prec-5 LTR-CA 1415, prec-5 LTR-CA 1878, prec-5 LTR-NC 2086, prec-5 LTR-p6 2292, prec-5 LTR-PR 2549, and prec-5 LTR-Nef 8902 (Fig. 20 B, plasmids depicted in orange, see Table 2 for HXB2 genomic RNA numbering). Minimal CMV promoter expressed the 5 LTR sub-genomic (sg) RNA from these constructs, and a BGH poly(a) was inserted at the end of the genome for proper transcription termination. In prec-nef 8902, which was the longest plasmid, 3 segments of HVI-1 genome were deleted; thus, the CMV promoter expressed a sgrna starting from R (nt 456; HXB2 genome numbering) and ending prior to the U3 (nt 8902 in HXB2 DNA numbering or 8846 nt in HXB2 genomic RNA). The rest of the plasmids harbored more or less of the HIV-1 genome, but all lacked any 3 LTR sequence (Fig. 20B, plasmids depicted in orange). It should be noted that from here on, the term prec-nfl-3 LTR will be used instead of prec_nfl_hiv-1 because the former was considered to be a more appropriate name in publication for RNA packaging studies. As explained before, precnfl-3 LTR expresses sgrna, starting at the primer binding sequence (PBS) and ending with the 3 LTR (Fig. 20B and Fig. 21A, B, plasmid depicted in blue). Both 5 TR sgrnas and nfl-3 LTR sgrnas harbor the Ψ packaging elements for proper RNA packaging. HIV-1 mrna expression from the CMV promoter may be somewhat reduced from that observed from the HIV-1 U3 promoter, but the 5 LTR RNA still contains the TAR RNA and 117

120 is stimulated by Tat. In contrast, the nfl-3 LTR lacks the sequence involved in abortive transcription that needs to be rescued by Tat for elongation HIV-1 replication system involving bi-partite HIV-1 genomic RNA where only one contributes to HIV-1 coding sequence. Research on genomic RNA packaging during HIV-1 assembly has focused primarily on signature RNA sequences or secondary structures in the 5 UTR that bind Gag precursor proteins and mediate encapsidation (6,8,48,48,144,144,154,154,167,167). Mapping potential RNA packaging elements within the HIV-1 coding region is more challenging, considering confirmatory mutagenesis requires synonymous substitutions to alter an RNA structure/sequence without affecting the retrovirus proteome. Alternatively, many studies have employed genomic RNAs that lack regions of the retroviral open reading frames and that require helper vectors to express viral proteins for particle assembly (63,169,212). Thus, key genetic elements for RNA encapsidation could be absent in these viral RNAs. Furthermore, uncoupling of the retroviral genomic and messenger RNAs may affect both HIV-1 RNA trafficking and translation in the cell, leading to unbalanced stoichiometry of viral protein components, and may thus reduce yields of infectious virus particles. However, subdividing the HIV-1 genome into two sgrna fragments in our bipartite system provides a unique opportunity to introduce any modification on the RNA genome expressed from prec-5 LTR plasmids without concern for the underlying coding sequence since the proteins are delivered in trans from prec-5 LTR-nfl (Fig. 21). 293T 118

121 producer cells were co-transfected with both plasmids (Fig. 21A, B), and 48 hours post transfection, cells and supernatants were harvested for further experiments. If both the 5 LTR-nfl and nfl-3 LTR sgrnas are encapsidated at equal efficiencies, 50% of the virus particles will be heterodiploid for both sgrnas (based on Hardy-Weinberg equilibrium X 2 +Y 2 +2XY) (Fig. 21B and Fig 22A). Due to lack of the U3-R or R-U5 sequences, the homodiploid viruses with two copies of 5 LTR-nfl or two copies of nfl-3 LTR sgrnas are unable to complete reverse transcription following de novo entry into a host cell (depicted in Fig. 21 A-C). In contrast, infection with the heterodiploid virus leads to completion of reverse transcription, re-constitution of a full-length wild type (WT) genome, and proviral DNA integration (Fig. 21 A, D). The entire HIV-1 proteome originates from the nfl-3 LTR sgrna following de novo infection with heterodiploid virus, whereas the 5 LTR-nfl sgrna only serves as a template for trna Lys,3 binding and synthesis of (-) strand strong-stop DNA (Fig. 21D). Consequently, the prec-5 LTR-nfl vector provides a platform to introduce mutations within the HIV-1 coding region and determine the impact of these mutations on 5 LTR-nfl sgrna incorporation into virus particles (relative to the nfl-3 LTR sgrna). As described in this and the next chapter, we have introduced some large deletions, multiple point mutations, and insertions into the coding region of 5 LTR-nfl sgrna without impacting RNA packaging or infectivity, whereas other mutations have had significant effects. Although the elongating HIV-1 DNA could jump between the nfl- 3 LTR and 5 LTR-nfl sgrna templates during reverse transcription, our high level of 119

122 infectivity with or without deletions suggest that these recombination events occur at a relatively low frequency (estimated at 10%). Encapsidation of the sgrnas can be measured using quantitative reverse transcription/real time PCR (qrt-pcr). We can also measure the relative infectivity of virus (produced from the dual transfected 293T cells) in target cells. In this system, theoretically both sgrnas from prec-5 LTR-nfl and prec-nfl-3 LTR in transfected cells, should be able to serve as mrna templates for translation of HIV-1 structural proteins. However, when cells transfected with some of selected prec-5 LTR plasmids were probed for Gag protein using western blot analysis, the results showed that the truncated 5 LTR sgrnas do not produce truncated Gag precursor proteins. In contrast, nfl-3 LTR sgrna expressed the Gag protein efficiently when the plasmids were used in a separate transfection (Fig. 23) Locating novel cis-acting RNA elements in the HIV-1 coding sequence necessary for grna encapsidation. Based on these observations in , we surmised that a novel site within the coding sequence of the HIV-1 genome was necessary for efficient RNA packaging. To crudely map the region that contributes to grna packaging, truncated prec-5 LTR vectors were co-transfected with the prec-nfl- 3 LTR vector into 293T cells (Fig. 21A, B), and the packaging efficiency of 5 LTR sgrnas in assembled virions was examined. As described previously, the produced virus particles can harbor both or individual nfl-3 LTR sgrna and the truncated 5 LTR sgrnas (Fig. 21C). 48 hours post-transfection viruses and cells were harvested to be used for further 120

123 experiments. For all co-transfections, Gag protein expression in cells and capsid (CA) in viruses was monitored by western blot and ELISA (Fig. 21D, E); and RT activity was measured in the cell-free supernatant (Fig. 21E). Transfections with prec-nfl-3 LTR alone, or co-transfections of the prec-nfl-3 LTR with the various prec-5 LTR vectors produced similar amounts of virus based on RT activity (Fig. 21E). A stock of WT NL4-3 HIV-1 and virus derived from 293T transfections was serially diluted to measure RT activity and to infect U87-CD4-CXCR4 cells (Fig. 21 C and Fig. 24D). We have previously shown that RT activity of a wild type HIV-1 stock is a strong correlate of its infectious titer (166). Based on the linear regression equation from RT activity versus infectious titer of an NL4-3 virus stock (Figure 24B), we estimated that each virus from co-transfections would have a similar infectious titer as shown in Fig. 24E. However, the actual serial dilution/infection in U87-CD4-CXCR4 revealed a significant drop in infectious titer with viruses that should have contained the shorter 5 LTR (NC 2085, CA 1878, CA 1415, or MA 1208 ) sgrna compared to the longer 5 LTR (Nef 8902, RT 2845, PR 2549, and p ) sgrna (Figure 24F). Based on this deletion mapping, it appears that a 207 nt sequence within the p1- p6 coding region (between NC 2085 and p in the 5 LTR sgrna) was responsible for >350-fold increase in infectious titers. Even inclusion of another 6600 nt of sequential HIV-1 genome sequence on the 5 LTR sgrna (e.g., Nef 8902; Fig. 21A) failed to further increase infectivity (Fig. 24F). However, shortening the 5 LTR sgrna from NC 2085 to CA 1878 (Fig. 20A) resulted in an additional 10-fold loss of infectivity, i.e. >3000 fold less 121

124 infectious than the heterodiploid virus potentially containing the 5 LTR-p and nfl- 3 LTR sgrnas (Fig. 24F) The GRPE RNA element is necessary for infectivity via effects on genomic RNA packaging. Since reduced infectivity with the 5 LTR sgrna truncations was not due to disruption of virus particle formation or release, we suspected that grna encapsidation might have been compromised. First, we confirmed that 5 LTR and 3 LTR sgrnas were efficiently expressed in the cells by performing quantitative reverse transcription/real time PCR (qrt-pcr) with 5 and 3 LTR specific primers and probes. For the unspliced 5 LTR sgrna, we employed a primer for reverse transcription that binds to 5 end of gag (only found in unspliced mrna) followed by real time qpcr primers and probes specific for the U5 region. For the nfl-3 LTR sgrna, the RNA was reverse transcribed with a poly (dt) primer followed by q-pcr using primer/probe sets specific for U3 region (Fig. 26A). Cellular-associated viral RNA levels, normalized to 18s ribosomal RNA levels (also measured by qrt-pcr), are shown in Fig. 25. Amounts of cellular 5 LTR sgrna expressed from the transfected vector were not affected by the length of the RNA transcript or co-transfection with the prec-nfl-3 LTR vector (expressing the nfl-3 LTR sgrnas). It is important to note that these sgrna levels were measured from extracts of the same transfections used to produce the virus for infectivity assays in Fig. 24 and to measure viral sgrna content in Fig. 26. To ensure that viral RNAs were transported from the nucleus to the cytoplasm, cells were partitioned into cytoplasmic and nuclear fractions. Nuclear membrane integrity can fragment during fractionation and reduces purity of nuclear and 122

125 cytoplasmic RNA. Thus, the presence of only spliced β-actin mrna in the cytoplasmic fraction confirmed efficient separation from the nuclear fraction (Fig. 25A). Spliced β- actin mrna was absent in the cytoplasmic fraction but detected in total cellular RNA and nuclear fractions (data not shown). There were, however, abundant and similar levels of the unspliced 5 LTR sgrnas and nfl-3 LTR sgrna in the cell cytoplasm (Fig. 25C). HIV-1 Rev binding to the Rev-responsive element (RRE) in the env gene rescues unspliced and partially spliced HIV-1 RNA transcripts from the nucleus for transport to the cytoplasm. It is unlikely that Rev would increase transport of the 5 LTR sgrna transcripts because all but 5 LTR-Nef 8902 sgrna lack the RRE as well as all known splice acceptor sites. In mono-transfections with the prec-5 LTR vectors and in the absence of Rev (provided by the prec-nfl-3 LTR), we still detected high levels of unspliced HIV-1 RNA in cytoplasmic fractions (data not shown). The level of sgrnas in virus was measured to determine packaging efficiency of the different 5 LTR sgrnas (Fig. 26 B) relative to nfl-3 LTR sgrna (Fig. 26C). A dramatic shift was observed in 5 LTR sgrna packaging when this RNA was extended into the p6 region (3 end) of the gag coding sequence (i.e., 27 fold increase with 5 LTR-p sgrna packaging over 5 LTR-NC 2085 and 51 fold over 5 LTR-MA 1208 sgrna packaging) (Fig. 26B). However, extension of the HIV-1 RNA genome beyond nt 2292 (to near full length) did not augment grna packaging. Relative encapsidation of the different 5 LTR sgrna directly correlated with the level of virus infectivity (compare Fig. 26B and 24F). Up to now, the initial results suggest the existence of a packaging signal in the 123

