A Dissertation Entitled Cauliflower Mosaic Virus P6 Protein Interactions

Home , Bromoviridae, Cauliflower mosaic virus, Caulimovirus, Mosaic virus

A Dissertation

Entitled

Cauliflower mosaic virus P6 protein interactions: a complex story

Lindy M. Lutz

Submitted to the Graduate Faculty as partial fulfillment of the requirements for Doctor of Philosophy Degree in Biology

______

Dr. Scott M. Leisner, Committee Chair

______

Dr. Patricia R. Komuniecki, Dean College of Graduate Studies

The University of Toledo May 2014

Copyright © 2014, Lindy Michelle Lutz This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Cauliflower mosaic virus P6 protein interactions: a complex story

Lindy M. Lutz

Submitted to the Graduate Faculty as partial fulfillment of the requirements for

Doctor of Philosophy Degree in Biology

The University of Toledo May 2014

Cauliflower mosaic virus (CaMV), one of the top ten viruses from a molecular plant pathology standpoint, is a plant pararetrovirus whose 8 kb circular double-stranded DNA genome encodes 7 different proteins (P1-P7). CaMV P6, encoded by gene VI has been implicated in a variety of functions such as: translational transactivation, host range control, symptom formation, host hypersensitive responses, RNA silencing suppressor activity, inclusion body (IB) formation and virus infectivity. Because of its multifunctional nature, P6 interacts with many host, and viral proteins including itself. P6 self-association appears to involve four domains (D1-D4). D3 has been implicated in viral infectivity and contains two RNA binding domains, separated by a highly conserved 34 amino acid spacer called D3b. CaMV mutants harboring a deletion of D3b are non- infectious, indicating its importance for viral propagation.

iii

To further analyze D3b, full-length P6 constructs were generated that harbored single amino acid substitutions within this region. In general, the mutants bound less efficiently to the individual P6 domains than wild type.

Mutations near the amino-terminal end of D3b had a more detrimental effect on self-association domain binding than those near the central portion. Since P6 is an IB protein, we hypothesized that mutations in D3b may influence IB formation. P6 IBs are thought to start out as small aggregations of protein (most likely P6) and ribosomes. They acquire additional materials (viral proteins and nucleic acids) to enlarge to form small bodies. Small bodies are then thought to fuse together to form larger, mature IBs. All mutant P6s formed IBs when expressed as green fluorescent protein (GFP) fusions in transgenic cells.

However, the mutant P6s that were most reduced in binding also showed decreased IB size. Hence, the mutations in D3b appear to affect the fusion of small IBs into larger ones. It is possible that IB size is important because it correlated with differences in virus host range. CaMV strain W260 has a much wider host range and more efficiently infects host plants when compared to the

CM1841 isolated. Our most recent data show that CM1841 IBs are smaller than those formed by W260 P6. In addition, P6 mutants that showed decreased binding to self-association domains and smaller IB sizes also exhibited much lower total viral DNA levels in inoculated leaves. This was also reflected by systemic symptom formation. Hence, less efficient binding correlates with smaller IB size and reduced local and systemic infection for the mutants. Taken together, these data suggest that fusion of small IBs into larger ones is important

iv for proper viral infections to occur and we have possibly identified mutants in this process. In addition, these data suggest that IB formation is required for viral infection rather than merely being a consequence of it

The CaMV genome encodes seven viral proteins including P6. P6 has been reported to interact with two other viral proteins in addition to itself.

Therefore, we also examined P6 for its ability to interact with the other viral gene products. P6 was found to interact with the aphid transmission factor (P2), the virion-associated protein (P3), reverse transcriptase protein (P5), and the protein of unknown function (P7). P2 was previously reported to control the difference in

IB stability between CM1841 and W260. Our data indicate that P2 from both viruses bound equally well to P6. The CM1841 P2 is less stable than its W260 counterpart. Taken together, this would suggest that the differences in IB stability for W260 and CM1841 mediated by P2 are due to variation in P2 protein stability rather than P6 binding. Binding of P6 to P3 could help the latter protein form complexes necessary for aphid transmission and virus cell-to-cell movement. P5 has a tri-partite structure with an N-terminal protease domain, a central reverse transcriptase (RT) and a C-terminal RNase H domain. Our pull-down results showed P6 could interact with full-length P5. Based on our preliminary pull- down analyses, P6 could bind inefficiently to the protease but more efficiently to the RT-RNase H (termed P5MC) portion of P5. Perhaps this interaction plays a role in P5 RT regulation. Interestingly, P5MC interactions with P5 showed a similar pattern to the P6 interactions. P5MC was able to self-associate well, but and interacted weakly with full-length P5 and the protease. P6 also interacted

v with P7, but the significance of this interaction is unknown. Perhaps P7 aids P6 in regulating an aspect of translational transactivation, but this is mere speculation.

In addition, P6 can also interact with a variety of host factors. In collaboration with Dr. James Schoelz at the University of Missouri, we found three Arabidopsis proteins: CHUP1, C2CDMT, and FIT that interact with full- length P6. Interestingly, of the four domains involved in P6 self-association, only

D2 and D4 bind to CHUP1 and C2CDMT. However, FIT was able to bind to all

P6 self-association domains but best to D2. Given that it binds to other host factors, we might speculate that D2 of P6 maybe acts as a host interface domain.

In summary P6 interacts with a large number of both viral and host proteins. P6 self-association is needed for proper IB formation and efficient infection. P6 interactions with each of the other viral proteins may be to modulate proper interactions of these proteins with their appropriate partners. Finally, P6 interactions with host factors may play a role in inhibiting host defenses, modulating systemic symptom formation, or mediating inter and intra cellular movement.

My family and friends, I am forever grateful to each and every one of you.

Words cannot adequately capture the degree of gratefulness I have for you. Your enduring love, support and understanding have enabled me to continue my work and reach this point. To my great friends, Reina Blake and Allison Haberman: thank you. You have always been there when I needed you most. To my family: my parents, John and Debi; my uncles, Ralph and Luther; my aunts, Linda and

Martha, and my cousins: Scott, Sue, Christy and Cheryl for all their patience, support and encouragement over the years.

vii

Acknowledgments

I would like to express my sincere appreciation to my advisor, Dr. Scott Leisner for giving me the opportunity to complete my Ph.D in his lab. His guidance, coupled with just the right amount of freedom, has been the source of my success in this endeavor. I offer my sincerest gratitude to my committee members: Dr.

John Gray, Dr. Lirim Shemshedini, Dr. Donald Ronning, and Dr. Jim Locke.

Their never-ending patience, valuable insights and suggestions have made this work enjoyable and prosperous. I have had the good fortune to work alongside many lab members and would like to acknowledge them all, in particular Dr.

Wendy Zellner and Dr. Gaurav Raikhy. The value of their mentorship and friendship cannot be overstated. There have been many colleagues and close friends that have supported me in this journey; without them I would have never gotten here. Brianne Sturt-Gillispie, Anish Purohit and Lynn Salazar. I can’t thank you enough for your constant friendship, support and countless acts of kindness over the years. My accomplishments are a direct reflection of every person that has supported me in one way or another over the past years and without whom I would not have succeeded.

viii

Table of Contents

Abstract………………………………………………………...……….……...... iii Acknowledgments…………………………………………...…………..……...viii

Table of Contents…………………………………………...……………..…...... xi

List of Tables………………………………………….…...……………….…....xii

List of Figures ………………………………………….…………...……..…....xiv

List of Abbreviations………………………………………………………...... xvii

1 Introduction………………………………………………………….….....1

2 Materials and Methods……...………………………..…..……………....27

3 Mutations Within A 35 Amino Acid Region Of P6 Influence Self- Association, Inclusion Body Formation, And Caulimovirus Infectivity………………………………………………………………...35

3.1 Abstract…………….…………………..…………...……...... 35 3.2 Introduction……………………..…..…………..…………....37 3.3 Materials and Methods……………...………………………..40

3.4 Results……………….……………………………………….49

3.5 Discussion ……………………...…………………………....64

3.6 Acknowledgments…………………………………………....69

3.7 References …………………………………………………...70

3.8 Supplemental Data …………………………………………..76

4 Cauliflower mosaic virus Major Inclusion Body Protein interacts with the Aphid Transmission Factor, the Virion-Associated Protein, and Gene VII Product…………………………...……………....77

4.1 Abstract.……………………… …………………………...... 77 4.2 Introduction………………………….…………..………...... 78 4.3 Materials and Methods, Results and Discussion…..……...... 79

4.4 Acknowledgements…………………………………………..85

4.5 References…………………………………………...... 86

4.6 Supplemental Data………………………………………...... 91

5. Additional CaMV Experiments: P6 and Reverse Transcriptase Protein (P5).....……....……………………………...…………………....92 5.1 Abstract……………………………………………………....92

5.2 Introduction………………....……………………...... ……....93

5.3 Materials and Methods...... 98

5.4 Results………………………………………………...... 100

5.5 Discussion ……………………………………………...... 109

6. Interactions of CaMV P6 with host proteins...... 113

6.1 Abstract…………………………………………………...... 113

6.2 Introduction………………………………………………....114

6.3 Materials and Methods...... 119

6.4 Results……………………………………………………....120

6.5 Discussion………………………………………………...... 126

7. Discussion/Future Work……………………………………………...... 130

References ………………………………………………...... 141

A Identification and analysis of viruses infecting Pelargonium ……….....160

A.1 Abstract………………………………………………...... 160

A.2 Introduction………………………………………………...161

A.3 Materials and Methods………………………………...... 164

A.4 Results………………………………………………...... 170

A.5 Discussion………………………………………………….181

A.6 Acknowledgments...... 185

A.7 References…………………………………………...... 186

A.8 Supplemental Data…………………………………………193

B Clones………………………………………………………………...... 195

List of Tables

3-S1: Primers used in the D3b study……………………………………………..76

4-S1: Primer used to amplify CaMV genes II, III, and VII……………………...91

A-1: Accession analyzed for PFBV, PLPV, ToRSV...... 172

A-S1: Primers used in PFBV and PLPV study…………………………………193

B-1: CaMV1841 and W260 full-length P6s in pENTR………………………...195

B-2: CaMV1841, W260, D4 P6 domains in pENTR…………………………...196

B-3: CaMV1841 P6 Mutants in pENTR………………………………………..197

B-4: CaMV1841 P2 full-length and domains in pENTR…………………….....198

B-5: CaMV1841 and W260 P5 full-length and domains in pENTR…………...199

B-6: CaMV1841 and W260 P3/P7 domains in pENTR………………………..200

B-7: CaMV1841 and W260 genes subcloned into pDEST10………………….200

B-8: CaMV1841 and W260 genes subcloned into pDEST15………………….201

xii

B-9: PFBV genes subcloned into pDEST15…………………………………....202

B-10: CaMV1841 genes subcloned into pDEST17…………………………….203

B-11: CaMV1841 and W260 genes subcloned into pEG202…………………..204

B-12: CaMV1841 and W260 genes subcloned into pJG4-5………………...... 205

B-13: CaMV1841 genes subcloned into pMAL………………………………..206

B-14: CaMV1841 genes subcloned into pSITE………………………………...208

B-15: CaMV pSITE constructs transformed into Argobacterium……………...211

xiii

List of Figures

1-1: Flow chart of main events during viral infection…………………………….3

1-2: The multiplication cycle of CaMV…………………………………………...7

1-3: The Cauliflower mosaic virus genome……………………………………...10

1-4: Movement protein (MOV), virion-associated protein (VAP), and insect

transmission(ATF)…………………………………………….....12

1-5: Structural elements of CaMV coat protein (P4)……………...……………..13

1-6: Structural elements of CaMV reverse transcriptase protein (P5)…………...14

1-7: CaMV P6 IBs by transmission electron micrograph………………………..16

1-8: Structural elements of the major IB protein (P6)………………………...... 19

1-9: Structural elements of P6 and P6 self-association domains………………...19

1-10: CaMV P6 self-association domains………………………………………20

1-11: Structural elements of CaMV coat protein, P6, eIF3g, and RL24………..22

3-1: Schematic diagram of P6 and location of mutations…………………….....51

xiv

3-2: Role of the D3b region in CaMV infectivity and protein binding………....53

3-3: Interactions of D3b mutant P6s with P6 self-association domain, D1……..54

3-4: Interactions of D3b mutant P6s with P6 self-association domain, D4…...... 56

3-5: Fluorescent inclusion bodies (IBs) formed by CM1841 P6 in Nicotiana

benthamiana……………………………………………………...58

3-6: Fluorescence microscopy of mutant P6 IBs………………………………..60

3-7: Propagation of D3b mutant viruses in turnips……………………………...62

4-1: Interaction of the gene III (P3) and gene VII products (P7) with the

gene VI product (P6) of Cauliflower mosaic virus………………81

4-2: Interaction of the gene II (P2) and gene VI products (P6) of

Cauliflower mosaic virus………………………………………...84

5-1: Structural elements of Retro and Pararetrovirus pol proteins……………....94

5-2: Multiple sequence alignment (MSA) and phylogenetic tree of the

Protease domain within different pol proteins………………....102

5-3: Multiple sequence alignment (MSA) and phylogenetic tree of the

RT domain within different pol proteins……………………….103

5-4: Multiple sequence alignment (MSA) and phylogenetic tree of the

RNase H domain within different pol proteins…………………104

5-5: Maltose binding protein (MBP) pull-down analysis of P5MC domain

interactions……………………………………………………...105

5-6: Maltose binding protein (MBP) pull-down analysis of P5 domains

with P6.………………………………………………………....106

5-7: Maltose binding protein (MBP) pull-down and fluorescence microscopy

analysis of P5MC interactions with P6…………………………108

5-8: Fluorescence microscopy of P5MC with P6 mutants……………………...109

6-1: Structural elements of Arabidopsis thaliana CHUP1……………………..117

6-2: Structural elements of Arabidopsis thaliana C2CDMT…………………...118

6-3: Structural elements of Arabidopsis thaliana FIT……………………….....118

6-4: Interaction of CaMV P6 and Arabidopsis thaliana CHUP1 protein………121

6-5: Interaction of CaMV P6 and Arabidopsis thaliana C2CDMT protein…....122

6-6: Interaction of CaMV P6 and C2CDMT homology from Nicotiana

benthamiana (C2*)………………………………….……….....123

6-7: Interaction of CaMV P6 and Arabidopsis thaliana FIT protein…………...124

6-8: Figure 6-8: Interaction of CaMV D4 strain P6 with Arabidopsis thaliana

xvi

CHUP1, C2CDMT, and FIT proteins, and C2CDMT

Homology protein from Nicotiana benthamiana…………...... 125

7-1: Flow chart on CaMV IB formation………………………………………..138

A-1: Self-association interactions of Pelargonium flower break virus

(PFBV) and Pelargonium line pattern virus (PLPV) coat

proteins (CPs)…………………………………………………...174

A-2: Identification of domains involved in self-association interactions

of PFBV and PLPV CPs………………………………………..175

A-3: PFBV p7 interactions……………………………………………………...177

A-4: PFBV p7 domain interactions involved in self-association…………….....179

A-5: Cross-interaction amongst the PFBV and PLPV CPs………………….....181

A-6: Summary of interactions amongst Pelargonium flower break virus

(PFBV) and Pelargonium line pattern virus

coat proteins(CPs) and PFBV CP with p7……………………..185

A-S1: Schematic presentation of full-length PFBV and PLPV CPs and there

fragments………………………………………………………..194

A-S2: Schematic presentation of full-length PFBV p7 and its fragments……...194

xvii

List of Abbreviations

ANOVA…………….Analysis of variance bHLH………………..Basic helix-loop-helix DNA binding domain BMV………………...Beet mosaic virus

CaMV……………….Cauliflower mosaic virus CHUP1……………...Chloroplast unusual positioning 1 CP…………………...Coat protein C2CDMT…………...Calcium-dependent-induced targeting protein

DRB4……………….RNA binding protein 4 DCL4……………….DICER-LIKE-4 DPI………………….Days post-inoculation DBD………………...DNA binding domain DLPs………………..Double layer particles

E.coli……………….. Escherichia coli eIF3g……………….. Eukaryotic translation initiation factor 3, subunit g ER…………………...Endoplasmic Reticulum

FIT…………………..Fe-deficiency-induced transcription factor 1

GFP………………….Green fluorescent protein GST………………….Glutathione s transferase

HBV…………………Hepatitis B virus HIV-1………………..Human Immunodeficiency Virus -1 HSP90 ………………Heat shock protein 90

IB……………………Inclusion body

MMLV………………Moloney murine leukemia virus MP…………………...Movement protein MSA…………………Multiple sequence alignment MBP…………………Maltose binding protein xviii

MES…………………Monohydrate 2-(N-morpholino) ethanesulfonic acid MMLV………………Moloney Murine Leukemia Virus MNSV ………………Melon necrotic spot virus

NPTII………………..Neomycin phosphotransferase II NSP2……………….. Nonstructural protein 2 NSP5………………...Nonstructural protein 5 NLS………………….Nuclear localization signal NES………………….Nuclear export signal

OPGC……………….Ornamental Plant Germplasm Center ORF…………………Open reading frame

P-domain……………Protruding domain PCR…………………Polymerase chain reaction PFBV………………..Pelargonium flower break virus PLPV………………..Pelargonium line pattern virus PMSF………………..Phenylmethylsulfonyl-fluoride PD…………………...Plasmodesmata PM…………………..Plasma membrane P1…………………...Movement protein P2…………………...Aphid transmission protein P3……………….…..Virion associated protein P4…………………...Coat protein P5…………………...Reverse Transcriptase Protein P6…………………...Major inclusion body protein P7…………………...Unknown protein

R-domain……………RNA-binding domain RT…………………...Reverse transcriptase RT-PCR……………..Reverse transcriptase-polymerase chain reaction RBD………………....RNA-binding domain RV…………………...Rotavirus RNase H……………..Ribonuclease H RFP………………….Red fluorescent protein

S-domain……………Central shell domain SDS…………………Sodium dodecyl sulfate

Tukey’s HSD………..Tukey’s honestly significantly different test TCV…………………Turnip crinkle virus TAV…………………Translational transactivation TBS………………….Tris-buffered saline

xix

TP…………………....Terminal protein TAD…………………Transcription activation domain TMV………………...Tobacco mosaic virus TF…………………...Transcription factor ToRSV……………...Tomato ringspot virus

Chapter 1

Introduction

1.1 Importance of plant and animal viruses

New emerging along with old viruses still threaten the lives of plants, animals and humans. Influenza, HIV and Tomato spotted wilt are just a few of the virus infections that not only require continuous monitoring but are also responsible for thousands of deaths annually and billions of dollars lost in crop production worldwide. A virus is an infectious agent with a simple composition that can multiply only in living cells of animals, plants, or bacteria. All viruses are obligate parasites utilizing host cellular machinery to reproduce. The viral genetic material i.e., RNA or DNA can adapt rapidly to host cellular conditions and viral infections can be either localized or spread to many distant locations in the host.

Recently there has been a thrust of research focused on how viruses infecting animals, humans and plants manipulate their host to increase their likelihood of survival and dissemination. A better understanding of how viruses manipulate their host and their means of transmission might aide in new drug therapy of mammalian viruses and controlling plant viruses. 1

1.2 Virus infection

To infect a host, both plant and animal viruses must perform a number of tasks (2011; Rothnie, 1994) (Fig. 1-1). First, a virus must gain entry into the host cell. For animal viruses this is accomplished through cell damage, receptor- mediated endocytosis, and direct membrane fusion. Plant viruses gain entry via mechanical damage. Since the cell membrane in plants is surrounded by a rigid cell wall, plant viruses require a wound for their initial entrance into a plant cell.

Wounds in plants can occur naturally, such as in the branching of lateral roots.

They may also be the result of agronomic or horticultural practices such as grafting or pruning, or other mechanical means. Plant viruses can also be transmitted by fungal, nematode, or parasitic plant infections; or by insects. These organisms create wounds permits the virus to pass through the cell wall. These organisms carry and can pass or transmit the virus.

Once inside a cell, the viral particle must be uncoated to allow for the release of the viral nucleic acid (Leclerc, 1999). The viral nucleic acid is then expressed and the viral genome is replicated. The newly synthesized viral genome is packaged into newly assembled virus particles to allow for movement either from cell-to-cell or by distances via vascular tissues or transmitted to new hosts, generating new infections.

Figure 1-1. Flow chart of main events during viral infection.

1.3 Plant and animal inclusion bodies

Unlike eukaryotic cells, where genome replication and transcription occurs in the nucleus, many RNA and some DNA viruses carry out viral genome replication and transcription in the cytoplasm (Boon, 2010). These viruses form organelle-like structures or compartments to increase replication efficiency and form a protective barrier from host defenses (Boon, 2010; Netherton, 2011).

Viruses can form either membrane-bound or non membrane-bound cytoplasmic compartments. These compartments are found in both plant and animal viruses.

Membrane-bound compartments are formed on different organelles within the cell, which occur in both animal and plant viruses (Netherton, 2011). Plant

3 viruses: tymovirus, Beet mosaic virus (BMV), and members of the tombusviridae form spherules at the chloroplast and peroxisome and double-membrane vesicles from the endoplasmic reticulum (ER), respectively. Animal viruses such as members of the nodaviridae and members of the alphavirus group form spherules at the endosome, lysosome and mitochondria. In contrast, flaviviruses form double membrane vesicles from the ER, while arteriviruses and coronaviruses form vesicle packets, derived from the ER membrane.

In contrast, some viruses form non membrane-bound compartments, which are also known as viroplasms or inclusion bodies (IBs). These IBs are dense, amorphous, aggregates mainly composed of viral nucleic acids and proteins generated from the virus genome. IBs are thought to be sites of viral DNA/RNA and protein synthesis, virus particle assembly and storage, which aids in viral propagation (Haas, 2005; Mazzolini, 1989; Rothnie, 1994). IBs can be found in either the nucleus or cytoplasm of infected cells (Bonneville, 1989; Covey, 1981;

De Tapia, 1993; Schoelz et al., 1986; Stratford, 1989). For example, members of the herpesvirus and adenovirus groups form nuclear IBs called cowdry type A and

B respectively (Burns, 2010) in their animal host cells. Plant viruses, in the rhabdovirus group also form nuclear IBs.

Cauliflower mosaic virus (CaMV) a plant pararetrovirus and Rotavirus (RV) a dsRNA animal virus, form electron dense cytoplasmic IBs to facilitate their lifecycles (Bonneville, 1989; Covey, 1981; Covey, 1985; Hu, 2012; Matthews,

1991; Mazzolini, 1989; Patton, 2006; Pesavento, 2006). These IBs are thought to be the sites of protein synthesis, replication, genome packaging and particle 4 assembly and storage. IBs may be mainly formed by a single viral protein or multiple polypeptides. RV utilizes the interaction of nonstructural protein 2 and 5

(NSP2 and NSP5) to form functional IBs (Patton, 2006; Pesavento, 2006).

However, RVs form intermediate virus particles known as double layer particles

(DLPs), which contain the newly synthesized viral genome (Arnold, 2009). It is not until the DLPs bud into the ER, that the mature viral particles are assembled, and released by lysis to move cell-to-cell. In contrast, CaMV utilizes a single protein termed gene VI product P6 to form functional IBs. Additionally, CaMV forms mature virus particles within IBs that are then able to accumulate within

IBs, move cell to cell or to other host.

1.4 Cauliflower mosaic virus Background

Pararetroviruses are a serious threat to both human and plant health. For example, Hepatitis B virus (HBV) affects approximately 350 million people globally and contributes to an estimated 620,000 deaths annually (2011). In the

U.S. alone, HBV caused chronic disease in 1.4 million people resulting in 3,000 deaths in 2011. Similarly, CaMV is in the top ten plant viruses of importance either from a scientific or economical standpoint (Scholthof, 2011).

From a economical standpoint, CaMV is the second leading virus disease in canola in England and becoming increasing prevalent in the middle east, especially Iran (Hunter, 2002). CaMV has narrow host range mainly infecting plants within the Brassicaceae and Solanaceae. However the Brassicaceae are among the most economically important crops in the world (Farzadfar, 2007) and

CaMV is of most importance due to its widespread distribution in temperate regions All CAMV isolates are able to infect a wide variety of Brassiciaceae plants however, only W260 and D4 CAMV strains can infect plants within the

Solanaceae family to include Nicotiana and Datura species. In addition, CaMV can infect Arabidopsis thaliana the model organism used to study virus replication, RNA silencing, and infection studies (Agama, 2002; Love, 2005;

Shepherd, 1979). CAMV can induce a variety of systemic symptoms including mosaics [irregular color pattern (light/dark green or yellow) over the entire leaf area], vein cleaning or banding which is caused by the tissues adjacent to the veins to become chlorotic (yellow in color), necrosis, where the plant tissue dies causing the tissue to turn brown, chlorotic lesion, which are yellow spots that form in specific locations over the small veins in the leaf, stunting and reduced fertility (Shepherd, 1981; Zijstra, 1996).

From a scientific standpoint, CaMV was the first dsDNA virus to be identified. This virus was also the first to produce IBs that contained virus-like particles. In addition, it was the first plant virus shown to replicate by reverse transcription and use a particular mode of translation. Lastly, the CaMV 35S promoter is used worldwide in plant biotechnology to express particular genes of interest at high levels. Therefore, the well-known molecular characteristics of

CaMV makes it a good model system to better understand reverse transcription, translational transactivation, and inclusion body formation, which will lead to more effective and novel ways of controlling retroviruses, pararetroviruses and rotaviruses.

Figure 1-2. The multiplication cycle of CaMV. The main steps of the viral cycle are: aphid-mediated entry of the virus into the host cell, NLS mediated transport of CaMV particles to the nuclear pore, import of the viral DNA into the nucleus, reparation of DNA sequence discontinuities and association with histones to form a minichromosome, transcription of the viral DNA by cellular RNA polymerase II, translation of the 19S and 35S RNA and spliced versions, replication of the genome and morphogenesis of viral particles in the elctron- dense viroplasms, and cell-to-cell movement of virus particles through tubules, targeting to the nucleus and aphid uptake. (Haas, 2002)

To understand how CaMV infects a host it is important to review the life cycle of this pathogen (Fig. 1-2). To infect a host, CaMV must gain entry into the host cell either by aphids or mechanical damage. Aphids carry out “non- circulative transmission” where the virus particle is attached only to the exterior mouthparts (stylet) and is released into a new host cell through a wound created by the aphid stylet (aphid transmission of Cauliflower mosaic virus, drucker). 7

Once the virus particle enters the host cell, its targeted to the nuclear pore by the nuclear localization signal (NLS) that is exposed on the surface of the mature virion (Haas, 2002). The virus particle then uncoats allowing the viral DNA to be released into the nucleus. Unlike adenoviruses, retroviruses, and some herpesviruses, CaMV genome does not integrate into the host chromosomes

(Morissette, 2010). The CaMV DNA associates with host histones to form independent minichromosomes (Covey, 1985). The now nuclear, circular, viral

DNA genome is transcribed by host RNA polymerase II from two promoters: the

19S and the 35S (Haas, 2002). The 19S promoter is the “weaker” of the two, while the 35S promoter is strong, constitutive, and expressed in most plants.

Thus, the 35S promoter is used in many plant biotechnology applications (Hull,

2000). Transcription from these two promoters gives rise to two transcripts; the

19S and the 35S RNAs (Covey, 1985; Rothnie, 1994). Both transcripts are polyadenylated at the same site following an AAUAAA signal motif (Hull, 2002;

Pooggin, 2001; Ryabova, 2000)

The subgenomic 19S RNA is monocistronic, encoding the gene VI product

(P6), which is the major IB protein and the most highly abundant viral protein infected cells (Cecchini, 1997; Haas, 2002). Whereas the 19S transcript serves as only a mRNA function, the 35S RNA serves a dual role in virus-infected cells.

The dual-function 35S RNA is longer than the CaMV genome, containing direct repeats at the ends, typical of retroid elements (Haas, 2002).

Thus, this RNA is copied into DNA by a reverse transcriptase encoded by the

CaMV genome (P5). Reverse transcription of the 35S RNA is thought to occur in 8

P6 IBs, but whether the process occurs within virus particles inside IBs or freely in IBs and the DNA is then incorporated into empty virus particles is unclear.

CaMV uses the methionine initiator tRNA as a primer for initiating first strand synthesis. After synthesis of the first strand, the P5-associated RNase H degrades the 35S RNA, but a purine rich region is retained that serves as a primer for initiating second strand synthesis.

The other function of the 35S RNA is to serve as a polycistronic mRNA for synthesis of all the viral proteins, including P6. The 35S RNA is capped allowing for cap-dependent translation. However, translation is complicated by a 5’ leader sequence that contains an elaborate secondary structure. This leader sequence, also known as the large intergenic region, controls the efficiency of translation of the downstream open reading frames

(ORFs). The leader sequence contains up to nine possible small ORFs; coding generally 1-4 amino acids that control translation efficiency of the downstream protein-coding gens. This long 5’ leader sequence can be bypassed by a mechanism known as ribosomal shunting to permit host ribosomes to initiate translation off the 35S RNA is mediated by a process called translational transactiviation accomplished by the P6 protein (See below).

Figure 1-3. The Cauliflower mosaic virus genome. Gray arrows indicate the 35S and 19S viral transcripts; black arrows, seven ORFs; IR spotted area, intergenic regions; roman numerals, the identity and function of the CaMV ORFs.

The CaMV genome contains seven genes (gene I-VII), encoding seven proteins (P1-P7) (Covey, 1985; Haas, 2002; Mason, 1987; Rothnie, 1994;

Shepherd, 1979) (Fig. 1-3). All of these proteins are synthesized from the 35S

RNA, including P6, which is also made from the 35S RNA. The order of ORFs on the 35S RNA starts with gene VII at the 5’end, and then genes I-VI. Gene VI is the last gene encoded at the 3’end of the 35S RNA.

Gene VII, first true ORF on the 35S RNA encodes P7, a small hydrophilic polypeptide. Mutations that delete most or all of gene VII appear to have no effect on viral infection, while mutations that eliminate the gene VII start codon are detrimental to virus infection. Therefore, it is unclear what role P7 plays in viral infection. Although it can be synthesized in yeast and Escherichia coli

(E.coli), P7 has not been detected in infected plant tissue, suggesting that it is unstable.

Gene I encodes the cell-to-cell movement protein P1 (Perbal, 1993). P1 forms tubules projecting through plasmodesmata (PD) and CaMV particles are observed inside these tubules. The N-terminal and central portions of P1 are required for tubule formation (Hohn, 2013; Huang, 2001; Thomas, 1999). In addition to its role in tubule formation, the central region of P1 also contains signals for targeting of the protein to the PD and a central RNA binding domain

(Koonln, 1991; Thomas, 1995)The function of the RNA-binding domain is unknown. The C-terminal region of P1 contains coiled-coil domain that permits the protein to form timers (Stavolone, 2005) (Fig. 1-4). Interestingly, the C- terminal region is also exposed on the inner surface of the P1 tubules. This is interesting because the C-terminus is the part of P1 that binds to the virion- associated protein (P3) (Hohn, 2013). This may be what allows CaMV particles to be loaded into the P1 tubules.

The fact that strains of CaMV, such as CM4-184, that harbors a deletion of most of gene II, and CM1841 which harbors a gene II mutation causing production of an P2 is dispensable for virus infection. However, neither CM4-

184, nor CM1841 are aphid transmissible and so P2 has been termed the aphid transmission factor. Gene II encodes the small protein, P2. The N-terminus of P2 is thought to interact with factors in the aphid stylet. The C-terminus contains two coiled-coil domains that permit P2 to self-associate, or bind to P3 (Blanc,

1993; Blanc et al., 1993; Hebrard, 2001; Hohn, 2013; Leh, 1999) (Fig. 1-4). 11

Figure 1-4. Movement protein (MOV), virion-associated protein (VAP) and insect transmission (ATF). Domain structures of CaMV VAP, ATF and MOV proteins. Coiled-coil domains direct homo-oligomerization and hetero- oligomerization (Hohn, 2013)

Gene III is essential for virus infection, but dispensable for virus replication (Haas, 2002; Jacquot, 1998; Leclerc, 2001). Gene III encodes P3, also called the virion-associated protein, hence, P3 interacts with virus particles (Hull,

2002). The C-terminal end of P3 interacts with virus particles and with double- stranded DNA. The central portion of P3 is dispensable and can be deleted without affecting virus infection. The P3 N-terminal region is a coiled-coil domain that interacts with other P3s, permitting the protein to form either a trimer or tetramer the depending on whether it is bound to virus particles (Hohn, 2013;

Stavolone, 2005) (Fig. 1-4). The P3 N-terminus is exposed on the surface of virus particles and it serves as an “arm” through which other viral proteins can attach to virions. Both P1 and P2 bind to P3 (Leh, 1999, 2001; Stavolone, 2005). Hence,

P3 is essential for both virus cell-to-cell movement as mediated by P1, and for aphid transmission, accomplished by P2.

Figure 1-5. Structural elements of CaMV coat protein (P4). P4 is 489aa in length and the various domains within P4 are indicated above.

The remainder of the CaMV genome and its encoded proteins resemble those of a retrovirus (Kobayashi and Hohn, 2003; Rothnie, 1994). For example, the CaMV protein most closely related to the retroviral gag protein is the P4 product. P4, encoded gene IV, is the protein mainly making up the bulk of CaMV particles (Cheng, 1992; Himmelbach, 1996). CaMV particles exhibit a triple- layered appearance with a hollow, layer consists of predicted antiparallel B-barrel structures. Layers two and three contain a lysine-rich region as one would expect for residues that interact with DNA (Fig. 1-5). These layers also contain a putative Cys/His motif that is highly conserved and binds to RNA in a non sequence specific manner. Mutations within either the lysine-rich region or the putative zinc finger abolish infectivity. P4 is also post-translationally modified.

