Roles of Snrk1, ADK, and APT1 in the Cellular Stress Response and Antiviral Defense

Home , Adenosine kinase, Adenosine nucleosidase, ADK (gene)

Roles of SnRK1, ADK, and APT1 in the Cellular Stress Response and Antiviral Defense

Dissertation

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate School of

The Ohio State University

Gireesha T. Mohannath, M.S.

Plant Cellular and Molecular Biology Graduate Program

The Ohio State University

2010

Dissertation Committee:

Dr. David M. Bisaro, Advisor

Dr. Erich Grotewold

Dr. Venkat Gopalan

Dr. Jyan-Chyun Jang

Gireesha T. Mohannath

2010

ABSTRACT

Members of the SNF1/AMPK/SnRK1 family of kinases are highly conserved.

Representatives include SNF1 kinase (sucrose non-fermenting 1) in yeast, SnRK1

(SNF1-related kinase 1) in plants, and AMPK (AMP-activated protein kinase) in animals.

These Ser/Thr kinases play a central role in the regulation of metabolism. In response to nutritional and environmental stresses that deplete ATP, they turn off energy-consuming biosynthetic pathways and turn on alternative ATP-generating systems as part of the cellular stress response (CSR). However, the mechanisms that activate these enzyme complexes are not completely understood.

These kinases function as heterotrimeric complexes. The catalytic  subunit consists of an N-terminal kinase domain with an activation loop that contains a conserved threonine residue which must be phosphorylated for activity. Following phosphorylation by upstream kinase(s), 5'-AMP is known to allosterically stimulate AMPK activity and to inhibit its inactivation due to dephosphorylation of  subunit at Thr172 by protein phosphatase 2C (PP2C). Direct allosteric stimulation of the SnRK1 complex by 5'-AMP has yet to be demonstrated. However, 5'-AMP has been shown to suppress dephosphorylation of SnRK1 by PP2C.

Adenosine kinase (ADK) phosphorylates adenosine to 5'-AMP. ADK and SnRK1 play key roles in antiviral defense, and geminivirus AL2 and L2 proteins inactivate both

ii kinases. Because ADK generates 5'-AMP that is known to activate CSR, this study hypothesizes that ADK contributes to rapid activation of the CSR.

In support of this hypothesis, we showed that ADK and SnRK1 form a complex in vivo. This study also demonstrates an increase in SnRK1 activity in transgenic

Arabidopsis plants over-expressing ADK, and a 4-6 fold in vitro stimulation of ADK by

SnRK1. Interestingly, ADK stimulation does not require SnRK1 kinase activity. We conclude that SnRK1 and ADK mutually stimulate each other to rapidly activate CSR.

SnRK1 phosphorylates ADK in vitro, and we speculate that this phosphorylation might be playing a role in negative regulation of the SnRK1-ADK complex.

Adenine phosphoribosyl transferase 1 (APT1) also generates 5'-AMP. We hypothesize that APT1 might also play a role in activating CSR. In vitro and in vivo data is presented in this thesis to demonstrate interaction between SnRK1 and APT1. Also,

SnRK1 is shown to phosphorylate APT1 in vitro. We speculate that SnRK1, ADK, and

APT1 might form a ternary complex that could play critical roles in detecting and responding to cellular stress.

We also show that four different geminivirus proteins, AL2, L2, AV2, and C1 interact with SnRK1, ADK, and APT1, and discuss the possible consequences of these interactions. Finally, we hypothesize that, as an antiviral defense, SnRK1 might play a role in regulating protein synthesis, and in support of this idea we observed phosphorylation of translation initiation factors eIF-2 and eIF-(iso)4E by SnRK1 in vitro. Future experiments have been proposed to confirm this hypothesis.

iii

Taken together, the findings from this study further our understanding on the potential roles of SnRK1, in conjunction with ADK and APT1, in sensing and mediating responses to various kinds of biotic and abiotic stresses.

Dedicated to my family and teachers

ACKNOWLEDGEMENTS

Mathrudevobhava, Pithrudevobhava, and Gurudevobhava,

(The mother, the father, and the teacher to be revered as God)

I begin my acknowledgements by expressing the deepest gratitude to my advisor, Dr.

David Bisaro, for his support, critical comments, and the invaluable knowledge I received from him during my tenure in his lab. I will be forever grateful to him for encouraging me to be innovative and exploratory. He greatly helped me to become a better scientist.

I am sincerely grateful to my committee members Drs. Erich Grotewold, Venkat

Gopalan and J.C. Jang for their time, unwavering support, guidance and advice. I have been fortunate to have these accomplished scientists as my committee members as their input into my research was very useful and has been highly valued.

From bottom of my heart I thank my current and former lab colleagues Dr. Kenn

Buckley, Dr. Cody Buchmann, Dr. Priya Raja, and Jamie Wolf, for their consistent moral and technical support throughout my tenure in the lab. Specifically, I cannot forget the technical help and advice I received from Drs. Buckley and Buchmann during my early days as a graduate student. I am immensely thankful to Veena Patil (my wife), Dr. Youn

Lee, and Allie Varner for helping me in my project. I highly value their contributions to my projects. I would also like to acknowledge the other current and past members of the

Bisaro lab who have been helpful in various ways: Xiaojuan Yang, Dr.

Hui Wang, Dr. Lin Hui, Dr. Shaheen Asad, Dr. Mohammed Mubeen, Isaac Heard,

Sizhun Lee, Jeff Ostler, Brad Sanville, and other undergraduate students.

Definitely I won‘t forget the help and the various constructive conversations I had with students, postdocs, and faculty members from different departments especially from

Molecular Genetics/Plant Cellular and Molecular Biology (PCMB). I am also thankful to current PCMB staff: Eduardo Acosta, Rene Reese, Laurel Shannon, and Joan Leonard for their help. A special acknowledgement to Dr. Biao Ding for his permission to use his confocal and fluorescent microscopes. Also, my sincere thanks to the staff of the biotechnology center: Melinda Parker, Diane Furtney, Dave Long, Scott Hines, Joe

Takayama and MCDB secretary, Jan Zinich for all of their help, and I offer my sincere acknowledgements to ABRC for their Arabidopsis mutants and DNA clones.

I owe a great debt of gratitude to all of my family members: My mother, father, brother, sister, and their family members for their kind and compassionate support throughout my career. Without their support I wound not have made it U.S. for my education. Although for my efforts I am the only person who receives Ph.D., my wife is an exemplary partner in all of my efforts and I share with her whatever I receive. It would have been impossible to pursue my doctoral studies but for my wife‘s moral support, tolerance, understanding, and help. Many of my teachers, friends and relatives, in conjunction with my family, gave me the strength to propel myself forward through the most difficult and trying times of my graduate career. I don‘t know how I would have managed without their support.

vii

I am extremely grateful to PCMB and CLSE for funding me every quarter throughout my Ph.D. years by employing me as a graduate teaching associate. I think because of their support I became a better teacher and received ‗Molecular Genetics

Teaching Award‘. Finally, I bow with reverence to The Ohio State University.

Go Bucks!

viii

VITA

1993 – 1997………………. B.S., Agricultural Sciences, UAS, Dharwad, India

1997 – 1999………………. M.S., Genetics & Plant Breeding, UAS, Bangalore, India

1999 – 2003………………. Research Associate, UAS, Bangalore, India

2003 – present……………... Graduate Teaching and Research Associate, the Ohio State

University.

(UAS: University of Agricultural Sciences)

PUBLICATIONS

Research Publications:

1. Buchmann, R.C., Asad, S., Wolf, J.N., Mohannath, G., and Bisaro, D.M. (2009). Geminivirus AL2 and L2 proteins suppress transcriptional gene silencing and cause genome-wide reductions in cytosine methylation. J. Virol. 83(10): 5005-5013) (featured in the “Spotlight” section of the journal)

2. Girish, T.N., Gireesha, T.M. Vaishali, M.G., Hanamareddy, B.G., and Hittalmani, S. (2006). Response of a new IR50/Moroberekan recombinant inbred population of rice (Oryza sativa L.) from an indica x japonica cross for growth and yield traits under aerobic conditions. Euphytica 152: 149-161.

3. Nadaradjan, S., Gireesha, T. M., Sheshashayee, M. S., Shankar, A. G., Prasad, T. G., and Udayakumar, M. (2003). Progress in genetic polymorphism studies via molecular markers in groundnut (Arachis hypogaea L.) - A review. J. Plant Biol. 30 (3): 285 – 292.

4. Toorchi, M., Shashidhar, H.E., Gireesha, T.M., and Hittalmani, S. (2003). Performance of backcross transgressant doubled haploid lines of rice under contrasting moisture regimes: Yield components and Marker heterozygosity. Crop Science 43:1448-1456.

5. Toorchi, M., Shashidhar, H.E., Hittalmani, S., and Gireesha, T.M. (2002). Rice root morphology under contrasting moisture regimes and contribution of molecular marker heterozygosity. Euphytica 126(2): 251-257.

6. Gireesha, T. M., Shashidhar, H. E., and Hittalmani, S. (2000), Genetics of root morphology and related traits in an indica-indica based mapping populations of rice (Oryza sativa L.). Res. on Crops 1(2): 208-215.

FIELDS OF STUDY

Major Field: Plant Cellular and Molecular Biology

TABLE OF CONTENTS

Abstract ...... ……………………………………..ii

Dedication ...... v

Acknowledgements ...... vi

Vita ...... ix

List of Tables………………………………………………………………………...…xvii

List of Figures ...... xviii

Chapters:

CHAPTER 1………………………………………………………………………………1

INTRODUCTION………………………………………………………………………...1

1.1 Geminiviruses………………………………………………………………………..1

1.1.1 Geminivirus classification………………………………………………………...2

1.1.2 Geminivirus nomenclature, gene organization, and gene functions……………...4

1.1.3 Geminivirus rolling circle replication and infection cycle………………………. 7

1.2 Geminivirus pathogenicity factors………………………………………………. .8

1.2.1 AL2 of TGMV and L2 of BCTV………………………………………………… 8

1.2.2 AV2, precoat protein of certain geminiviruses………………………………… 11

 C1, a protein encoded by satellite DNA ……………………………………. .11

1.3 SnRK1, a Ser/Thr protein kinase………………………………………………...12

1.3.1 Protein kinases, key molecular switches involved in the regulation of cellular

processes………………………………………………………………………....12

1.3.2 SnRK1, a member of SNF1-AMPK family of kinases………………………..…13

1.3.3 Classification of SnRK1 in plants………………………………………………..14

1.3.4 Regulation of SnRK1 expression and activity…………………………………..15

1.3.5 Role of increased AMP/ATP ratio in activating SNF1-AMPK family of

kinases………………………………………………………………...………….16

1.3.6 SNF1-AMPK family kinases function as heterotrimeric complexes………….....16

1.3.7 Mechanisms of SnRK1 activation and known SnRK1 targets………………..…17

1.3.8 Role of SnRK1 in antiviral defense…………………………………………...…19

1.4 Adenosine kinase (ADK) and Adenine phosphoribosyltransferase (APT)……....20

1.4.1 ADK generates 5'-AMP, and is involved in cytokinin metabolism………….…..20

1.4.2 ADK plays a critical role in methylating geminivirus genomes as part of the host

defense……………………………………………………………………...……22

1.4.3 Roles of ADK in mammals………………………………………………………23

1.4.4 APT generates 5'-AMP, and is involved in cytokinin metabolism………………23

1.5 Defense responses against viruses…………………………………………….…24

1.5.1 PKR-mediated interferon (IFN) response, a key antiviral defense in animals…..24

1.5.2 PKR pathway in plants…………………………………………………………..27

1.5.3 Roles of translation initiation factors in plant antiviral defense………………...28

CHAPTER 2…………………………………………………………………………..…39

xii

AN IN VIVO COMPLEX CONTAINING SNF1-RELATED (SnRK1) AND

ADENOSINE KINASE IN ARABIDOPSIS……………………………………………..39

2.1 Introduction……………………………………………………………………….….39

2.2 Results………………………………………………………………………….…….43

2.2.1 Development of a gel-based SnRK1 assay……………………………………..….43

2.2.2 SnRK1 and ADK interact in yeast and plant cells…………………………………45

2.2.3 SnRK1 co-immunoprecipitates with native ADK, and co-purifies with overexpressed ADK……………………………………………………………………….….46