126 sequences at the 3 end of gag. This region was therefore termed the genomic RNA packaging enhancer or GRPE element. Interestingly, the levels of efficient nfl-3 LTR sgrnas in virus particles remained relatively constant, despite reduced 5 LTR sgrna packaging (Fig. 26C). We had expected to observe a slight increase in nfl-3 LTR sgrna in the absence of 5 LTR sgrna packaging to compensate for reduced RNA packing within the virus particles. Instead, levels of nfl- 3 LTR sgrna were slightly reduced with poor encapsidation of the 5 LTR sgrna suggesting potential cooperativity between these two sgrnas during packaging. Furthermore, transfecting just the prec-nfl-3 LTR vector produced equivalent amounts of virus particles (based on RT activity; Fig. 21E) but harboring approximately 5-fold less nfl-3 LTR sgrna than that observed when the 5 LTR sgrnas were present. Such cooperativity, in packaging may reflect heterodimerization of the sgrnas. However, reduced nfl-3 LTR sgrna encapsidation in the absence of 5 LTR sgrnas may also be related to the packing of virus particles with host RNA species such as spliced viral RNA, 7SL and SRP cellular RNA (72,117,244) Recombination between two sgrna during reverse transcription of de novo infection. Although the elongating HIV-1 DNA during reverse transcription could jump between the nfl-3 LTR and 5 LTR-nfl sgrna templates and lead to dead virus production, our high level of infectivity with or without deletions suggest that these recombination events occur at a relatively low frequency (estimated at 10%) (13). As a control, the Gag AUG was deleted in some of the 5 LTR sgrnas (Fig.28) such that any 124

127 recombination would result in dead virus and yet, we did not observe an effect on RNA packaging (Fig. 27A) or infectivity (Fig. 27B) suggesting that this low frequency recombination was not affecting the primary results within our system. 125

128 Figure 20. Cartoon showing bi-partite plasmids. (A) HIV-1 DNA genome. (B) prec-5 LTR plasmids are shown in orange. The length of each plasmid is shown based on the map of pnl4-3 genome (A). prec-nfl-3 LTR is depicted in blue color. This plasmid was used as a complementing vector for co-transfection with prec-5 LTR plasmids and is lacking 5 LTR region. 126

129 Figure 21. Complementation system used for packaging studies and infectious virus production. (A) Series of prec-5 LTR HIV-1 vectors were constructed to express different lengths of HIV-1 sub-genomic (sg) RNA. The plasmid depicted in orange harbors the 5 LTR followed by various HIV-1 coding sequence. The prec-nfl-3 LTR (near full length or nfl) HIV-1 vector (in blue) lacks the 5 LTR, and is used to co-transfect 293T cells and complement the prec-5 LTR. (B) Both vectors express 5 capped, 3 polya HIV- 1 mrna species for the full complement HIV-1 proteins in the cells and produce virus particles indistinguishable from those derived for transfection with full length proviral DNA constructs (e.g. pnl4-3) (74). 293T transfectants produce three virus types (C) containing either two 5 LTR sgrnas (x), two nfl-3 LTR sgrnas (y), or one of each (heterodiploid or xy). As described in Fig. S2, the 5 LTR and nfl-3 LTR sgrnas in the heterodiploid virus can complement each other during reverse transcription to generate a WT, full-length proviral DNA where the coding sequence is only derived from the nfl- 3 LTR, while only the U5 and R uncoding regions are derived from the 5 LTR-nfl. At 48- hours post-transfection, virus produced from 293T cells transfected with prec-nfl-3 LTR and prec-5 LTR was monitored by Western blots using anti-p24 and anti-β actin (as control for cellular protein expression) in cell lysates or in virus-containing supernatants (D). (E) Virus production was also measured in supernatants by measuring RT activity using a radiolabelled assay or by quantifying CA p24 using an antigen capture assay. 127

130 128

131 Figure 22. Schematic of the reverse transcription following host cell entry with homoand heterodiploid viruses. (A) Viruses 1 and 2 are homodiploid containing only the 5 LTR or nfl-3 LTR sgrnas, while virus 3 represents a heterodiploid virus. (B) With virus 1, trna Lys,3 could bind to the PBS at the end of the nfl-3 LTR sgrna, but this binding is likely to be inefficient considering the extended interactions between the trna Lys,3 and sequences upstream and downstream of the PBS. There is no template for the synthesis of (-) strand DNA from the 3 terminus of trna Lys,3 annealed to the PBS. (C) In virus 2, trna Lys,3 can bind to the 5 LTR sgrna species at the PBS as well as through extended interactions. trna Lys,3 initiates the synthesis of (-) strand strong stop DNA catalyzed by HIV-1 RT. This (-) strand strong stop DNA can be identified (by PCR) for ~72 hrs in cells exposed to homodiploid virus 1 (data not shown). (D) In heterodiploid virus 3, reverse transcription is initiated by binding of trna Lys,3 to PBS of 5 LTR-nfl sgrnas. The RNA template is degraded by RT RNase H activity and frees the (-) strand strong stop DNA to anneal the nfl-3 LTR sgrna to complete (-) strand DNA synthesis. The process of reverse transcription then continues as outlined by the intrastrand model of retroviral reverse transcription. 129

132 Figure 23. The level of Gag precursor polyprotein measured in 293T cells transfected with truncated prec-5 LTR Plasmids. 293T cells were transfected with truncated prec- 5 LTR-MA 1208, NC 2085, p6 2292, RT 2845, Nef 8902, and prec-nfl-3 LTR plasmids. Then, 48-h post transfection, cells were lysed and monitored by western blotting using anti-p24. Anti-β actin was used to show cellular protein expression. The Gag 55KDa precursor polyprotein was not produced from prec-5 LTR plasmids with various lengths but was encoded from prec-nfl-3 LTR (right lane). 130

133 Figure 24. Comparing the expected and actual infectious titers of virus derived from co-transfected 293T cells. Wild type NL4-3 HIV-1 was serially diluted (1:4) and used to measure RT activity and to infect U87.CD4.CXCR4 (A). RT activity was plotted against the level of virus production for each dilution to determine a linear regression formula (B). There is a direct and significant correlation between TCID 50 values and the RT activity (166), indicating that the latter is a strong surrogate of infectious titer. Virus produced from the co-transfected 293T cells was again diluted to measure RT activity (C) or to infect U87.CD4.CXCR4 cells (D). Panel E provides the estimated infectious titers based on analyses from panels B and C. The actual TCID 50 values (F) were derived from the serial dilution/infections of the U87.CD4.CXCR4 cells (D) using the standard limiting protocol (166). The values are plotted at log 10 infectious units/ml 131

134 Figure 25. The levels of 5 LTR and nfl-3 LTR sgrna within the cell and cytoplasm following co-transfections. 293T cells were co-transfected with prec-nfl-3 LTR and the various prec-5 LTR vectors. Cellular, cytoplasmic, and nuclear RNA was fractionated as described in the Materials and Methods. As a control (A), unspliced and spliced β-actin RNA transcripts were RT-PCR amplified from the cytoplasmic fraction as previously described (98). The various 5 LTR sgrnas and nfl-3 LTR sgrna was qrt-pcr amplified from cellular RNA (B) or cytoplasmic RNA (C) extracts using the primer sets/probes described in Fig. 4A. Copy numbers of the 5 LTR and nfl-3 LTR sgrnas are presented in both panels B and C. 132

135 Figure 26. Relative packaging of the HIV-1 sub-genomic RNA in virus derived from cotransfected cells. (A) Schematic representation of primers and probes used to specifically PCR amplify and quantify the 5 LTR sgrna (orange) and nfl-3 LTR sgrna (blue) following cdna synthesis (See Table S2 for primer details). Viral RNA was extracted from cell-free supernatants of transfected cells. Copy number of 5 LTR sgrna (B) and nfl-3 LTR sgrna (C) were determined by qrt-pcr, compared to qrt-pcr amplification of in vitro transcribed HIV-1 RNA of known copy number (10 4 to copies), and presented as relative to the viral RT activity, establishing values relative to RT activity. 133

136 Figure 27. RNA packaging efficiency was not influenced by deletion of a Gag AUG start codon in the 5 LTR sgrnas. Gag AUG initiation codon was deleted in the prec-5 LTR- CA 1878, NC 2085, RT 2845, and Nef These plasmids were used to transfect 293T cells along with prec-nfl-3 LTR as described in the text and supplementary methods. Viral RNA was extracted from cell-free supernatants of transfected cells. Copy number of WT and AUG 5 LTR sgrnas were determined by qrt-pcr and normalized to the viral RT activity and presented in panel A. Virus infectivity was measured by first normalizing for RT activity, serially diluting as described in Fig. 3C, and then adding to U87.CD4.CXCR4 cells, i.e. a standard TCID 50 assay. The level of infectious virus is presented as log 10 infectious units/ml in panel B. 134

137 3.4 Discussion Studies starting in the late 1970s discovered that avian and murine retroviruses preferentially packaged full length viral RNA over the pool of cellular RNAs or spliced viral mrna (97,146,162,217,233,254). Ground breaking papers by Shank and Linial (151,224) were first to described a 300 bp region upstream of the gag-pol genes in avian sarcoma virus (ASV) necessary for the preferential packaging of the ASV RNA genome. After 30 years of research using more sophisticated molecular tools, this same region is still characterized as the essential packaging element for almost all retroviruses including spleen necrosis virus (254), Murine leukemia virus (Mann et al., 1983), and HIV-1 (144). Despite advances in molecular and structural analysis, the RNA packaging element is largely mapped to the 5 UTR. Only limited probing and analysis of genomic RNA sequences that extend into the retroviral coding sequence have been performed due in part to difficulties in modifying the RNA primary sequence without altering the HIV-1 proteome. Several years ago, our laboratory adopted a yeast-based HIV-1 cloning strategy for HIV-1. Development of this system required the deletion of the 5 LTR ( RU5) from the HIV-1 genome due to yeast recombination at repeat genetic elements (e.g. the LTRs), hence the construction of the prec_nfl_hiv-1. Transfection of this plasmid along with a complementing plasmid resulted in production of heterodiploid virus that was infectious upon de novo infection. A complete reverse transcription scheme in infected cells is expected to re-constitute the HIV-1 proviral DNA for production of fully 135

138 replication competent virus. However, when prec_nfl_hiv-1 construct was cotransfected with the pcmv_cplt complementary construct (expressing a short sgrna containing R-U5-PBS-MA) in a previous study, the resulting virus was minimally infectious compared to WT NL4-3 virus (<0.01% infectivity) (74). Although the absolute level of viral RNA was similar to that in WT HIV-1 particles, we determined that this reduced infectivity was attributable to poor packaging of this short cplt RNA compared to the nfl-3 LTR sgrna, despite the fact that both sgrnas harbored Ψ. We hypothesized that cplt RNA was not packaged efficiently because it lacked additional element(s) due to deletion of almost all the HIV-1 coding sequence. Therefore, we optimized this dual complementary system in order to investigate this possibility. In the new system, we used the prec_nfl_hiv-1 (was termed prec-nfl3 LTR in our study) for protein production which allowed us to use another plasmid named prec-5 LTR only for RNA expression and reverse transcription initiation. In this system, same as the previous one, the resulting virus was infectious upon de novo infection only if it contained both subgenomic RNAs (5 LTR-nfl and nfl-3 LTR sgrna) to complete the reverse transcription. We performed a series of RNA truncations within 5 LTR-nfl sgrna to identify and characterize a novel RNA packaging element at the 3 end of gag. We noticed that wild type RNA packaging and infectivity of virus produced from this dual transfection was rescued by extending the short cplt or 5 LTR sgrna to the 3 end of gag gene (1700 nt from the Ψ packaging element). Based on these observations, we discovered a novel RNA packaging site within the coding sequence of the HIV-1 genome. This element was termed the Genomic RNA Packaging Enhancer or GRPE element. Interestingly, we 136