P4 is glycosylated, although the function of this modification is unknown. P4 is also phosphorylated near the N-terminal region and this modification apparently regulates the activity of the N-terminal NLS (Champagne, 2004). P4 is also proteolytically processed at both ends. The N-terminal 76 amino acids are toxic and are thought to play a role in maintaining P4 in the cytoplasm. These amino acids are excised by the P5 protease. Further processing of the N, and additional

13 proteolytic cleavages at several sites near the C-terminal are mediated by proteases, but their source, either viral or host is unclear.

Gene V encodes the multifunctional P5 protein, which is analogous to the pol gene of retroviruses. P5 is essential for viral replication and mutations within this protein abolish infectivity. Like the pol protein, P5 contains an N-terminal aspartic proteinase, a central reverse transcriptase (RT) and a C-terminal RNase H domain (Fig 1-6). However, unlike retroviruses, CaMV P5 does not contain an integrase function. The full-length P5 protein is inactive with respect to RT function (Takatsuji, 1992). The N-terminal protease must be removed in order for the RT function to be active. The protease has been proposed to be active as a dimer and to self-process P5 to release the functional RT-RNase H segment.

However, it is unclear if the reverse transcriptase dimerizes or if it remains as a monomer. P5 protease also processes P4 N-terminus and P7.

Figure 1-6. Structural elements of CaMV reverse transcriptase protein (P5). P5 is 679aa in length and the various domains within P5 are indicated above.

The 3’-most ORF encoded by the 35S RNA is gene VI product, P6 (Haas,

2002). While P6 does not share sequence homology with, and is not a membrane protein like, retroviral env proteins, it is a protein located on the outside of viral particles. In addition to its role as the major CaMV inclusion body protein, P6 has been implicated in a variety of functions including: translational

14 transactivation (TAV), viral genome replication, virion assembly, virus-host interactions (host range control, resistance breakage, host hypersensitive responses, symptom formation) RNA silencing suppressor activity, and virus infectivity (Anderson et al., 1991; Bonneville, 1989; Himmelbach, 1996;

Kobayashi, 1998; Love, 2007; Mazzolini, 1989; Qui, 1997; Schoelz, 1986, 1988;

Schoelz, 1991; Stratford, 1989; Wintermantel, 1993).

CaMV IBs are primarily composed of P6 and contain the other viral proteins (Covey, 1981). However, unlike those of some other plant viruses, P6

IBs contain no cellular membranes or organelles (Cecchini, 1997). P6 IBs are electron dense, amorphous to round in shape, and are thought to be the site of

CaMV genome replication, viral protein synthesis, virion assembly and storage.

P6 IBs are not solid, but hollow, and these holes are where virion assembly and storage are thought to occur (Fig 1-7). CaMV IBs can vary in size depending upon the viral isolate and the host plant (Shalla, 1980). In addition, number virions free in the cytoplasm relative to those present in IBs also varies between different CaMV isolates.

Electron microscopy and immunogold staining revealed that the precursor coat protein is present in “small bodies” but absent from the mature IBs

(Champagne, 2004) (Fig 1-7). As stated previously, P4 is post-translationally processed to allow for the formation of mature virions and this modification cleaves both the N-and C-terminus of P4. Therefore, the precursor coat protein only present in “small bodies” helps support that CaMV IBs are the site of virion assembly and storage. Importantly, the P5 protease is responsible for P4 15 processing. P5 protease also processes P5 to activate the reverse transcriptase functionality. Hence, one might expect “small bodies” to be inefficient at reverse transcribing CaMV RNA to DNA.

Figure 1-7. CaMV P6 IBs by transmission electron micrograph. Immunogold labeling performed on thin sections with purified mouse polyclonal antibody (anti-N-terminal domain) as primary antibody followed by colloidal gold (10nm)- conjugated goat anti-mouse IgG+IgM. Localization of the N-terminal domain of CaMV pre-CP. A small body (SB) structure neighboring larger electro-dense inclusion body (IB) is specifically labeled (arrow). Particles present over the SB and virions (V). (Champagne, 2004)

Using transgenic Arabidopsis lines, (Geri, 2004) showed that expression of P6 alone can induce the formation of IB-like structures. These data indicate that the other viral gene products do not play a primary role in the formation of

IBs. Transgenic plants expressing only P6 also show virus-like symptoms,

16 implicating this protein as a major symptom determinant. Furthermore, a strong correlation between symptom phenotype severity and P6 levels was observed

(Matthews, 1991). In addition, the transgenic expression of gene VI from different CaMV isolates induced different types of symptoms.

Furthermore, P6 is an important player in determining host range, symptom formation and host hypersensitivity response (Broglio, 1995; Daubert,

1990; Schoelz and Shepherd, 1988; Wintermantel, 1993). However, the relationship among these various processes are very complex. Most work to analyze these processes has been done with three CaMV isolates. CaMV D4 strain has a wide host range affecting both solanaceous and brassica species.

CaMV W260 can affect a limited number of solanaceous species and brassica species whereas CaMV1841 is limited to only infecting brassica species. For example, CaMV D4 strain can induce a systemic mosaic in both Nicotiana clevelandii and Nicotiana edwarsonii where as CaMV 1841 and W260 remain symptomless (Schoelz and Shepherd, 1988). Interestingly, W260 strain induces systemic cell death in N. clevelandii and host hypersensitive response in N. edwardsonii. Moreover, mutations within gene VI or CaMV genomes harboring chimeric ORFs derived from different strains influence symptom formation and severity. Infecting Datura stramonium with wild-type CaMV D4 strain or D4 mutants containing either a point mutation at the N-terminus or C-terminus of gene VI show systemic mosaic symptoms and necrotic local lesions, This indicates P6s impact on host-specific symptom changes (Daubert, 1990). In addition, domains within gene VI influence systemic infection of Nicotiana

17 bigelovii when these domains were exchanged between CaMV strains D4, W260 and CM1841 (Wintermantel, 1993). Chimeric viruses between CM1841 and either D4 or W260 induced systemic infection but CM1841 alone did not.

However, sequence variation between gene VI W260 and D4 at the N-terminus are responsible for systemic infection and host range.

Additionally, P6 regulates protein synthesis through a process called translational transactivation (TAV) (De Tapia, 1993). TAV permits translation of all seven ORFs encoded by the polycistronic 35S RNA. This processes occurs after the translation of one polypeptide off the 35S RNA, the large ribosome (60S) stays bound to the RNA, which allows for re-initiation of protein synthesis at the next ORF (Bonneville, 1989; Bureau, 2004; Leh, 2000; Park, 2001). P6 also stabilizes the interactions between the components of the host translation machinery, which will be discussed below.

In recent years, P6 has been shown to influence host RNA silencing pathways to enhance viral infectivity. P6 contains NLS that permit it to affect host nuclear processes. This nuclear localization is essential for inhibition of host

RNA silencing pathways. P6 influences the trans-acting short interfering RNA

(tasiRNA) pathway of CaMV host plants. Furthermore, P6 binds directly to

Double-Stranded RNA Binding Protein 4 (DRB4) (Haas, 2008). DRB4 is a double-stranded RNA binding protein that binds DICER-LIKE-4 (DCL4) and aids in the DICER cleavage of double-stranded RNAs. P6 binding to DRB4 has been proposed to be a possible mechanism by which the tasiRNA pathway is inhibited. 18

Figure 1-8. Structural elements of CaMV major IB protein (P6). Gray boxes indicates nuclear export signal (NES) or nuclear localization signal (NLS), Blue box indicates translational transactivation (TAV), Red box indicates RNA- binding domains (RB), and Green box indicates punative zinc binding domain (ZF).

Gene VI product P6 is composed of 520 amino acids that forms alpha- helical motif and is the least conserved protein within the CaMV genome. It contains a nuclear export signal (NES) at the N-terminus, TAV domain, two

(NLS), two RNA binding domains, and a punative zinc finger at the C-terminus

(Fig. 1-8).

Figure 1-9. Structural elements of P6 and P6 self-association domains. The P6 self-association domains are indicated above.

In order to form IBs CaMV P6 must self-associate. There are four independent domains (D1-D4) involved in P6 self-association (Haas, 2005; Li,

2002) (Fig.1-9). The N and C termini (domains D1 and D4, respectively) as well as two internal regions (domains D2 and D3) bind to the full-length P6. Both D1 and D3 are important for self-association and viral infection (Agama, 2002; Haas,

2005; Li, 2002; Schoelz, 1988; Wintermantel, 1993). Domain D1 shows self- association while domain D3 binds to the other three domains but not itself (Li, 19

2002). Each independent domain is responsible for the variety of P6 functions, which will be discussed below. D1 can self-associate and bind to D3. D3 can bind to both D1 and D4. And D4 can bind to D3 (Fig. 1-10). Each independent domain is responsible for the variety of P6 functions, which will be discussed below.

Figure 1-10. CaMV P6 self-association domains. P6 self-association domains are indicated above and the arrow indicate P6 interaction domains.

Domain D1 (amino acids 1-110) plays a role in virus-host interactions and influences CaMV host range (Agama, 2002; Li, 2002; Schoelz, 1988;

Wintermantel, 1993). D1 contains the determinant that permits certain CaMV isolates to overcome resistance in particular Arabidopsis thaliana ecotypes

(Agama, 2002) and contains a nuclear export signal (amino acids1-31) that prevents the accumulation of monomeric form of P6 to accumulation in the nucleus (Haas, 2005). The alpha helix motif of a leucine zipper located in D1

(amino acids 4-31) is required for P6 self-association.

Domain D2 (amino acids 106-253) contains the minimal portion of P6 necessary for translational transactivation (mini-TAV) (De Tapia, 1993). The mini-TAV is unable to bind ssRNA but contains a dsRNA-binding domain

(amino acids 140-182) (De Tapia, 1993). D2 physically interacts with ribosomal

20 proteins L13 and L18 that bind to the N-terminal region of the mini-TAV

(Bureau, 2004; Leh, 2000). L13 and L18 compete for binding at the same region on the mini-TAV, which suggests the interaction between the mini-TAV and L13 and L18 occur at different stages of translation. L13 and L18 can also interact with one another (Bureau, 2004; Park, 2001). The mini-TAV also contains a hydrophilic motif located in the central region (Cerritelli, 1998). Deletion of this motif abolished the mini-TAV activity. D2 also contains a NLS at the C- terminus.

Domain D4 (amino acids 414-520) influences virus-host interactions

(Agama, 2002; Schoelz, 1988; Wintermantel, 1993). For example, point mutations within C-terminal region of D4 affected virus host range is affected

(Daubert, 1990), whereas viruses harboring C-terminal deletions of D4 showed either impaired movement or were non-infectious (Turner, 1996). D4 also contains a zinc finger binding domain (amino acid 455-480) that has an unknown function.

Domain D3 (amino acids 249-379) is important for P6 self-association (Li,

2002). Full-length P6 lacking D3 is incapable of self-association and binding to the other independent domains. The inability of D2 and D4 to bind to full-length

P6 lacking D3 suggests either the position where D2 binds was deleted or this deletion altered P6 conformation hindering interaction with D2 and D4 which may influence virus-host interactions (Li, 2002). In addition, deletion of the D3 coding portion of gene VI from the CaMV genome resulted in a non-infectious virus, indicating, that this region is also required for systemic viral infections. 21

Figure 1-11. Structural elements of CaMV coat protein, P6, eIF3g, and RL24. The size of each protein in amino acids is indicated above. The doted lengths in red indicate the portions where P4, eIF3g, and RL24 interact with P6.

D3 can be divided into three regions, which have different functions (Fig.

1-11). The N and C-terminal regions (amino acids 249-308 and 344-379; D3a and D3c respectfully) each contain a non-sequence specific RNA-binding domains (De Tapia, 1993). The central portion of D3 (amino acids 309-343;

D3b) is a highly conserved spacer region separating D3a and D3c, which contain a NLS. D3a contains the binding sites for the ribosomal protein L24 and the translation factor eIF3g (Park, 2001). Although the complete mechanism of P6s role in TAV is not fully understood, it has been proposed that this region of P6 keeps eIF3g bound to the ribosome during elongation by shuttling eIF3 between

40S and 60S ribosomal subunits (Bureau, 2004; Park, 2001). Furthermore, the last 11 amino acids of D3b and D3c contain the coat protein (P4) binding site

(Ryabova, 2000, 2002). To date, the function of D3 has been minimally characterized.

1.5 Cauliflower mosaic virus P6 interactions

As mentioned, P6 can self-associate, which is paramount for IB formation and the ability of CaMV to carry out viral propagation. In addition, P6 can interact with other viral proteins and host proteins. Previous studies have shown that P6 can interact with P1 and P4 (Himmelbach, 1996; Li, 2002). The P6-P1 interaction might occur to prevent P1 from forming tubules until it reaches the cell periphery. P1 tubules are only observed in PD. The P4-P6 interaction has been suggested to play a role in managing viral assembly (Himmelbach, 1996).

P6 might act as a chaperone to ensure proper folding of the capsid protein. The portion of P4 that P6 binds overlaps with the region of P4 that DNA binds. It is possible that P6 binds to P4 and holds it in an “open” state. Viral DNA could then interact with this open P4. This would displace P6 and permit viral DNA to be incorporated into the assembling virus particle.

P6 can also interact with host proteins. P6 directly interacts with components of the host translational machinery to carry out TAV. P6 directly interacts with translation initiation factor 3 subunit g (eIF3g). P6 also interacts with large ribosomal subunit proteins L18, L13, and L24 (Bureau, 2004; Park,

2001; Ryabova, 2002). As mentioned previously, L18 and L13 compete to bind to a sequence specific site on the mini-TAV. L24 and eIF3g bind to the same site within P6 domain D3 and compete with each other for this interaction as well.

L24 binding could be extremely important as acts a bridging protein between the

40S and 60S ribosomal subunits, something that could be essential for reinitiation of translation of the downstream ORFs on the 35S RNA.

In addition to interacting with host factors for TAV, P6 also binds to plant proteins involved in defense. As mentioned previously, P6 binds directly to

DRB4 and this interaction is thought to disrupt host silencing pathways. RL13 is also a component of Arabidopsis thaliana RNA polymerase V (Bureau, 2004).

So perhaps by binding to RL13, P6 sequesters this protein away from the nucleus and interferes with transcriptional gene silencing as well.

1.6 Hypothesis

During the course of both plant and animal virus infections, IBs are formed (Knipe, 1990b; Matthews, 1991). IBs are thought to play a pivotal role in virus infection, as they are often sites of virus genome replication and virion assembly. However, whether these structures are actually required for viral infection or are a consequence of the disease process is unclear and a detailed understanding of how IBs function in viral infection is still lacking. Because

CaMV IBs do not contain cellular organelles or membranes like those of other viruses, their simpler organization makes them a useful model for studying the formation, function and role of these structures in virus infection. By elucidating the design principles of CaMV IBs, we may uncover insights that could be applicable to other viruses infecting both animal and plant hosts. These insights could lead to novel strategies for controlling viral infections.

CaMV IBs are mainly composed of a single protein, P6. Therefore, if

IBs are required for infection, mutation of P6 self-association domains could compromise IB structure and possibly viral propagation. On the other hand, if

IBs are merely a consequence of viral infection, then we might not expect mutations that impair P6 self-association to have a drastic effect on CaMV propagation.

It is possible that mutation of the self-association domains may completely impair IB formation. If this is true, then if these mutants are impaired in propagation, that would suggest that IBs are required for viral infection. Alternatively, if the mutants are incapable of forming IBs and yet, propagate like wild type, then IBs are likely not required for infection.

CaMV IBs appear to go through a progression from “small bodies” to large IBs. The “small bodies” are apparently not as effective at proteolytic processing of viral proteins than the large ones. Therefore, one possible outcome for our mutants would be that if P6 self-association is impaired, the formation of large IBs from “small bodies” might be compromised. If this is true, then the viruses trapped in “small body” formation would propagate inefficiently and infection would be impaired. Therefore our first hypothesis is that CaMV P6 IBs are required for virus infection and that mutation of P6 self- association domains should impair virus propagation and IB formation.

In addition to its role in IB formation, P6 has also been proposed to have a chaperone function. P6 has been suggested to help assemble virus particles. If

P6 is a chaperone, then it may bind to viral proteins other than P4 and the 25 movement protein, P1. P6 has been suggested to stabilize P3, P4, and P5

(Kobayashi, 1998). This suggests that P6 interacts with P3 and P5 as well as the published interaction with P4. Finally, P4 does not directly interact with P1 for cell-to-cell movement yet virus particles do move through plasmodesmata. P1 is attached to virus particles via P3, which forms an “arm” projecting from virions.

Therefore, it is possible that because P6 interacts with P4 and P1 (Himmelbach,

1996; Li, 2002) that it may also interact with P3 to help form the complexes permitting cell-to-cell movement. Because viral reverse transcription occurs in

IBs (Mazzolini, 1989), it is possible that P5 interacts with P6. P2 has been suggested to stabilize P6 IBs, again suggesting an interaction among these two proteins. Finally, P7 is the first ORF encoded by the 35S RNA and is the only one that can be synthesized without TAV activity. Therefore, P7 may play a role in regulating aspects of TAV activity and perhaps, this suggests that P7 interacts with P6. Therefore, our second hypothesis is that P6 interacts with all of the

CaMV proteins.

Finally, P6 is an important CaMV protein influencing virus-host interactions. For it to mediate many activities, it is likely that P6 interacts with a variety of host factors. Therefore, our third hypothesis is that P6 interacts with several host factors and that these interactions may also influence the efficiency of virus propagation.

Chapter 2

Material and Methods

Plasmids, Enzymes, Yeast Media, and Yeast and Virus Strains

Yeast plasmids pEG202, pJG4-5, reporter plasmid pSH18-34, and yeast strain EGY48 were gifts from Dr. Roger Brent (Molecular Sciences Institute,

Berkeley, CA). Additionally, pEG202 and pJG4-5 were converted into gateway compatible vectors . Yeast media, SD minimal base and SD Gal/Raf along with

DO supplement media were purchased from CLONTECH (Palo Alto, CA). SalI fast digest restriction enzyme used was purchased from thermo scientific (Wayne

MI). All other restriction enzymes, Taq DNA polymerase, and T4 DNA Ligase was purchased from Promega Corporation (Madison WI) and were used as stated by the manufacturer. Pfu DNA Polymerase was purchased from Stratagene (La

Jolla, CA) and Phusion DNA polymerase was purchased from New England

BioLabs (Ipswich, MA). The cloned genome of CaMV isolate CM1841

(pCaMV10) was provided by Dr. S.H. Howell (Iowa State University, IA)

(GARDNER 1981). The cloned genome of CaMV isolate W260

(SCHOELZ/SHEPHERD 1988) were provided by Dr. J.E. Schoelz (University of

Missouri, MO)

Construction of clones

All primer sequences used to amplify the genes within these studies are found in Appendix B. The PCR products containing EcoRI and XhoI restriction sites at the 5’ and 3’ end, respectfully, were inserted into the EcoRI and XhoI sites in both yeast two-hybrid vectors, pEG202 and pJG4-5 (Gyuris, 1993). CaMV

CM1841 gene VI and regions of VI encoding the self-association domains were described previously (Li, 2002). The primers sequences used to amplify genes and gene fragments for cloning into pENTR are described in Appendix B. These constructs were then sub-cloned into Gateway® compatible yeast two-hybrid vectors, GWpEG202, GWpJG4-5.

Site-directed mutagenesis constructs for yeast two-hybrid

Site-directed mutagenesis of D3b was performed using the QuikChange®

XL Site-Directed Mutagenesis Kit from Stratagene (La Jolla, CA). Single- stranded DNA oligonucleotide primers were designed to individually change each of the five pre-determined amino acids to an alanine. These primers were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA) and are listed in Supplementary Table 3-S1. Site-directed mutagenesis reactions were preformed according to the manufacturer’s specifications. The template used was

P6 domain D3 inserted in pJG4-5, described previously in (Li, 2002). After each

PCR reactions, the samples were digested with DpnI included in the kit and the

28 plasmids were introduced into E. coli by transformation according to the manufacturer’s specifications. Transformants harboring the designated mutations were confirmed by sequencing (Plant-Microbe Genomics Facility, Columbus,

OH). The plasmid pJG4-5 harboring a full-length version of gene VI lacking D3

(FL-ΔN; described in Li and Leisner, 2002) was digested with NcoI. The various amino acid substitution mutants were excised from pJG4-5 by NcoI digestion and the D3 fragment harboring the mutations were ligated into FL-ΔN. This reconstitutes a full-length gene VI harboring D3b mutations in pJG4-5.

Constructions were confirmed by sequencing.

Yeast two-hybrid analysis

Recombinant plasmids were introduced into Saccharomyces cerevisiae

EGY48 containing the β-galactosidase reporter plasmid pSH18-34 using the

LiOAc method (Ausubel 1998 and CLONTECH). First the pEG202 plasmids containing the gene of interest fused to the DNA binding domain (DBD) were introduced into EGY48 containing pSH18-34 and selected on synthetic defined

(SD)/-Ura-His media. Colonies grown on –Ura-His media were used to create yeast lines containing pEG202 plasmids. Second, the pJG4-5 plasmids containing the gene of interest fused to the B42 activation domain were introduced into the yeast lines containing the pEG202 plasmids. The transformants were plated on either SD Base/GAL/RAF/-Ura-His-Trp or SD

Base/GAL/RAF/-Ura-His-Trp-Leu. Colonies grown on SD Base/GAL/RAF/-

Ura-His-Trp-Leu were used to measure β-galactosidase Activity. Colonies grown

29 on SD Base/GAL/RAF/-Ura-His-Trp were used when measuring reporter gene activity of negative controls as previously described in Li and Leisner 2002.

Pull-down Assays

The primer sequences used to amplify CaMV genes cloned into pENTR and used in pull-down assays are shown in Appendix B. These constructs were then subcloned into Gateway® compatible bacterial expression vectors,

GWpMAL-c2x and pDest15. Proteins used in the pull-down assays were expressed in E. coli. Bacteria expressing MBP- and GST-tagged proteins were grown in 250 ml LB cultures (bacterial cultures expressing MBP-tagged proteins were supplemented with glucose) until the optical density (OD) reached 0.6. E. coli containing the expression plasmids were induced at 37°C for 2 hours with

1mM isopropyl B-D-1-thiogalactopyranside (IPTG). For pull-down experiments employing P5 constructs, induction was at 16°C for 12 hours using 0.1 mM IPTG.

Cultures were centrifuged at 8000rpm for 10 minutes and the pellets were resuspended in 10ml of MBP column buffer (20mM Tris-HCl, pH7.4, 200mM

NaCl, 1mM EDTA) with protease inhibitor cocktail (Sigma Aldrich) 1mM phenylmethylsulfonyl-fluoride (PMSF) (Fisher Scientific), and 10 U of RNaseA

(Five Prime, San Francisco, CA). Cells were lysed using a French press and centrifuged at 19,600g for 25 minutes. The supernatant was collected and protein concentration was analyzed by Bradford assay. Equal amounts of protein were used MBP pull-down assay were performed as described by (Hapiak, 2008).

Fluorescence microscopy

Constructs used in fluorescence microscopy studies were subcloned into the binary pSITE vectors containing either a C-terminal GFP or RFP tag.

Constructs were confirmed by sequencing (Genewiz, NJ). Constructs were then introduced into Agrobacterium tumefaciens strain AGL-1 by electroporation.

Transformants were then selected on appropriate antibiotics (100mg/ml

Spectinomycin), screened for inserts by colony PCR using gene specific primers

(see appendix B).

The A. tumefaciens containing the pSITE constructs and others containing the pKYLX7-P19 construct (expressing the P19 silencing suppressor) were grown in 3 ml LB starter cultures containing the appropriate antibiotic (100mg/ml

Spectinomycin for pSITE constructs and 50mg/ml Kanamycin for the pKYLX7-

P19 construct) at 28°C for 24 hours. The P19 silencing suppressor was used to ensure expression of the agrointroduced constructs. The starter culture (1 ml) was added to 40 ml of LB with the appropriate selective antibiotic for 24 hours at

28°C. Each Agrobacterium culture was permitted to reach an OD at 600 nm of

0.8-1.0 before induction. Bacteria were then centrifuged at 8000g at 4°C for 10 minutes. The supernatant was discarded and the pellet was resuspended in 20 ml of induction medium (20 mM MES [monohydrate 2-(N-morpholino) ethanesulfonic acid] pH5.4, 60 mM sucrose, 55 mM glucose, and 2 mM acetosyringone dissolved into N-N-dimethylformamide) for 16 hours at 28°C.

Agroinfiltration experiments were performed on Nicotiana benthamiana. Holes were generated in the abaxial surface of young leaves in plants approximately 4-6

31 weeks old with 4 true leaves using a 14.5 gauge syringe needle. Infiltrated left sizes vary due to growth and seasonal conditions. Plants approximately 4-6 week old were used. Depending upon the experiment either one leaf was infiltrated per construct along with the 19 plasmid at a 1:1 ratio (1ml for each plasmid) or three leaves were infiltrated per construct along with the 19 plasmid at again 1:1 ratio (2ml for each plasmid). If co-infiltrations were completed, constructs were still infiltrated at a 1:1 ratio. The aim was to infiltrate the entire surface area of each leaf. Each experiment was repeated in triplicate

Three days post-infiltration the N. benthamiana leaves were observed using Leica SP8 Confocal Laser Scanning Microscope at the University of

Toledo. Circular sections of the infiltrated leaves were cut out, placed on coverslips and visualized on the confocal microscope. GFP and RFP were excited at wavelengths of 488 and 543 nm, respectively. Images were acquired sequentially line-by-line when using multiple flourophores simultaneously. The

Leica oil immersion 63x lens was used to acquire images at a resolution of 1024- by-1024 pixels. Images obtained were chosed at random however, 10-15 fields were analyzed per construct by Z-stack. Z-stacks varied in size due to the thickness of the plant tissue. Images were processed using Leica software.

Quantitative analysis of confocal images

The Leica processing software was used to measure the area of inclusion bodies for wild-type P6 for both CaMV strains 1841 and W260. The R-program was used to remove all outliers and perform unpaired t-test. Approximately 250

IBs were measured per construct tested. 32

The IBs formed by both wild type CM1841 and the point mutant P6s were measured using ACCESS software to determine the average area of IBs per construct. The R-programs was again used to remove all outliers and used to perform a one-way analysis of variance (ANOVA) and TukeyHSD test.

Protein expression analyses

To determine expression levels for wild type and mutant P6s, agroinfiltrated N. benthamiana leaves were harvested and ground in extraction buffer as described in Angel et al. (2013). Extracts were centrifuged at 2000g for

10 minutes. Supernatants were passed through a 0.45 mm filter to remove debris and centrifuged at 13,000 x g in a microcentrifuge for 5 minutes. The supernatants were then transferred to a mixed with 2x SDS loading dye and boiled

10 minutes. Samples were separated by electrophoresis through a 10% polyacrylamide gel at 180 volts in a Mini-PROTEAN® 3 Cell (Bio-Rad

Laboratories, Inc., Hercules, CA). Proteins were then electrophoretically transferred to a nitrocellulose membrane in a Mini Trans-Blot® Transfer Cell

(Bio-Rad Laboratories, Inc.). The membrane was washed with 1 X TBS (20 mM

Tris-HCl, pH 7.6, 0.14 M NaCl), and blocked with 5% non-fat dry milk for 2 hours at room temperature. The membrane was cut into an upper and lower portion to allow for detection of both P6:GFP (upper) and neomycin phosphotransferase II (lower) on the same blot. The Neomycin phosphotransferase II gene is encoded on the same piece of T-DNA as the P6-

GFP constructions and so serves as a good loading control.

The membranes were incubated with goat-anti-GFP (Santa Cruz,

Biotechnology, Santa Cruz, CA) primary antibody (1:1000 dilution) or rabbit- anti-NPTII (neomycin phosphotransferase II; Sigma-Aldrich, ST Louis, MO;

2mg/ml). The membranes were washed three times with 1 X TBS containing

0.05% Tween-20, and incubated with secondary antibody donkey-anti-goat

(Sigma-Aldrich, ST Louis, MO) at 1:5000 dilution or goat-anti-rabbit (Santa

Cruz, Biotechnology, Santa Cruz, CA) at 1:5000 dilution, respectively. The membranes were washed with 1X TBS containing 0.05% Tween-20, covered with luminal/peroxide chemiluminescent reagent (Millipore, Inc., Billerica, MA) and exposed to X-ray film. Films (HyBlot CL Autoradiography Film, Danville

Scientific, Inc.) were developed using a Konica SRX-101® developer.

Chapter 3

Mutations Within A 35 Amino Acid Region Of P6 Influence Self-Association, Inclusion Body Formation, And Caulimovirus Infectivity

3.1 Abstract

The gene VI product (P6) of the plant pararetrovirus Cauliflower mosaic virus is a multifunctional protein, essential for infectivity that self-associates to form cytoplasmic inclusion bodies (IBs). Previously, P6 was found to contain four regions involved in self-association termed D1-D4. D3, a central region of

P6, binds to D1 and D4 but not to itself and is required for viral infectivity.

Functionally, D3 could be separated into three regions: an N- and a C-terminal

RNA-binding domain separated by a central spacer region (D3b). This paper focuses on D3b, which is predicted to be mainly alpha-helical in structure and more highly conserved than the other portions of D3. D3b binds to full-length P6 as well as D1 and D4, although interactions with the complete polypeptide are more efficient than with the individual regions. Full-length P6s with single amino acid substitutions within the D3b region showed a variety of binding

35 characteristics with D1and D4. All of the mutant P6s showed a reduction in D1 binding. Single amino acid substitutions at the N- and C-terminal ends of the major predicted a-helix within D3b greatly reduced binding efficiency to D4.

Mutations near the middle of D3b generally showed more wild type levels of interaction with the D4. Full-length CM1841 P6 fused with green fluorescent protein (GFP) forms IBs when transiently expressed in Nicotiana benthamiana leaves. GFP-tagged full-length P6s harboring the D3b mutations all formed IBs as well. However, there were significant differences in IB size among the mutants. Those mutants that most strongly affected binding to D4, induced the formation of smaller IBs. The mutants not strongly affected in binding to D4, induced IBs that were the same size as those observed in plants expressing wild type P6. Viruses harboring these same mutations were all infectious, but they varied in their degree of infection efficiency. Viruses harboring the mutations that did not strongly affect D4 binding showed near wild type levels of viral DNA in inoculated leaves. Viruses harboring mutations that strongly reduced D4 binding, showed reduced levels of viral DNA in inoculated leaves. These results were also reflected in systemic infection data. Mutants not strongly impaired in

D4 binding showed near-wild type levels of systemic infection, while mutants that were reduced in D4 binding showed diminished systemic infection. Taken together, these data suggest that mutations influencing P6 self-association lead to altered IB formation and a reduction of virus infection.

3.2 Introduction

The proteins encoded by many viral genomes are often multi-functional

(Hull, 2002). A good example of this is the major inclusion body protein (P6)

(Covey and Hull, 1981) of the plant pararetrovirus Cauliflower mosaic virus

(CaMV) (Hull, 2002). CaMV P6 is encoded by gene VI and has been implicated in a variety of functions such as: translational transactivation (TAV), host range determination, movement, replication, and silencing suppression) (Bonneville et al., 1989; Haas et al., 2008; Kobayashi and Hohn, 2003; Love et al., 2007;

Schoelz et al., 1986; Schoelz and Wintermantel, 1993). It is likely that many of these activities are mediated via interactions of P6 with viral and host proteins.

Indeed, P6 has been shown to interact with CaMV proteins: P1 (movement protein) and P4 (coat protein) (Hapiak et al., 2008a; Himmelbach et al., 1996).

Similarly, P6 also interacts with a variety of host proteins, such as ribosomal proteins RL13, RL18 and RL24 as well as translation factor eIF3g (Bureau et al.,

2004; Leh et al., 2000; Park et al., 2001). These interactions have been implicated in translational transactivation. P6 also interacts with the plant myosin CHUP1

(Angel et al., 2013).

In addition to interactions with other proteins, P6 also self-associates

(Haas et al., 2005; Li and Leisner, 2002). P6 self-association involves four regions, termed D1-D4, all of which can bind to the full-length protein. The N- terminal 110 amino acids (D1) is essential for P6 self-association (Haas et al.,

2005). D1 can self-associate, independent of the rest of P6. D2 (amino acids

156-253) may bind weakly to D3 (Li and Leisner, 2002). D3, the central portion of P6 (amino acids 249-379) binds well to D1 and D4 (amino acids 414-520), but not itself. Deletion of D3 within the context of the CaMV genome resulted in a non-infectious virus. Hence, D3 is an essential part of P6 and warrants further investigation.

D3 possesses a tripartite organization: The N-terminal portion of D3

(amino acids 249-308; D3a) contains a non-sequence specific RNA-binding domain (De Tapia et al., 1993). This region also contains the binding sites for the large ribosomal subunit protein RL24 and translation factor eIF3g (Park et al.,

2001). The central portion (amino acids 309-343; termed D3b) contains part of the P4 binding site at it’s C-terminal end. D3b also is part of a region previously proposed to play a role in P6 interactions. Deletion of the TAV domain from P6 results in a protein incapable of enhancing translation of proteins from the CaMV polycistronic 35S RNA (De Tapia et al., 1993). Addition of this deleted form of

P6 to TAV assays appears to act as a dominant-negative, interfering with TAV activity. Deletion of a P6 region containing D3b, in addition to the TAV domain inhibits the dominant-negative activity. The C-terminal portion (amino acids 344-

379; D3c) also contains part of another non-sequence specific RNA-binding domain and a portion of the capsid protein (P4) binding site (De Tapia et al.,

1993; Ryabova et al., 2002).