2.2.4 SnRK1 phosphorylates ADK in vitro………………………………………...……49

2.2.5 ADK stimulation does not require SnRK1 kinase activity…………………..…….50

2.2.6 The activities of SnRK1 and ADK are mutually enhanced in vivo..…………..…..52

2.3 Discussion……………………………………………………………………………54

2.4 Materials and Methods…………………………………………………………..….56

2.4.1 Protein expression in N. benthamiana…………………………………………….56

2.4.2 Protein expression in E. coli…………………………………………………….…57

2.4.3 Kinase assays………………………………………………………………...…….58

2.4.4 Yeast two-hybrid analysis……………………………………………………….…59

2.4.5 BiFC analysis of interactions…………………………………………………..…..59

xiii

2.4.6 Immunoprecipitation and Immunodetection…………………………………..…..61

CHAPTER 3………………………………………………………………………..……81

ADENINE PHOSPHORIBOSYL TRANSFERASE 1 (APT1) INTERACTS WITH SnRK1 AND ADK……………………………..…………………………………….….81

3.1 Introduction………………………………………………………………..………....81

3.2 Results……………………………………………………………………….……….84

3.2.1 APT1 interacts with SnRK1 in the cytoplasm………………………………….… 84

3.2.2 SnRK1-KD phosphorylates APT1 in vitro…………………………………..….…85

3.2.3 Endogenous SnRK1 copurifies with overexpressed APT1 in N. benthamiana……87

3.2.4 SnRK1, APT1, and ADK interact with each other in the cytoplasm………….…..87

3.3 Discussion……………………………………………………………………………89

3.4 Materials and Methods…………………………………………………………….....92

3.4.1 Plant material, agroinfiltration, and bimolecular fluorescence complementation (BiFC) ………………………………………………………………………………...…92

3.4.2 Protein expression and purification………………………………………………..93

3.4.3 SnRK1 kinase assay……………………………………………………………..... 95

3.4.4 Protein sequence comparison using BLAST analysis……………………….….…96

CHAPTER 4……………………………………………………………………………105 GEMINIVIRUS PATHOGENICITY FACTORS INTERACT WITH SNRK1, ADK, AND APT1, A PUTATIVE COMPLEX THAT LIKELY PLAYS A ROLE IN ANTIVIRAL DEFENSE……………………………………………………………….105 4.1 Introduction…………………………………………………………………………105

4.2 Results……………………………………………………………………………....108

4.2.1 AL2, L2, AV2, and C1 interact with SnRK1, ADK, and APT1 in BiFC analyses……………………………………………………………………………....…108

4.2.2 AL2 forms nuclear speckles with SnRK1 but not with ADK…………….…...…111

xiv

4.2.3 AL2 interacts with APT1 in an in vitro GST-pull down assay………………...…111

4.2.4 SnRK1 phosphorylates translation initiation factors eIF-2 and eIF-(iso)4E in vitro…………………………………………………………..…………………………112

4.3 Discussion………………………………………………………………………..…114

4.4 Materials and Methods………………………………………………………….…..120

4.4.1 Plant material, agroinfiltration, and bimolecular fluorescence complementation (BiFC) ………………………………………………………………………………….120

4.4.2 Protein expression and purification………………………………………………121

4.4.3 SnRK1 kinase assay and GST pull down assay ……………………………….…123

4.4.4 Protein sequence comparison using BLAST analysis……………………………124

CHAPTER 5……………………………………………………………………………137

DISCUSSION……………………………………………………………………..……137

5.1 SnRK1-ADK-APT1 interactions and their consequences……………………….…137

5.1.1 SnRK1 and ADK form a complex in vivo that could rapidly activate CSR……...137

5.1.2 Activation of ADK by SnRK1 is independent of its kinase activity: a novel role for

SNF1-AMPK-SnRK1 family kinases………………………………………………..…140

5.1.3 SnRK1-ADK-APT1, a putative complex that might play significant roles in responses to various kinds of stresses………………………………………………….142

5.2 SnRK1, ADK, and APT1 could be ―high value‖ targets for geminiviruses and perhaps to all plant viruses……………………………………………………………………....144

5.2.1 Different geminivirus proteins interact with SnRK1, ADK, and APT1 …………144

5.2.2 One of the SnRK1 antiviral roles could be to block protein synthesis……….…..149

5.2.3 SnRK1 pathway plays significant role in antiviral defense and geminiviruses appear to be evolved to counteract this pathway …………………………………….…...……153

BIBLIOGRAPHY………………………………………………………………………155

xvi

LIST OF TABLES

Table 2.1 Interactions between ADK, SnRK1, and SnRK1-KD in the yeast two-hybrid system ………………………………………………………………………………..….77

Table 2.2 ADK2 of Arabidopsis thaliana has three evolutionarily conserved potential SnRK1 sites………………………………………………………………………...…….78

Table 3.1 APT1 of Arabidopsis thaliana has two evolutionarily conserved potential SnRK1 site ………………………………………………………………………..……104

Table 4.1 AteIF-2 has three evolutionarily conserved potential SnRK1 sites....…..…135

Table 4.2 AteIF-4E/AteIF-(iso)4E has an evolutionarily conserved potential SnRK1 site………………………………………………………………………………………136

xvii

LIST OF FIGURES

Figure 1.1 Geminivirus gene organization………………………………………….……31

Figure 1.2 Geminivirus replication………………………………………………………33

Figure 1.3 AL2 and L2 proteins……………………………………………………...…..35

Figure 1.4 Role of eIF2 in translation initiation…………..……………………….…….36

Figure 1.5 Genes encoding eIF4F components in Arabidopsis thaliana……………..…38

Figure 2.1 Gel-based SnRK1 assay …………………………….…………………….…62

Figure 2.2 SnRK1 and ADK interact in the BiFC assay ……………………………..…65

Figure 2.3 SnRK1 co-immunoprecipitates with ADK ………………..…………..…….66

Figure 2.4 Endogenous SnRK1 activity copurifies with ADK in N. benthamiana...…..68

Figure 2.5 ADK is phosphorylated by SnRK1 in vitro……………………………….….69

Figure 2.6 SnRK1-KD and SnRK1-KD-K49R enhance ADK activity in vitro ………...71

Figure 2.7 ADK and SnRK1 activities increase in parallel in vivo…………………..…73

Figure 2.8 GST-SnRK1-KD expressed in E. coli is poorly active and does not phosphorylate ADK in vitro………………………………………………………….…..75

Figure 3.1 APT1 interacts with SnRK1 in the cytoplasm……………………………..…97

Figure 3.2 SnRK1 phosphorylates APT1 in vitro………………………………..………98

Figure 3.3 Endogenous SnRK1 activity copurifies with APT1 in N. benthamiana……100

Figure 3.4 SnRK1, APT1 and ADK interact with each other in the cytoplasm…….….101

Figure 3.5 APT1 might coimmunoprecipitates with ADK……………………………..102

xviii

Figure 3.6 Overall hypothesis………………………………………………………....103

Figure 4.1 AL2, L2, AV2, and C1 interact with SnRK1, ADK and APT1 in BiFC analyses…………………………………………………………………………………125

Figure 4.2 AL2-SnRK1 complexes accumulate in nuclear speckles…………………...128

Figure 4.3 AL2 interacts with APT1 in a GST pull-down assay…………………….…130

Figure 4.4 SnRK1 phosphorylates AteIF-2 and AteIF-(iso)4E in vitro………………131

Figure 4.5 APT plays a role in cytokinin metabolism………………………………....132

Figure 4.6 Methyl cycle………………………………………………….…………..…133

Figure 4.7 Significance of SnRK1-ADK-APT1 complex in blocking protein synthesis as part of antiviral defense………………………………………………………………...134

xix

CHAPTER 1

INTRODUCTION

1.1 Geminiviruses

Viruses are ultramicroscopic microorganisms that replicate only within the cells of living hosts. They contain single- or double-stranded RNA or DNA as genetic material, which is normally encased in a protective coat made of proteins or lipoproteins, and some have a surrounding envelope. They use the host machinery and energy for their replication and spread. Viruses are recognized as one of the major pathogens causing diseases to diverse eukaryotes. The first description of a plant virus disease was made as early as 752 A.D.

(Saunders et al., 2003).

The family Geminiviridae contains viruses with one (monopartite) or two

(bipartite) circular single-stranded DNA (ssDNA) genomes (Bisaro, 1996; Jeske, 2009).

These viruses derive their name from their existence as twin icosahedral (geminate) particles. Each paired particle encapsidates a single molecule of ssDNA genome 2.5 to

3.0 kb in length. These genomes are expressed and replicated from double-stranded DNA

(dsDNA) replicative forms (RFs), which are organized as minichromosomes complexed with histone proteins (Pilartz and Jeske, 1992; Gutierrez 1999; Hanley-Bowdoin et al.,

2000). Geminiviruses are simple viruses that rely on the host for replication and transcription because they don‘t encode any polymerases.

Geminiviruses cause significant yield loses to several crop plants throughout the world (Hull, 2002; Mansoor et al., 2003). Specifically they have been regarded as the most destructive pests for important crops such as beans, corn, potato, cotton, pepper, sugar beet, sweet potato, and tomato (Moffatt 1999). Because most of these plant species are staple food crops for people depending on subsistence agriculture in tropical and subtropical countries, combating geminivirus epidemics significantly contributes to social and economic stability in these areas (Jeske, 2009).

1.1.1 Geminivirus classification

Geminiviruses belong to the family Geminiviridae which is divided into four genera, namely, Begomovirus, Curtovirus, Mastrevirus and Topocovirus. This classification is based on genome organization, host range, and insect vectors (Fauquet et al., 2003).

Begomovirus includes more than hundred species, with Bean golden yellow mosaic virus

(originally Bean golden mosaic virus-Puerto Rico) as the type species. These viruses infect dicotyledonous plants, and are transmitted by whiteflies (Bemicia tabaci Genn.)

(Fauquet et al., 2003). The begomoviruses are further divided into Old World and New

World groups. Most Old World begomoviruses have bipartite genomes although some have monopartite genomes. By contrast, all New World begomoviruses have bipartite genomes. Bipartite viruses‘ genomes consist of two DNA molecules (called genome A and genome B) each about 2.5 kb, and both are required for infection. The DNA sequences of these two molecules are quite different with the exception of a common region (CR) about 200 nt long (Hamilton et al., 1983; Stanley 1983). Some well-

2 characterized bipartite begomoviruses include Tomato golden mosaic virus (TGMV),

African cassava mosaic virus (ACMV), Squash leaf curl virus (SqLCV) and Cabbage leaf curl virus (CabLCV). Among monopartite begomoviruses, Tomato yellow leaf curl virus (TYLCV) and Tomato leaf curl virus (ToLCV) are well-studied.

Viruses belonging to the Mastrevirus and Curtovirus genera have monopartite genomes and are transmitted by leafhopper vectors. However, mastreviruses primarily infect monocotyledonous plants while the curtoviruses infect dicotyledonous plants.

Maize streak virus (MSV) and Beet curly top virus (BCTV) are the type species of these genera, respectively (Palmer and Rybicki, 1998). Topocuvirus is a recently designated genus by the International Committee on Taxonomy of Viruses (ICTV). Tomato pseudo- curly top virus (TPCTV) is its lone member and type species (Pringle, 1999a; Pringle

1999b). Its genome organization and host specificity (dicotyledonous plants) are similar to the curtoviruses, but it is transmitted by a treehopper (Micrutalis malleifera).

Interestingly, some of the monopartite begomoviruses support a satellite DNA called DNA, which cannot replicate on its own but can significantly enhance the virulence of associated helper viruses (Cui et al., 2004; Saeed et al., 2007) (for more details on C1 please section 1.2.3). This report includes studies on proteins from

TGMV, BCTV, TYLCV (C1 coded by the associated satellite DNA), and ToLCV.

1.1.2 Geminivirus nomenclature, gene organization, and gene functions

Currently there are two nomenclature systems to designate geminivirus genes, which are oriented in both the directions on their circular, dsDNA replicative forms. One system names genes based on whether they are coded by the virion (V) (positive) or complementary (C) (negative) sense strand. The other names genes based on the direction they are oriented, rightward (R. viral sense, clockwise) or leftward (L, complementary sense, counterclockwise) (Figure 1.1). In both systems, genes and gene products are numbered, and in case of bipartite viruses, the prefix A or B is added to a gene depending on whether it is located on DNA A or DNA B. We use the R and L system in this document.

All geminiviruses have a non-coding intergenic region (IR) that contains the origin of replication and divergent promoters. In bipartite geminiviruses, part of IR (the

CR) is shared by both DNA A and DNA B. Interestingly, the IR is methylated as part of an epigenetic host defense response (Raja et al., 2008). The IR contains the origin of replication and is flanked by divergent RNA polymerase II promoters that drive transcription of the leftward and the rightward genes. The origin of replication contains an inverted repeat that forms a hairpin loop. The IR and hairpin structure are important for viral replication, and these are highly conserved not only in all geminiviruses, but also in the related nanoviruses of Circoviridae family (Lazarowitz, 1987; Hull 2002).