139 noticed that this region also house a conserved cis-acting element known as ribosomal frameshift signal (RFS). Removing the GRPE, from 5 LTR-nfl sgrna did not impact virus production from transfected cells, but resulted in a >50 fold loss in packaging of 5 LTRnfl sgrna relative to nfl-3 LTR sgrna, paralleled by the loss in virus infectivity. Even with this two log decrease, residual virus infectivity levels suggests that the Ψ/5 UTR is required for low levels of grna packaging. The next chapter of this thesis proposal focuses on a thorough biochemical analysis to map the exact GRPE sequence/rna structure, to identify potential RNA (host or HIV-1) co-factors for its function (e.g. RNA binding proteins), and finally, to determine the possible functional relationship between the GRPE and RFS. 137

140 CHAPTER 4 THE NOVEL PACKAGING ENHANCER ELEMENT (GRPE) OVERLAPS WITH RIBOSOMAL FRAMESHIFT SIGNAL AND MEDIATES AN INTERPLAY BETWEEN GRNA PACKAGING AND FRAMESHIFTING Mastooreh Chamanian 1,3, Katarzyna J. Purzycka 2, Paul T. Wille 3, David McDonald 1, Stuart F.J. Le Grice 2, Eric J. Arts 1,3 * 1 Department of Molecular Biology and Microbiology, Case Western Reserve University, Cleveland, OH, 44106, USA 2 RT Biochemistry Section, HIV Drug Resistance Program, Frederick National Laboratory for Cancer Research, National Cancer Institute, Fredrick, MD, 21702, USA 3 Division of Infectious Diseases, Department of Medicine, Case Western Reserve University, Cleveland, OH, 44106, USA Figures 28, 29, 30, 32, 33, 34, and 35 are adapted with modification from Chamanian et al (A cis-acting element in retroviral genomic RNA links Gag-Pol ribosomal frameshifting to selective viral RNA encapsidation, Cell Host and Microbe, Feb 2013 (42)). 138

141 4.1 Preface Using truncated plasmids in a complementing system, we have identified a region within the 3 end of the HIV-1 gag coding region that contributes to RNA packaging and subsequently infection. In this chapter, we continued to investigate the role of GRPE in genome encapsidation in more detail by mutation and deletion analysis of this region. Previous work suggested that knowledge of RNA elements structure is required for understanding its role in biological processes. Therefore, we also analyzed the secondary structure of GRPE. Due to the tight relationship between RNA dimerization and packaging, the dimerization ability of sgrna was also tested in vitro. Moreover, we partially investigated the association of a cellular protein involved in ribosomal frameshifting with RNA packaging. The RNA synthesis and SHAPE experiments were performed by the author and Katarzyna Purzycka in the lab of Dr. Stuart F.J. Le Grice at the National Cancer Institute (NCI). All secondary and 3D structural analyses were performed by Katarzyna Purzycka. Dimerization assays were done and analyzed by Paul T. Wille. The GFP expressing plasmids for microscopy and full-length genome studies were kindly provided by David McDonald. Keith Olszens and David McDonald aided in photography. The remaining experiments and analyses were performed by the author. 139

142 4.2 Abstract Using a dual grna complementation system, high levels of grna packaging were shown to be dependent on a cis acting RNA element (GRPE) found within the p1-p6 domain of the gag coding sequence and overlapping with the ribosomal frameshift sequence (RFS). A series of deletions and mutations were undertaken in GRPE/RFS element of prec-5 LTR plasmids for further investigation of the role of GRPE in RNA encapsidation. The results showed that deletions or mutations that disrupted the two conserved GRPE stem-loops (slippery sequence and frameshift stimulatory stem), diminished grna packaging and infectivity >50 fold. This GRPE may function independent of genomic position considering efficient sgrna packaging was observed even with a deletion of intervening gag sequences between Ψ and the GRPE. Necessity of the two GRPE stem-loops for grna packaging was confirmed by RNA structure analyses using high-throughput SHAPE. Finally, partial downregulation of the release factor 1 (erf1) protein, known to bind near the GRPE and to stop Gag translation, resulted in production of defective virus particles containing up to 20-times more grna than wild type virus. Thus, encapsidation of grna may be specific for those HIV-1 RNAs employed for Gag-Pol translation following the ~5% ribosomal frameshifting event. These findings may explain the key control mechanism limiting two grnas per virion. 140

143 4.3 Introduction Several groups have studied the secondary and tertiary structure of retroviral grnas in the 5 UTR as well as proximal coding regions involved in grna dimerization, packaging, and translation (1,61,122,154). In addition to usage of its linear sequence for gene expression, RNA also folds into structures to play vital roles in different biological functions. RNA structure can control the access to the information stored in a linear RNA sequence or can act directly as regulatory elements. Like many natural RNAs, the function of the HIV-1 genome is tightly linked to its structure and to RNA-protein and RNA-RNA interactions for regulating viral replication. In the HIV-1 RNA genome, the canonical packaging element (Ψ) is located in the 5 UTR, and folds into four closely spaced stem loops (Fig. 8). During the process of RNA packaging, viral proteins and HIV-1 grna must co-localize in the host cell and move to the inner plasma membrane to form budding viral particles. The sequence of intracellular trafficking events of shuttling RNA:Gag complex to the viral assembly site where the genomic RNA is encapsidated is still not well understood. Confocal microscopy studies suggest that the unspliced genomic RNA is recognized and captured by Gag in the cytoplasm, followed by their transport to the inner plasma membrane (201). A single interaction with grna may nucleate Gag multimerization during assembly (34,140) or multiple Gag proteins may preassemble into oligomeric arrays in the cytoplasm prior to binding grna (237). Regardless of the known events during virus assembly, it is unclear why only one grna 141

144 dimer is bound to one of 2000 Gag dimers forming a new virus particles (68) rather than one grna dimer per Gag protein. Pseudotyping systems are commonly employed for the study of an RNA element, where viral proteins are expressed from HIV-1 mrna that cannot serve as grna and where the grna often lacks the majority of HIV-1 coding sequence and/or contains foreign sequence elements. As was explained in Chapter 3, here, the HIV-1 g/mrna was split into two subgenomic (sg) RNA species: one containing the entire HIV-1 genome and coding sequence but lacking the 5 LTR and was named prec-nfl-3 LTR, while the other one contained a group of RNA expressing vectors that lacked the 3 LTR and thus were termed prec-5 LTR. Both HIV-1 sgrnas must be encapsidated for subsequent de novo infection and virus propagation. In this system, prec-nfl-3 LTR is responsible for production of the complete viral proteins, while the 5 LTR-nfl sgrna only acts as a template for initiation of reverse transcription. This allows us to evaluate mutations/deletions for effects on packaging, independent of the underlying coding sequence. Significantly, the nfl-3 LTR sgrna acts as internal control for accurate, relative measures of efficient grna encapsidation in the heterodiploid virus particle. Using this system, a site at the 3 end of gag at the p1-p6 domains (located ~1200 nt downstream from Ψ) was shown to increase the grna encapsidation. This region was found to include both the GRPE element as well as the ribosomal frameshift signal. Like other retroviruses, HIV-1 utilizes the eukaryotic translational machinery for viral protein synthesis during infection of its target cells. Unspliced HIV-1 RNAs serve both as mrna for translation and as grna for encapsidation (28,60,132,155). Gag 142

145 proteins are translated from the first reading frame on unspliced HIV-1 RNA, and a -1 ribosomal frameshift occurs at a ~5% frequency leading to translation of the Gag-Pol precursor proteins (124,259). Frameshifting is dependent on two essential elements: a ribosomal stimulatory hairpin where the ribosome stalls and an upstream slippery polyu sequence, which facilitates and regulates ribosomal pausing and -1 nt shifting (30,74,76). Recently, Watts et al. (255) demonstrated that the polyu sequence is also found in a stable stem loop, termed Stem Loop (SL) P2. In most retroviruses (aside from lentiviruses), a similar region in their genomic RNA forms a pseudoknot that regulates ribosomal pausing and suppression of translation termination rather than a ribosomal shift (16,17,121,124). Recent studies suggest that a transient lowering of cellular ph may drive the infrequent assembly of this pseudoknot, where the ribosome may be stalled, resulting in suppression of translation termination (116). In this chapter we will also examine whether or not the GRPE element has a role in coordination of packaging versus translation. These findings may provide the first evidence as to why the ~2000 Gag proteins (per virus particle) only encapsidates two HIV-1 grnas instead of equal molar ratios of the abundant unspliced HIV-1 mrnas. 143

146 4.4 Results HIV-1 genomic RNA encapsidation is reduced with GRPE deletion. Based on our truncation analyses in , we mapped a putative packaging determinant designated the Genomic RNA Packaging Enhancer element (or GRPE) to a nt RNA sequence at the 3 end of gag gene. Since RNA structure is generally involved in post-transcriptional RNA activities, a linear GRPE sequence is difficult to map with truncation analyses and a conserved GRPE RNA structure may require surrounding RNA sequence. To confirm the role of the putative GRPE in genome encapsidation, two fragments containing GRPE and surrounding sequences were deleted from prec-5 LTR- RT 2845 plasmid to generate 5 LTR-RT and 5 LTR-RT sgrnas (Fig. 28A). Encapsidation of the 5 LTR-RT 2845 sgrna was reduced >20 fold with the two deletions of the putative GRPE (Fig. 28B), while the nfl-3 LTR sgrnas were encapsidated at similar levels (Fig. 33A). The reduced 5 LTR sgrna in the absence of the GRPE corresponded to a >80-fold decrease in virus infectivity (Fig. 28C). Based on 3 truncation and deletion analyses, we could map the GRPE to ~200 nt between 1956nt to 2188nt (RNA# 1500nt-1732nt) in the HIV-1 genome which also encompassed RFS. Again, the 5 LTR sgrna neither contributes appreciably to HIV-1 protein production (as shown in Fig. 23), nor impacts virus particle formation and release, but must be copackaged with the nfl-3 LTR sgrna for subsequent virus propagation. 144

147 In order to test the importance of GRPE in the full-length construct, we deleted a fragment containing GRPE from a pnl4-3 HIV-1 plasmid. However, investigating the impact of the GRPE deletion in the full length HIV-1 construct is problematic due to simultaneous deletion of the Gag open reading frame which prevents virus production. To compensate, Gag and Gag-Pol proteins were provided in trans from pcmv R8.91 helper plasmid co-transfected with WT and GRPE pnl4-3 HIV-1. pcmv R8.91 is a common HIV-1 Gag-Pol expression vector which is described in Fig. 28D. Virus from transfections were harvested, equalized for RT activity, and then lysed to measure grna content. The wild type pnl4-3 can express Gag-Pol but as a control, we examined the grna packaging from 293T cells transfected with pnl4-3 +/- pcmv R8.91. The results in Fig. 28E indicate that virus particles derived from the wild type (with or without the additional exogenous Gag-Pol) versus pnl4-3 GRPE contained 7-fold more grna (p<0.001). Thus we showed that deletion of the GRPE in the full-length virus leads to a similar defect in RNA packaging as observed in our bi-partite genome system. Due to the 3 -Gag/GRPE deletion, the virus derived from these 293T transfections were limited to a single round of infection and could not be propagated as with the bipartite genome system. 145