In addition to performing its many tasks mentioned above, P6 is also the major inclusion body (IB) protein and its self-association is likely essential for the formation of these structures (Covey and Hull, 1981). Viruses infecting both 38 animal and plant hosts induce the formation of IBs (Knipe, 1990a; Martelli and

Russo, 1977). However, it is unclear whether IB formation is important for infection, or if it is merely a consequence of infection. For CaMV, IBs are where viral genome replication, protein synthesis, and virion assembly are thought to occur (Haas et al., 2002; Hull, 2002; Mazzolini et al., 1989). Therefore, altering

P6 self-association would be expected to affect IB formation and may reduce viral propagation. In this paper this possibility is investigated.

The interactions of D3b with full-length P6 and its self-association domains D1 and D4 were investigated. Single amino acid substitution mutations were generated in D3b to examine their effects on P6 binding to the self- association domains. These same mutations were also tested for their effects on

IB formation. Finally, viruses harboring these substitution mutations were tested for their ability to propagate in inoculated leaves and to cause a systemic infection. Here we predict that substitution mutations with the strongest effects on binding to the P6 self-association domains would have impaired IB formation and reduced viral infection, both in inoculated leaves and in the ability to infect the host systemically.

3.3 Material and Methods

Secondary Structure Prediction and Calculation of Sequence Variability.

The secondary structure of the D3b region of CaMV isolate CM1841

(Gardner et al., 1981) was predicted using the Garnier-Robson model in the

Protean software contained within the Lasergene Software package. Helical wheels were constructed using the same software. Multiple sequence alignment of the gene VI nucleotide and amino acid sequences encoding D3 for 20 CaMV isolates (given in Supplementary Figure 2) was performed using the Megalign

Program within the Lasergene Software package. The number of sites (either amino acid or nucleotide) that varied within each segment of D3 were totaled and divided by the total number of residues (either amino acids or nucleotides) within that segment to determine the number of variable sites per residue.

Enzymes, Plasmids, Bacteria and Yeast.

All restriction enzymes Taq DNA polymerase, and T4 DNA Ligase was purchased from Promega Corporation (Madison, WI) and were used as stated by the manufacturer’s protocol. Pfu DNA Polymerase was purchased from

Stratagene (La Jolla, CA). The cloned genome of CaMV isolate CM1841

(pCaMV10) was provided by Dr. Stephen Howell (Iowa State University, Ames,

IA). Yeast plasmids pEG202, pJG4-5, and pSH18-34, along with yeast strain

EGY48, were gifts from Dr. Roger Brent (Molecular Sciences Institute, Berkeley,

CA). Expression vectors pSITE4NB-eIF3g, pSITE2NB-W260P6, P19 and A.

40 tumefaciens strain AGL-1 were gifts from Dr. James Schoelz (Columbia, MO).

The pENTR plasmid was purchased from Invitrogen (Invitrogen, Carlsbad CA).

The pSITE plasmids, pSITE-2NB, pSITE-4CA, and pSITE-4NB were purchased from TAIR. Yeast media was purchased from Clonetech (Palo Alto, CA).

Gateway cloning

Full-length gene VI lacking a stop codon was PCR amplified from CaMV isolate CM1841 (pCaMV10) (Gardner et al., 1981) using primers GW CaMVP6-

1F and GW CaMVP6-2CR (Supplementary Table 1). The PCR product was then cloned into pENTR TOPO-D vector (Invitrogen, Carlsbad CA) following manufacturers protocol. The P6 insert was confirmed by sequencing (DNA

Sequencing Core Facility, Ann Arbor, MI). Site-directed mutagenesis was performed on this construct, see below for details. After each construct was confirmed to contain the appropriate point mutation, it was then mobilized into pSITE2NB using Gateway technology LR Clonase II reaction (Invitrogen,

Carlsbad CA), following manufacturers protocol. This resulted in agroinfectious plant vectors capable of expressing full-length wild type or mutant P6s fused at their C-terminus to green fluorescent protein (GFP). Constructs were sequenced confirm the insert was correct and retained the point mutation.

Site-directed mutagenesis and yeast two-hybrid analyses.

Site-directed mutagenesis of D3b was performed using the QuikChange®

XL Site-Directed Mutagenesis Kit from Stratagene (La Jolla, CA). Single- stranded DNA oligonucleotide primers were designed to change each of five pre-

41 determined amino acids to alanines. These primers were synthesized by

Integrated DNA Technologies, Inc. (Coralville, IA) and are listed in

Supplementary Table 3-1. Site-directed mutagenesis reactions were performed according to the manufacturer’s specifications in an Eppendorf Mastercycler

Thermocycler. The template used was P6 domain D3 inserted in pJG4-5, described previously (Li and Leisner, 2002). After each PCR reaction, the samples were digested with DpnI included in the kit and the plasmids were introduced into E. coli by transformation, according to the manufacturer’s specifications. Transformants harboring the designated mutations were confirmed by sequencing (Plant-Microbe Genomics Facility, Columbus, OH).

The plasmid pJG4-5 harboring a full-length version of gene VI but with a deletion of D3 (FL-DN; described in (Li and Leisner, 2002) was digested with NcoI. The various amino acid substitution mutants were excised from pJG4-5 by NcoI digestion and the D3 fragments harboring the mutations were ligated into FL-DN.

This reconstitutes a full-length gene VI harboring D3b mutations in pJG4-5.

Constructions were then confirmed by sequencing.

Site-directed mutagenesis of full-length P6 cloned into pENTR was performed using the QuikChange® XL Site-Directed Mutagenesis Kit from

Stratagene (La Jolla, CA). The same primers in Supplementary Table 1 were used to generate the same point mutations within the D3b coding region of full- length gene VI. Constructs were then confirmed by sequencing.

Yeast two-hybrid analyses.

Yeast transformations were performed according to the LiOAc method

(Ausubel et al., 1993; Gyuris et al., 1993). Yeast lines containing pEG202 plasmids harboring gene VI D1, D2, D4, and Gene I as well as pJG4-5 containing full-length gene VI were described previously (Li and Leisner, 2002). Yeast lines harboring pEG202 plasmids were established first and the pJG4-5 plasmids were then introduced second. Transformants were then evaluated for their ability to grow in the absence of leucine and plated onto leucine-containing media to check the transformation efficiency. To measure the efficiency of binding, the β- galactosidase activity of four individual colonies per transformation was determined using the method of (Ausubel et al., 1993; Gyuris et al., 1993).

Enzyme activities were graphed using Excel.

Cauliflower mosaic virus mutant construction and infection.

Each of the pJG4-5 plasmids harboring the single amino acid substitution mutant P6s were digested with NcoI and the D3 fragments (395 bp) were isolated and ligated into NcoI-digested pCaMV10-NH-BB-N plasmid (Li and Leisner,

2002). The resulting plasmids were then digested with SpeI and SacI (sites are present within genes VII and VI, respectively) and inserted into those sites within pCaMV10-NcoI-DE, to generate CaMV genomes containing the gene VI single amino acid mutations within D3b of gene VI. The viral mutations were then confirmed by sequencing.

To construct a viral genome lacking the majority of D3b, a fragment of gene VI was amplified from pCaMV10 (nucleotides 6448-6709 on the CaMV sequence) with the primers KL-1F and DI-2R, the PCR product was digested with

EcoRI and XhoI and ligated into those sites within pJG4-5, to generate the plasmid pD3D5. A second fragment (nucleotides 6788-7350) from the CaMV genome was amplified using the DI-1F and FL-2R PCR primers, digested with

XhoI and inserted into the XhoI site of pD3D5 to generate the plasmid pD3D53. pD3D53 contains the complete C-terminal half coding region of gene VI from nucleotides 6448-7350 but lacking virtually all of D3b. Effectively amino acids

311-342 are replaced by a GAG codon. The pD3D53 plasmid was then digested with NcoI and the D3 (302 bp) fragment was ligated into that site of the pCaMV10-NH-BB-N plasmid. The resulting plasmids were then digested with

SpeI and SacI and inserted into those sites within pCaMV10-NcoI-DE. The deletion was confirmed by sequencing.

The wild type and mutant CaMV genomes were excised from the vector after digestion with SalI and inoculated onto plants. Turnips (Brassica rapa L. cv.

Just Right) approximately 6 week old were mechanically inoculated with 10 µg of

SalI digested virus clones per plant in a solution of 2X SSC and washed celite.

Plants were propagated as described previously (Agama, 2002a).

PCR Detection of CaMV.

To simply detect the presence of virus (e.g., the D3b deletion mutant versus wild type virus), inoculated and systemic leaf tissue was harvested at 35 days post-inoculation (DPI). Leaf tissue was ground and total DNA was extracted 44

(Li and Leisner, 2002). CaMV was detected by PCR using the GIIF and GIIR primers as described previously.

To quantify virus levels, real-time PCR was used. The inoculated leaves of both wild type and mutant virus-infected plants were harvested at 35 DPI.

Inoculated leaves for each individual plant pooled and treated as one sample.

Five infected plants were analyzed per virus. Inoculated leaf tissue was ground, treated with Proteinase K (250ug/ml), and total viral DNA was isolated using the

Qiagen DNeasy plant mini kit (Valencia, CA) as described in (Cecchini et al.,

2002).

Total viral DNA isolated from inoculated leaves was assayed by quantitative PCR using Bio-rad Real-time PCR machine. Reactions were carried out in triplicate in a total volume of 50 ml using Promega qPCR kit, according to the manufacturers specifications. Each reaction contained 5 ng of total DNA extracted from inoculated leaves and the primer concentrations were 0.2 mM.

Primers used to amplify CaMV genome were Q7-F and Q1-R derived from regions found in ORFs VII and I respectively as described by Love et al. (2005).

To normalize these data, primers 18S-F and 18S-R were used to amplify 18S ribosomal RNA gene (Love et al., 2005). The qPCR results were analyzed using the comparative threshold cycle method and normalized against 18S rRNA. The

R-program (Team, 2005) was used to remove all outliers and GraphPad Prism 5

Software was used to perform a one-way analysis of variance (ANOVA) and

Tukey HSD test to determine statistical significance of differences between the mutants and wild type. 45

Fluorescence microscopy.

Agrobacterium tumefaciens strain AGL-1 and pKYLX7 expressing either: the Tomato bushy stunt virus P19 protein; the W260 P6 fused to GFP or eIF3g fused to red fluorescent protein (RFP) all driven by the 35S promoter as well as

Nicotiana benthamiana seeds were provided by Dr. James Schoelz (University of

Missouri) (Angel et al., 2013; Chakrabarty et al., 2007; Schardl et al., 1987). All the above constructs as well as the wild type and mutant P6 expressing pSITE constructions were introduced into AGL-1 A. tumefaciens by electroporation

(Mattanovich et al., 1989). After confirming the identity of the A. tumefaciens strains, the bacteria were grown in 3 ml starter cultures containing LB with the appropriate selective antibiotic (100 mg/ml spectinomycin for pSITE constructs and 50 mg/ml kanamycin for pKYLX7 constructs) at 28°C for 24 hours. The starter culture (1 ml) was then added to 40 ml of LB with selective antibiotic for

24 hours at 28°C. The spectrophotometrically-measured optical density at 600 nm 8000g at 4°C for 10 minutes. The pellets were resuspended in 20 ml of induction medium [20 mM MES (monohydrate 2-(N-morpholino)ethanesulfonic acid) (pH 5.4), 60 mM sucrose, 55 mM glucose, and 2 mM acetosyringone dissolved in N-N-dimethyl-formamide]. The induced Agrobacteria were induced for 16 hours at 28°C. Agrobacterium strains harboring the P19 and P6-GFP vectors were mixed in a 1:1 ratio and used for agroinfiltrations. Agroinfiltrations were done on the abaxial surface of Nicotiana benthamiana leaves of 4-6 week old plants as described in (Angel et al., 2013). All experiments consisted of infiltrating 3 leaves per plant. Each experiment was repeated in triplicate.

Three days post infiltration N. benthamiana leaves were observed using

Leica SP8 Confocal Laser Scanning Microscope at the University of Toledo.

GFP and RFP were excited at wavelengths of 488 and 543nm, respectively

(Chakrabarty et al., 2007). Images were acquired sequentially line-by-line when using multiple fluorophores in concert. The Leica oil immersion 63x lens was used to acquire images at a resolution of 1024-by-1024 pixels. Images were processed using Leica software. Twelve Z-stacks were taken and analyzed per construct.

The Leica processing software was used to measure the area of inclusion bodies for 1841 and W260 P6. The R-program was used to remove all outliers and perform the unpaired t-test. ASSESS software (American Phytopathological

Society Press) was used to determine the average area of IBs formed by wild type and mutant P6s. The R-program was used to remove all outliers and preform a one-way analysis of variance (ANOVA) and TukeyHSD test. Mutant IB size was then normalized to the size of wild type IBs, which was set to 1.0 in our studies.

Protein expression analyses

To determine expression levels for wild type and mutant P6s, agroinfiltrated N. benthamiana leaves were harvested and ground in extraction buffer as described in Angel et al. (Angel et al., 2013). Extracts were centrifuged at 2000 x g for 10 minutes. Supernatants were passed through a 0.45 mm filter to remove debris and centrifuged at 13,000 x g in a microcentrifuge for 5 minutes.

The supernatants were then transferred to a mixed with 2x SDS loading dye and boiled 10 minutes. Samples were separated by electrophoresis through a 10% 47 polyacrylamide gel at 180 volts in a Mini-PROTEAN® 3 Cell (Bio-Rad

Laboratories, Inc., Hercules, CA). Proteins were then electrophoretically transferred to a nitrocellulose membrane in a Mini Trans-Blot® Transfer Cell

(Bio-Rad Laboratories, Inc.). The membrane was washed with 1 X TBS (20 mM

GFP constructions and so serves as a loading control.

The membranes were incubated with goat-anti-GFP (Santa Cruz,

Biotechnology, Santa Cruz, CA) primary antibody (1:1000 dilution) or rabbit- anti-NPTII (neomycin phosphotransferase II; Sigma-Aldrich, ST Louis, MO; 2 mg/ml). The membranes were washed three times with 1 X TBS containing

0.05% Tween-20, and incubated with secondary antibody donkey-anti-goat

(Sigma-Aldrich, ST Louis, MO) at 1:5000 dilution or goat-anti-rabbit (Santa

Scientific, Inc.) were developed using a Konica SRX-101® developer.

3.4 Results

Identification of a short conserved amino acid segment within P6 and its requirement for infectivity.

Previous work indicated that domain D3 is essential for P6 self-association

(Li and Leisner, 2002), so this region was examined in greater detail. D3 has a tripartite structure with a central portion (termed D3b) located between N- terminal (termed, D3a) and C-terminal (termed, D3c) RNA-binding domains (De

Tapia et al., 1993) (Fig. 3-1). Multiple sequence alignment of the gene VI nucleotide sequences encoding D3 for 20 CaMV isolates showed that D3b was less variable (0.200 variable sites/nucleotide) than either the D3a or D3c regions

(0.359 and 0.361 variable sites/nucleotide, respectively). This bias was even more dramatic when the amino acid sequences were compared. The D3a and D3c portions of D3 contained 0.311 and 0.297 variable amino acid positions per amino acid, while the value for the central portion was 0.114. This suggests that D3b

(P6 amino acids 309-343) was more well-conserved than the two RNA binding domains.

Since regions on either side of D3b are known to participate in other interactions (De Tapia et al., 1993; Park et al., 2001)and we were more concerned with P6 self-association, our attention was drawn to D3b. In addition to being conserved, the D3b region is quite polar (21/35 amino acids are polar) and charged (4/35 amino acids are acidic, while 10/35 are basic). Interestingly, computer prediction of secondary structure for D3b suggests that this region is

49 mainly a-helical. In fact, D3b is predicted to be two helices separated by a four amino acid turn. The N-terminal helix appears to be amphipathic, while the C- terminal helix does not.

Interactions of D3b with full length P6 and its self-association domains.

Deletion of the D3 coding region within gene VI in CaMV resulted in a non-infectious virus (Li and Leisner, 2002). A similar result was observed with viruses lacking D3b (Fig. 3-2A). Not only was the virus undetectable in upper non-inoculated leaves of turnips infected with the D3b mutant at 65 days post- inoculation (DPI) (0 out of 10 plants inoculated), but virus was not found by PCR in the inoculated leaves by this time. Of course, viral DNA was easily detectable in both the inoculated and upper non-inoculated leaves of plants infected with wild type virus.

Figure 3-1. Schematic diagram of P6 and location of mutations. The 520 amino acid P6 protein is indicated. Hatched areas, P6 regions involved in self- association) (Haas et al., 2005; Li and Leisner, 2002): D1 (amino acids 1-110); D2 (amino acids 156-253); D4 (amino acids 414-520); shaded area, D3 (amino acids 249-379) D3a and D3c are indicated (note both contain RNA-binding domains); black area, D3b examined in this study. Numbers in italics: upper, number of variable nucleotide positions within that portion of gene VI per nucleotide; lower, number of variable amino acid positions within that portion of P6 per amino acid. The amino acid sequence of the D3b region is shown below the P6 cartoon as well as the amino acid changes for the various mutants, single letter amino acid designations are given. The D3b region is mainly a-helical (black) with an intervening turn (gray) below the sequences as predicted by Garnier-Robson model in the Protean software contained within the Lasergene Software package. Helical wheels for both helices (predicted by the Protean software package) are indicated below the secondary structure prediction; black, non-polar amino acids; dark gray, uncharged polar amino acids; light gray, acidic amino acids; white, basic amino acids; amino acid numbers are given; bold italic numbers, amino acids mutated.

D3 was previously shown to interact with D1, D4, and possibly inefficiently with D2, but not with itself (Li and Leisner, 2002). Therefore the ability of D3b to interact with P6 and these three interaction regions was investigated. Yeast two-hybrid data indicated that D3b interacts efficiently with full-length P6 (Fig. 3-2B-E). Interestingly, D3b interacted with the D1 and D4 P6 self-association domains. D3b interactions with D2 were inconsistent. Therefore, this interaction was not pursued further. However, D3b alone interacted with D1 and D4 to induce lower levels of β-galactosidase activity than when binding to the full-length protein. However, all (except negative control) showed leucine- independent growth, indicative of an interaction.

Binding of D3b mutants to P6 interaction domains.

To examine the role of D3b in P6 interactions in more detail, single amino acid substitutions were generated in this portion of gene VI in the context of the full-length protein. Each of these point mutations changed either charged (E, K, or R; single letter amino acid abbreviations) or a non-polar (L, V) amino acid side chains to alanines (Fig. 3-1). None of these mutants, when co-transformed with empty pEG202, permitted either leucine-independent growth, or gave β- galactosidase activity (for colonies grown on plates containing leucine) above control levels (Data not shown).

Figure 3-2. Role of the D3b region in CaMV infectivity and protein binding. A, Gel electrophoresis of PCR products from virus-infected plants. L indicates 100 bp ladder. I represents inoculated leaves; U, upper non-inoculated leaves for representative plants inoculated with either the D3b deletion mutant (Mutant) or with wild type (Wild Type) virus. DNA indicates the PCR performed with pCaMV10 DNA; M, procedure performed on mock-inoculated plant; arrow indicates position of CaMV PCR amplification product. B, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity (for C, D, and E). Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 ; numbers to the left of each pair of constructions correspond to the plates in C and the β-galactosidase assays shown in E. C, Growth of yeast transformants on media with (+L) and without (-L) leucine. D, key for the plates in C. E, β-galactosidase activity of yeast transformants expressing constructs as represented in B. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in B and plates in C.

Based on leucine-independent growth, the D1 P6 self-association domain interacted with all five of the D3b amino acid substitution mutants (Fig. 3-3). In general, all of the D3b mutant proteins interacted with D1. However, β-

53 galactosidase activity generated from interaction of D1 with each D3b mutant was lower than interaction of D1 with the wild type protein. The two mutations near the N-terminus of D3b (E312A and L316A) showed a much more drastic decrease in β-galactosidase activity than the other three mutants (K318A, V324A and

R328A).

Figure 3-3. Interactions of D3b mutant P6s with P6 self-association domain, D1. A, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity (for B, C, and D). Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 ; numbers to the left of each pair of constructions correspond to the plates in C and the β-galactosidase assays shown in B. B, β-galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A and plates in C. C, Growth of yeast transformants on media with (+L) and without (-L) leucine. D, key for the plates in C.

The β-galactosidase activity generated from D4 interacting with the P6 mutants somewhat resembled the results with D1 (Fig. 3-4). Again, mutations at the N-terminal end of D3b (E312A and L316A) exhibited a strong reduction on

D4 binding with the mutant P6s based on β-galactosidase activity. Interestingly, the β-galactosidase generated from the R328A mutation (at the other end of the tested region) in combination with D4 was also low and similar to that of the

E312A mutant. The two remaining, more centrally-located showed much greater

β-galactosidase activity than the other mutants when interacting with D4. When interacting with D4, the first (K318A) exhibited essentially wild type level activity, while the second (V324A), showed substantial activity but was reduced approximately nine-fold.

Figure 3-4. Interactions of D3b mutant P6s with P6 self-association domain, D4. A, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity (for B, C, and D). Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 ; numbers to the left of each pair of constructions correspond to the plates in C and the β-galactosidase assays shown in B. B, β-galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A and plates in C. C, Growth of yeast transformants on media with (+L) and without (-L) leucine. D, key for the plates in C.

Analysis of fluorescent-labeled IBs in plant cells.

P6 from CaMV isolate W260 tagged at the C-terminal end with GFP, is able to form fluorescent-IBs (Harries et al., 2009). To determine if P6 from

CaMV isolate CM1841, the virus we have been working with, forms similar IBs,

CM1841 P6 was tagged at the C-terminal end with GFP. Like its W260 56 counterpart, leaves agroinfiltrated with CM1841 P6 showed a punctate distribution of GFP fluorescence whereas GFP alone exhibited a diffuse cytoplasmic pattern (Fig. 3-5). GFP alone was agroinfiltrated to confirm the punctate distribution shown with CM1841 P6 was IB formation not GFP aggregation due to over expression. IBs are associated with the host translational machinery and W260 P6-GFP was found to co-localize with the translation factor eIF3g in agroinfiltrated cells (Angel et al., 2013). Therefore co-localization studies were conducted between translation factor eIF3g and CM1841 P6-GFP.

The co-localization pattern of eIF3g with CM1841 P6-GFP was similar to that observed with eIF3g-W260 P6-GFP. This strongly suggests that CM1841 P6-

GFP puncta are authentic IBs. During the course of these studies, a difference in

IB size between both CaMV strains (CM1841 and W260) was noticed.

Interestingly, statistical analysis showed CM1841 putative IBs were significantly smaller than those formed by their W260 counterpart.

Figure 3- 5. Fluorescent inclusion bodies (IBs) formed by CM1841 P6 in Nicotiana benthamiana. Fluorescent constructions were agroinfiltrated into N. benthamiana leaves. Three days post-infiltration, the leaves were examined by fluorescence microscopy. Magnification bar for A and C is 10 mm. A. Distribution of green fluorescent protein (GFP) alone (top three panels); W260 P6 fused to GFP (center three panels) and CM1841 P6 fused to GFP (bottom three panels). Left panels show GFP fluorescence alone (GFP); center panels, show bright field (BF) images of the same panels; and right, bright field overlay of GFP fluorescence (MERGE). B. Average size of W260-GFP and CM1841-GFP IBs. Z-stacks obtained by fluorescence microscopy were delineated and analyzed for size differences. Error bars indicate standard error of the mean and the star indicates a statistically-significant difference (p>0.0001). C. Co-localization of P6 and eIF3g. Left panels; P6-GFP fluorescence, middle panels; eIF3g-RFP, and right panels; overlay of GFP and RFP fluorescence. Top 3 panels, with W260 P6- GFP; bottom three panels, with CM1841 P6.

Since CM1841 P6 was capable of forming IBs, the D3b mutants were then examined to determine if the mutations influenced IB size or other characteristics.

The D3b mutant P6s were tagged at the C-terminal end with GFP. As previously shown with wild type CM1841 P6, leaves agroinfiltrated with all of the D3b mutants showed punctate distribution of GFP fluorescence, indicative of IB formation (Fig. 3-6). However, IB size varied between wild-type P6 and the D3b mutants. In addition, distinct differences in IB size were observed among the D3b mutants. The D3b mutants can be divided into two classes based on IB size when compared to wild-type. The mutants K318A and V324A showed IBs similar in size to those induced by wild-type P6. In contrast, the E312A, L316A, and the

R328A mutants formed the second class, and showed significantly smaller IBs when compared to wild-type. Western blot analysis showed that all five mutants were expressed in N. benthamiana leaves.

Effects of D3b mutations on CaMV infectivity.

The D3b mutations described above were incorporated into the CM1841 genome. The virus genomes were then inoculated to turnips at ten plants per virus and the infections were then permitted to progress. Inoculated leaves were harvested at 36 DPI and total viral DNA levels were determined by Real-time

PCR. Based on inoculated leaf levels, the D3b mutants could be classified into two categories: those that were not significantly different from wild type (K318A and V324A) and those that were different from wild type (E312A, L316A and

R328A). The latter mutants showed dramatically reduced levels of inoculated leaf viral DNA (Fig. 3-7A). 59

A GFP BF MERGE B

1.2! a"

WT WT a"

1! a,b"

0.8! b,c"

E312A E312A 0.6! c" c" 0.4! Relative IB Size IB Size Relative

0.2! L316A L316A

0! WT! E312A! L316A! K318A! V324A! R328A! TukeyHSD:"P"value"<"0.015"

K318A K318A C

V324A V324A GFP:P6

NTPII R328A R328A

Figure 3-6. Fluorescence microscopy of mutant P6 IBs. Fluorescent constructions were agroinfiltrated into N. benthamiana leaves. Three days post- infiltration, the leaves were examined by fluorescence microscopy. Magnification bar for A and C is 10 mm. A. Distribution of fluorescence for wild type CM1841 P6 and the D3b mutants. Left panels, GFP fluorescence (GFP); middle panels, bright field (BF) images of the same panels; and right, bright field overlay of left panels. Magnification bar is 10 mm. B. Average size of IBs formed by P6 mutants relative to those formed by wild type is given. This was determined from the z-stacks obtained by fluorescence microscopy. Average values normalized to the size of wild type IBs, that was set to 1 and standard error of the mean are given. Different letters indicate statistically significant differences (P> 0.015). C. Levels of wild type and mutant P6s expressed by agroinfiltration in N. benthamiana leaves. Upper panel, western blot of wild type and D3b mutant P6 protein using anti-GFP antibodies. The neomycin phosphotransferase type II gene was encoded on the same T-DNA as the gene VIs and NTPII served as a loadingcontrol.

Plants inoculated with the wild type and all mutant viruses generally showed systemic symptoms by 30-36 days post-inoculation (dpi). All of the plants inoculated with wild type virus exhibited vein clearing and vein banding 60 symptoms characteristic of CaMV infection. By 36 dpi, a large number of leaves per plant showed these symptoms (Fig. 3-7B). Plants inoculated with the K318A mutant also showed vein clearing and vein banding symptoms indistinguishable from wild type, with a similar large number of symptomatic leaves. Likewise the

V324A mutant showed symptoms identical to wild type on systemically-infected leaves. However, the number of systemically-infected leaves per plant was slightly less than wild type.

A! 4

a a 3

2 a

1 Normalized Viral DNA Levels DNA NormalizedViral b b b

WT E312A L316A K318A V324A R328A Tukey HSD pValue <0.001

B 10! 9! a a,b 8! b 7! 6! 5! 4! c c 3! 2! d 1! 0! Number Of SystemicNumber Of Leaves Per Plant WT! E312A! L316A! K318A! V324A! R328A! Tukey HSD pValue <0.03

Figure 3-7. Propagation of D3b mutant viruses in turnips. A, propagation of virus in inoculated leaves as determined by real-time PCR. Inoculated leaves were harvested, DNA was isolated and real time PCR was performed, normalizing virus DNA levels against levels of 18S rDNA genes. Average values and standard error of the mean are given. Different letters indicate statistically- significant differences, while same letters indicate no difference. B, Average number of leaves exhibiting systemic symptoms per plant. A total of ten plants were inoculated with each virus and the average number of leaves showing systemic symptoms per plant was determined at 35 days post-inoculation. Both average and standard error of the mean are indicated; different letters indicate statistically-significant differences; same letters, no difference.

The more dramatic differences were exhibited by the E312A, L316A and

R328A mutants. For the most part, the systemic symptoms displayed by plants infected with these mutants were mainly chlorotic lesions on systemic leaves.

The L316A mutant appeared to be the most restricted, exhibiting only a few chlorotic lesions on some systemically-infected leaves. Turnips infected with the

E312A and R328A mutants showed two types of symptoms. For E312A, about half of the plants showed exclusively chlorotic lesions on systemically-infected leaves. The other half showed vein clearing and vein banding symptoms typical of wild type. In both cases, the number of leaves showing systemic symptoms was the same but was significantly reduced compared to plants infected with wild type virus. For R328A, about three-fourths of the plants showed chlorotic lesions on systemically-infected leaves and approximately one-fourth showed vein- clearing and vein banding symptoms characteristic of wild type. Again, the number of leaves showing either type of systemic symptom was reduced compared to wild type. Interestingly, the viruses from plants inoculated with the

E312A, L316A, and R328A mutants that showed systemic chlorotic lesions exhibited these symptoms on older systemically-infected leaves, but not young leaves. We analyzed the R328A mutant and discovered that the young leaves on these plants did contain virus, but their levels were much lower than in the older systemically-infected leaves that were symptomatic. Interestingly, as the asymptomatic young leaves aged, they eventually exhibited chlorotic lesions.

The presence of wild-type symptoms for the E312A and R328A mutants could be due to those viruses reverting within these plants. To investigate this, systemically-infected leaves were harvested for all of the mutant viruses, the gene

VI region was PCR-amplified and sequenced. The sequence data showed that all of the plants inoculated with the wild type virus harbored wild type sequence throughout the entire D3b region. The viral genomes amplified from infected with the K318A and V324A mutants all retained their corresponding mutations.

The viral genomes from the L316A mutants also all retained their mutations as well. For the E312A and R328A mutants, the viral genomes of plants exhibiting the chlorotic lesion phenotype showed that those viruses retained their respective mutations. However, in plants infected with the E312A and R328A mutant viruses that showed wild type symptoms, the viral genomes harbored a reversion of their respective mutations back to the wild type sequence.

3.5 Discussion

The Cauliflower mosaic virus P6 protein is an excellent example of a multifunctional protein since it plays a role in many activities required for viral infection (Bonneville et al., 1989; Haas et al., 2008; Kobayashi and Hohn, 2003;

Love et al., 2007; Schoelz et al., 1986; Schoelz and Wintermantel, 1993). Since many viral processes occur in inclusion bodies (Hull, 2002; Knipe, 1990a;

Martelli and Russo, 1977), it is likely that the integrity of these structures is important for infectivity. Indeed the stability of inclusion bodies does appear to influence the CaMV life cycle (Anderson et al., 1992b). Hence, P6 self- association may be a key determinant for some of the functions this protein performs. Previously, four portions of P6 (D1-D4) were found to play a role in self-association (Haas et al., 2005; Li and Leisner, 2002). Here we investigated a region termed D3 (amino acids 249-379) involved in P6 self-association.

D3 has been implicated in a variety of activities and is essential for viral infectivity (Li and Leisner, 2002). D3 contains the two P6 non-sequence specific

RNA binding domains, separated by a spacer region. The upstream RNA-binding

64 domain (amino acids 249-309) also plays a role in interactions with eIF3g and large ribosomal subunit protein 24 (Park et al., 2001). The downstream RNA- binding domain (343-379) has also been implicated in interactions with the P4

CaMV capsid protein (Ryabova et al., 2002). Interestingly, the spacer region

(which we have termed D3b) is the most highly conserved portion of D3.

Therefore, this region of unknown function was investigated for its potential role in P6 self-association and viral infection.

D3b interacted with full-length P6. The inefficient interaction of D3b with the individual domains may suggest that these domains require a particular structural context for appropriate interactions with D3b. These data indicate that

D3b likely plays role in P6 self-association functions identified for D3 (Li and

Leisner, 2002). Underscoring the importance of D3b was the observation that

CaMVs lacking this region were non-infectious. This also agrees with the data by

Kobayashi and Hohn (Kobayashi and Hohn, 2003) who showed that a larger deletion of P6, spanning all of D3b and beyond (amino acids 306-366), was non- infectious.

Because deletion mutations can have drastic effects on protein structure, single amino acid substitutions were generated within D3b that were not predicted to change protein secondary structure. These mutants were then evaluated for their effects on binding to the various P6 self-association domains, IB formation, and viral infection. The two most N-terminal mutations had a detrimental effect on binding to both D1 and dramatically reduced association with D4. The C- terminal-most mutation showed lower than wild-type activity in binding to D1 65 and D4 as well. The two central mutations appeared less severe in their effects on binding to D4 and to some extent D1. Taken together, these data suggest that mutations in D3b influence binding of P6 to domains involved in self-association and likely influence self-interaction.

If these mutations affect P6 self-association, they may also affect the formation of IBs. Therefore, the ability of CM1841 P6 to form IBs was examined by fluorescence microscopy. The majority of our prior work was performed with

CaMV isolate CM1841. Therefore, we performed this study using the same viral isolate to provide internal consistency among experiments. The ability of

CM1841 P6 to form fluorescent IBs was compared with that of W260 P6, as a reference. Just like W260 P6, the CM1841 counterpart did result in fluorescent puncta within agroinfiltrated cells (Angel et al., 2013; Harries et al., 2009). In addition, both the W260 and CM1841 P6s were also able to co-localize with eIF3g, indicative of IBs. To determine if the D3b mutants affected IB formation, fluorescence microscopy was performed. All of the mutants generated punctate distributions of GFP indicative of IBs. These studies indicated that the mutants fell into two classes. Some mutants (K318A and V324A) generated fluorescent

IBs that were not significantly different (on average) in size compared to those produced by wild type. Other mutants (E312A, L316A, and R328A) induced formation of IBs that were significantly smaller in size than wild type.