The bipartite begomoviruses encode five to six genes on DNA A and two in the

DNA B. In DNA A, one or two genes are in the rightward orientation (termed AR) and four genes are in the leftward orientation (termed AL). DNA B encodes one each in

4 rightward orientation (termed BR) and leftward orientation (termed BL) (Figure 1.1). The leftward oriented genes of DNA A are considered early genes because they are required during the early events of virus life cycle (replication and gene expression), whereas the rightward genes are late genes involved in DNA encapsidation and movement. Among early genes, AL1 (also called Rep protein) is the replication initiator protein, as is required for viral DNA replication. AL1 mediates origin recognition by binding to origin of replication, where it recruits host replication factors including proliferating cell nuclear antigen (PCNA), replication protein A (RPA), and replication factor C (RPC) (Fontes et al., 1994; Luque et al., 2002; Castillo et al., 2003; Singh et al. 2006). It is required for cleavage (site and strand-specific) and ligation events that initiate and terminate rolling circle replication (Laufs et al., 1995), and also, likely serves as a replicative helicase

(Choudhary et al., 2006; Clerot et al., 2006). To aid DNA synthesis, AL1 interacts with host retinoblastoma (Rb) protein to antagonize E2F-mediated repression of cellular replication genes such as PCNA (Nagar et al., 1995; Ach et al., 1997; Kong et al., 2000;

Egelkrout et al., 2001), and modulates metabolism through interactions with kinases

(Shen and Hanley-Bowdoin, 2006; Shen et al., 2009). AL3 is a replication enhancer which interacts with AL1, and mutation in AL3 (also known as Replication Enhancer protein REn) leads to a significant reduction in viral DNA accumulation (Sunter et al.,

1990). Recently, it has been shown that interaction of AL3 with a host NAC domain protein is responsible for enhanced viral replication (Selth et al., 2005).

AL2 is involved in transcriptional activation of late viral genes and thus is sometimes designated TrAP (Transcriptional Activator Protein) (Sunter and Bisaro,

1991) (for more details on AL2 please see section 1.2.1). The AL4 or AC4 of Old World

5 begomoviruses acts as a silencing suppressor by binding to single-stranded small interfering RNAs (siRNAs) (Chellappan et al., 2005). It also interacts with Arabidopsis shaggy-related protein kinase (AtSKeta), a component of brassinosteroid signaling pathway (Piroux et al., 2007), but the consequences of this interaction remain to be established. AR1 is the coat protein (CP) which is basic in nature and non-specifically binds to nucleic acids. BR1 is the nuclear shuttle protein (NSP) and BL1 is the movement protein (MP) that associates with plasmodesmata. In bipartite begomoviruses, only BL1 and BR1 are required for cell-to-cell movement and systemic spread of the virus

(Gardiner et al., 1988; Liu et al., 1997; Qin et al., 1998). Also, BL1 and BR1 interact with each other and it has been speculated that BR1 conveys viral genomes to BL1 for movement between cells (Rojas et al., 2005).

The genome organization of monopartite begomoviruses is similar to that of DNA

A of bipartite begomoviruses (Figure 1.1). The L1 and L3 genes are functional homologs of AL1 and AL3 but L2, unlike AL2, is not a transcriptional activator (Stanley et al.,

1992; Sunter et al., 1994; Hormuzdi and Bisaro, 1995). However AL2 and L2 share similar mechanisms in blocking host adaptive and innate antiviral defenses (Wang et al.,

2003; Hao et al., 2003; Wang et al., 2005; Yang et al., 2007; Raja et al., 2008; Buchmann et al., 2009). Lacking a B component, the rightward genes of monopartite viruses perform all movement functions. The R1 gene of BCTV encodes the coat protein, which along with R3, is required for virus movement. The protein encoded by the gene R2 is needed for accumulation of ssDNA, the encapsidated form of viral DNA (Hormuzdi and Bisaro,

1993).

1.1.3 Rolling circle replication and infection cycle of geminivirus

As mentioned above, geminiviruses depend on host replication machinery because they do not encode a DNA polymerase. The viral proteins (Rep and REn) reprogram the host cell by activating the cellular replication machinery irrespective of the cell cycle stage, and redirect this machinery to the viral genome to drive replication. Replication occurs via double-stranded DNA (dsDNA) intermediates in the nuclei of host cells by a rolling circle replication mechanism (RCR) which is a two step process (Stenger et al., 1991;

Gutierrez, 1999; Hanley-Bowdoin et al., 2000) (Figure 1.2). In the first step, ssDNA (+) is used as the template for synthesis of the complementary (-) strand, leading to formation of ds replicative forms (dsRF). In the second step, (+) ssDNA are synthesized using dsRF. The nicking-closing functions of Rep are required to initiate (+) strand synthesis, and for the synthesis of (-) strand and the post-synthesis ligation of the new 3' end to 5' end (to circularize) (Laufs et al., 1995). Therefore, AL1 is sometimes called ‗nick and stick‘ protein. The newly generated (+) ssDNA can be immediately packaged into virions or template additional (-) strands. However, dsRF, containing either open or covalently closed (supercoiled) circular forms, are organized into minichromosomes complexed with histone proteins (Bisaro et al., 1982; Pilartz and Jeske, 1992). Host methyltransferases appear to target certain histone proteins presumably to block viral DNA replication and/or gene expression (Raja et al., 2008).

As noted above, transmission of geminiviruses from plant to plant takes place via various species of leafhoppers, treehoppers, or a single species of whitefly, depending upon the genus. This transmission is persistent and circulative regardless of the particular insect species. The specificity between virus and insect vector is exclusively determined

7 by the coat protein (Briddon et al., 1990; Hohnle et al., 2001). While feeding on the phloem of the host plants, these insects also take up virus capsids. These pass through the midgut epithelium into the homocoel, and then enter the salivary glands to be delivered to new host plants when the insects next feed. These insects use their stylets to directly inject geminiviruses into phloem cells, presumably nucleated companion and phloem parenchyma cells (Jeske, 2009). Most geminiviruses are limited to phloem although some

(e.g. TGMV, CaLCuV) are found in leaf mesophyll tissues and epidermis (Rushing et al.,

1987).

1.2 Geminivirus Pathogenicity Factors

1.2.1 AL2 of TGMV and L2 of BCTV

The AL2 gene of Tomato golden mosaic virus (TGMV), a bipartite begomovirus, codes for transcriptional activator protein (TrAP), named after its first known function (Sunter and Bisaro, 1991). As the name suggests, it is involved in the transcriptional activation of late viral genes and therefore is indispensable to the virus infection cycle (Sunter and

Bisaro, 1992). However, L2, a positional homolog of AL2, is a gene encoded by Beet curly top virus (BCTV), a monopartite curtovirus. Unlike AL2, L2 is not a transcriptional activator and is dispensable for the virus infection cycle.

AL2 protein has a basic region near the N-terminus that contain a nuclear localization signal (NLS), a central zinc finger-like motif (C-X1-C-X4-H-X2-C) and an acidic C-terminal region that contains the minimal transcription activation domain

(Hartitz et al., 1999; Figure 1.3). The central CCHC motif is required for zinc and DNA binding, and the cysteine and histidine residues present in this motif are conserved in most of the begomovirus AL2 proteins (Hartitz et al.,1999; Dong et al., 2003). In comparison, L2 shares only the central CCHC motif and lacks the function of transcriptional activation (Figure 1.3). The self-interacting AL2 localizes to the nucleus and this dimerization, mediated by the central CCHC motif, is required for proper subnuclear localization and for efficient transcriptional activation (Yang et al., 2007). By contrast, L2 does not dimerize although it contains the central CCHC motif. L2 also shows nuclear localization in transient assays when fused to double GFP (Gireesha et al., unpublished data). As pathogenicity factors, AL2 and L2 share some, but not all, mechanisms (Raja et al., 2010).

AL2 is a non-canonical transcription factor in that it only weakly binds ssDNA and dsDNA in a sequence non-specific manner, and is likely recruited to responsive promoters by interactions with cellular proteins (Hartitz et al., 1999; Lacatus and Sunter,

2008; Lacatus and Sunter, 2009). AL2 acts as a RNA silencing suppressor both in transcription-dependent and transcription-independent manners. AL2 does not bind siRNA or miRNA (Chellappan et al., 2005; Wang et al., 2005). Several studies have demonstrated that the ability of AL2 to reverse post transcriptional gene silencing

(PTGS) and inhibit its systemic spread requires an intact NLS, the CCHC motif, and the transcription activation domain, suggesting that silencing suppression in these contexts requires AL2 to stimulate transcription. More specifically, it is presumed that AL2 activates the transcription of cellular genes that negatively regulate silencing pathways

(van Wezel et al., 2002; Dong et al., 2003; Trinks et al., 2005). This mode of silencing

9 suppression has been termed transcription-dependent, although interactions with silencing pathway components through the activation domain cannot formally be ruled out (Bisaro, 2006).

In a transcription-independent manner, AL2 and L2 can block transcriptional

(TGS) and post-transcriptional gene silencing (PTGS) and can reverse TGS that is previously established (Wang et al., 2005; Buchmann et al., 2009). These activities of

AL2 and L2 are accomplished, at least in part, through interaction with and inactivation of ADK, an important plant kinase (Wang et al., 2003; Wang et al., 2005). The ability to suppress RNA silencing is shared by AL2 and L2 proteins of most geminiviruses

(Brigneti et al., 1998; Voinnet et al., 1999; van Wezel et al., 2002; Van Wezel et al.,

2003;Wang et al., 2005). In particular, methylation of viral DNA and associated histone proteins is a key mechanism of host defense against DNA viruses, and these viral proteins have evolved to block this methylation presumably by targeting different members of the methylation pathway (Raja et al., 2008; Raja et al., 2010). Also, both

AL2 and L2 interact with and inactivate SnRK1, one more important plant kinase (Hao et al., 2003). Interactions with SnRK1 inhibit the cellular stress response, which appears to be a component of plant basal defenses (Hao et al., 2003) (For more details please section

1.3.8).

Inactivation of SnRK1 by AL2 and L2 is responsible for the enhanced susceptibility (ES) phenotype caused by these viral proteins in transgenic Nicotiana benthamiana plants when challenged by TGMV, BCTV, and TMV viruses (Sunter and

Bisaro, 2001; Hao et al., 2003). This ES is mainly characterized by a reduction in mean latent period (1 to 9 days depending on the virus tested) and by a decrease in the

10 inoculum concentration required to infect these transgenic plants (ID50 reduced 6- to 60- fold). However, disease symptoms and virus replication levels were not significantly increased.

1.2.2 AV2, pre-coat protein of certain geminiviruses

Geminiviruses replicate in the nucleus and therefore their systemic infection requires movement from cell-to-cell as well as from the nucleus to cytoplasm. The proteins required for movement are encoded by the B genome of the bipartite viruses, and two virion-sense genes of monopartite begomoviruses. These virion-sense genes code for the pre-coat (AV2 or V2) and coat proteins (CP) (Briddon et al., 1989; Rothenstein, et al.,

2007). The movement proteins encoded by plant viruses are often involved in pathogenesis and are known to help virus overcome host defense responses (Brough et al., 1988; Pascal et al., 1993; Brigneti et al., 1998). AV2 has been shown to be a pathogenicity determinant in Papaya leaf curl virus (PaLCuV) and Cotton leaf curl khokhran virus (CLCuKV), and is also known to induce a hypersensitive response (HR) in some cases (Mubin et al., 2010). In addition, AV2 of Tomato yellow leaf curl virus

(TYLCV) can suppress post-transcriptional gene silencing (PTGS) (Zrachya et al., 2007).

In this study, we have made an attempt to identify interacting partners of Tomato leaf curl

New Delhi virus (ToLCNDV) AV2.

1.2.3 C1, a protein encoded by satellite DNA 

DNA beta (DNA) is a circular single-stranded DNA satellite (~1.35 kb) with a single open reading frame that codes for a protein termed C1. These satellite DNAs have been

11 found to be associated with some monopartite begomovirus infections (Saeed et al.,

2007). So far, more than 100 full length DNA  genomes have been cloned and sequenced (Mansoor et al., 2006). Interestingly, they do not share sequence homology with begomoviruses except for the conserved step-loop structure that forms the replication of origin in all geminiviruses. However, different DNA s share 44-99% sequence identity among themselves (Briddon et al., 2003). These satellite DNAs depend on their helper virus for replication and encapsidation but contribute to the production of symptoms and enhanced accumulation of the helper virus (Saeed et al., 2007). For example, Tomato yellow leaf curl China virus (TYLCCNV) does not cause severe symptoms on its own but when associated with DNA can cause severe symptoms (Cui et al., 2004). C1 is a pathogenicity determinant and a suppressor of RNA silencing

(Saunders et al., 2000; Briddon and Stanley 2006). It can also block and reverse transcriptional gene silencing at least in part by inhibiting methylation (Yan et al., unpublished data). C1 can also substitute DNA B genome of bipartite begomoviruses in systemic infection, suggesting a role for this protein in virus movement (Saeed et al.,

2007)

1.3 SnRK1, a Ser/Thr Protein Kinase

1.3.1 Protein kinases, key molecular switches involved in the regulation of cellular processes

Kinases (phosphotransferases) are enzymes that transfer phosphate groups from high energy donor molecules, such as ATP, to substrates in a process called phosphorylation.