148 Figure 28. Effect of deleting the putative GRPE element on packaging of the 5 LTR sgrna and full-length grna. (A) The putative GRPE was deleted as a and nt region in the p1-p6 coding regions of gag within the prec-5 LTR-RT 2845 which was then co-transfected with prec-nfl-3 -LTR in 293T cells. (B) The grnas were quantified in virus particles by qrt-pcr, and presented as copies/ul. The levels of 5 LTR- RT , 5 LTR-RT sgrnas as compared to 5 LTR-RT 2845 sgrna. Virus infectivity was measured by first normalizing for RT activity, serially diluting as described previously, and then adding to U87.CD4.CXCR4 cells, i.e. a standard TCID 50 assay. The level of infectious virus is presented as log 10 infectious units/ml in panel C which shows a corresponding decrease in infectivity with GRPE deletion. (D) A similar GRPE region was also deleted from the WT pnl4-3 construct and co-transfected with the pcmv R8.91 vector. The grnas were quantified in virus particles by qrt-pcr, and presented as copies/ul in panel E for NL4-3 grna.. 146

149 4.4.2 Positional dependence of GRPE for grna encapsidation. To further map the GRPE, investigate its position dependence, and determine its minimal structure, four different intervening sequences between the GRPE and Ψ in the 5 LTR sgrnas (2292, 2270, and 2183 in p6) were deleted by yeast recombination/gap repair to generate 5 LTR-p6 868, 5 LTR-p6 716, 5 LTR-p6 706, and 5 LTR-p6 996 sgrnas. The extract deletion sites in the 5 LTR sgrnas are shown in Fig. 29A. The resulting RNAs lack most of the gag MA/CA/NCp7 coding sequence but retain different p1/p6 regions housing the putative GRPE. The packaging efficiencies were investigated with the same complementing system method as described above. Quantitation of the sgrnas in virus particles revealed that the 868 nt deletion within the gag coding sequence of 5 LTR sgrna (Figure 29A) maintained wild type grna packaging levels (i.e., only 1.8 fold less than 5 LTR_RT 2845 sgrna). Despite a similar sequence length, this 5 LTR-p6 868 sgrna was packaged 59-fold more than the 5 LTR_MA 1208 sgrna harboring only the Ψ signal (Fig. 29B). As expected, 5 LTR-p6 996 sgrna was not efficiently packaged, since it did not include a portion of the GRPE (Fig. 29B). Surprisingly, deleting a smaller region in 5 LTR-p6 706 and 5 LTR-p6 716 sgrna (Fig. 29A) significantly reduced sgrna packaging (Fig. 29B) despite retaining the entire linear GRPE sequence found in 5 LTR- 868 sgrna with wild type packaging efficiency. Again, the level of nfl-3 LTR sgrna encapsidation remained constant with all 5 LTR sgrna mutants (Fig. 33B). Reduced packaging of the 5 LTR sgrna with these internal deletions resulted in decreased infectivity of U87.CD4.CXCR4 cells (Fig. 29C), similar to that described with 147

150 the 3 truncations and the GRPE deletions. However, in the case of the 5 LTR-p6 868 sgrna, increased infectivity (10 3 IU/ml) was less than expected considering the high level of 5 LTR-p6 868 sgrna packaging (Fig. 29B) relative to the nfl-3 LTR sgrna levels (Fig. 33B) and relative to virus particle counts (based on RT activity; data not shown). As discussed later, intervening RNA sequences (also hidden in the coding sequence) may impact other sgrna functions aside from encapsidation, such as trna Lys,3 placement, or initiation of reverse transcription SHAPE technology for probing RNA secondary structure of GRPE. To better understand the behavior of these grna constructs, domains within and around the GRPE were structurally characterized using high-throughput selective 2' hydroxyl acylation analyzed by primer extension or hshape (255). This chemo-enzymatic probing strategy (Fig. 17) was developed based on the discovery that ribose 2'-hydroxyl group in RNA has a nucleophilic reactivity to certain chemical probes which is gated by local nucleotide flexibility (Fig. 17). SHAPE offers the benefit of using nontoxic agents, interrogation of all four nucleotides, and long nucleotide coverage from 300 to 600 nucleotides per experiment, which was the problem with classical probing systems. In this approach, RNA is exposed to a chemical probe such as 1-methyl-7-nitro-isatoic anhydride (1M7) and N-methylisatoic anhydride (NMIA), resulting in preferential acylation of the ribose 2 -OH moieties within single-stranded or conformationally flexible regions. In contrast, the regions that are constrained by base pairing are unreactive at nucleotides (179,180). The sites of chemical modification in RNA are 148

151 subsequently detected as stops during reverse transcription mediated primer extension (171,257). Next, the products are fractionated by automated capillary electrophoresis (CE). RNA secondary structures are predicted by introducing constraints derived from automated CE into ShapeFinder (249). ShapeFinder is transformation and analysis software that processes and converts the electropherograms into nucleotide reactivity tables. The normalized reactivity values are then imported into RNAstructure (v5.3) to be converted into pseudo-energy constraints that are incorporated into the RNA secondary structure folding algorithm (196,209). It has been shown that the RNA secondary structure obtained by combining SHAPE probing and folding algorithms significantly improves the structure prediction accuracy compared to structures obtained using either method alone (71,153). It should be noted that prediction of overall 5 UTR still relies on nucleotide protection, mutagenesis, and biochemical experiments since NMR technique is not capable to analyze a multi-hairpin RNA or a larger size RNA (larger than about 40-50nt). This made the SHAPE method the most suitable and accurate method for our purpose. Full-length 5 LTR sgrna constructs including--5 LTR-p6 868, 5 LTR-p6 716, 5 LTR-p6 706, and 5 LTR-p6 996 sgrnas were synthetized in vitro and used for structural probing. Even though the full-length in vitro transcribed 5 LTR sgrna constructs were used for structural probing, acylation profiles presented here are restricted to regions involved in RNA packaging (Fig. 30). Fig. 30A shows the secondary structure analyses of GRPE for 5 LTR-p6 868 sgrna using SHAPE. This structure reveals a wild type structure of GRPE/RFS that closely resembles the SHAPE-derived RNA 149

152 structure obtained using grna isolated from HIV-1 particles (255). This structure is characteristic of a type C three-way junction (143) connecting a stem and two stem loops (P2 and P3). P2 stem loop (P2SL) contains the slippery sequence flanked at the 3 end by the continuous, P2 stem loop (P3SL), which is essential for ribosomal frameshifting (76) (Fig. 12 and Fig. 30).. The 3D structure prediction of wild type GRPE/RFs was obtained (Fig. 30B) by employing the RNAComposer server that uses the RNA FRABASE dictionary. This newly developed method is parallel to computational linguistics which provide a fully automated prediction of RNA 3D structure from secondary structures based on the concept of translation machinery as explained in 2.9. For the 5 LTR-p6 707 and 5 LTR-p6 716 sgrna which showed low packaging efficiency, the 3 ends were truncated to nt 2270 and 2183 (Fig. 29A), respectively, suggesting these sequences may stabilize an RNA structure. Probing of these two protein-free sgrnas showed substantial destabilization of P2SL (Fig. 30C, D). Particularly, the 5 LTR-p6 707 and 5 LTR-p6 716 sgrna showed high 1M7 reactivity within the residues forming the P2SL (as compared with wild type HIV-1 grna or 5 LTR- 868 sgrna). Moreover, nucleotides show enhanced reactivity compared to 5 LTR- 868 sgrna, suggesting they no longer base pair with the segment spanning nts (Fig. 30C, D). In contrast, nt , which were reactive in the 5 LTRp6 868 sgrna were unreactive in these sgrna and are instead predicted to be involved in forming an extended, discontinuous P3 hairpin with a 3nt bulge. These data suggest 150

153 that although the GRPE may function independently of sequence context, both the P2SL and P3SL must be maintained for enhanced packaging activity. As a final control, we used 5 LTR-p6 996 sgrna in which 996 nt of the gag sequence ( ) was deleted to remove the linear sequence encoding P2SL while retaining part of the sequence involved in the formation of P3SL domain. SHAPE analyses predicted an extended, stable P3SL, similar to that assumed within 5 LTRp6 716 and 5 LTR-p6 707 sgrnas (Fig. 30E). 151

154 Figure 29. Effect of deleting the region separating GRPE and Ψ on packaging of the 5 LTR sgrna. (A) Four regions were deleted separating the putative GRPE from Ψ. The nucleotides that were deleted are shown below each plasmid schematic. All deletions start from nt 1091 and end at different sites upstream of GRPE in plasmids that end at p6 coding region. (B) These four plasmids and prec-5 LTR-RT 2845 were co-transfected with prec-nfl-3 -LTR in 293T cells. The resulting RNAs from deleted plasmids are 5 LTRp6 868, 5 LTR-p6 716, 5 LTR-p6 706, 5 LTR-p6 996 sgrnas. The grnas were quantified in virus particles by qrt-pcr and presented as copies/ul for the 5 LTR sgrnas and for the nfl-3 LTR sgrna in Fig. 33B. (C) Virus infectivity was measured by first normalizing for RT activity, serially diluting as described in Fig. 3C, and then adding to U87.CD4.CXCR4 cells, i.e. a standard TCID 50 assay. The level of infectious virus is presented as log 10 infectious units/ml in panel. 152

155 153

156 Figure 30. SHAPE analysis of GRPE/RFS element in 5 LTR sgrna with deletions of the region separating GRPE and Ψ. In vitro transcribed 5 LTR sgrna were modified with1m7, reverse transcribed using fluorophore-labaled probes and separated by capillary electrophoresis. Reactivity at each nucleotide was calculated by subtraction of the reverse transcription of unmodified RNA from that of 1M7 modified RNA, using Shapefinder software. (A) Secondary structure of GRPE in wild type RNA 5 LTR-p6 868 (B) 3D structure of GRPE in wild type RNA and 5 LTR-p Secondary structure of GRPE in (C) 5 LTR-p6 716 (C) 5 LTR-p6 706 (D) 5 LTR-p LTR-p6 868 sgrna contains a GRPE/RFS structure similar to the RFS of the SHAPE-derived RNA structure obtained using grna isolated from HIV-1 particles and is shown in Fig. 17. (255). SHAPE data provided high 1M7 reactivity for residues forming the P2SL showed in 5 LTRp6 706 and 5 LTR-p6 716 sgrna. A base pairing of that the nucleotides and that existed in wild type P2SL in opened in these sgrnas due to high reactivity. In contrast, the reactivity of nt reduced in 5 LTR-p6 706 and 5 LTR-p6 716 sgrnas as compared to wild type and 5 LTR-p6 868 sgrnas. These nucleotides are instead predicted to be involved in forming an extended, discontinuous P3 hairpin with a 3 nt bulge. Only the sequence involved in the formation of P3SL domain of 5 LTR-p6 996 sgrna was analyzed by SHAPE since P2SL is deleted. A stable P3SL, similar to what was assumed within 5 LTR-p6 716 and 5 LTR-p6 707 sgrnas was predicted. Nucleotides are colored according to their SHAPE reactivities using the scale. Nts numbers are based on HXB2 DNA numbering and are shown with a dash in the structures. 154