Some of these mutations also influenced virus propagation and systemic infection. The virus levels for the K318A and V324A mutants in inoculated leaves were not significantly different from wild type. However, the levels of 66 virus for the E312A, L316A, and R328A mutants were significantly reduced compared to wild type. This is interesting since a region of P6 spanning all of

D3b along with additional downstream sequences was proposed to be essential in basic viral replication (Kobayashi and Hohn, 2003). The inoculated leaf propagation data were also reflected in the formation of systemic symptoms. The central mutations (K318A and V324A) appeared essentially wild type in their ability to elicit viral systemic symptoms. However, the N- (E312A and L316A) and C-terminal (R328A) mutations were debilitated in their ability to induce systemic symptoms. However, in some plants, the E312A and R328A mutations reverted.

Taking these data together, we conclude that the mutations that most strongly affect D3b binding to the P6 self-association domains have the strongest reduction in IB size and viral inoculated leaf DNA levels. This translates into a reduced ability to induce a systemic infection. One possible explanation for all of these data revolves around the ability of P6 to form IBs. The formation of P6 IBs begins with the generation of small clusters of ribosomes associated with electron-dense material, but containing few or no virus particles (Martelli and

Russo, 1977). Those clusters are thought to then accrete to form small and then larger IBs that do contain virus particles. Interestingly, larger IBs appear to have different properties from smaller ones. Smaller P6 IBs, termed, “small bodies”

(Champagne et al., 2004) contain unprocessed (N-terminal-containing) P4 coat protein, while the larger IBs do not. The N-terminal 76 amino acids of P4 are removed by the protease located at the N-terminal end of P5. This protease must

67 be excised from P5, in order for the rest of the protein, reverse transcriptase/RNase H enzyme, to be activated (Hohn and Futterer, 1997;

Takatsuji et al., 1992; Torruella et al., 1989). If the N-terminus of P4 is present in

“small bodies”, then it is likely that the protease is not very active in those structures and probably has not been processed from P5 efficiently. Hence,

“small bodies” would be expected to have less active reverse transcriptase/RNase

H activity compared with large IBs. All of the D3b mutants analyzed in this study likely have the ability to form “small bodies” because a punctate distribution of fluorescent-P6 is observed. However, the mutants (E312A, L316A, and R328A) most strongly affected in binding to the P6 self-association domains also form the smallest IBs in agroinfiltrated leaves. This suggests that the E312A, L316A, and

R328A may be “trapped” in the “small body” stage and are unable to efficiently fuse to form larger IBs. Hence, these three mutants should have the lowest replication in inoculated leaves, which is exactly what is observed. Because of the lower levels of virus in inoculated leaves, the efficiency of systemic infection would also be compromised in the small body trapped mutants. These data suggest that fusion of “small bodies” into larger IBs may require P6 interactions that involve D3b. These data also suggest that for CM1841 at least, the ability to form proper IBs is important for the infection process. Hence, IB formation by

CM1841 is apparently required for viral infection rather than being merely a consequence of it.

In summary, we find that mutations within a self-association domain of

CaMV P6 protein that affect binding to the other self-association domains result

68 in smaller IBs, reduced propagation in inoculated leaves and a more restricted systemic infection.

3.6 Acknowledgments

The authors thank Drs. Richard Komuniecki and Song-Tao Liu

(University of Toledo, Toledo, OH, USA) for the vectors used in this study as well as Dr. Roger Brent (Molecular Sciences Institute, Berkeley, CA) for plasmids pEG202 and pJG4-5, along with yeast strain EGY48 harboring pSH18-

34. The authors also thank Dr. Wendy Zellner of the USDA-ARS for her help with statistical analyses. This work was supported in part by NIH Grant number

1R15AI50641-01 and USDA-ARS Specific Cooperative Agreement: 58-3607-1-

193.

3.7 Reference

Agama, K., Beach, S., Schoelz, J., Leisner, S.M., 2002. the 5' third of Cauliflower

mosaic virus gene VI conditions resistance breakage in Arabidopsis

ecotype Tsu-0. Phytopathology 92, 190-196.

Anderson, E.J., Trese, A.T., Sehgal, O.P., Schoelz, J.E., 1992. Characterization of

a chimeric Cauliflower mosaic virus isolate that is more severe and

accumulates to higher concentrations the either of the strains from which it

was derived. Mol. Plant-Microbe Interact. 5, 48-54.

Angel, C.A., Lutz, L., Yang, X., Rodriguez, A., Adair, A., Zhang, Y., Leisner,

S.M., Nelson, R.S., Schoelz, J.E., 2013. The P6 protein of Cauliflower

mosaic virus interacts with CHUP1, a plant protein which moves

chloroplasts on actin microfilaments. Virology 443, 363-374.

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith,

J.A., Struhl, K., 1993. Current Protocols in Molecular Biology. Greene

Publishing Associates and John Wiley & Sons, INC, Cambridge MA

USA.

Bonneville, J.M., Sanfacon, H., Futterer, J., Hohn, T., 1989. Posttranscriptional

trans-activation in Cauliflower mosaic virus. Cell 59, 1135-1143.

Bureau, M., Leh, V., Haas, M., Geldreich, A., Ryabova, L., Yot, P., Keller, M.,

2004. P6 protein of Cauliflower mosaic virus, a translation reinitiator,

interacts with ribosomal protein L13 from Arabidopsis thaliana. J. Gen.

Virol. 85, 3765-3775.

Cecchini, E., Geri, C., Love, A.J., Coupland, G., Covey, S.N., Milner, J.L., 2002.

Mutations that delay flowering in Arabidopsis de-couple symptom

response from Cauliflower mosaic virus accumulation during infection.

Mol. Plant Pathol. 3, 81-90.

Chakrabarty, R., Banerjee, R., Chung, S.M., Farman, M., Citovsky, V.,

Hogenhout, S.A., Tzfira, T., Goodin, M., 2007. pSITE vectors for stable

integration or transient expression of autofluorescent protein fusions in

plants: probing Nicotiana benthamiana-virus interactions. Mol. Plant-

Microbe Interact. 20, 740-750.

Champagne, J., Benhamou, N., Leclerc, D., 2004. Localization of the N-terminal

domain of Cauliflower mosaic virus coat protein precursor. Virology 324,

257-262.

Covey, S., Hull, R., 1981. Transcription of Cauliflower mosaic virus DNA.

Detection of transcripts, properties, and location of the gene encoding the

virus inclusion body protein. Virology 111, 463-474.

De Tapia, M., Himmelbach, A., Hohn, T., 1993. Molecular dissection of the

Cauliflower mosaic virus translation transactivator. EMBO J. 12, 3305-

3314.

Gardner, R.C., Howarth, A.J., Hahn, P., Brown-Luedi, M., Shepherd, R.J.,

Messing, J., 1981. The complete nucleotide sequence of an infectious

clone of Cauliflower mosaic virus by M13mp7 shotgun sequencing.

Nucleic Acids Res. 9, 2871-2888.

Gyuris, J., Golemis, E., Chertkov, H., Brent, R., 1993. Cdi1, a human G1 and S

phase protein phosphatase that associates with Cdk2. Cell 75, 791-803.

Haas, G., Azevedo, J., Moissiard, G., Geldreich, A., Himber, C., Bureau, M.,

Fukuhara, T., Keller, M., Voinnet, O., 2008. Nuclear import of CaMV P6

is required for infection, and suppression of the RNA silencing factor

DRB4. EMBO, J. 27, 2102-2112.

Haas, M., Bureau, M., Geldreich, A., Yot, P., Keller, M., 2002. Cauliflower

mosaic virus: still in the news. Mol. Plant Pathol. 3, 419-429.

Haas, M., Geldreich, A., Bureau, M., Dupuis, L., Leh, V., Vetter, G., Kobayashi,

K., Hohn, T., Ryabova, L., Yot, P., Keller, M., 2005. The open reading

frame VI product of Cauliflower mosaic virus is a nucleocytoplasmic

protein: its N terminus mediates its nuclear export and formation of

electron-dense viroplasms. Plant Cell 17, 927-943.

Hapiak, M., Li, Y., Agama, K., Swade, S., Okenka, G., Falk, J., Khandekar, S.,

Raikhy, G., Anderson, A., J., P., Zellner, W., Schoelz, J.E., Leisner, S.M.,

2008. Cauliflower mosaic virus gene VI product N-terminus contains

regions involved in resistance-breakage, self-association, and interactions

with movement protein. Virus Res. 138, 119-129. 72

Harries, P.A., Palanichelvam, K., Yu, W., Schoelz, J.E., Nelson, R.S., 2009. The

Cauliflower mosaic virtus protein 6 forms motile inclusions that traffic

along actin microfilaments and stabilize microtubules. Plant Physiol. 149,

1005-1016.

Himmelbach, A., Chapdelaine, Y., Hohn, T., 1996. Interaction between

Cauliflower mosaic virus inclusion body protein and capsid protein:

Implications for viral assembly. Virology 217, 147-157.

Hohn, T., Futterer, J., 1997. The proteins and functions of plant pararetroviruses:

knowns and unknowns. Cr. Rev. Plant Sci. 16, 133-161.

Hull, R., 2002. Matthews' Plant Virology, fourth ed. Academic Press, San Diego.

Knipe, D., 1990. Virus-host interactions, in: Fields, B.N., Knipe, D.M. (Eds.),

Fields Virology. Raven Press, New York, pp. 293-316.

Kobayashi, K., Hohn, T., 2003. Dissection of Cauliflower mosaic virus

transactivator/viroplasmin reveals distinct essential functions in basic

virus replication. J. Virol. 77, 8577-8583.

Leh, V., Yot, P., Keller, M., 2000. The Cauliflower mosaic virus translational

transactivator interacts with the 60S ribosomal subunit protein L18 of

Arabidopsis thaliana. Virology 266, 1-7.

Li, Y., Leisner, S.M., 2002. Multiple domains within the Cauliflower mosaic

virus gene VI product interact with the full-length protein. Mol. Plant-

Microbe Interact. 15, 1050-1057.

Love, A.J., Laird, J., Holt, J., Hamilton, A.J., Sadanandom, A., Milner, J.J., 2007.

Cauliflower mosaic virus protein P6 is a suppressor of RNA silencing. J.

Gen. Virol. 88, 3439-3444.

Love, A.J., Yun, B.W., Laval, V., Loake, G.J., Milner, J.J., 2005. Cauliflower

mosaic virus, a compatible pathogen of Arabidopsis, engages in three

distinct defense-signaling pathways and activates rapid systemic

generation of reactive oxygen species. Plant Physiol. 139, 935-948.

Martelli, G.P., Russo, M., 1977. Plant Virus Inclusion Bodies. Adv. Virus Res.

21, 175-266.

Mattanovich, D., Rucker, F., da Camara Machado, A., Laimer, M., Regner, F.,

Steinkellner, H., Himmler, G., Katinger, H., 1989. Efficient transformation

of Agrobacterium spp. by electroporation. Nucleic Acids Res. 17, 6747.

Mazzolini, L., Dabos, P., Constantin, S., Yot, P., 1989. Further evidence that

viroplasms are the site of Cauliflower mosaic virus genome replication by

reverse transcription during viral infection. J. Gen. Virol. 70, 3439-3449.

Park, H.-S., Himmelbach, A., Browning, K.S., Hohn, T., Ryabova, L.A., 2001. A

plant viral reinitiation factor interacts with the host translational

machinery. Cell 106, 723-733.

Ryabova, L.A., Pooggin, M.M., Hohn, T., 2002. Viral strategies of translation

initiation, ribosomal shunt and reinitiation. Prog. Nucleic Acid Res. 72, 1-

39.

Schardl, C.L., Byrd, A.D., Benzion, C., Altschuler, M.A., Hildebrand, D.F., Hunt,

A.G., 1987. Design and construction of a versatile system for the

expression of foreign genes in plants. Gene 61, 1-11.

Schoelz, J., Shepherd, R.J., Daubert, S., 1986. Region VI of Cauliflower mosaic

virus encodes a host range determinant. Mol. Cell. Biol. 6, 2632-2637.

Schoelz, J.E., Wintermantel, W.M., 1993. Expansion of viral host range through

complementation and recombination in transgenic plants. Plant Cell 5,

1669-1679.

Takatsuji, H., Yamauchi, H., Watanabe, S.-I., Kato, H., Ikeda, J.-E., 1992.

Cauliflower mosaic virus reverse transcriptase. J. Biol. Chem. 267, 11579-

11585.

Team, R.D.C., 2005. R: A language and environment for statistical computing,

reference 2† index version 2.2.1. . R Foundation for Statistical Computing,

Vienna, Austria. .

Torruella, M., Gordon, K., Hohn, T., 1989. Cauliflower mosaic virus produces an

aspartic proteinase to cleave its polyproteins. EMBO J. 8, 2819-2825.

3-8. Supplementary Data

Supplementary Table 3-1. Primers used in the D3b study.

A. Primers for generating P6-GFP constructions. GW CaMVP6-1F: CACCATGGAGAACATAGAAAAACTCCT GW CaMVP6-2CR ATCCACTTGCTTTGAAGACGTGGT

B. Primers used for site-directed mutagenesis. E312A-1F: AATAATCTCCAGGCGATCAAATACCTTCCCAAGAAGGT E312A-2R: ACCTTCTTGGGAAGGTATTTGATCGCCTGGAGATTATT

L316-1F: AATCTCCAGGAGATCAAATACGCTCCCAAGAAGGTTAAAGATGCA L316-2R: TGCATCTTTAACCTTCTTGGGAGCGTATTTGATCTCCTGGAGATT

K318-1F: CAGGAGATCAAATACCTTCCCGCGAAGGTTAAAGATGCAGTC K318-2R: GACTGCATCTTTAACCTTCGCGGGAAGGTATTTGATCTCCTG

V324A-1F: CCCAAGAAGGTTAAAGATGCAGCCAAAAGATTCAGGACTAACTGC V324-2R: GCAGTTAGTCCTGAATCTTTTGGCTGCATCTTTAACCTTCTTGGG

R328A-1R: GATGCAGTCAAAAGATTCGCGACTAACTGCATCAAGAACACAG R328-2R: CTGTGTTCTTGATGCAGTTAGTCGCGAATCTTTTGACTGCATC

Five pairs of primers were designed to convert the targeted amino acid (in primer name) to alanine. Primers with suffix -1F are forward primers, while those ending in -2R are reverse primers; the mutated nucleotides are underlined.

C. Primers used for making the D3b deletion in the CaMV genome. KL-1F (6448) CACGCTAGAATTCAAAAGGC DI-2R: (6709) CTCCTCGAGATTATTACTCG DI-1F: (6788) ATATTTCTCGAGATCAGAAG FL-2R: (7350) TGCGTCATCCCTCGAGTCAG

D. Primers used for real-time PCR. Q7-F: AGCGGTCAAAATATTGCTTA Q1-R: AACTTACCGTATGVTAGATTACCT

18S-F: CGTGATCGATGAATGCTACC 18S-R: GGGGTTTGTTGCACGTATTA

Chapter 4

Cauliflower mosaic virus Major Inclusion Body Protein Interacts with the Aphid Transmission Factor, the Virion-Associated Protein, and Gene VII Product

4.1 Abstract

The Cauliflower mosaic virus (CaMV) gene VI product (P6) is a multifunctional protein essential for viral infection. In order to perform its various tasks, P6 interacts with both viral and host factors, as well as forming electron-dense cytoplasmic inclusion bodies. Here we investigate the interactions of P6 with three CaMV proteins: P2 (aphid transmission factor), P3 (virion- associated protein), and P7 (protein of unknown function). Based on yeast two- hybrid and maltose-binding protein pull-down experiments, P6 interacted with all three of these CaMV proteins. P2 helps to stabilize P6 inclusion bodies.

Although the P2s from two CaMV isolates (W260 and CM1841) differ in the ability to stabilize inclusion bodies, both interacted similarly with P6. This suggests that inclusion body stability may not be dependent on the efficiency of

P2-P6 interaction. However, neither P2 nor P3 interacted with P7 in yeast two- hybrid assays.

4.2 Introduction

Cauliflower mosaic virus (CaMV) is a plant pararetrovirus that is one of the top ten viruses in molecular plant pathology (Scholthof et al., 2011). The eight kbp genome encodes seven potential polypeptides, designated P1-P7 (Haas et al., 2002; Hull, 2002). P7 is the first protein encoded by the 35S RNA, but the function of this molecule is unknown (Dixon et al., 1986). P1 serves as the cell- to-cell movement protein while P2 serves as a factor permitting CaMV to be transmitted by aphids (Armour et al., 1983; Thomas et al., 1993). However neither P1 nor P2 can bind to virus particles directly, but require interactions with another protein, P3 (also called virion-associated protein or VAP), that serves as a bridge linking P1 or P2 to virions (Leh et al., 2001; Stavolone et al., 2005). Thus,

P3 binds to P4, the virus capsid protein. CaMV is a DNA virus that replicates through an RNA intermediate (Haas et al., 2002; Hull, 2002) and accomplishes this through a POL-like protein, similar to retroviruses, which is the function of

P5. P6 is the major protein making up electron-dense cytoplasmic inclusion bodies (IBs) (Covey and Hull 1981) and it is responsible for many other functions as well. Interestingly, P2 influences the stability of P6 IBs (Anderson et al.,

1992).

P6 IBs are thought to be the sites of virus genome replication, protein synthesis, and virus particle assembly/accumulation (Haas et al., 2002). Since P6 is the most abundant protein present within IBs and most viral functions occur 78 within those structures, it is perhaps not surprising that the protein is multifunctional. P6 has been implicated in virtually every aspect of viral infection. P6 aids viral genome expression by permitting translation of all of the viral proteins from the 35S RNA through a process called translational transactivation (De Tapia et al., 1993). P6 appears to be involved in virus genome replication, movement and virus particle assembly (Kobayashi and Hohn, 2003;

Schoelz et al., 1991; Himmelbach et al., 1996). The consequences of these processes are that P6 influences CaMV symptoms on virus-infected plants. P6 affects interactions with the host including influencing host range, and can impair plant defenses by inhibiting gene silencing (Schoelz, 1988; Love et al, 2007).

4.3 Material and Methods, Results, and Discussion

Since P6 may act as a molecular chaperone (Himmelbach et al., 1996) in virus particle assembly, we investigated whether P6 also interacted with P3, that is associated with virions. CM1841 gene III was amplified by polymerase chain reaction (PCR) using Taq DNA polymerase (Promega Corporation, Madison,

WI), the primers 88G-1, 88G-2, and pCaMV10 (Gardner et al., 1981) as the template. All primer sequences used in this work are given in Supplementary

Table 1. The PCR products were then inserted into the EcoRI and XhoI sites within the yeast two-hybrid vectors pEG202 and pJG4-5 (Ausubel et al., 1993;

Gyuris et al., 1993). Full-length gene VI inserted in pJG4-5, was described earlier

(Li and Leisner, 2002) as were the yeast transformation procedures. Yeast transformants expressing P3 fused to the LexA DNA binding domain (DBD) and

P6 attached to the B42 transcription activation domain (TAD) grew on leucine- deficient media and exhibited β-galactosidase activity (Fig. 4-1). To confirm these interactions biochemically, the P3 coding region was amplified with the primers CaP3GWF and CaP3GWR. The PCR product was then inserted into

Gateway-compatible pMAL-c2X and maltose binding protein (MBP) pull-down assays were performed as described in (Hapiak et al., 2008b). These data confirmed that P6 binds to P3.

Our data, together with the observation that P6 stabilizes P3 in protoplasts

(Kobayashi et al., 1998), suggest that the P3-P6 interaction plays a role in viral infection. Perhaps P6 aids in assembling capsids “decorated” with P3 by bringing

P3 and P4 together. The P3 and P4 proteins are known to interact with each other

(Leclerc et al., 2001). Such decorated virus particles could then associate with P1 to permit cell-to-cell movement.

Figure 4-1. Interaction of the gene III (P3) and gene VII products (P7) with the gene VI product (P6) of Cauliflower mosaic virus. A, Diagram of constructions tested for interaction with the yeast two-hybrid system. Constructs are not drawn to scale. Black box, LexA DNA-binding domain (from pEG202); hatched box, B42 transcription activation domain (from pJG4-5); white boxes, full-length CaMV proteins; P3 (129 amino acids long), P6 (520 amino acids), P7 (96 amino acids). Numbers in bold to the left of each pair of constructs correspond to β-galactosidase assay data shown in B and yeast growth in C. B, β- galactosidase activity of yeast transformants expressing constructs shown in A. The bar graph indicates average β-galactosidase activity (on abscissa) for 3 independent yeast colonies along with the standard deviation. Numbers at the bottom (ordinate) correspond to the transformants in C. C, growth of yeast transformants on media with (left) and without (right) leucine. The streaks are numbered (key in D) to correspond to the constructs shown in A and the β- galactosidase activities shown in B. D, key for the plates in C. E, Maltose binding protein pull-downs of P3 or P7 with P6. Load; amount of protein initially loaded onto column; Flow-through, proteins that do not bind to the amylose column and are washed off; Elution, proteins eluting off the amylose resin. P3/P6, P3-Maltose binding protein fusion polypeptide expressed in Escherichia coli, was mixed with P6 expressed in E. coli, loaded on an amylose column, eluted with maltose and probed with a P6 antibody. P7/P6, P7-Maltose binding protein fusion polypeptide expressed in E. coli, was mixed with P6 expressed in E. coli, loaded on an amylose column, eluted with maltose and probed with a P6 antibody. pMAL/P6, Maltose binding protein expressed in E. coli, was mixed with P6 expressed in E. coli, loaded on an amylose column, eluted with maltose and probed with a P6 antibody.

As a result of the interactions observed among P6, P1 (Hapiak et al.,

2008b) and P3 (described above), we hypothesized that the gene VI product also interacts with P2, which, in turn, interacts with P3 (Leh et al., 1999). CM1841 gene II was amplified by PCR using the CaP2GWF and CaP2GWR primers. The gene II product was inserted into Gateway-compatible pEG202, pJG4-5 or pMAL-c2X. Yeast transformants expressing gene II from CM1841 (expressed from pEG202) and CM1841 P6 (expressed from pJG4-5) grew on leucine- deficient media (Fig. 4-2) and showed β-galactosidase activity. The P2-P6 interaction was confirmed biochemically, by MBP pull-down experiments as described above

P2 of CaMV isolate W260 was shown to have a greater stabilizing effect on P6 inclusion bodies than its CM1841 counterpart (Anderson et al., 1992a).

Therefore, we investigated if W260 P2 interacted with P6 in a manner different from its CM1841 counterpart. CaMV W260 gene II was amplified by PCR using the CaWP2F and CaWP2R primers with pCaMVW260 (Schoelz and Shepherd,

1988) as a template, then inserted into the EcoRI site of pEG202. Yeast transformants expressing W260 gene II (expressed from pEG202) and CM1841

P6 (expressed from pJG4-5) grew on leucine-deficient media (Fig. 4-2) and showed β-galactosidase activity. Interestingly, the W260 P2-P6 and CM1841 P2-

P6 interactions were similar based on leucine-independent growth and β- galactosidase activity. Hence, a difference in interaction with P6 for the W260 and CM1841 P2s was not detected, even though the inclusion body stabilization mediated by the P2 proteins was different. One possible explanation for this, is

82 that CM1841 P2 is less stable than its W260 counterpart (Blanc et al., 1993;

Nakayashiki et al., 1993) and this characteristic rather than the P2-P6 interaction is responsible for the difference in IB stability.

Another important function of P6 is translational transactivation, the reinitiation of ribosomes on the polycistronic 35S RNA to make the various viral proteins . Interestingly, the first open reading frame on the 35S RNA encoding a polypeptide is the hypothetical gene VII product (P7). The majority of gene VII can be deleted without obvious effects on viral infection (Dixon et al., 1986;

Dixon et al., 1983). However, mutagenesis of the initiation codon delays viral symptoms viruses harboring this mutation revert at a high frequency. While P7 can be expressed in yeast, it is undetectable in virus-infected plants, suggesting that it may be unstable (Wurch et al., 1990). Supporting this hypothesis is the observation that CaMV P5 protease cleaves P7 in vitro (Guidasci et al., 1992).

Since P6 regulates CaMV translation and P7 is the first polypeptide encoded by the 35S RNA, we speculated that both proteins may interact.

Figure 4-2. Interaction of the gene II (P2) and gene VI products (P6) of Cauliflower mosaic virus. A, Diagram of constructs tested for interaction with the yeast two-hybrid system. Black box, LexA DNA-binding domain; hatched box, B42 transcription activation domain; white (CM1841 proteins) and gray (W260 proteins) boxes, full-length CaMV genes; P2 (159 amino acids long), P6 (520 amino acids), P7 (96 amino acids). Numbers in bold to the left of each pair of constructs correspond to β-galactosidase assay data shown in B and yeast growth in C. B, β-galactosidase activity of yeast transformants expressing constructs shown in A. C, growth of yeast transformants on media with (left) and without (right) leucine. D, key for the plates in C. E, Maltose binding protein pull-downs of P2 with P6. P2/P6, P2-Maltose binding protein fusion polypeptide expressed in Escherichia coli, was mixed with P6 expressed in E. coli, loaded on an amylose column, eluted with maltose and probed with a P6 antibody. pMAL/P6, Maltose binding protein, was mixed with P6, loaded on an amylose column, eluted with maltose and probed with a P6 antibody.

To test this possibility, CM1841 gene VII was amplified using the primers

GVII-1F and GVII-2R and inserted into the EcoRI and XhoI sites within the yeast two-hybrid vector pEG202. Yeast transformants expressing P7 (fused to the

LexA DBD) and P6 (attached to the B42 TAD) grew on leucine-deficient media and exhibited β-galactosidase activity (Fig. 4-1). The P7 coding region was

84 amplified with the primers CaP7GWF and CaP7GWR, inserted into Gateway- compatible pMAL-c2X (Raikhy et al., 2011b) and MBP pull-down assays were performed which confirmed P6 binds P7. By interacting with P6, P7 may permit the former protein to monitor translation efficiency and more appropriately regulate translational transactivation. Alternatively, P7 may regulate the activity of the P5 protease as the first polypeptide encoded by some RNA viruses that process polyproteins is a regulator of the viral protease (Goldbach and Wellink,

1996). Yeast two-hybrid analyses failed to detect interactions of P2 or P3 with P7

(Fig. 4-1,4-2).

In summary, we have shown that P6 interacts with P2 and P3, two proteins essential for virus movement. We have also found that P6 interacts with a protein of unknown function, P7.

4.4 Acknowledgments

The authors thank Drs. Richard Komuniecki and Song-Tao Liu

34. This work was supported in part by NIH Grant number 1R15AI50641-01 and

USDA-ARS Specific Cooperative Agreement: 58-3607-1-193.

4.5 References

Anderson, E.J., Trese, A.T., Sehgal, O.P., Schoelz, J.E., 1992. Characterization of

a chimeric Cauliflower mosaic virus isolate that is more severe and

accumulates to higher concentrations than either of the strains from which

it was derived. Mol. Plant-Microbe Interact. 5, 48-54.

Armour, S.L., Melcher, U., Pirone, T.P., Lyttle, D.J., Essenberg, R.C., 1983.

Helper component for aphid transmission encoded by region II of

Cauliflower mosaic virus DNA. Virology 129, 25-30.

Blanc, S., Cerutti, M., Chaabihi, L.C., Devauchelle, G., Hull, R., 1993. Gene II

product of an aphid-nontransmissible isolate of Cauliflower mosaic virus

expressed in a baculovirus system possesses aphid transmission factor

activity. Virology 192, 651-654.

Covey, S., Hull, R., 1981. Transcription of Cauliflower mosaic virus DNA.

Detection of transcripts, properties, and location of the gene encoding the

virus inclusion body protein. Virology 111, 463-474.

De Tapia, M., Himmelbach, A., Hohn, T., 1993. Molecular dissection of the

Cauliflower mosaic virus translation transactivator. EMBO J. 12, 3305-

3314.

Dixon, L., Jiricny, J., Hohn, T., 1986. Oligonucleotide-directed mutagenesis of

Cauliflower mosaic virus DNA using a repair-resistant nucleoside

analogue: identification of an agnogene initiation codon. Gene 41, 225-

231. 86

Gardner, R.C., Howarth, A.J., Hahn, P., Brown-Luedi, M., Shepherd, R.J.,

Messing, J., 1981. The complete nucleotide sequence of an infectious

clone of Cauliflower mosaic virus by M13mp7 shotgun sequencing.

Nucleic Acids Res. 9, 2871-2888.

Goldbach, R.W., Wellink, J., 1996. Chapter 3: Comoviruses: molecular biology

and replication. In The plant viruses: polyhedral virions and bipartite RNA

genomes, B.D. Harrison, and A.F. Murant, eds. (New York, Plenum

Press), pp. 35-75.

Guidasci, T., Mougeot, J.L., Lebeurier, G., Mesnard, J.M., 1992. Processing of

the minor capsid protein of the Cauliflower mosaic virus requires a

cysteine proteinase. Res. Virol. 143, 361-370.

Gyuris, J., Golemis, E., Chertkov, H., Brent, R., 1993. Cdi1, a human G1 and S

phase protein phosphatase that associates with Cdk2. Cell 75, 791-803.

Haas, M., Bureau, M., Geldreich, A., Yot, P., Keller, M., 2002. Cauliflower

mosaic virus: still in the news. Mol. Plant Pathol. 3, 419-429.

Hapiak, M., Li, Y., Agama, K., Swade, S., Okenka, G., Falk, J., Khandekar, S.,

Raikhy, G., Anderson, A., Pollock, J., Zellner, W., Schoelz, J., Leisner, S.

M., 2008. Cauliflower mosaic virus gene VI product N-terminus contains

regions involved in resistance-breakage, self-association and interactions

with movement protein. Virus Res. 138, 119-129.

Himmelbach, A., Chapdelaine, Y., Hohn, T., 1996. Interaction between

Cauliflower mosaic virus inclusion body protein and capsid protein:

Implications for viral assembly. Virology 217, 147-157.

Hull, R., 2002. Matthews' Plant Virology. Academic Press, London.

Kobayashi, K., Hohn, T., 2003. Dissection of Cauliflower mosaic virus

transactivator/viroplasmin reveals distinct essential functions in virus

replication. J. Virol. 77, 8577-8583.

Kobayashi, K., Tsuge, S., Nakayashiki, H., Mise, K., Furusawa, I., 1998.

Requirement of Cauliflower mosaic virus open reading frame VI product

for viral gene expression and multiplication in turnip protoplasts.

Microbiol. Immunol. 42, 377-386.

Leclerc, D., Stavolone, L., Meier, E., Guerra-Peraza, O., Herzog, E., Hohn, T.,

2001. The product of ORF III in Cauliflower mosaic virus interacts with

the coat protein through its C-terminal proline rich domain. Virus Genes

22, 159-165.

Leh, V., Jacquot, E., Geldreich, A., Haas, M., Blanc, S., Keller, M., Yot, P., 2001.

Interaction between the open reading frame III product and the coat

protein is required for transmission of Cauliflower mosaic virus by aphids.

J. Virol. 75, 100-106.

Li, Y., Leisner, S.M., 2002. Multiple domains within the Cauliflower mosaic

virus gene VI product interact with the full-length protein. Mol. Plant-

Microbe Interact. 15, 1050-1057.

Love, A.J., Laird, J., Holt, J., Hamilton, A.J., Sadanandom, A., Milner, J.J., 2007.

Cauliflower mosaic virus protein P6 is a suppressor of RNA silencing. J.

Gen. Virol. 88, 3439-3444.

Raikhy, G., Krause, C., Leisner, S.M., 2011. The Dahlia mosaic virus gene VI

product N-terminal region is involved in self-association. Virus Res 159,

69-72.

Scholthof, K.-B. G., Adkins, S., Czosnek, H., Palukaitis, P., Jacquot, E., Hohn, T.,

Hohn, B., Saunders, K., Candresse, T., Ahlquist, P., Hemenway, C.,

Foster, G.D., 2011. Top 10 plant viruses in molecular plant pathology.

Mol. Plant Pathol. 12, 938-954.

Schoelz, J.E., Shepherd, R.J., 1988. Host range control of Cauliflower mosaic

virus. Virology 162, 30-37.

Schoelz, J.E., Goldberg, K.-B.; Kiernan, J., 1991. Expression of Cauliflower

mosaic virus (CaMV) gene VI in transgenic Nicotiana bigelovii

complements a strain of CaMV defective in long-distance movement in

nontransformed N. bigelovii. Mol. Plant-Microbe Interact. 4, 350-355.

Stavolone, L., Villani, M.E., Leclerc, D., Hohn, T., 2005. A coiled-coil interaction

mediates Cauliflower mosaic virus cell-to-cell movement. Proc. Natl.

Acad. Sci. USA 102, 6219-6224.

Thomas, C.L., Perbal, C., Maule, A.J., 1993. A mutation of Cauliflower mosaic

virus gene I interferes with virus movement but not virus replication.

Virology 192, 415-421.

Wurch, T., Kirchherr, D., Mesnard, J.-M., Lebeurier, G., 1990. The Cauliflower

mosaic virus open reading frame VII product can be expressed in

Saccharomyces cerevisiae but is not detected in infected plants. J. Virol.

64, 2594-2598.

4.6 Supplementary Data

Table 4-1. Primer used to amplify CaMV genes II, III, and VII.

aPrimer names ending with suffixes ending in F are forward primers, while those with R are reverse primers (synthesized by Integrated DNA Technologies, Coralville, IA. Primers with GW before the F or R suffix are GatewayR compatible primers. bNucleotide sequences of primers, written 5’ to 3’, nucleotides differing from the CM1841 sequence (Gardener et al., 1981; GenBank Accession Number V00140) are written in bold, while restriction enzyme sites (EcoRI in forward primers, XhoI in reverse primers) are underlined and stop codons are given in italics. cLocation of 5’ nucleotide on the CM1841 genomic sequence. For the forward GatewayR primers, the location of the first nucleotide encoded by the CM1841 genomic sequence is indicated in italics.