Protein kinases comprise the greater part of the known kinases, and these enzymes add inorganic phosphate to either themselves (autophosphorylation) and/or to other protein substrates. The reversal of phosphorylation (removal of phosphate) is called dephosphorylation, and the enzymes that carry out this function are called phosphatases.

Most of the known protein kinases phosphorylate serine or threonine residues and the remaining ones phosphorylate either tyrosine or histidine residues. Post-translational modification of proteins through reversible phosphorylation is a universal mechanism for regulating diverse biological functions (Smith and Walker, 1996). Addition or removal of phosphates to proteins can induce variety of changes, including changes in protein conformation, protein binding, protein recognition, enzyme activity, etc. Therefore, protein kinases are critical ―molecular switches‖ involved in the regulation of various cellular functions.

1.3.2 SnRK1, a member of SNF1-AMPK family of kinases

Members of the SNF1/AMPK family of kinases are highly conserved. Representatives include SNF1 kinase (sucrose non-fermenting1) in yeast, SnRK1 (SNF1-related kinase1) in plants, and AMPK (AMP-activated protein kinase) in animals (Halford and Hardie,

1998; Polge and Thomas, 2007). These Ser/Thr kinases play a central role in the regulation of metabolism, and in response to nutritional and environmental stresses that deplete ATP, they turn off energy-consuming biosynthetic pathways and turn on alternative ATP-generating systems (Halford and Hardie, 1998; Hardie et al., 1998;

Radchuk et al., 2006; Baena-Gonzalez et al., 2007; Polge & Thomas, 2007). This is often

13 called cellular stress response (CSR). In plants, for example, SnRK1 kinases phosphorylate and inactivate key enzymes that control steroid and isoprenoid synthesis, nitrogen assimilation for amino acid and nucleotide synthesis, and sucrose biosynthesis

(Polge and Thomas, 2007; Sugden et al., 1999b).

The kinase catalytic domains of SnRK1, AMPK and SNF1 share 62-64 % amino acid identity (Halford and Hardie, 1998), and also share some similarities with animal calmodulin-dependent protein kinases (CaMKs) and plant calmodulin domain protein kinases (CDPKs) (Halford and Hardie, 1998). Therefore, the SNF1-AMPK family has some overlap in substrate specificity with CaMKs and CDPKs. For example, nitrate reductase in spinach appears to be regulated through phosphorylation by both CDPKs and

SnRK1 (Douglas et al., 1997).

1.3.3 Classification of SnRK1 in plants

Arabidopsis SnRK1 kinases have been placed in the CDPK-SnRK superfamily, which consists of seven types of serine-threonine protein kinases (Hrabak et al., 2003). There are 38 members in the SnRK family, subdivided into three groups called SnRK1, SnRK2, and SnRK3 based on sequence similarity and domain structure (Halford and Hardie,

1998; Hrabak et al., 2003). The SnRK1 subgroup is the most closely related to SNF1 of yeast and AMPK of mammals. SnRK1, SnRK2 and SnRK3 groups have three, ten, and twenty five members, respectively. Only two SnRK1 members called SnRK1.1 (also

Arabidopsis kinase 10; AKIN10) and SnRK1.2 (also called AKIN11) are expressed. Both the genes encode protein kinases that are 512 amino acid long (~ 58 kDa), and these

14 kinases display 89% amino acid identity in the N-terminal portion (aa 1-343), which contains the kinase catalytic domain. Numerous exchanges are found in the C-terminal third between these proteins which contain regulatory domains; nevertheless, they share

64% identity in this region (Hao et al., 2003). Both Arabidopsis SnRK1 (AKIN10 and

AKIN11) can complement the yeast SNF1 mutants, suggesting a similar function for these proteins (Bhalerao et al., 1999). SnRK1.2 (AKIN11) is the protein kinase used in the studies described here, and throughout this document it is referred to as SnRK1 for simplicity.

1.3.4 Regulation of SnRK1 expression and activity

SnRK1 is subject to transcriptional and post-transcriptional regulation in response to growth and development, and various stress conditions (Halford et al., 2003). It is known to be expressed throughout the plant but the expression is not uniform. For example, in potato, expression is highest in stolons as they start developing into tubers, and is lowest in leaves (Man et al., 1997). However, in maturing tubers SnRK1 expression declines. In these studies, 40-fold higher more SnRK1 kinase activity was observed in mini-tubers than in mature tubers although transcript levels were approximately the same (Man et al.,

1997), indicating post-transcriptional regulation. In Arabidopsis, SnRK1 activity changes under phosphate starvation (nutrient stress) conditions without changes in gene expression and intriguingly, the activity of SnRK1.2 (AKIN11) was reduced by 35% to

40% in stark contrast to a 100% increase in SnRK1.1 (AKIN10) activity (Fragoso et al.,

2009). However, inducing metabolic stress in Arabidopsis cells with 2-deoxyglucose, an inhibitor of glycolytic ATP production, led to increased SnRK1 activity (Harthill et al.,

2006). Also, challenging plants with geminiviruses (a biotic stress) appears to upregulate

SnRK1 expression in Nicotiana benthamiana (G. Sunter, unpublished data). However, the exact nature of the signals that lead to regulation of SnRK1 expression is not yet known.

1.3.5 Role of increased AMP/ATP ratio in activating SNF1-AMPK family of kinases

In eukaryotic cells, the levels of AMP, ADP and ATP are in the state of dynamic flux, influenced by diverse cellular stimuli. Under optimal conditions, the typical AMP: ATP ratios are approximately in the order of 1:100 (Hardie et al., 1998). Thus, cells maintain very low levels of AMP in a fully energized status. But under stress conditions, the ratio of AMP:ATP increases, leading to the cellular stress response (CSR). The increase in

AMP concentration activates SNF1-AMPK kinases which subsequently turn off ATP consuming biosynthetic pathways and turn on alternative ATP-generating systems

(Hardie et al., 1998; Halford and Hardie, 1998). However, the mechanisms that activate

SNF1/AMPK/SnRK1 complexes are not completely understood and differences are likely to exist between kingdoms.

1.3.6 SNF1-AMPK family kinases function as heterotrimeric complexes

The members of SNF1/AMPK/SnRK1 kinases function as heterotrimeric complexes comprised of an  catalytic subunit, a  subunit and a  subunit that appears to bridge the

 and  subunits. Yeast has a single  subunit (SNF1), but multiple genes encoding , ,

16 and  subunits are typically found, suggesting multiple combinations of isoforms and potential for complex and subtle regulation (Hardie, 2007). Specifically, Arabidopsis has two genes each coding for  (SnRK1.1 and SnRK1.2) and  subunits while three genes coding for  subunits (Polge and Thomas, 2007). The catalytic  subunits consist of an

N-terminal kinase domain with an activation loop that contains a conserved threonine residue that must be phosphorylated for activity. The C-terminal domain is required for interaction with the  and  subunits. Interestingly, the  subunits contain tandemly repeated cystathionine -synthase motifs, which act in pairs to form Bateman motifs that bind ATP or AMP in a mutually exclusive manner, providing a structural basis for energy sensor function (Scott et al., 2004).

1.3.7 Mechanisms of SnRK1 activation and known SnRK1 targets

First, all members of the SNF1-AMPK family are activated by upstream kinases through phosphorylation of the Thr residue in the conserved T-loop of  subunit (activation loop)

(Shen et al., 2009). This results in autophosphorylation followed by subsequent phosphorylation of the downstream targets. In Arabidopsis, Geminivirus Rep-interacting kinase 1 and 2 (GRIK1 and GRIK2) activate SnRK1 by phosphorylating T176 in the T- loop (Shen et al., 2009), and these kinases are related to yeast SNF1- and mammalian

AMPK-activating kinases (Shen and Hanley-Bowdoin, 2006). Following phosphorylation by upstream kinase/s, 5'-AMP (also referred to as AMP in this document) allosterically stimulates AMPK activity and inhibits its inactivation due to dephosphorylation of  subunit at Thr172 (activation loop) by protein phosphatase 2C (PP2C) (Davies et al.,

1995; Suter et al., 2006; Sanders et al., 2007). Recent evidence indicates that these allosteric effects are mediated by AMP binding to the -subunit (Sanders et al., 2007).

Direct allosteric stimulation of the SnRK1 complex has yet to be demonstrated. However,

AMP has been shown to suppress dephosphorylation of SnRK1 by mammalian PP2C

(Sugden et al., 1999a). Because ADK generates 5'-AMP, this study hypothesizes that

ADK contributes to rapid activation of the CSR.

In plants, very few SnRK1 substrates and interacting partners have so far been identified. In vitro, SnRK1 phosphorylates and inactivates four important biosynthetic enzymes (Polge and Thomas, 2007): (i) 3-hydroxymethyl-3-methylglutaryl-CoA reductase (HMGR), which catalyzes the NADH-dependent reduction of 3-hydroxy-3- methylglutaryl-coenzyme A (HMG-CoA) to mevalonic acid (Dale et al., 1995 a & b); (ii) sucrose phosphate synthase (SPS), which catalyzes sucrose biosynthesis (McMichael et al., 1995); (iii) nitrate reductase (NR), which catalyzes the first step of nitrogen assimilation into amino acids (Douglas et al., 1995; Bachmann et al.,1996b); (iv) trehalose phosphate synthase 5 (TPS5), which is involved in the synthesis of trehalose-6- phosphate, a signaling sugar that regulates plant metabolism and development (Harthill et al., 2006). However, SnRK1 inactivation of some of these targets require 14-3-3 protein binding to the phosphorylated sites (Bachmann et al., 1996a; Moorhead et al., 1996;

Sugden et al., 1999b; Ikeda et al., 2000; Huber et al., 2002; ). In wheat, 14-3-3 proteins also bind to autophosphorylated SnRK1 and this binding does not affect SnRK1 activity in vitro (Ikeda et al., 2000). Interestingly, SnRK2.8, a member belonging to SnRK2 family has been shown to phosphorylate three different proteins belonging to 14-3-3 family (Shin et al., 2007). Although SnRK1 interacts with 14-3-3 proteins in wheat, it is

18 not known whether SnRK1 phosphorylates those proteins (Ikeda et al., 2000). SnRK1 also phosphorylates barley heat shock protein 17 (BHSP17), but the consequences of this phosphorylation remain to be established (Slocombe et al., 2004).

Recent evidence indicates that SnRK1 integrates diverse energy and stress signals on a global level by inducing extensive changes in the transcriptome that have the effect of promoting catabolism and inhibiting anabolic pathways (Baena-Gonzalez et al., 2007:

Baena-Gonzalez and Sheen, 2008). SnRK1 also indirectly controls carbohydrate metabolism by modulating the transcription of sucrose synthase (sucrose degradation) and -amylase (starch degradation) (Purcell et al., 1998; Laurie et al., 2003). Therefore,

SnRK1 protein kinases are considered global regulators of carbon metabolism.

1.3.8 Role of SnRK1 in antiviral defense

From earlier studies it is evident that SnRK1 plays a role in innate antiviral defense against geminiviruses and RNA viruses (Sunter et al., 2001; Hao et al., 2003). Nicotiana benthamiana transgenic plants expressing an antisense SnRK1 (earlier referred to as

SNF1) displayed enhanced susceptibility (ES) to geminivirus (TGMV and BCTV) infection (Hao et al., 2003). Such an ES does not increase the disease symptoms, rather, it is characterized by significant decreases in mean latent period (~10 to 16% for TGMV and 13 to 27% for BCTV) and by a decrease in the amount of virus inoculum required to infect the transgenic plants (ID50;50% infectious dose observed at 25 to 250-fold dilution of the standard inoculum dose of OD600=1) (Hao et al., 2003). Similar ES was observed when transgenic N. benthamiana plants expressing geminivirus AL2 (truncated: AL21-100,

19 lacking the transcription activation domain) and L2 proteins, were infected with DNA and RNA viruses (Sunter et al., 2001). The ES in this study was characterized by a reduction in the mean latent period (from 1 to 9 days) and by a decrease in the ID50 (6-to

60-fold), without enhancement of disease symptoms or virus load (Sunter et al., 2001).

