Overexpression of the Turnip Crinkle Virus Replicase Exerts Opposite Effects On

Home , Carmotetraviridae, Cytorhabdovirus, Luteovirus, Machlomovirus, Sclerodarnavirus, Turnip crinkle virus

Overexpression of the Turnip Crinkle Virus Replicase Exerts Opposite Effects on

the Synthesis of Viral Genomic RNA and a Novel Viral Long Non-Coding RNA

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the

Graduate School of The Ohio State University

Shaoyan Zhang, M.S.

Graduate Program in Plant Pathology

The Ohio State University

2020

Dissertation Committee:

Dr. Feng Qu, Advisor

Dr. Maria S. Benitez Ponce

Dr. Renukaradhya Gourapura

Dr. Scott P. Kenney

Dr. Tea Meulia

2020

ABSTRACT

Plant viruses routinely inflict serious losses in the yield of food and cash crops, ruining the livelihood of farmers, and leading to hunger and malnutrition in many parts of the globe. Our research focuses on a group of viruses known as positive sense (+) RNA viruses that are the most common disease-causing viruses in crops worldwide. All (+)

RNA viruses use single-stranded RNAs as the genome to encode a small number of viral proteins on the genomic RNA. Despite the enormous diversity among the proteins these viruses encode, all (+) RNA viruses encode at least one protein, the viral RNA-dependent

RNA polymerase (RdRp), that directs the replication of the viral genome. Existing evidence indicates that the expression of RdRp in virus-infected cells is tightly regulated at low levels, but the consequences of perturbing expression of RdRp have not been well understood. Therefore, our research concentrates on understanding how increasing the

RdRp expression levels of a (+) RNA virus could affect the viral RNA replication dynamics.

In Chapter one, we review the existing literature about RdRps encoded by different

(+) RNA viruses. we first give an overview of the replication process of (+) RNA viruses, then summarize various mechanisms that control the expression levels of RdRp. We also touch on viral derived long noncoding RNAs (lncRNAs), which are often 3’ co-terminal with genomic RNAs of (+) RNA viruses. We introduce turnip crinkle virus (TCV), the

i primary model virus used in my thesis research. At the end of Chapter one, we raise the question that how increasing the RdRp expression levels of a (+) RNA virus could affect the viral RNA replication dynamics. We hypothesize that overexpression of the p88

RdRp encoded by TCV compromises TCV replication and upsets the relative accumulation levels of different TCV RNAs. This hypothesis is rigorously tested in

Chapters two and three.

Previous work in our lab found that p28, the auxiliary replication protein of TCV, trans-complemented a defective TCV lacking p28, yet repressed the replication of another TCV replicon encoding wildtype p28. In Chapter Two we showed that p88, the

TCV-encoded RdRp, readthrough product of p28, likewise trans-complemented a p88- defective TCV replicon, but repressed the replication of another TCV replicon encoding wild-type p88. Surprisingly, lowering p88 protein levels enhanced trans- complementation, but weakened repression. Repression by p88 was not simply due to protein over-expression, as deletion mutants missing 127 or 224 N-terminal amino acids accumulated to higher levels but were poor repressors. Finally, both trans- complementation and repression by p88 were accompanied by preferential accumulation of subgenomic RNA2, and a TCV specific lncRNA. Our results suggest that repression of

TCV replication by p88 may manifest a viral mechanism that regulates the ratio of genomic and subgenomic RNAs based on p88 abundance.

Many positive sense (+) RNA viruses encode long noncoding RNAs (lncRNAs) that play important roles in their infections. A distinguishing feature of these lncRNAs is that they are produced through 5'-to-3' degradation of viral genomic or subgenomic RNAs, by

ii exoribonucleases of host cells. In Chapter Three, we further investigated the lncRNA discovered in Chapter Two and found that this TCV-borne lncRNA was produced by a replication-based mechanism. This lncRNA, designated tiny TCV subgenomic RNA

(ttsgR), was mapped to the last 283 nucleotides of TCV genomic RNA. It accumulated to high levels in cells of Nicotiana benthamiana plants in which TCV replication took place in the presence of overexpressing the RdRp p88. Moreover, ttsgRNA replicated robustly from templates as short as itself, without the need for any other TCV RNAs, as long as both of the TCV replication proteins, p28 and p88, were provided in trans from nonviral sources. Accordingly, both (+) and (-) sense forms of ttsgR were detected using a strand- specific RT-PCR procedure. ttsgRNA replication did not entail any 5’ RNA secondary structure but required the presence of a G3(A/U)4 motif at the 5’ terminus. Furthermore, it strictly relied on the integrity of the CCC motif at the 3’ terminus. Both of these structural features are shared by TCV genomic and subgenomic RNAs. These findings established that ttsgRNA was the product of a replication-based mechanism, and identified a novel strategy for the biogenesis of lncRNAs associated with (+) RNA viruses.

In Chapter four, we extend our RdRp research to a distinct (+) RNA flock house virus

(FHV), which infects plants, insects, and also mammalian cells. Unlike TCV, FHV encodes one single replication protein – the RdRp – that directs the entire replication cycle. Our preliminary experiments nevertheless demonstrated that FHV RdRp, when over-expressed in N. benthamiana cells from a non-viral source, was able to complement

iii the replication of FHV mutants lacking their own RdRp. More importantly, the same

RdRp potently repressed the replication of FHV replicons encoding a functional RdRp.

Collectively, my thesis research demonstrated that over-expressing RdRps of (+)

RNA viruses appears to have a conserved, repressive effect on the replication of the cognate viruses. These results, combined with findings by others in our lab, are consistent with a working model postulating that viral RdRps and possibly other replication proteins, dynamically regulate the viral replication process depending on the intracellular concentration of these proteins. Manipulating the concentration of these proteins could potentially control the replication of viruses. These findings are expected to provide an important guide for novel virus control strategies.

DEDICATION

To my parents for their unconditional love and support,

to my advisor for his patient guidance and fruitful help

ACKNOWLEDGMENTS

I cannot begin to express my gratitude to my advisor, Dr. Feng Qu, for his guidance, critiques, and advices throughout my PhD studies. Hundreds of hours of discussions with him cultivated my logical thinking and sense of science. His passion and dedication for science are inspiring and set a standard that I will always strive to observe. My deep and heartfelt thanks to my committee members Drs. Maria Soledad Benitez Ponce,

Renukaradhya Gourapura, Scott Kenney, and Tea Meulia. It has been my privilege to have had the opportunity to be guided by all of you! Although your own research differs from my focus, you understood my research with a surprising depth, and are always willing to spend time to help and guide me. I am also very grateful to Dr. Monica

Lewandowski for her wise suggestions whenever I run into questions and troubles.

I’m fortunate to work with previous and current members in Qu’s Lab. I am deeply appreciative of Junping Han, Xiuchun Zhang, Xiaofeng Zhang, Qin Guo, Rong Sun,

Xiaolong Yao, Fides Zaulda, Limin Zheng, Camila Perdoncini Carvalho and Tu Huynh for their friendship and collaborations. The assembly line we developed to scale up

Nicotiana benthamiana planting will be a legend in our Lab.

A steadily growing environment is vital for research that like mine uses plants heavily, I am fortunate to have found the greenhouse and culture rooms are usually well maintained, and for that I am thankful for Bob James for maintaining greenhouse and

vi keeping greenhouse supplies handy. Thanks to Lee Wilson for his always readiness to solve any mechanical problem in the greenhouse, culture room and other matters.

Thanks to MCIC for technical support in sequencing and confocal imaging. Thanks to the Labs of Drs. Lucy Stewart, Peg Redinbaugh, Sally Miller for generous equipment sharing. Thanks to department for tuition assistance.

It wouldn’t have been easy for a foreign student to study and live abroad, if not for the wise counsel, generous help and support of Mark Jones, Jane Todd, Kristen Willie,

Jody Whitter, Andrea Kaszas, Maria Elena, Fiorella Carter, Santosh Dhakal, Sankar

Renu, Hugo Pantigoso, Luis Huezo, Kaylee South, Chengsong Hu, Yi Han, Pia

Golborne, Lu Zhao, Zhenyu Li, Pailing Liu, Yiyun Lin, Timothy Frey, Ram Khadka,

Wanderson Moraes, Cecilia Freitas, Hong Hanh Tran, Deogracious Massawe, Brian

Hodge, Seyed Mousavi, DeeMarie Troyer-Marty, Sizo Mlotshwa, Wirat Pipatpongpinyo,

Mingde Liu, Ke Li, Yixuan Hou, Su, and Daowen Hou. I absolutely treasure our friendship and I am so grateful to have them in my life.

And finally, I cannot begin to express my love and gratitude to my mom and dad, who gave me the best education they can ever afford and care for me more than themselves. To my little sister, who always makes me laugh.

vii

VITA

June, 2011…………………...………… B.S. Agriculture, Hainan University, China

June, 2014……………. M.S., Molecular Plant Pathology, Hainan University, China

July, 2015……………………...… Visiting Scholar, Department of Plant Pathology,

Food Animal Health Research Program, The Ohio State University, U.S.A

August, 2015 to present...... Graduate Research Associate,

Department of Plant Pathology, The Ohio State University, U.S.A

PUBLICATIONS

Guo, Q., Zhang, S., Sun, R., Yao, X., Zhang, X.-F., Tatineni, S., Meulia, T., and Qu, F.

2020. Superinfection exclusion by p28 of turnip crinkle virus is separable from its replication function. Molecular Plant-Microbe Interactions 33:364-375.

Sun, R., Zhang, S., Zheng, L., and Qu, F. 2020. Translation-independent roles of RNA secondary structures within the replication protein coding region of turnip crinkle virus.

Viruses 12:350.

viii

Zhang, S., Sun, R., Guo, Q., Zhang, X.-F., and Qu, F. 2019. Repression of turnip crinkle virus replication by its replication protein p88. Virology 526:165-172.

Zhang, X.-F., Zhang, S., Guo, Q., Sun, R., Wei, T., and Qu, F. 2018. A new mechanistic model for viral cross protection and superinfection exclusion. Frontiers in plant science

9:40.

Zhang, X.-F., Sun, R., Guo, Q., Zhang, S., Meulia, T., Halfmann, R., Li, D., and Qu, F.

2017. A self-perpetuating repressive state of a viral replication protein blocks superinfection by the same virus. PLoS pathogens 13:e1006253.

FIELDS OF STUDY

Major Field: Plant Pathology

Specialization: Molecular Virology

TABLE OF CONTENTS

ABSTRACT ...... i DEDICATION...... v ACKNOWLEDGMENTS ...... vi VITA...... viii TABLE OF CONTENTS ...... x LIST OF FIGURES ...... xiv LIST OF TABLES ...... xv CHAPTER 1...... 1 Literature Review ...... 1 Overview ...... 1 The replication process of (+) RNA viruses ...... 3 Expression level of RdRp is tightly regulated through various mechanisms in (+) RNA viruses...... 5 Viral subgenomic (sg) RNAs and long non-coding (lnc) RNAs ...... 11 TCV as a unique model for studying the effect of RdRp overexpression...... 15 Research Hypothesis and Objective ...... 17 Significance of my thesis research ...... 18 CHAPTER 2: Repression of Turnip Crinkle Virus Replication by Its Replication Protein p88 ...... 20 Abstract ...... 20 Keywords ...... 20 Introduction ...... 21 Results ...... 22 p88 expressed independently of replication trans-complements the replication of p88-defective TCV replicons...... 22 p88-2HA represses the replication of a TCV replicon encoding intact p88...... 24

The C-terminal 2HA tag does not appreciably affect p88 activities...... 26 Trans-complementation by p88-2HA tolerates limited N-terminal truncation...... 26 p88 mutants capable of trans-complementation are also strong repressors of replication...... 28 Replicational repression by p88 mutants partially correlates with the induction of p28 inclusions...... 29 p88 with a C-terminal GFP tag forms large inclusions similar to p28-GFP...... 31 Discussion ...... 32 Materials and methods ...... 35 Constructs...... 36 Agro-infiltration...... 36 RNA extraction and Northern blotting...... 36 Protein extraction and Western blotting...... 37 Detection of (-) gRNA with semi-quantitative RT-PCR...... 37 Confocal microscopy...... 37 Acknowledgments ...... 38 Figures ...... 39 CHAPTER 3: Replication-Based Biogenesis of a Long Noncoding RNA Originated from Turnip Crinkle Virus ...... 48 Abstract ...... 48 Introduction ...... 49 Results ...... 52 Overexpression of p88 unveils a novel TCV RNA...... 52 Biogenesis of ttsgRNA is replication dependent...... 54 ttsgRNA initiates at a discrete position of TCV RNA, and it exists in both viral sense and complementary sense forms...... 57 An RNA stem loop flanking the 5’ end of ttsgR does not contribute to ttsgR biogenesis...... 58 A shorter ttsgR template with 28 extra nt at the 5’ end reveals specific sequence requirements for ttsgR accumulation...... 59 A ttsgRNA template with an exact 5’ terminus directs efficient ttsgR synthesis...... 61 A 5’ terminal G3(A/U)4 motif is critically important for ttsgRNA accumulation. ... 61

The 3’ terminal CCC triplet is essential for efficient ttsgRNA accumulation...... 64 The GGGUAAA motif is needed for ttsgRNA synthesis from infectious TCV RNA templates...... 66 Discussion ...... 67 Replication-based production of ttsgRNA...... 67 Role of p88 overexpression in ttsgRNA...... 69 Conservation and repair of ttsgRNA terminal characteristics...... 70 Material and Methods...... 71 Constructs...... 71 Agro-infiltration...... 73 RNA extraction and Northern blotting...... 73 Protein extraction and Western blotting...... 73 Mapping of ends sequence of ttsgR...... 74 Acknowledgement ...... 75 Figures...... 76 Table ...... 87 CHAPTER 4: Repression of Flock House Virus replication by its replication protein pA: Preliminary Findings...... 99 Abstract ...... 99 Introduction ...... 99 Results ...... 102 An agro-infiltration-delivered FHV RNA1 replicon replicates in N. benthamiana leaf cells...... 102 Tag-free pA is fully functional when provided in trans but lacks repressive activity...... 103 An N-terminally tagged pA is less capable of replicational complementation, whereas a C-terminally tagged pA is more potent at repression replication...... 104 Tag-free B1 also represses the replication of FHV RNA1 replicons...... 105 Sequential infiltration reveals the potent suppression function of pA...... 105 Protein A prefers to replicate pA defective replicons over replicon with long insertion...... 106 Discussion ...... 107

xii

Material and methods ...... 109 Constructs...... 110 Agro-infiltration...... 111 RNA extraction and Northern blotting...... 112 Acknowledgement ...... 112 Figures ...... 113 Table ...... 118 Bibliography ...... 123

xiii

LIST OF FIGURES

Figure 2.1. TCV p88 partially trans-complements a p88-defective replicon...... 39 Figure 2.2. The level of p88-2HA protein correlates negatively with its trans- complementation activity, but positively with its replicational repression activity...... 41 Figure 2.3. The C-terminal 2HA tag does not appreciably affect p88 activities...... 42 Figure 2.4. Trans-complementation mediated by various deletion mutants of p88-2HA. 43 Figure 2.5. Replicational repression by p88-2HA mutants...... 45 Figure 2.6. Replicational repression by p88-2HA variants partially correlates with their ability to trans-aggregate G11-p28...... 46 Figure 3.1: Overexpression of p88 unveils a novel TCV RNA...... 76 Figure 3.2: Biogenesis of ttsgRNA is replication dependent...... 78 Figure 3.3: Circularization RACE mapping the ends of ttsgR...... 80 Figure 3.4: Specific sequence requirement for ttsgR accumulation...... 82 Figure 3.5: A 5’ terminal G3(A/U)4 motif is critically important for ttsgRNA accumulation...... 84 Figure 3.6: The 3’ terminal CCC triplet is essential for efficient ttsgRNA accumulation...... 85 Figure 3.7: The GGGUAAA motif is needed for ttsgRNA synthesis from infectious TCV RNA templates...... 86 Figure 4.1 Replication of FHV RNA1 in N. benthamiana...... 113 Figure 4.2 The impacts of trans-expression of pA and its variants on replication of various RNA1 replicons revealed by northern blotting...... 115 Figure 4.3 The impacts of trans-expression of B1 and B2 on replication of various RNA1 replicons revealed by northern blotting...... 116 Figure 4.4 Sequential infiltration unmasked the replication repression function of tag free pA...... 117 Figure 4.5 Sequential infiltration reveals pA preferentially replicated more authentic RNA1 by northern blotting...... 117

xiv

LIST OF TABLES

Table 3.1: gBlocks used in this study ...... 87 Table 3.2: Primers used in this study ...... 93 Table 3.3: Primers used for circularization RACE ...... 98 Table 4.1: Constructs used in this study ...... 118

CHAPTER 1.

Literature Review

Overview

As we have been witnessing right now, viruses like SARS-CoV-2, human immunodeficiency virus (HIV) (Hargreaves et al. 2020), and influenza viruses

(Hutchinson 2018) cause devastating illnesses and deaths in human beings. We are also well aware that viruses cause severe diseases in both domesticated and wild animals, and some of those viruses often spill over to infect humans, leading to pandemics that cripple economic activities in human societies. Therefore, it seems self-evident as to why we must study viruses that infect humans and animals – so we can find ways to contain and eradicate these viruses.

Lesser known to the public are the losses caused by plant viruses to food and cash crops worldwide, with annual economic fallout easily surpassing billions of dollars (Wei et al. 2010b). It is well documented that rice, the number one grain for human consumption, is plagued with numerous virus diseases that threaten the food security of billions of people and endanger the livelihood of subsistent rice farmers in Southeast Asia and Africa (Hibino 1996; Yang et al. 2017). Another example is the maize lethal necrosis disease that is causing near 100% losses in maize production in a number of

African countries for the last ten years (Redinbaugh and Stewart 2018). However, the

1 importance of plant virus research lies not just on finding solutions for crop virus diseases

– it has historically helped to shed light on basic principles of virus life cycles shared by all viruses. The first identified virus was a plant virus – tobacco mosaic virus (TMV). Not until discovery of TMV and proven of its pathogenicity in plants did people realize there is a totally new category of pathogen that can infect almost all living organisms on earth

(Kelman 1995).

My research focuses on a group of plant viruses known as positive sense (+) RNA viruses. They are so named because these viruses use single-stranded, positive sense

RNA as the carrier of their genetic information. These RNA genomes are defined as positive sense because all viral protein open reading frames (ORFs) are identifiable on the genomic RNA themselves. Indeed，many or all of these proteins are directly translated from the RNA genomes. (+) RNA viruses are the most common disease- causing viruses in crop plants worldwide (Gergerich and Dolja 2006).

Unlike cellular parasites (e.g. bacteria, fungi, nematodes, protists), viruses lack the basic machineries and components needed for protein and nucleic acid synthesis, hence need to enter host cells, and repurpose the host cell machineries to synthesize viral proteins and replicate viral genomes. (+) RNA viruses are no exception. Nevertheless, all

(+) RNA virus genomes encode at least one essential protein, known as viral RNA- dependent RNA polymerase (RdRp), or viral replicase, that directs the replication of the viral genome itself, thus sustaining viral infections in successively invaded cells.

Therefore, understanding how viral replicases manipulate host cell machineries to

2 achieve viral replication could reveal characteristics shared by all (+) RNA viruses, which could then become targets for antiviral therapy.

This rationale forms the basis of my thesis research. Specifically, my research has concentrated on understanding how perturbing the expression levels of the replicase of a

(+) RNA virus could affect the viral replication dynamics. Below I will summarize the existing knowledge relevant to my research. I will then formulate my central hypothesis, and outline the specific objectives aimed at testing the hypothesis.

The replication process of (+) RNA viruses

The (+) RNA virus replication cycle begins with the release of the (+) RNA genome into the cytoplasm of the infected cell, allowing it to engage host cell ribosomes to translate proteins needed for genome replication, among them the viral replicase or

RdRp. Some (+) RNA viruses, such as the plant-infecting potyviruses and animal- infecting flaviviruses, translate nearly all viral proteins as a single polyprotein precursor that is then proteolytically processed into multiple mature proteins, among them RdRp.

In addition to RdRp, most (+) RNA viruses also translate one or more other proteins, known as auxiliary replication proteins (ARP), that function together with RdRp in the replication process. Examples of ARP include p126 encoded by tobacco mosaic virus

(TMV), 1a encoded by brome mosaic virus (Kroner et al. 1990), 6K2 encoded by potyviruses (Restrepo-Hartwig and Carrington 1994; Schaad et al. 1997), as well as NS1,

NS3, NS4A and NS4B encoded by flaviviruses (Mazeaud et al. 2018).

Most animal and plant-infecting (+) RNA viruses replicate their genome in ARP and/or RdRp induced membrane invaginations or vesicles (Ahola 2019). Viruses from different families utilize different organelle membrane to form replication factories

(Salonen et al. 2005). Picornaviruses, flaviviruses, and bromoviruses prefer endoplasmic reticulum (ER) membranes, togaviruses utilize endosomes and lysosomes, tombusviruses utilize peroxisomes and chloroplasts, nodaviruses and carmoviruses utilize mitochondria, while different potyviruses may exploit membranes of different organelles, including ER, chloroplast, or peroxisome (Wei et al. 2010a).

The replication of (+) RNA viruses begins with the synthesis of a negative sense (-) copy of the (+) RNA genome. The (-) RNA then serves as the template for the synthesis of (+) RNA. Many models have been proposed to explain how progeny (+) RNA genomes are synthesized. Prominent among them are the stamping machine model

(SMM), and the geometric replication model (GRM) (Martinez et al. 2011). In the SMM, the founding (+) RNA only templates one round of (-) RNA synthesis, and the (-) RNA then templates the synthesis of all (+) RNA progenies (Sanjuán and Domingo-Calap

2016). As no more (-) RNA is synthesized, stamping machine model predicts a huge accumulation imbalance between RNAs with different polarities. In contrast to SMM, the

GRM postulates an equal efficiency replication in both RNA polarities that is similar to semi-conservative replication in DNA replication. Since the mutations from either polarity will be passed on and on, GRM anticipates a much higher mutation rate

(Martinez et al. 2011). Experimental results (Quinkert et al. 2005) showing asymmetric accumulation of viral RNA with different polarities favors the SMM. However, from an

4 evolutionary perspective, higher mutation rates resulting from GRM might also benefit the viruses.

After replication, the newly synthesized (+) RNA genomes are assembled into virus particles (virions) made of virus-encoded capsid proteins and, in plants, spread into neighboring cells with the assistance of viral encoded cell-to-cell movement proteins.

Genomes of some (+) RNA plant viruses were also found to move cell-to-cell in the form of RNA-MP complexes (Taliansky et al. 2008). The relatively slow infections of consecutive cells in the initially infected tissues eventually allow virions to enter vascular bundles, whereupon viruses quickly spread to distal tissues through rapid systemic movement, causing the entire host plant to become infected. Transmission of plant viruses between different host plants is mostly mediated by arthropod vectors, although other forms of transmission such as fungi mediated transmission are also known to occur.

Expression level of RdRp is tightly regulated through various mechanisms in (+)

RNA viruses

RdRp is one of the most essential proteins encoded by the genomes of (+) RNA viruses. It is responsible for the synthesis of both strands of viral RNA. In some extreme cases, such as fungus-infecting sclerodarnavirus (Hu et al. 2014), narnavirus (Hillman and Cai 2013), and mitovirus (Hillman et al. 2018), which spread between different host individuals through hypha fusion (known as anastomosis), RdRp is the only protein encoded on the (+) RNA genome. RdRp of most (+) RNA viruses lack proof-reading activities, meaning they incur relatively high numbers of replication errors. The current

5 estimates of RdRp-introduced mutation rates range from 3.5 × 10−6 to 4.8 ×

10−4 site/nucleotide/cycle (Sanjuán and Domingo-Calap 2016). It is speculated that the high mutation rates allow viruses to cope with environmental changes, and to counter or evade host antiviral defense responses. Many viruses also produce subgenomic messenger RNAs (sgRNAs). Their synthesis is through RdRp-mediated transcription.

The translation of ARP and RdRp occurs on ribosomes of host cells. Viral genomic

RNA frequently contains additional structural features that favor viral protein translation at the expense of host cell proteins. It has been shown that a sequence within TMV 5’ untranslated region (UTR) called the omega element enhanced the translation of reporter genes by 1- to 3.8- fold in vivo (Gallie and Kado 1989). This element was found to enhance the recruitment of a critical host translation initiation factor eIF4F (Gallie 2002).

Similar translation-enhancing elements were identified in UTRs of many viruses including Tobacco etch virus (Carrington and Freed 1990), turnip crinkle virus (Qu and

Morris 2000), and tobacco necrosis virus (Shen and Miller 2004), among others.

Viruses have evolved various strategies to optimize the expression of RdRp relative to other proteins, including ARPs. These strategies include embedding a stop codon within the RdRp ORF that is then overcome through programmed translational readthrough or frameshifting, implanting an RNA element within RdRp leading sequence that is then targeted by a viral encoded protein to suppress translation initiation.

Many viruses encode RdRp as C-terminal extensions of ARP, with the full-length

RdRp being translated at much lower levels, often less than 10% of ARP (Miras et al.