Chapter 5

Additional CaMV Experiments: P6 and Reverse Transcriptase Protein (P5).

5.1 Abstract

Reverse transcriptase enzymes play an essential role in viral genome replication in both retro and pararetroviruses. They contain three domains: an N- terminal aspartic acid protease, a central reverse transcriptase domain (RT), and a

C-terminal RNase H domain, in general. CaMV, a plant pararetrovirus uses a RT enzyme to replicate its genome. Based on sequence analysis, CaMV P5 has a similar amino acid sequence to retroviruses, HIV and MMLV, than pararetrovirus

HBV. In this study I focused on further understanding P5s ability to interaction with itself and P6. From our pull-down analysis we were able to show that P6 can interaction with CaMV RT-RNase H region (P5MC) and weakly interact with the protease and full-length P5. In addition, we also showed P5 can self-associate via the RT-RNase H domains.

5.2 Introduction

Retro- and pararetroviruses encode a versatile DNA polymerase called reverse transcriptase (RT), which plays an essential role in viral genome replication (Rothnie, 1994). Retroviral RT is part of a larger protein called pol that is a multi-domain protein. Pol proteins are multi-domain enzymes that generate cDNA from an RNA template. In retroviruses, RT copies viral RNA into DNA (Das, 2004; Rothnie, 1994; Wintermantel, 1993). The newly synthesized DNA is then integrated into the host genome. On the other hand, pararetroviruses, are DNA viruses that replicate their genomes through an RNA intermediate, employing an RT, and typically do not integrate into the host genome (Rothnie, 1994; Takatsuji, 1992). All RT’s have at least three sequential biochemical activities DNA synthesis; RNA-dependent DNA polymerase, ribonuclease H (RNase H) function, and DNA-dependent DNA polymerase

(Rothnie, 1994). RT is unique among related DNA polymerizing enzymes in its ability to utilize either RNA or DNA as a template. Although RTs are functionally similar they are in fact structurally diverse (Goff, 1990), including heterodimer and monomeric enzymes. RT occurs in a wide range of living organisms including: viruses, prokaryotes and eukaryotes (Xiong, 1990)

HIV-1 (Human Immunodeficiency Virus) and MMLV (Moloney Murine

Leukemia Virus) are two widely studied retroviruses that utilize RT to replicate their genomes (Das, 2004; Tasara, 1999) (Fig.5-1). Although these two viruses function in a similar fashion, their RTs are architecturally different (Das, 2004).

HIV-1 RT is a heterodimer of p66 and p51 subunits (Tasara, 1999). The p66 93 subunit contains RT and RNase H domain whereas, the p51 subunit is produced by the proteolytic cleavage of the p66 subunit, which lacks the RNase H domain.

The RT-RNase H domains are thought to interact with the RT domain forming the heterodimer. On the other hand, MMLV RT is monomeric as all five domains are located on a single polypeptide and all the functions are found in the same protein

(Das, 2004).

Figure 5-1. Structural elements of Retro and Pararetrovirus pol proteins. The length of CaMV, HIV-1, MMLV, and HBV are indicated in amino acids as shown above. In addition the various domians within each protein are indicated above.

In addition to classical retroviruses and retrotransposons, RTs are found in both animal and plant pararetroviruses. The most commonly studied pararetrovirus RTs are Hepatitis B virus (HBV) and Cauliflower mosaic virus

(CaMV), respectively (Fig.5-1). Unlike retroviruses and CaMV, HBV RT is very distinct and is composed of four domains (terminal protein (TP), spacer region, RT domain, and RNase H domain) (Jones, 2013). Although, the HBV

94 enzyme functions are similar to all other RTs, this protein does not contain a protease domain at the N-terminus. Instead, it contains a terminal protein (TP), which is separated from the RT domain by a nonessential/nonconserved spacer region. The HBV RT also contains a C-terminal RNase H domain (Hu, 2004).

The TP region within the RT is used as a protein primer, in contrast to CaMV and retroviruses that use a specific tRNA to initiate replication (Hu, 2004).

Additionally, TP contains a T3 motif at its C-terminal end, which is important for binding, RNA packaging and protein priming (Jones, 2013). The RT domain of

HBV is homologous to retroviral RTs to include HIV-1. HBV RT shows significant homology to short motifs in retroviral RTs that form a well-defined catalytic core. The RNase H domain has been shown to be important for pregenomic RNA packaging. Furthermore, previous studies have shown RT activation is triggered by host chaperones, most specifically heat shock protein 90

(Hsp90) (Hu, 2004). Hsp90 is essential and required for RT function, protein-

RNA interaction, and RNA packaging, which will be discussed later.

CaMV, a plant pararetrovirus, uses RT to replicate its genome (Takatsuji,

1992). The CaMV encodes a protein called P5, which consists of three domains: an N-terminal protease domain, a central RT domain, and a C-terminal RNase H domain (Rothnie, 1994). Unlike retroviruses, CaMV P5 does not contain an integrase domain. Similar to other retrovirus pols such as HIV, MMLV, and Rous sarcoma virus, an aspartic protease is located on the N-terminus CaMV of the P5 protein (Takatsuji, 1992). A proposed mechanism suggests that P5 is cleaved autocatalytically to release the N-terminal protease domain (Takatsuji, 1992).

Previous x-ray crystallographic studies have shown HIV-1 protease to dimerize

(Wlodawer, 1989). The CaMV protease has also been proposed to dimerize

(Torruella, 1989). Full-length P5 containing the protease is inactive for RT activity (Takatsuji, 1992). However, the P5 RT domain is activated by proteolytic processing catalyzed by the P5 protease (Le Grice, 1988). In addition to self- processing to activate P5 RT, the P5 protease also cleaves other CaMV proteins.

The aspartic protease domain within the pol of retroviruses processes the gag and gag-pol polyprotein precursors, which leads to the rearrangement and assembly of mature virus particles (Guerra-Peraza, 2000; Mervis, 1988). Since

CaMV P4 (capsid protein) is the CaMV equivalent of retroviral gag proteins

(Kobayashi and Hohn, 2003), it would make sense for P5 protease to also cleave capsid protein. Indeed, P5 protease cleaves P4 to remove the N-terminus and to trigger the formation of mature virions (Takatsuji, 1992; Torruella, 1989).

To date, not much is known about the CaMV RT. All RTs contain a conserved motif YXDD. CaMV and MMLV RT contain YVDD at this location whereas, HIV-1 and HBV RT contain YMDD instead (Wakefield, 1992).

Previous studies, performed by making point mutations within the YXDD region of HIV-1 RT, have shown reduced activity. Therefore, this amino acid motif is essential for all polymerase function because it forms part of the active site.

Interestingly, the change of YMDD to YVDD in HIV-1 and MMLV RT has been suggested to increase fidelity of the enzyme (Kaushik, 2000; Wakefield, 1992), although other data dispute that hypothesis (Back, 1996).

Previous results suggest that CaMV genome replication by reverse transcription occurs within inclusion bodies (IBs) (Mazzolini, 1989). For example, the 35S RNA, which is the template for P5 RT, was detected in purified viroplasms. However, whether reverse transcription of the viral genome occurs in the IB matrix or within virion-like particles is unclear. Reverse transcription of the viral RNA in all known retroviruses as well as HBV has been shown to occur within the capsid (Hu, 2004; Rothnie, 1994). Therefore, it is possible that CaMV replication could also occur within virion-like particles (Kobayashi and Hohn,

2003). However, it has also been suggested that CaMV replication occurs in viral precursor particles, or in the IB matrix and must be completed before capsid maturation (Himmelbach, 1996).

In addition to IBs being the site of viral replication, the major IB protein

P6 has also been implicated in viral replication and the stabilization of P5

(Kobayashi and Hohn, 2003; Torruella, 1989). Most importantly, P6 domains D2 and D3 are required for viral replication, as well as CaMV gene products P4, P5.

Therefore it’s not surprising that P6 stabilizes both P4 and P5 (Kobayashi, 1998).

CaMV protein was expressed in turnip protoplasts, P4, and P5 were not detected unless co-expressed with P6.

The ability of P6 to stabilize viral gene products involved in replication suggests P6 could be a chaperone protein. As mentioned previously, Hsp90 plays an essential role in the function of HBV RT (Hu, 2004)Hsp90 is a molecular chaperone protein associated with the folding of signal transducing proteins

(Buchner, 1999; Caplan, 1999; Cho, 2000). Hsp90 works in concert with other 97 chaperone proteins to help refold denatured proteins into a competent state (Cho,

2000). The molecular chaperone complex, Hsp90 interaction with HBV RT, helps facilitate the RT-RNA interaction.

To date not much is known about CaMV P5. Therefore I wanted to further investigate P5 function. To do this, I applied a bioinformatics approach to examine the relationship of various P5 domains with similar regions of pol proteins from other viruses. I also examined if P5 can self-associate.

Determining if CaMV P5 forms a heterodimer or is monomeric will allow us to better understand its function. Because of its stabilizing effect, I also investigated if P6 interacts with P5. These data will provide further insights regarding P5 functions during the CaMV life cycle.

5.3 Materials and Methods.

Bioinformatics.

The predicted amino acid sequence of the aspartic acid protease, reverse transcriptase, and RNase H domains of P5 of two CaMV isolates CM1841

(Genbank accession number NP_056728.1) and W260 (Genbank accession number AF161742.1) as well as those domains of the HIV (Genbank accession number NP_0578494), MMLV (Genbank accession number NP_057933.2) and

HBV (Genbank accession numbers NP_647602.2 for RT domain and

BAL61052.1 for RNase H domain) pol proteins were aligned employing the

ClustalW contained within the MegAlign program of the DNASTAR Lasergene 7 software package (Madison, WI).

P5 Constructions and Protein Analysis.

Clones of full-length P5, the P5-N-terminal proteinase (P5N), and P5- middle and C-terminus (P5 MC) were generated by PCR. Full-length P5 was generated using primers GW CAMVP5N-1F - 5'

CACCATGATGAATCATCTACTTCTG 3' and GW P5MC – 2R – 5’

AATCTTATCTTCGGATTTCAATTA 3’, P5N was generated using GW

CAMVP5N-1F - 5' CACCATGATGAATCATCTACTTCTG 3' and GW

CAMVP5N -2R - 5' TTATTCTGATAACCTCCTCCC3', while P5MC was generated using GWP5MC -1F – 5’ CACCAATCCACTAGAAGAAATTGC 3’ and GWP5MC – 2R – 5’ AATCTTATCTTCGGATTTCAATTA 3’. A greater than full-length clone of CM1841 (p1.5 CaMV) (Callaway, 1996)was used as the template for constructing these clones as it contained a continuous gene V. Note, pCaMV10, the plasmid used a template for all of the other constructions made in this thesis contains the pBR322 cloning vector in the center of gene V and so could not be used for the experiments in this chapter. All of the above PCR products were inserted into pENTR and then mobilized into Gateway® compatible pMAL-c-2X or pDEST according to the instructions provided with the

Gateway system.

After insertion of P5N into pMAL-c-2X, a single amino acid change was generated using the Stratagene Quickchange II site-directed mutagenesis kit using the primers: Gene V D46A-1F (5’

GAGCTTCACTGTTTTGTAGCCACGGGAGCAAGCTTATGC 3’) and Gene V

D46A-2R (5’ GCATAAGCTTGCTCCCGTGGCTACAAAACAGTGAAGCTC

3’) according to the manufacturer’s specifications. This mutation was reported to inactivate the P5 protease (Kohl, 1989). This was important to examine because after insertion of P5MC into pENTR, the P5 coding sequences were also mobilized into pSITE-4CA vector. This plasmid expresses a red fluorescent protein (RFP)-P5MC protein with the fluorophore located at the N-terminal end of P5MC. All plasmid constructions were confirmed by sequencing and more detailed information regarding these constructions is given in Appendix 2.

Maltose binding protein (MBP) pull-down assays between P5 and its domains as well as P5-P6 interactions were performed as described in Lutz et al.

(2012). GFP-tagged wild type and mutant CM1841 P6 as well as GFP-tagged

W260 P6 were described in Chapter 3. Fluorescence microscopy to analyze the cellular distribution of P5MC and its possible co-localization with P6 was performed as described in Chapter 2.

5.4 Results

For decades it has been known that CaMV uses reverse transcription to replicate it’s genome and fulfill its life cycle (Hohn, 1985; Hull, 1983; Pfeiffer,

1983). To do this, proteolytic processing must first occur to make the CaMV RT active. This process is thought to occur within IBs (Torruella, 1989). However, little is known about P5 self-association or its interacting partners. To further study this I first examined the relationship of CaMV P5 proteinase, RT, and

RNase H domains to those of other CaMVs as well as to those of retroviruses.

The HIV RT domain is dimeric in its functional form, separated from its protease,

100 while the MMLV domain is monomeric. Multiple sequence alignment (MSA) were performed and phylogenetic trees were generated on the individual pol domains (protease, RT, and RNase H) using two CaMV isolates 1841 and W260 and distantly related mammalian viruses HIV-1, MMLV, and HBV to search for putative phylogenetic relationships.

The MSA of the protease domain showed one very short conserved region throughout all the RTs examined (Fig. 5-2A). This region is comprised of the sequence motif V/LDTGA that is believed to make up the aspartic proteinase active site. The phylogenetic tree grouped the RTs into two clades (Fig. 5-2B).

One clade contained the two CaMV isolates CM1841 and W260 (that shared

>95% amino acid sequence identity, while the second contained the MMLV and

HIV proteases (that shared 31.8% amino acid sequence identity). Based on amino acid sequence identity, a proteinase domain could not be identified in the HBV pol protein. The CM1841 protease domain shared 23.5% amino acid sequence identity to the HIV protease and 10.9% to that of MMLV.

101

Figure 5-2. Multiple sequence alignment (MSA) and phylogenetic tree of the Protease domain within different pol proteins. A. MSA of CM1841, CMW260, HIV, and MMLV protease domains using Megalign (DNAstar). Sequences conserved with the CM1841 are highlighted in black. B. Phylogenetic tree of sequences aligned in A.

The MSA of the RT domains showed individual and pairs of amino acids conserved within the domains (Fig. 5-3A). The two regions that stood out were the sequence motif YV/MDD and TAFT. YXDD (where X is V or M) is believed to make up part of the RT active site and hence, is believed to be essential for polymerase function. Methionine is present in all CaMV strains and MMLV, whereas Valine is present in HIV and HBV. The TAFT sequence motif is conserved among all CaMV strains examined, along with HIV. However,

MMLV contains FAFE at these positions while HBV shares no sequence similarities. The phylogenetic tree grouped the RTs into two clades (Fig. 5-3B).

One clade contained the two CaMVs as well as MMLV and HIV that are distantly related, while the second contained only HBV. The CM1841 RT domain shared

102

30.1% and 29.7% amino acid sequence identity to the HIV and MMLV region, respectively, but only 17.3% to the corresponding region of the HBV protein.

Figure 5-3. Multiple sequence alignment (MSA) and phylogenetic tree of the RT domain within different pol proteins. A. MSA of CM1841, CMW260, HIV, MMLV, and HBV protease domains using Megalign (DNAstar). Sequences conserved with the CM1841 are highlighted in black. B. Phylogenetic tree of sequences aligned in A.

The MSA of the RNase H domains showed less obvious conservation than the protease and RT domains (Fig. 5-4A). The phylogenetic tree grouped the

RNase H domain into two clades (Fig. 5-4B). Clade one contained the two

CaMV isolates CM1841 and CMW260, while the second clade contained HIV,

MMLV and HBV. The CM1841 RNase H domain shared 20.1% amino acid sequence identity to MMLV, but only 13.2% and 10.5% with HIV and HBV

103 respectively. Overall, it would appear that CM1841 P5 amino acid sequence is more similar to both MMLV and HIV pol than HBV.

Figure 5-4. Multiple sequence alignment (MSA) and phylogenetic tree of the RNase H domain within different pol proteins. A. MSA of CM1841, CMW260, HIV, MMLV, and HBV protease domains using Megalign (DNAstar). Sequences conserved with the CM1841 are highlighted in black. B. Phylogenetic tree of sequences aligned in A.

Based on the sequence analysis above, CaMV P5 protease domain has similar amino acid identity to HIV than MMLV or HBV. The CaMV P5 RT domain has similar amino acid identity to HIV and MMLV than HBV, whereas the RNase H domain has similar amino acid identity to MMLV than HIV or

HBV. Because the RT domain has similar amino acid identity to both MMLV and HIV and these two RT proteins are present in different forms, I wanted to further investigate whether or not CaMV RT could self-associate and/or interact with its other domains. Therefore, I focused my studies on examining the

104 interactions of the RT-RNase H portion of P5 (P5MC; P5 central and C-terminal portions). I studied the interactions of P5MC with full-length P5, P5MC fragment, and the N-terminal protease domain by MBP pull-down assays.

MBP pull-down results show P5MC can bind to full-length P5 (Fig. 5-5).

Furthermore, P5MC can self-associate and this interaction appeared to be the strongest compared to the other constructs. Finally, P5MC was able to bind to the protease (P5N) and an inactive protease mutant (P5N MUT D46A). However, the interaction seems weak due to decreased band intensity.

Figure 5-5. Maltose binding protein (MBP) pull-down analysis of P5MC domain interactions. LOAD; protein initially loaded onto column; WASH, proteins that do not bind to the amylose column and are washed off; ELUTION, proteins eluting off the amylose resin. For labels at the right of each row, the first indicates the protein fused to MBP, the second indicated the protein fused to glutathione-s-transferase (GST). All proteins were collected from the column after the treatments indicated, separated by SDS-PAGE, blotted onto nitrocellulose and detected using an anti-GST antibody. Those samples with MBP alone served as a negative control.

105

Because CaMV reverse transcription occurs in IBs(Torruella, 1989), I next investigated if full-length and portions of P5 could interact with P6, the major IB protein (Fig. 5-6). From the MBP pull-down analysis, my results show P6 can interact with full-length P5 as well as to wild type (P5N) and mutant (P5N MUT) protease. However, the binding of P6 to P5N and P5N MUT appeared somewhat weaker than binding of these domains to P5MC, based on band intensity. There was no interaction between MBP alone and P6.

Figure 5-6. Maltose binding protein (MBP) pull-downs analysis of P5 domains with P6. LOAD; protein initially loaded onto column; WASH, proteins that do not bind to the amylose column and are washed off; ELUTION, proteins eluting off the amylose resin. For labels at the right of each row, the first indicates the protein fused to MBP, the second indicated the protein P6 expressed in E. coli. All proteins were collected from the column after the treatments indicated, separated by SDS-PAGE, blotted onto nitrocellulose and detected using an anti-P6 antibody (Hapiak et at 2008). Those samples with just MBP alone served as a negative control.

Since full-length P5 as well as the N-terminal protease domain interact with P6, albeit weakly, I next tested if P5MC could also bind P6. In MBP pull- down experiments, I found that P5MC interacted with P6 (Fig.5-7A). However,

MBP alone did not pull down P6. Interestingly, the ability of P6 to bind to P5MC

106 appeared stronger than the binding to P5N or full-length P5. To analyze the interaction in cells, I performed fluorescence microscopy (Fig. 5-7B).

Fluorescence microscopy data showed that P5MC fused to red fluorescent protein

(P5MC:RFP) was widely distributed throughout cells. However, some

P5MC:RFP co-localized with either CM1841- or CMW260-P6:GFP. P5MC:RFP did not co-localize with every P6 IB.

Previously, we generated a set of P6 mutants that varied in their ability to bind to the P6 self-association domains, to form large IBs, and to induce infections when present within the context of a viral genome (Chapter 3).

Therefore, I wanted to determine if these P6 mutants would affect P5s ability to co-localize with P6. To do this I employed fluorescence microscopy. My microscopy results showed that the P6 mutations did not affect P6s ability to interact with P5 (Fig. 5-8). Furthermore, I was unable to determine if the mutant

P6s interacted differently with P5. P5 co-localized with the larger P6 IBs, which is compatible to the wild type P6 results from Fig. 5-6.

107

Figure 5-7. Maltose binding protein (MBP) pull-down and fluorescence microscopy analysis of P5MC interactions with P6. A. LOAD; protein initially loaded onto column; WASH, proteins that do not bind to the amylose column and are washed off; ELUTION, proteins eluting off the amylose resin. For labels at the right of each row, the first indicates the protein fused to MBP, the second indicated the protein P6 expressed in E. coli. All proteins were collected from the column after the treatments indicated, separated by SDS-PAGE, blotted onto nitrocellulose and detected using an anti-P6 antibody (Hapiak et at 2008). Those samples with just MBP alone served as a negative control. B. Colocalization of CaMV P6-GFP (CM1841 and CMW260) and P5MC-RFP in N,benthamiana leaves by confocal microscopy. Expression of CaMV P6-GFP and P5MC-RFP at 3dpi. Left panels: CaMV P6-GFP (CM1841 and CMW260) alone, Middle panels: P5MC-RFP expressed in the same cells as P6-GFP, and Right panels: overlay of left and middle panels.

108

Figure 5-8. Fluorescence microscopy of P5MC with P6 mutants. Colocalization of CaMV P6 mutants-GFP and P5MC-RFP in N,benthamiana leaves by confocal microscopy. Expression of CaMV P6 mutants-GFP and P5MC-RFP at 3dpi. Left panels: CaMV P6 mutants-GFP alone, Middle panels: P5MC-RFP expressed in the same cells as P6 mutants-GFP, and Right panels: overlay of left and middle panels.

5.5 Discussion

Retro- and pararetro-viruses encode an RNA-dependent DNA polymerase that facilitates their genome replication within a host. CaMV, a plant pararetrovirus, utilizes just such an enzyme to replicate its genome.

Unfortunately, there are still many unknowns about CaMV RT. In this study I aimed to further understand CaMV P5 function using bioinformatics techniques and protein binding studies.

CaMV P5 contains three domains: an N-terminal aspartic acid protease, a central reverse transcriptase domain (RT), and a C-terminal RNase H domain.

This organization is also observed in the pol proteins of retroviruses and the last two domains are observed in the pararetrovirus HBV. A protease domain is not obvious in the HBV pol sequence. The MSA analysis showed that the CaMV P5 109 protease is most similar to that within HIV pol while the RNase H domain is most similar to that within the MMLV pol. However, the RT domain of P5 is most similar to both HIV and MMLV pol proteins. It is interesting that P5 is more identical to both HIV and MMLV retroviruses than HBV, a pararetrovirus. These data may suggest that even though CaMV is classified as a pararetrovirus, the P5 gene may have emerged from a retrovirus via a separate evolutionary path than that which gave rise to HBV.

Since HIV and MMLV sequence identity are similar to CaMV, inferences on protein structure could not be made by an MSA alone. Therefore, P5 self- interaction experiments were employed. To date little is known about P5-P5 interactions and P5 self-association. My pull-down assays show that P5MC was able to interact with full-length P5 and with the N-terminal protease domain. I also demonstrated P5MC-P5MC self-interaction.

This led me to speculate that P5MC interactions with full-length P5 may occur between the MC-MC domains, because that was apparently the strongest interaction that I detected. Due to the apparent weakness of the interactions of

P5MC with P5N, and P5N Mutant, I believe that these may be transient interactions. However, P5 activities occur in IBs (see below), a molecularly crowded environment and so these interactions may actually be stronger in virus infected cells. The P5MC self-association data suggest that this region could be functional as a dimer. Early work on the CaMV P5 RT activity showed that deletion of the N-terminal protease, but without removal of the RNase H domain gave rise to a functional enzyme. Therefore, the actual functional form of P5

110 responsible for RT activity is currently unclear. Future work will need to be done to separate the P5MC into the individual RT and RNase H domains to investigate this further. It will be necessary to examine the interactions of these individual domains and to then measure the RT activity of various combinations of portions of P5.

Previous studies have shown that CaMV RT activity occurs within IBs made primarily of P6 (Mazzolini, 1989; Pfeiffer, 1983). In addition if P6 is unable to properly form IBs, this results in virus that is either of reduced infectivity or is non-infectious. This suggests that P5 replication of the viral genome is dependent on IBs. Furthermore, P6 stabilizes P5 when both proteins are transiently expressed in protoplasts. Taken together, these data suggested that

P5 may interact with P6. My recent pull-down data confirmed that full-length P5 interacted with P6 as did the P5 protease domain. However, both of these interactions appeared to be rather weak. The P5MC region interacted with P6 more strongly than the full length or the protease domain based on pull-down data. Hence, these data may suggest that the proteolytic processing of P5 that occurs during the course of a viral infection (Takatsuji, 1992) and that activates

RT activity, may also render the protein more able to bind to P6. Perhaps this plays a regulatory role modulating P5 activity. Perhaps P6 influences the efficiency of the RT reaction or the fidelity. P6 has been suggested to play a chaperone-like role in virus particle formation (Himmelbach, 1996).

Furthermore, the pol protein of HBV, another pararetrovirus, requires host cell chaperones for RT activity. Perhaps P6 fulfills the role that molecular chaperones

111 may have. One reason why this may be important is because replication of HBV does not occur in IBs, while it does for CaMV. One might speculate that the

CaMV P6 may act as a barrier impeding the acquisition of host cell chaperones into IBs. In this case then P6 may serve the chaperone type role to permit proper

P5 function.

112

Chapter 6

Interactions of CaMV P6 with host proteins.

6.1 Abstract

CaMV gene VI product P6 forms in large amorphous, electron dense IBs.

These IBs are thought to be the site of viral protein synthesis, genome replication and encapsidation. P6 has been shown to interact with a multitude of host factors, which play a role in translational transactivation to cell-to-cell movement. P6 has also been shown to interact with microfilaments to allow for intra cellular movement of IBs. Our collaborator Dr. James Schoelz from the University of

Missouri paid for a commercial yeast two-hybrid screen between Arabidopsis proteins that could potential interact with CaMV P6. From this screen, three proteins were identified: CHUP1, C2CDMT, and FIT. In this study I focused on confirming these three proteins could interact with CM1841 and which domain specifically interacted with these host proteins. In addition, I examined if these proteins could also interaction with a strain of CaMV termed D4 that has a broader host range than CM1841. Our yeast-two hybrid results showed, all three

113 host factors can interact with P6 from CM1841 but not D4. However, the D2 domain of P6 interacted more efficiently with all three-host proteins than the other individual self-association domains. This suggest that D2 may be the interface between P6 and host factor interactions.

6.2 Introduction

Viruses are obligate parasites that require a host to fulfill their life cycle

(Koonin, 2006). The host provides an environment for viruses to propagate and because of this, there must be many interactions between viral and host factors.

For example, plant viruses spread from one cell to another by modifying the function of normal cellular junctions called plasmodesmata (PD). This method of spread is unique to plant viruses as most mammalian viruses, spread from cell-to- cell by exocytosis and cell lysis (Patton, 2006; Pesavento, 2006; Stoye, 2012).

All plant viruses contain movement proteins (MPs) permitting viral nucleic acids to move from cell to cell.

Therefore, viral proteins must interact with host cell PD components to permit the spread of plant viruses. Other plant viral proteins are multifunctional and likely interact with many host factors to mediate their many activities. As mentioned previously, CaMV P6 is a good example of a multifunctional protein.

P6 mediates a variety of functions from silencing suppression to translational transactivation. Both of these activities involve interactions of P6 with host factors. P6 can interact with the DRB4 protein involved in the siRNA pathway in

Arabidopsis thaliana, presumably interfering with silencing. P6 also interacts with large ribosomal subunit proteins L13, L18, and L24, as well as with 114 eukaryotic translation initiation factor 3, subunit g, all of which are thought to play a role in translational transactivation. P6 also forms IBs where most of the processes of viral propagation occur. Virus particles are formed in IBs, but it is unclear how virions actually reach the PD for cell-to-cell movement. It is possible that like RNA viruses such as Tobacco mosaic virus (TMV), in which replication factories are transported to the PD, perhaps the CaMV IBs are transported there as well. This would suggest that CaMV IBs and P6 in particular

(because P6 makes up the bulk of CaMV IBs) may interact with components of the plant cell cytoskeleton to permit IB transport within cells.

Finally, P6 has been known for some time, to be a symptom determinant

(Anderson, 1991; Hapiak, 2008; Schoelz, 1986). The types of symptoms caused by different CaMV isolates are dictated by the P6 that they harbor. Some viruses induce systemic mosaics on their host plants, while other cause the formation of systemic necrosis or no symptoms at all (Leisner unpublished data). Interestingly, expression of P6 alone in transgenic plants can induce symptoms that mimic those of the virus from which the gene VI was derived (Cecchini, 1997). For example, transgenic Arabidopsis plants expressing P6 from CM1841 show mosaic symptoms typical of that viral isolate. In contrast, transgenic Arabidopsis expressing P6 from CMW260 produce more extreme mosaic and stunting symptoms indicative of the more severe nature of this viral isolate. Interestingly, transgenic plants expressing P6 from the D4 isolate of CaMV are asymptomatic.

Arabidopsis plants infected with D4 typically show no symptoms even though the virus can easily be detected by PCR and plant skeleton hybridization techniques

115

(Dr. Scott Leisner, personal communication). These data suggest that P6 likely interacts with host factors involved in some aspect of chlorophyll or chloroplast function. It is also possible that rather than interacting with these components directly; P6 may interact with plant signaling components regulating chloroplast function. Taken together, these observations suggest that CaMV P6 may interact with a variety of host factors and that these host factors may play a role in virus movement, symptom formation, and cell signaling. Thus, a screen of potential plant interactors with CaMV P6 could provide useful information regarding the many facets of how this protein can carry out so many different functions.

Therefore, our collaborator, Dr. James Schoelz from the University of

Missouri, paid for a commercial yeast two-hybrid screen of Arabidopsis proteins interacting with CaMV P6 to be done. Many P6 interacting proteins were identified in this screen. Three very interesting proteins identified were: chloroplast unusual positioning 1 (CHUP1), calcium-dependent membrane targeting protein with a C2 domain (C2CDMT), and Fe-deficiency-induced transcription factor 1 (FIT). It is possible that interactions of P6 with CHUP1 and/or C2CDMT may help facilitate viral movement. It is possible that through its interactions with FIT, P6 may facilitate symptom formation.

CHUP1 is a type XI myosin protein that anchors chloroplasts to the plasma membrane (PM) and contributes to movement of chloroplasts on the actin cytoskeleton in response to changes in light intensity (Oikawa, 2003, 2008).

CHUP1 has four important functional domains (Fig. 6-1). The N-terminus targets

CHUP1 to the chloroplast outer envelope. The second domain consists of a 116 coiled-coil motif, which interacts with the plasma membrane and is important for homo-dimerization of the protein. The third domain binds to F-actin, presumably mediating the movement of chloroplasts along the cytoskeleton. Finally the C- terminal domain contains a proline-rich domain that interacts with profilin and actin, along with a leucine zipper that may interact with other unknown plant proteins (Oikawa et al 2003, 2008). CHUP1 may play a role in intracellular movement of IBs to the PD.

Figure 6-1. Structural elements of Arabidopsis thaliana CHUP1. CHUP1 is 1004aa in length and the various domains within CHUP1 are indicated above.

C2 is a membrane targeting two-domain protein 360 amino acids in length

(Kim, 2008; Takahashi, 1997)This protein contains an N-terminal C2 domain and a C-terminal proline-rich domain (Fig, 6-2). C2 domains are approximately 130 amino acids long and bind to calcium ions (Nalefski, 1996). The C2 domain may play a signaling role but its function, like that of the proline-rich domain is unclear. Studies by our collaborator show that C2 associates with the PM

(Schoelz unpublished data). In addition, other preliminary data suggest that C2 interacts with the base of tubule structures composed of CaMV P1 protein. If confirmed, these data suggest that C2 could play a role in CaMV cell-to-cell movement.

117

Figure 6-2. Structural elements of Arabidopsis thaliana C2CDMT. C2CDMT is 360aa in length and the various domains within C2CDMT are indicated above.

FIT is a 360 amino acid transcription factor (TF) that contains a sequence- specific basic helix-loop-helix DNA binding domain (bHLH) (Bauer, 2007)(Fig.

6-3). FIT is located within the nucleus, where it regulates the expression of iron transporters, and therefore, regulates iron uptake. In addition FIT is responsive to ethylene and nitric oxide permitting iron uptake to be regulated by these signaling compounds. Plants require the essential micronutrient iron (Fe) for photosynthesis and chlorophyll biosynthesis. Plants with iron deficiency exhibit leaf chlorosis and stunted growth.

Figure 6-3. Structural elements of Arabidopsis thaliana FIT. FIT is 318aa in length and the various domains within FIT are indicated above.

The purpose of the work described in this chapter was to accomplish three tasks regarding P6 host-protein interactions. First, I attempted to confirm the interactions of CHUP1, C2CDMT, and FIT with P6 initially identified in the commercial screen initiated by Dr. Schoelz. Second, assuming I could confirm

118 binding, I dissected P6 to identify the domains responsible for interactions with the three host factors. Finally, I determined if there were differences in binding of

CHUP1, C2CDMT, and FIT to P6s from two isolates of CaMV. These two isolates were: CM1841 that is symptomatic on Arabidopsis and D4 that is asymptomatic. It is important to note that the work describing the interactions of

P6 with CHUP1 has already been published, I and Dr. Leisner are co-authors

(Angel, 2013)Furthermore, the P6-C2CDMT interaction study is currently in preparation for publication, Dr. Leisner and I are co-authors on that manuscript as well.

6.3 Materials and methods

Constructions and yeast two-hybrid analysis.

Clones of portions of CHUP1, C2CDMT, the Nicotiana benthamiana homolog of C2CDMT (C2CDMT*), and FIT, were generated by PCR and inserted into Gateway® compatible pJG4-5 by Dr. Schoelz’s laboratory and provided to us. CM1841 P6 and domains D1-D4 cloned into pEG202 were described previously (Li, 2002). Full-length P6 from CaMV isolate D4 was generated by PCR using the primers GWCaMVVI-1F (5’

CACCATGGAGAACATAGAAAAACTCC 3’) and GWCaMVVI-2R (5’

AGATGTCACATCAATCCACTTGCTTTG 3’) and inserted into Gateway® compatible pEG202. Constructions were confirmed by sequencing. Yeast two- hybrid analyses were performed as described in (Hapiak, 2008).