Conversely, transgenic plants overexpressing SnRK1 displayed enhanced resistance (ER) where in the ID50 was not observed upon infecting the plants with standard inoculum doses of BCTV (Hao et al., 2003). Interestingly, transgenic N. benthamiana plants expressing SnRK1-interaction defective AL2 (AL2 Δ33-43 in a 1-114 background) failed to exhibit ES when infected with geminiviruses (Hao et al., 2003). All these observations suggested that the cellular stress response (CSR) mediated by SnRK1 is a component of innate antiviral defense. Through their pathogenicity factors, geminiviruses are able to counteract CSR by inhibiting SnRK1. However, the specific downstream effects of

SnRK1 inactivation by the geminivirus proteins are not known. Attempts to reveal the potential SnRK1 downstream targets most relevant to viruses are described in this study.

1.4 Adenosine Kinase (ADK) and Adenine Phosphoribosyltransferase (APT)

1.4.1 ADK generates 5'-AMP, and is involved in cytokinin metabolism

Adenosine kinase (ADK; EC 2.7.1.20; adenosine 5'-phophotransferase) is an abundant purine kinase present in all eukaryotic cells examined to date. In the presence of magnesium, this enzyme catalyzes the transfer of -phosphate from ATP or GTP to adenosine to produce 5'-AMP (Moffatt et al., 2000). Arabidopsis has two ADK genes,

ADK1, ADK2 which share 89% nucleotide sequence identity and 92% amino acid

20 identity. Both ADKs can metabolize adenosine and cytokinin substrates although the former seems to be preferred over the later (Moffatt et al., 2000). ADK1 and ADK2 are expressed constitutively in Arabidopsis plants with the highest steady-state mRNA levels found in stem and root, but generally ADK1 transcript level is higher than that of ADK2.

ADK enzyme activity levels correlate well with protein expression levels in stems, leaves, and flowers, but weakly so in root, siliques, and dry seeds (Moffatt et al., 2000).

ADK plays a critical role in the adenine and adenosine salvage pathways and is important for the synthesis of nucleic acids and nucleotide cofactors. It also plays major roles in sustaining the methyl cycle, the gravitropic response, and cytokinin (CK) regulation in plants (von Schwartzenberg et al., 1998; Moffatt et al., 2000; Moffatt et al.,

2002; Young et al., 2006). ADK‘s roles in sustaining methyl cycle (see section 1.4.3) and cytokinin regulation are components of host defense against geminiviruses (Wang et al., 2003; Wang et al., 2005; Raja et al., 2008; Baliji et al., 2010).

Cytokinins are phytohormones that play a critical role in regulating the plant cell proliferation and differentiation, and also control diverse processes in plant growth and development (Sakakibara, 2006). These hormones are N6-substituted adenine derivates which exist as free base, nucleoside, and nucleotide forms (von Schwartzenberg et al.,

1998; Sakakibara, 2006). The phosphorylated nucleotide form is inactive compared to other forms, and therefore interconversion of these forms determines bioactive pool levels of cytokinins (Sakakibara 2006; Baliji et al., 2010). ADK converts the active nucleoside form of cytokinin to an inactive nucleotide form, and this appears to be part of the host defense against geminivirus infection (Baliji et al., 2010). Because cytokinins promote cell division and differentiation, it would be advantageous for DNA-containing

21 geminiviruses to keep these hormones in an active state to facilitate their replication and multiplication. In fact, geminivirus proteins AL2 and L2 interact with and inactivate

ADK (Wang et al., 2003) and this inactivation leads to increased expression of cytokinin- responsive genes (Baliji et al. 2010). Also, exogenous application of cytokinin results in increased susceptibility to geminivirus infection characterized by reductions in mean latent period and enhanced viral replication (Baliji et al., 2010).

1.4.2 ADK plays a critical role in methylating geminivirus genomes as part of the

host defense

Geminiviruses encapsidate single-stranded (ss) DNA genomes that replicate in plant cell nuclei through double-stranded (ds) DNA intermediates that associate with cellular histone proteins to form minichromosomes (Pilartz and Jeske, 1992; Pilartz and Jeske,

2003). Plants methylate both the viral DNA and associated histone proteins as part of an epigenetic host defense (Raja et al., 2008). ADK is required for efficient production of the methyl donor S-adenosyl methionine (SAM) and therefore, methylation-deficient adk mutant plants display hypersensitivity to geminivirus infection (Moffatt et al., 2002; Raja et al., 2008). This is mostly due to the decreased ability of the mutant plants to methylate the viral genome (Raja et al., 2008). As discussed above, inhibition of ADK also leads to increased cytokinin activity that plays a role in virus replication (Baliji et al., 2010). As a counterdefense mechanism, geminivirus proteins AL2 and L2 act to inhibit this methylation, partly by inactivating ADK (Wang et al., 2003: Raja et al., 2008). Through this inactivation, these viral proteins also act as RNA silencing suppressors of both transcriptional gene silencing (TGS) and post-transcriptional gene silencing (PTGS)

(Wang et al., 2005: Yang et al.,2007; Buchmann et al., 2009). Overall, the geminiviruses derive multiple benefits by inactivating ADK and one of the aims of this study is to reveal one more potential benefit of ADK inactivation, that in turn might inactivate

SnRK1-controlled innate antiviral defense mechanisms.

1.4.3 Roles of ADK in mammals

In mammals, adenosine kinase helps regulate intra- and extracellular levels of adenosine which has widespread effects on the cardiovascular, nervous, respiratory, and immune systems (Fox and Kelley, 1978; Berne, 1993; Chang et al., 1983). For example, elevated levels of adenosine in the heart and brain are associated with the attenuation of ischemic injury, and inhibition of ADK has been shown to markedly increase the therapeutic effect of adenosine in the animal models (Jiang et al., 1997; Kowaluk, et al., 1998; Martin et al.,

1997; Tatlisumak et al., 1998). Interestingly, a recent study has identified adenosine as a central player in mediating some of the effects of acupuncture (Goldman et al., 2010).

Therefore, most studies of ADK in mammals have been focused on its inhibitors to derive therapeutic benefits. (McGaraughty et al., 2001; Wiesner et al., 1999; Zhu et al.,

2001)

1.4.4 APT generates 5'-AMP, and is involved in cytokinin metabolism

Adenine phosphoribosyltransferase (APT: pyrophosphate phospho-D-ribosyltransferase;

EC 2.4.2.7) is a constitutively expressed enzyme that salvages adenine in one-step to 5'-

AMP. This reaction involves Mg2+-dependent transfer of the phosphoribosyl group of 5-

23 phosphorylribosyl 1-pyrophosphate (PRPP) to adenine (Allen et al., 2002). The

Arabidopsis genome has five ‗APT-like‘ or ‗putative APT‘ genes designated APT1,

APT2, APT3, APT4 and APT5. Of these, evidence of gene expression is available only for

APT1, APT2 and APT3 (Allen et al., 2002). APT1 metabolizes adenine at a rate 30-50 times faster than APT2 and APT3, respectively (Allen et al., 2002), and APT1 deficiency leads to male sterility in Arabidopsis (Gaillard et al., 1998). APT, like ADK, also contributes to 5'-AMP pools and thus might also play a role in activating the cellular stress response (CSR), which is triggered by increased AMP/ATP ratio.

APT is also involved in the metabolism of cytokinins (Allen et al., 2002;

Sakakibara, 2006), the phytohormones which play a crucial role in various phases of plant growth and development (Sakakibara, 2006). APT plays a role in converting the active form (base) of cytokinin to the inactive form (nucleotide). Therefore, we investigated if APT is also a target of geminivirus counterdefense strategies.

1.5 Defense Responses Against Viruses

1.5.1 PKR-mediated interferon (IFN) response, a key antiviral defense in animals

In eukaryotes, viruses need to overcome various host-acquired humoral and cell-mediated immune responses to cause an infection in their hosts (Gale and Katze, 1998). In animals, they must first counteract the innate antiviral defense provided by the cellular interferon

(IFN) system. IFNs are a family of cell-secreted cytokines found in vertebrates that induce the expression of several genes some of which have been extensively characterized (Sen and Ransohoff, 1993; Muller et al., 1994; Jaramillo et al., 1995). IFN-

24 induced gene products play antiviral roles primarily by blocking viral gene expression at the transcriptional (blocking RNA transcription, degrading viral transcripts, etc) and translational levels. Protein kinase RNA-activated (PKR) is a key IFN-induced gene product which limits viral protein synthesis and thus plays a role in providing the first level of host defense against viruses (Gale and Katze, 1998). PKR, general control non- derepressible-2 (GCN2) of yeast, and reticulocyte-specific HR1 are serine-threonine protein kinases that belong to the eukaryotic translation initiation factor-2 subunit (eIF-

2) protein kinase family (Chen et al., 1991; Dever et al., 1992; Clemens et al., 1993;

Hinnebusch, 1994).

eIF-2 is a heterotrimeric complex comprising an  subunit, a  subunit, and a  subunit (Kimball, 1999). It forms a complex with GTP and mediates the binding of the methyl initiator tRNA to the 40S ribosomal subunit, to form a pre-initiation complex

(PIC). PIC then binds to the 5' end of the mRNA through eIF3 and eIF4 complexes to form translation initiation complex (Kimball 1999; Alberts et al., 2002; Robaglia and

Caranta, 2006; Figure 1.4 A), which then begins scanning along the mRNA. When an

AUG codon is recognized, eIF5 binds to eIF2 and stimulates the hydrolysis of eIF2- bound GTP. This results in a conformational change in eIF2 causing its release from the

40S ribosomal subunit. The 60S ribosomal subunit then joins the 40S to form a complete ribosome that begins protein synthesis. (Figure 1.4 B; Kimball, 1999; Robaglia and

Caranta, 2006).

Because eIF-2 binds very tightly to GDP, eIF-2B, a guanine nucleotide exchange factor, is required to cause GDP release so that a new GTP molecule can bind eIF-2 to enable its reuse. However, phosphorylation of the  subunit of eIF-2 causes it to bind

25 tightly to eIF-2B. This binding inactivates eIF-2B and thus, prevents reuse of eIF-2.

Relatively, eIF-2 is more abundant than eIF-2B in cells, and therefore, even a fraction of phosphorylated eIF-2 can trap nearly all of the eIF-2B. This prevents the reuse of the non-phosphorylated eIF-2 and greatly slows protein synthesis (Alberts et al., 2002).

PKR is expressed ubiquitously at low levels virtually in all mammalian cells, but its expression level is increased by IFN (Baier et al., 1993; Clemens et al., 1993). Binding to either ds or ss RNA containing regions of extensive secondary structure leads to dimerization and rapid activation of PKR. This activation results in PKR autophosphorylation and subsequent phosphorylation and inactivation of eIF-2 (Mellits et al., 1990; Green and Mathews, 1992; Clemens et al., 1994; McCormack and Samuel,

1995).

Like many other kinases, PKR is subject to transcriptional and post- transcriptional regulation. However, PKR, unlike many other members of protein kinase superfamily, is also regulated by viral-encoded inhibitory molecules (Clemens and Elia,

1997). Virtually all mammalian viruses have developed diverse strategies to successfully counteract the deleterious effects of PKR-mediated protein synthesis inhibition (Gale and

Katze, 1998). Different virus-encoded inhibitors block PKR-mediated eIF-2 phosphorylation at different steps, starting from transcriptional regulation of PKR to counter-regulation of eIF-2.

1.5.2 PKR pathway in plants

In plants, a PKR gene or its functional homolog remains to be identified, although PKR- like activity has been demonstrated in in vitro studies using wild type (51S) and mutant

(51A) eIF-2 substrates (Chang et al., 1999), and during Tobacco mosaic virus (TMV) and Potato spindle tuber viroid infections (Hiddinga et al., 1988; Crum et al., 1988;

Langland et al., 1996). Like mammalian PKR, a plant protein with PKR-like activity

(pPKR) also binds to dsRNA, and monoclonal antiserum to the human PKR as well as antiserum to a conserved dsRNA-binding domain present on mammalian PKR cross- reacts with pPKR. Also, polyclonal antiserum to pPKR cross-reacts with mouse PKR and human PKR (Langland et al., 1995). Nevertheless, the gene encoding pPKR remains to be identified.