2017). This is achieved through stop codon readthrough or suppression of termination.

During canonical translation, the stop codon is decoded by a release factor that terminates the translation. However, in cases of readthrough, a stop codon (usually UAG or UAG) is recognized by an anticodon with a single nucleotide mismatch, thus suppressing the termination, allowing for continued translation until the next stop codon. Many viruses were shown, or bioinformatically predicted to encode an RdRp through readthrough of the cognate APR. These viruses are clustered in Grammaflexiviridae (Howitt et al. 2001),

Carmotetraviridae, Togaviridae, Virgaviridae, and Tombusviridae families (Firth and

Brierley 2012; Miras et al. 2017). The best-known example is the p126 readthrough producing p183 in TMV (Pelham 1978).

The readthrough rates vary among different viruses, about 10% in TMV (Beier et al.

1984), 6.4% in Sindbis virus, and 7.6% in Venezuelan equine encephalitis virus (Firth and Brierley 2012). However, these data were all derived from in vitro translation experiments. The results of these experiments may vary depending on experimental conditions. It has been reported that increasing the concentration of tRNATyr boosted

TMV readthrough rate by 3.5 times (Beier et al. 1984). Another example comes from

Providence virus (Walter et al. 2010). Providence virus can infect both plants and insects and replicates in mammalian cell lines (Jiwaji et al. 2019). The translational readthrough of its RdRP was dramatically influenced by the in vitro translation system used (Walter et al. 2010). In a rabbit reticulocyte extract system, the readthrough product dominated, while in a wheat germ extract system pre-readthrough product dominated (Walter et al.

2010). How the translational readthrough is regulated in vivo has yet to be thoroughly examined. It may be subjected to additional layers of temporal regulation. In many

7 viruses, RNA secondary structures and/or sequence motifs that interact with each other, often across long intervening sequences, were shown to influence the efficiency of translational readthrough (Sun et al. 2020).

Programmed ribosomal frameshift (frameshift in short) is another mechanism that translates RdRp as C-terminal extensions of ARP in many (+) RNA viruses. When a ribosome gets close to the stop codon of ARP, it slips back (or rarely forward) to bypass the stop codon and continues to translate an intact RdRp. Similar to readthrough, the resultant full-length RdRp is translated at much lower levels, often less than 20% of ARP

(Firth and Brierley 2012; Miras et al. 2017). Frameshift differs from readthrough in that, instead of using a cognate anticodon to continue translation, it prompts the ribosome to slip to another coding frame using a slippery sequence to avoid the in-frame ARP stop codon (Gao and Simon 2015). Efficient frameshift usually requires a 7-nt slippery motif and a downstream stable pseudoknot or stem-loop that are linked by a 5-9 nt spacer

(Moureau et al. 2015).

For viruses that encodes large polyprotein precursors, translational frameshifting has been found to serve the role of diverting a subset of ribosomes away from the viral genomic RNA before the complete genome is translated (Finch et al. 2015; Firth et al.

2010; Li et al. 2014). Since in these viruses the RdRp domain is frequently located within the C-terminus half of the polyprotein, such frameshift events may play the role of limiting the RdRp translation relative to other viral proteins (Li et al. 2014) (Napthine et al. 2019).

In summary, both programmed translational read-through and frameshifting serve the purpose of keeping the RdRp at relatively low levels. Indeed, such an outcome may be evolutionary advantageous to (+) RNA viruses, because additional innovative strategies that serve the same purpose have been identified from diverse (+) RNA viruses. For example, some viruses use a strategy known as polymerase slippage to synthesize a subset of mRNA that contain a single nt insertion or deletion. These mRNA cause the reading frame of the polyprotein to shift for one nt at the point of the insertion/deletion, leading to the termination of polyprotein translation before encountering the RdRp coding region (Atkins et al. 2016). Polymerase slippage has been observed in viruses of the family Potyviridae (White 2015). Aside from minimizing the RdRp translation, polymerase slippage also plays an important role in viral cell-to-cell movement by synthesizing the movement protein P3N-PIPO (Wei et al. 2010b).

Finally, RdRp translation in some viruses is additionally regulated by other proteins encoded by the same virus. MS2 and Qβ are (+) RNA viruses that infect bacteria (Cui et al. 2017). Unlike viruses that infect eukaryotes, their genomic RNAs are polycistronic.

The translation of RdRp is under the control of an RNA element of approximately 20-nt long, located immediately upstream of the RdRp coding region (Rumnieks and Tars

2014). This RNA element, referred to as an RNA operator, folds into a stem-loop structure and encompasses the start codon for RdRp translation (Gralla et al. 1974). The operator serves as a high-specificity binding site for the dimeric form of the capsid protein (CP) of the same virus, allowing CP to shut down RdRp translation completely.

As a result, CP of these viruses acts as a potent regulator of RdRp levels, maintaining

RdRp at a certain concentration threshold.

Similar regulatory circuits were also found in viruses of eukaryotes that use virus- encoded proteins to regulate translational frameshifting. Two well-characterized examples are encephalomyocarditis virus (EMCV) (Napthine et al. 2017) and Theiler's murine encephalomyelitis virus (TMEV) (Napthine et al. 2019). The slippery sequence that mediates translational frameshift in EMCV and TMCV is located within the 2B coding region, which is immediately downstream of 2A coding region. It was shown that the 2A protein binds to a stem-loop structure of the viral RNA located 13–14 nt downstream of slippery sequence. The RNA-protein interaction blocks in-frame translation to favor -1 frameshift, hence dramatically increases the frameshift rate.

Similar results were also shown in porcine reproductive and respiratory syndrome virus where nsp1β also activates the downstream frameshift within the nsp2 coding region by binding to viral RNA (Li et al. 2014).

Multipartite viruses may also use protein-RNA interaction to control the translation of their RdRps. Brome mosaic virus (BMV) is a tripartite (+) RNA virus. Its

RNA1, RNA2, and RNA3 encode the 1a replication accessary protein, 2a RdRp, and 3a

MP and 3b CP, respectively. Researchers used an experimental system to assess the replication efficiency of chimeric RNA1 and 2 in which the viral proteins (1a or 2a) were replaced with a green fluorescent protein (GFP) coding sequence (Yi et al. 2007).They found that when 1a was overexpressed, the expression of GFP from chimeric RNA1 and

RNA2, but not RNA3, were dramatically reduced. Sequence analysis reveals an RNA

10 secondary structure shared by RNA1 and RNA2, referred to as the B-Box. Deletion of B-

Box rescued the expression of GFP in this system (Yi et al. 2007). These results strongly indicate the interaction between B-Box and 1a inhibits the translation of 1a and 2a.

Viral subgenomic (sg) RNAs and long non-coding (lnc) RNAs

In addition to the full-length genomic RNA copies, many (+) RNA viruses also produce shorter versions of viral RNA that are usually not incorporated into virions.

These RNAs include sgRNAs that serve as mRNA for the translation of viral proteins with specific functions, but also various RNAs with no obvious coding capacities, known as long noncoding RNA (lncRNA). The sequence of sgRNAs are usually 3’ co-terminal with the genomic RNA of the same virus. This feature allows them to bring ORFs that are distal to 5’ end in genomic RNA (gRNA) closer to the 5’ end of their own mRNA, enhancing the translation of the cognate proteins. There is no doubt that sgRNAs are synthesized by RdRp (Sztuba-Solińska et al. 2011). Many mechanisms have been proposed to explain how sgRNA is produced in different viruses. For sgRNAs that only share the 3’-proximal sequence with the gRNA, two different models have been proposed: the internal initiation model, and the premature termination model (Sztuba-

Solińska et al. 2011). For sgRNAs that share both 3’ and 5’ sequences with the gRNA, there are the leader-primed transcription model and the discontinuous template synthesis model.

In the internal initiation model (Koev and Miller 2000), (-) gRNA containing an internal promoter is first synthesized. Then the internal promoter is recognized by RdRp

11 to initiate (+) sgRNA synthesis. This model satisfactorily explains the sgRNA synthesis in TMV, BMV, barley yellow dwarf virus (BYDV), and Sindbis virus. In contrast, the premature termination model (White 2002) relies on a stop signal in the (+) gRNA to terminate the replication of a fraction of (-) gRNA, resulting in the production of (-) sgRNA. Then the (-) sgRNA is used to template (+) sgRNA synthesis. The key difference between these two models is whether a (-) sgRNA intermediate is used for the synthesis of the (+) sgRNAs. The stop signal usually involves long distance RNA-RNA interactions within the viral genome. For example, the genomic RNA of Tomato bushy stunt virus (TBSV) has the capacity to form at least three long-distance base pair interactions that form the physical barriers for the synthesis of (-) sgRNAs of different lengths (Choi et al. 2001; Choi and White 2002; Lin and White 2004). This kind of intramolecular long-distance interaction were also described in Flock house virus

(Lindenbach et al. 2002) and Potato virus X (Kim and Hemenway 1999). Interestingly, the bipartite red clover necrotic mosaic virus (RCNMV) utilizes a unique intermolecular interaction to produce a stop signal. The synthesis of a (-) sgRNA from RCNMV gRNA1 requires base pairing between gRNA1 and gRNA2 (Guenther et al. 2004).

In addition to sgRNAs, many viruses have now been found to produce a class of

RNAs that are also 3’ coterminal to viral genomes but are unlikely to encode any meaningful proteins due to their short sizes. They are typically between 150 and 400 nt in length, with their 5’ ends rarely extending beyond the 3’ UTRs of viral genomes

(Gunawardene et al. 2019). They are nevertheless designated as long noncoding RNAs

(lncRNAs) to distinguish them from viral small interfering RNAs (vsiRNAs) that are

12 typically <30 nt. Viral lncRNAs have attracted a surge of interest of late, due to their perceived roles in viral pathogenesis. Many viral lncRNAs are thought to counteract the host defense responses that target viral genomes, such as RNA decay mediated by the exoribonuclease Xrn1 of host cells (Moon et al. 2012), as well as RNA silencing (Moon et al. 2015). They may also play a role in repressing the type I interferon-mediated antiviral defense in animal virus-infected cells (Chang et al. 2013), and possibly also the

Toll antiviral pathway in insect vectors (Pompon et al. 2017). Some lncRNAs have also been found to be involved in viral replication, dissemination, and host adaption (Slonchak and Khromykh 2018).

Many animal and human-pathogenic flaviviruses are known to produce lncRNAs.

These include Murray Valley encephalitis virus (MVEV), dengue virus (DENV), West

Nile virus (WNV), and Zika virus (Lin et al. 2004; Scherbik et al. 2006; Urosevic et al.

1997). Similarly, a number of plant viruses have been found to produce lncRNAs as well.

Examples include lncRNAs associated with BYDV (Kelly et al. 1994; Koev and Miller

2000), Soybean dwarf virus (Yamagishi et al. 2003), RCNMV, maize chlorotic mottle machlomovirus (Scheets 2000), tobacco necrosis virus D (TNV-D) (Tombusviridae), and beet necrotic yellow vein virus (BNYVV, Benyvirus). Two unique 5’ co-terminal lncRNAs referred to as low-molecular-weight tristeza 1 and 2 (LMT1, LMT2) have also been identified in citrus tristeza virus (CTV)-infected host plants (Che et al. 2001; Kang et al. 2019).

The biogenesis of flavivirus-associated lncRNAs has been characterized in WNV

(Pijlman et al. 2008), yellow fever virus (Silva et al. 2010), DENV (Chapman et al.

2014), and Zika (Akiyama et al. 2016). Separately, lncRNAs encoded by BNYVV

(Dilweg et al. 2019), RCNMV (Iwakawa et al. 2008), and TNV-D (Gunawardene et al.

2019) have also been thoroughly investigated. These lncRNAs all resulted from the 5’ to

3’ processing of longer viral RNAs (e.g. genomic or subgenomic RNAs) by host 5’ exonucleases such as XrnI, frequently occurring in the processing bodies present in the host cell cytoplasm. Processing bodies are ribonucleoprotein complexes consisting of ribosome-free mRNAs, nonsense RNAs, microRNAs and more than 125 host proteins, including Xrn1in animals and yeast or Xrn4 in plants that degrades uncapped mRNAs.

Xrn1/Xrn4-mediated degradation plays an important antiviral role since many RNA virus genomes are not capped.

Even though flavivirus genomes are capped at their 5’ end, they readily produce lncRNAs (termed sfRNA, short for subgenomic flaviviral RNA) through Xrn1 degradation in infected cells. It was speculated that flaviviral RNA genomes are first decapped by decapping enzymes in the processing bodies before they are exposed to

Xrn1 processing (Pijlman et al. 2008). In support of the involvement of Xrn1 in WNV sfRNA biogenesis, its production is decreased in Xrn1-depleted cells (Pijlman et al.

2008). Xrn1 degradation alone is insufficient for sfRNA biogenesis, as at least one stable

RNA secondary structure is needed to stall Xrn1-mediated decay. Such Xrn1-resistant structures are abundantly available in the 3’ UTR of flaviviruses. Some flaviviruses with multiple stable RNA secondary structures were found to produce up to four distinct sfRNAs (Slonchak and Khromykh 2018).

Available evidences strongly suggest that most plant virus-related lncRNAs are likewise products of 5’ to 3’ exonuclease degradation. For example, the BYNVV- associated noncoding sgRNA3 and its variants lack the 5’ cap present in the viral genomic RNA, a characteristic cited as evidence for their origination from processes other than replication or transcription (Peltier et al. 2012). On the other hand, the genomes of RCNMV and TNV-D are both uncapped, making them potentially vulnerable to Xrn4 degradation. Like flaviviruses, nuclease-resistant RNA structures were identified in both viruses. Unlike flaviviruses, protein-coding sgRNAs of these viruses may also serve as substrates for Xrn4 decay. Another interesting case is the biogenesis of the

BYDV-associated lncRNA. In addition to the lncRNA, BYDV also produces two sgRNAs. The 5’ sequence of gRNA and sgRNA are highly conserved, but such conserved sequence is absent at the 5’ end of the lncRNA (Kelly et al. 1994; Zavriev et al. 1996). This suggests that BYDV lncRNA is not the product of viral replication.

Despite their apparent resistance to Xrn1/Xrn4-mediated decay, it is not known whether some lncRNAs could also be generated by the viral replication process. This is an important question given the pro-virus role of these lncRNAs. Indeed there is preliminary evidence to suggest that the RCNMV-associated lncRNA SR1f might also be produced through viral replication (Iwakawa et al. 2008; Newburn and White 2020), as

(-) RNA complementary to SR1f has been detected with Northern blotting.

TCV as a unique model for studying the effect of RdRp overexpression

Many (+) RNA viruses have been used as models to study (+) RNA virus replication cycles. These model viruses include bacteriophage Qβ, the tripartite plant-infecting brome mosaic virus (BMV), and the human-infecting poliovirus (PV). Studying of these model viruses deepened our understanding of the replication of (+) RNA viruses, and potentially provide us new means to prevent and treat viral diseases more efficiently.

The model virus we are using is turnip crinkle virus (TCV). TCV is a small (+) RNA plant virus that counts model plants Arabidopsis and Nicotiana benthamiana as hosts.

Infection of TCV can be easily achieved by rub-inoculating plants with the sap of infected-plants or in vitro-transcribed TCV genomic RNA. Alternatively, a more common method we use in our lab is to deliver a DNA infectious clone through

Agrobacterium infiltration (agro-infiltration). This method allows us to manipulate the

TCV genome easily. The TCV genome is a nonsegmented, single-stranded (+) RNA of

4,054 nt. It encodes five proteins p28, p88, p8, p9 and p38 (Fig 2.1A). Among them, p28 and p88 are directly translated from the genomic RNA and serve as ARP and RdRp. P8 and p9 are two small movement proteins translated from the positive sense of subgenomic RNA1, which are produced during virus replication. Subgenomic RNA2 encodes p38 which is the TCV capsid protein and also serves as an RNA silencing suppressor.

Only p28 and p88 are needed for TCV replication in infected cells. Suppressing host

RNA silencing machinery by TCV CP or a heterologous silencing suppressor bolsters

TCV replication by preventing RNA degradation. P88 is encoded through the programmed readthrough of the p28 stop codon. The readthrough is mediated by a stem-

16 loop structure immediately downstream of the stop codon, and a long-distance RNA kissing-loop interaction (Cimino et al. 2011; Sun et al. 2020). The readthrough efficiency of p88 in an in vivo assay is ~1% relative to p28 (White et al. 1995). Previous studies in our lab found that overexpression of p28 in the form of a fusion protein with C-terminal

GFP or double HA (2HA) tags completely shut down the replication of TCV in the same cells (Zhang et al. 2017). A follow-up study pinpointed a critical region in p28 (from amino acid residue 171 to 182) that are essential for p28 to repress TCV replication (Guo et al. 2020).

Since p88 contains the entire p28 at its N terminus, and p88 expression during TCV replication is tightly regulated by a programmed translational readthrough event, an interesting question is whether p88 overexpression would repress TCV replication in a manner similar to p28 overexpression. We are further interested in learning whether p88 overexpression would perturb the relative accumulation levels of different TCV RNAs -- genomic RNA, the two subgenomic RNAs, and possibly TCV-borne lncRNA. Another related question is whether p88 also plays a role in regulating the replication process in addition to its role as RdRp. These are the questions that I strived to address with my thesis research.

Research Hypothesis and Objective

Previous studies with TCV and other (+) RNA viruses strongly suggest that the expression level of RdRp in (+) RNA viruses is tightly regulated and kept at a relatively low level. Perturbing the RdRp expression level should have detrimental effects on viral

17 replication. Based on these observations, I hypothesized that overexpression of the p88

RdRp encoded by TCV compromises TCV replication and disrupts the relative accumulation levels of different TCV RNA species. To rigorously test this central hypothesis, I pursued the following specific objectives:

Objective 1: Determine how overexpression of p88 and its various truncated forms impact the replication of two different TCV replicons.

Objective 2: Determine whether p88 overexpression perturbs the relative levels of

TCV genomic and subgenomic RNAs and exposes potential TCV-associated lncRNAs.

Objective 3: Characterize a TCV lncRNA unmasked by p88 overexpression.

Significance of my thesis research

My thesis research, and the research results I obtained, have important implications in several aspects of the viral replication cycle. First, they contribute to our basic understanding of why the expression of viral replication proteins has to be tightly regulated in order for viruses to achieve maximal success in host cells. I showed that, contrary to the expectation, p88 overexpression severely diminished the replication of

TCV genomic RNA. This suggests that mechanisms such as programmed translational readthrough or frameshifting are evolutionarily selected to control the accumulation of

RdRp relative to other replication proteins (e.g. ARP). This insight suggests a novel strategy to control (+) RNA infections, by overexpressing viral RdRp.

Second, the identification and characterization of a TCV-associated lncRNA, which we call ttsgRNA, uncovers the first viral lncRNA that is produced through a replication-

18 dependent path. This is in contrast to all of the previously reported viral lncRNAs, which are all produced through exonuclease-mediated degradation of longer viral RNAs. More importantly, characterization of the sequence and structural requirements of ttsgRNA biosynthesis allowed me to reveal novel mechanisms through which viral sgRNA can be generated.

Finally, my findings could offer potential guidance to the control and management of some of the more devastating viral diseases of crop plants. Although TCV is not associated with substantial economic losses, it represents a large family of viruses that includes maize chlorotic mottle virus, which is one of the causal agents of maize lethal necrosis, a disease of the staple crop maize that has been responsible for greater than 70% losses of maize production in Kenya, Uganda, Tanzania, and other eastern African countries (Redinbaugh and Stewart 2018). Knowledge learned through my research could offer tangible solutions to this destructive crop disease. Sin combination of screening cheap chemical compounds that could stabilize RdRp could potentially provide a cure for those devastating virus diseases.

CHAPTER 2.

Repression of Turnip Crinkle Virus Replication by Its Replication Protein p88

Abstract

We recently reported that p28, one of the two turnip crinkle virus (TCV) replication proteins, trans-complemented a defective TCV lacking p28, yet repressed the replication of another TCV replicon encoding wild-type p28 (Zhang et al. 2017). Here we show that p88, the TCV-encoded RNA-dependent RNA polymerase, likewise trans-complemented a p88-defective TCV replicon, but repressed one encoding wild-type p88. Surprisingly, lowering p88 protein levels enhanced trans-complementation, but weakened repression.

Repression by p88 was not simply due to protein over-expression, as deletion mutants missing 127 or 224 N-terminal amino acids accumulated to higher levels but were poor repressors. Finally, both trans-complementation and repression by p88 were accompanied by preferential accumulation of subgenomic RNA2, and a novel class of small TCV RNAs. Our results suggest that repression of TCV replication by p88 may manifest a viral mechanism that regulates the ratio of genomic and subgenomic RNAs based on p88 abundance.

Keywords

RNA virus, plant virus, replication, repression by replication protein

Introduction

Viruses with single-stranded, positive sense (+) RNA genomes are the most common viral pathogens of plants, animals, and humans. Examples of important (+) RNA viruses include the plant-infecting tobacco mosaic virus, animal-infecting foot and mouth disease virus, and human pathogens poliovirus and Zika virus. The genomes of most (+) RNA viruses are relatively small (<15,000 nucleotides [nt]), encoding a limited number of proteins that are frequently multifunctional. A recently revealed example is the p28 protein encoded by the plant-infecting turnip crinkle virus (TCV), which exhibited opposite functions depending on protein concentration and terminal modifications (Zhang et al. 2017).

TCV is a small (+) RNA plant virus that counts model plants Arabidopsis and

Nicotiana benthamiana as hosts (Cao et al. 2010; Kuhlmann et al. 2016; Zhang et al.

2012). Its genome of 4,054 nt encodes five proteins (Fig. 2.1A). The 5′ proximal p28 and its read-through product (p88), both needed for genome replication, are translated directly from the genomic RNA (gRNA). The p8 and p9 movement proteins (MPs), and p38 capsid protein (CP), are translated from sub-genomic RNAs (sgRNA1 and 2) produced during viral replication (Fig. 2.1A). We recently reported that the auxiliary replication protein p28, upon C-terminal fusion with a double HA (2HA) tag, or a fluorescent protein

(GFP or mCherry), completely shut down the replication of TCV in the same cells

(Zhang et al. 2017). We further demonstrated that p28 free of any terminal modifications also potently repressed the replication of a TCV replicon, and p28 mutants with

21 weakened repressive activity were more competent at trans-complementing the replication of a p28-defective TCV replicon (Zhang et al. 2017).

In the current study, we examined p88, the TCV-encoded RNA-dependent RNA polymerase (RdRp), for its ability to trans-complement the replication of TCV mutants in which the p88 coding region was disrupted, and to repress the replication of TCV replicons that encode functional p88 of their own. Our results showed that TCV p88 exhibited concentration-dependent repression of TCV replication, and that this repressive activity inversely correlated with the ability of the same protein to trans-complement the replication of a p88-defective TCV mutant. Interestingly, the repressive activity of p88 preferentially suppressed TCV gRNA amplification but favored sgRNA accumulation.

These findings suggest that p88 may play a role in regulating the relative ratio of gRNA and sgRNAs through concentration-dependent changes in its function.

Results

p88 expressed independently of replication trans-complements the replication of p88-defective TCV replicons. In a previous study, a defective TCV replicon was generated by changing the p28 stop codon to UAC (White et al. 1995). As a result, genomic RNA of that defective replicon could translate only p88 but not the smaller p28, thus was unable to launch replication on its own (White et al. 1995). Nevertheless, p88 produced by that defective replicon trans-complemented the replication of a reciprocal mutant encoding only p28. However, it remains to be determined whether p88 expressed

22 from a non-replicating mRNA could complement TCV replication in trans. To address this question, we generated a p88-expressing binary construct, designated 2X35S::p88-

2HA (Fig. 2.1B). The p88 expression from this construct is driven by the strong cauliflower mosaic virus 35S promoter with duplicated enhancers (2X35S). In addition, a double HA epitope tag (2HA) was added to the C-terminus of p88 to facilitate the detection of p88 protein (Fig. 2.1B).

To test whether p88 expressed from this non-replicating construct trans-complements the replication of TCV, the construct was delivered into leaf cells of Nicotiana benthamiana plants along with Δp88_sg2R, a mutant TCV replicon in which the p88 coding sequence was disrupted by a four nucleotide (nt) deletion (Fig. 2.1B). Note that

Δp88_sg2R and the other two TCV replicon constructs used in the current study all harbored the 2X35S promoter that launches the transcription of first batch of TCV gRNA. Also, in all experiments described in the current study, a construct expressing the p19 silencing suppressor of tomato bushy stunt virus (TBSV) was always included to mitigate RNA silencing-mediated transcript degradation (not shown). As shown in Fig.

2.1C (top panel, lanes 3 and 7), replication of the defective Δp88_sg2R replicon was partially restored by p88-2HA. Nevertheless, the complementation was inefficient, and favored accumulation of TCV sgRNAs, especially sgRNA2, and a class of smaller TCV

RNA of unknown size(s) and origin, which we call tiny TCV sgRNA (ttsgRNA) hereafter.