119

6.4 Results

CaMV P6 has been shown to interact with a multitude of host proteins.

To further examine the host proteins interacting with P6, a commercial yeast two- hybrid analysis identified three potential interactors: CHUP1, C2CDMT, and FIT.

Below we provide data regarding these interactions.

The portion of CHUP1 consisting of the N-terminal 115 amino acids was originally identified in the commercial yeast two-hybrid screen to interact with

P6. To confirm this, we performed a separate yeast two-hybrid analysis and found that indeed this portion of CHUP1 does indeed interact with P6.

Biochemical studies performed by our collaborator, also showed that the full- length CHUP1 could also interact with P6 (Angel, 2013).

To determine which portion of P6 interacted with CHUP1, I employed constructs developed previously in our laboratory. P6 contains four self- association domains, termed D1-D4. Together these four domains span >85 percent of P6. Neither D1, nor D3 interacted with CHUP1 as determined by the inability of yeast transformants to grow on media lacking leucine and by the lack of β-galactosidase activity. However, P6 self-association D2 and to a lesser extent D4 both interacted with CHUP1 using the yeast two hybrid system. Based on β-galactosidase activity, D4 bound to CHUP1 about as well as full-length P6, while D2 associated approximately four-fold better.

120

Figure 6-4. Interaction of CaMV P6 and Arabidopsis thaliana CHUP1 protein. A, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity in B. Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 and CHUP1 (1-363aa) numbers to the left of each pair of constructions correspond to the β-galactosidase assays shown in B. The plus and minus signs indicate growth of yeast transformants on (-L) leucine independent plates. B, β- galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A.

I next investigated the ability of the 360 amino acid protein C2CDMT

(C2) to interact with P6. My independent yeast two-hybrid analysis confirmed that C2 interacted with P6 and as based on leucine-independent growth and β- galactosidase activity (Fig. 6-5). Similar to CHUP1, yeast two-hybrid results showed C2CDMT interacted with P6 self-association domains D2 and D4, but not

D1 or D3. D2 showed the highest ®−γαλαχτοσιδασε activity compared to full- length P6 and D4.

Since Nicotiana benthamiana is a host used for many virus experiments and one used for our fluorescence microscopy studies, I investigated the properties of P6 binding to the N. benthamiana homolog of C2CDMT homology

(C2*). Based on my yeast two-hybrid analysis, C2* was able to interact with P6 121

(Fig. 6-6). However, the C2*-full-length P6 interaction was much less efficient than the C2-P6 interaction. Unlike C2, C2* bound well to all four P6 domains

D1-D4. In fact, all four domains bound to C2* better than the full-length P6 did.

Like C2, the P6 domain that C2* bound best was D2.

Figure 6-5. Interaction of CaMV P6 and Arabidopsis thaliana C2CDMT protein. A, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity in B. Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 and C2CDMT (C2); numbers to the left of each pair of constructions correspond to the β-galactosidase assays shown in B. The plus and minus signs indicate growth of yeast transformants on media lacking leucine. B, β-galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A.

122

Figure 6-6. Interaction of CaMV P6 and C2CDMT homolog from Nicotiana benthamiana (C2*). A, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity in B. Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 and numbers to the left of each pair of constructions correspond to the β-galactosidase assays shown in B. The plus and minus signs indicate growth of yeast transformants on media lacking leucine. B, β-galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A.

I also investigated the ability of FIT to interact with P6. My data confirmed the original yeast two-hybrid study as I detected interactions between

FIT with full-length P6 (Fig. 6-7). Similarly to C2*, FIT interacted with all four of the P6 individual self-association domains. Furthermore, FIT interacted more efficiently domain D2 than with the full-length P6 based on β-galactosidase activity. Interactions of the three other domains (D1, D3, and D4) with FIT were comparable with that of the full-length P6 based on β-galactosidase activity.

123

Figure 6-7. Interaction of CaMV P6 and Arabidopsis thaliana FIT protein. A, Schematic diagram of the constructs tested for leucine independent growth and β- galactosidase activity in B. Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; white box, full-length or portions of CaMV P6 and numbers to the left of each pair of constructions correspond to the β-galactosidase assays shown in B. The plus and minus signs indicate growth of yeast transformants on leucine-deficient media. B, β-galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β- galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A.

CaMV isolates vary widely in their ability to infect different hosts

(Anderson, 1992; Schoelz, 1988). Some CaMV isolates like CM1841 are limited to infecting hosts mainly in the Brassicaceae. Other isolates, like D4 are able to infect plants in the Solanaceae as well as the Brassicaceae. Please keep in mind here that D4 in this context is referring to a viral isolate and not to the P6 self- association domain D4. Therefore, it would stand to reason that the host factors that CM1841 P6 interacts with may be quite different or the interactions themselves may have different properties than P6 from another virus, like D4.

Because we had recently cloned the full-length D4 P6 in yeast two-hybrid vectors, we analyzed the interaction of this protein with CHUP1, C2, C2*, and

FIT described in the previous yeast two-hybrid studies. In yeast two-hybrid

124 studies, full-length D4 P6 self-associated very efficiently, based on leucine- independent growth and ®−γαλαχτοσιδασε activity. However, further yeast two-hybrid analyses were unable to detect an interaction of full-length D4 P6 with

CHUP1, C2, C2* or FIT. As initial yeast two-hybrid results showed no interaction between D4 P6 from and the host proteins, I did not carry out further studies to determine if these host proteins could interact with the individual self- association domains of this CaMV strain.

Figure 6-8. Interaction of CaMV D4 strain P6 with Arabidopsis thaliana CHUP1, C2CDMT, and FIT proteins, and C2CDMT Homology protein from Nicotiana benthamiana. A, Schematic diagram of the constructs tested for leucine independent growth and β-galactosidase activity in B. Black box, LexA DBD in pEG202; hatched box, B42 TAD in pJG4-5; grey shaded box, full-length CaMV P6 D4 strain and numbers to the left of each pair of constructions correspond to the β-galactosidase assays shown in B. The plus and minus signs indicate growth of yeast transformants on (-L) leucine independent plates. B, β- galactosidase activity of yeast transformants expressing constructs as represented in A. The bar graph shows average β-galactosidase units for three different experiments along with the standard deviation. Numbers at the bottom correspond to the construction pairs as presented in A.

125

6.5 Discussion

Viruses are obligate parasites, which use a variety of host factors to carry out their life cycles (Haas, 2008; Hapiak, 2008; Love, 2007; Palanichelvam, 2000;

Wang, 2012). Some of these host factors are required for virus intra- and intercellular movement. Specialized plant viral movement proteins aid transit of infectious material through PD. The host cytoskeleton is an important component for intracellular transport of both plant and animal viruses. Movement of TMV involves both the microtubule and actin cytoskeleton.

The two host proteins, CHUP1 and C2CDMT could play a role in CaMV intracellular and/or cell-to-cell movement. Because CHUP1 is a type of myosin

(Oikawa, 2003), it could play a role in transporting viruses along actin filaments within the cell. Preliminary data from our collaborator indicates that C2CDMT apparently binds to PD. Our yeast two-hybrid results showed CaMV P6 interacts in a similar fashion to both CHUP1 and C2CDMT. Both proteins interacted more efficiently with domain D2. D2 of P6 may be a major interface of P6 with host proteins as not only do CHUP1 and C2CDMT bind to this domain but components of the host translational machinery do as well.

Since CHUP1 is a myosin protein the interaction between itself and P6 might explain how IBs move intracellularly along actin microfilaments. We hypothesize that this interaction is key for allowing P6 to move along actin microfilaments via the myosin activity of CHUP1. This could permit P6s to find and interact with one another, which ultimately leads to their aggregation into large IBs. The observation that P6 binds to the CHUP1 coiled-coil domain 126 through D2 is significant. Theoretically P6 could self-associate via D1 or D3 and yet could still simultaneously bind to CHUP1 via D2. This would link IBs to the microfilaments through the actin/chloroplast binding domain of CHUP1.

Interestingly, transgenic plant cells expressing a truncated CHUP1 protein deficient in the ability to move chloroplasts this can also inhibit the transport of

P6 IB complexes.

On the other hand, the function of the interaction C2CDMT with P6 in viral infection is less clear. Our lab has previously shown by yeast two-hybrid that P1 and P6 interact. In addition previous fluorescence microscopy studies showed C2CDMT binds to P1 tubules at the PD. Therefore, we speculate that this interaction could help facilitate the interaction between P1 and P6 to allow for cell-to-cell movement. One can envision a situation where P6 IBs containing virus particles move along actin filaments via CHUP1. Since actin filaments are directed outward towards the cell periphery, CHUP1 could take the IBs there.

Interactions of P6 with C2CDMT at the PD could then aid in the concentration of virus particles near tubules composed of the CaMV movement protein, P1. The proximity of virus could then aid in the binding of P3 on virus particles to the inner surface of P1 tubules. This would then help in the incorporation of virions into P1 tubules.

It is possible that the interaction of P6 with C2CDMT may be important for functions independent of movement. For example, C2CDMT is regulator in programmed cell death triggered by biotic stresses (Costa, 2013). Perhaps P6 binding to C2CDMT modulates cell death responses mediated by the latter 127 protein. This may help the virus to infect the plant host more efficiently and prevent the host from defending itself, leading to a productive infection.

Interestingly, C2CDMT from Arabidopsis thaliana and the homolog of this protein from Nicotiana benthamiana (C2CDMT*) showed different patterns of binding to CaMV P6. While C2CDMT to mainly the D2 and D4 domains of

P6, C2CDMT* bound to all four domains, with an emphasis on D2. This shows that virus-host interactions can be quite different even with the same protein homolog from two different hosts. Interestingly, C2CDMT* interacts more efficiently with P6 than C2CDMT. This is intriguing since N. benthamiana is not a host for the CM1841 isolate of CaMV. We speculate this is due to differences in amino acid sequence between both C2CDMT, which could potential allow for increased binding to P6.

We also believe that the final P6-interacting host protein identified in our study may play a major role in symptom formation. FIT is a transcription factor that regulates the expression of iron transporter genes. Plants harboring FIT mutations are iron deficient and show chlorosis. Since a common symptom induced by CaMV is chlorosis, perhaps P6 inhibits FIT activity leading to iron deficiency. We speculate that because P6 is mainly cytoplasmic, it binds to FIT sequestering the transcription factor in the cytoplasm resulting in down-regulation of iron transporters, iron deficiency and ultimately chlorosis. This is an interesting possibility in light of our findings that the P6 protein of the D4 isolate of CaMV is asymptomatic when expressed as a transgene in Arabidopsis and this protein is also unable to interact with FIT.

128

CaMV isolates can vary substantially in their host ranges and in the symptoms that they induce. Therefore, my yeast two-hybrid studies also examined the interactions of using CM1841 strain, we also looked at the host proteins above with P6 from CaMV isolate D4. The D4 isolate has a broader host range than CM1841, but is asymptomatic on Arabidopsis thaliana. Likewise,

Arabidopsis thaliana expressing CM1841 P6 as a transgene showed virus-like symptoms. However, transgenic Arabidopsis thaliana expressing P6 from the D4 isolate exhibited no symptom formation. While P6 from the D4 isolate of CaMV showed efficient self-association in yeast two-hybrid assays, we were unable to detect interactions of this protein with any of the host proteins described above.

We speculate that the inability of D4 P6 to interact with CHUP1, C2CDMT,

C2CDMT* or FIT, could be due to differences in the amino acid sequence compared to CM1841. Alternatively, D4 P6 self-association may be so efficient that the protein may be locked into a confirmation that prohibits interaction with the host proteins. Regardless of the reason, we speculate that the inability of D4

P6 to interact with FIT may explain the lack of symptoms in D4-infected

Arabidopsis. D4 P6 may not be able to sequester FIT in the cytoplasm, this would permit FIT to normally regulate the transcription of iron transporters leading to no iron deficiency and chlorosis.

129

Chapter 7

Discussion/Future Work

Since viral genomes are small and encode only a few proteins, some animal and plant viruses encode multifunctional proteins to facilitate their life cycle. In animal viruses, the influenza A virus NS1 protein is known to be a multifunctional protein, that inhibits host immune responses as well as modulating viral replication, protein synthesis, and host-cell physiology (Hale,

2008). Like the NS1 protein, the Hepatitis B virus (HBV) X protein modulates a number of activities including: transcription, protein degradation, and signaling pathways (Murakami, 2001). In addition X protein affects cell cycle checkpoints, cell death, and carcinogenesis. Other multifunctional proteins found in animal viruses include rotavirus NSP4, which is involved in disruption of plasma membrane integrity, inhibition of sodium absorption, and remodeling of the cellular microtubule and actin networks (Hu, 2012). In addition NSP4 exhibits a number of activities affecting replication, transcription, and morphogenesis.

Plant viruses also utilize multifunctional proteins. For example, potyviruses encode a multifunctional protein known as helper component- proteinase (HC-Pro). HC-Pro has autoproteolytic activity (Urcuqui-Inchima,

130

1999) and is involved in aphid transmission as well as viral genome replication.

HC-Pro plays a role in both types of viral movement: both cell-to-cell and systemic. Finally, it is a suppressor of posttranscriptional (PTGS) and virus- induced gene silencing (Plisson et al., 2003). Lastly and more importantly,

Cauliflower mosaic virus gene VI product P6 is also multifunctional. CaMV P6 has been implicated in a variety of functions such as; translational transactivation

(TAV), host range control, symptom formation, RNA silencing suppressor activity, viral infectivity, and is the major structural component of inclusion bodies (IBs) (2011; Bonneville, 1989; Cecchini, 1997; Citovsky, 1991; De Tapia,

1993; Haas, 2005; Harries, 2008; Hunter, 2002; Jozwiakowski, 2009; Schoelz,

1986, 1988).

Many animal and plant viruses carry out viral genome replication in the cytoplasm of infected cells (Boon, 2010). Therefore, to increase replication efficiency and protect against host defenses these viruses form IB-like structures.

Depending upon the virus, it can form either membrane-bound or non membrane- bound IBs and these structures may form within the nucleus, or the cytoplasm

(Boon, 2010; Netherton, 2011). My research focused on non membrane-bound

IBs that form within the cytoplasm of infected cells. To study non membrane- bound IBs I used CaMV as our model system.

A major goal of this project was to elucidate the formation of IBs. In addition, this research would provide further understanding towards answering the question if the formation of IBs is required for viral infection or is merely a consequence, due to protein overproduction and aggregation. As mentioned 131 above CaMV P6 is the major IB protein. CaMV IB formation is likely a complex process requiring proper higher order structures of P6 via the four self-association domains. From previous data and the results from my study I believe there are three events that occur in order to form mature functional IBs. In order, I believe these events include: nucleation of P6 IBs, the formation of small IBs (sIBs), and finally, the aggregation of sIBs into mature IBs. Below I will discuss my speculations regarding the three events I believe occur in IB formation (Fig. 7-1).

Early in virus infection, P6 is mainly synthesized off the 19S RNA. This free P6 is then able to enter the nucleus via the importin α-pathway using the

NLSs located in domains D2 and D3 (Haas, 2008). This allows P6 to inhibit the

TasiRNA pathway by binding to DRB4 and disrupting DICER activity (Haas,

2008; Love, 2012). So early in infection the free P6 inhibits a host defense, which is a common strategy used by many viruses. I speculate that as the infection progresses and P6 levels in cells start to rise, then P6s can start to self-associate.

If these interactions involve binding of P6s to the D2 or D3 domains, this interaction would potentially mask the NLSs within P6, keeping it localized to the cytoplasm. Since the N-terminus of P6 is apparently required for IB formation

(Haas, 2005), it is possible that free P6s can bind to P6 N-terminus as it is synthesized from the 19S RNA. The newly synthesized P6s could form a 19S

RNA-ribosome-P6 complex allowing for the free P6s to bind to the complex ultimately leading to the formation of sIBs.

Once the P6 progresses into sIBs I speculate that several different processes occur. First, I believe that the 35S RNA is incorporated into the sIBs. 132

Next, P6 carries out translational transactivation (TAV) activity synthesizing all the viral proteins. Finally, the newly synthesized proteins are incorporated into the sIB. I believe that P7 may play a role in the first process, namely, the incorporation of the 35S RNA into IBs. To date P7s function is still unknown.

What we do know is mutations removing the P7 start codon have a dramatic effect reducing virus infection (Wurch, 1990). Furthermore, P7 is the only protein on the 35S RNA that can be made without the activity of P6 TAV activity.

In addition, from my yeast two-hybrid I showed an interaction between P7-P6.

Therefore, the interaction between P7-P6 might suggest a function for P7. P7 might help bring the 35S RNA into the sIBs. As P7 is being synthesized, I postulate it can be bound to P6 pulling the P7-ribosome-35S RNA complex into the sIBs.

Once the 35S RNA is pulled onto the sIBs, I speculate that all viral proteins are synthesized from the 35S RNA on the periphery of sIBs through P6s

TAV activity. Therefore, P6s ability to bind to ribosomal subunits gives rise to sIB formation and TAV activity. Our data and those of other labs indicate that P6 can bind to all CaMV viral proteins. P6s ability to bind to all viral proteins may indicate two addition functions. Previous results showed P6 binds to P4 suggested this interaction could play a role in virus assembly (Himmelbach).

Therefore, this could suggest P6 acts as a type of chaperone to facilitate proper folding of all viral proteins. If this is true P6 could act in a similar fashion as

HSP90 in the Hepatitis B virus (Cho, 2000; Hu, 2004). HSP90 is required for proper function of the HBV RT. In addition, P6 may prevent proteins aggregating

133 after being synthesized on the 35S RNA and facilitate proper protein-protein interactions. Furthermore, P6s interaction with all viral proteins may also help stabilize IBs. Schoelz et al (1992) previously showed that IBs were more stable in the presence of W260 P2 than CM1841 P2. In addition, I showed that P6 can bind to P2s of both strains (CM1841 and CMW260) in a similar fashion.

Therefore, I believe that it is the difference in CM1841 P2 protein stability rather than its ability to bind P6 that responsible for the difference in IB stability mediated by the P2 of this protein and the W260 counterpart. Therefore, not only could the interaction between P6-P2 help stabilize IBs, this interaction may help facilitate proper folding of P2. We also showed P6 can bind to P3. This interaction is likely to help facilitate the proper folding of P3 to allow for its proper interactions with P1, P2, and P4. If P3 were not properly folded one would speculate virion assembly and virus movement would be impaired.

As previously mentioned, CaMV proteins are apparently synthesized in the sIBs. However, evidence suggests that the proteins present within sIBs are not processed efficiently, if at all. P4 is normally synthesized with an N-terminal sequence that is removed by the P5 protease prior to producing a functional protein (Torruella, 1989). Champagne et al. (2004) reported that P4 harboring its

N-terminal extension is present within sIBs, but not large ones. However, the P4

N-terminal extension apparently does not prevent the formation of virus particles, as they are observed in sIBs. It is likely that the P5 protease is not active within sIBs. Therefore, if the P5 protease is not active in sIBs then the P5 reverse transcriptase is likely not active. Takatsuji et al (1992) reported the proteolytic

134 processing of the P5 protease is essential for reverse transcriptase activation.

These data also suggest that fusion of sIBs into larger ones somehow activates the protease, permitting processing of P5 to activate the reverse transcriptase and removal of the N-terminal extension of P4.

In order for sIBs to fuse into larger IBs, they need to take advantage of host mechanisms to move around and find other sIBs. These mechanisms would most likely involve the host cytoskeleton. IBs have been shown to move along the actin cytoskeleton and smaller ones are apparently more motile than large ones (Harries, 2008). However, a motor protein potentially responsible for this process was identified by collaboration between our laboratory and that of Dr.

James Schoelz (University of Missouri). We showed that P6 can bind to the non- muscle myosin CHUP1. Furthermore, a dominant-negative CHUP1 completely inhibited the movement of P6 IBs in cells. This interaction may help move sIBs intra-cellularly, possibly facilitating fusion to form larger IBs and possibly to the plasmodesmata for cell-to-cell movement (see below).

I hypothesize that sIBs aggregate to form mature IBs, where functional P5 reverse transcriptase can synthesize viral DNA and mature virus particles would be present. At this stage the mature IBs permit processing of the viral proteins and nucleic acids and allow for the assembly of mature structures.

As a consequence of this hypothesis, I also predict that viral propagation would decrease if mature IBs weren’t able to form. Therefore, I predict that viruses harboring point mutations within P6 that disrupt the aggregation of sIB

135 into mature IBs would cause reduced viral propagation. Certain P6 mutants

(E312A, L316A, and R328A) showed a decrease in binding to the N- and C- terminus self-association domains. Furthermore, these same mutants also showed a decrease in IB size suggesting these mutant viruses could not accrete into mature IBs efficiently. These mutants also showed impaired viral propagation in inoculated leaves and systemic infection, and reduction in viral DNA levels.

Hence, my data indicate that certain mutations within the D3b region of P6 appear to impair mature IB formation showed a decrease in viral propagation. Thus, these data suggest that these mutant viruses are unable to properly aggregate into mature IBs. Since these mutants showed decreased viral DNA levels, this suggest that the P5 protease is inactive in these mutants and hence, both reverse transcriptase and P4 protein processing are impaired. Just as predicted. I believe that the mutations within P6 could also disrupt self-association and the ability of

P6 to form higher order structures (such as mature IBs) as efficiently, which in turn could disrupt or weaken its interactions with other host and viral proteins.

Previous studies showed P6 stabilizes P5 when they are transiently expressed in protoplast. In addition, I showed P5MC region can interact with P6 more strongly than the protease domain and full-length P5. These data suggest that P6 may help keep P5 in its active conformation. In addition, these results help further support that P6 stabilizes P5, which could potentially be mediated through P6 functioning as a chaperone protein similarly to HSP90 in HBV.

In addition to P6s interaction with viral proteins during IB formation, P6 can also interact with a multitude of host proteins. As I previously mentioned, for

136

P6 to carry out TAV activity on the 35S RNA, it must interact with host large ribosomal proteins (RL13, RL18, RL24) and eIF3g. In addition, P6 plays a role in RNA silencing suppressing by presumably binding to DRB4. This allows

CaMV to inhibit host defenses and efficiently carry out its life cycle. Moreover,

P6 can also interact with proteins that aid in movement. As previously mentioned, P6IBs can move along actin microfilaments using motor proteins such as CHUP1 (Angel, 2013; Harries, 2008). The ability of P6 to move along actin microfilament would explain how mature IBs reach the plasmodesmata (PD).

Once the IBs are at the PD, C2, a calcium binding protein can bind to P6

(Schoelz, unpublished data). Fluorescence microscopy studies showed C2 can co- localize with P1 at the PD. However, I was unable to detect an interaction between P1 and C2 by yeast two-hybrid. Therefore, it is unclear, if C2s interaction with P6 and/or possibly P1 helps facilitate virus movement. Lastly P6, which plays an important role in symptom formation, was shown to interact with a transcription factor that regulates iron (FE) uptake. Interestingly, CaMV D4 strain is asymptomatic on Arabidopsis thaliana and the P6 of this virus was unable to interact with FIT. In contrast, CM1841 strain induces symptoms on

Arabidopsis thaliana and P6 of this virus was able to interact with FIT.

Therefore, I hypothesize that P6-FIT interaction may be one mechanism by which

P6 causes symptom formation in infected plants.

137

Figure 7-1. Flow chart on CaMV IB formation.

In summary, our data imply that aggregation of sIBs into mature IBs is required for efficient viral replication and propagation. Our data also suggest that sIBs and mature IBs facilitate different processes. Therefore, IB size does play a role in efficient viral propagation. This may explain why CMW260 has a broader 138 host range than CM1841 since CMW260 formed larger IBs than its counterpart

CM1841.

More importantly some of the P6 mutations, characterized by me, that seem to be sequestered in the sIB form, will allow us to further study IB formation and the functions of sIBs. If my hypothesis is correct, that different size IBs have different functions, then it should be possible to detect differences in IB composition. For example, we would expect to observe a higher proportion of unprocessed P4 or P5 for the mutant viruses that we think are trapped in sIBs when compared to viruses that produce normal sized IBs. The P6 mutants that I have characterized could be used to examine unprocessed P4 and P5 proteins and compared to wild type. To do this, we could use electron microscopy and probe for the N-terminus of P4 to determine if it is present. We could also isolate IBs and then perform Western blot analyses to determine if there is a higher proportion of unprocessed P4 in the mutants trapped in sIB formation rather than wild type. We could also analyze P5 proteolytic processing in a similar manner, isolating IBs, performing Western blot analyses and probing for the protease. We expect this to be the case since the mutants that seemed to be trapped in sIB formation should have a decrease in total mature virus particles present in IBs.

However, other factors could explain the reduced infectivity of the P6 mutant viruses. For example, the point mutations that we generated within P6 may disrupt self-association and the ability of P6 to form higher order structures.

This could be important for TAV function since this activity apparently requires appropriate P6 self-association involving D3. The altered IB structure could thus 139 lead to inefficient TAV function, reduced viral protein synthesis and impaired viral genome replication.

140

References

2011. Hepatitis B Information for Health Professionals, in: National Center for

HIV/AIDS, V.H., STD, and TB Prevention (Ed.), Division of Viral

Hepatitis. Centers for Disease Control and Prevention, Atlanta, GA.

Agama, K.A., Beach, S., Schoelz, J., Leisner, S.M., 2002. The 5' third of

Cauliflower mosaic virus gene VI conditions resistance breakage in

Arabidopsis ecotype Tsu-0. Phytopathology 92, 190-196.

Anderson, E.J., Qui, S.G., Schoelz, J.E., 1991. Genetic analysis of determinants of

disease severity and virus concentration in cauliflower mosaic virus.

Virology 181, 647-655.

Anderson, E.J., Trese, A., Sehgal, O., Schoelz, J.E., 1992. Characterization of a

Chimeric Cauliflower Mosaic Virus isolate that is more severe and

accumulates to higher concentrations than either of the strains from which

it was derived. Molecular plant microbe interactions 5, 48-54.

Angel, C.A., Lutz, L., Yang, X., Rodriguez, A., Adair, A., Zhang, Y., Leisner,

S.M., Nelson, R.S., Schoelz, J.E., 2013. The P6 protein of Cauliflower

mosaic virus interacts with CHUP1, a plant protein which moves

chloroplasts on actin microfilaments. Virology 443, 363-374.

Arnold, M.a.P., J., 2009. Rotavirus Antagonism of the Innate Immune Response.

Viruses 1, 1035-1056.

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith,

J.A., Struhl, K., 1993. Current Protocols in Molecular Biology. Greene

141

Publishing Associates and John Wiley & Sons, INC, Cambridge MA

USA.

Back, N.K., Nijhuis, M., Keulen. W., Boucher, C.A., Oude Essink, B.O., Van

Kuilenburg, A.B., Van Gennip, A.H., Berkhout, B., 1996. Reduced

replication of 3TC-resistant HIV-1 variants in primary cells due to a

processivity defect of the reverse transcriptase enzyme. . EMBO 15, 4040-

4049.

Bauer, P., Ling, H.Q., Guerinot, M.L., 2007. FIT, the FER-LIKE IRON

DEFICIENCY INDUCED TRANSCRIPTION FACTOR in Arabidopsis.

Plant Physiology Biochemistry 5, 260-261.

Blanc, S., Cerutti, M., Chaabihi, H., Louis, C., Devauchelle, G., Hull, R., 1993.

Gene II Product of an Aphid-Nontransmissible Isolate of Cauliflower

Mosaic Virus Expressed in a Baculovirus System Possesses Aphid

Transmission Factor Activity. Virology 192, 651-654.

Bonneville, J.M., Sanfacon, H., Futterer, J., Hohn, T., 1989. Posttranscriptional

trans-activation in cauliflower mosaic virus. Cell 59, 1135-1143.

Boon, J.A., Diaz, A., Ahiquist, P., 2010. Cytoplasmic Viral Replication

Complexes. Cell Host & Microbe 8, 77-85.

Broglio, E.P., 1995. Mutational analysis of cauliﬂower mosaic virus gene VI:

changes in host range, symptoms, and discovery of transactivation-

positive, noninfectious mutants. . Mol Plant Microbe Interact 8, 755-760.

142

Brunt, A., Crabtree, K., Dallwitz, M., Gibbs, A., Watson, L., 1999. Viruses of

plants: descriptions and lists from the VIDE database. Oxford University

Press, New York, NY.

Buchner, J., 1999. Hsp & Co. - a holding for folding. Trends Biochem Sci 4, 136-

141.

Bureau, M., Leh, V., Haas, M., Geldreich, A., Ryabova, L., Yot, P., Keller, M.,

2004. P6 protein of Cauliflower mosaic virus, a translation reinitiator,

interacts with ribosomal protein L13 from Arabidopsis thaliana. J. Gen.

Virol. 85, 3765-3775.

Burns, T., Breathnach, S., Cox, N., Griffiths, C., 2010. Virus Infections, Rook's

Textbook of Dermatology. Blackwell Publishing.

Callaway, A., Liu, W., Andrianov, V., Stenzler, L., Zhao, J., Wettlaufer, S.,

Jayakumar, P., Howell, S., 1996. Characterization of Cauliflower mosaic

virus (CaMV) resistance in virus-resistant ecotypes of Arabidopsis. Mol

Plant Microbe Interact 9, 810-818..

Caplan, A., 1999. Hsp90's secrets unfold; new insights from structural and

functional studies. Trends Biochem Sci 9, 262-268.

Cecchini, E., Geri, C., Love, A.J., Coupland, G., Covey, S.N., Milner, J.L., 2002.

Mutations that delay flowering in Arabidopsis de-couple symptom

response from Cauliflower mosaic virus accumulation during infection.

Mol. Plant Pathol. 3, 81-90.

Cecchini, E., Gong, Z., Geri, C., Covey, S., Milner, J.J., 1997. Transgenic

Arabidopsis lines expressing gene VI from cauliflower mosaic virus

143

variants exhibit a range of symptom-like phenotypes and accumulate

inclusion bodies. Molecular plant microbe interactions 10, 1094-1101.

Cerritelli, S., Fedoroff, O., Reid, B., Crouch, R., 1998. A common 40 amino acid

motif in eukaryotic RNases H1 and caulimovirus ORF VI proteins binds

to duplex RNAs. Nucl. Acids Res 26, 1834-1840.

Chakrabarty, R., Banerjee, R., Chung, S.M., Farman, M., Citovsky, V.,

Hogenhout, S.A., Tzfira, T., Goodin, M., 2007. pSITE vectors for stable

integration or transient expression of autofluorescent protein fusions in

plants: probing Nicotiana benthamiana-virus interactions. Mol. Plant-

Microbe Interact. 20, 740-750.

Champagne, J., Benhamou, N., Leclerc, D., 2004. Localization of the N-terminal

domain of cauliflower mosaic virus coat protein precursor. Virology 324,

257-262.

Cheng, R.H., Olson, N., Baker, T., 1992. Cauliflower mosaic virus: A 420 subunit

(T= 7), multilayer structure. Virology 186, 655-668.

Cho, G., Park, S-G., Jung, G., 2000. Localization of HSP90 binding sites in the

human hepatitis B virus polymerase. Biochemical and Biophysical

Research Communications 269, 191-196.

Citovsky, V., and P. Zambryski, 1991. How do plant virus nucleic acids move

through intercellular connections. BioEssays 13, 373-379.

Costa, M., Nobre, M., Becker, J., Masiero, S., Amorim, M., Pereira, L., Coimbra,

S. , 2013. Expression-based and co-localization detection of

144

arabinogalactan protein 6 and arabinogalactan protein 11 interactors in

Arabidopsis pollen and pollen tubes. BMC Plant Biology 13, 1-19.

Covey, S., Hull, R., 1981. Transcription of cauliflower mosaic virus DNA.

Detection of transcripts, properties, and location of the gene encoding the

virus inclusion body protein. Virology 111, 463-474.

Covey, S.N., Hull, R., 1985. Cauliflower mosaic virus: Pathways of infection.

BioEssays 3, 160-163.

Das, D.a.G., M., 2004. The crystal structure of the monomeric reverse

transcriptase from moloney murine leukemia virus. Structure 12, 819-829.

Daubert, S., Routh, G., 1990. Point mutations in cauliflower mosaic virus gene VI

confer host-specific symptom changes. Molecular plant microbe

interactions 3, 341-345.

De Tapia, M., Himmelbach, A., Hohn, T., 1993. Molecular dissection of the

cauliflower mosaic virus translation transactivator. EMBO 12, 3305-3314.

Dixon, L., Jiricny, J., Hohn, T., 1986. Oligonucleotide-directed mutagenesis of

cauliflower mosaic virus DNA using a repair-resistant nucleoside

analogue: identification of an agnogene initiation codon. Gene 41, 225-

231.

Dixon, L.K., Koenig, I., Hohn, T., 1983. Mutagenesis of cauliflower mosaic virus.

Gene 25, 189-199.

Farzadfar, S., Ahoonmanesh, A., Mosahebi, G.H., 2007. Occurrence and

distribution of cauliflower mosaic virus on cruciferous plants in iran. Plant

Pathology Journal 6, 22-29.

145

Gardner, R.C., Howarth, A.J., Hahn, P., Brown-Luedi, M., Shepherd, R.J.,

Messing, J., 1981. The complete nucleotide sequence of an infectious

clone of cauliflower mosaic virus by M13mp7 shotgun sequencing.

Nucleic Acids Res. 9, 2871-2888.

Geri, C., Love, A. J., Cecchini, E., Barrett, S. J., Laird, J., Covey, S. N., Milner, J.