157 4.4.4 Effect of GRPE secondary structure on grna encapsidation, dimerization, and virus infectivity. Over the years, it has becoming increasingly clear that retroviral RNA packaging and dimerization, regardless of the primary sequence, is dependent on RNA structural motifs of the packaging determinants and their interaction(s) at the structural level (RNA-RNA interactions during dimerization and RNA-protein interactions during packaging) [reviewed in (60,175)]. To further probe the impact of the P2 and P3 stem-loops in RNA packaging, we performed site-directed mutagenesis based on Mfold RNA structure predictions. Mutations were introduced (Fig. 31A) in the GRPE element of 5 LTR-RT 2845 sgrna (Fig. 31B) in order to change the primary sequence, but retain the stability and general structure of both P2SL and P3SL (mu1, mu2, and mu4 in Fig. 31 and 32A). Other mutations in the 5 LTR-RT 2845 sgrna dramatically changed the overall structure of GRPE (mu3 and mu5 in Fig. 31 and 32 A). Similar to the dual complementation system that was used previously, the prec- 5 LTR-RT 2845 and plasmids prec-5 LTR-RT-mu (1-5) plasmids were co-transfected in 293T cell along with prec-nfl-3 LTR plasmid. RNA packaging efficiency of the 5 LTR RNAs expressed from these plasmids were determined by qrt-pcr of RNA content in the assembled virions. The results in Fig. 32B show that mutations (Fig. 31 and 32 A) causing minor rearrangement in the general structure of both P2 and P3SL (mu1, mu2, and mu4) had minimal effect on 5 LTR sgrna encapsidation (1.1, 2.2 and 3.2 fold respectively) and only slight loss in infectivity (Fig. 32C). In contrast, when M-fold predicted significant changes to P2 and P3SL via 4 nt and 3 nt substitutions in mu3 and mu5 sgrnas, respectively, (Fig. 31 and 32A), 13.5 and 28-fold reductions were observed in packaging 155

158 which corresponded to a 560- and 1400-fold loss of infectivity, respectively (Fig. 32B, C). Similar to previous experiments, the levels of nfl-3 LTR sgrna packaging remained constant with all the mu 5 LTR sgrnas (Fig. 33C). These studies provided the strongest evidence for the importance of the P2 and P3SL for sgrna packaging. Additionally, 3D prediction was used, as described in 4.4.4, to predict the effect of these mutations on the tertiary structure of GRPE (Fig. 32) in vitro dimerization assay to determine the impact of GRPE on sgrna dimerization. Dimerization is described as an early event during grna capture/transport and linked to the 5 UTR (51,126,193,230). To determine the impact of GRPE deletion on grna dimerization, we performed in vitro RNA dimerization experiments using 3 LTR and 5 LTR RNAs +/- GRPE (all containing the DIS) (Fig. 34). We observed efficient in vitro dimerization of all sgrnas (either 5 or 3 LTR containing RNAs) regardless of the presence or absence of the GRPE (Fig. 34). In fact, the shortest 5 LTR RNA (MA-1208) dimerized with high efficiency (>70% in loose dimers) but could not be packaged efficiently (>51-fold decrease). Oligonucleotides annealing to DIS could disrupt dimer formation (Fig. 34B). However, these in vitro findings may or may not reflect HIV-1 grna dimerization in the cell during encapsidation Exploring the relationship between HIV-1 mrna translation and grna packaging. The GRPE overlaps with the sequences and/or structures involved in ribosomal frameshifting, but more studies are needed to better understand the factors 156

159 involved in linkage of translation with processes involved in grna packaging. The translation termination factor, erf1 binds to the termination codon, stops translation via interactions with the ribosome, and helps to initiate the nonsense RNA decay (NMD) pathway (186). Previous studies have indicated that sirna knockdown of erf1 increases Gag-Pol synthesis relative to Gag, suggesting increased ribosomal frameshifting and loss of infectivity (137). To explore the possible role of erf1 and ribosomal frameshifting on grna packaging, we downregulated cellular levels of erf1 with sirna (Fig. 35A), but failed to observe an increase in Gag-Pol in virus particles (data not shown). Next, we measured levels of 5 LTR sgrnas (by qrt-pcr) relative to RT activity (or relative to p24 content of the virus). When erf1 was depleted with sirna knockdown, the amount of 5 LTR sgrna increased >10-fold as compared to virus produced in the presence of scrambled sirna (Fig. 35B). Based on the assumption of 2000 Gag dimers per wild type virus particles, knockdown of erf1 appeared to have produced virus particles containing genomic RNAs. We performed the same erf1 knockdown experiment with HIV molecular clone (NL4-3 based) that carried a Gag-interdomain green fluorescent protein (igfp) (120). HIV Gag-iGFP was transfected into 293T cells to produce virus harboring a Gag-GFP fusion. Reductions in erf1 (Fig. 35C) appeared to result in virus particles with more grna per Gag molecule (Fig. 35D) as well as brighter fluorescence (compare panel E and F in Fig. 35) suggesting enlarged virus particles containing more of Gag, Gag-Pol, and grna. 157

160 Figure 31. Prediction of GRPE 3D strcuture of in the 5 LTR-RT-mutant sgrnas. (A) Multiple point mutations were introduced to the predicted P2 and P3 stem-loops within the GRPE/RFS sequence. (B) The map of prec-5 LTR-RT 2845 vector. (C) The RNA secondary and 3D structures in this GRPE/RFS region of these five mutant 5 LTR sgrna templates were predicted by M-fold ( and RNAComposer server, respectively. 158

161 Figure 32. The role of RNA secondary structure in the GRPE/RFS region on HIV-1 RNA packaging. (A) The RNA secondary structures in this GRPE/RFS region of these five mutant 5 LTR sgrna templates as was shown in Fig. 31C. The five mutant 5 LTR sgrna templates were expressed in co-transfected 293T cells. Supernatants from these transfected cells were harvested to measure 5 LTR sgrna (B) and nfl-3 LTR sgrna levels (Fig. 33C) by q-rt-pcr. The same virus-containing supernatant was then equalized for RT activity, serially diluted and then used to infect U87.CD4.CXCR4 cells. The relative infectivity, as measured by TCID 50 assay, is presented as Log 10 infectious units/ml (C). Data in (B) and (C) are presented as mean ± SEM. 159

162 Figure 33. The levels of nfl-3 LTR sgrna were measured by q-rt-pcr. (A) Total viral RNA were isolate form supernatants of co-trasnected 293T cells. Q-RT-PCR using probe U3 wsa performed to measure the levels of nfl-3 LTR sgrna which was used as positive control to transfect the 293T cells. The levels of 5 LTR sgrna from the same virus-containing supernatant are shown in Fig. 28B. Supernatants from these transfected cells were harvested to measure nfl-3 LTR sgrna levels by q-rt-pcr. (B-C) Similar to panel A, the levels of nfl-3 LTR sgrna in panel B and C are associated with expreminets in Fig. 29 and Fig. 32, respectively. Overall, the levels of nfl-3 LTR sgrna reamined relatively contant for all the experiments. and the RNA level in virus particle is shown as the last bar. Data are presented as mean ± SEM. 160

163 Figure 34. In vitro dimerization of the 5 LTR and nfl-3 LTR sgrna. 5 LTR RNAs, specifically 5 LTR-MA 1280, 5 LTR-RT 2845, 5 LTR- RT-mu5, 5 LTR- RT , 5 LTR- RT , and 5 LTR-p6 868 and a shorter version of the nfl-3 LTR sgrna (starting with nt 634 and extending up to 2292 nt; 3 LTR-P ) were produced by in vitro transcription. For homodimerization, 200 nm of RNA denatured was placed on ice, incubated at 37 C for 30 min at room temperature in Loose Dimer buffer as described in the methods below (229). For denaturation curves, aliquots of the dimers were incubated at the indicated temperatures for 10 min and then placed back on ice until electrophoresis on 0.8% high resolution agarose gels. The black labeled sgrnas contain the GRPE while the GRPE is deleted in the red labeled sgrnas. All sgrnas contain the Ψ and DIS. Panel A is two examples of dimers derived from p6 868 and MA LTR RNAs and then incubated at increasing temperature prior to electrophoresis and imaging (AlphaImager EP, Alpha Innotech). (B) The same conditions for loose dimer formation (see above and in the methods below) were performed using 5 LTR-MA 1280 sgrna and oligonucleotides that are sense or antisense to the dimerization initiation sequence (DIS). Image of the electrophoresed products show the denatured monomer, dimer, and folder monomer sgrna. All of the gels containing the dimer melting analyses (e.g. panel A) for each sgrna were analyzed and band quantified using AlphaView (Alpha Innotech) software. The percent dimer versus temperature for each RNA was plotted in panel C. 161

164 Figure 35. The effect of erf1 knockdown on virus production and grna encapsidation. A pool of three sirnas against erf1 (see the supplementary experimental procedures) effectively reduces cellular erf1 expresion level by ~70% prior to 293T transfections as indicated by the Western blot analysis (A and C). 293T cells were co-transfected with the prec-nfl-3 LTR along with either prec-5 LTR-MA 1208, prec-5 LTR-p6 868 and prec- 5 LTR-RT 2845 or transfected with HIV Gag-iGFP. It is important to note that erf1 knockdown reduced all protein levels in the cell 2-fold within 72 h. Virus produced from these sirna-treated, co-transfected cells were equalized for RT activity and then used to infect U87.CD4.CXCR4 cells. When derived from cells with erf1 knockdown, virus was non-infectious, even at the highest dilution (not shown). Viral RNA was extracted from the cell free supernatant and the 5 LTR sgrna or NL4-3 grna was then measured qrt- PCR and presented as relative to p24 antigen content (B and D). In the presence of wild type levels of erf1 (SD sirna treatment), the virus produced from the co-transfected cells harbored approximately LTR sgrna copies per 2000 molecule of p24 (or the estimated size of one HIV-1 particle). With the erf1 knockdown and the WT prec-5 LTR- RT 2845, there is approximately 19 5 LTR sgrna copies per 2000 p24 molecules. Virus derived from transfections with the HIV Gag-iGFP molecular clone (erf1 knockdown E; WT erf1 - F) was sucrose-cushion purified, spread on poly-l-lysine coverslips, and images captured on a Deltavision RT epifluorescent microscope system. 162

165 4.5 Discussion Using the dual transfection system, we determined that a novel site aside from the Ψ signal, within the 3 end of gag coding region of the HIV-1 genome was necessary for efficient RNA packaging. By performing a series of RNA truncations, deletions, and point mutations within 5 LTR-nfl sgrna, the sequence for GRPE was narrowed down to a region between 2015 and 2175 nt which also contains the ribosomal frameshift signal (RFS). We assumed that this element had been missed due to its importance for the ribosomal -1 frameshifting function that is mediated by the overlapping RFS. In our system, the nfl-3 LTR sgrna lacks the TAR hairpin and poly(a) loops, which have shown to have modest effects on grna encapsidation (48,167,212,218). Vrolijk et al. proposes that these elements (located upstream of SD) must work in coordination with the major packaging signal, found only in the unspliced HIV-1 RNA (251). We examined the GRPE effect in the context of a full length HIV-1 genome (with both LTRs intact) and found that the deletion of GRPE still reduced grna packaging (Fig. 28E). The GPRE deletion withun the bipartite genome system resulted in 14-fold reduction in grna packaging whereas a 7-fold reduction was observed with the deletion of GRPE in WT NL4-3. The slightly reduced effect of the GRPE with an intact 5 LTR suggests a modest impact of the TAR on grna packaging: ~2-fold as reported (Clavel and Orenstein, 1990; McBride and Panganiban, 1996; Russell et al., 2003). Additionally, the RNA secondary structures of GRPE in WT RNA and in RNAs containing internal deletions were analyzed by high throughput SHAPE technique. These 163