We next assessed whether p88-2HA, together with a tag-free p28 that was also expressed from a non-replicating construct, could complement the replication of

[p28stop]_sg2R, a TCV replicon encoding neither p28 nor p88 (Fig. 2.1B). By comparing lanes 4, 8, 9, and 10 in Fig. 2.1C, we concluded that both p88-2HA and p28 were needed to achieve a partial complementation of [p28stop]_sg2R. Finally, the expression of p88-

2HA protein, as well as the replication-dependent production of mCherry protein, were confirmed with Western blotting, using HA and mCherry antibodies, respectively (Fig.

2.1C, bottom three panels). Therefore, p88-2HA provided in trans was partially replication-competent. It preferentially replicated sgRNAs but not genomic RNA.

p88-2HA represses the replication of a TCV replicon encoding intact p88. We wondered why the complementation by p88-2HA was not as efficient as self-replication of TCV_sg2R (Fig. 2.1C). Since the transiently expressed p88-2HA was expected to accumulate at much higher levels than p88 translated from TCV replicons, we reasoned that the over-expressed p88-2HA protein could partition into two states, with the first supporting replication, but the second repressing replication. This possibility was attractive to us because (i), the same p88-2HA protein partially repressed the replication of a GFP-expressing TCV replicon in a previous study (Zhang et al. 2017); and (ii), p28-

2HA, the 2HA-tagged form of the smaller replication protein p28, completely repressed

TCV replication (Zhang et al. 2017). Note that p28 constitutes the N-terminal 1/3 of p88

(Fig. 2.1A).

One prediction of this possibility is that lowering intracellular p88-2HA levels would discourage the genesis of the repressive p88-HA state, leaving more p88-2HA in the replication-competent state. To test this prediction, we managed to decrease the p88-2HA protein levels by removing both enhancers from the 2X35S promoter, yielding the

24 construct Core35S::p88-2HA (Fig. 2.2A). As shown in Fig. 2.2B, the level of p88-2HA protein was substantially reduced in cells receiving this new construct (compare lanes 2 and 3), and it remained low in the presence of TCV replicons (compare lanes 6, 7, 8 and

9). Consistent with our prediction, decreased p88-2HA expression from the

Core35S::p88-2HA construct led to improved trans-complementation of the p88- defective replicon (Fig. 2.2C, compare lanes 14-16 and 17-19. Note that the triplicate lanes 5-7, 8-10, 11-13, 14-16, and 17-19, represent three independent samples).

Both the 2X35S::p88-2HA and Core35S::p88-2HA constructs were then used to test whether p88-2HA repressed the replication of a TCV replicon encoding its own p88. As shown in Fig. 2.2C (lanes 8-10), higher p88-2HA protein level caused a dramatic reduction in both gRNA and sgRNA levels of TCV_sg2R, suggesting that this level of p88-2HA strongly repressed the replication-dependent synthesis of both gRNA and sgRNAs. Strikingly, this repression was greatly reduced at the lower p88-2HA protein level (Fig. 2.2B, lane 7, and Fig. 2.2C, lanes 11-13). Together the data in Fig. 2.2 demonstrated that (i), p88-2HA both complemented and repressed TCV replication in a concentration dependent manner; and (ii) higher p88-2HA protein levels correlated with lower complementation but higher repression. Therefore, the limited complementation by p88-2HA is likely caused by the intrinsic repressive activity of p88. Importantly, both trans-complementation and repression by p88-2HA led to a preferential accumulation of sgRNA2 and ttsgRNA (lanes 8-19). This selective amplification is particularly strong at lower p88-2HA levels (lanes 11-13, 17-19), suggesting that repression by p88-2HA preferentially targeted the synthesis of TCV gRNA.

The C-terminal 2HA tag does not appreciably affect p88 activities. In addition to being transiently overexpressed, p88-2HA also differs from the virus-encoded p88 by having a C-terminal 2HA tag. We thus needed to assess whether this tag compromised the function of p88, rendering it more potent at repression yet less competent at trans- complementation. This assessment is of particular importance in light of a previous report showing that activity of the RdRp of a related virus (p92 of TBSV) is strongly affected by changes within the last five amino acid (aa) residues (Wu and White 2007). To this end, we generated a new construct to express p88 free of any terminal tags, under control of the Core35S promoter (Core35S::p88. Fig. 2.3). Northern blotting results in Fig. 2.3 showed that, compared to p88-HA under control of the same promoter (Core35S::p88-

2HA), the tag-free p88 was actually slightly less competent at complementing the replication of Δp88_sg2R (lane 10 versus 11), but exhibited a similar repressive activity toward TCV_sg2R (lane 6 versus 7). Therefore, the 2HA tagged p88-2HA was functionally analogous to the tag-free p88, hence can be used as the surrogate of the latter.

Trans-complementation by p88-2HA tolerates limited N-terminal truncation. We recently reported that supplying p28 alone restored the replication of a mutant TCV replicon in which the translation of both p28 and p88 was disrupted by an early frameshift (fs) mutation (Zhang et al. 2017. Also see Fig. 2.4A). We speculated that translation of p88 (and p28) could reinitiate from the second in-frame start codon 36 amino acid (aa) downstream, and that the resulting N-terminally truncated p88 (but not p28) could still be functional. Another study by Rajendran and colleagues (2002) found

26 that removal of the entire p28-coding region from p88 enhanced the replication function of p88 in vitro. These findings prompted us to assess whether trans-complementation by p88-2HA in vivo required p88 to be full-length. To this end, we generated a series of N- terminal deletions within the p88 coding sequence (Fig. 2.4A). The first mutant,

2X35S::p88fs-2HA, contained the same frame-shift mutation described above, but in the

2X35S::p88-2HA backbone. All other mutants were based on the Core35S::p88-2HA backbone (Fig. 2.4A). ΔN36, ΔN127, and ΔN224 removed the N-terminal regions upstream of the three internal methionine residues at positions 37, 128, and 225, respectively, whereas Δ250 removed the entire p28 sequence.

As expected, both 2X35S::p88fs-2HA and Core35S::p88ΔN36-2HA produced a protein of the same size, which is slightly smaller than the full-length p88-2HA (Fig.

2.4B, top panel, lanes 2-4; also lanes 10-12). Nevertheless, the former (lane 2) accumulated to modestly higher levels, possibly due to the strong 2X35S promoter.

Surprisingly, ΔN127 and ΔN224 deletion mutants accumulated to very high levels despite the weak Core35S promoter (lanes 5, 6, 13, 14). Finally, the ΔN250 mutant lacking the entire p28 portion accumulated to a level slightly higher or comparable to the full-length p88-2HA driven by Core35S promoter (compare lanes 7 and 15 with lanes 3 and 11, respectively).

Trans-complementation was observed only with the p88ΔN36-2HA mutant, expressed from the 2X35S::p88fs-2HA and Core35S::p88ΔN36-2HA constructs (Fig.

2.3B, bottom panel, lanes 10 and 12). Compared to the full-length p88-2HA, the ΔN36 mutant protein was less efficient at complementing the synthesis of gRNA (lanes 10-12. *

27 in lane 11 denotes a weak gRNA signal detectable only in the presence of full-length p88-2HA). Interestingly, this mutant protein was substantially more efficient than p88-

2HA at facilitating the accumulation of sgRNA2 and ttsgRNA (lanes 10-12), suggesting that the N-terminal 36 aa region is critical for gRNA synthesis activity of TCV RdRp.

None of the three mutants with larger deletions (ΔN127, ΔN224, and ΔN250) were capable of trans-complementation (lanes 13-15).

The inability to detect TCV gRNA in p88ΔN36-2HA-mediated complementation prompted the question of whether gRNA synthesis by p88 was completely abolished by the ΔN36 deletion. To test this, we attempted to detect the gRNA-specific, negative-sense

(-) RNA using a more sensitive semi-quantitative reverse transcription (RT)-PCR procedure (Plaskon et al. 2009). As shown in Fig. 2.4C, p88ΔN36-2HA (lane 3), but not p88ΔN250-2HA (lane 4), facilitated the accumulation of a very low level of (-) gRNA.

This experiment also confirmed that the full-length p88-2HA more efficiently complemented gRNA synthesis (compare lanes 2 and 3). Together these results indicated that deletion of N-terminal 36 aa of p88 diminished, but did not abolish, gRNA accumulation; yet strongly stimulated the synthesis of sgRNA2 and ttsgRNA.

p88 mutants capable of trans-complementation are also strong repressors of replication. We next examined the deletion mutants for their ability to repress the replication of TCV_sg2R. Interestingly, the same two mutants (2X35S::p88fs-2HA and

Core35S::p88ΔN36-2HA) that exhibited partial trans-complementation (see above) also exerted a strong repression on TCV_sg2R replication (Fig. 2.5, top panel, lanes 10 and

12). Strikingly, both mutant proteins severely repressed the accumulation of gRNA, but

28 actually enhanced the accumulation of sgRNA2 and ttsgRNA, leading to an RNA accumulation profile extremely similar to that of trans-complementation (compare Figs.

2.4 and 2.5). A similar, but not identical, RNA accumulation profile was also observed when trans-complementation and repression were achieved using low levels of full- length p88-2HA (Core35S::p88-2HA; Figs. 2.2, 2.3, 2.4, and 2.5). Furthermore, the repression is unlikely due to non-specific protein over-expression, as both ΔN127 and

ΔN224 mutant proteins accumulated to greatly elevated levels but exerted weak repression on TCV_sg2R replication that did not cause the preferential accumulation of sgRNA2 and ttsgRNA (lanes 13 and 14). On the other hand, the ΔN250 mutant, despite relatively low accumulation, caused a near complete blockage of TCV_sg2R replication, affecting all of the TCV RNA species (lane 15).

Replicational repression by p88 mutants partially correlates with the induction of p28 inclusions. How did p88-2HA and some of its deletion mutants repress the replication of TCV_sg2R? Since TCV p28-2HA repressed TCV replication as well

(Zhang et al., 2017), we wondered if they shared the same repression mechanism. p28 with a C-terminal GFP tag (p28-GFP) displayed strong repressive activity, and formed large, intense intracellular inclusions (Zhang et al., 2017). By contrast, p28 with an N- terminal G11 (the 11th and last β-strand of GFP) tag, referred to as G11-p28, abolished the repressive activity of p28, but enhanced its trans-complementation activity (Zhang et al., 2017). Consistent with the role of p28 inclusions in replicational repression, G11-p28, when co-expressed with the first ten β-strands of GFP (G1-10), failed to form any inclusions, and was instead diffusely distributed throughout the cytoplasm (Zhang et al.,

2017. Also see Fig. 2.6A). This (G11-p28 + G1-10) combination was further used as a reporter system to demonstrate trans-aggregation by p28, as the presence of p28 inclusions in the same cell pulled G11-p28 into the inclusions, changing the distribution pattern of the latter (Zhang et al., 2017).

We next used this reporter system to assess whether p88-2HA and its deletion mutants could trans-aggregate G11-p28. As shown in Fig. 2.6A, simultaneous expression of G11-p28 and G1-10 led to evenly distributed diffuse green fluorescence in N. benthamiana cells. Note that the cell nuclei also fluoresced but very faintly (white arrows). As expected, adding p28-HA or untagged p28 caused the majority of green fluorescence to congregate into very bright inclusion dots (Fig. 2.6B, panels 1 and 2). The remaining G11-p28 also formed tiny dots that lined the cell boundaries discontinuously.

Strikingly, high level of p88-2HA expression driven by the 2X35S promoter led to the formation of even larger inclusions, accompanied by a near complete loss of diffuse fluorescence (Fig. 2.6B, panel 3), indicating that this level of p88-2HA caused strong protein aggregation, and the consequent inclusions became magnets to coalesce nearly all diffuse G11-p28. By contrast, lower p88-2HA expression driven by the Core35S promoter still induced the formation of G11-p28 inclusions, though also left a fraction of

G11-p28 in the diffuse state (Fig. 2.6B, panel 4). Therefore, the full-length p88-2HA likely self-associated in a concentration-dependent manner to form protein aggregates, which in turn trans-aggregated G11-p28.

The p88ΔN36-2HA protein still induced G11-p28 aggregation in some cells (Fig.

2.6B, panel 5). However, it was clearly less capable of doing so than the full-length p88-

2HA, as more G11-p28 fluorescence remained diffused. This trend was even more apparent with cells entered by the three mutants harboring larger N-terminal deletions, where most of the G11-p28 green fluorescence remained in the diffused form (Fig. 2.6B, panel 6-8). Note that the Δ127 and Δ224 mutants repressed TCV replication only marginally, whereas ΔN250 was the strongest repressor of all (Fig. 2.5). Therefore, while the ability to trap G11-p28 in protein aggregates could explain the repressive power of some mutants, the same explanation did not apply to the p88ΔN250-2HA mutant.

p88 with a C-terminal GFP tag forms large inclusions similar to p28-GFP. To directly address the question of whether transiently overexpressed p88 forms large protein aggregates by itself, we then tagged the full-length p88, as well as the deletion mutants described earlier, with a C-terminal GFP tag. Confocal images in Fig. 2.6C showed that while free GFP exhibited a diffused distribution in agro-infiltrated cells

(panel 1), the p88-GFP fusion protein, regardless of the strength of the promoters used, formed large, irregular inclusions that are strikingly similar to those of p28-GFP (Fig.

2.6C, panels 2-4). Interestingly, the GFP-tagged ΔN36, ΔN127, and ΔN224 mutants failed to form large inclusions of p88-GFP. Instead, they formed substantially smaller dots that evenly aligned the cell boundaries (Fig. 2.6C, panels 5-7). Finally, ΔN250-GFP protein showed a predominantly diffused distribution (panel 8). Overall, these results are consistent with a role of protein aggregation in the repression of TCV replication by transiently expressed p88 and some of its mutants. However, it is important to note that protein aggregation does not explain the preferential accumulation of sgRNA2 and

31 ttsgRNA – an observation best explained by the potential RdRp activity of overexpressed p88.

Discussion

The goals of the current study were to determine whether the TCV-encoded p88 replication protein, when expressed from a non-replicating mRNA, trans-complements the replication of TCV mutants lacking p88-coding capacity, and whether it interferes with the replication of TCV replicons encoding an intact p88. We recently reported that p28, the smaller TCV replication protein, trans-complemented the replication of a TCV replicon lacking its own p28 but repressed the replication of another replicon encoding a functional p28 (Zhang et al., 2017). We further established that p28 formed large, intense intracellular inclusions that impeded trans-complementation, but correlated with repression. However, it is not known whether these properties are also shared by p88, the

TCV-encoded RdRp.

To resolve these questions, replication-independent expression of p88 and its deletion mutants were administered in N. benthamiana cells, along with two different TCV replicons differing in their capacity to produce cognate p88. Based on their ability to trans-complement and/or repress TCV replication, these p88 variants can be sorted into four categories. First, the p88ΔN250-2HA mutant, in which the entire p28 portion of p88 was deleted, failed to trans-complement the replication of the p88-defective Δp88_sg2R replicon, but almost completely repressed the replication of the p88-producing

TCV_sg2R replicon. The inability of this protein to substitute for p88 in N. benthamiana cells contrasts with the previous in vitro results where a C-terminal tagged p88ΔN250 being more potent than the full-length p88 in replication (Rajendran et al. 2002), suggesting that TCV replication in vivo may entail additional requirements. Conversely, this protein may also use a distinct mechanism to repress TCV replication, as it did not drastically alter the diffuse distribution of G11-p28 in most cells, and its GFP-fused form likewise exhibited a diffuse distribution.

The second category of p88 mutants includes p88ΔN127-2HA and p88ΔN224-2HA, neither was capable of trans-complementation. Strikingly, both were only marginally repressive, despite very high levels of protein accumulation. This is particularly noteworthy, as the p88ΔN224-2HA protein was only 26-aa longer than the highly repressive p88ΔN250-2HA. Yet this short N-terminal extension seemed to have dramatically stabilized the protein while simultaneously diminished its repressive power.

The simplest explanation would be that the 26-aa region (aa #225-250) shielded a repressive domain within p88ΔN250-2HA, probably through intra-molecular interactions, preventing the latter from engaging in inter-molecular interactions with full-length p88 produced from the TCV_sg2R replicon. The exact nature of these interactions will be investigated in future studies.

The third category consists of the p88ΔN36-2HA protein translated from two different constructs (2X35S::p88fs-2HA and Core35S::p88ΔN36-2HA), and the full- length p88-2HA protein at lower concentration (the Core35S::p88-2HA construct). Due to suboptimal construct configurations (translational re-initiation in the “fs” construct, or

33 the weak Core35S promoter in the other two constructs), these p88 variants all accumulated relatively low levels of proteins. Nevertheless, they all partially trans- complemented the replication of the p88-defective replicon, leading to a unique TCV

RNA profile with barely detectable levels of gRNA, but highly abundant sgRNA2 that rivaled or exceeded the control TCV_sg2R replicon, and the accumulation of the new ttsgRNA (Fig. 2.4). Interestingly, these three p88 forms also repressed the replication of

TCV_sg2R to similar extents and giving rise to a similar RNA profile of diminished gRNA accumulation, but elevated sgRNA accumulation (Fig. 2.5). This latter observation strongly suggests that these p88 proteins did not strictly repress TCV_sg2R replication, but rather preferentially enhanced the amplification of sgRNAs.

What would happen if the level of p88 protein were further reduced? This is an important question because in actual TCV infections, the p88 accumulation level is expected to be much lower still, as the read-through of the p28 stop codon is very inefficient. It seems reasonable to speculate that the very low p88 level at the early stage of TCV infection might be evolutionarily selected to ensure a sufficient amount of gRNA is synthesized in order to sustain virus infections. Conversely, a slight increase at a later stage of intracellular infection might then permit the accumulation of sgRNAs, facilitating virion assembly and intercellular spread. It could thus be inferred that during

TCV replications, change in p88 concentration might exert a regulative role by adjusting the ratio of gRNA and sgRNAs according to the progression of infection.

Finally, the highly expressed p88-2HA (from the 2X35S::p88-2HA) constitutes a category of its own as it exerted a very strong repression on TCV_sg2R that diminished

34 the accumulation of both gRNA and sgRNA levels. By contrast, it was much less capable of trans-complementing Δp88_sg2R than its Core35S-driven counterpart, despite a substantially higher protein level. We reason that the very high level of p88-2HA could precipitate a repressive state in a manner similar to p28 (Zhang et al. 2017), by coalescing into large intracellular inclusions that could in turn trans-aggregate p28/p88 produced by the TCV_sg2R replicon, as well as p28 produced by Δp88_sg2R. Consistent with this idea, trans-aggregation of G11-p28 was much more strongly induced by 2X35S-driven p88-2HA than by Core35S-driven p88-2HA (Fig. 2.6).

In summary, this study confirmed that p88, as well as a minimally truncated form of p88 (ΔN36), was capable of trans-complementing the replication of a p88-defective TCV mutant in a relatively inefficient manner that favors sgRNAs over gRNA. Lowering the expression levels of p88 caused a modest increase of complementation efficiency, revealing an inverse relationship between trans-complementation and p88 concentration.

We further demonstrated that p88 and its various variants also repressed TCV replication by primarily targeting the accumulation of TCV gRNA, and this repressive activity positively correlated with p88 protein concentration. These results led to a hypothesis that p88 exerts a concentration-dependent control of gRNA/sgRNA ratio during the TCV replication cycle. This hypothesis will be tested with additional follow-up research.

Materials and methods

Constructs. Constructs 2X35S::p88-2HA, TCV_sg2R and Δp88_sg2R were described in previous studies (Zhang et al. 2015; Zhang et al. 2017). Core35S::p88-2HA was generated by deleting the duplicated enhancer sequence from2X35S::p88-2HA, retaining only the 99-nt core promoter. The [p28stop]_sg2R mutant replicon construct was modified from TCV_sg2R by changing the methionine-coding AUG at the aa position 36 of p28 to UAG, thus introducing an in-frame stop codon in both p28 and p88 coding sequence. All of the deletion mutants of p88-2HA were generated on the

Core35S::p88-2HA backbone. The identity of all new constructs was verified with

Sanger sequencing.

Agro-infiltration. Upon verification, all of the constructs were introduced into

Agrobacterium tumefaciens strain C58C1 with electroporation (Qu et al. 2003). To carry out the experiments described in the Results section, various combinations of

Agrobacterium suspensions were mixed together and delivered into N. benthamiana leaves as described (Qu et al. 2003; Zhang et al. 2015; Zhang et al. 2017). A p19- expressing Agrobacterium strain was included in all combinations to alleviate RNA silencing-mediated mRNA degradation.

RNA extraction and Northern blotting. Total RNA was extracted from agro- infiltrated N. benthamiana leaves using the Direct-zol RNA Miniprep kit (Zymo

Research, Irvine, CA). To ensure consistency, four equivalent leaf sections derived from infiltrated leaves of four different plants were pooled before RNA extraction. The RNA extraction procedure included a DNase treatment step that removes DNA contamination.

The RNA was then quantified with NanoDrop and subjected to Northern blotting as described (Zhang et al. 2015; Zhang et al. 2017).

Protein extraction and Western blotting. Protein was extracted from the same benthamiana leaf pool using RIPA buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl, 1 mM PMSF, and 1X ProBlockTM Gold Plant Protease Inhibitor Cocktail, the last two reagents should be added immediately before use). Western blotting was carried out as described (Zhang et al. 2015; Zhang et al. 2017). Anti-HA and anti-mCherry antibodies were purchased from ThermoFisher Scientific.

Detection of (-) gRNA with semi-quantitative RT-PCR. The procedure for detecting (-) strand of TCV gRNA was adapted from Plaskon et al. (2009). A TCV (-) strand specific primer with a unique 5′ end tag (5′-

GGAACGTCATGGTGGTGACAAAAACAGCGCTCGCAGT-GGGACT-3′) was used to prime the reverse transcription. Subsequently, a tag-specific primer (5′-

GGAACGTCATGGTGGTGACAAAA-3′) and a TCV-specific reverse primer (5′-

GGACAAAAGAGATCGCCTGGTC-3′) were used to amplify an (-)-strand-specific

PCR product. Semi-quantification was achieved using an RT-PCR product of N. benthamiana EF1α mRNA as the reference.

Confocal microscopy. Confocal microscopic observations were carried out using a

Leica Confocal microscope (TCS SP5) available through Molecular and Cellular Imaging

Center at the Ohio Agricultural Research and Development Center, The Ohio State

University.

Acknowledgments

We thank the labs of Drs. Lucy Stewart and Peg Redinbaugh for generous equipment sharing. This study was supported by a SEEDS grant from the Ohio Agricultural

Research and Development Center, Graduate Assistantships from OSU and OARDC to

Q.G. and R.S., respectively, as well as tuition assistances to S.Z., R.S., and Q.G. from the

Department of Plant Pathology, OSU. X.-F. Z was supported in part by China

Scholarship Council.

Figures

Figure 2.1. TCV p88 partially trans-complements a p88-defective replicon. A:

TCV genome organization. The nucleotide (nt) coordinates of the open reading frames encoding p28, p88, and capsid protein (CP) were highlighted. The two subgenomic

RNAs, sgRNA1 and 2, serving as mRNA for movement proteins (p8 and p9) and CP respectively, were also depicted. B: Constructs used in experiments leading to C. All constructs were delivered into N. benthamiana cells using agro-infiltration. The first two support replication-independent expression of p88 (p88-2HA) and p28, with their transcription controlled by the strong 35S promoter with duplicated enhancers. The next three were TCV replicons with the CP ORF replaced by that of mCherry. TCV_sg2R replicates on its own as both p28 and p88 ORFs are intact. Δp88_sg2R is p88-defective

39 as the p88 ORF was terminated prematurely by a 4-nt deletion at nt #1402.

[p28stop]_sg2R translates neither p28 nor p88, as both ORFs were terminated by a stop codon introduced at aa #36. C: Partial trans-complementation by replication-independent expression of p88 and p28 revealed by Northern and Western blottings. EB: ethidium bromide stained agarose gel serving as the loading control for Northern blotting.

Figure 2.2. The level of p88-2HA protein correlates negatively with its trans- complementation activity, but positively with its replicational repression activity. A:

Constructs with the same p88-2HA ORF under control of promoters of different strengths. 2X35S: 35S promoter with duplicated enhancers. Core35S: core 35S promoter without the enhancers. B: Protein levels of the replication-independent p88-2HA and replication-dependent mCherry, in the absence or presence of two different replicons differing in their ability of producing cognate p88, as determined by Western blotting. C:

The efficiency of p88-2HA-mediated trans-complementation (lanes 14-19) and

41 repression (lanes 8-16), with two different p88-2HA protein levels, as determined by

TCV-specific Northern blotting.

Figure 2.3. The C-terminal 2HA tag does not appreciably affect p88 activities.

The Core35S::p88 construct expressing a tag-free p88 was compared side-by-side with two other constructs that express varying levels of p88-2HA, a C-terminally tagged p88, for its ability to repress, as well as trans-complement, TCV replication.

Figure 2.4. Trans-complementation mediated by various deletion mutants of p88-2HA. A: diagrams of different deletion mutants along with the promoters used.

Although the p88fs-2HA contained a one-nt deletion after fourth aa residue resulting in frameshift (fs), it permitted translational re-initiation at aa #37 (see results in B). B:

Partial trans-complementation by the ΔN36 mutants as revealed by Western and

Northern blottings. C: Detection of (-) strand gRNA with strand-specific, semi- quantitative RT-PCR (see Materials and Methods for details).