Arabidopsis does possess a GCN2 ortholog (AtGCN2) that mediates Arabidopsis eIF-2 (AteIF-2) phosphorylation in response to amino acid and purine starvation, UV, cold shock, and wounding (Lageix et al., 2008). However, infection of Arabidopsis plants with Turnip yellow mosaic virus or Turnip crinkle virus did not lead to AteIF-2 phosphorylation (Zhang et al., 2008). Another member of the PKR pathway, P58IPK, an ortholog of PKR-inhibitor, has been identified in plants, and interestingly, it plays a role in viral pathogenesis (Bilgin et al., 2003). Specifically, viral infection of P58IPK-silenced

Arabidopsis and N. benthamiana plants led to host death and this response was associated with phosphorylation of eIF-2. Apparently, viral induction of a pPKR activity, if not

IPK balanced by P58 , causes host cell death.

Due to the apparent absence of PKR (and some components of mammalian translational machinery) in plants as revealed by Arabidopsis and rice genome data, it has

27 been suspected that eIF-2 phosphorylation might not be an important regulatory mechanism in plants (Browning, 2004). Arabidopsis encodes two eIF-2 genes (Zhang et al., 2008), called eIF-2 and eIF-22, which share 84% amino acid identity. eIF-2 and eIF-22 encode proteins of approximate mass of 38 kD and 28 kD, respectively. There is no evidence on whether eIF-22 is a functional protein. By contrast, eIF-2 has been shown to encode a functional protein (Browning, 1996; Zhang et al., 2008; Lageix et al.,

2008), and therefore was selected for this study. In the present study, we propose that

SnRK1 may play a role similar to PKR in plants and participate in the inhibition of protein synthesis upon infection by a virus, based on in vitro phosphorylation data obtained using partially purified SnRK1 kinase domain and Arabidopsis eIF-2and eIF-

(iso)4E.

1.5.3 Roles of translation initiation factors in plant antiviral defense

Naturally occurring virus-resistance genes can provide efficient barriers against viral infections. However, recessive virus-resistance genes are far more prevalent than dominant-resistance genes compared to resistance genes for fungal or bacterial pathogens

(Robaglia and Caranta, 2006). Unlike pathogenic fungi or bacteria, viruses encode very few proteins (4-10) and recruit various host components to complete their infection cycle.

It has long been proposed that any mutation or loss of host component/s required for virus replication would lead to resistance, and in fact the data accumulated during the last two decades corroborate this hypothesis (Fraser, 1986; Robaglia and Caranta, 2006).

Analysis of recessive resistance genes identified so far in natural crops has revealed a

28 group of proteins belonging to the eukaryotic initiation factor 4E (eIF4E) gene family, an essential component of the translational machinery (Robaglia and Caranta, 2006).

eIF4E is a small protein that binds to the 5' cap structure (m7G group) of mRNAs and associates with eIF4G (a scaffold for other members of the translation initiation complex) to form what is called the eIF4F complex (Figures 1.4 A) (Browning 2004;

Robaglia and Caranta, 2006). Interaction of the eIF4F complex with mRNA is considered the first step in the initiation of translation (Rogers et al., 2002). This complex, along with mRNA, then binds to the 40S ribosomal subunit through the eIF-3 complex, to form a translation initiation complex ultimately leading to protein synthesis (Robaglia and

Caranta, 2006; Figure 1.4). The eIF4F complex is common to all eukaryotes, but plants have an additional complex called eIF(iso)4F, comprised of eIF(iso)4E and eIF(iso)4G

(Browning et al., 1992) (Figure 1.5). The eIF(iso)4F and eIF4F complexes have similar in vitro activities (Browning 1996). Arabidopsis has three eIF4E genes (eIF4E1, eIF4E2 and eIF4E3) and one eIF(iso)4E (Figure 1.5), all of which have been shown to play a role in antiviral defense. We studied eIF-(iso)4E because it is unique to plants.

Most recessive resistance linked to eIF4E in plants has been identified against potyviruses, the largest group of plant viruses. These viruses have single-stranded, positive sense RNA genomes. The 5' end of the genomes lacks a cap structure and instead is bound by a virus-encoded protein called VPg. The 3' end of the genome has a poly (A) tail. VPg of several potyviruses is known to bind eIF4E and this interaction is essential for viral infection (Leonard et al., 2000). A possible role of eIF4E in potyvirus replication may be to allow circularization of genomic RNA by interacting with VPg and polyA

29 binding protein (PABP), similar to its normal role of binding the 5' cap structure and 3'- polyA associated PABP.

The recessive resistance observed in several plants against potyviruses corresponds to mutations in eIF4E members. Examples include the pepper (Capsicum annum) pvr2 resistance gene against Potato virus Y (PVY) and Tobacco etch virus (TEV)

(Ruffel et al.,2002), an ethyl methane sulfonate (EMS)-induced mutation affecting susceptibility to TEV, Turnip mosaic virus (TuMV) and Lettuce mosaic virus (LMV) in

Arabidopsis (Lellis et al., 2002; Duprat et al., 2002), the sbm1 gene for pea resistance to

Pea seed-borne mosaic virus (PSbMV), and the pot-1 gene for tomato resistance to PVY and TEV (Nicaise et al., 2003; Gao et al., 2004; Ruffel et al., 2005), etc. This dissertation suggests a novel role for eIF4E and/or eIF(iso)4E in geminivirus pathogenesis.

Figure 1.1 Geminivirus gene organization

Maps of the gene organization of Tomato golden mosaic virus (TGMV, a bipartite, begomovirus), Tomato yellow leaf curl virus (TYLCV, a monopartite begomovirus), and

Beet curly top virus (BCTV, a curtovirus). The solid arrows indicate the positions of viral genes with the approximate molecular mass of each encoded protein given in kilodaltons

(kD). Viral genes are designated by a number and the direction of transcription from the double stranded intermediate: leftward (L, complementary sense) or rightward (R, viral sense). An asterisk and a hatched box within the intergenic region (IR) indicate the position of the conserved hairpin and the common region (CR), respectively (Bisaro,

1996).

Rep: replication initiator protein, TrAP: transcriptional activation protein, REn: replication enhancer, NSP: nuclear shuttle protein, CP: coat or capsid protein.

Figure 1.1 Geminivirus gene organization.

Figure 1.2 Geminivirus replication.

Geminivirus DNA replication occurs in two stages. First, the ssDNA is converted into dsDNA replicative form (RF). The dsRF serves as template for viral gene expression and for rolling circle replication to produce new single-stranded DNA (ssDNA) products.

Both stages require various cellular factors. Progeny ssDNA can either re-enter the DNA replication pool, associate with coat protein, or be transported outside the nucleus and to the neighboring cells with the help of the viral movement protein BR1. The role of other viral proteins in the replication is indicated by black arrows, and for a detailed description please see section 1.1.2 (Figure provided by D.M. Bisaro).

Figure 1.2 Geminivirus replication.

Figure 1.3 AL2 and L2 proteins

The diagram depicts domain organization of AL2 protein and its positional homolog L2 protein. The transcriptional activator protein AL2 has a basic region near the N-terminus, a zinc finger-like motif (C-X1-C-X4-H-X2-C) in the center, and an acidic C-terminal region that contains the minimal transcription activation domain. In comparison, L2 shares only the central CCHC motif.

Figure 1.3 AL2 and L2 proteins

Figure 1.4 Role of eIF2 in translation initiation.

A. eIF-2 is a heterotrimeric complex comprising an  subunit, a  subunit, and a 

subunit. It forms a complex with GTP and mediates the binding of the methyl

initiator tRNA to the 40S ribosomal subunit, to form a pre-initiation complex

(PIC). PIC then binds to the 5' end of the mRNA through eIF3 and eIF4

complexes to form translation initiation complex.

B. The translation initiation complex, once formed, begins scanning along the

mRNA. When an AUG codon is recognized, eIF5 binds to eIF2 and stimulates the

hydrolysis of eIF2-bound GTP. This results in a conformational change in eIF2

protein enabling its release from the 40S ribosomal subunit. The 60S ribosomal

subunit then joins the 40S to form a complete ribosome that begins protein

synthesis. Because eIF-2 binds very tightly to GDP, eIF-2B, a guanine nucleotide

exchange factor, is required to cause GDP release so that a new GTP molecule

can bind eIF-2 thus enabling its reuse. (Adapted from Kimball, 1999; Robaglia

and Caranta, 2006).

Figure 1.4 Role of eIF2 in translation initiation.

Figure 1.5 Genes encoding eIF4F components in Arabidopsis thaliana.

Translation initiation factor eIF-4E associates with eIF4G (a scaffold for other components of the translation initiation complex) to form the eIF4F complex. The eIF(iso)4F complex that is unique to plants, is composed of eIF(iso)4E and eIF(iso)4G.

Genes encoding proteins of eIF4F complex belong to a small gene family. In Arabidopsis thaliana, three genes named eIF-4E1, eIF-4E2 and eIF-4E3 code for the proteins of the eIF4E subfamily, and one gene codes for eIF(iso)4E. Experimentally proven interactions are shown by arrows.

(Adapted from Robaglia and Caranta, 2006)

Figure 1.5 Genes encoding eIF4F components in Arabidopsis thaliana.

CHAPTER 2

AN IN VIVO COMPLEX CONTAINING SNF1-RELATED KINASE

(SnRK1) AND ADENOSINE KINASE IN ARABIDOPSIS

2.1 Introduction

The evolutionarily conserved SNF1/AMPK/SnRK1 family of protein kinases includes

SNF1 kinase (sucrose non-fermenting 1) in yeast, AMPK (AMP-activated protein kinase) in animals, and SnRK1 (SNF1-related kinase 1) in plants. These serine/threonine kinases play a central role in the regulation of metabolism by responding to cellular energy charge, as sensed by relative AMP and ATP concentrations (Halford and Hardie, 1998;

Hardie et al., 1998; Halford et al., 2003; Hardie, 2007; Polge and Thomas, 2007; Halford and Hey, 2009). Due to the action of adenylate kinase, nutritional and environmental stresses that deplete ATP lead to increased AMP levels, and when the AMP:ATP ratio is elevated the SNF1/AMPK/SnRK1 kinases become active, turning off energy-consuming biosynthetic pathways while turning on alternative ATP-generating systems. For example, in plants, SnRK1 phosphorylates and inactivates key enzymes that control steroid and isoprenoid biosynthesis, nitrogen assimilation for amino acid and nucleotide synthesis, and sucrose synthesis (Sugden et al., 1999a). SnRK1 also plays a role in

39 metabolic signaling, and phosphorylates trehalose-6-phosphate synthase enzymes implicated in signaling processes (Harthill et al., 2006; Jossier et al., 2009). Remarkably, in addition to directly altering the activity of key enzymes, SnRK1 has also been shown to integrate diverse energy and stress signals on a global scale by inducing extensive changes in the transcriptome that have the effect of promoting catabolism and inhibiting anabolic pathways (Baena-Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008).

Given their dependence on the host both for biosynthetic machinery and energy, it is not surprising that viruses have been linked to SnRK1 function. This was first recognized in plants, where SnRK1-mediated responses were identified as a component of innate antiviral defense that is antagonized by the geminivirus pathogenicity proteins AL2 and

L2, which target and inactivate SnRK1 (Hao et al., 2003; Baliji et al., 2007). The geminivirus Rep protein (AL1) has also been shown to interact with the SnRK1 activating kinases GRIK1 and GRIK2 (Shen and Hanley-Bowdoin, 2006). More recently interactions between AMPK and several mammalian viruses have been described, including HIV, SV40, hepatitis C virus, and human cytomegalovirus (Kudchodkar et al.,

2007; Kumar and Rangarajan, 2009; Zhang and Wu, 2009; Mankouri et al., 2010).

SNF1/AMPK/SnRK1 kinases function in vivo as heterotrimeric complexes comprised of an  catalytic subunit, a  subunit, and a  subunit that bridges the  and

subunits. Budding yeast has only a single  subunit (SNF1), but multiple genes encoding and subunits are typical, suggesting multiple combinations of isoforms and potential for complex and subtle regulation (Hardie, 2007; Polge and Thomas, 2007).

The  subunits consist of an N-terminal kinase domain that contains a conserved activation loop, and a C-terminal domain required for interaction with the and 

40 subunits. The subunits contain tandemly repeated cystathionine -synthase (CBS) motifs, which act in pairs to form Bateman motifs that bind ATP or AMP in a mutually exclusive manner, providing a structural basis for energy sensor function (Scott et al.,

2004; Xiao et al., 2007).