J. , 2004. Arabidopsis mutants that suppress the phenotype induced by

transgene-mediated expression of cauliflower mosaic virus (CaMV) gene

VI are less susceptible to CaMV- infection and show reduced ethylene

sensitivity. Plant Mol Biol 56, 111-124.

Goff, S.P., 1990. Retroviral reverse transcriptase: synthesis, structure, and

function. J. Acquir. Immune Defic. Syndr. 3, 817-831.

Goldbach, R.W., Wellink, J., 1996. Chapter 3: Comoviruses: molecular biology

and replication, in: Harrison, B.D., Murant, A.F. (Eds.), The plant viruses:

polyhedral virions and bipartite RNA genomes. Plenum Press, New York,

pp. 35-75.

Guerra-Peraza, O., De Tapia, M., Hohn, T., Hemmings-Mieszczak, M., 2000.

Interaction of the cauliflower mosaic virus coat protein with the

pregenomic RNA leader. Virology 74, 2067-2072.

Guidasci, T., Mougeot, J.L., Lebeurier, G., Mesnard, J.M., 1992. Processing of

the minor capsid protein of the cauliflower mosaic virus requires a

cysteine proteinase. Res. Virol. 143, 361-370.

Gyuris, J., Golemis, E., Chertkov, H., Brent, R., 1993. Cdi1, a human G1 and S

phase protein phosphatase that associates with Cdk2. Cell 75, 791-803.

146

Haas, G., Azevedo, J., Moissiard, G., Geldreich, A., Himber, C., Bureau, M.,

Fukuhara, T., Keller, M., Voinnet, O., 2008. Nuclear import of CaMV P6

is required for infection, and suppression of the RNA silencing factor

DRB4. EMBO, J. 27, 2102-2112.

Haas, M., Bureau, M., Geldreich, A., Yot, P., Keller, M., 2002. Cauliflower

mosaic virus: still in the news. Mol. Plant Pathol. 3, 419-429.

Haas, M., Geldreich, A., Bureau, M., Dupuis, L., Leh, V., Vetter, G., Kobayashi,

K., Hohn, T., Ryabova, L., Yot, P., Keller, M., 2005. The open reading

frame VI product of Cauliflower mosaic virus is a nucleocytoplasmic

protein: its N terminus mediates its nuclear export and formation of

electron-dense viroplasms. Plant Cell 17, 927-943.

Hale, B.G., Randall, R., Ortin, J., Jackson, D., 2008. The multifunctional NS1

protein of influenza A viruses. General Virology 89, 2359-2376.

Hapiak, M., Li, Y., Agama, K., Swade, S., Okenka, G., Falk, J., Khandekar, S.,

Raikhy, G., Anderson, A., Pollock, J., Zellner, W., Schoelz, J., Leisner. S.,

2008. Cauliflower mosaic virus gene VI product N-terminus contains

regions involved in resistance-breakage, self-association and interactions

with movement protein. Virus Research 138, 119-129.

Harries, P.A., Palanichelvam, K., Yu, W., Schoelz, J.E., Nelson, R.S., 2009. The

Cauliflower mosaic virus protein P6 forms motile inclusions that traffic

along actin microfilaments and stabilize microtubules. Plant Physiol. 149,

1005-1016.

147

Hebrard, E., Drucker, M., Leclerc, D., Hohn, T., Uzest, M., Froissart, R., Strub, J-

M., Sanglier, S., Van Dorsselaer, A., Padilla, A., Labesse, G., Blanc, S.,

2001. Biochemical characterization of the helper component of

Cauliflower mosaic virus. Virology 75, 8538-8546.

Himmelbach, A., Chapdelaine, Y., Hohn T., 1996. Interaction between

cauliflower mosaic virus inclusion body protein and capsid protein:

implication for viral assembly. Virology 217, 147-157.

Hohn, T., 2013. Plant pararetroviruses: interactions of cauliflower mosaic virus

with plants and insects. Current Opinion in Virology 3, 1-10.

Hohn, T., Futterer, J., 1997. The proteins and functions of plant pararetroviruses:

knowns and unknowns. Cr. Rev. Plant Sci. 16, 133-161.

Hohn, T., Hohn, B., Pfeiffer, P., 1985. Reverse transcription in CaMV. Trends

Biochem Sci 10, 205-209.

Hu, J., Flores, D., Toft, D., Wang, X., Nguyen, D., 2004. Requirement of heat

shock protein 90 for human hepatitis B virus reverse transcriptase

function. Virology 78, 13122-13131.

Hu, L., Crawford, S., Hyser, J., Estes, M., Prasad, V., 2012. Rotavirus non-

structural proteins: structure and function. Current Opinion Virology 2, 1-

Huang, Z., Han, Y., Howell, S., 2001. Effects of movement protein mutations on

the formation of tubules in plant protoplasts expressing a fusion between

the green fluorescent protein and cauliflower mosaic virus movement

protein. Molecular Plant-Microbe Interact. 14, 1026-1031.

148

Hull, R., 2002. Matthews' Plant Virology, fourth ed. Academic Press, San Diego.

Hull, R., Covey, S., 1983. Does cauliflower mosaic virus replicate by reverse

transcriptase. Trends Biochem Sci 8, 119-121

Hull, R., Covey, S., Dale, P., 2000. Genetically modified plants and the 35S

promoter: assessing the risks and enhancing the debate. Microbial Ecology

in Health and Disease 12, 1-5.

Hunter, P.J., Jones, J.E., Walsh, J.A., 2002. Involvment of Beet western yellos

virus, Cauliflower mosaic virus, and Turnip mosaic virus in internal

disorders of stored white cabbage. Phytopathology 92, 816-826.

Jacquot, E., Geldreich, A., Keller, M., Yot, P., 1998. Mapping regions of the

cauliflower mosaic virus ORF III product required for infectivity.

Virology 242, 395-402.

Jones, S., Hu, H., 2013. Hepatitis B virus reverse transcriptase: diverse functions

as classical and emerging targets for antiviral intervention. Emerging

Microbes and Infections 2, 1-9.

Jozwiakowski, S., Connolly, B., 2009. Plasmid-based lacZa assay for DNA

polymerase fidelity: application to archaeal family-B DNA polymerase.

Nucleic Acids Research 37, 1-7.

Kaushik, N., Chowdhury, K., Pandey, V.N., Modak, M.J., 2000. Valine of the

YVDD motif of moloney murine leukemia virus reverse transcriptase: role

in the fidelity of DNA synthesis. Biochemistry 39, 5155-5165.

Kim, Y.-C., Kim, S-Y., Choi, D., Ryu, C-M., Park, J., 2008. Molecular

characterization of a pepper C2 domain-containing SRC2 protein 149

implicated in resistance against host and non-host pathogens and abiotic

stresses. Planta 227, 1169-1179.

Knipe, D., 1990a. Virus-host interactions, in: Fields, B.N., Knipe, D.M. (Eds.),

Fields Virology. Raven Press, New York, pp. 293-316.

Knipe, D., 1990b. Virus-host interactions. Raven Press, New York.

Kobayashi, K., Hohn, T., 2003. Dissection of Cauliflower mosaic virus

transactivator/viroplasmin reveals distinct essential functions in basic

virus replication. J. Virol. 77, 8577-8583.

Kobayashi, K., Tsuge, S., Nakayashiki, H., Mise, K., Furusawa, I., 1998.

Requirement of cauliflower mosaic virus open reading frame VI product

for viral gene expression and multiplication in turnip protoplasts.

Microbiol. Immunol. 42, 377-386.

Kohl, N., Emini, E., Schleif, W., Davis, L., Heimbach, J., Dixon, R., Scolnick, E.,

Sigal, S., 1989. Active human immunodeficiency virus protease is

required for viral infectivity. PNAS 85, 4686-4690.

Koonin, E., Senkevich, T., Dolja, V., 2006. The ancient virus world and evolution

of cells. Biology Direct, 1-29.

Le Grice, S.F., Mills, J., Mous, J., 1988. Active site mutagenesis of the AIDS

virus protease and its alleviation by trans complementation. EMBO 8,

2547-2553.

Leclerc, D., Chapdelaine, Y., Hohn, T., 1999. Nuclear Targeting of the

Cauliflower Mosaic Virus Coat Protein. Virology 73, 553-560.

150

Leclerc, D., Stavolone, L., Meier, E., Guerra-Peraza, O., Herzog, E., Hohn, T.,

2001. The product of ORF III in cauliflower mosaic virus interacts with

the viral coat protein through its C-terminal proline rich domain. Virus

Genes 22, 159-165.

Leh, V., E. Jacquot, A. Geldreich, T. Hermann, D.Leclerc, M. Cerutti, P.Yot, and

M. Keller, S. Blanc, 1999. Aphid transmission of cauliflower mosaic virus

requires the viral PIII protein. EMBO 18, 7077-7085.

Leh, V., Jacquot, E., Geldreich, A., Haas, M., Blanc, S., Keller, M., Yot, P., 2001.

Interaction between the Open Reading Frame III Product and the Coat

Protein Is Required for Transmission of Cauliflower Mosaic Virus by

Aphids. Virology 75, 100-105.

Leh, V., Yot, P., Keller, M., 2000. The cauliflower mosaic virus translational

transactivator interacts with the 60S ribosomal subunit protein L18 of

Arabidopsis thaliana. Virology 266, 1-7.

Li, Y., Leisner, S.M., 2002. Multiple domains within the Cauliflower mosaic

virus gene VI product interact with the full-length protein. Mol. Plant-

Microbe Interact. 15, 1050-1057.

Love, A., Geri, C., Laird, J., Carr, C., Yun, B-W., Loake, G., Tada, Y.,

Sadamamdom, A., Milner, J., 2012. Cauliflower mosaic virus Protein P6

Inhibits Signaling Responses to Salicylic Acid and Regulates Innate

Immunity. PLOS One 7, 1-12.

151

Love, A.J., Laird, J., Holt, J., Hamilton, A.J., Sadanandom, A., Milner, J.J., 2007.

Cauliflower mosaic virus protein P6 is a suppressor of RNA silencing. J.

Gen. Virol. 88, 3439-3444.

Love, A.J., Yun, B., Laval, V., Loake, G., Milner, J., 2005. Cauliflower mosaic

virus, a Compatible Pathogen of Arabidopsis, Engages Three Distinct

Defense-Signaling Pathways and Activates Rapid Systemic Generation of

Reactive Oxygen Species. . Plant Physiology 139, 935-948.

Martelli, G.P., Russo, M., 1977. Plant Virus Inclusion Bodies. Adv. Virus Res.

21, 175-266.

Mason, W., Taylor, J., Hull, R., 1987. Retroid virus genome replication. Adv

Virus Res 32, 35-96.

Mattanovich, D., Rucker, F., da Camara Machado, A., Laimer, M., Regner, F.,

Steinkellner, H., Himmler, G., Katinger, H., 1989. Efficient transformation

of Agrobacterium spp. by electroporation. Nucleic Acids Res. 17, 6747.

Matthews, R.E.F., 1991. Plant Virology, Third ed. Academic Press, San Diego.

Mazzolini, L., R.Dabos, S. Constantin, P.Yot., 1989. Further evidence that

viroplasms are the site of cauliflower mosaic virus genome replication by

reverse transcription during viral infection. General Virology 70, 3439-

3449.

Mervis, R.J., Ahmad, N., Lillehoj, E.P., Raum, M.G., Salazar, F.H., Chan, H.W.,

Venkatesan, S., 1988. The gag gene products of human immunodeficiency

virus type 1: alignment within the gag open reading frame, identification

152

of posttranslational modifications, and evidence for alternative gag

precursors. Virology 62, 3993-4002.

Morissette, G., Flamand, L., 2010. Herpesviruses and Chromosomal Integration.

Virology 84, 12100-12109.

Murakami, S., 2001. Hepatitis B virus X protein: a multifunctional viral regulator.

Gastroentrerol 36, 651-660.

Nakayashiki, H., Tsuge, S., Kobayashi, K., Okuno, T., Furusawa, I., 1993.

Reasons for the low accumulation level of aphid transmission factor

protein in infected leaves with an aphid-non-transmissible cauliflower

mosaic virus isolate, CM1841. J. Gen. Virol. 74, 2469-2472.

Nalefski, E., Falke, J., 1996. The C2 domain calcium-binding motif: Structural

and functional diversity. Protein Science 5, 2375-2390.

Netherton, C.L., Wileman, T., 2011. Virus factories, double membrane vesicles

and viroplasm generated in animal cells. Current Opinion Virology 5, 381-

387.

Oikawa, K., Kasahara, M., Kiyosue, T., Suetsugu, N., Kagawa, T., Takahashi, F.,

Kaegae, T., Niwa, Y., Kadota, A., Wada, A., 2003. Chloroplast unusual

positioning 1 is essential for proper chloroplast positioning. Plant Cell 15,

2805-2815.

Oikawa, K., Yasmasato, A., Kong, S-G., Kasahara, M., Nakai, M.,Takahashi, F.,

Ogura, Y., Kagawa, T., Wada, A., 2008. Chloroplast outer envelope

protein CHUP1 is essential for chloroplast ancharage to the plasma

membrane and chloroplast movement. Plant Physiology 148, 829-842.

153

Palanichelvam, K., Cole, A.B., Shababi, M., Schoelz, J.E. , 2000. Agroinfiltration

of Cauliflower mosaic virus gene VI elicits hypersensitive response in

Nicotiana species. Molecular plant microbe interactions 13, 1275-1279.

Park, H.-S., Himmelbach, A., Browning, K.S., Hohn, T., Ryabova, L.A., 2001. A

plant viral reinitiation factor interacts with the host translational

machinery. Cell 106, 723-733.

Patton, J.T., Silverstri, L.S., Tortorici, M.A., Vasquez-Del Carpoi, R.,

Taraporewala, Z.F., 2006. Rotavirus Genome Replication and

Morphogenesis: Role of the Viroplasm, Reoviruses: Entry, Assembly and

Morphogenesis. Springer-Verlag, Berlin Heidelberg, pp. 169-187.

Perbal, M., Thomas, L., Maule, A., 1993. Caulflower mosaic virus gene I product

(P1) forms tubular structures which extend from the surface of infected

protoplasts. Virology 195, 281-285.

Pesavento, J.B., Crawford, S.E., Estes, M.K., Prasad, V., 2006. Rotavirus

Proteins: Structure and Assembly, Reoviruses: Entry, Assembly and

Morphogenesis. Spinger-Verlag Berlin Heidelberg, pp. 189-219.

Pfeiffer, P., and Hohn, T., 1983. Involvement of reverse transcription in the

replication of cauliflower mosaic virus: a detailed model and test of some

aspects. Cell 33, 781-789.

Plisson, C., Drucker, M., Blanc, S., German-Retana, S., Le Gall, O., Thomas, D.,

Bron, P., 2003. Structural characterization of HC-Pro, a plant virus

multifunctional protein. The Journal of biological chemistry 278, 23753-

23761.

154

Plyte, S.E., Kneale, G.G., 1991. The role of tyrosine residues in the DNA-binding

site of the Pf1 gene 5 protein. Protein Eng. 4, 553-1560.

Pooggin, M., Futterer, J., Skryabin, K., Hohn, T., 2001. Ribosome shunt is

essential for infectivity of cauliflower mosaic virus. Proc. Natl. Acad. Sci

98, 886-891.

Qui, S.G., Wintermantel, W. M., Sha, Y., Schoelz, J.E., 1997. Light-dependent

systemic infection of solanaceous species by cauliflower mosaic virus can

be conditioned by a viral gene encoding an aphid transmission factor.

Virology 227, 180-188.

Rothnie, H., Y. Chapdelaine, and T. Hohn, 1994. Pararetroviruses and

retroviruses: A comparative review of viral structure and gene expression

strategies. Advanced Virus Res. 44, 1-67.

Ryabova, L., Pooggin, M., Dominguez, D., Hohn, T., 2000. Continuous and

Discontinuous Ribosome Scanning on the Cauliflower Mosaic Virus 35 S

RNA Leader Is Controlled by Short Open Reading Frames. Biological

Chemistry 275, 37278-37284.

Ryabova, L.A., Pooggin, M.M., Hohn, T., 2002. Viral strategies of translation

initiation, ribosomal shunt and reinitiation. Prog. Nucleic Acid Res. 72, 1-

39.

Schardl, C.L., Byrd, A.D., Benzion, C., Altschuler, M.A., Hildebrand, D.F., Hunt,

A.G., 1987. Design and construction of a versatile system for the

expression of foreign genes in plants. Gene 61, 1-11.

155

Schoelz, J., Shepherd, R.J., Daubert, S., 1986. Region VI of cauliflower mosaic

virus encodes a host range determinant. Mol. Cell. Biol. 6, 2632-2637.

Schoelz, J.E., Goldberg, K.-B., Kiernan, J., 1991. Expression of cauliflower

mosaic virus (CaMV) gene VI in transgenic Nicotiana biglovii

complements a strain of CaMV defective in long-distance movement in

nontransformed N. biglovii. Molecular plant microbe interactions 4, 350-

355.

Schoelz, J.E., Shepherd, R.J., 1988. Host range control of cauliflower mosaic

virus. Virology 162, 30-37.

Schoelz, J.E., Wintermantel, W.M., 1993. Expansion of viral host range through

complementation and recombination in transgenic plants. Plant Cell 5,

1669-1679.

Scholthof, K., Adkins, S., Czosnek, H., Palukaitis, P., Jacquot, E., Hohn, T.,

Hohn, B., Saunders, K., Candresse, T., Ahlquist, P., Hemenway, C.,

Foster, G., 2011. Top 10 plant viruses in molecular plant pathology.

Molecular plant pathology 12, 938-954.

Shalla, T., Shepherd, R., Petersen, L., 1980. Comparative cytology of nine

isolates of cauliflower mosaic virus. Virology 102, 381-388.

Shepherd, R.J., 1979. DNA plant viruses. Annual Review of Plant Physiology 30,

405-423.

Shepherd, R.J., 1981. Cauliflower mosaic virus. Commonw. Myol. Inst./Assn.

Appl. Biol 243, 1-6.

156

Stavolone, L., Villani, M. E., Leclerc, D., Hohn, T., 2005. A coiled-coil

interaction mediates cauliflower mosaic virus cell-to-cell movement.

Proceedings of the National Academy of Sciences 102, 6219-6224.

Stoye, J., 2012. Studies of endogenous retroviruses reveal a continuing

evolutionary saga. Nature Reviews Microbiology 10, 395-406.

Stratford, R., Covey, S., , 1989. Segregation of cauliflower mosaic virus

movement protein RNA-binding domain. Virology 172, 451-459.

Takahashi, R., Shimosaka, E. , 1997. cDNA sequence analysis and expression of

two cold-regulated genes in soybean. . Plant Science 123, 93-104.

Takatsuji, H., Yamauchi, H., Watanabe, S.-I., Kato, H., Ikeda, J.-E., 1992.

Cauliflower mosaic virus reverse transcriptase. J. Biol. Chem. 267, 11579-

11585.

Tasara, T., Amacker, M., Hubscher, U. , 1999. Intramolecular chimeras of the p51

subunit between HIV-1 and FIV reverse transcriptases suggest a

stabilizing function for the p66 subunit in the heterodimeric enzyme.

Biochemistry 38, 1633–1642.

Team, R.D.C., 2005. R: A language and environment for statistical computing,

reference 2† index version 2.2.1. . R Foundation for Statistical Computing,

Vienna, Austria.

Thomas, C., Maule, A.J., 1995. Identification of the cauliflower mosaic virus

movement protein RNA binding domain. Virology 206, 1145-1149.

157

Thomas, C., Maule, A.J., 1999. Identification of inhibiory mutants of cauliflower

mosaic virus movement protein function after expression in insect cells.

Virology 9, 7886-7890.

Torruella, M., Gordon, K., Hohn, T., 1989. Cauliflower mosaic virus produces an

aspartic proteinase to cleave its polyproteins. EMBO J. 8, 2819-2825.

Turner, D.S., McCallum, D.G., Covey, S. N., 1996. Roles of the 35S promoter

and multiple overlapping domains in the pathogenicity of the

pararetrovirus cauliflower mosaic virus. Virology 70, 5414-5421.

Urcuqui-Inchima, S., Walter, J., Drugeon, G., German-Retana, S., Haenni, A.,

Candresse, T., Bernardi, F., Le Gall, O., 1999. Potyvirus helper

component-proteinase self-interaction in the yeast two-hybrid system and

delineation of the interaction domain involved. Virology 258, 95-99.

Wakefield, J., Jablonski, S., Morrow, C., 1992. In vitro enzymatic activity of

human immunodeficiency virus type 1 reverse transcriptase mutants in the

highly conserved YMDD amino acid motif correlates with the infectious

potential of the proviral genome. Virology 66, 6806-6812.

Wang, R.Y., Li, K., 2012. Host factors in the replication of positive-strand RNA

viruses. Chang Gung Med 35, 111-124.

Wintermantel, W.M., Anderson, E. J., Schoelz, J.E. , 1993. Identification of

domains within gene VI of cauliflower mosaic virus that influence

systemic infection of Nicotiana biglovii in a light-dependent manner.

Virology 196, 789-798.

158

Wlodawer, A., Miller, M., Jaskolski, M., Sathyanarayana, B., Baldwin, E., Weber,

I., Selk, L., Clawson, L., Schneider, J., Kent, S.B., 1989. Conserved

folding in retroviral proteases: crystal structure of a synthetic HIV-1

protease. Science 245, 616-621.

Wurch, T., Kirchherr, D., Mesnard, J.-M., Lebeurier, G., 1990. The cauliflower

mosaic virus open reading frame VII product can be expressed in

Saccharomyces cerevisiae but is not detected in infected plants. J. Virol.

64, 2594-2598.

Wurch, T., Kirchherr, D., Mesnard, J-M., Lebeurier, G., 1990. The cauliflower

mosaic virus open reading frame VII product can be expressed in

Saccharomyces cerevisiae but is not detected in infected plants. Virology

64.

Xiong, Y.a.E., T., 1990. Origin and evolution of retroelements based upon their

reverse transcriptase sequences. EMBO 9, 3353-3362.

Zijstra, C., Scharer-Hernandez, N., Gal, S., Hohn, T., 1996. Arabidopsis thaliana

expressing the cauliflower mosaic virus ORF VI transgene has a late

flowering phenotype. Virus Genes 13, 5-17.

159

Appendix A

Pelargonium flower break virus coat protein self-associates and binds to the p7 movement protein

A.1 Abstract

Pelargoniums are a staple crop of the floriculture industry. A number of viruses infect these plants causing a variety of damage. Two common pathogens are Pelargonium flower break virus (PFBV) and the related Pelargonium line pattern virus (PLPV). In this paper, the interactions of the PFBV and PLPV coat proteins (CPs) were investigated. PFBV CP was found to self-associate. CP was divided into three structural domains and examined for self-association. All three

PFBV CP domains self-associated and interacted with each of the others, as well as with the full-length PFBV CP. Investigation of the PLPV CP self-interactions also showed a similar pattern. Since PFBV and PLPV are found in co-infections, interactions amongst their CPs were investigated. Full-length PFBV CP interacted with full-length PLPV CP. However binding among the specific interaction domains differed. PFBV CP was also found to interact with the p7 movement protein. Additional studies showed that the p7 protein self-associated,

160 which was mainly mediated through the C-terminal region. Taken together, these data suggest that the PFBV CP is a highly versatile protein with respect to its interactions, binding to itself, to p7, and to the CP of a related virus.

A.2 Introduction

Pelargonium species are important floriculture crops

(http://www.nass.usda.gov/Statistics_by_State/Ohio/Publications/Reports_by_Titl e/flor2012.pdf). Carmoviruses, such as Pelargonium flower break virus (PFBV) and Pelargonium line pattern virus (PLPV), are prevalent pathogens of these ornamental plants (Alonso and Borja, 2005; Bouwen and Maat, 1992; Brunt et al.,

1999). PFBV symptoms range from stunting of the plants and/or chlorotic spots on leaves to white flower break (Brunt et al., 1999). PLPV induces chlorotic/greenish lesions and line patterns on leaves (Brunt et al., 1999; Castano and Hernandez, 2005, 2007).

PFBV particles are icosahedral, each containing a positive-sense, single- stranded, unipartite RNA genome (Brunt et al., 1999; Rico and Hernandez, 2004).

The 3,923 nucleotide PFBV genome contains five open reading frames coding for

(in order from 5’ to 3’): a 27/86 kDa (p27/p86) replicase protein, two movement proteins (MPs) that are 7 and 12 kDa in size (termed p7 and p12, respectively), and a single 38 kDa capsid protein (CP).

p7 and p12 are required for virus cell-to-cell movement (Martinez-Turino and Hernandez, 2011) and are synthesized from a 1.7 kb sub-genomic RNA (Rico and Hernandez, 2009). PFBV p7 shows a high degree of amino acid sequence 161 conservation with other carmoviruses (Rico et al., 2006). Carmoviral p7 equivalent proteins are typically organized into three regions: a non-ordered, poorly-conserved N-terminal region; a central well-conserved α-helical region that functions as an RNA-binding domain (RBD); and a conserved C-terminal region with β-sheet folding potential (Akgoz et al., 2001; Marcos et al., 1999;

Navarro et al., 2006; Vilar et al., 2005; Vilar et al., 2002). Using the RBD as a landmark (Vilar et al., 2005; Vilar et al., 2002), it is possible to distinguish these three regions within PFBV p7.

The PFBV CP is synthesized from a 1.4 kb transcript (Rico and

Hernandez, 2009). The PFBV CP shares motifs with other carmoviral CPs, which are typically subdivided into an N-terminal RNA-binding domain (R), a central shell domain (S), and a C-terminal protruding domain (P) (Carrington et al., 1987;

Lommel et al., 2005). In addition to structurally protecting the viral genome within a virus particle, carmoviral CPs have other functions. For example, the

Turnip crinkle virus (TCV) CP is required for systemic movement (Hacker et al.,

1992). The P domain of the Melon necrotic spot virus (MNSV) CP is important for vector transmission (Ohki et al., 2010). Furthermore, this same region of the

Cucumber necrosis virus CP functions as a chloroplast transit signal that is apparently important for infection (Xiang et al., 2006). Finally, the PFBV CP has been shown to be a silencing suppressor (Martinez-Turino and Hernandez, 2009).

PFBV often can be found in mixed infections with PLPV (Alonso and

Borja, 2005; Bouwen and Maat, 1992). In one European study, the percentage of

162 plants infected with both viruses was greater than that of plants infected with

PFBV alone (Alonso and Borja, 2005). In that study, PLPV was the most common virus identified. Because PLPV is also a member of the Tombusviridae, it produces virus particles, and has a genome structure similar to PFBV (Castano and Hernandez, 2005). The 3,884 nt PLPV genome encodes a 27/87 kDa replicase protein at the 5’ end. Next are two small open reading frames of 7 kDa and 6 kDa that are putative MPs. A 37 kDa CP is encoded at the 3’ end of the

PFBV genome.

Even though PFBV and PLPV are similar, the two viruses are quite distinct. For example, the PFBV and PLPV CPs share only 33.5% sequence identity, even though the three domains (R, S, and P) can be easily distinguished.

An examination of a population of PFBV genomes indicates that the three domains of the CP show different degrees of sequence heterogeneity (Castano et al., 2011). However, for PFBV CP, the R domain harbored the lowest percentage of polymorphic positions, followed by S and then with P showing the highest amount.

In the study described here, we determined the prevalence of viruses in an

American collection of Pelargonium species. We then examined the interactions of the PFBV CP because of its multiple functions. We tested self-interaction and identified the domains involved. Because the CP has been implicated in virus systemic spread, we examined its interactions with the MPs. Finally, because

PFBV and PLPV can be found in co-infections, we studied the interactions of CPs for these two viruses. 163

A.3 Material and Methods

Plant propagation, RNA isolation and reverse transcriptase-polymerase chain reaction (RT-PCR).

Pelargonium Accessions: 223, 230, 239, 264, 337, 344, 364, 366, 382,

453, 479, 485, 569, 604, 799, 1063, 1084, 1107, 1281, 1321, 1325, 1482, 1724, and 1728 were obtained from the Ornamental Plant Germplasm Center (OPGC),

Columbus, OH (http://opgc.osu.edu/home). These plants were propagated in

Sungro® soilless media in a greenhouse with natural lighting, at a temperature of

22°C.

Total RNA was isolated from OPGC Accessions using the Concert Plant

RNA Reagent (Invitrogen Corporation, Carlsbad, CA), according to the manufacturer’s specifications. Isolated RNA was used as a template to synthesize cDNA employing the M-MLV (Promega Corporation, Madison, WI) reverse transcriptase and random hexamer primers (from Integrated DNA Technologies

(IDT), Coralville, IA) according to the manufacturer’s specifications. Note: all primers used in this paper were purchased from IDT and primer information is provided in Supplementary Table A-1.

The cDNA was then used for virus detection. PCR reactions (25 ml) used the cDNA and GoTaq (Promega Corporation), according to the manufacturer’s specifications. The PCR program employed an initial denaturation step at 94°C for 2 min, followed by 28 cycles of: denaturation, annealing and elongation (30 sec for each step) at 94°C; 55°C; and 72°C, respectively. The reactions were then

164 examined by electrophoresis through an agarose gel. The DNA bands were visualized by ethidium bromide staining and UV illumination.

For the detection of PFBV, we employed the PFBV-1F and PFBV-1R primers; for PLPV, we used the PLPV-1F and PLPV-1R primers; for Tomato ringspot virus (ToRSV), we used the ToRSV-1F and ToRSV-1R primers. We also tested the Accessions by RT-PCR for a variety of other potential

Pelargonium pathogens: Artichoke Italian latent virus, Cucumber mosaic virus,

Elderberry latent virus, Moroccan pepper virus, Pelargonium leaf curl virus,

Pelargonium zonate spot virus, Tobacco ringspot virus, and Tomato spotted wilt virus.

PFBV and PLPV CP constructions.

Total RNA, isolated from Pelargonium accession OPGC 453, was converted to cDNA as above. A full-length fragment encoding the PFBV CP was generated by PCR, using primers FCP-F and FCP-R (Supplementary Fig. A-1).

The full-length PLPV CP coding sequence was produced in a similar fashion employing the LCP-F and LCP-R primers, using cDNA generated as above from

RNA isolated from Pelargonium accession number OPGC 230. Note: all CP and p7 constructions generated for this paper, were confirmed by sequencing

(University of Michigan Sequencing Core, Ann Arbor MI). The full-length

PFBV and PLPV CP coding sequences were then inserted into the EcoRI sites of the pMAL-c2X (New England Biolabs) and pGST41a Escherichia coli expression vectors for pull-down analyses. The full-length PFBV CP sequence was inserted

165 into the EcoRI and XhoI sites of both of the yeast two-hybrid vectors pEG202 and pJG4-5 (Ausubel et al., 1993; Gyuris et al., 1993). The full-length PLPV CP sequence was inserted into the with EcoRI and NcoI sites of the yeast two-hybrid vector pEG202. The full-length PLPV CP sequence was also inserted into a modified form of pJG4-5 (pJG4-5Mod, that contains extra restriction enzyme sites for BglII, NcoI and SmaI between the EcoRI and XhoI sites; (Raikhy et al.,

2011a)) at the EcoRI and NcoI sites. The PFBV and PLPV CP genes differed slightly from the published sequences (GenBank Accession numbers NC_005286 and NC_007017) at 5 (V14A, K38N, I47M, V53I, and A310V) and 9 amino acid positions (A8V, S213N, T236N, T262N, I266V, E283V, K295R, T301A and

D302E), respectively.

The PFBV R domain-coding region (encoding amino acids 1-73) was PCR amplified using primers FCP-F and FCP-1R, while the portion encoding the S domain (encoding amino acids 70-240) was generated using the FCP-2F and

FCP-2R primers, and the P domain-coding region (encoding amino acids 241-

346) was generated using the FCP-3F and FCP-R primers. The PCR products were then inserted into the EcoRI and XhoI sites of pEG202 and pJG4-5.

The PLPV R domain-coding region (encoding amino acids 1-62) was PCR amplified using primers LCP-F and LCP-1R, while the portion encoding the S domain (encoding amino acids 63-233) was generated using the LCP-2F and

LCP-2R primers, and the P domain-coding region (encoding amino acids 234-

339) was generated using the LCP-3F and LCP-R primers (Supplementary Fig. A-

1). DNA fragments encoding the PLPV CP R and P domains were inserted into 166 the EcoRI and NcoI, sites of pEG202 and pJG4-5Mod, while the S domain was inserted into the NcoI and XhoI sites of these two vectors.

PFBV p7 constructions.

Full-length PFBV p7 was PCR amplified using primers p7-F and p7-R and cDNA generated from RNA purified from OPGC 453 as above

(Supplementary Fig. A-2). The sequence of the full-length clone of the PFBV p7 gene used in our experiments was identical to GenBank NC_005286, except for a single T to C substitution at nucleotide number 39 within the p7 coding region that did not result in an amino acid change. The p7 gene was then inserted into the BamHI and XhoI sites within pEG202, or into the BglII and XhoI sites of pJG4-5Mod. Full-length p7 was also amplified with the p7GW-r and p7-R primers and inserted into Gateway® compatible protein expression destination vectors derived from pMAL-c2X and pDEST5 (Raikhy et al., 2011).

The PFBV p7 N-terminal domain coding region (encoding amino acids 1-

23) was PCR amplified using primers p7-F and p7D1-R, while the portion encoding the central RNA-binding domain (encoding amino acids 20-42) was generated using the p7D2-F and p7D2-R primers, and the p7 C-terminal coding region (encoding amino acids 39-68) was generated using the p7D3-F and p7-R primers (Supplementary Fig. A-2). Following amplification, the central and C- terminal fragments were inserted into the NcoI and XhoI sites within pEG202 and pJG4-5mod, while the N-terminal fragment was inserted into the BamHI and NcoI sites within those vectors.

167

Yeast two-hybrid analysis.