Figure 2.5. Replicational repression by p88-2HA mutants depicted in Figure

2.3A, as determined by Northern and Western blottings.

Figure 2.6. Replicational repression by p88-2HA variants partially correlates with their ability to trans-aggregate G11-p28. A: Schematic depiction of the G11-p28 trans-aggregation assay. The diffuse distribution pattern of G11-p28, a p28 variant with an N-terminal G11 (the 11th β-strand of GFP) tag, could be revealed with the co- expression of G1-10 (the first 10 β-strands of GFP). The right-hand image shows a confocal microscopy image. B: various p28 and p88-2HA derivatives altered the distribution of G11-p28 to varying extents (see text for details). C. p88 deletion mutants tagged at the C-termini with GFP revealing their intracellular behaviors.

CHAPTER 3

Replication-Based Biogenesis of a Long Noncoding RNA Originated from Turnip

Crinkle Virus

Abstract

Many positive sense (+) RNA viruses encode long noncoding RNAs (lncRNAs) that play important roles in their infectious cycle. A distinguishing feature of these lncRNAs is that they are produced through 5'-to-3' degradation of viral genomic or subgenomic

RNAs, by exoribonucleases of host cells. We report here a new lncRNA encoded by the plant-infecting turnip crinkle virus (TCV) that is produced by a replication-based mechanism. This lncRNA, designated tiny TCV subgenomic RNA (ttsgR), was mapped to the last 283 nucleotides of TCV genomic RNA. It accumulated to high levels in cells of Nicotiana benthamiana plants in which TCV replication took place in the presence of overexpressing p88, the TCV-encoded RNA-dependent RNA polymerase (RdRp).

However, ttsgRNA replicated robustly from templates as short as itself, without the need for any other TCV RNAs, as long as both of the TCV replication proteins, p28 and p88, were provided in trans from nonviral sources. Accordingly, both (+) and (-) sense forms of ttsgR were detected using a strand-specific RT-PCR procedure. ttsgRNA replication did not entail any 5’ RNA secondary structure but required the presence of a G3(A/U)4 motif at the 5’ terminus. Furthermore, it strictly relied on the integrity of the CCC motif at the 3’ terminus. Both of these structural features are shared by TCV genomic and

48 subgenomic RNAs. These findings established that ttsgRNA was the product of a replication-based mechanism, and identified a novel strategy for the biogenesis of lncRNAs associated with (+) RNA viruses.

Introduction

Long noncoding RNAs (lncRNAs) have recently drawn increasing attentions due to their widespread presence in both animal and plant cells, and their potential regulatory roles in diverse cellular processes (Quinn and Chang 2016). Many single-stranded, positive sense (+) RNA viruses, despite their relatively small genomes, have also been found to produce lncRNAs. They include many animal infecting flaviviruses such as

Murray Valley encephalitis virus (MVEV), Dengue virus (DENV), West Neil virus

(WNV), and Zika Virus (Lin et al. 2004; Scherbik et al. 2006; Urosevic et al. 1997), and plant-infecting barley yellow dwarf virus (BYDV; Kelly et al. 1994; Koev and Miller

2000), soybean dwarf virus (SDV; Yamagishi et al. 2003), red clover necrotic mosaic virus (RCNMV; (Iwakawa et al. 2008), maize chlorotic mottle machlomovirus (MCMV;

Scheets 2000), tobacco necrosis virus D (TNV-D;(Gunawardene et al. 2019), and beet necrotic yellow vein virus (BNYVV; (Peltier et al. 2012). lncRNAs derived from these viruses are typically 3’ co-terminal with genomic RNA of their parental viruses, at a length of 200 to 500 nt.

These virus-derived lncRNAs have lately attracted a surge of interest, due to their potential roles in viral pathogenesis. Many of them are thought to counteract the host defense responses that target viral genomes, such as RNA decay-mediated by the exoribonuclease Xrn1 of host cells (Moon et al. 2012), as well as RNA silencing (Moon

49 et al. 2015). They may also play a role in repressing the type I interferon-mediated antiviral defense in animal virus-infected cells (Chang et al. 2013), and possibly also counters the Toll antiviral pathway in insect vectors (Pompon et al. 2017). Some lncRNAs have also been found to be involved in viral replication, dissemination, and host adaption (Slonchak and Khromykh 2018).

The biogenesis of flavivirus-associated lncRNAs has been characterized in WNV

(Pijlman et al. 2008), yellow fever virus (YFV; Silva et al. 2010), DENV (Chapman et al.

2014), ZIKA (Akiyama et al. 2016). Separately, lncRNAs encoded by BNYVV (Dilweg et al. 2019), RCNMV (Iwakawa et al. 2008), and TNV-D (Gunawardene et al. 2019) have also been thoroughly investigated. These lncRNAs all resulted from the 5’ to 3’ processing of longer viral RNAs (e.g. genomic or subgenomic RNAs) by host 5’ exonucleases such as XrnI, frequently occurring in the processing bodies present in the host cell cytoplasm.

Xrn1 processing (Pijlman et al. 2008). In support of the involvement of Xrn1 in WNV sfRNA biogenesis, its production is decreased in Xrn1-depleted cells (Pijlman et al.

2008). Xrn1 degradation alone is insufficient for sfRNA biogenesis, as at least one stable

RNA secondary structure is needed to stall Xrn1-mediated decay. Such Xrn1-resistant structures are abundantly available in the 3’ UTR of flaviviruses. Some flaviviruses with

50 multiple stable RNA secondary structures were found to produce up to four distinct sfRNAs.

BYDV-associated lncRNA. In addition to the lncRNA, BYDV also produces two sgRNAs. The 5’ sequence of gRNA and sgRNA are highly conserved, but such conserved sequence is absent at the 5’ end of the lncRNA (Kelly et al. 1994; Zavriev et al. 1996). This suggest that BYDV lncRNA is not the product of viral replication.

Here we report the identification and characterization of a lncRNA derived from turnip crinkle virus (TCV). The existence of this lncRNA, designated tiny TCV subgenomic RNA (ttsgRNA), was unveiled by overexpressing p88, the TCV-encoded

RNA-dependent RNA polymerase (RdRp) in cells replicating a TCV replicon (Zhang et al. 2019a). TCV is a small plant-infecting (+) RNA virus in the same family as the above mentioned RCNMV and TNV-D. Its 4,054 kilobase (kb) genome encodes two proteins directly involved in the replication process: the p28 auxiliary protein, and the p88 RdRp, the latter being the product of a translational readthrough event that overcomes the p28

51 stop codon (White et al. 1995). During infection of host cells, TCV also produces two subgenomic RNAs that encode movement proteins and the coat protein, respectively. In the current study, we used a novel approach to examine the biogenesis of ttsgRNA in the absence of other TCV RNAs and discovered that a replication-based mechanism was responsible for ttsgRNA production. Our results revealed a novel viral lncRNA biogenesis strategy that does not involve 5’-to-3’ exonuclease degradation of longer viral

RNAs.

Results

Overexpression of p88 unveils a novel TCV RNA. We previously reported that the overexpression of p88, the TCV-encoded RdRp, repressed the replication of TCV replicons (Zhang et al. 2019a). Furthermore, the extent of p88-mediated repression correlated positively with p88 protein levels. In the current study, we first corroborated this observation by attaching the p88 ORF to a new promoter that is tobacco microspore- specific (Oldenhof et al. 1996), which mainly driven transcription in microspore cells, hence expected to be minimally active in leaf cells of N. benthamiana (Fig. 3.1B,

Micro::p88). This new construct, along with the 2X35S::p88 and Core35S::p88 constructs reported in the previous study (Zhang et al. 2019a), were brought into N. benthamiana leaves, alone or in combination with various TCV replicons, using

Agrobacterium infiltration (agro-infiltration). The expression levels of p88 were then assessed with an anti-HA antibody because the p88 ORF was C-terminally tagged with a double HA tag in all three constructs. Note that a p19-expressing construct was included in most infiltrations to counteract RNA silencing (see exceptions below).

Western blotting with an HA antibody showed that the Micro::p88 construct indeed drove lower p88 expression than the other two constructs (Fig. 3.1C). Note that p88 at this level still constituted an overexpression relative to its level in TCV infections. As expected, the lower p88 expression caused a much less potent repression on the co- introduced TCV_sg2R replicon, permitting the accumulation of higher levels of TCV gRNA and sgRNAs than the other two p88-expressing constructs (Fig. 3.1D, lanes 6-8).

Conversely, the lower p88 expression also allowed for modestly better complementation of a mutant TCV replicon lacking its own p88 (Δp88_sg2R; Fig. 3.1D, lanes 9-12). These results confirmed the dual functions of p88 established in the earlier study (Zhang et al.,

2019).

Surprisingly, p88-mediated repression and replicational complementation of TCV replicons were both accompanied by prominent accumulation of a previously overlooked,

TCV-specific RNA that is substantially smaller in size than TCV sgRNA2 (Fig. 3.1D; lanes 6-8, 10-12, bottom). Upon closer inspection, the same RNA was also present in

TCV_sg2R-only infections, albeit at much lower levels than other TCV RNAs (Fig.

3.1D, lane 5). By contrast, overexpression of p88 from the three non-viral constructs caused an over-accumulation of TCV sgRNA2, and more importantly this novel TCV

RNA. Hereafter we refer this novel TCV RNA as tiny TCV subgenomic RNA, or ttsgRNA.

Was the RdRp function of p88 responsible for ttsgRNA overaccumulation? To answer this question, we tested three p88 mutants for their impacts on ttsgRNA levels.

The first mutant, p88(mGDD), was an RdRp-null mutant due to point mutations that replaced the conserved Gly-Asp-Asp (GDD) motif to Val-Ala-Ala (VAA). The other two

53 mutants, p88ΔN36 and p88ΔN127, harbored N-terminal deletions of 36 and 127 amino acid (aa) residues, respectively. Our previous study (Zhang et al. 2019a) established that while p88ΔN36 retained both the repressive and complementing activities of p88 to a substantial extent, p88ΔN127 was still modestly repressive but incapable of complementing p88-defective replicons. As shown in Fig. 3.1E, relative overaccumulation of ttsgRNA occurred in the presence of p88, and more prominently p88ΔN36, but not in the presence of p88(mGDD) or p88ΔN127. These results provided the earliest hint that ttsgRNA are generated by the RdRp activity of p88.

Since the TCV_sg2R replicon used in above experiments lacked the CP/silencing suppressor gene of wildtype (wt) TCV, we next tested whether wt TCV infections were also accompanied by ttsgRNA accumulation. A faint signal was indeed detectable at the position expected for ttsgRNA (Fig. 3.1F, lanes 8-9, arrow), along with one additional band below it, and two more bands above. Expressing p19 without p88 did not enhance the intensity of these signals (lanes 4-5). By contrast, overexpressing p88 alone caused a dramatic reduction in TCV gRNA and sgRNA levels, and a relative increase in ttsgRNA abundance (lanes 10-11). Interestingly, expression of both p19 and p88 caused the wt

TCV to dramatically over-accumulate sgRNA2, as well as ttsgRNA, and another slightly larger RNA species (lanes 6-7). Taken together these experiments demonstrated that ttsgRNA was a novel TCV RNA that over-accumulated in the presence of excess p88.

Biogenesis of ttsgRNA is replication dependent. We next investigated how ttsgRNA was produced during TCV infections, and why its accumulation was bolstered by p88 overexpression. The small size of ttsgRNA reminded us of RCNMV SR1f RNA, and TNV-D svRNA, two lncRNAs produced by viruses highly similar to TCV

(Gunawardene et al. 2019; Iwakawa et al. 2008). Since both of these lncRNAs were shown to be products of stalled exonuclease degradation of longer viral RNAs, including gRNA and sgRNA, we asked whether ttsgRNA could also be stalled degradation products of TCV gRNA or sgRNAs.

To first assess the potential of TCV sgRNAs as the source for ttsgRNA, we abolished the synthesis of TCV sgRNA1, sgRNA2, or both, by introducing mutations in the promoter regions of the sgRNAs (Wang et al. 1999; Wang and Simon 1997; Wu et al.

2010). Specifically, the Δsg1 mutant contained multiple nucleotide substitutions that disrupts the base-paired stem needed for sgRNA1 transcription (Fig. 3.2A, top), whereas the Δsg2 mutant harbored a 12-nt deletion in the stem-loop structure necessary for sgRNA2 transcription (Fig. 3.2A, bottom). Furthermore, the Δsg1&2 mutant combined the mutations in Δsg1 and Δsg2. As shown in Fig. 3.1B, the Δsg1, Δsg2, and Δsg1&2 mutants abolished the accumulation of TCV sgRNA1, sgRNA2, and both, respectively.

However, in the presence of overexpressed p88, all of the mutants accumulated ttsgRNA to levels comparable to the TCV_sg2R control. Therefore, TCV sgRNA degradation were not the primary source of ttsgRNA.

In contrast to sgRNAs, whether TCV gRNA degradation led to ttsgRNA accumulation could not be easily investigated with mutagenesis. We hence turned to an alternative approach, asking whether ttsgRNA could be produced from various 3’ co- terminal, defective TCV templates whose replication could not occur unless both p28 and p88 were provided in trans. As shown in Fig. 3.2C, these defective templates included:

(i) [p28stop]_sg2R that could translate neither p28 nor p88 due to an extra stop codon at the aa position 36 of the p28/p88 ORF (Zhang et al., 2019); (ii) sg2R_Temp that was

55 expected to transcribe sgRNA2 only; and (iii) ttsgR_Temp that begins at nt position 3386

– immediately downstream of the mCherry insertion site.

The replication of [p28stop]_sg2R could be complemented by providing both p28 and p88 in trans (Zhang et al., 2019; Fig. 3.2D, lanes 3 and 6). Notably, ttsgRNA accumulated in [p28stop]_sg2R infections only when its replication was successfully complemented (compare lanes 3 and 6). Similarly, both the sg2R_Temp and ttsgR_Temp could replicate to detectable levels when both p28 and p88 were provided in trans. Again, ttsgRNA was detectable only when these templates replicated (Fig. 3.2B, lanes 11 and

16). To emphasize, ttsgRNA was undetectable when p28 or p88 was provided separately, or when p28 was co-delivered with a p88(mGDD) mutant (Fig. 3.2D and 2E). Together these results indicated that ttsgRNA accumulation depended on viral RNA replication.

Nevertheless, it could still be derived from the degradation of other longer RNAs that underwent replication. Therefore, to unambiguously determine whether ttsgRNA originated from replication or stalled degradation, we designed a new template with a more precise 5’ end.

To guide the design of a more precise ttsgR template, we first tried to map the 5’ end of the ttsgRNA with a set of four different probes that hybridized to different positions of

TCV RNA (Fig. 3.2C, bottom). As shown in Fig. 3.2F, ttsgRNA was readily detectable with the 3997R oligo probe, but not with the 3696Rc probe, or two other probes that hybridized to more upstream positions of TCV RNA. These experiments narrowed down the 5’ end of ttsgRNA to a position between 3696 and 3997 of TCV, suggesting a ttsgRNA size of no more than 358 nts.

ttsgRNA initiates at a discrete position of TCV RNA, and it exists in both viral sense and complementary sense forms. To map the 5’ end of ttsgRNA precisely, we adopted a circularization RACE protocol (Shivaprasad et al. 2005) to generate circularized ttsgRNA. Since the presence of a complementary sense (-) form was often used as one of the criteria to verify the replication of viral RNAs, we decided to adopt a highly strand-specific RT-PCR procedure (Plaskon et al. 2009) to separately amplify the cDNA of ttsgRNA and its (-) form, should the latter be present. The specific experimental steps were schematically summarized in Fig. 3.3A, and details given in

Materials and Methods. As shown in Fig. 3.3B, specific, reverse-transcription (RT)- dependent fragments were detected in samples that combined the ttsgR_Temp with p28 and p88, but not with p28 and p88(mGDD) (compare lanes 8, 10, and 12). However, there were two notable surprises: (i) both (+) and (-) forms of ttsgRNA were successfully amplified; (ii) RNA circularization was unnecessary for the amplification of both forms, indicative of the existence of tandem ttsgRNA dimers of both (+) and (-) senses. These two surprises strongly suggested that ttsgRNA underwent active replication in the presence of p28 and p88.

The PCR products highlighted by arrows in Fig. 3.3B were purified and cloned. For each of the PCR products, more than 20 individual clones were sequenced. As shown in

Fig. 3.3C, regardless of the sense of the RNA, or whether the RNA was circularized, more than 50% of clones in each category shared a 3’-5’ junction that joins the 3’ end

CCC with GGG beginning with nt #3772 of TCV. The rest of clones mostly contained 1-

2 extra G/C between the two ends. It should be noted that circularized and non- circularized RNA samples yielded similar sequencing results, which indicated the

57 preexistence of ttsgR molecular that were 5’ and 3’ ends covalently ligated. Such molecular could be ttsgR dimers arisen from ttsgR replication. Hence, the junction sequences were probably mostly derived from dimeric RNAs. As a result, some of the

RNA species, especially those with rare junction identities, might replicate as dimeric rather than monomeric forms. Nevertheless, the fact that most clones had #3772 joining

#4054 led us to conclude that the ttsgRNA is predominantly 283 nt long, most likely arisen from RdRp-mediated amplification that initiates at nt #3772.

An RNA stem loop flanking the 5’ end of ttsgR does not contribute to ttsgR biogenesis. The nt #3772 initiation site of ttsgR falls inside the end loop of an RNA secondary structure found previously to be important for TCV virion assembly (origin of assembly or OAS; (Qu and Morris 1997); Fig. 3.3D). Separately, similar RNA secondary structures were deemed important for TCV sgRNA synthesis through the premature termination mechanism (Wu et al. 2010). Conversely, stable RNA secondary structures were also found to be needed to stall exonuclease degradation, ensuring the accumulation of various viral lncRNAs (Gunawardene et al. 2019; Iwakawa et al. 2008; Pijlman et al.

2008). We hence assessed the potential role of the OAS stem-loop in ttsgRNA accumulation, by introducing mutations that disrupt the base paired stem, in the ttsgR_Temp background. As shown in Fig. 3.3D, the ttsgR_mL and -mR mutants contained mutations within the left and right side of the stems, respectively, that were expected to abolish most of the base pairing, whereas the ttsgR_mLR restored the base pairing, but not the original sequence, by combining the mutations of the first two mutants. Figure 3E showed that none of the mutants significantly compromised the

58 accumulation of ttsgR. Therefore, this stem-loop structure is not essential for ttsgR biogenesis.

A shorter ttsgR template with 28 extra nt at the 5’ end reveals specific sequence requirements for ttsgR accumulation. The precise mapping of the ttsgR 5’ end allowed us to design a new template, referred to as ttsgR_S, that begins with TCV nt #3744, extending the ttsgR sequence for 28 nt to the upstream (Fig. 3.4A). As shown in Fig.

3.4B (lanes 6-7), ttsgR_S permitted (p28 + p88)-dependent accumulation of RNAs similar to ttsgRNA in size (compare lanes 6-7 with 3-4). Curiously, the accumulated

RNAs appeared to contain an additional, slower migrating species, at higher abundance than ttsgRNA. Sequencing results showed that RNAs that initiated at nt #3772 and nearby positions accounted for 7 of the 18 clones sequenced (Fig. 3.4C), whereas those that initiated at the very end of the template (nt #3744) accounted for 11 of the 18 clones.

Furthermore, most of these clones contained three extra nts GAG (G as guanine, as adenine) at their 5’ ends, which were presumably derived from the sequence of the adjacent XhoI site (CTCGAG, C as cytosine, T as thymine) in the vector. Therefore, with ttsgR_S as the template, ttsgR synthesis could initiate at two sites, with both types of

RNA remaining sufficiently stable to permit detection. Together these data strongly support the view that ttsgRNA originated from RdRp-dependent synthesis, rather than stalled degradation.

We next used the ttsgR_S template to investigate the structure and/or sequence requirements for ttsgR accumulation. While the OAS stem-loop could still be formed with the ttsgR_S template, its dispensability for ttsgR accumulation as demonstrated in

Fig. 3.3D prompted us to search for alternative structures that could potentially influence

59 ttsgR biogenesis. We found that the GGG triplet at the 5’ end of ttsgR, along with a few nts downstream, could potentially pair with 8 nts at the 3’ terminus to form an imperfect stem (Fig. 3.4A). We hence interrogated this putative structure with three mutants depicted in Fig. 3.4A. Specifically, mutant S_m5p changed the 5’ end GGG to CCA, whereas mutant S_m3p changed the 3’ end CCC to UGG, both disrupting the basal stem of the structure. Finally, S_m5p3p combined these mutations to restore the stem with the altered sequences.

Judging from the Northern blotting, mutants S_m3p and S_m5p3p, both sharing the

CCC to UGG change at the 3’ end, completely abolished the accumulation of both RNA species (Fig. 3.4B, lanes 11-17). This result suggested that the identity of the CCC triplet was essential for the accumulation of both ttsgRNA and the slightly larger RNA initiating from the upstream GAG (see above). By contrast, the GGG to CCA mutation in the

S_m5p mutant appeared to have diminished the accumulation of ttsgRNA, but not the larger RNA (Fig. 3.4B, lanes 9-10). We then used sequence analysis to discern the ends of S_m5p-derived RNA (Fig. 3.4D). Consistent with the Northern blotting results, most of the clones (13 out of 17) contained cDNA that initiated in the vicinity of the upstream

GAG. Strikingly, none of the clones initiated at the original initiation site, which now has

CCA instead of GGG. Similarly, informative are three clones whose cDNA initiated from two different downstream sites. Like the initiate sites at nt #3772 (GGG) and 3744

(GAG), both of the downstream initiation sites (at nt #3803 and 3964) had at least two

Gs, followed by at least three U/A residues. This information became important later when we attempted to decipher the conserved sequence motifs at the 5’ ends of all replicating TCV RNAs. In summary, the CCC triplet at the 3’ end was strictly needed for

60 ttsgRNA biosynthesis, whereas the 5’ end GGG or GAG appeared to be slightly more flexible.

A ttsgRNA template with an exact 5’ terminus directs efficient ttsgR synthesis.

To completely rule out the involvement of the OAS stem-loop and possibly exonuclease degradation in ttsgR accumulation, we then produced another ttsgR template, ttsgR_S2, that has the exactly same sequence as ttsgR itself, including the GGG triplet at the 5’ end

(Fig. 3.5A, left). As shown in Fig. 3.5B (lanes 4-6), ttsgR_S2 permitted (p28 + p88)- dependent accumulation of ttsgRNA. We further verified the identity of the accumulated

RNA with sequence analysis and found that 7 of the 17 clones contained the expected ttsgRNA sequence. The other 10 clones, each representing one sequence, all contained slight variations at the junction. As mentioned earlier, some of the variant might replicate in the dimeric form. Together these results demonstrated that a ttsgRNA template without any extra upstream sequence or structure readily directed its own replication, and produced sufficient stable RNA to allow consistent detection.

A 5’ terminal G3(A/U)4 motif is critically important for ttsgRNA accumulation.

Because ttsgR_S2 template directed faithful amplification of ttsgRNA, it would be an ideal vehicle for interrogating potential terminal motifs that bring the template to the p28/p88-based replication machinery. Specifically, it could be used to further define the

5’ end of ttsgRNA, which earlier experiments (Fig. 3.4) hinted that the GGG triplet was critical but could be substituted with GAG. Furthermore, recovered sequences compiled in Fig. 3.3C and D collectively showed that in all but one amplified variant, the

GGG/GAG triplet was followed by four A/U residues. Incidentally, this G3(A/U)4 motif was also evident at the 5’ ends of TCV gRNA and two sgRNAs (GGGUAAU,

GGGAUAU, GGGUAAU, respectively). We hence examined the 5’ GGGUAAA motif for its potential role in ttsgRNA generation using a set of four mutants.

The first mutant, designated S2_m1, mutated 3 nts to change GGGUAAA into

AGGCAAG (Fig. 3.5A, 2nd diagram). Note that in the sequence of TCV gRNA, these three mutations maintained the aa sequence of CP, hence could later be adopted to test the role of ttsgRNA in the context of TCV replicons. Additionally, the nearest G residue in the upstream vector sequence was changed to an A (green, underlined) – a change also shared by S2_m2, _m3, and _m4 (Fig. 3.5A). When brought into plant cells with p28 and p88, the S2_m1 mutant still produced an easily detectable level of ttsgRNA, although its size appeared to be slightly larger (Fig. 3.5B, lanes 7-9). Consistently, none of the recovered sequences initiated at the original starting site, indicating that the GGGUAAA motif was indeed critical. Instead, most of sequences initiated at the underlined A in the vector (Fig. 3.5D). Strikingly, 18 out of 28 recovered sequences acquired at least two Gs de novo, and 9 out of those also acquired an extra A (Fig. 3.5D, rows 2-6). As a result, 9 sequences had GGAAA as the 5’ end, 6 had GGAAAA, and 3 had GGGAAAA.

Moreover, 3 sequences initiated at a downstream GGUAAU site (nt #3803), with an extra

G added de novo. To further highlight the importance of the U/A stretch immediately following GG or GGG, neither of the two “GGCA” sequences almost immediately downstream of nt #3772 was used as an initiation site. Collectively these results suggested two important points: (i) at least two Gs, followed by three A/Us, are needed at the 5’ end for an efficient ttsgRNA template; (ii) with appropriate contexts, such 5’ end motifs could be regenerated de novo with the addition of extra nts during replication.