Mechanisms that activate SNF1/AMPK/SnRK1 complexes are not completely understood and while differences may exist between kingdoms, the well-studied AMPK provides a general model. First, all SNF1/AMPK/SnRK1 enzymes are activated by upstream kinases that phosphorylate the catalytic  subunit in the highly conserved activation loop at threonine 172 in AMPK and analogous residues in SNF1 and SnRK1

(Hawley et al., 1996; Wilson et al., 1996; Sugden et al., 1999b). Multiple upstream activating kinases have been identified in mammals, yeast, and plants (Sutherland et al.,

2003; Hawley et al., 2005; Woods et al., 2005; Shen et al., 2009). This essential phosphorylation event triggers AMPK autophosphorylation at several sites, although the consequences of this cascade are not clear. Once AMPK is activated, 5'-AMP allosterically stimulates its activity and also inhibits its inactivation by protein phosphatase 2C (PP2C)-catalyzed dephosphorylation of threonine 172 (Davies et al.,

1995; Suter et al., 2006; Sanders et al., 2007). These effects are mediated by AMP binding to the  subunit and are antagonized by ATP (Scott et al., 2004; Sanders et al.,

2007). Direct allosteric stimulation of the SnRK1 complex by AMP has not been demonstrated, although AMP has been shown to suppress dephosphorylation of plant

SnRK1 by PP2C (Sugden et al., 1999b). Thus, while some differences may exist between these kinase complexes, it is abundantly clear that 5'-AMP promotes

SNF1/AMPK/SnRK1 activity while ATP has the opposite effect.

With a total of nearly 40 members in Arabidopsis, the SnRK family has undergone considerable expansion and diversification into three subfamilies: SnRK1,

SnRK2, and SnRK3 (Halford et al., 2003; Halford and Hey, 2009). Most of the growth has occurred in the SnRK2 and SnRK3 subfamilies, which are unique to plants. These kinases are considerably diverged from SNF1 and AMPK and, unlike SnRK1, those tested are unable to complement yeast snf1 mutants. In Arabidopsis, there are two expressed SnRK1 genes which encode SnRK1.1 (also called Arabidopsis kinase 10;

AKIN10) and SnRK1.2 (AKIN11). In addition to a similar length (512 amino acids),

SnRK1.1 and SnRK1.2 share 89% sequence identity in the N-terminal kinase domain and

64% in the C-terminal domain. They also share ~50% overall identity with SNF1 and

AMPK, and this increases to 60-65% in the kinase catalytic domain. The work described in this study was carried out with SnRK1.2 and its isolated kinase domain, which for simplicity are referred to as SnRK1 and SnRK1-KD, respectively.

Adenosine kinase (ADK) is an abundant purine kinase present in all eukaryotes.

ADK catalyzes the transfer of -phosphate from ATP to adenosine to generate 5'-AMP.

Arabidopsis has two ADK proteins, ADK1 and ADK2, which share 92% amino acid identity (Moffatt et al., 2000). The studies described here were performed with ADK2, which for convenience is referred to as ADK. ADK is important for adenine and adenosine salvage and the synthesis of nucleic acids and nucleotide cofactors. It is also involved in the gravitropic response and cytokinin regulation in plants, and in plants and yeast has been shown to maintain methyltransferase activities by sustaining the methyl cycle that generates S-adenosyl methionine (von Schwartzenberg et al., 1998; Lecoq et al., 2001; Moffatt et al., 2002; Young et al., 2006). ADK's role in supporting DNA and

42 histone methylation, and its ability to phosphorylate and inactivate the adenosine analogue cytokinin, are components of the host antiviral response to DNA viruses, which geminivirus AL2 and L2 proteins counter by interacting with and inhibiting ADK (Wang et al., 2003; Wang et al., 2005; Baliji et al., 2007; Raja et al., 2008; Buchmann et al.,

2009; Baliji et al., 2010; Raja et al., 2010).

Because 5'-AMP generated by ADK promotes SnRK1 activity, and the geminivirus pathogenicity proteins AL2 and L2 inhibit SnRK1 and ADK, we speculated that inhibition of both kinases might serve as a dual strategy to suppress SnRK1-mediated antiviral responses (Hao et al., 2003; Wang et al., 2003). This led us to explore the relationship between these kinases. Here we present several lines of evidence which demonstrate that SnRK1 and ADK form an in vivo complex, and that their activities are mutually stimulatory. In addition, we describe a novel mechanism for ADK stimulation by SnRK1 that is independent of SnRK1 kinase activity.

2.2 Results

2.2.1 Development of a gel-based SnRK1 assay

Most SNF1/AMPK/SnRK1 assays measure the phosphorylation of small peptides that contain conserved consensus sites specifically recognized by these, and not other, kinases. To more simply detect and measure SnRK1 kinase activity, we developed a gel- based assay by fusing the SAMS peptide (Davies et al., 1989), which contains such a

43 consensus site, with glutathione-S-transferase to create GST-SAMS (phosphorylation site underlined). A similar fusion protein with the target serine substituted by alanine (GST-

SAMA) was used as a negative control (Figure 2.1 A). Both GST-SAMS and GST-

SAMA were expressed in E. coli and partially purified by glutathione-agarose chromatography. Because it proved difficult to express full-length SnRK1, the SnRK1 kinase domain (SnRK1-KD; amino acids 1-343) was used for in vitro experiments.

Arabidopsis SnRK1-KD, and a kinase-inactive form containing an arginine for lysine substitution in the ATP binding site (SnRK1-KD-K49R) (Hao et al., 2003), were constructed as N-terminal, double hemagglutinin peptide-six histidine (HA2His6) fusion proteins. Proteins were expressed in Nicotiana benthamiana leaf cells using TRBO, a

Tobacco mosaic virus-based vector, following delivery of vector constructs by agroinfiltration (Lindbo, 2007a, b). Expressed proteins were partially purified from leaf extracts by nickel-NTA chromatography, and autophosphorylation assays with 32P-ATP showed that recombinant SnRK1-KD isolated from N. benthamiana was active (Figure

2.1 B). As expected, SnRK1-KD-K49R lacked detectable activity.

Phosphorylation of GST-SAMS and GST-SAMA substrates was tested with varying amounts of SnRK1-KD and SnRK1-KD-K49R. In the presence of 32P-ATP,

GST-SAMS was phosphorylated by SnRK1-KD and levels correlated with the amount of kinase added, showing that this gel-based assay is sensitive and quantitative (Figure 2.1 C and D). By contrast, GST-SAMA was not phosphorylated by SnRK1-KD, confirming that neither GST or the SAMA peptide is a SnRK1 substrate. SnRK1-KD-K49R failed to phosphorylate either GST-SAMS or GST-SAMA, validating its use as a kinase negative control (Figure 2.1 C). Highlighting the sensitivity of GST-SAMS, it was also possible

44 to detect SnRK1 activity in crude extracts from N. benthamiana leaves agroinfiltrated with non-replicating plasmid vectors expressing full-length SnRK1 or SnRK1-KD from the Cauliflower mosaic virus 35S promoter, although considerably more activity was observed with the latter (Figure 2.1 E). The activity observed in control extracts obtained from plants expressing -glucuronidase (GUS) is likely due to endogenous N. benthamiana SnRK1.

2.2.2 SnRK1 and ADK interact in yeast and plant cells

To assess whether SnRK1 and ADK can physically associate, we first tested their ability to interact in the yeast two-hybrid system using strain Y190, which contains His3 and

LacZ reporter genes (Durfee et al., 1993). As summarized in Table 2.1, robust growth in synthetic complete medium lacking histidine, indicating interaction, was observed when cells co-expressed full-length Arabidopsis SnRK1 and ADK, or SnRK1-KD and ADK.

In all cases cells were also positive for LacZ activity, confirming the interactions. The

ADK:SnRK1/SnRK1-KD interactions were detected regardless of whether the proteins were expressed as bait or prey, and no interactions were observed with the negative control proteins CAT and p53. These results show that ADK and SnRK1 specifically interact, and that the N-terminal SnRK1 kinase domain (amino acids 1-343) is sufficient for interaction.

Bimolecular fluorescence complementation (BiFC) was employed to examine whether SnRK1 and ADK interact in plant cells, and where in the cell complexes accumulate (Hu et al., 2002). Constructs expressing full-length Arabidopsis SnRK1 and

ADK, or the control proteins DCL4 and DRB4, fused to the N- or C-terminal portions of yellow fluorescent protein (YN and YC, respectively) were introduced into N. benthamiana leaf cells by agroinfiltration. Cells expressing oppositely tagged proteins

(i.e., YN + YC fusion proteins potentially capable of reconstituting active yellow fluorescent protein; YFP) were viewed under a confocal microscope 48 h later (Yang et al., 2007). DCL4 and DRB4 are known interaction partners (Nakazawa et al., 2007).

Histone 2B fused to red fluorescent protein (RFP-H2B) served as a marker for the nucleus.

No signal was detected in negative control experiments in which only one of the fusion proteins was expressed (data not shown). Additionally, there was no signal when

ADK or SnRK1 were co-expressed with oppositely tagged DCL4 or DRB4, although

YFP fluorescence indicating DCL4:DRB4 interaction was detected in both the cytoplasm and the nucleus when DCL4 and DRB4 were co-expressed (Figure 2.2). ADK is a predominantly cytoplasmic protein, whereas SnRK1 is found in the cytoplasm and the nucleus. When oppositely tagged SnRK1 and ADK fusion proteins were tested, strong

YFP fluorescence was observed in the cytoplasm (Figure 2.2), indicating that these kinases form extranuclear complexes in vivo.

2.2.3 SnRK1 co-immunoprecipitates with native ADK, and co-purifies with overexpressed ADK

To further verify that SnRK1 and ADK form an in vivo complex, immunoprecipitation experiments were performed. These studies employed a polyclonal antibody raised

46 against Arabidopsis ADK2 (-ADK) (Wang et al., 2003), and crude extracts from one month-old, wild-type Arabidopsis (Col-0) plants. An antibody against the phosphorylated threonine 172 residue in the AMPK activation loop (-pT172, Santa

Cruz Biotechnology), corresponding to threonine 176 in SnRK1, was used to detect the plant protein. The pT172 antibody was previously demonstrated to cross-react with activated SnRK1 (Jossier et al., 2009; Shen et al., 2009). Using -pT172 as probe, a signal corresponding to full-length SnRK1, which migrates with an apparent molecular weight of ~75 kDa, was detected in Western blots of immunoprecipitates obtained with

-ADK (Figure 2.3 A). As expected, the ADK antibody was also capable of precipitating endogenous ADK protein, as confirmed by Western blot and by activity assays that measure the production of labeled AMP from adenosine and 32P-ATP

(Figure 2.3 A and 2.3 B) (Wang et al., 2003). By contrast, a negative control antibody raised against Arabidopsis Trans-Membrane Domain Protein-1 (-TMD1) failed to bring down either SnRK1 or ADK, indicating that both proteins were specifically precipitated by the ADK antibody (Figure 2.3 A and B). To confirm the presence of SnRK1 in -

ADK precipitates, aliquots of immune complexes were tested for their ability to phosphorylate GST-SAMS and GST-SAMA, as described above. As judged by specific phosphorylation of the GST-SAMS substrate, SnRK1 activity was present in immunoprecipitates obtained with -ADK, but not with -TMD1 (Figure 2.3 C).

However, because of the long exposure time required (2days vs. 2 h in Figure 2.1) some background phosphorylation of GST-SAMA was also observed with immunoprecipitates obtained with -ADK.

As an additional test of their physical association in vivo, we asked whether

SnRK1 and ADK could be co-purified. HA2His6-tagged Arabidopsis ADK was overexpressed in N. benthamiana leaf cells using the TRBO vector as described above.

Following nickel-NTA chromatography, HA2His6-ADK preparations contained only one major protein species, as detected by Coomassie Blue stain (Figure 2.4 A). The ADK preparations were then tested for SnRK1 activity using the GST-SAMS assay. An activity capable of phosphorylating GST-SAMS was present in HA2His6-ADK preparations, but not in preparations of similarly tagged, expressed, and purified green fluorescent protein (GFP) (Figure 2.4B). Again, some background phosphorylation of

GST-SAMA was observed with ADK preparations due to the prolonged exposure required. The most likely explanation for this result is that an endogenous N. benthamiana SnRK1 activity co-purified with the Arabidopsis ADK protein. In reciprocal experiments, we did not detect the co-purification of ADK with over-expressed

SnRK1-KD or SnRK1-KD-K49R. This is most likely due to the much lower expression levels of the SnRK1 proteins (5-10 µg/ml compared to 500-600 µg/ml) (Figure 2.6 A).

Taken together, the yeast two-hybrid and BiFC studies, as well as co- immunoprecipitation and co-purification experiments, provide strong evidence for a cytoplasmic SnRK1:ADK complex that appears to be conserved across diverse plant species.