166 internal deletions resulted in movement of GRPE proximal to 5 UTR. We showed that efficient grna packaging was maintained with point mutations or deletions that had minor effects on the stability of P2 and P3SL. These observations suggest the dependence of efficient RNA packaging on the GRPE motif at the structural level. Aside from the newly described role of two stem-loops of GRPE in grna packaging, these structures were also shown to be necessary for ribosomal frameshifting, a mechanism conserved among all lentiviruses. We are currently exploring in what way the GRPE may factor into this regulation, and promote grna packaging. Our preliminary in vitro studies suggest that the GRPE does not impact RNA dimerization of either the 5 or 3 LTR RNAs whereas disruption of the DIS does prevent dimerization as previously described (230). Although loose RNA dimer formation may have some relation to the in vivo event, there are obviously cellular and viral factors that are absent in vitro (e.g. NC region of Gag) that are required for maturation of tight dimers and encapsidation (126,193). As described herein, we show that the GRPE cannot act alone for grna packaging but likely involves the Ψ and related 5 UTR sequences. Overlap between the GRPE and the RFS suggested a possible link between translation and grna packaging. In relation to translation termination, erf1 mediates peptidyl-trna hydrolysis at the peptidyl transferase center of the ribosome to terminate translation. erf1 also nucleates a host protein complex involved in 3 UTR sensing and the NMD pathway (114). In 2010, Kobayashi et al. identified erf1 from a genome-wide sirna knockdown screen as a putative candidate for interactions with ribosomal 164

167 frameshift site (137). In their studies, erf1 depletion led to increased ratio of Gag- Pol:Gag in HIV-infected cells and production of non-infectious virus. As a preliminary study, we reduced erf1 expression levels with sirnas in cells subsequently transfected with our wild type and GRPE mutant HIV-1 constructs. Reducing erf1 levels increased encapsidation of HIV-1 grna relative to CA p24 levels in the cell-free supernatant. These virus-like particles may contain >20 times more grna than wild type virus based on an assumption of 2000 Gag dimers per virus particle. When using a virus containing GagiGFP, knockdown of erf1 appeared to mediate production of large virus particles with more Gag (brighter virus particles) and more grna. Based on these preliminary results and positioning of the GRPE, we propose that ribosomal frameshifting may prevent the NMD pathway and designate this unspliced HIV-1 mrna (destined for translation of Gag-Pol) as grna for encapsidation. Thus, even slight increases in ribosomal frameshifting and Gag-Pol synthesis could lead to dramatic increases in grna levels and release of noninfectious virus particles. A model is proposed in Chapter 6 for the interplay between RNA packaging and ribosomal frameshifting. Although the RNA packaging element is often described as a region with conserved RNA hairpin structures, structural studies on this region involving combination of bioinformatics, enzymatic probing, phylogenetic and intra-molecular UVcrosslinking approaches have revealed a structural transition between branched multiple hairpin (BMH) and Long distance interaction (LDI) alternative conformations. A more recent study using nuclear magnetic resonance (NMR) proposed a similar riboswitch that coordinates the dimerization and translation.they showed a similar 165

168 structure for BMH conformation in which the AUG start codon base pairs with the sequences in U5 to promote RNA packaging. However, they proposed an alternative and different LDI structure in which the DIS interacts with U5 (instead of forming a long stem in the original LDI form) to inhibit the dimerization and promote the translation stem (154). Nonetheless, the reason that only one RNA dimer is incorporated per Gag molecules in each virus despite the presence of abundant viral unspliced RNA in the cytoplasm has remained relatively undefined (68). The overlap of GRPE and the RFS suggest a possible link between translation and grna packaging. Based on our model (explained in detail in Chapter 6), we propose that the selection of unspliced RNA for encapsidation is an extremely organized and tightly regulated process, which in part is organized by the GRPE/RFS located at the 3 end of gag. Our model proposes a control of HIV-1 translation, ribosomal frameshifting, and grna packaging by the GRPE/RFS. We show that the GRPE cannot act alone for grna packaging but likely involves the Ψ and the flanking 5 UTR sequences. We suggest that P2 and P3SL of the GRPE can interact with the 5 UTR or may form a pseudoknot, as observed with other retroviruses (116). Studies are underway to probe for higher order RNA structures that may interact with viral or host proteins for the regulation of optimal grna packaging. 166

169 CHAPTER 5 THE LENTIVIRUS TRANSDUCTION EFFICIENCY IS ENHANCED BY INCORPORATION OF GRPE ELEMENT IN THE LENTIVIRAL VECTORS Mastooreh Chamanian 1,2, Paul T. Wille 2, Eric J. Arts 1,2 1 Department of Molecular Biology and Microbiology 2 Division of Infectious Diseases, Department of Medicine Case Western Reserve University, Cleveland, OH, 44106, USA *To be submitted for publication after additional experiments are performed. 167

170 5.1 Preface Studies performed in Chapters 5 describe improvement of lentiviral vectors by enhancing packaging of the transfer vector RNA from the transfected produced cell line. It involves improvement of packaging efficiency by addition of Genomic RNA packaging Enhancer Element (GPRE) derived from HIV-1 NL4-3 into a lentivector that has been used for studies related to murine bone marrow cell transduction. The plv-mnd-palag lentivector and K562 cells were kindly provided by Yuan Lin, a former graduate student in Stanton Gerson s lab. The work presented here was performed by the author. This is an ongoing study in our lab for assessing the application of GRPE in other lentivectors, following the identification of this element. In this chapter, we will be presenting the initial results from this study. The following experiments concerning the application of this element in other lentiviruses. i.e. SIV, IAEV and HIV-1 as well as commercially available HIV-1 and FIV lentivectors are currently being performed by Paul Wille. 168

171 5.2. Abstract Lentiviral vectors have been increasingly used in both nonclinical and preclinical research. They are capable of introducing therapeutic genes into a broad range of mammalian cell types, including non-dividing cells, with stable expression. Lentivirus grna packaging has been the subject of intense investigation owing to the impact it may have on different aspects of HIV-1 replication/pathogenesis, development of new drug targets, and enhancement of lentiviral gene delivery or vaccine efficacy. Still, with this system, many improvements are needed to be able to address major safety concerns regarding the prevention of replication-competent lentivirus (RCL), insertional mutagenesis, and vector mobilization before initiating large clinical trials. One of the factors that increases the theoretical risk for insertional mutagenesis in a gene transfer system is using a relatively high multiplicity of infection (MOI) which has been used in the very few initial clinical trials to raise clinical efficacy and gene expression. Due to the risk of insertional recombination, multiple plasmids are used in gene delivery system in which the unnecessary cis-acting sequences have been removed in order to minimize the viral gene expression. As a result, the currently used lentivectors lack the GRPE element that was identified in Chapters 3 and 4. Here, we inserted the GRPE into the HIV- based lentiviral vectors and achieved enhancement of RNA packaging and subsequent transduction efficiency by employment of the GRPE. These findings have important implications for the improvement of lentivirus-based gene delivery systems. 169

172 5.3. Introduction Retroviruses have been preferentially used in gene transfer delivery systems for gene therapy applications since the early 1980 s by utilizing their ability to integrate into the host genome (99,240). A modified viral genome, carrying a gene of interest serves as a vector to deliver and express the transgene in the host cell. Retroviral vectors derived from gamma-retroviruses can only transduce actively dividing cells (214). This is because the pre-integration complex (PIC) containing the viral (-) ssdna only gains access to the host genome during mitosis when nuclear membrane breakdown takes place. Recently, lentiviral vectors have been used in place of traditional retroviral vectors due to their unique ability to stably transduce both dividing and non-dividing cells, an attribute that significantly expands the host cell range (148,183). The PIC of lentiviruses can traverse the intact membrane of the nucleus. This complex is recognized and docked by the nuclear import machinery, enabling the PIC complex to pass into the nucleus (231). Another advantage of using lentiviral vectors is the lower risk for insertional mutagenesis due to their different integration pattern compared to gamma-retroviral vectors (58,147). Gamma-retroviruses have been shown to integrate near transcription start sites and CpG islands which dramatically increases the chance of insertional mutagenesis, gene activation and enhancer like function (41). However, even though lentiviruses favor active transcriptional units and local hotspot, they still have less chance at activating proto-oncogenes (41,223,260). Development of safe and effective 170

173 gene transfer systems that can stably transduce both dividing and nondividing cells is the ideal character for the delivery of genes into humans. A major obstacle inherent to many gene delivery systems is inefficient expression of the vector transgene as well as concerns for safety. Some general approaches have been widely investigated to overcome these problems, including enhancement of target gene transcription through incorporation of a transcription enhancer, inclusion of post-transcriptional elements, such as WPRE, that facilitate RNA processing and RNA export, or using tissue-specific cellular promoters (82,264). In addition to high level synthesis of RNA, efficient post-transcriptional regulation is a potent target to improve vector gene expression. One of the advantageous approaches was elimination of introns in the transgene transcript to reduce the size of the vector RNA to adjust to the packaging capacity of the vector viruses with limited packaging size (250). However the elimination of introns can significantly diminish protein yield per molecule of transgene transcript because the process of splicing promotes the translation of intron-containing genes (46). Therefore targeting other steps of post transcription replication such as packaging can be very beneficial. Most of the packaging systems contain three major plasmids: 1) Transfer vector that contains the gene of interest, 2) Packaging vector for expression of structural proteins that package the vector transcript into virion. 3) Envelope vector which contains envelope of viruses such as VSVG. Self-renewal and the capacity for differentiation of hematopoietic stem cells (HSCs) have made them the primary targets for the permanent correction of genetic 171

174 defects in all lineages derived from these cells (81). HSC can regenerate all lineages of blood cell, over a long period of time after in vivo transplantation in the recipients. In this chapter, we employed the genomic RNA packaging Enhancer (GRPE) element from the 3 end of HIV-1 gag with the goal of increasing the transgene output from a HIV-based lentiviral vector and preventing formation of defective interfering particles (DIP) containing partial or defective genomic RNA. Increased packaging efficiency is expected to improve the transduction efficiency and transgene expression. From our initial results, we noticed increased packaging in the RNA from transfer plasmid that contained the GRPE element as compared to the RNA from transfer plasmid lacking GRPE. Transformation levels in target cells were increased accordingly with addition of GRPE. Application of GRPE to improve lentiviral vectors is a potentially useful approach for gene delivery in a broad array of clinical and research applications. 172

175 5.4 Results Generation of lentiviral vectors including the GRPE and flanking sequences. In the past few years, lentiviral vectors have been increasingly used for the genetic modification of primary cells, followed by employing them for a variety of applications. In vivo imaging studies of cells have become increasingly popular in order to investigate cell migration and function in animal models. Imaging was first described in murine hepatocytes liver cells which were transduced with the green fluorescent protein (GFP) (199). Researchers have been routinely using imaging to determine in vivo cell distribution and transgene expression in various models such as mice and nonhuman primates (150,222,241,253). In addition to what has been done up to now to enhance the transgene expression and transduction efficiency, improvement of genome encapsidation in produced viruses from transfection is another distinct approach to consider. The goal of the experiments described in this chapter was to investigate the requirement of GRPE for efficient packaging of RNA from transfer RNA from transfection and subsequently optimized transduction of dividing and nondividing cells by vector particles. We used one of the lentiviral vectors, plv-mnd-palg, that has been used for transplantation of hematopoietic cell (HTC) in the laboratory of Gerson and colleagues (150,207) (Fig. 36C). The molecular design of the plv-mnd-palg vector was modified from a previous version by Lin et al. (150) in order to improve the efficiency of viral gene transfer and transgene expression in HSC. HTC transplantation has become a routine 173