These points were both further reinforced by the next three mutants. The S2_m2 mutant was originally designed to cause more extensive disruption to the putative hairpin shown in Fig. 3.4, formed by long-distance interactions between two termini of ttsgRNA.

It nevertheless also changed the GGGUAAA motif into ACCGUAA (Fig. 3.5A, middle diagram). However, there were still four U/As following the middle G. This mutant replicated robustly in N. benthamiana cells in the presence of p28 and p88 (Fig. 3.5B, lanes 10-12). Among the 21 cDNA sequences analyzed, approximately one half (11 clones; Fig. 3.5E, rows 2-8) initiated at the middle G at position #3775, but most of them also acquired additional Gs, or underwent other small modifications (e.g. U to A at

#3776, rows 7-8), to restore the G3(U/A)4 motif. The other half of the sequences (10 clones), like those of S2_m1, adopted the upstream A as the initiation site, but again acquired a varying number of Gs or Gs plus A to reconstitute the G3(U/A)4 motif.

Notably, in these 10 sequences all of the mutated nts were retained, suggesting that substantial 5’-3’ base pairing, with the possible exception of three terminal G:C pairs, was not essential for ttsgRNA replication.

Many of the RNAs generated from the mutant S2_m3 were notable exceptions to the above generalization. Similar to S2_m2, this mutant was also created to disrupt the putative 5’-3’ interactions. This mutant led to the accumulation of high levels of ttsgRNA

(Fig. 3.5B, lanes 13-15), with the original mutations retained in all 20 clones sequenced, indicating that the putative interaction as depicted in Fig. 3.5A was not needed for ttsgRNA synthesis. However, a majority of the recovered cDNAs initiated from vector sequence AA, which is two nts upstream of the position of insertion. Additionally, although some of the sequences acquired one or more Gs de novo, none of them had

63 more than two As after the Gs. Nevertheless, as noted earlier, these junction sequences may simply reflect the joints of dimeric replicons, rather than true 5’ termini of replicating monomers. Importantly, the single sequence with most clones (five) had the authentic 5’ end initiating at #3772 (Fig. 3.5F, row 2), conforming to the G3(A/U)4 consensus.

Finally, the mutant S2_m4, which combined all mutations in m2 and m3, also accumulated abundant ttsgRNA (Fig. 3.5B, lanes 16-18). Among the 22 clones sequenced, 11 contained cDNA sequences that initiated at the single G at nt #3775 that preceded UAAA, but with two Gs added de novo (Fig. 3.5G). The other 11 sequences initiated from various upstream sites, with or without extra G/A added de novo. However, all of them contained at least three As in the vicinity of starting sites. Overall the data obtained with this set of mutants provided strong evidence for a conserved 5’ end motif of G3(U/A)4.

The 3’ terminal CCC triplet is essential for efficient ttsgRNA accumulation. With the 5’ end requirement defined we now return to the 3’ end to determine the conserved features important for ttsgRNA biosynthesis. We learned earlier that mutating all three of the Cs completely abolished ttsgRNA accumulation (Fig. 3.4B). We hence decided to focus our attention on the middle C only. Although experiments in Figs. 4 and 5 ruled out the involvement of extensive 5’-3’ base-pair interactions, the potential contribution of the three terminal G:C pairs could not be discounted, as they could always be reconstituted with extra Gs acquired de novo. Therefore, we investigated the 3’ middle C in conjunction with the 5’ middle G.

We first determined whether changing the middle G within the 5’ GGG into a U or A affected ttsgRNA levels, with mutants S2a-5u and -5a. Note that the ttsgR_S2a template here differed slightly from the ttsgR_S2 used earlier, with two nt changes that eliminated the G residue immediately upstream of ttsgR coding sequence (Fig. 3.6A). As shown in

Fig. 3.6B (lanes 9 and 15), in both cases ttsgRNA were abundantly detected. This was expected as more dramatic alterations within this terminus, as documented in Figs. 3.4 and 3.5, were well tolerated. Interestingly, neither mutation was retained in a majority of the sequenced clones. Rather, they were either reverted to G (e.g. Fig. 3.6C, row 2), or simply discarded by the most abundant ttsgRNA variants (e.g. Fig. 3.6D, row 2). A few of S2a-5u-derived clones had an A instead of U at the mutation site, though all of them appeared to use alternative initiation sites that were upstream of GAG (Fig. 3.6C, rows

10-13). Similarly, two of the S2a-5a descendants that retained the A mutation also appeared to have initiated at upstream sites. These results were surprising because GAG was found in earlier experiments as an alternative initiation site (Fig. 3.4). This may suggest that GAG could only be used as an initiation site under certain sequence contexts.

Overall these results corroborated earlier results showing mutations in the vicinity of 5’

GGG were well tolerated and easily corrected by the replication process itself. More importantly, they bolstered the importance of the G3(A/U)4 motif at the 5’ end.

By contrast, changing the middle C of the 3’ CCC to either U and A (the S2a-3u and

S2a-3a mutants), either alone or in combination with changes at the 5’ middle G, dramatically reduced, but curiously did not completely abolish, the ttsgR accumulation

(Fig. 3.6B, lanes 5, 7, 11, 13, 17, 19). These results strongly suggested that the middle C, and possibly all three Cs at the 3’ end, were absolutely needed for efficient ttsgR

65 synthesis. They further suggested that the repair of such mutations, though involving only single nt alterations, was difficult and rare, occurring only in a small fraction of cells. We nevertheless attempted to sequence the ttsgRNA progeny in tissues receiving these mutants. The C to U mutation in all S2a-3u-derived sequences was reverted to C (Fig.

3.6E). Likewise, the C to A mutation in the S2a-3a mutant was reverted to C as well.

Interestingly, a small number of the S2a-3a descendants appeared to make do with just two terminal Cs instead of three. Together these results verified that integrity of the C triplet at the 3’ terminus was essential for ttsgRNA replication.

The GGGUAAA motif is needed for ttsgRNA synthesis from infectious TCV

RNA templates. Now that we have established the essential role of the GGGUAAA motif at the 5’ end of ttsgRNA, we wondered if this motif was also needed for ttsgRNA synthesis during TCV replication, from either TCV_sg2R or wt TCV templates. To this end, we introduced the three point mutations in the S2-m1 mutant into the TCV_sg2R replicon, as well as wt TCV, creating sg2R_m1 and wt_m1 mutants (Fig. 3.7A). The mutants were delivered into N. benthamiana leaf cells with or without p88 to determine whether ttsgR production was affected by the mutations. As shown in Fig. 3.7B, while

TCV_sg2R and wt TCV infections in the presence of overexpressed p88 were accompanied with ttsgRNA accumulation to levels that rivaled sgRNA2 (lanes 6-7, 14-

15), similar infections with the mutated forms of both infectious TCV constructs led to a dramatic decrease of ttsgRNA, to barely detectable levels (lanes 10-11, 18-19). These results clearly demonstrated that the GGGUAAA motif was also required for ttsgRNA production from infectious TCV templates.

We next wanted to resolve whether production of ttsgRNA conferred any fitness benefit to TCV, or the host plants. To this end, we used wt TCV and the wt-m1 mutant to inoculate N. benthamiana plants, and closely monitored the infected plants for six weeks.

The plants infected with wt TCV and wt-m1 did not show any significant difference in the timing and severity of symptoms. The titers of viral RNA as determined for multiple plants, using Northern blotting, were very similar as well. We are currently assessing the stability of the mutations in wt-m1 by sequencing the virus progeny RNA collected from infected plants at several time points (2, 4, and 6 weeks post inoculation).

Discussion

Replication-based production of ttsgRNA. Viral lncRNAs have been reported for numerous (+) RNA viruses. They have been well investigated for a number of human- infecting flaviviruses, including DENV, WNV, and Zikavirus, which are collectively known as long noncoding subgenomic flavivirus RNAs (sfRNAs) (Zhang et al. 2019b).

Viral lncRNAs have also been widely observed in (+) RNA plant virus infections, including BNYVV, RCNMV, and TNV-D (Gunawardene et al. 2019; Iwakawa et al.

2008; Peltier et al. 2012). With the notable exception of lncRNAs associated with citrus tristeza virus (CTV), most of these lncRNAs share the following characteristics: (i) they are typically less than 500 nt in length, and often heterogenous at the 5’ end; (ii) they are derived from the 3’ termini of parental viruses; (iii) they are produced by exonuclease degradation of longer viral RNAs that stalls at various stable secondary structures; and

(iv) they mostly function to benefit the parental viruses.

The ttsgRNA, a TCV-related lncRNA reported here, bears most of these characteristics but is notably different in one key aspect: unlike the previously examined viral lncRNAs, ttsgRNA is clearly produced by the viral replication process. This conclusion is well supported by several lines of evidence. First, ttsgRNA could be produced from a template that was exactly the same length as the final product, in the absence of any other longer viral RNA, but in the presence of viral replication proteins p28 and p88. Second, p28 and/or p88-mediated template stabilization was unlikely the cause for ttsgRNA accumulation because replacing p88 with a RdRp-null mutant

[p88(mGDD)] completely abolished it. Third, both (+) and (-) strands of ttsgRNA were detected, suggesting that the (+) template was copied into (-) intermediates through replication. Finally, dimeric forms of ttsgRNA were readily detectable, which could only have been produced through inaccurate copying, possibly on non-covalently circularized templates (Herold and Andino 2001; Li and Nagy 2011).

While replication-based production has not been observed for other viral lncRNAs, this production mode should not be surprising given the functional conservation of viral lncRNAs. Namely, if these lncRNAs play important pro-viral roles during infections, the mechanisms of their biogenesis can be expected to evolve independently, converging on similar end products. Another important point we wish to make is that none of the previously reported viral lncRNAs have been examined using procedures similar to ours, invoking an in vivo system that is independent of the replicating parental viruses. It may not be unthinkable that some of those earlier lncRNAs could be produced/stabilized by a combination of replication and stalled degradation. Indeed, a most recent report found that the (-) from of RCNMV lncRNA could be detected in infected tissues, suggesting

68 that this lncRNA may have resulted from such a combination (Newburn and White

2020). Separately, one of the lncRNA species derived from Japanese encephalitis virus

(JEV) was found to act as a potent template for JEV RdRp in vitro (Chen et al. 2018).

Role of p88 overexpression in ttsgRNA. It is worth noting that ttsgRNA accumulation is rather inconspicuous in infections with wt TCV or TCV replicons that encode their own p28 and p88. However, ttsgRNA levels were dramatically elevated by p88 overexpression. We reported earlier that p88 overexpression depressed the accumulation of TCV gRNA, but elevated the accumulation of sgRNA, especially sgRNA2 (Zhang et al., 2019). The fact that both sgRNA2 and ttsgRNA were preferentially amplified by p88 overexpression hints at a regulative valve controlled by p88 protein levels. It is possible that higher p88 levels, which could occur later during cellular infection by TCV, serves as a switch that turns down/off gRNA replication in favor of sgRNA and ttsgRNA replication (transcription). How can this be achieved? We speculate that higher p88 levels may lower the p28:p88 ratio, favoring the formation of smaller membrane-borne spherules in which RNA replication occurs (Ertel et al. 2017). It is further possible that such smaller spherules act as constraints to limit the size of the

RNA templates they enclose, leading to preferential amplification of smaller RNA species. Finally, the finding that RdRp overexpression caused preferential accumulation of shorter viral RNAs also reminds us of the limitations of experimental systems that express replication proteins from nonviral sources, and use shorter RNAs (e.g defective interfering RNA or modified RNA3 of brome mosaic virus) as viral RNA surrogates

(Janda and Ahlquist 1993; Panavas and Nagy 2003).

Conservation and repair of ttsgRNA terminal characteristics. Since ttsgRNA is produced through replication, it is no surprise that its termini share sequence/structure characteristics with TCV gRNA and subgenomic RNAs. In particular, the CCC motif at the 3’ terminus is essential for ttsgRNA accumulation because altering this motif by as few as one nt led to dramatic decrease in ttsgRNA levels. Furthermore, the mutations at the middle C invariably reverted to the original identity in ttsgRNA variants recovered from the poorly accumulating pools, suggesting that: (i) the repair (reversion), possibly occurring during (-) strand synthesis, was very inefficient, possibly occurring in only a fraction of cells; (ii) subsequent replication could ensue only when the mutations were reverted to the original C. Successful repair of deleted 3’ ends have been reported for

TCV-associated satellite RNAs (Carpenter and Simon 1996; Nagy et al. 1997).

However, those repairs occurred in the presence of a helper TCV RNA that donated the repair templates.

Intriguingly, changes at the 5’ end of ttsgRNA were very efficiently repaired, suggesting that mutations at the 5’ end had a very small (if any) impact on the synthesis of (-) strands, the latter than served as the template for replication-dependent repair of the

5’ end. Interestingly, the 5’ end repair was very flexible, accomplished through either the use of alternative starting sites, or de novo addition of extra nts. Nevertheless, most successful variants restored a 5’ end composed of 3 Gs followed by 4 A/Us. Strikingly, similar G3(A/U)4 motifs are found at the 3’ ends of TCV gRNA and sgRNAs. However, unlike sgRNAs, ttsgRNA amplification did not require a strong secondary structure to brake the synthesis of a (-) strand intermediate (Wang et al., 1997; Wu et al., 2010). It is possible that low levels of ttsgRNA-corresponding (-) strands became stabilized in the

70 smaller spherules formed by p88 overexpression, as discussed above. Nevertheless, ttsgRNA offers a dramatically simplified alternative on which certain aspects of viral replication can be examined.

In contrast to previously reported viral lncRNAs that were found to have diverse pro- virus and anti-defense roles, we have yet to delineate a definitive role for ttsgRNA.

However, it is possible such a role manifests only under certain growth or stressed conditions, and maybe in certain hosts. These conditions will be thoroughly investigated in the near future. In summary, we have discovered and characterized a novel viral lncRNA that is produced by viral replication. Further dissection of this and other viral lncRNAs is expected to clarify their potential functions and offer new targets for virus control and management.

Material and Methods

Constructs. Constructs 2X35S::p88-2HA (p88), Core35S::p88, p88ΔN36, p88ΔN127, p28, TCV_sg2R, Δp88_sg2R, and [p28stop]_sg2R, were described in previous studies (Zhang et al. 2015; Zhang et al. 2017). Construct Micro::p88 was modified from Core35S::p88 by replacing the 2X35S promoter with a microspore- specific promoter derived from Nicotiana tabacum (Oldenhof et al. 1996). Construct p88(mGDD) was made by digesting the Core35S::p88 construct with HpaI and SalI, and replacing the 130-nt region encompassing the GDD coding sequence (GGAGACGAT) with a synthesized DNA fragment (gBlock-mGDD) in which the GDD-coding sequence was changed to GTTGCAGCT (Integrated DNA Technology, Carolville, IA). The replacement was accomplished using Gibson Assembly ligation (New England Biolabs,

Ipswich, MA). The sequences of this and other gBlock fragments are provided in Table 1.

Mutant replicon constructs Δsg1, Δsg2, and Δsg1&2, sg2R_m1, and wt-m1 were similarly created by replacing various regions in TCV_sg2R with different gBlock fragments (Table 1) or PCR fragments generated by mutation containing primers (Table

2). Specifically, the gBlock fragment for the Δsg1 construct contained 5 single nt mutations that disrupted the stem-loop structure of the sgRNA1 promoter (see Fig. 3.2A and Table 1 for details). The gBlock fragment for Δsg2 construct contained a 12 nt deletion that disrupted the stem-loop structure of the sgRNA2 promoter (see Fig. 3.2A and Table 1 for details). Both gBlocks were used for making Δsg1&2 construct.

Fragments for sg2R-m1 and wt-m1 constructs were amplified from TCV_sg2R or

TCV_wt using four primers, among which the second forward contained 3 single nt mutation in the G3(T/A)4 motif (see Fig. 3.7A and Table 2).

Construct sg2R_Temp was made by digesting the TCV_sg2R construct with XhoI and AatII and replacing the large 5’ deletion with a gBlock fragment (Table 1) that restores the sgRNA2 promoter structure (starting from TCV nt #2468). Similarly, template constructs ttsgR_Temp, and ttsgR_S were made by digesting TCV_sg2R construct with XhoI and SpeI and replacing the deleted region with gBlock fragments that encompass the 3386-3950, and 3744-3950 regions of TCV gRNA, respectively.

Constructs ttsgR_mL, ttsgR_mR, and ttsgR_mLR are variants of ttsgR_Temp, each made with gBlock fragments that contained mutations shown in Fig. 3.2D. The remaining ttsgR template constructs were all modified from ttsgR_S by digesting with XhoI and XbaI, and reinserting PCR fragments that further modified the template sequence from 5’ end, and/or introduced mutations at selected positions. The oligodeoxynucleotide primers used

72 to produce these mutant fragments are included in Table 2. The identities of all new constructs were verified with Sanger sequencing.

Agro-infiltration. Upon verification, all of the constructs were introduced into

Agrobacterium tumefaciens strain C58C1 with electroporation (Qu et al. 2003). To carry out the experiments described in the Results section, various combinations of

Agrobacterium suspensions were mixed together to a final OD600 = 0.5 of each

Agrobacterium suspension and delivered into N. benthamiana leaves as described (Qu et al. 2003; Zhang et al. 2015; Zhang et al. 2017). A p19-expressing Agrobacterium strain was included in most combinations to alleviate RNA silencing-mediated mRNA degradation.

RNA extraction and Northern blotting. Total RNA was extracted from agro- infiltrated N. benthamiana leaves using the Direct-zol RNA Miniprep kit (Zymo

Research, Irvine, CA). To ensure consistency, six equivalent leaf sections derived from infiltrated leaves of three different plants were pooled before RNA extraction. The RNA extraction procedure included a DNase treatment step that removes DNA contamination.

The RNA was then quantified with NanoDrop and subjected to Northern blotting as described (Zhang et al. 2015; Zhang et al. 2017).

73 described (Zhang et al. 2015; Zhang et al. 2017). Anti-HA antibody was purchased from

ThermoFisher Scientific (Waltham, MA).

Mapping of ends sequence of ttsgR. A circularization RACE protocol (Shivaprasad et al., 2005) was adopted to map both the 5’ and 3’ ends of the ttsgRNA variants. A strand-specific RT-PCR (Plaskon et al., 2009) was incorporated into the RACE process to differentiate RNAs of opposite polarities. Briefly, 5 µg of DNA-free total RNA were first subjected to decapping, or removal of pyrophosphates, by incubating in 12.5 U RppH

(New England BioLabs) with 20 U RNase Inhibitor (ThermoFisher Scientific) and 2 µL of 10X Thermopol Buffer (NEB) in a total volume of 20 μl for 1 hour at 37 °C . The pyrophosphate-removed RNA was then circularized by adding 20 μl 10X T4 ligase buffer, 20 U T4 RNA ligase (NEB), 2 μL 10 mM ATP, 40 μL 50% PEG8000 and nuclease-free water to 200 μL and incubating in 16 °C water bath for overnight. Then

RNA was column purified by using an RNA concentrator kit (Zymo Research). Only samples treated with ttsgR_Temp have undergone circularization treatment; other samples were directly processed to next steps.

The purified RNA (2 µg) was then reverse transcribed by a TCV strand specific primer with a unique 5′ end tag. Subsequently, a tag-specific primer and a downstream forward TCV-specific primer were used to amplify an (+)-strand-specific PCR product.

Similarly, a tag-specific primer and an upstream reverse TCV-specific primer were used to amplify an (-)-strand-specific PCR product. All the primers used for reverse transcription and strand specific PCR are list in Table 3. The PCR products were gel purified and cloned to pJET1.2 cloning vector (ThermoFisher Scientific) for sequencing.

Acknowledgement

We thank the labs of Drs. Lucy Stewart, Peg Redinbaugh, Sally Miller for generous equipment sharing. This study was supported by an NSF award, a SEEDS grant from the

Ohio Agricultural Research and Development Center, Graduate Assistantships from OSU and OARDC to R.S., as well as tuition assistances to S.Z. and R.S. from the Department of Plant Pathology, OSU.

Figures.

Figure 3.1: Overexpression of p88 unveils a novel TCV RNA. (A) Schematic representation of the TCV genome with encoded proteins shown as rectangles. The two subgenomic RNAs, sgRNA1 and 2, serving as mRNA for movement proteins (p8 and p9) and CP respectively, were also depicted. ORFs of p9, CP positions to TCV genome were

76 marked. (B) Constructs used in experiments leading to C and D. All constructs were delivered into N. benthamiana cells using agro-infiltration. Both TCV replicon constructs use the 2X35S promoter and T35S terminator to drive the transcription of the infectious

TCV RNA. In both constructs the TCV cDNA was modified by replacing the CP ORF with that of mCherry. In TCV_sg2R, both p28 and p88 ORFs are intact, whereas in

Δp88_sg2R the p88 ORF was terminated prematurely by a 4-nt deletion at nt #1402.

Three different promoters (2X35S, Core35S, Micro) were used to drive the overexpression of p88 to different extents. The p88 coding region in these three constructs were all fused to a C-terminal double HA epitope to facilitate their detection with an HA antibody. (C) Different expression levels of three p88 constructs revealed by Western blotting. Anti-HA antibody was used to detect HA tagged p88. (D)

Repression and replicational complementation of TCV replicons by overexpression of p88 to different extent. Northern blotting was used to determine the accumulation of TCV

RNA species. EB: ethidium bromide stained agarose gel serving as the loading control for Northern blotting. (E) The effects of RdRp-null and RdRp functioning p88 variants on ttsgR accumulation during TCV_sg2R replication revealed by northern blotting. (F)

Overexpressing of p88 unmasked on replication of wt TCV revealed by northern blotting.

Figure 3.2: Biogenesis of ttsgRNA is replication dependent. (A) Diagram shows the mutation in sg1 promoter, and deletion in sg2 promoter in the new TCV replicons.

Substitutions in sg1 promoter are shown in red. Twelve-nt deletion in sg2 promoter are indicated by a red line. (B) Abolish TCV sgRNA(s) production have modest effects on

78 ttsgR production revealed by northern blotting. (C) Constructs used in experiments leading to D, E, and F. All constructs were modified from TCV_sg2R. [p28stop]_sg2R only translates the first 36 aa of p88 before encountering an introduced stop codon, thus it is defective in replication. Constructs sg2R_Temp 5’ trimmed to TCV position 2468 was designed to transcribe sgRNA2 only. Construct ttsgR_Temp were 5’ trimmed to immediately downstream of mCherry insertion site – TCV nt position 3386. The arrows in the bottom of the panel shows the relative position of probes used in the experiments leading to F. (D) ttsgR independent accumulation in the absence of gRNA revealed by northern blotting. (E) Partial repeat of experiments leading to D. (F) Coarse mapping of the 5’ end of the ttsgR using a set of probes annealing to different positions of ttsgR_Temp transcript revealed by northern blotting.

Figure 3.3: Circularization RACE mapping the ends of ttsgR. (A) Mapping procedure diagram: A combination of Circularization RACE and strand-specific RT-PCR

(amplifying positive sense as an example). Transcript of ttsgR_Temp is shown as gradient rectangles, whereas dark terminus indicates the 5’ end and light the 3’ end. Dark

80 arrow with red tail (a unique tag) is to show the reverse transcription primer which is located upstream of PCR forward primer (dark arrow above transcript). The relative position of these two primers indicate they cannot prime amplification without circularization. (B) Electrophoresis results of strand-specific RT-PCR products. Arrows indicate the expected bands. (C) A summary of sequencing results of strand-specific circularization RACE. Numbers above the sequences indicate the corresponding nts’ position in the TCV genome. Extra nts at 3’-5’ junction are shown in orange (D) Putative

RNA stem loop structure in the transcripts of ttsgR_Temp and its variants. “-” indicates base pairing. Substitutions in the stem-loop in ttsgR_Temp variants are shown in red.

Disrupting of base pair is shown by longer distance between the two nts. Numbers indicate the pointing nts’ positions in the TCV genome. Number in the box shows the 5’ end of the template sequence that derived from TCV. The 5’ end of ttsgR was highlighted in light blue. (E) The RNA stem loop flanking the 5’ end of ttsgR is not essential for ttsgR biogenesis revealed by northern blotting.

Figure 3.4: Specific sequence requirement for ttsgR accumulation. (A) Putative RNA base pairing between the 3’ and 5’ end of ttsgR and its variants. The construct ttsgR_S and its variants have only 28 extra nts at the 5’ end of ttsgR as indicated by the number in a box. The 5’ end of ttsgR was highlighted in light blue. Substitutions in the triple G-C pair are shown in red. (B) The 3’ C triplet is critical for ttsgR replication revealed by northern blotting. Double arrows in Lane 6 indicate two unseparated bands, an arrow in

Lane 9 indicats a band bigger than that of ttsgR. (C-D) Summaries of mapping results of

82 ttsgR derived from template ttsgR_S and S_m5p, respectively. TCV derived sequences are shown in light blue, plasmid derived in green, introduced mutation in red, extra nts at

3’-5’ junction in orange. Numbers above the sequences indicate the pointing nts’ positions in the TCV genome.