Yeast two-hybrid analyses were performed as describe previously (Raikhy et al., 2011a). Briefly yeast lines were first established containing the pEG202 plasmid constructions described above. Constructions were introduced into yeast strain EGY48 (Ausubel et al., 1993; Gyuris et al., 1993) harboring pSH18-34 (a

β-galactosidase reporter plasmid) by lithium acetate transformation. The pJG4-5 or pJG4-5mod plasmids were then subsequently introduced into these yeast lines.

Yeast transformants were evaluated for their ability to grow on media lacking leucine and for β-galactosidase activity as described by (Raikhy et al., 2011a).

Cauliflower mosaic virus gene VI product (P6) expressed from pEG202 or pJG4-

5, used as a specificity control for yeast two-hybrid studies, was described previously (Li and Leisner, 2002).

Pull-down experiments.

To confirm protein interactions biochemically, maltose-binding protein

(MBP) pull-down experiments were performed batch-wise as described in

(Raikhy et al., 2011a). Briefly, full-length PFBV or PLPV CPs or PFBV p7 were expressed in Escherichia coli as MBP fusions. MBP alone was also expressed in

E. coli from the same vector to act as a negative control. Full-length PFBV or

PLPV CPs or PFBV p7 were also expressed as glutathione-s-transferase (GST) fusions. GST alone was used as a negative control. MBP-fused proteins from E. coli were co-incubated with the GST-fused protein at 4°C overnight. Note, all E. coli protein extracts were treated with 35 units RNase A (5 Prime, Inc.,

168

Gaithersburg, MD) for 1 hr (according to the manufacturer’s specifications) prior to co-incubation to ensure that the interactions identified are not mediated by

RNA. Incubations were then mixed with amylose resin. A portion of the co- incubated proteins prior to incubation with the resin served as a control and was termed the “Load.” The resin was subsequently washed with Washing Buffer (20 mM Tris pH 7.5, 200 mM NaCl, 1mM EDTA) and a portion of the flow-through was collected and termed, “Wash.” After washing, the MBP-bound proteins were removed from the resin by co-incubation with maltose (Washing solution +10 mM Maltose). This fraction, representing proteins bound to the MBP-fused protein was termed, “Elute.” Samples of each fraction were then mixed with sample buffer, boiled, and subjected to electrophoresis through a 10% polyacrylamide gel. Proteins were then transferred onto nitrocellulose, protein gel blot analyses were performed, and the blot was probed with primary antibodies binding to GST (Santa Cruz Biotechnology, Santa Cruz, CA at a dilution of

1:50,000) and detected with conjugated secondary antibodies (goat anti-rabbit antibodies at a dilution of 1:7,500). Protein distribution on the blots was determined using the HyGLO chemiluminescence kit (Denville Scientific,

Metuchen, NJ), according to the manufacturer’s specifications.

Protein purifications.

Interaction of PFBV p7 with PFBV CP was performed with purified proteins instead of cell extracts. MBP-tagged CP and GST-tagged p7 were expressed in E. coli BL21 codon plus (DE3) cells induced at 16° C. Cells expressing MBP PFBV CP were lysed by French Press in wash buffer (20mM 169

Tris-HCl pH7.4, 200mM NaCl, 1mM EDTA, 1mM TCEP (Tris(2- carboxyethyl)phosphine hydrochloride; Soltec Ventures, Beverly, MA)) supplemented with Protease Inhibitor Cocktail (Sigma-Aldrich Corporation, St.

Louis, MO), 1mM phenylmethanesulfonylfluoride (PMSF; Thermo Fisher

Scientific, Pittsburgh, PA), and RNaseA (35 U). Protein was purified by FPLC using HiTrapMBP column (GE Healthcare Life Sciences, Piscataway, NJ) according to the manufacturer’s specifications. Protein was eluted from the columns using elution buffer (wash buffer supplemented with 10mM Maltose).

Cells expressing GST PFBV p7 were lysed by French Press in cell lysis buffer

(50mM Tris-HCl pH7.4, 150mM NaCl, 1mM TCEP, 5% glycerol, 0.1% Triton-

X100 and supplemented with Protease Inhibitor Cocktail, 1mM PMSF, and

RNaseA (35 U). Protein was purified by FPLC using HiTrapGST column (GE

Healthcare Life Sciences) according to the manufacturer’s specifications.

Columns were washed with wash buffer (50mM Tris-HCl pH7.4, 300mM NaCl,

1mM TCEP, and 5% glycerol) and protein was eluted from columns using elution buffer (wash buffer supplemented with 50 mM Reduced Glutathione). Protein concentrations were determined spectrophotometrically. Protein interaction studies were performed as described above in pull-down assays.

170

A.4 Results

Virus prevalence in Ornamental Plant Germplasm Center (OPGC)

Accessions

To examine the type and prevalence of viruses in an American collection of Pelargonium species, cuttings obtained from the Ornamental Plant Germplasm

Center (OPGC; Columbus, OH) were analyzed for the presence of Pelargonium viruses. The OPGC collection contains a variety of different Pelargonium types ranging from scented to ivy geraniums (http://opgc.osu.edu/home). A total of 24

OPGC accessions (Table A-1). were evaluated in our study. These accessions were originally obtained by the OPGC mainly from collections from Illinois and

Pennsylvania (indicated in Table 1). The plants were chosen based on their exhibition of virus-like symptoms and some asymptomatic plants were also examined.

RNA was isolated from the leaves of each accession and reverse transcriptase-polymerase chain reaction (RT-PCR) analyses were performed.

Virus was detected in 19 of the 24 accessions as identified by RT-PCR producing bands of the correct size. The identity of the PCR products was confirmed by sequencing. PFBV was the most abundant virus in the population, being detected in 15 accessions alone or in combination with either PLPV (two cases) or ToRSV

(one case). ToRSV was not detected alone, but PLPV was in 3 cases. All accessions tested by RT-PCR were also negative for eight additional viruses (data not shown).

171

Table A-1: Accession analyzed for PFBV, PLPV, ToRSV.

*Ornamental Plant Germplasm Center (OPGC; at Ohio State University) accessions analyzed. **Source listed on the OPGC website: (http://opgc.osu.edu/home) I, Illinois; N, not known; P, Pennsylvania. ***Virus detected from cuttings from the OPGC accessions; PFBV, Pelargonium flower break virus; PLPV, Pelargonium line pattern virus; ToRSV, Tomato ringspot virus. Yes, virus detected by reverse transcriptase-polymerase chain reaction; No, not detected. #Accessions from which the PFBV and PLPV genes were obtained.

Analysis of PFBV Coat Protein Self-Association

Given the prevalence of PFBV in the collection that we analyzed, this virus was investigated in more detail. Initial studies focused on the PFBV CP and its ability to self-associate. Yeast transformants expressing the PFBV CP from

172 both pEG202 (as a LexA DNA-binding domain (DBD) fusion protein) and pJG4-

5 (as a B42 transcription activation domain (TAD) fusion protein) grew on leucine-deficient media and showed β-galactosidase activity (Fig. A-1a,b). In contrast PFBV CP did not show interaction with either the unfused LexA DBD

(expressed from empty pEG202), or with the unfused B42 TAD (expressed from empty pJG4-5). In addition, this interaction was specific since PFBV CP did not interact with Cauliflower mosaic virus (CaMV) gene VI product (Li and Leisner,

2002), an unrelated protein. Therefore, the PFBV CP specifically self-associates.

The self-association of PFBV CP was also confirmed biochemically in maltose- binding protein (MBP) pull-down assays (Fig. A-1d) but it did not bind GST alone.

To identify the region responsible for self-association, the PFBV CP gene was divided into three portions: the R (amino acids 1-73), S (amino acids 74-

240), and P domains (amino acids 241-346) based on the predicted sequence and the structure of other carmoviral CPs (Carrington et al., 1987; Rico and

Hernandez, 2004, 2006). Full-length PFBV CP interacted with all three portions,

R, S, and P as indicated in yeast two-hybrid assays by growth on leucine-deficient media and β-galactosidase activity of the transformants (Fig. A-2a,b).

Furthermore, all three PFBV CP domains were able to self-associate and interact with the other domains.

173

Figure A-1. Self-association interactions of Pelargonium flower break virus (PFBV) and Pelargonium line pattern virus (PLPV) coat proteins (CPs). (a) Diagram of constructions tested: constructions in column under pEG202, proteins expressed from that plasmid fused to LexA; constructions in column under pJG4-5, proteins expressed from that plasmid (or a modified form pJG4- 5Mod for PLPV CP) fused to B42; black boxes, full length coat protein for either PFBV (in b) or PLPV (in c); textured boxes in 4 and 5 indicate full length Cauliflower mosaic virus (CaMV) P6 serving as a negative control to indicate specificity of CP interactions; dashes indicate where empty vector was used; numbers correspond to β-galactosidase activity results in b or c. (b) Yeast two- hybrid analysis of PFBV CP self-interaction: numbers correspond to constructions shown in a; graph shows average and standard deviation of β-galactosidase activity for a set of constructions; + indicates leucine independent growth for that set of constructions; - indicates inability to grow on media lacking leucine. (c) Yeast two hybrid analysis of PLPV CP self-interaction. (d) Maltose binding protein (MBP) pull-down analyses of CP interactions. Load, protein loaded onto the amylose column; wash, non-specifically bound protein washed off the column by buffer; elute, protein that is eluted from the column with maltose, indicative of interaction. For labels at the right end of each row, the first indicates the protein fused to MBP, the second indicates the protein fused to the glutathione-s- transferase (GST) tag. All proteins were collected from the column after the treatments indicated, separated by SDS-PAGE, blotted onto nitrocellulose and detected using an anti-GST antibody. Those samples with just GST express an unfused GST protein to serve as a negative control.

174

Figure. A-2. Identification of domains involved in self-association interactions of PFBV and PLPV CPs. (a) Diagram of constructions tested, for either PFBV (in b; 346 amino acids) or PLPV (in C; 339 amino acids); R, N-terminal RNA-binding region (amino acids 1-73 for PFBV CP, 1-62 for PLPV CP); S, central shell domain (amino acids 74- 240 for PFBV CP, 63-233 for PLPV CP); P, C-terminal protruding domain (amino acids 241-346 for PFBV CP, 234-339 for PLPV CP); dashes indicate where empty vector was used. (b) Yeast two-hybrid analysis of PFBV CP domains. (c) Yeast two hybrid analysis of PLPV CP domains.

Analysis of PFBV CP-p7 Interactions

The CP for Turnip crinkle virus (TCV), another carmovirus, is required for systemic movement (Hacker et al., 1992). We reasoned that the same requirement may be true of the PFBV CP. Therefore, we investigated whether the

PFBV CP interacted with the viral movement proteins (p7 or p12). While inconsistent results were obtained when we examined CP-p12 interactions (data not shown), we found that the PFBV CP did indeed interact with p7. Yeast expressing the LexA-DBD-CP and B42-p7 grew on leucine-deficient media and 175 exhibited β-galactosidase activity (Fig. A-3a,b). Yeast expressing either fusion protein alone with the corresponding empty vector were unable to grow on leucine-deficient media and showed negligible β-galactosidase activity. We also confirmed p7-CP interaction biochemically, using proteins purified from E. coli, and used in MBP pull-down assays (Fig. A-3c). Interaction of p7 with the CP domains and CP with the p7 domains was inconclusive (data not shown).

Examination of p7 self-association

We also investigated if p7 could self-associate. Yeast transformants expressing both LexA DBD-p7 and B42 TAD-p7 (Fig. A-3d) showed both leucine-independent growth and β-galactosidase activity (Fig. A-3e). In contrast,

PFBV p7 did not show interaction with LexA DBD, the B42 TAD alone, or with

CaMV P6. The self-association of p7 was confirmed biochemically by an MBP pull-down assay (Fig. A-3f). The p7-GST fusion protein was retained on the amylose matrix by the MBP-p7 fusion but not by GST alone.

176

Figure A-3. PFBV p7 interactions. (a) Diagram of constructions tested in yeast two-hybrid analyses of full-length PFBV CP (346 amino acids) interactions with PFBV p7. (b) Yeast two-hybrid analysis of PFBV CP-p7 interactions. (c) Maltose binding protein (MBP) pull down analysis of CP-p7 interactions. Either purified MBP alone (MBP) or purified MBP-PFBV CP (CP) was co-incubated with purified glutathione-s- transferase-tagged p7 (P7) and added to an amylose column. Columns were then washed, eluted by the addition of maltose and analyzed by gel blot analysis using anti-GST primary antibodies. Note, only the eluted fractions are shown. (d) Diagram of constructions tested in yeast two-hybrid analyses of full-length p7 (white boxes; 68 amino acids) self-association. (e) Yeast two hybrid analysis of PFBV p7 self-association. (f) MBP pull down analysis of p7 self-association.

To examine the portions of p7 involved in self-association, the protein was divided into three portions: N-terminal (amino acids 1-23), middle (amino acids

20-42) and C-terminal (amino acids 39-68) segments. These three regions for the

PFBV p7 protein were analyzed for interaction with the full-length protein (Fig.

A-4a). The only yeast transformants showing leucine-independent growth and β- galactosidase activity (Fig. 4b) were those expressing p7 fused to the TAD and

177 the C-terminal region connected to the DBD. Based on the yeast two-hybrid results, the C-terminal segment alone interacted efficiently with full-length p7.

The remaining co-transformant combinations were unable to grow on leucine- deficient media and lacked β-galactosidase activity.

To determine which portion of p7 the C-terminus binds, interactions were examined between the C-terminus and the other regions (Fig. A-4c). Yeast transformants expressing the C-terminal domain of p7 fused to the LexA DBD showed leucine-independent growth and β-galactosidase activity (Fig. 4d) when expressing either the central portion or the C-terminus. However, yeast transformants expressing the DBD-p7 C terminal portion and either the N- terminus fused to the TAD, or the B42 TAD alone, showed no leucine- independent growth nor β-galactosidase activity.

Investigation of PLPV CP interactions

Since PLPV is related to PFBV we chose to determine if the PLPV CP showed a pattern of interactions similar to it’s PFBV counterpart. Yeast transformants expressing the full-length PLPV CP from both yeast two-hybrid vectors were b-galactosidase-positive and grew on leucine-deficient media (Fig.

A-1c). The PLPV CP did not interact with the negative controls, nor with the

CaMV P6. Therefore, the PLPV CP specifically self-associates. The self- association of PLPV CP was also confirmed biochemically by MBP pull-down assays (Fig. A-1d). The PLPV CP-GST fusion protein was retained on an

178 amylose column when co-incubated with MBP-PLPV CP fusion protein but not

GST alone.

Figure A-4. PFBV p7 domain interactions involved in self-association. (a) Diagram of constructions tested in yeast two-hybrid analyses, full length, full length PFBV p7 N, N-terminal region (p7 amino acids 1-23); M, middle portion (amino acids 20-42); C, C-terminal region (amino acids 39-68); dashes indicate where empty vector was used; numbers correspond to β-galactosidase activity results in b. (b) Yeast two-hybrid analysis of PFBV p7 domains. (c) Diagram of p7 constructions tested in yeast two-hybrid analyses. (d) Yeast two-hybrid analysis of PFBV p7 domains.

To identify the region(s) of the PLPV CP responsible for self-association, the CP gene was also divided into the R (amino acids 1-62), the S (amino acids

63-233), and the P domains (amino acids 234-339) as above. Yeast two-hybrid analysis of the PLPV CP domains showed results similar to their PFBV counterparts (Fig. A-2c). All three domains interacted with full-length PLPV CP.

179

In addition, all three PLPV CP domains were capable of self-interaction as well as association with the other domains.

Interaction of PFBV and PLPV CP domains

Because PFBV is occasionally found in mixed infections with PLPV

((Alonso and Borja, 2005); Table 1), we investigated whether the CPs for these two viruses could cross-interact (Fig. 5a). Yeast two-hybrid analyses indicated that the full-length PFBV CP interacted with the PLPV CP, as indicated by growth on leucine-deficient media and β-galactosidase activity (Fig. A-5b).

These results were confirmed biochemically by MBP pull-down experiments (Fig.

A-1d).

The domains involved in PFBV-PLPV CP binding were then identified

(Fig. A-5a,b). The PFBV R domain interacted with the full-length PLPV CP as well as with the R, S, and P domains. The PFBV CP S domain interacted with only the PLPV R domain, but not the full-length PLPV CP, nor the other domains. Finally, the PFBV P domain interacted with the PLPV R and P domains, but not full-length PLPV CP or the S domain.

180

Figure A-5. Cross-interaction amongst the PFBV and PLPV CPs. (a) Diagram of constructions tested: PFBV CP and domains, white boxes; PLPV CP and domains, (b) Yeast two-hybrid analysis of PFBV CP and domains interacting with PLPV CP and domains.

A.5 Discussion

Pelargonium species are susceptible to a number of viruses, including

PFBV and PLPV (Alonso and Borja, 2005; Bouwen and Maat, 1992; Brunt et al.,

1999). We examined an American collection of Pelargonium varieties and found that PFBV was the most common virus detected. This is in contrast to European studies, that reported PLPV as the more abundant virus infecting Pelargoniums

(Alonso and Borja, 2005). This difference is likely do to the limited number of sources for Pelargoniums sampled in our study. We also detected PLPV either alone or in mixed infections with PFBV.

To better understand PFBV and PLPV, the interactions of the multifunctional CP were examined, starting with self-association. The CPs for both PFBV and PLPV specifically self-associated. These interactions are likely

181 essential for virus particle formation. Typically, carmovirus CPs are divided into an N-terminal R domain, a central S domain and the C-terminal P domain

(Carrington et al., 1987). Interestingly, each domain interacted with each other domain for both PFBV and PLPV CPs. For other carmoviruses, R-R domain interaction has been suggested to play a role in assembly of three CP dimers on viral RNA (Carrington et al., 1987). The R domain may also interact with the S domain at certain interfaces within carmoviral capsids. Genetic evidence for

PFBV CP suggests that the R and P domains interact, since certain pairs of mutations are co-variant amongst the two domains (Rico et al., 2006). However, a similar pattern of covariation was not observed with PLPV (Castano et al.,

2011). The S domains self-associate to form the core of carmovirus particles

(Carrington et al., 1987). Certain co-variant sites between the S and P domains of

PLPV CP also suggest that these domains interact in some way (Castano et al.,

2011). P domains on carmoviral CPs interact to form the protruding spike-like structures emerging from the virions (Carrington et al., 1987). It is intriguing that our yeast two-hybrid analysis confirmed these inter-domain interactions for both

PFBV and PLPV CPs. Our data could permit design of peptides to interfere with virus particle formation and hence, virus infection.

Since PFBV and PLPV are found in mixed infections (Alonso and Borja,

2005; Carrington et al., 1987; Castano and Hernandez, 2005; Castano et al., 2009;

Rico and Hernandez, 2004; Rico et al., 2006), we examined the possibility of cross-association amongst their CPs. Even though the PFBV and PLPV CPs share only 33.5% amino acid sequence identity, they were capable of cross-

182 associating. This may be because carmoviral CPs can have unrelated amino acid sequences, yet still show similar protein structures (Carrington et al., 1987).

However, the details of domain cross-association are different for the putative

PFBV-PLPV CP interactions when compared to self-association. For self- association of either the PFBV or PLPV CP, all three domains (R, S and P) all bind to themselves and to each other. In contrast, for cross association, the S domain of one virus does not bind either the S or the P domain for the other virus.

Interaction among the S domains forms the basis of the virus particle (Carrington et al., 1987). So even though PFBV-PLPV CPs cross-react and could possibly form virus-like particles, it is unlikely that such virions would be as stable as those composed of a single type of CP.

The function of PFBV-PLPV CP interactions is unknown. However, the P domain of PFBV CP contains silencing suppressor activity (Martinez-Turino and

Hernandez, 2009). If the PLPV CP has a similar function, perhaps a PFBV-PLPV

CP dimers (or higher order structure) connected by R binding to the S or P domains or P-P interactions, could inhibit gene silencing in more than one way.

This may explain why plants infected with both PFBV and PLPV display symptoms different from plants infected with a single virus isolate. Synergistic interactions amongst different viruses are often mediated at the level of silencing suppression (Pruss et al., 1997; Latham & Wilson, 2008).

Some carmoviruses, like TCV, require the CP for exit from vascular tissues and invasion of surrounding cells to develop a systemic infection (Cao et al., 2010). If the same is true of PFBV, then our discovery of CP-p7 interaction 183 could shed some light on this process. The interaction of the PFBV p7 with the

CP is also somewhat reminiscent to what is observed in the Bromoviridae.

Viruses in this latter group require the CP as well as the MP for systemic spread

(Allison et al., 1990; Hull, 2002). Interestingly, the MPs of viruses in the

Bromoviridae are much larger (~30 kDa) compared to the PFBV small movement protein (~7 kDa) (Hull, 2002). This may somewhat explain why PFBV p7 self- associates (see below), to provide a larger surface for interaction with the CP.

Our analyses of PFBV p7 showed that it had the capacity to self-associate.

The p7 equivalent protein in MNRV also has been reported to cooperatively self- associate (Genoves et al., 2006, 2009; Navarro et al., 2006). Taken together with our results, self-association may be a general characteristic of carmoviral p7 equivalent proteins. Our analyses suggest that p7 self-association is mainly mediated by the C-terminus of p7. The C-terminal portion of p7 is highly conserved among carmoviruses (Rico et al., 2006). This conservation is probably reflected by protein structure, as this region has the potential to fold into β-sheet secondary structures in aqueous solution (Vilar et al., 2005; Vilar et al., 2002).

Since β-sheet domains can mediate protein-protein interactions (Plyte and Kneale,

1991), it is possible that they are performing this function to facilitate p7 self- association.

In summary, we report that the CPs of PFBV and PLPV self-associate through three separate domains, each of which appears to participate in the interaction (Fig. A-6) The PFBV and PLPV CPs can also cross-associate, but the domain interactions are different from those for self-association. PFBV CP also 184 interacts with the movement protein p7. Finally, p7 self-associates through the C- terminal domain. These data provide new insights into the interactions among

PFBV proteins that may lead to novel ways for controlling the virus, either by affecting virion structure or movement.

Figure A-6. Summary of interactions amongst Pelargonium flower break virus (PFBV) and Pelargonium line pattern virus coat proteins (CPs) and PFBV CP with p7. The three regions of p7 and of the PFBV and PLPV CPs are indicated. Double- headed arrows, protein interactions; arrows starting and ending on the same portion of the proteins, self-association of that particular domain; vertical brackets, interaction of full-length proteins where domain interactions could not be discerned.

A.6 Acknowledgments

This work was supported in part by USDA-ARS Specific Cooperative

Agreement: 58-3607-1-193.

185

A.7 References

Akgoz, M., Nguyen, Q.N., Talmadge, A.E., Drainville, K.E. & Wobbe, K.K.

(2001). Mutational analysis of Turnip crinkle virus movement protein p8.

Mol Plant Pathol 2, 37-48.

Allison, R., Thompson, C. & Ahlquist, P. (1990). Regeneration of a functional

RNA virus genome by recombination between deletion mutants and

requirement for Cowpea chlorotic mottle virus 2a and coat protein genes

for systemic infection. Proc Natl Acad Sci USA 87, 1820-1824.

Alonso, M. & Borja, M. (2005). High incidence of Pelargonium line pattern

virus infecting asymptomatic Pelargonium spp. in Spain. Eur J Plant

Pathol 112, 95-100.

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith,

J.A. & Struhl, K. (1993). Current Protocols in Molecular Biology.

Cambridge MA USA, Greene Publishing Associates and John Wiley &

Sons, INC.

Bouwen, I. & Maat, D.Z. (1992). Pelargonium flower break and Pelargonium

line pattern viruses in the Netherlands; purification, antiserum preparation,

serological identification, and detection in Pelargonium by ELISA. Neth J

Plant Pathol 98, 141-156.

186

Brunt, A., Crabtree, K., Dallwitz, M., Gibbs, A. & Watson, L. (1999). Viruses

of plants: descriptions and lists from the VIDE database. New York, NY,

Oxford University Press.

Canizares, M.C., Marcos, J.F. & Pallas, V. (2001). Molecular variability of

twenty-one geographically distinct isolates of Carnation mottle virus

(CarMV) and phylogenetic relationships within the Tombusviridae family.

Arch Virol 146, 2039-2051.

Cao, M., Ye, X., Willie, K., Lin, J., Zhang, X., Redinbaugh, M.G., Simon,

A.E., Morris, T.J. & Qu, F. (2010). The capsid protein of Turnip crinkle

virus overcomes two separate defense barriers to facilitate systemic

movement of the virus in Arabidopsis. J Virol 84, 7793-7802.

Carrington, J.C., Morris, T.J., Stockley, P.G. & Harrison, S.C. (1987).

Structure and assembly of Turnip crinkle virus, IV. Analysis of the coat

protein gene and implications of the protein subunit primary structure. J

Mol Biol 194, 265-276.

Castano, A. & Hernandez, C. (2005). Complete nucleotide sequence and

genome organization of Pelargonium line pattern virus and its relationship

with the family Tombusviridae. Arch Virol 150, 949-965.

Castano, A. & Hernandez, C. (2007). Biological activity of transcripts from

cDNA of Pelargonium line pattern virus. Acta Virologica 51, 271-274.

187

Castano, A., Ruiz, L., Elena, S.F. & Hernandez, C. (2011). Population

differentiation and selective constraints in Pelargonium line pattern virus.

Virus Res 155, 274-282.

Castano, A., Ruiz, L. & Hernandez, C. (2009). Insights into the translational

regulation of biologically active open reading frams of Pelargonium line

pattern virus. Virology 386, 417-426.

Genoves, A., Navarro, J.A. & Pallas, V. (2006). Functional analysis of the five

Melon necrotic spot virus genome-encoded proteins. J Gen Virol 87,

2371-2380.

Genoves, A., Navarro, J.A. & Pallas, V. (2009). A self-interacting carmovirus

movement protein plays a role in binding viral RNA during cell-to-cell

movement and shows an actin cytoskeleton dependent location in cell

periphery. Virology 395, 133-142.

Gyuris, J., Golemis, E., Chertkov, H. & Brent, R. (1993). Cdi1, a human G1

and S phase protein phosphatase that associates with Cdk2. Cell 75, 791-

803.

Hacker, D.L., Petty, I.T., Wei, N. & Morris, T.J. (1992). Turnip crinkle virus

genes required for RNA replication and virus movement. Virology 186, 1-

Hull, R. 2002. Matthews' Plant Virology, 4th Ed. Academic Press, San Diego.

188

Krczal, G., Albouy, J., Damy, I., Kusiac, C., Deogratias, J., Moreau, J.,

Berkelman, B. & Wohanka, W. (1990). Transmission of Pelargonium

flower break virus by irrigation systems and by thrips. Plant Dis 79, 163-

166.

Latham, J.R. & Wilson, A.K. (2008). Transcomplementation and synergism in

plants: implications for viral transgenes? Mol Plant Pathol 9, 85-103.

Li, W.-z., Qu, F. & Morris, T.J. (1998). Cell-to-cell movement of Turnip crinkle

virus is controlled by two small open reading frames that function in trans.

Virology 244, 495-416.

Li, Y. & Leisner, S.M. (2002). Multiple domains within the Cauliflower mosaic

virus gene VI product interact with the full-length protein. Mol Plant-

Microbe Interact 15, 1050-1057.

Lommel, S.A., Martelli, G.P., Rubino, L. & Russo, M. (2005). Family

Tombusviridae. In Virus taxonomy: 8th report of the international

committee on taxonomy of viruses, C.M. Faquet, M.A. Mayo, J. Maniloff,

U. Desselberger & L.A. Ball, eds. San Diego, CA, Academic Press, pp.

907-936.

Marcos, J.F., Vilar, M., Perez-Paya, E. & Pallas, V. (1999). In vivo detection,

RNA-binding properties and characterization of the RNA binding domain

of the p7 putative movement protein from Carnation mottle virus

(CarMV). Virology 255, 354-365.

189

Martinez-Turino, S. & Hernandez, C. (2009). Inhibition of RNA silencing by

the coat protein of Pelargonium flower break virus: distinctions from

closely related suppressors. J Gen Virol 90, 519-525.

Martinez-Turino, S. & Hernandez, C. (2011). A membrane-associated

movement protein of Pelargonium flower break virus shows RNA-binding

activity and contains a biologically relevant leucine zipper-like motif.

Virology 413, 310-319.

Navarro, J.A., Genoves, A., Climent, J., Sauri, A., Martinez-Gil, L.,

Mingarro, I. & Pallas, V. (2006). RNA-binding properties and membrane

insertion of Melon necrotic spot virus (MNSV) double gene block

movement proteins. Virology 356, 57-67.

Ohki, T., Akita, F., Mochizuki, T., Kanda, A., Sasaya, T. & Tsuda, S. (2010).

The protruding domain of the coat protein of Melon necrotic spot virus is

involved in compatibility with and transmission by the fungal vector

Olpidium bornovanus. Virology 402, 129-134.

Plyte, S.E. & Kneale, G.G. (1991). The role of tyrosine residues in the DNA-

binding site of the Pf1 gene 5 protein. Protein Eng 4, 553-1560.

Pruss, G., Ge, X., Shi, X.M., Carrington, J.C. & Bowman Vance, V. Plant

viral synergism: the potyviral genome encodes a broad-range

pathogenicity enhancer that transactivates replication of heterologous

viruses Plant Cell 9, 859-868.

190

Raikhy, G., Krause, C. & Leisner, S. (2011). The Dahlia mosaic virus gene VI

product N-terminal region is involved in self-association. Virus Res 159,

69-72.

Rico, P. & Hernandez, C. (2004). Complete nucleotide sequence and genome

organization of Pelargonium flower break virus. Arch Virol 149, 641-651.

Rico, P. & Hernandez, C. (2006). Infectivity of in vitro transcripts from a full-

length cDNA clone of Pelargonium flower break virus in an experimental

and a natural host. J Plant Pathol 88, 103-106.

Rico, P. & Hernandez, C. (2009). Characterization of the subgenomic RNAs

produced by Pelargonium flower break virus: identification of two novel

RNA species. Virus Res 142, 100-107.

Rico, P., Ivars, P., Elena, S.F. & Hernandez, C. (2006). Insights into the

selective pressures restricting Pelargonium flower break virus genome

variability: evidence for host adaptation. J Virol 80, 8124-8132.

Salanki, K., Gellert, A., Huppert, E., Naray-Szabo, G. & Balasz, E. (2004).

Compatibility of the movement protein and the coat protein of

cucumoviruses is required for cell-to-cell movement. J Gen Virol 85,

1039-1048.

Takeda, A., Kaido, M., Okuno, T. & Mise, K. (2004). The C terminus of the

movement protein of Brome mosaic virus controls the requirement for

191

coat protein in cell-to-cell movement and plays a role in long-distance

movement. J Gen Virol 85, 1751-1761.

Vilar, M., Sauri, A., Marcos, J.F., Mingarro, I. & Perez-Paya, E. (2005).

Transient structural orgering of the RNA-binding domain of Carnation

mottle virus p7 movement protein modulates nucleic acid binding.

ChemBioChem 6, 1391-1396.

Vilar, M., Sauri, A., Monne, M., Marcos, J.F., von Heijne, G., Perez-Paya, E.

& Mingarro, I. (2002). Insertion and topology of a plant viral movement

proten in the endoplasmic reticulum membrane. J Biol Chem 277, 23447-

23452.

Xiang, Y., Kakani, K., Reade, R., Hui, E. & Rochon, D. (2006). A 38-amino-

acid sequence encompassing the arm domain of cucumber necrosis virus

coat protein functions as a chloroplast transit peptide in infected plants. J

Virol 80: 7952-7964.

192

A.7 Supplementary Data

Table A-S1: Primers used in PFBV and PLPV study.

*Primer names with suffixes ending in -F are forward primers while those with -R are reverse primers. **Nucleotide sequences of primers, written 5’ to 3’, nucleotides differing from PFBV coat protein or p7 sequences or PLPV coat protein sequences are given in small letters while restriction enzyme sites are underlined and stop codons are given in italics. ***Location of 5` nucleotide on PFBV or PLPV genomic sequence. #Amino acid position encoded by primer, numbers indicate either first amino acid (forward primer) or last amino acid (reverse primer). &GenBank accession numbers from which the primers were designed: NC_005268, PFBV; NC_007017, PLPV; NC_003840, ToRSV.

193

Figure A-S1: Schematic presentation of full-length PFBV (white) and PLPV (shaded) CPs and their fragments analyzed in this study. The primers (given in Supplementary Table 1) indicate those used to generate the CP/CP fragments by PCR.

Figure A-S2: Schematic presentation of full-length PFBV p7 and its fragments analyzed in this study. The primers (given in Supplementary Table 1) indicate those used to generate the p7/p7 fragments by PCR.

194

Appendix B

Clones

Table B-1: CaMV1841 and W260 full-length P6s in pENTR

195

Table B-2: CaMV1841, W260, D4 P6 domains in pENTR

196

Table B-3: CaMV1841 P6 Mutants in pENTR

197

Table B-4: CaMV1841 P2 full-length and domains in pENTR

198

Table B-5: CaMV1841 and W260 P5 full-length and domains in pENTR

199

Table B-6: CaMV1841 and W260 P3/P7 domains in pENTR

Table B-7: CaMV1841 and W260 genes subcloned into pDEST10

200

Table B-8: CaMV1841 and W260 genes subcloned into pDEST15

201

Table B-9: PFBV genes subcloned into pDEST15

202

Table B-10: CaMV1841 genes subcloned into pDEST17

203

Table B-11: CaMV1841 and W260 genes subcloned into pEG202

204

Table B-12: CaMV1841 and W260 genes subcloned into pJG4-5

205

Table B-13: CaMV1841 genes subcloned into pMAL

206

207

Table B-14: CaMV1841 genes subcloned into pSITE

208

209

210

Table B-15: CaMV pSITE constructs transformed into Argobacterium

211

212