176 approach to treat patients with hematopoietic malignancies and bone marrow failure, but hematopoietic stem cell transplantation using conventional lentiviruses containing therapeutic genes suffers from difficulties of culturing and enriching HSC ex vivo. In order to optimize the lentiviral vector, a drug resistance gene which was made by the P140K point mutant of human alkylguanine DNA-transferase (MGMT) was inserted for drug selection (Fig. 36C). The plv-mnd-palg vector also contains the GFP and firefly luciferase genes (Fig. 36C) which have been routinely used in the past few years for HSC transduction to allow non-invasive visualization of transduced HSC homing, trafficking, and localization and in vivo selection over a long period of time. In this vector, the luciferase gene was linked to the MGMT-P140K to monitor homing, trafficking, and expansion after transplantation and drug selection over time. The optimized plv-mnd- PALG lentivector was used in experiments conducted to understand these processes under conditions of low MOI of lentiviral gene transfer which had not been well characterized at that time. In vivo selection was done in HSC transduced with lentiviruses containing the drug resistance gene MGMT-P140K to enrich for transduced and transgene-expressing HSCs. In the study by Lin et al. (150) the dynamic of selection process in vivo was monitored by a novel imaging technique, bioluminescence imaging (BLI). In their study, mice bone marrow cells were transplanted two consecutive times with HPC containing the resistance MGMGT gene and the transplanted recipients were monitored by periodic bioluminescence imaging for up to 9 months. They demonstrated that MGMT-P140K gene transfer followed by drug selection allows robust HSC 174

177 enrichment and stable gene expression in transduced bone marrow in animal models (150). In order to investigate the effect of GRPE on the lentivector transduction efficiency in our study, three fragments containing partial and complete GRPE and surrounding sequences were amplified from HIV-1 NL4-3 (Fig. 36A, B) and cloned into the plv-mnd-palag lentiviral vector (Fig. 36C). The first fragment contained HIV-1 sequence from nt 2085 to nt 2289, therefore lacking the slippery sequence in P2 stem loop of the GRPE/ ribosomal frameshift site (RFS) (Fig. 36B, shown in pink). The second fragment (nt ) had the complete GRPE that was shown in a previous chapter to increase the packaging of genomic RNA significantly (Fig. 36B, shown in blue). The last fragment (nt ) also contained the GRPE, but had more sequences upstream of GRPE (Fig. 36B, shown in green) GRPE increases the packaging efficiency of lentivirus RNA. 293T cells were transfected with plv-mnd-palag transfer vectors along with packaging (pcmv R8.91) and VSVG envelope (pmd.g) vectors at a ratio of 1:3:3 (Fig. 37). The supernatants were harvested and concentrated by ultracentrifugation 48 hours post transfection and were subjected to RT activity assay of the serially diluted cell-free supernatant to measure lentiviral production as explained previously. In the next step, lentiviruses were equalized for RT activity, and then lysed to examine the presence of transfer vector RNA in virus particles. Quantitative reverse transcription PCR (q-rt-pcr) was utilized to quantify the levels of transfer vector RNA using probe/primer set that binds to the HIV-1 175

178 U5 region which only exists in the transfer vector RNA. An increase in packaging of transfer vector RNA was observed when regions nt and nt , both containing the complete GRPE, were included in this RNA (Fig. 38). However, addition of nt fragment which lacked the P2 stem loop of GRPE didn t have any effect. PALAG+ ( ) +GRPE RNA and PALAG+ ( ) +GRPE RNA were packaged by 11 and 3.3 fold increase in comparison to the PALAG RNA (Fig. 38) GRPE enhances the transduction efficiency of lentivirus via grna packaging. To investigate the effect of transfer vector RNA packaging on lentivirus transduction, lentiviruses produced from 293T cells were used to transduce target cells (Fig. 37) using serially diluted MOI. Although 293T and K562 erythroleukemia cell lines were selected to be used as the target, the initial results here just show the infectivity of 293T cells. We first analyzed cells for transgene expression by measurement of luciferase expression 72 hours post infection. The detection of luciferase expression is highly sensitive for transgene expression in the whole cell population. We observed an increase of luciferase expression from RLU of approximately 28,000 to 67,000 and 125,000 at MOI of 1 by addition of fragments nt and nt respectively. The luciferase expression in these PALAG+( ) +GRPE and PALAG+( ) +GRPE lentiviruses were higher than PALAG and PALAG+( ) -GRPE at lower MOI of infection (Fig. 39A) To determine the transgene expression per each cell, 293T cell in the other plate were also collected 72 hours post infection for measurement of GFP expression by flow 176

179 cytometry. The results of GFP expression are shown as percentage of GFP positive cells, considering the positive cells in PALAG lentiviruses at MOI of 1 as 100%. Fig. 39B shows that at the MOI of 1, the transgene positivity is increased from 100% in PALAG(-GRPE) to an average of 200% by addition of GRPE in PALAG+( ) +GRPE and PALAG+( ) +GRPE lentiviruses. The difference in transgene expression is more pronounced at the MOI of 1/256. Similar to results from luciferase assay, the addition of a fragment that contained nt caused more significant augmentation of RNA packaging than a fragment containing nt Interestingly, the region that was responsible for formation of wild type structure of GRPE and maintenance of a wild type packaging level in 5 LTR-p6 868 sgrna [as described in sections and (Fig. 29 and 30)] also contained the sequence between nt 1958 and nt These observations suggest that this amount of flanking sequence is required for the wild type conformation of GRPE and therefore adequate packaging of the RNA through binding of this element with viral or cellular proteins or other RNA structures. Although more experiments such as using primary and hematopoietic cells which are the primary targets for gene therapy, are required, the initial data would suggest an enhancement in lentivirus gene expression and transduction by addition of the GRPE in the transfer vector, likely resulting from formation of lentiviral particles with optimized genome material. Observation of reciprocal packaging between different retroviruses in which viruses such as HIV-1, HIV-2 and SIV can package each other s genome suggest that the evolutionary related retroviruses might contain similar genome packaging, including Ψ 177

180 and GRPE (4,213,262). Thus, after identification of GRPE in HIV-1 we started to look for the related element located in similar location in other lentiviruses such as HIV-2, SIV, FIV, and EIAV. 178

181 Figure 36. Lentiviral vector constructs. (A) HIV-1 map of 5 LTR and gag gene that contains the GRPE/RFS signal. (B) Three fragments from the 3 end of gag were amplified from HIV-1 NL4-3 and inserted in a site downstream of Ψ in the lentiviral vector plv-mnd-palag. One fragment (shown in red) contained only the P2 stem loop of GRPE (as was shown in Fig. 20). The other fragment (blue) has HIV-1 sequences from nt 1958 to nt 2289 and therefore include the entire GRPE. The last fragment (green) contains GRPE and additional upstream sequence up to nt (C) The construct of tricistronic lentiviral vector plasmid (plv-mnd-palag) contained GFP, firefly luciferase gene, and MGMT-P140K gene linked by FMDV 2A cleavage sites and was controlled by an MND promoter (150). 179

182 Figure 37. Lentivirus-based gene delivery system. Transfer vector, envelope vector, and packaging vector were used to transfected 293T cells. Progeny lentiviruses were collected hr post-transfection and used to infect the target cells. 180

183 Figure 38. Quantitative analysis of the RNA content of lentiviral particles released from transfected 293T producer cells. (A) Quantitation of GRPE or +GRPE transfer vector RNA encapsulated into newly formed lenitviruses by q-rt-pcr. Mock control is RNA levels from the transfection without plasmid. Error bars indicate SD from at least three independent experiments. (B) Schematic map of the fragments containing partial or complete GRPE, which were incorporated into the transfer vector RNA. The levels of viral grna were normalized to the RT activity for virus production from transfection. 181

184 Figure 39. Transduction efficiency of Lentiviruses increase by addition of GRPE Transduction experiments by transducing 293T with serially diluted lentiviruses PALAG- GRPE, PALAG+( )-GRPE PALAG+( )+GRPE, and PALAG+( )+GRPE. The transfer gene expression was measured by (A) luciferase activity and (B) flow cytometry for GFP protein expression. 182

185 5.5 Discussion Lentivirus vector based delivery of various genes has become widespread for preclinical and clinical research. HIV-1 derived lentiviral vectors provide a long-term expression of short hairpin RNA (shrna), transgenes (basic research), or the therapeutic gene (clinical) in a wide-range of cell types with sustained long-term expression of the transgenes which makes them a great tool in gene therapy (148,183). Efficient transfer and sustained expression of transgenes are among the most important factors to consider in gene delivery. As the transition of therapeutic applications for lentiviral vectors from preclinical animal models to corrective therapy in humans has started, it is vital to optimize the lentiviral vector design and production, to eliminate potential hazards such as insertional mutagenesis, vector mobilization, and generation of replication competent lentiviral vectors (RCLs). The first obstacle observed in gene delivery systems using simple retroviral vector occurred by development of mutagenesis in three patients from a SCID-X1 clinical trial as a consequence of the insertional mutation (45,106). All of these putative risks increase with using high MOI of lentiviruses for transduction of cells to compensate for the low gene expression efficiency in the transduced cells from which gene delivery system is suffering. Infection of cells with high MOI of lentiviruses has had unpredicted consequences regarding to insertional mutagenesis and the immune response. 183

186 During development of HIV-1 based lentiviral vectors in the past two decade, many of improvements have been accomplished with the goal of increasing safety, transgene expression and transduction efficiency. These manipulations of lentiviral vector have resulted in generation of the lentiviral system that includes three to four expression cassettes, each harboring different portion of HIV-1 genome separately, in order to reduce the possibility for formation of RCLs. The packaging vector provides the structural and enzymatic protein expression, while the envelope vector provides the envelope for entry into the host cell. The gene of interest (transgene) is located in the transfer vector which also harbors important cis acting elements such as PBS, major packaging signal Ψ, PPT, central polypurine tract (cppt), rev-response element (RRE), and the woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) required for efficient expression of the gene in the transduced cell. Ψ is placed only in the transfer vector to exclude the packaging of other RNAs into the virion, except the transfer vector RNA. PPT and cppt are required for reverse transcription and efficient integration respectively (70,263). The WPRE sequence at the 3 side of transgene enhances overall mrna transcript stability and overall transgene expression (265) due to efficient RNA processing. However, due to the risk of possible genetic recombination with wild-type lentiviruses and insertional mutagenesis, most of the HIV-1 genome is removed from the current version of the transfer vector. Even though the 3 end of the gag coding region contains the signal for ribosomal frame shifting, it is excluded from the transfer vector because the transfer vector RNA is not for Gag/Gag-pol production; but it is for the packaging vector. Due to the deletion of this region, the genomic 184