Figure 3.5: A 5’ terminal G3(A/U)4 motif is critically important for ttsgRNA accumulation. (A) Putative RNA base pairings between the 3’ and 5’ end of ttsgR and its variants from templates with an exact 5’ terminus of ttsgR. Substitutions in the triple G-C pair are shown in red. Plasmid sequence is shown in green. (B) Template as short as ttsgR itself accumulates in the presence of p28 and p88 revealed by northern blotting. (C-G)

Summaries of mapping results of ttsgR derived from template ttsgR_S2, S2_m1, S2_m2,

S2_m3 and S2_m4, respectively. TCV derived sequences are shown in light blue, plasmid derived in green, introduced mutation in red, extra nts at 3’-5’ junction in orange.

Numbers above the sequences indicate the pointing nts’ positions in the TCV genome.

Figure 3.6: The 3’ terminal CCC triplet is essential for efficient ttsgRNA accumulation. (A) Putative RNA base pairings between 3’ C triplets and 5’ G triplets in ttsgR and its variants from templates ttsgR_S2a and its variants. Substitutions in the second G-C pair are shown in red. Plasmid sequence is shown in green. (B) Mutating single nt in the triple C dramatically reduces the accumulation of ttsgR revealed by northern blotting. (C-F) Summaries of mapping results of ttsgR derived from template

5u, 5a, 3u and 3a, respectively. TCV derived sequences are shown in light blue, plasmid derived in green, introduced mutation in red, extra nts at 3’-5’ junction in orange, the reversions of mutated C in 3u 3a are also shown in orange. Numbers above the sequences indicate the pointing nts’ positions in the TCV genome.

Figure 3.7: The GGGUAAA motif is needed for ttsgRNA synthesis from infectious

TCV RNA templates. (A) Schematic representation of the TCV replicon m1 that mutates the GGGUAAA motif. (B) Mutation in GGGUAAA motif abolished ttsgR production revealed by northern blotting. (C) Comparison of replication efficiency in systemic infected leaves of wt TCV and m1 mutant.

Table 1: gBlocks used in this study

Construct gBlock sequence (5’ → 3’) Enzyme site p88mGDD GTCGCATGTCCGGAGATGTTAACACAGCCTTGGGCAACTGCCTACTGGCTTGCTCTATC Core35S::p88

ACCAAGTACTTAATGAAGGGAATCAAATGCAAATTAATCAACAACGttGcaGTGTGTGCT digested by

GTTCTTCGAAGCTGATGAAGTCGACAGGGTGCGCGAAAGGCTGC HpaI & SalI

Δsg1 TGTGCTGTTCTTCGAAGCTGATGAAGTCGACAGGGTGCGCGAAAGGCTGCATCATTGGA TCV_sg2R

TCGACTTTGGGTTTCAATGCATAGCGGAAGAACCACAATACGAATTGGAGAAAGTTGA digested by

ATTTTGCCAGATGTCCCCTATTTTCGATGGTGAAGGGTGGGTCATGGTCAGAAACCCCC BamHI &

GTGTGAGCCTCTCCAAGGACAGCTACAGCACCACACAATGGGCGAATGAGAAAGATGC SalI

AGCCAGATGGTTGGCTGCCATCGGAGAGTGTGGCTTGGCTATTGCAGGTGGCGTACCAG

TGTTACAATCATATTATTCTTGCCTGAAGAGGAATTTTGGACCCCTGGCCGGGGACTAC

AAGAAGAAGATGCAAGATGTTTCCTTTGATAGTGGATTCTACAGGTTATCCAAGAACGG

GATGAGGGGCAGCAAAGACGTGTCCCAAGATGCTAGGTTCAGCTTTTAtCGcGGcTTCG

GCTACACTCCAGACGAGCAGGAAGCGCTTGAGGAGTACTACGACAACCTCGAACTGCT

CTGTGAGTGGGACCCCACcGGtTATAAAGAAGAACTTAGTGATAGATGGATCCTGAACG

AATTCCCTACAACTCTCTAA

Δsg2 GtTATAAAGAAGAACTTAGTGATAGATGGATCCTGAACGAATTCCCTACAACTCTCTAA TCV_sg2R

GCGACAGCGACGCAACAGGAAAACGGAAGAAAGGCGGAGAGAAAAGTGCGAAGAAG digested by

AGATTGGTAGCTAGCCACGCGGCTAGCTCTGTTTTAAACAAGAAAAGAAATGAAGGTT BamHI &

CTGCTAGTCACGGGGGTACTTGGGTTATTGTTGCTGATAAAGTGGAAGTCTCAATCAAC NcoI

TTCAACTTCTAATCAGACATGTCAGTGcccgacgtccccGTGGGTAATATATGCTTTCTACAAC

TCTCTCTCACTGGTCCTCCTACTTTGTCATCTGATTCCTGAAATCAAACCGATTCACACA

TCCTACAACACACACGACTCATCGAAGCAGCAACACATAAGCATCAACACTGGAAACG

GAAAATAACCATGGTGAGCAAGGGCGAGGAGGACAACATG

Δsg1&2 gBlocks for Δsg1 and Δsg2 were both used TCV_sg2R

digested by

SalI & NcoI

88 sg2R_Temp CATTTCATTTGGAGAGGACCTCGAGGCTAGCTCTGTTTTAAACAAGAAAAGAAATGAA TCV_sg2R

GGTTCTGCTAGTCACGGGGGTACTTGGGTTATTGTTGCTGATAAAGTGGAAGTCTCAAT digested by

CAACTTCAACTTCTAATCAGACATGTCAGTGCCCGACGTCCCCGTGGGTAATATATGCT AtaII & XhoI ttsgR_Temp CATTTCATTTGGAGAGGACCTCGAGTGGCCACCTACGGCCAAGGAGCCAATGATGCCG TCV_sg2R

CCCAACTCGGTGAAGTGCGAGTCGAGTACACCGTGCAGCTCAAGAACAGAACTGGCTC digested by

AACCAGCGACGCCCAGATTGGGGACTTCGCAGGTGTTAAGGACGGACCCAGGCTGGTT SpeI & XhoI

TCATGGTCCAAGACCAAGGGGACAGCTGGGTGGGAGCACGATTGTCATTTTCTCGGAA

CCGGAAACTTCTCGTTGACATTGTTCTACGAGAAGGCGCCGGTCTCGGGGCTAGAAAAC

GCAGACGCCTCTGACTTCTCGGTCCTGGGAGAAGCCGCAGCAGGTAGTGTCCAATGGG

CAGGAGTGAAGGTAGCAGAAAGGGGACAAGGCGTGAAAATGGTCACAACTGAGGAGC

AGCCAAAGGGTAAATGGCAAGCACTCAGAATTTAGTACGGTAATAGTGTAGTCTTCTCA

TCTTAGTAGTTAGCTCTCTCTTATATTAAGAAAAGAAAACAAAAACCCCCAGGTCGCTT

TATTTTGACCTGTGTTAGGGACCAAAAACGGTGGCAGCACTGTCTAGCTGCGGGCATTA

GACTGGAAAACTAGTGCTCTTTGGGTAACCACTA

89 ttsgR_mL CATTTCATTTGGAGAGGACCTCGAGTGGCCACCTACGGCCAAGGAGCCAATGATGCCG TCV_sg2R

CCCAACTCGGTGAAGTGCGAGTCGAGTACACCGTGCAGCTCAAGAACAGAACTGGCTC digested by

AACCAGCGACGCCCAGATTGGGGACTTCGCAGGTGTTAAGGACGGACCCAGGCTGGTT SpeI & XhoI

TCATGGTCCAAGACCAAGGGGACAGCTGGGTGGGAGCACGATTGTCATTTTCTCGGAA

CCGGAAACTTCTCGTTGACATTGTTCTACGAGAAGGCGCCGGTCTCGGGGCTAGAAAAC

GCAGACGCCTCTGACTTCTCGGTCCTGGGAGAAGCCGCAGCAGGTAGTGTCCAATGGG

CAGGAGTGAAGGTAGCAGAAAGGGGACAAGGCGTGAAAATGGTCACAAgTcAcGAcgAc

ggAAAGGGTAAATGGCAAGCACTCAGAATTTAGTACGGTAATAGTGTAGTCTTCTCATC

TTAGTAGTTAGCTCTCTCTTATATTAAGAAAAGAAAACAAAAACCCCCAGGTCGCTTTA

TTTTGACCTGTGTTAGGGACCAAAAACGGTGGCAGCACTGTCTAGCTGCGGGCATTAGA

CTGGAAAACTAGTGCTCTTTGGGTAACCACTA ttsgR_mR CATTTCATTTGGAGAGGACCTCGAGTGGCCACCTACGGCCAAGGAGCCAATGATGCCG TCV_sg2R

CCCAACTCGGTGAAGTGCGAGTCGAGTACACCGTGCAGCTCAAGAACAGAACTGGCTC digested by

AACCAGCGACGCCCAGATTGGGGACTTCGCAGGTGTTAAGGACGGACCCAGGCTGGTT SpeI & XhoI

TCATGGTCCAAGACCAAGGGGACAGCTGGGTGGGAGCACGATTGTCATTTTCTCGGAA

CCGGAAACTTCTCGTTGACATTGTTCTACGAGAAGGCGCCGGTCTCGGGGCTAGAAAAC

GCAGACGCCTCTGACTTCTCGGTCCTGGGAGAAGCCGCAGCAGGTAGTGTCCAATGGG

CAGGAGTGAAGGTAGCAGAAAGGGGACAAGGCGTGAAAATGGTCACAACTGAGGAGC

AGCCAAAGGGTAAATccgAAcgAgTgAcAATTTAGTACGGTAATAGTGTAGTCTTCTCATC

TTAGTAGTTAGCTCTCTCTTATATTAAGAAAAGAAAACAAAAACCCCCAGGTCGCTTTA

TTTTGACCTGTGTTAGGGACCAAAAACGGTGGCAGCACTGTCTAGCTGCGGGCATTAGA

CTGGAAAACTAGTGCTCTTTGGGTAACCACTA ttsgR_mLR CATTTCATTTGGAGAGGACCTCGAGTGGCCACCTACGGCCAAGGAGCCAATGATGCCG TCV_sg2R

CCCAACTCGGTGAAGTGCGAGTCGAGTACACCGTGCAGCTCAAGAACAGAACTGGCTC digested by

AACCAGCGACGCCCAGATTGGGGACTTCGCAGGTGTTAAGGACGGACCCAGGCTGGTT SpeI & XhoI

TCATGGTCCAAGACCAAGGGGACAGCTGGGTGGGAGCACGATTGTCATTTTCTCGGAA

CCGGAAACTTCTCGTTGACATTGTTCTACGAGAAGGCGCCGGTCTCGGGGCTAGAAAAC

GCAGACGCCTCTGACTTCTCGGTCCTGGGAGAAGCCGCAGCAGGTAGTGTCCAATGGG

CAGGAGTGAAGGTAGCAGAAAGGGGACAAGGCGTGAAAATGGTCACAAgTcAcGAcgAc

ggAAAGGGTAAATccgAAcgAgTgAcAATTTAGTACGGTAATAGTGTAGTCTTCTCATCTTA

GTAGTTAGCTCTCTCTTATATTAAGAAAAGAAAACAAAAACCCCCAGGTCGCTTTATTT

TGACCTGTGTTAGGGACCAAAAACGGTGGCAGCACTGTCTAGCTGCGGGCATTAGACT

GGAAAACTAGTGCTCTTTGGGTAACCACTA ttsgR_S GTTCATTTCATTTGGAGAGGACCTCGAGAAATGGTCACAACTGAGGAGCAGCCAAAGG TCV_sg2R

GTAAATGGCAAGCACTCAGAATTTAGTACGGTAATAGTGTAGTCTTCTCATCTTAGTAG cut by

TTAGCTCTCTCTTATATTAAGAAAAGAAAACAAAAACCCCCAGGTCGCTTTATTTTGAC SpeI & XhoI

CTGTGTTAGGGACCAAAAACGGTGGCAGCACTGTCTAGCTGCGGGCATTAGACTGGAA

AACTAGTGCTCTTTGGGTAACCACTA

Table 2: Primers used in this study

Construct Primers for inserting fragment (5’ → 3’) Backbone and

enzyme sites

Micro::p88 F: GCTTGTTAGGCCTAAACCTGCAGGAGATCCAGATTTATAGGGTCCTAATG Core35S::p88 digested

R: TATGTTGTGTTGAGAATTCTCGAGGTTAATCACTAAGAAATTTAGGGTTG by XhoI & SdaI

Sg2R-m1 F1: CTTCTAATCAGAMATGTCAGTGC TCV_sg2R digested

(Use TCV_sg2R R1: TTTGGCTGCTCCTCAGTTGTG by AatII & SpeI as template) F2: CACAACTGAGGAGCAGCCAAAaGGcAAgTGGCAAGCACTCAGAA

R2: GATTTTAGTGGTTACCCAAAGAGCA wt-m1 F1: CTTCTAATCAGAMATGTCAGTGC TCV_sg2R digested

(Use TCV_wt as R1: TTTGGCTGCTCCTCAGTTGTG by AatII & SpeI template) F2: CACAACTGAGGAGCAGCCAAAaGGcAAgTGGCAAGCACTCAGAA

R2: GATTTTAGTGGTTACCCAAAGAGCA

S_m5p F: CATTTCATTTGGAGAGGACCTCGAGAAATGGTCACAACTGAGGAG- ttsgR_S digested by

CAGccaAACCATAAATGGCAAG XhoI & XbaI

R: GAGACTGGTGATTTTTGCGGACTCT

S_m3p F: CTATATAAGGAAGTTCATTTCATTTGG ttsgR_S digested by

R: ACTGGTGATTTTTGCGGACTCTAGAccaCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

S_m5p3p F: CATTTCATTTGGAGAGGACCTCGAGAAATGGTCACAACTGAG- ttsgR_S digested by

GAGCAGccaAACCATAAATGGCAAG XhoI & XbaI

R: ACTGGTGATTTTTGCGGACTCTAGAccaCAGGCCCCCCCCCC-

GCGCGAGGAGGGAGGCTATCTT ttsgR_S2 F: CATTTCATTTGGAGAGGACCTCGAGGGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2_m1 F: CATTTCATTTGGAGAGGACCTCaAaGGcAAgTGGCAAGCACTCAGAA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2_m2 F: TTTCATTTGGAGAGGACCTCaAaccgtAAaGGCAAGCACTCAGAATTTA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2_m3 F: TTTCATTTGGAGAGGACCTCaAGGGTAAATccgtAGCACTCAGAATTTA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2_m4 F: TTTCATTTGGAGAGGACCTCaAaccgtAAaccgtAGCACTCAGAATTTA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2a F: CATTTCATTTGGAGAGGACCTtaAGGGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2a_3u F: CATTTCATTTGGAGAGGACCTtaAGGGTAAATGGCAAGCACTCAGAA ttsgR_S cut digested

R: ACTGGTGATTTTTGCGGACTCTAGAGaGCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

S2a_3a F: CATTTCATTTGGAGAGGACCTtaAGGGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: ACTGGTGATTTTTGCGGACTCTAGAGtGCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

S2a_5u F: CATTTCATTTGGAGAGGACCTtaAGtGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2a_5u3u F: CATTTCATTTGGAGAGGACCTtaAGtGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: ACTGGTGATTTTTGCGGACTCTAGAGaGCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

S2a_5u3a F: CATTTCATTTGGAGAGGACCTtaAGtGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: ACTGGTGATTTTTGCGGACTCTAGAGtGCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

S2a_5a F: CATTTCATTTGGAGAGGACCTtaAGaGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: GAGACTGGTGATTTTTGCGGACTCT XhoI & XbaI

S2a_5a3u F: CATTTCATTTGGAGAGGACCTtaAGaGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: ACTGGTGATTTTTGCGGACTCTAGAGaGCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

S2a_5a3a F: CATTTCATTTGGAGAGGACCTtaAGaGTAAATGGCAAGCACTCAGAA ttsgR_S digested by

R: ACTGGTGATTTTTGCGGACTCTAGAGtGCAGGCCCCCCCCCC- XhoI & XbaI

GCGCGAGGAGGGAGGCTATCTT

Table 3: Primers used for circularization RACE

Sense Type Primer name and sequence (5’ → 3’) Used in

Tag-TCV3905R: ttsgR_temp, (+) RT-primer GGAACGTCATGGTGGTGACAAAAGGTCCCTAACACAGGTCAAAAT ttsgR_S, S_m5p,

PCR-For TCV3963F: TGGGTAACCACTAAAATCCCGAA S_m3p,

PCR-Rev Tag-ssRT: GGAACGTCATGGTGGTGACAAAA S_m5p3p

Tag-TCV3797F: (-) RT-primer GGAACGTCATGGTGGTGACAAAATAGTACGGTAATAGTGTAGTCTTC ttsgR_temp PCR-F Tag-ssRT: GGAACGTCATGGTGGTGACAAAA

PCR-R TCV3794R: TTCTGAGTGCTTGCCATTTAC

Tag2-TCV3999F: ttsgR_S2 (-) RT-primer GGAACGTCATGGTGGTGACTATGTGACCTTCCGAACTAAAAG S2_m1-m4

PCR-For Tag-ssRT2: GGAACGTCATGGTGGTGACTAT S2a_3u, 3a

PCR-Rev TCV3980R: GATTTTAGTGGTTACCCAAAGAGCA S2a_5u, 5a

CHAPTER 4

Repression of Flock House Virus replication by its replication protein pA:

Preliminary Findings

Abstract

RNA dependent RNA polymerase (RdRp) and auxiliary replication protein (ARP) encoded by many positive sense (+) RNA viruses are essential for their replication. In our previous study we found that overexpression of the p88 RdRp or the p28 ARP of turnip crinkle virus (TCV) repressed the replication of TCV. Some (+) RNA viruses encode a single multi-functional RdRp that directs the entire replication cycle. One such virus is the Flock house virus (FHV). In this study, we investigated how overexpression of its

RdRp, the pA affects its replication. Our preliminary experiments demonstrated that pA, when over-expressed in N. benthamiana cells from a non-viral source, was able to complement the replication of FHV mutants lacking their own functional RdRp. More importantly, the same RdRp modestly repressed the replication of FHV replicons encoding a functional RdRp.

Introduction

Many RNA viruses, especially viruses with single stranded, positive sense (+) RNA genomes are among the most devastating pathogens of human beings. As a vivid reminder, we are currently experiencing the pandemic spread of one of such (+) RNA

99 virus, SARS-COV-2, the causative agent of COVID-19 (Hargreaves et al. 2020). Since its emergence less than six months ago, COVID-19 has infected millions of people and claimed hundreds of thousands of lives (https://coronavirus.jhu.edu/map.html). A lesser known fact is that many (+) RNA viruses are similarly devastating for their ability to wipe out our crops and herds, such as cassava streak brown mosaic virus infecting cassava (Tomlinson et al. 2018) and porcine reproductive and respiratory syndrome virus infecting pigs (Renukaradhya et al. 2015). Despite the difference in host range, genome size and genome organization, all (+) RNA viruses encode an RNA dependent RNA polymerase (RdRp) that is essential for the replication of viral genomic (g)RNAs, and in some viruses the transcription of subgenomic (sg)RNAs. In addition to RdRp, many (+)

RNA viruses, such as viruses in plant-infecting tombusviruses (Souza and Carvalho

2019) and animal-infecting flaviviruses (Mazeaud et al. 2018), also encode at least one auxiliary replication protein (ARP), that is also needed for the multiplication of viral genomic RNAs.

In our previous studies using turnip crinkle virus (TCV) as a model, we found that p28, the TCV encoded ARP, trans-complemented a defective TCV lacking p28, yet repressed the replication of another TCV replicon encoding wildtype p28 (Zhang et al.

2017). Surprisingly, the TCV encoded RdRp, also known as p88, likewise exhibited both trans-complementary and repressive activities when tested with replicons encoding defective or wildtype p88 (Zhang et al. 2019). However, unlike repression caused by p28 which acted on all TCV RNA species, the repression caused by high levels of p88 preferentially subdued longer TCV RNAs, leaving smaller RNA species over- accumulated. It has yet to be examined whether the sole RdRp encoded by viruses like

100

FHV also acts on its own genomic RNA in a manner similar to TCV RdRp, namely functioning as a repressor or trans-complementor depending on whether the viral RNA encodes its own RdRp.

To address this question, here we used FHV as a model and examined how pA, the

FHV encoded RdRp, would affect the replication of various of FHV RNA1 replicons.

FHV is a small bipartite (+) RNA virus. Its ~3.1 kilobase (kb) genomic RNA1 encodes the pA RdRp. Its 1.4 kb genomic RNA2 encodes the pα, which is the capsid protein (CP) precursor that is further cleaved into mature CP after provirion assembly (Fig. 4.1A).

RNA2 is not needed for RNA1 replication in individual infected cells. During replication,

RNA1 also templates the production of RNA3, a sgRNA 3’ co-terminal with RNA1

(Guarino et al. 1984). RNA3 encodes two proteins B1 and B2 (Fig. 4.1A). B1 has the same amino acid (aa) sequence as the last 102-aa residues of pA, whose function is currently unknown. B2 is an RNA silencing suppressor (Li et al. 2002), whose coding region overlaps with that of pA and B1 but is decoded in the +1 reading frame relative to pA.

FHV was first isolated from the grass grub Costelytra zealandica (Dearing et al.

1980), but its genomic RNAs also replicate in insect cell lines (Selling and Rueckert

1984), cells of model plant Nicotiana benthamiana (Selling et al. 1990), some mammalian cell lines (Ball et al. 1992), and cells of budding yeast (Price et al. 2000).

Since FHV RNA1 is capable of self-replication, our study only focused on RNA1. We adopted an Agrobacterium (Agro) infiltration method to deliver DNA-launched FHV

RNA1 replicons and pA transient expression construct to N. benthamiana leaf cells to examine the effects of overexpressing pA on replication of various RNA1 replicons. Our

101 preliminary data showed over-expressing pA in trans, was able to complement the replication of FHV mutants lacking their own functional RdRp, and only modestly repressed the replication of FHV replicons encoding a functional RdRp.

Results

An agro-infiltration-delivered FHV RNA1 replicon replicates in N. benthamiana leaf cells. To adopt the agro-infiltration method, FHV RNA1 cDNA was cloned into a binary vector pAIDE. The resultant replicon FR_FL was then modified to generate a serial of FHV RNA1 replicons, including Acc65I RE site containing replicon

FR1_Acc65I; defective replicons FR1_Def, FR1(fs)_FL, and FR1(5’rp)_FL; and fluorescent protein expressing replicons FR1_sgG and FR1_sgR (Fig. 4.1B). Transient expression constructs expressing FHV RNA1 encoded proteins were also generated by cloning into the binary vector pAIDE (Fig. 4.1C). Each construct was agro-infiltrated into

N. benthamiana leaves to test their self-replicability. note in all experiments described in the current study, a construct expressing the p19 silencing suppressor of tomato bushy stunt virus (TBSV) was always included in the agro-infiltration to mitigate RNA silencing-mediated transcript degradation. Our northern blot results (Fig. 4.1D) showed full-length FHV RNA1 replicon FR1_Fl was able to self-replicate producing strong bands of RNA1 and subgenomic RNA3 (Lane 3). So was FR1_Acc65I (Lane 4), which contains 3 nts insertion after protein A stop codon. Two fluorescent gene containing replicons, in which the fluorescent proteins are expressed as N-terminal B2 fused protein

(Price et al. 2000), were only able to effectively replicate recombined RNA3 but not

RNA1 (Lane 5, 6). Surprisingly, two replicons FR1(5’rp)_Fl and FR1(5’rp)_Acc65I,

102 which were made by replacing the FHV 5’ untranslated region (UTR) sequence with a

40-nt sequence derived from the 5’ UTR of carnation mottle virus (CarMV), were able to replicate though less efficiently (Lane 11, 12). Other mutant replicons were all defective as expected. They are FR1_Def, in which the last 205 nt of RNA1 was deleted; two pA frameshift replicons, in which the translation of pA was expected to stop at the 113th aa

(Price et al. 1996); and two mutant replicons in which the conserved GDD motif of RdRp was mutated to STD (Van Wynsberghe et al. 2007). We also tested four pA transient expression constructs. They were untagged pA, GDD motif mutated pA, C-terminal G11-

Flag tagged pA, N-terminal G11-Flag tagged pA. These latter constructs could not replicate because they lacked the key cis-acting elements present within the UTRs of

FHV RNA1 (lanes 13-16).

Tag-free pA is fully functional when provided in trans but lacks repressive activity. Next, we tested how co-infiltration of pA transient expression construct with an

FHV RNA1 replicon would affect the replication of the replicon. Our results (Fig. 4.2A) showed providing pA in trans had limited effects on wildtype FHV replicon that encodes intact pA (compare lanes 4-5 to lanes 10-11), but effectively complemented the replication of replicons that were defective in pA, including the frameshift and GDD mutants, which upon pA complementation replicated to levels similar to wildtype

(compare lanes 6-7 to lanes 12-13). These results also suggesting mutated pAs expressed from the replicons, had no dominate-negative effects that prevent trans-expressed pA from complementing replication.

Providing pA in trans also significantly increased the replication of FR1(5’rp)_Fl and

_Acc65I, which is evident for the potent accumulation of RNA1 and RNA3, especially

103

RNA3 (Fig. 4.2A, compare lanes 14-15 to lanes 8-9). This result suggested that replacing

FHV 5’ UTR with that of CarMV 5’ UTR probably had a major effect on the translation of pA from RNA1, but a relatively small effect on the replicability of RNA1 (Fig. 4.2A, compare lanes 14-15 to lanes 12-13). Some of these results were confirmed in a separate experiment as showed in Fig. 4.2B. Furthermore, deletion of the last 205 nts including partial pA coding region and 3’UTR totally abolished the replication of FHV RNA1, which is probably due to the loss of essential cis-acting RNA elements, as providing pA in trans did not rescue its replication (Fig. 4.2A, compare lanes 19-20 to lanes 9-10).

An N-terminally tagged pA is less capable of replicational complementation, whereas a C-terminally tagged pA is more potent at repression replication. In our previous studies on TCV, we found N-terminal tagging had modest impact on p88’s ability to suppress replication of TCV replicons (Zhang et al. 2019). But N terminal tagging effectively abolished the repression function of p28, while C terminal tagging reinforced the repression function of p28 (Zhang et al. 2017). Since pA embodies functions of a RdRp and an ARP, we wonder whether N-terminal or C-terminal tagging would unmask the repression function of pA.

To this end, we replaced the untagged pA with N-terminal or C-terminal G11-Flag tagged pA in co-infiltration test. Our results showed N-terminal tagged pA still did not show any significant repression effect (Fig. 4.2C), but C-terminal tagged pA showed substantial increase in suppression function as evident by the reduce in RNA1 and RNA3 accumulation in co-infiltrated samples (Fig. 4.2D, compare lane 10 to 3, 12 to 4). We also found N terminal tagged pA failed to complement replication of pA defective replicons.

104

This result indicated N-terminal tagging may abolished the translation or disrupted the

RdRp function of pA. Such loss of function was not observed in C-terminal tagged pA.

Tag-free B1 also represses the replication of FHV RNA1 replicons. B1 and B2 are translated from RNA1-derived RNA3. And B1 presents the last 102-aa residues of pA, but its function remains enigmatic. We speculated that B1 may also involve in FHV replication. Therefore, we examined how B1 overexpression would affect the self- replication of FHV RNA1. B2 is a known RNA silencing suppressor, which protects

FHV RNAs from Dicer mediated degradation. Here, we also tested B2 as well. As shown in Fig. 4.3A, co-infiltration with B1, substantially reduced the replication of most replicons (compare lanes 12-13,17-18 to 3-4, 8-9), but slightly increased the replication of the fluorescent gene recombined RNA3 in FR1_sgR replicon (compare lane14 to 5).

B2 could still increase the FHV RNA stability even though p19 was additionally provided, which indicated that B2 and p19 may targeting different component of RNA silencing pathway.

Sequential infiltration reveals the potent suppression function of pA. A speculated biological function of ARP or RdRp induced replication suppression is to avoid accumulation of mutation by restricting the replication of progenies or secondary invading viruses that are the same as primary infecting virus. From here, we hypothesize that pre-delivery of pA and other proteins involved in FHV replication may pose a strong repression effect on secondary delivered FHV replicon. To test this hypothesis, constructs expressing FHV RNA1 encoded proteins were first delivered into N. benthamiana leaves individually, then replicon FR1_Acc65I was delivered at second day. Our results (Fig.

4.4A) showed pA preferentially suppressed the replication of RNA1 but enhanced the

105 replication of RNA3. Such preferentially accumulation of smaller RNA species has also been recorded in overexpressing p88 in our TCV study (Zhang et al. 2019). B1 also repressed the replication of FR1_Acc65I, but in a different way, in which shorter RNA3 were the main target.

Why would pA expressed from replicon failed to exert suppression function? One possible explanation is the proteins expressed from RNA3 could potentially interact with pA that blocks the replication repression domains in pA. So, here we tested the impact of predelivery of different combinations of pA, B1, and B2 on the replication of one-day- later delivered FR1_Acc65I. Our results (Fig. 4.4B) showed pA and B1 combination had strongest repression effect on both RNA1 and RNA3. And the discriminate suppression we saw earlier caused by single protein was not observed here. Adding B2 to pA, B1, or their combination all reduced the repression function of pA and B1. Here, we also showed pA(mGDD) could dramatically shut down the replication of FHV RNA1

Protein A prefers to replicate pA defective replicons over replicon with long insertion. We have made several pA defective replicons. Since pA is also responsible for

RNA3 transcription, defective in pA theoretically makes these pA defective replicons fail to produce B1 and B2 but could still produce defective pA. We want to test how these replicons produced pAs would impact the replication of secondary delivered replicon. To distinguish the defective replicon, we used FR1_sgG as secondary replicon because it has

~700 insertion. The previous observation that pA preferentially accumulate smaller RNA species was confirmed here, as the RNA3r was accumulated to a higher level (Fig. 4.5A, lanes 13-14 to lanes 3-4). We were surprised to learn that pA translated from FR1_sgG abandoned FR1_sgG to replicate defective replicons FR1(fs) and FR1(mGDD).

106

Discussion

Protein A is the sole FHV encoded protein that is required for FHV replication. The goals of current study were to determine whether pA expressed from a non-replicating mRNA could trans-compliment the replication of FHV RNA1 replicons encoding defective pA and whether it could suppress the replication of RNA1 replicons encoding an intact pA. Under these goals, we extended our investigation on the counter-intuitive functions of viral encoded replication proteins that has been revealed in our previous study using TCV as a model (Zhang et al. 2019; Zhang et al. 2017). In the TCV studies, we showed that p28 the ARP trans-complemented the replication of TCV replicon lacking its own p28 but repressed the replication of replicon encoding intact p28. The

TCV encoded RdRp also worked in a similar way. Different from TCV in which two viral proteins are encoded for replication, FHV just encode pA for its replication. Here, we investigated whether our previous discovery on TCV replication proteins is transposable to virus that encodes sole replication protein. And we chose FHV to test.

We took advantage of FHV’s ability to replicate in the leaf cells of model plant N. benthamiana (Selling et al. 1990), on which we have developed a mature system in TCV studies. We first made 11 FHV RNA1 replicons and tested their automatous replication in

N. benthamiana. In this experiment, we confirmed the automatous replication of FR1_Fl,

FR1_Acc65I; preferential replication of recombined RNA3 in fluorescent gene inserted replicons such as FR1_sgG, FR1_sgR; inefficient replication of 5’UTR replaced replicons; and no replication of pA defective replicons including FR1_Def, FR1(fs)_Fl,

107 and FR1(mGDD)_Fl. Then we co-delivered these replicons with pA and its variants to test the impact of trans-expression pA on replication of those replicon.

The trans-complementation by pA was easily achieved by tag-free pA and C-terminal tagged pA, but not N-terminal tagged pA. N-terminal sequences of pA contain a transmembrane domain that is critical for mitochondrial localization and replication spherule formation (Miller and Ahlquist 2002), it is reasonable to speculate that N- terminal tagging affects the function of these sequence. Another possibility is N-terminal tagging makes the RNA unable to translate or pA unstable. In our previous study on

TCV, we found p28 trans-complementation resembled the replication of wt TCV, but p88 trans-complementation preferentially replicated small RNA species, specifically sgRNA2 and the 283-nt ttsgR (Zhang et al. 2019; Zhang et al. 2017). In current study, the trans-complementation by pA was efficient for replication of both RNA1 and RNA3, similar to what was observed with TCV p28 trans-complementation of TCV replication.

We failed to detect repression effect of pA using untagged pA. Then we managed to detect repression effect by using G11-Flag tagged pA. And at last, we succeeded to detect repression effect of tag-free pA through a sequential infiltration technique. Tagging pA is inspired by our previous experience on TCV that the repression function of p28 was reinforced by C-terminal tagging and diminished by N-terminal tagging (Zhang et al.

2017). It is assumed C-terminal tagging stabilized p28 thus increased its concentration, which was beneficial for repression function. Whether C-terminal tagging would increase the stability of pA is unknown. However, predelivery of pA one day before replicon infiltration could potentially prepare pA cached in the cells. Failing to exert repression

108 effect in co-infiltration experiment indicate FHV may have a mechanism to regulate the expression of pA even from a non-replicating mRNA source.

Trans expressing p28 and p88 exerted different repression pattern on replication of

TCV replicon. In case of p28, it almost indiscriminately abolished replication of all TCV

RNA species; in contrast, repression caused by trans expressing p88 specific targeted on the replication of longer TCV RNAs, leaving TCV sgRNA2 and ttsgR fairly unchanged or even increased (Zhang et al. 2019; Zhang et al. 2017). In this study, we showed that trans expression of pA has a similar effect to FHV RNA1 replication as it of p88 to TCV replication. It mainly targets RNA1, but preferentially accumulates RNA3.

The function of B1 is still a mystery. Though it has been shown to be nonessential for replication (Ball 1995), this small protein containing the entire C-terminal 102 aa residuals of pA, causes substantial repression on RNA1 replication when translated from a non-replicating mRNA. More importantly, the indiscriminate repression pattern of B1 is different from that of pA.

In summary, this study confirmed pA was capable of effectively trans-complement the replication of pA defective replicons including FR1(fs)_Fl and FR1(mGDD)_Fl in a pattern similar to that of p28, the ARP of TCV. Repression function of pA was only observed in C-terminally tag pA and predelivery of tag free pA, in a pattern similar to that of p88, the RdRp of TCV. Our results added another example that viral encoded replication protein(s) have dual opposite function in replication.

Material and methods

109

Constructs. FR1_Def was made by inserting a PCR fragment amplified from FHV

RNA1 infectious clone to the XhoI and XbaI linearized binary vector pAIDE. The insertion was accomplished by using Gibson Assembly ligation (New England Biolabs,

Ipswich, MA). FR1_Def contains a deletion of the last 205 nt of FHV RNA1 and an insertion of the last 12 nt of hepatitis D virus ribozyme. FR1_Fl was made by digesting

FR1_Def with ApaI and XbaI and replacing the shorter digested product with a fragment consisted of the deleted 205 nt in FR1_Def and a ribozyme sequence derived from tobacco ringspot virus. FR1_Acc65I was made by digesting FR1_Fl with ApaI and XbaI and replacing with two overlapping fragments that contain C to G mutation at nt right downstream of pA stop codon to “GTA” insertion right after mutation. The mutation and deletion introduced a unique Acc65I site in the context. The introduction of Acc65I changed one aa and inserted one aa in the B2 coding region, had no effect on pA and B1 coding region. FR1_sgG and FR1_sgR were made by inserting EGFP or mCherry coding sequence at the Acc65I site of FR1_Acc65I. Two pA frameshift replicons

FR1(fs)_Fl and FR1(fs)_sgR based on FR1_Fl and FR1_sgR were made by inserting

TAGG at FHV RNA nt 379 through a synthesized DNA fragment (gBlock, Integrated

DNA Technology, Corallville, IA). The insertion introduced a stop codon, a new AvrII site and disrupted the downstream reading frame. Similarly, FR1(mGDD)_Fl was made by digesting the FR1_Fl construct with SmaI (two sites on FHV RNA1) and replacing the

446-nt region encompassing the GDD coding sequence (GGTGATGAT) with a gBlock in which the GDD-coding sequence was changed to AGTACTGAT. Such modification also introduced a ScaI site. FR1(5’rp)_Acc65I and FR1(5’rp)_Fl were made by digesting FR1_Acc65I and FR1_Fl with XhoI and BamHI and replicating the cut-off with

110 a gBlock. The gBlock eliminated the EagI site at nt 375 and BamHI site at nt 384 without changing the aa sequence of pA, replaced the 40-nt 5’ UTR with the last 40 nt of CarMV

5’ UTR, and modified the GCTTTC of CarMV 5’UTR to introduced a new BamHI site.

The pA transient expression construct was made by digesting FR1(5’rp)_Ac655I with

Acc65I and XbaI to removing the 3’ UTR and the TRSV ribozyme, and re-ligating through two annealing primers using Gibson Assembly. GDD mutated pA transient expression construct pA(mGDD) was made by modifying pA in a way like that of making FR1(mGDD)_Fl. The pA N-terminally tagged transient expression construct

G11-Flag-pA was modified from pA through the pre-created BamHI site to drop in a

202-nt gBlock. Similarly, the pA C terminally tagged transient expression construct pA-

G11-Flag was made by digesting pA with SacI and Acc65I and inserting a 220-nt gBlock. B1 and B2 were made by replacing the pA coding sequence using BamHI and

XbaI digestion (CarMV 5’UTR was remained as in pA) with B1 or B2 coding sequence obtained from PCR. The identities of all new constructs were verified with Sanger sequencing. A brief summary of the constructs and how they were made are listed in

Table 4.1.

Agro-infiltration. Upon verification, all of the constructs were introduced into

Agrobacterium tumefaciens strain C58C1 with electroporation (Qu et al. 2003). To carry out the experiments described in the Results section, various combinations of

Agrobacterium suspensions were mixed together to a final OD600 = 0.5 of each

Agrobacterium suspension and delivered into N. benthamiana leaves as described (Qu et al. 2003; Zhang et al. 2015; Zhang et al. 2017). A p19-expressing Agrobacterium strain

111 was included in most combinations to alleviate RNA silencing-mediated mRNA degradation.

RNA extraction and Northern blotting. Total RNA was extracted from agro- infiltrated N. benthamiana leaves using the Direct-zol RNA Miniprep kit (Zymo

The RNA was then quantified with NanoDrop and subjected to Northern blotting as described (Zhang et al. 2015; Zhang et al. 2017).

Acknowledgement

We thank Dr. Shouwei Ding for kindly providing FHV infectious clones. We thank the labs of Drs. Lucy Stewart, Peg Redinbaugh, Sally Miller for generous equipment sharing. This study was supported by a SEEDS grant from the Ohio Agricultural

Research and Development Center, Graduate Assistantships from OSU and OARDC to

R.S., as well as tuition assistances to S.Z. and R.S. from the Department of Plant

Pathology, OSU.

112

Figures

Figure 4.1 Replication of FHV RNA1 in N. benthamiana. (A) Schematic representation of the FHV genome with encoded proteins shown as rectangles. The subgenomic RNA, RNA3 serving as mRNA for B1 and B2 were also depicted. Relative position of GDD motif in pA was highlighted in blue. (B) FHV RNA1 replicon constructs used in this study. FR1_Fl transcribes full-length FHV RNA1. FR1_Def transcribes RNA1 with last 205 nt deleted.

FR1_Acc65I has three nts inserted after pA stop codon in frame of B2, creating an Acc65I

RE site. FR1_EGFP and mCherry express EGFP and mCherry as B2 fusion proteins, respectively. The frameshift (fs) mutant expresses only the first 113 aa residuals of pA before encounter an inserted stop codon and an additional nt causing frameshift. The 5’ UTR replacement mutant (5’rp) has the 5’ UTR replaced with the last 40 nts of carnation mottle

113 virus 5’UTR. (C) Protein A, B1, and B2 transient expression constructs used in this study.

G11-Flag-pA, pA N-terminal fused with the 11th beta sheet of GFP and Flag. Similarly, pA-

G11-Flag is C-terminal fused. In these transient expression constructs, the 3’UTR of RNA1 has been removed, while the 5’ UTR of FHV has been change to last 40 nts of CarMV

5’UTR. All constructs including replicons and transient expression were delivered into N. benthamiana cells using agro-infiltration. All constructs use the 2X35S promoter and T35S terminator to drive the transcription. (D) Demonstration of self-replication of these constructs in N. benthamiana revealed by northern blotting. RNA3r, the recombined RNA3 results from fluorescent protein gene insertion.

114

Figure 4.2 The impacts of trans-expression of pA and its variants on replication of various RNA1 replicons revealed by northern blotting. (A, B) Trans-complementation mediated by pA. (C) N-terminal tagging diminished pA mediated trans-complementation.

(D) C-terminal tagging enhanced pA’s function in replication repression.

115

Figure 4.3 The impacts of trans-expression of B1 and B2 on replication of various

RNA1 replicons revealed by northern blotting. (A) B1 trans-represses the replication of

RNA1 replicons. (B) B2 trans-expression slightly increase the accumulation of RNA1 and

RNA1 derived RNA3.

116

Figure 4.4 Sequential infiltration unmasked the replication repression function of tag free pA. (A) The effects of pre-infiltration of pA, B1, and B2 on FR1_Acc65I replication.

(B) Combination of pA and B1 has most potent effect on replication repression.

Figure 4.5 Sequential infiltration reveals pA preferentially replicated more authentic

RNA1 by northern blotting.

117

Table 4.1: Constructs used in this study

Construct gBlock sequence or PCR primers (5’ → 3’) Note

FR1_Def F: TTCATTTGGAGAGGACCTCGAGGTTTTCGAAACAAATAAAACAGAAAAGC pAI digested

R: ACTGGTGATTTTTGCGGACTCTAGAAGCTCTCCCTTAGCCATCCGA by XhoI, XbaI

(Underlined part of reverse primer annealed to FR1 # 3896-3903)

FR1_Fl F: GGATTATTGCAAAATGGTACTACGA (FR1 #2397) FR1_Def

R: ACTGGTGATTTTTGCGGACTCTAGAAGCTCTCCCTTAGCCATCCGAGTG digested by

Apa I (FR1 nt

2432), XbaI

FR1_Acc65I F1: GGATTATTGCAAAATGGTACTACGA (FR1 #2397) FR1_Fl

R1: TGGGGGTACcTCACTTCCGGTTGTTGGAAG digested by

F2: AACCGGAAGTGAgGTACCCCCACCCGCAAA Apa I (FR1 nt

R2: ACTGGTGATTTTTGCGGACTCTAGAAGCTCTCCCTTAGCCATCCGAGTG 2432), XbaI

118

FR1_sgG F: CTTCCAACAACCGGAAGTGAGATGGTGAGCAAGGGCGAGGA FR1_Acc65I

R: CAGTTTTGCGGGTGGGGGTTACTTGTACAGCTCGTCCAT digested by

Acc65I

FR1_sgR Same primers as that for FR1_sgG, using different template Same as above

FR1(fs)_Fl CATTTCATTTGGAGAGGACCTCGAGGTTTTCGAAACAAATAAAACAGAAAAG FR1_Fl

CGAACCTAAACAATGACTCTAAAAGTTATTCTTGGAGAACACCAGATCACCCG digested by

AACTGAATTGTTAGTCGGGATTGCAACCGTATCTGGGTGCGGTGCCGTAGTGTA XhoI, BamHI

CTGCATATCCAAGTTCTGGGGCTATGGGGCAATTGCGCCCTATCCTCAGAGTGG (FR1 nt 2432)

FR1(fs)_sgR AGGGAACCGAGTTACACGCGCATTGCAACGGGCTGTCATTGACAAAACGAAGA FR1_Fl

CCCCGATAGAGACACGTTTCTATCCGCTTGACAGCCTGCGTACCGTGACGCCTA digested by

AGCGTGTCGCAGACAACGGGCACGCCGTTTCAGGGGCCGTACGTGATGCCGCA XhoI, BamHI

CGTCGTTTGATCGACGAGTCCATCACGGCCTAGGGTTGGAGGATCCAAATTTGA (FR1 nt 2432)

GGTCAACCCCAA (TAGG insertion at 379-380, introduces a frameshift, a stop codon,

and a new AvrII site)

119

FR1(mGDD)_Fl ATGAACATTGGTATTACCCCGGGCGGAACCCGACTGAGATCGCCGACGGTGTTT FR1_Fl

GTGAGTTTGTTAGTGACTGTGATGCTGAAGTCATAGAAACTGACTTCTCCAACC digested by

TCGATGGCAGGGTTTCCAGCTGGATGCAAAGAAACATCGCCCAAAAGGCCATG SmaI

GTTCAAGCATTCCGCCCAGAATACAGAGATGAGATCATTTCATTCATGGACACG

ATAATCAATTGTCCAGCTAAAGCTAAACGCTTTGGTTTCCGATATGAGCCTGGT

GTAGGCGTTAAAAGTGGAAGTCCAACAACCACGCCACATAACACCCAATACAA

TGGATGTGTCGAATTTACAGCTCTGACCTTTGAGCATCCTGATGCTGAACCTGA

AGATTTGTTCCGTTTAATCGGACCGAAGTGCAGTACTGATGGTCTTTCCCGGGC

TATCATTCAAAAATCA

FR1(5’rp)_Acc65I CATTTCATTTGGAGAGGACCTCGAGTCTTACATACATTATATATTCTAACCAT FR1_Ac655I

CTGGATCCACTATGACTCTAAAAGTTATTCTTGGAGAACACCAGATCACCCGA digested by

ACTGAATTGTTAGTCGGGATTGCAACCGTATCTGGGTGCGGTGCCGTAGTGTAC XhoI, BamHI

FR1(5’rp)_Fl TGCATATCCAAGTTCTGGGGCTATGGGGCAATTGCGCCCTATCCTCAGAGTGGA FR1_Fl

GGGAACCGAGTTACACGCGCATTGCAACGGGCTGTCATTGACAAAACGAAGAC digested by

CCCGATAGAGACACGTTTCTATCCGCTTGACAGCCTGCGTACCGTGACGCCTAA XhoI, BamHI

120

GCGTGTCGCAGACAACGGGCACGCCGTTTCAGGGGCCGTACGTGATGCCGCAC

GTCGTTTGATCGACGAGTCCATCACGGCAGTTGGAGGTTCCAAATTTGAGGTCA

ACCCCAA pA CTTCCAACAACCGGAAGTGAGGTACCACTGTCTCTAGAGTCCGCAAAAATCA FR1_Ac655I

CCAGT digested by

Acc65I, XbaI pA(mGDD) Same as the one of FR1(mGDD)_Fl pA digested by

SmaI

G11-Flag-pA CATTATATATTCTAACCATCTGGATCCACAACAATGGGTAGAGACCACATGGT pA digested by

GCTGCACGAGTACGTGAACGCAGCAGGAATCACAGGAGATGGTTCTGGATCAG BamHI

GTGATTACAAGGATGACGACGATAAGGACTATAAGGACGATGATGACAAAGGC

TCTGGAAGTGCATCAGGATCCACTATGACTCTAAAAGTTATTC pA-G11-Flag AGAACGCCTCCGAAAGCTGGAGCTCAGCCACAGCCTTCCAACAACCGGAAGGG pA digested by

TGACGGAAGTGGTTCAGGAAGAGACCACATGGTGCTGCACGAGTACGTGAACG Acc65I, SacI

CAGCAGGAATCACAGGAGATGGTTCTGGATCAGGTGATTACAAGGATGACGAC

121

GATAAGGACTATAAGGACGATGATGACAAATAATGAGGTACCACTGTCTCTA

GAGTCCGCA

B1 F: TATATATTCTAACCATCTGGATCCACTATGTTAAACGATGCCAAGCAAAC pA digested by

R: GTGATTTTTGCGGACTCTAGAGACTTAGGTACCCTTCCGGTTGTTGGAAG BamHI, XbaI

B2 F: TATATATTCTAACCATCTGGATCCACTATGCCAAGCAAACTCGCGCTAAT pA digested by

R: GGTGATTTTTGCGGACTCTAGAGACTTAGGTACCCAGTTTTGCGGGTGGG BamHI, XbaI

122

Bibliography

Ahola, T. 2019. New phylogenetic grouping of positive-sense RNA viruses is concordant with replication complex morphology. Mbio 10:e01402-01419.

Akiyama, B. M., Laurence, H. M., Massey, A. R., Costantino, D. A., Xie, X., Yang, Y., Shi, P.-Y., Nix, J. C., Beckham, J. D., and Kieft, J. S. 2016. Zika virus produces noncoding RNAs using a multi-pseudoknot structure that confounds a cellular exonuclease. Science 354:1148-1152.

Atkins, J. F., Loughran, G., Bhatt, P. R., Firth, A. E., and Baranov, P. V. 2016. Ribosomal frameshifting and transcriptional slippage: From genetic steganography and cryptography to adventitious use. Nucleic Acids Res 44:7007- 7078.

Ball, L. A. 1995. Requirements for the self-directed replication of flock house virus RNA 1. Journal of virology 69:2722-2722.

Ball, L. A., Amann, J. M., and Garrett, B. K. 1992. Replication of nodamura virus after transfection of viral RNA into mammalian cells in culture. Journal of virology 66:2326-2334.

Beier, H., Barciszewska, M., Krupp, G., Mitnacht, R., and Gross, H. J. 1984. UAG readthrough during TMV RNA translation: isolation and sequence of two tRNAsTyr with suppressor activity from tobacco plants. The EMBO journal 3:351-356.

Cao, M., Ye, X., Willie, K., Lin, J., Zhang, X., Redinbaugh, M. G., Simon, A. E., Morris, T. J., and Qu, F. 2010. The capsid protein of turnip crinkle virus overcomes two separate defense barriers to facilitate systemic movement of the virus in Arabidopsis. Journal of virology 84:7793-7802.

Carpenter, C. D., and Simon, A. E. 1996. In vivo repair of 3′-end deletions in a TCV satellite RNA may involve two abortive synthesis and priming events. Virology 226:153-160.