187 packaging enhancer element or GRPE which overlaps with RFS is also missing from the transfer vector RNA, even though it has been shown to be necessary for efficient RNA encapsidation. The common system of HIV-1 and other lentivirus vector production is based on a transient transfection of 3 4 plasmids into 293T producer cells. The virus-containing supernatant from transfection is purified and centrifuged to be used for transduction of target cells. It has been shown that a large portion of produced lentiviral particles lack or contain a defective viral genome and therefore function as defective interfering particles (DIPs) (24). As a consequence, when lentiviruses containing DIPs are used for transduction, the transgene expression is not very efficient. Using lenviruses with maximal packaged genome for transduction is one measure, critical for efficient transduction. By enhancing the packaging of lentiviral (transfer) RNA genome in the viruses produced from transfection, the amount of DIPs in the population of lentivirus drops which consequently increases the transduction efficiency without perturbing normal function of the host cell genome. After discovering that the GRPE is absent from most single-cycle lentiviral systems used for gene therapy, biochemical studies, or vaccine development, we postulated that a novel, yet to be developed technique to increase transduction (and subsequently virus titer in transduced cells) is to optimize the packaging efficiency of transgene RNA by addition of GRPE. Preliminary studies suggest that introducing the GRPE within a HIV-based gene delivery vector, even out of sequence context; can result in increased transduction and gene delivery, which is directly related to increased grna 185

188 encapsidation. The requirement for HIV-1 GRPE was demonstrated by enhancement of the levels of + GRPE transfer vector RNA in progeny pseudo-viruses assembled from 293T cells as compared to the levels of GRPE transfer vector RNA. However, more experiments are required for achieving the efficient vector production and infectivity in hematopoietic and primary cells. Enhanced virus production from transfected cells would help in reducing the high multiplicity of infection (MOI) for transduction with lentiviruses which is being used currently to increase clinical efficacy, and gene expression which increases the theoretical risk for insertional mutagenesis. 186

189 CHAPTER 6 GENERAL DISCUSSION AND FUTURE DIRECTIONS 187

190 GENERAL DISCUSSION AND FUTURE DIRECTIONS HIV-1, like most retroviruses, must incorporate two copies of a positive-strand unspliced genome (g) RNA, both of which are important for genetic recombination during reverse transcription and proviral DNA synthesis (119). A highly ordered and precise process is involved in the packaging or encapsidation of grna into new virus particles. Unspliced HIV-1 RNA constitutes less than 1% of the total cellular mrna (24) yet is preferentially selected for packaging. Prior to packaging, intergenomic annealing initiates formation of loose non-covalent dimers of the unspliced HIV-1 RNA (192), which is then further condensed during encapsidation and association with HIV-1 nucleocapsid (NC) domain of Gag polyprotein (125). Stringent selectivity for packaging of the grna dimer is based on recognition of a cis-acting packaging element by zinc Knuckles of NC (24,60,145,191,192,208,252). Critical to our understanding of the packaging process are the RNA motifs that are involved in genome packaging. Numerous studies have suggested that the main packaging signal (also termed psi or Ψ) is located in an approximately 120 nucleotide region within the 5 UTR of HIV-1 RNA (5,60,111,132,144,169). Moreover, Ψ is thought to be located in other sequences such as trans-activating-responsive (TAR) stem-loop, splice donor site and primer binding sequences (48,50,135,167,212,218). This chapter summarizes the significance and future perspectives of the research presented in this dissertation. In closing, this dissertation has: (1) outlined the approaches aimed at identifying a novel cis-acting packaging element, aside from the 188

191 canonical RNA packaging element Ψ, which increases genomic RNA packaging by at least 10-fold; (2) determined that GRPE overlaps with a ribosomal frameshift site and suggested that low frequency frameshifting may designate the HIV-1 mrna as genomic RNA; (3) determined that the introduction of the GRPE into lentiviral vectors has the potential to dramatically increase gene delivery and improve tools for more effective gene therapy 6.1 Identification of an additional packaging element within HIV-1 coding region. We determined that a novel site, aside from the Ψ signal, within the 3 end of gag coding region of the HIV-1 genome, was necessary for efficient RNA packaging by using the dual transfection system described in Chapters 3 and 4. Several years ago, our laboratory adopted a yeast-based cloning strategy for cloning of any gene into a HIV-1 backbone (74). During the development of this system, we stumbled onto this new packaging determinant due to the unexpected and poor transformation efficiency/infectivity of viral particles carrying a short HIV-1 genomic RNA that lacked the HIV coding sequence. This site was therefore termed genomic RNA packaging enhancer or GRPE element. The importance of the two GRPE stem-loops for grna packaging was evident by the studies with point mutations and deletions of intervening sequences between the GRPE and 5 UTR. Based on the results, the sequence for GRPE was narrowed down to a region between 2015 and 2175 nt which also contains the ribosomal frameshift signal 189

192 (RFS). Thus, we suspected that this element had been missed due to its importance for the ribosomal -1 frameshifting function, which is mediated by the overlapping RFS. More mportantly, based on our result, GRPE appears to function outside of its grna position and can be re-positioned immediately upstream of the Ψ. As explained previously, the structure of stem loops in the Ψ signal is highly conserved in all complex retroviruses. Therefore, the existence of a GRPE element in the RNA of HIV-1 may also be extrapolated to other lentiviruses. The presence of similar regions at the 3 end of gag and its dual role as frameshift signal and GRPE element is currently being investigated in lentiviruses including HIV-2, feline immunodeficiency viruses (FIV), simian immunodeficiency virus (SIV), also non-primate lentiviruses including equine infectious anemia virus (EIAV). If the GRPE is shown to be present as a conserved structure in other lentiviruses, it would reinforce the link between RNA structure and function. Overall, the delineation of HIV-1 grna packaging determinants presented here should enhance our understanding of retroviruses packaging, which is imperative for development of antiviral drugs and lentiviruses used in human gene therapy. 6.2 Interplay between grna packaging and mrna translation/frameshifting It is known that during the process of RNA packaging, viral proteins and HIV-1 grna must co-localize in the host cell and move to the inner plasma membrane to form budding viral particles. The sequence of events in which RNA is recognized by Gag, 190

193 forms the RNA:Gag complex, and then traffics to the viral assembly site is still not well understood. Some studies on Gag trafficking have provided evidence that genome packaging by Gag might be dependent on the shuttling of Gag through the nucleus, mediated by NLS and NES signals in some retroviruses such as Rous sarcoma virus and HIV-1 (104,198). However, the role of HIV-1 Gag in nuclear export of viral RNA remains controversial and needs further validation. In a more recent study by Bennasser et al. (20) engineered RNA were used to demonstrate that dimerization of the HIV-1 genome takes place in the cytoplasm. If we assume that Gag preferentially binds to the genomic RNA in the form of dimer, this finding would seemingly suggest the location of Gag:RNA assembly to be in the cytoplasm. Moreover, confocal microscopy and recent live cell imaging studies by Jouvenet et al (127,128) showed that HIV-1 ribonucleoprotein complex is formed in a perinuclear/centrosomal site followed by its transport to the inner plasma membrane (201). It appears that the ribonucleoprotein complex contains a very small number (fewer than 12) of precursor Gag proteins and one copy of grna dimer (201). During assembly, the complex recruits approximately 4000 more Gag proteins to form the structure of new viral particle. Although there is basic knowledge about recognition of RNA by the NC domain of Gag, the location of Gag:RNA recognition, and trafficking of the complex by the HIV Matrix and cellular protein, there remains considerable interest in understanding the molecular basis for genome selection and the fate of genomic RNA in the cytoplasm of infected cells. A longstanding question in retroviral RNA biology is: How does Gag specifically direct the packaging of only one copy of the dimeric RNA genome? 191

194 Extensive studies have characterized the relationship between translation and encapsidation for the unspliced viral RNA (2,22,122,155). It appears that after transcription by cellular RNA polymerase II, the newly full length synthesized viral RNA can either serve as mrna for the production of the viral proteins or alternatively as a genome to be encapsidated in the newly formed virions. Depending on the retrovirus, the dual roles of viral RNA are mutually exclusive or temporally regulated. Different scenarios have been suggested for selection of an unspliced RNA for packaging or translation (38,100,155): 1) Translation and encapsidation machinery compete for utilization of genome-length RNA since the packaging signal Ψ becomes disrupted during the cap-dependent translation which melts the structure of 5 LTR during initiation. On the other hand, nucleocapsid binding to the RNA packaging signal is expected to arrest ribosome scanning and inhibit the translation. 2) The genomic-length RNA functions interchangeably as an mrna template and virion RNA. 3) mrna translation is a prerequisite for RNA packaging due to the requirement for viral proteins for the assembly process. The transition between translation and encapsidation is structures of the 5 UTR architecture (2,22,122,155). Depending on whether the 5 UTR adopts the Branched multiple hairpin (BMH) or Long range interaction (LDI) alternative conformations, the presentation of the dimerization initiation site (DIS) hairpin and packaging signal Ψ, and therefore their accessibility for NC binding, changes. BMH conformation promotes the dimerization, NC binding and subsequently RNA packaging due to the long-range RNA RNA interaction between AUG, and residues in U5, which results in exposure of the DIS 192

195 and Ψ hairpins (154). On the other hand, in LDI conformation poly(a) and DIS domains are base paired which makes it favorable for translation (122). As an outcome, this LDI- BMH riboswitch has been suggested to regulate the dimerization and translation (as shown in Fig. 11). In agreement with the LDI-BHM ribo-swicth proposal and based on the results from erf1 downregulation, we predicted a model that proposes the GRPE as a site for regulation of the encapsidation-frameshifting transition (Fig. 40). Our model proposes a control of HIV-1 translation, ribosomal frameshifting, and grna packaging by the GRPE/RFS. Based on the model, after unspliced viral RNA is exported to the cytoplasm, its 5 UTR region forms the LDI conformer and it therefore serves as the template for Gag and gag-pol protein translation (Fig. 40). If the mrna is used for Gag protein production, during translation, a complex involving erf1 may form at or near the GRPE/RFS for translation termination and consequently initiates mrna degradation by activation of the nonsense-mediated decay (NMD) pathway. During NMD degradation, Up-frameshift protein 1 (UPF1) which is a RNA decay enzyme (as shown in Fig. 13) accumulates on the extended 3 UTR through binding to the erf1 complex and thus prevents Gag precursor binding to this RNA for encapsidation (114). In <5% of translation events the unspliced HIV-1 mrna is used for Gag-pol translation by ribosomal frameshifting. In this case, the ribosome may shift -1 nt at the RFS, clear the GRPE, prevent recruitment of the Upf1- dependent decay complex, and promote interactions of cytoplasmic poly(a)-binding protein 1 (PABPC1) with erf1/erf3 (123) to eventually terminate Gag-Pol translation (Fig. 13). This rare event in HIV-1 translation would maintain a stable unspliced HIV-1 193

196 RNA where the GRPE may be free to participate in higher order RNA interactions (possibly with Ψ) and/or interact with HIV-1 Gag (or additional cellular factors) (Fig. 40). Binding of the free GRPE to Ψ and related 5 UTR sequences will probably result in riboswitch in conformation of that region from LDI (translation-competent) to BMH (packaging-competent). Thus, the unspliced mrna which was initially used as a template for Gag-Pol translation would also serve as genomic RNA. This model could provide an explanation for the selection of only one RNA dimer as genomic material per virion among the pool of unspliced RNA. 194

197 Figure 40. Model for translation and packaging of the unspliced HIV-1 RNA. Unspliced mrna are destined for (1) Gag translation and subsequent substrate for NMD pathway or (2) may serve as a template Gag-Pol translation following ribosomal frameshift and thus avoiding NMD (<5%). Ribosomal translocation across the unspliced mrna used for Gag-Pol translation may clear the RNA of cellular factors (e.g. erf1) and thus promote new secondary/tertiary HIV-1 RNA resulting in RNA-protein interactions necessary for grna dimerization following by packaging. MA: Matrix, CA: Capsid, NC: Nucleocapsid. 195