Carrington, J. C., and Freed, D. D. 1990. Cap-independent enhancement of translation by a plant potyvirus 5'nontranslated region. Journal of virology 64:1590-1597.

Chang, R.-Y., Hsu, T.-W., Chen, Y.-L., Liu, S.-F., Tsai, Y.-J., Lin, Y.-T., Chen, Y.-S., and Fan, Y.-H. 2013. Japanese encephalitis virus non-coding RNA inhibits activation of interferon by blocking nuclear translocation of interferon regulatory factor 3. Veterinary microbiology 166:11-21.

Chapman, E. G., Moon, S. L., Wilusz, J., and Kieft, J. S. 2014. RNA structures that resist degradation by Xrn1 produce a pathogenic Dengue virus RNA. eLife 3:e01892.

123

Che, X., Piestun, D., Mawassi, M., Yang, G., Satyanarayana, T., Gowda, S., Dawson, W. O., and Bar-Joseph, M. 2001. 5′-coterminal subgenomic RNAs in citrus tristeza virus-infected cells. Virology 283:374-381.

Chen, Y.-S., Fan, Y.-H., Tien, C.-F., Yueh, A., and Chang, R.-Y. 2018. The conserved stem-loop II structure at the 3' untranslated region of Japanese encephalitis virus genome is required for the formation of subgenomic flaviviral RNA. PloS one 13:e0201250.

Choi, I.-R., and White, K. A. 2002. An RNA activator of subgenomic mRNA1 transcription in tomato bushy stunt virus. Journal of Biological Chemistry 277:3760-3766.

Choi, I.-R., Ostrovsky, M., Zhang, G., and White, K. A. 2001. Regulatory activity of distal and core RNA elements in tombusvirus subgenomic mRNA2 transcription. Journal of Biological Chemistry 276:41761-41768.

Cimino, P. A., Nicholson, B. L., Wu, B., Xu, W., and White, K. A. 2011. Multifaceted regulation of translational readthrough by RNA replication elements in a Tombusvirus. PLOS Pathogens 7:e1002423.

Dearing, S., Scotti, P., Wigley, P., and Dhana, S. 1980. A small RNA virus isolated from the grass grub, Costelytra zealandica (Coleoptera: Scarabaeidae). New Zealand journal of zoology 7:267-269.

Dilweg, I. W., Gultyaev, A. P., and Olsthoorn, R. C. 2019. Structural features of an Xrn1- resistant plant virus RNA. RNA biology 16:838-845.

Ertel, K. J., Benefield, D., Castano-Diez, D., Pennington, J. G., Horswill, M., den Boon, J. A., Otegui, M. S., and Ahlquist, P. 2017. Cryo-electron tomography reveals novel features of a viral RNA replication compartment. eLife 6:e25940.

Finch, L. K., Ling, R., Napthine, S., Olspert, A., Michiels, T., Lardinois, C., Bell, S., Loughran, G., Brierley, I., and Firth, A. E. 2015. Characterization of ribosomal frameshifting in Theiler's murine encephalomyelitis virus. Journal of virology 89:8580-8589.

Firth, A. E., and Brierley, I. 2012. Non-canonical translation in RNA viruses. Journal of General Virology 93:1385-1409.

Firth, A. E., Blitvich, B. J., Wills, N. M., Miller, C. L., and Atkins, J. F. 2010. Evidence for ribosomal frameshifting and a novel overlapping gene in the genomes of insect-specific flaviviruses. Virology 399:153-166.

Gallie, D. R. 2002. The 5'-leader of tobacco mosaic virus promotes translation through enhanced recruitment of eIF4F. Nucleic Acids Res 30:3401-3411.

124

Gallie, D. R., and Kado, C. I. 1989. A translational enhancer derived from tobacco mosaic virus is functionally equivalent to a Shine-Dalgarno sequence. Proceedings of the National Academy of Sciences 86:129-132.

Gao, F., and Simon, A. E. 2015. Multiple Cis-acting elements modulate programmed -1 ribosomal frameshifting in Pea enation mosaic virus. Nucleic Acids Res 44:878- 895.

Gergerich, R. C., and Dolja, V. V. 2006. Introduction to plant viruses, the invisible foe. The plant health instructor 4.

Gralla, J., Steitz, J. A., and Crothers, D. M. 1974. Direct physical evidence for secondary structure in an isolated fragment of R17 bacteriophage mRNA. Nature 248:204- 208.

Guarino, L. A., Ghosh, A., Dasmahapatra, B., Dasgupta, R., and Kaesberg, P. 1984. Sequence of the black beetle virus subgenomic RNA and its location in the viral genome. Virology 139:199-203.

Guenther, R. H., Sit, T. L., Gracz, H. S., Dolan, M. A., Townsend, H. L., Liu, G., Newman, W. H., Agris, P. F., and Lommel, S. A. 2004. Structural characterization of an intermolecular RNA–RNA interaction involved in the transcription regulation element of a bipartite plant virus. Nucleic Acids Res 32:2819-2828.

Gunawardene, C. D., Newburn, L. R., and White, K. A. 2019. A 212-nt long RNA structure in the Tobacco necrosis virus-D RNA genome is resistant to Xrn degradation. Nucleic Acids Res 47:9329-9342.

Guo, Q., Zhang, S., Sun, R., Yao, X., Zhang, X. F., Tatineni, S., Meulia, T., and Qu, F. 2020. Superinfection exclusion by p28 of turnip crinkle virus is separable from its replication function. Molecular plant-microbe interactions : MPMI 33:364-375.

Hargreaves, J., Davey, C., Hargreaves, J., Davey, C., Auerbach, J., Blanchard, J., Bond, V., Bonell, C., Burgess, R., Busza, J., Colbourn, T., Cowan, F., Doyle, A., Hakim, J., Hensen, B., Hosseinipour, M., Lin, L., Johnson, S., Masuka, N., Mavhu, W., Mugurungi, O., Mukungunugwa, S., Mushavi, A., Phillips, A., Platt, L., Prost, A., Ruzagira, E., Seeley, J., Taramusi, I., and Yekeye, R. 2020. Three lessons for the COVID-19 response from pandemic HIV. The Lancet HIV.

Herold, J., and Andino, R. 2001. Poliovirus RNA replication requires genome circularization through a protein-protein bridge. Molecular cell 7:581-591.

Hibino, H. 1996. Biology and epidemiology of rice viruses. Annual review of phytopathology 34:249-274.

Hillman, B. I., and Cai, G. 2013. The family narnaviridae: simplest of RNA viruses. Advances in virus research 86:149-176.

125

Hillman, B. I., Annisa, A., and Suzuki, N. 2018. Chapter Five - Viruses of Plant- Interacting Fungi. Pages 99-116 in: Advances in virus research, vol. 100. M. Kielian, T. C. Mettenleiter and M. J. Roossinck, eds. Academic Press.

Howitt, R. L. J., Beever, R. E., Pearson, M. N., and Forster, R. L. S. 2001. Genome characterization of Botrytis virus F, a flexuous rod-shaped mycovirus resembling plant ‘potex-like’ virusesThe GenBank accession number of the sequence reported in this paper is AF238884. Journal of General Virology 82:67-78.

Hu, Z., Wu, S., Cheng, J., Fu, Y., Jiang, D., and Xie, J. 2014. Molecular characterization of two positive-strand RNA viruses co-infecting a hypovirulent strain of Sclerotinia sclerotiorum. Virology 464-465:450-459.

Hutchinson, E. C. 2018. Influenza virus. Trends in Microbiology 26:809-810.

Iwakawa, H.-o., Mizumoto, H., Nagano, H., Imoto, Y., Takigawa, K., Sarawaneeyaruk, S., Kaido, M., Mise, K., and Okuno, T. 2008. A viral noncoding RNA generated by cis-element-mediated protection against 5′→3′ RNA decay represses both cap- independent and cap-dependent translation. Journal of virology 82:10162-10174.

Janda, M., and Ahlquist, P. 1993. RNA-dependent replication, transcription, and persistence of brome mosaic virus RNA replicons in S. cerevisiae. Cell 72:961- 970.

Jiwaji, M., Matcher, G. F., de Bruyn, M.-M., Awando, J. A., Moodley, H., Waterworth, D., Jarvie, R. A., and Dorrington, R. A. 2019. Providence virus: An animal virus that replicates in plants or a plant virus that infects and replicates in animal cells? PloS one 14:e0217494.

Kang, S.-H., Sun, Y.-D., Atallah, O. O., Huguet-Tapia, J. C., Noble, J. D., and Folimonova, S. Y. 2019. A Long non-coding RNA of Citrus tristeza virus: Role in the virus interplay with the host immunity. Viruses 11:436.

Kelly, L., Gerlach, W. L., and Waterhouse, P. M. 1994. Characterisation of the subgenomic RNAs of an australian isolate of Barley yellow dwarf luteovirus. Virology 202:565-573.

Kelman, A. 1995. Contributions of plant pathology to the biological sciences and industry. Annual review of phytopathology 33:1-23.

Kim, K., and Hemenway, C. 1999. Long-distance RNA-RNA interactions and conserved sequence elements affect potato virus X plus-strand RNA accumulation. RNA (New York, N.Y.) 5:636.

Koev, G., and Miller, W. A. 2000. A Positive-Strand RNA Virus with Three Very Different Subgenomic RNA Promoters. Journal of virology 74:5988-5996.

126

Kroner, P. A., Young, B. M., and Ahlquist, P. 1990. Analysis of the role of brome mosaic virus 1a protein domains in RNA replication, using linker insertion mutagenesis. Journal of virology 64:6110-6120.

Kuhlmann, M. M., Chattopadhyay, M., Stupina, V. A., Gao, F., and Simon, A. E. 2016. An RNA element that facilitates programmed ribosomal readthrough in turnip crinkle virus adopts multiple conformations. Journal of virology 90:8575-8591.

Li, H., Li, W. X., and Ding, S. W. 2002. Induction and suppression of RNA silencing by an animal virus. Science 296:1319-1321.

Li, Y., Treffers, E. E., Napthine, S., Tas, A., Zhu, L., Sun, Z., Bell, S., Mark, B. L., van Veelen, P. A., van Hemert, M. J., Firth, A. E., Brierley, I., Snijder, E. J., and Fang, Y. 2014. Transactivation of programmed ribosomal frameshifting by a viral protein. P Natl Acad Sci USA 111:E2172-E2181.

Li, Z., and Nagy, P. D. 2011. Diverse roles of host RNA binding proteins in RNA virus replication. RNA biology 8:305-315.

Lin, H. X., and White, K. A. 2004. A complex network of RNA–RNA interactions controls subgenomic mRNA transcription in a tombusvirus. The EMBO journal 23:3365-3374.

Lin, K.-C., Chang, H.-L., and Chang, R.-Y. 2004. Accumulation of a 3′-terminal genome fragment in Japanese encephalitis virus-infected mammalian and mosquito cells. Journal of virology 78:5133-5138.

Lindenbach, B. D., Sgro, J.-Y., and Ahlquist, P. 2002. Long-distance base pairing in flock house virus RNA1 regulates subgenomic RNA3 synthesis and RNA2 replication. Journal of virology 76:3905-3919.

Martinez, F., Sardanyes, J., Elena, S. F., and Daros, J. A. 2011. Dynamics of a plant RNA virus intracellular accumulation: stamping machine vs. geometric replication. Genetics 188:637-646.

Mazeaud, C., Freppel, W., and Chatel-Chaix, L. 2018. The Multiples Fates of the Flavivirus RNA Genome During Pathogenesis. Front Genet 9.

Miller, D. J., and Ahlquist, P. 2002. Flock house virus RNA polymerase is a transmembrane protein with amino-terminal sequences sufficient for mitochondrial localization and membrane insertion. Journal of virology 76:9856- 9867.

Miras, M., Miller, W. A., Truniger, V., and Aranda, M. A. 2017. Non-canonical translation in plant RNA viruses. Frontiers in plant science 8

Moon, S. L., Dodd, B. J., Brackney, D. E., Wilusz, C. J., Ebel, G. D., and Wilusz, J. 2015. Flavivirus sfRNA suppresses antiviral RNA interference in cultured cells

127

and mosquitoes and directly interacts with the RNAi machinery. Virology 485:322-329.

Moon, S. L., Anderson, J. R., Kumagai, Y., Wilusz, C. J., Akira, S., Khromykh, A. A., and Wilusz, J. 2012. A noncoding RNA produced by arthropod-borne flaviviruses inhibits the cellular exoribonuclease XRN1 and alters host mRNA stability. RNA (New York, N.Y.) 18:2029-2040.

Moureau, G., Cook, S., Lemey, P., Nougairede, A., Forrester, N. L., Khasnatinov, M., Charrel, R. N., Firth, A. E., Gould, E. A., and de Lamballerie, X. 2015. New insights into flavivirus evolution, taxonomy and biogeographic history, extended by analysis of canonical and alternative coding sequences. PloS one 10:e0117849- e0117849.

Nagy, P. D., Carpenter, C. D., and Simon, A. E. 1997. A novel 3′-end repair mechanism in an RNA virus. Proceedings of the National Academy of Sciences 94:1113- 1118.

Napthine, S., Bell, S., Hill, C. H., Brierley, I., and Firth, A. E. 2019. Characterization of the stimulators of protein-directed ribosomal frameshifting in Theiler's murine encephalomyelitis virus. Nucleic Acids Res 47:8207-8223.

Newburn, L. R., and White, K. A. 2020. A trans-activator-like structure in RCNMV RNA1 evokes the origin of the trans-activator in RNA2. PLOS Pathogens 16:e1008271.

Oldenhof, M., De Groot, P., Visser, J., Schrauwen, J., and Wullems, G. 1996. Isolation and characterization of a microspore-specific gene from tobacco. Plant molecular biology 31:213-225.

Panavas, T., and Nagy, P. D. 2003. Yeast as a model host to study replication and recombination of defective interfering RNA of Tomato bushy stunt virus. Virology 314:315-325.

Pelham, H. R. 1978. Leaky UAG termination codon in tobacco mosaic virus RNA. Nature 272:469-471.

Peltier, C., Klein, E., Hleibieh, K., D’Alonzo, M., Hammann, P., Bouzoubaa, S., Ratti, C., and Gilmer, D. 2012. Beet necrotic yellow vein virus subgenomic RNA3 is a cleavage product leading to stable non-coding RNA required for long-distance movement. Journal of General Virology 93:1093-1102.

Pijlman, G. P., Funk, A., Kondratieva, N., Leung, J., Torres, S., van der Aa, L., Liu, W. J., Palmenberg, A. C., Shi, P.-Y., Hall, R. A., and Khromykh, A. A. 2008. A highly structured, nuclease-resistant, noncoding RNA produced by flaviviruses is required for pathogenicity. Cell Host & Microbe 4:579-591.

128

Plaskon, N. E., Adelman, Z. N., and Myles, K. M. 2009. Accurate strand-specific quantification of viral RNA. PloS one 4.

Pompon, J., Manuel, M., Ng, G. K., Wong, B., Shan, C., Manokaran, G., Soto-Acosta, R., Bradrick, S. S., Ooi, E. E., and Missé, D. 2017. Dengue subgenomic flaviviral RNA disrupts immunity in mosquito salivary glands to increase virus transmission. PLoS pathogens 13:e1006535.

Price, B. D., Rueckert, R. R., and Ahlouist, P. 1996. Complete replication of an animal virus and maintenance of expression vectors derived from it in Saccharomyces cerevisiae. Proceedings of the National Academy of Sciences 93:9465-9470.

Price, B. D., Roeder, M., and Ahlquist, P. 2000. DNA-Directed expression of functional flock house virus RNA1 derivatives in Saccharomyces cerevisiae, heterologous gene expression, and selective effects on subgenomic mRNA synthesis. Journal of virology 74:11724-11733.

Qu, F., and Morris, T. J. 1997. Encapsidation of turnip crinkle virus is defined by a specific packaging signal and RNA size. Journal of virology 71:1428-1435.

Qu, F., and Morris, T. J. 2000. Cap-independent translational enhancement of turnip crinkle virus genomic and subgenomic RNAs. Journal of virology 74:1085-1093.

Qu, F., Ren, T., and Morris, T. J. 2003. The coat protein of turnip crinkle virus suppresses posttranscriptional gene silencing at an early initiation step. Journal of virology 77:511-522.

Quinkert, D., Bartenschlager, R., and Lohmann, V. 2005. Quantitative analysis of the hepatitis c virus replication complex. Journal of virology 79:13594-13605.

Quinn, J. J., and Chang, H. Y. 2016. Unique features of long non-coding RNA biogenesis and function. Nature reviews. Genetics 17:47-62.

Rajendran, K. S., Pogany, J., and Nagy, P. D. 2002. Comparison of turnip crinkle virus RNA-dependent RNA polymerase preparations expressed in Escherichia coli or derived from infected plants. Journal of virology 76:1707-1717.

Redinbaugh, M. G., and Stewart, L. R. 2018. Maize lethal necrosis: An emerging, synergistic viral disease. Annual Review of Virology 5:301-322.

Restrepo-Hartwig, M. A., and Carrington, J. C. 1994. The tobacco etch potyvirus 6- kilodalton protein is membrane associated and involved in viral replication. Journal of virology 68:2388-2397.

Salonen, A., Ahola, T., and Kääriäinen, L. 2005. Viral RNA replication in association with cellular membranes. Pages 139-173 in: Membrane trafficking in viral replication. M. Marsh, ed. Springer Berlin Heidelberg, Berlin, Heidelberg.

129

Sanjuán, R., and Domingo-Calap, P. 2016. Mechanisms of viral mutation. Cell Mol Life Sci 73:4433-4448.

Schaad, M. C., Jensen, P. E., and Carrington, J. C. 1997. Formation of plant RNA virus replication complexes on membranes: role of an endoplasmic reticulum‐targeted viral protein. The EMBO journal 16:4049-4059.

Scheets, K. 2000. Maize chlorotic mottle machlomovirus expresses its coat protein from a 1.47-kb subgenomic RNA and makes a 0.34-kb subgenomic RNA. Virology 267:90-101.

Scherbik, S. V., Paranjape, J. M., Stockman, B. M., Silverman, R. H., and Brinton, M. A. 2006. RNase L plays a role in the antiviral response to West Nile virus. Journal of virology 80:2987-2999.

Selling, B. H., and Rueckert, R. R. 1984. Plaque assay for black beetle virus. Journal of virology 51:251-253.

Selling, B. H., Allison, R. F., and Kaesberg, P. 1990. Genomic RNA of an insect virus directs synthesis of infectious virions in plants. Proceedings of the National Academy of Sciences 87:434-438.

Shen, R., and Miller, W. A. 2004. The 3′ untranslated region of tobacco necrosis virus RNA contains a barley yellow dwarf virus-like cap-independent translation element. Journal of virology 78:4655-4664.

Shivaprasad, P. V., Akbergenov, R., Trinks, D., Rajeswaran, R., Veluthambi, K., Hohn, T., and Pooggin, M. M. 2005. Promoters, transcripts, and regulatory proteins of mungbean yellow mosaic geminivirus. Journal of virology 79:8149-8163.

Silva, P. A., Pereira, C. F., Dalebout, T. J., Spaan, W. J., and Bredenbeek, P. J. 2010. An RNA pseudoknot is required for production of yellow fever virus subgenomic RNA by the host nuclease XRN1. Journal of virology 84:11395-11406.

Slonchak, A., and Khromykh, A. A. 2018. Subgenomic flaviviral RNAs: What do we know after the first decade of research. Antiviral Research 159:13-25.

Sun, R., Zhang, S., Zheng, L., and Qu, F. 2020. Translation-independent roles of RNA secondary structures within the replication protein coding region of turnip crinkle virus. Viruses 12.

Sztuba-Solińska, J., Stollar, V., and Bujarski, J. J. 2011. Subgenomic messenger RNAs: Mastering regulation of (+)-strand RNA virus life cycle. Virology 412:245-255.

Taliansky, M., Torrance, L., and Kalinina, N. O. 2008. Role of plant virus movement proteins. Pages 33-54 in: Plant virology protocols: from viral sequence to protein function. G. D. Foster, I. E. Johansen, Y. Hong and P. D. Nagy, eds. Humana Press, Totowa, NJ.

130

Tomlinson, K. R., Bailey, A. M., Alicai, T., Seal, S., and Foster, G. D. 2018. Cassava brown streak disease: historical timeline, current knowledge and future prospects. Molecular plant pathology 19:1282-1294.

Urosevic, N., Van Maanen, M., Mansfield, J. P., Mackenzie, J. S., and Shellam, G. R. 1997. Molecular characterization of virus-specific RNA produced in the brains of flavivirus-susceptible and-resistant mice after challenge with Murray Valley encephalitis virus. Journal of General Virology 78:23-29.

Van Wynsberghe, P. M., Chen, H. R., and Ahlquist, P. 2007. Nodavirus RNA replication protein a induces membrane association of genomic RNA. Journal of virology 81:4633-4644.

Walter, C. T., Pringle, F. M., Nakayinga, R., de Felipe, P., Ryan, M. D., Ball, L. A., and Dorrington, R. A. 2010. Genome organization and translation products of Providence virus: insight into a unique tetravirus. Journal of General Virology 91:2826-2835.

Wang, J., and Simon, A. E. 1997. Analysis of the two subgenomic RNA promoters for turnip crinkle virus in vivo and in vitro. Virology 232:174-186.

Wang, J., Carpenter, C. D., and Simon, A. E. 1999. Minimal sequence and structural requirements of a subgenomic RNA promoter for turnip crinkle virus. Virology 253:327-336.

Wei, T., Huang, T. S., McNeil, J., Laliberte, J. F., Hong, J., Nelson, R. S., and Wang, A. 2010a. Sequential recruitment of the endoplasmic reticulum and chloroplasts for plant potyvirus replication. Journal of virology 84:799-809.

Wei, T., Zhang, C., Hong, J., Xiong, R., Kasschau, K. D., Zhou, X., Carrington, J. C., and Wang, A. 2010b. Formation of complexes at plasmodesmata for potyvirus intercellular movement is mediated by the viral protein P3N-PIPO. PLoS Pathog 6:e1000962.

White, K. A., Skuzeski, J. M., Li, W., Wei, N., and Morris, T. J. 1995. Immunodetection, expression strategy and complementation of turnip crinkle virus p28 and p88 replication components. Virology 211:525-534.

White, K. A. 2002. The premature termination model: a possible third mechanism for subgenomic mRNA transcription in (+)-strand RNA viruses. Virology 304:147- 154.

White, K. A. 2015. The polymerase slips and PIPO exists. EMBO reports 16:885-886.

131

Wu, B., and White, K. A. 2007. Uncoupling RNA virus replication from transcription via the polymerase: functional and evolutionary insights. Embo Journal 26:5120- 5130.

Wu, B., Oliveri, S., Mandic, J., and White, K. A. 2010. Evidence for a premature termination mechanism of subgenomic mRNA transcription in a carmovirus. Journal of virology 84:7904-7907.

Yamagishi, N., Terauchi, H., Kanematsu, S., and Hidaka, S. 2003. Characterization of the small subgenomic RNA of Soybean dwarf virus. Archives of Virology 148:1827- 1834.

Yang, X., Huang, J., Liu, C., Chen, B., Zhang, T., and Zhou, G. 2017. Rice stripe mosaic virus, a novel cytorhabdovirus infecting rice via leafhopper transmission. Frontiers in microbiology 7:2140.

Yi, G., Gopinath, K., and Kao, C. C. 2007. Selective repression of translation by the brome mosaic virus 1a RNA replication protein. Journal of virology 81:1601- 1609.

Zavriev, S. K., Hickey, C. M., and Lommel, S. A. 1996. Mapping of the red clover necrotic mosaic virus subgenomic RNA. Virology 216:407-410.

Zhang, X.-F., Guo, J., Zhang, X., Meulia, T., Paul, P., Madden, L. V., Li, D., and Qu, F. 2015. Random plant viral variants attain temporal advantages during systemic infections and in turn resist other variants of the same virus. Scientific Reports 5.

Zhang, X. F., Sun, R., Guo, Q., Zhang, S., Meulia, T., Halfmann, R., Li, D., and Qu, F. 2017. A self-perpetuating repressive state of a viral replication protein blocks superinfection by the same virus. PLoS Pathog 13:e1006253.

Zhang, S., Sun, R., Guo, Q., Zhang, X. F., and Qu, F. 2019a. Repression of turnip crinkle virus replication by its replication protein p88. Virology 526:165-172.

Zhang, X.-F., Sun, R., Guo, Q., Zhang, S., Meulia, T., Halfmann, R., Li, D., and Qu, F. 2017. A self-perpetuating repressive state of a viral replication protein blocks superinfection by the same virus. Plos Pathogens 13.

Zhang, X., Zhang, X., Singh, J., Li, D., and Qu, F. 2012. Temperature-dependent survival of turnip crinkle virus-infected arabidopsis plants relies on an rna silencing-based defense that requires DCL2, AG02, and HEN1. Journal of virology 86:6847-6854.

132

Zhang, Y., Zhang, Y., Liu, Z.-Y., Cheng, M.-L., Ma, J., Wang, Y., Qin, C.-F., and Fang, X. 2019b. Long non-coding subgenomic flavivirus RNAs have extended 3D structures and are flexible in solution. EMBO reports 20:e47016.

133