2.2.4 SnRK1 phosphorylates ADK in vitro

Analysis of the Arabidopsis ADK1 and ADK2 amino acid sequences revealed four potential SnRK1 phosphorylation sites that conform to the minimal recognition motif:

Hyd-(Basic/X)-X3-Ser/Thr-X3-Hyd, where Hyd indicates the hydrophobic residues M, L,

V, F, or I (Halford and Hardie, 1998). Three of these sites are evolutionarily conserved through mono- and dicotyledous plants and the moss Physcomitrella patens, and two appear to be conserved in nearly all eukaryotes, including Saccharomyces cerevisiae

(Table 2.2). Given that deep evolutionary conservation of consensus sites is predictive of bona fide kinase-substrate relationships (Budovskaya et al., 2005), we tested whether

ADK could be phosphorylated by SnRK1.

Kinase assays employed SnRK1-KD, SnRK1-KD-K49R, and ADK expressed in plant cells from the TRBO vector as HA2His6-tagged proteins. Thus, HA antibody (-

HA) could detect all three of the proteins. However, because these proteins are similar in size and migrate to the same position in polyacrylamide gels, when samples included

SnRK1-KD or SnRK1-KD-K49R and ADK, aliquots were removed before the addition of ADK for Western blot analysis with -HA. ADK antibody was then used to uniquely detect ADK. For the same reason, it was also necessary to obscure SnRK1-KD autophosphorylation in kinase assays that also contained ADK. This was accomplished by pre-incubating SnRK1-KD (or SnRK1-KD-K49R) with 0.5 mM unlabeled ATP in kinase buffer for 20 min, after which signal due to SnRK1-KD autophosphorylation was essentially undetectable (compare Figure 2.5 A and B). Following the pre-incubation,

ADK and the GST-SAMS or GST-SAMA control substrates were added along with 32P-

ATP.

In kinase reactions containing the pre-incubated SnRK1-KD and ADK, a labeled species corresponding to ADK was readily observed, whereas this band was absent in samples containing SnRK1-KD-K49R and ADK (Figure 2.5 B). That SnRK1-KD could also phosphorylate GST-SAMS confirmed that it remained active following the pre- incubation (Figure 2.5 C). Based on these results, we concluded that SnRK1 phosphorylates ADK in vitro.

2.2.5 ADK stimulation does not require SnRK1 kinase activity

Because ADK is a SnRK1 substrate, we investigated the impact of phosphorylation on

ADK activity. In these experiments, which employed HA2His6 fusion proteins expressed in N. benthamiana, ADK was pre-incubated for 10 min with 32P-ATP in the presence of a two-fold molar excess of SnRK1-KD, SnRK1-KD-K49R, or negative control proteins.

Following the addition of adenosine, its conversion to 5'-AMP was measured using thin layer chromatography to separate labeled substrate and product (Wang et al., 2003).

As noted previously, no ADK activity was detected in SnRK1-KD or SnRK1-KD-

K49R preparations, and only a relatively small increase in activity (two to three-fold) over reactions containing ADK alone was observed when the enzyme was incubated with the negative control proteins GFP or Adenine Phosphoribosyl Transferase-1 (APT1), suggesting that added protein might have a non-specific, stabilizing effect. By contrast, a substantial increase in ADK activity (more than 6-fold greater than ADK + GFP) was observed in the presence of SnRK1-KD. Surprisingly, a comparable increase was also

50 seen with SnRK1-KD-K49R (Figure 2.6 A). This suggested that SnRK1 stimulates ADK activity whether or not it is able to phosphorylate ADK.

Because all proteins used in these experiments, including the GFP and APT1 negative controls, were similarly expressed and purified from N. benthamiana extracts, it is unlikely that ADK stimulation was due to non-specific, contaminating proteins.

However, to eliminate the possibility that unknown plant proteins that might associate and co-purify with SnRK1-KD and SnRK1-KD-K49R were responsible, the active and inactive forms of the kinase domain were expressed as GST fusions in E. coli and purified. Due to the absence of upstream activating kinases, GST-SnRK1-KD expressed in E. coli displayed extremely low autophosphorylation activity and a similarly weak ability to phosphorylate GST-SAMS, and was unable to phosphorylate ADK to a detectable extent (Figure 2.8). Nevertheless, when incubated with ADK in kinase reactions, both GST-SnRK1-KD and GST-SnRK1-KD-K49R caused increases in ADK activity similar to the plant-expressed proteins (Figure 2.6 B).

Using proteins expressed in N. benthamiana, the stoichiometry of stimulation was determined by measuring ADK activity in the presence of increasing amounts of SnRK1-

KD. We found that ADK activity reached nearly 90% of maximum stimulation at a 1:1 molar ratio of SnRK1:ADK and no significant further increase was observed at molar ratios as high as 4:1 (Figure 2.6 C). A similar outcome was obtained when this experiment was repeated using GST-SnRK1-KD expressed in E. coli (Figure 2.6 C).

We then considered the reciprocal question, namely whether ADK or its product could modulate SnRK1-KD activity in vitro. However, no effect of added AMP (50-500

µM) or ADK (two-fold molar excess ADK + SnRK1-KD + 2 µM adenosine) on SnRK1-

KD activity was observed (data not shown). This is consistent with studies indicating that allosteric stimulation by AMP is mediated by the  subunit of heterotrimeric

SNF1/AMPK/SnRK1 complexes (Scott et al., 2004; Sanders et al., 2007).

Taken together, the results of these studies conclusively demonstrate that SnRK1-

KD and ADK physically interact, and indicate that a functional outcome of this interaction is a significant stimulation of ADK activity. Further, SnRK1 phosphorylation of ADK is not necessary for the increase in activity, which instead appears to be a result of direct, stoichiometric interaction between the two proteins. Thus, while there may be in vivo consequences of ADK phosphorylation by SnRK1, no in vitro effect was evident.

A converse in vitro stimulation of SnRK1-KD activity by ADK or AMP was not observed.

2.2.6 The activities of SnRK1 and ADK are mutually enhanced in vivo

As SnRK1-KD and SnRK1-KD-K49R increased ADK activity in vitro, it was of interest to determine whether SnRK1, which is expected to be in complex with  and

(ATP/AMP-sensing)  subunits, could stimulate ADK activity in vivo. To address this question, N. benthamiana suspension culture cells were treated with two metabolic inhibitors, 2-deoxyglucose and sodium azide (NaN3) (Sunter and Bisaro, 2003). These agents cause rapid depletion of cellular ATP and increase cellular SnRK1 activity

(Harthill et al., 2006).

Cells were grown in MS medium containing sucrose for seven days, after which

2-deoxyglucose and NaN3 were added to a concentration of 100 mM and 3 mM, respectively. After one hour, cell extracts were obtained and ADK activity was measured. SnRK1 activity, as determined with the GST-SAMS assay, was extremely low owing to the sucrose-containing medium in which the cells were grown (data not shown). However, a significant increase in ADK activity (~1.8-fold) was detected

(Figure 2.7 A).

We next analyzed in vivo SnRK1 activity following ectopic expression of ADK in transgenic plants. Because an over-abundance of ADK is cytotoxic (Moffatt et al., 2002),

Arabidopsis lines that express ADK from a dexamethasone (dex)-inducible promoter were constructed (McNellis et al., 1998; Buchmann et al., 2009). Transgenes were confirmed by genomic PCR, and expression was verified by Northern blot analysis or semi-quantitative RT-PCR (data not shown). Two independent lines (ADK-L5 and

ADK-L6) with enhanced ADK activity were chosen for study. Following dex or mock treatment, extracts were obtained and tested for ADK and SnRK1 activity. It was observed that dex-induced expression of the ADK transgene in lines ADK-L5 and ADK-

L6 caused a 2 to 2.5-fold increase in ADK activity, and a 1.5 to 2-fold increase in SnRK1 activity compared to mock-treated plants (Figure 2.7 B).

Consistent with our hypothesis that SnRK1:ADK complexes could exist in vivo for the purpose of mutual stimulation, we concluded from these experiments that increasing the activities of either SnRK1 or ADK led to a corresponding, parallel increase in the activity of the other. However, the mechanisms involved in mutual stimulation of

SnRK1 and ADK are not the same.

2.3 Discussion

That SNF1/AMPK/SnRK1 activities are enhanced by 5'-AMP is well established, and previous work from this laboratory has shown that the geminivirus AL2 and L2 pathogenicity factors interact with and inactivate both SnRK1 and ADK, which generates

5'-AMP. This led us to propose that inhibition of these kinases might be a dual strategy to inhibit antiviral SnRK1-mediated responses, prompting an evaluation of the relationship between SnRK1 and ADK (Hao et al., 2003; Wang et al., 2003). The experiments described here support our proposal, and provide strong evidence that these kinases physically associate in vivo. The results of two-hybrid experiments in yeast, and

BiFC, co-immunoprecipitation, and co-purification studies in Arabidopsis and N. benthamiana, indicate that SnRK1 and ADK interact and form cytoplasmic complexes.

Experiments in vitro with purified proteins further demonstrate that ADK can be phosphorylated by SnRK1-KD, and surprisingly that SnRK1-KD causes a stoichiometric, rather than enzymatic, enhancement of ADK activity. That stimulation occurs even with kinase inactive forms of SnRK1-KD confirms this and provides compelling evidence that interaction between the two kinases is direct. Finally, experiments in cultured plant cells and transgenic plants demonstrate that enhancing the activity of either kinase results in a parallel increase in the activity of the other, suggesting that SnRK1 and ADK activities are mutually stimulatory in vivo.

Due to cellular adenylate kinase activity, stresses that deplete intracellular ATP result in increased levels of AMP, which allosterically stimulates the

SNF1/AMPK/SnRK1 kinases (Hardie et al., 1998; Halford and Hardie, 1998; Hardie,

2007; Polge and Thomas, 2007; Halford and Hey, 2009). Our data clearly suggest that

ADK activity directly provides another source of AMP for this purpose, but do not address why ADK and SnRK1 complexes might be necessary. As a possible explanation we note that the obligatory first step in SNF1/AMPK/SnRK1 activation is phosphorylation of the conserved activation loop threonine by upstream activating kinases, after which a key role of AMP is to prevent dephosphorylation by PP2C (Davies et al., 1995; Suter et al., 2006; Sanders et al, 2007). It is interesting that in mammals, the activities of some upstream activating kinases appear to be constitutive and are not increased by stresses known to cause AMPK activation. In these cases, allosteric inhibition of dephosphorylation by AMP is likely a critical mechanism for stimulating and maintaining AMPK activity. If the same is true in plants, then SnRK1-ADK complex formation would be an effective means of increasing the local AMP concentration to inhibit dephosphorylation in response to stress. In support of this idea,

ADK activity has been shown to increase following infection of N. benthamiana plants with RNA and DNA viruses (Wang et al., 2003), and following abiotic stress in potato

(Katahira and Ashihara, 2006).

Our data demonstrate that SnRK1 stimulates ADK activity by direct physical interaction. That SnRK1 can stimulate ADK activity, and vice versa, raises the question of how SnRK1-ADK complex activities might be negatively regulated. While our in vitro studies did not provide an answer to this question, we did observe robust ADK phosphorylation by SnRK1-KD. Phosphorylation of ADK may be a common function of different SnRK kinase subfamilies, as in vitro and in vivo phosphorylation of ADK by

SnRK2.8 has been previously reported (Shin et al., 2007). The consequences of ADK phosphorylation in vivo are not known, and we did not detect an in vitro effect of ADK

55 phosphorylation by SnRK1-KD. However, it is known or suspected that in several instances, inhibition of target enzyme activity by SnRK1 phosphorylation is indirect, and requires subsequent binding of the phosphoryated site by one of a family of inhibitory 14-

3-3 proteins (Barker et al., 1996; Sugden et al., 1999; Kanamaru et al., 1999; Ikeda et al.,

2000; Huber et al., 2002). Whether in vivo phosphorylation by SnRK1 indirectly inhibits the activity of ADK in this or an analogous manner remains to be determined.

In conclusion, the data presented in this study provide the first evidence for the existence of SnRK1-ADK complexes in vivo, which may cause increases in local AMP concentrations that could lead to rapid and sustained activation of SnRK1. In addition, we also report a novel mechanism for SnRK1-mediated stimulation of ADK involving direct physical contact between the two kinases. It will be interesting to determine how this complex is regulated in vivo, and whether ADK also enters into complexes with

SNF1 and AMPK in other eukaryotes.

2.4 Materials and Methods

2.4.1 Protein expression in N. benthamiana

SnRK1-KD, SnRK1-KD-K49R, and ADK constructs have been previously described

(Hao et al., 2003; Wang et al., 2003). PCR products of Arabidopsis ADK, SnRK1-KD and SnRK1-KD-K49R genes were cloned into PacI and AvrII digested pJL-TRBO, a

Tobacco Mosaic Virus RNA-Based Over-expression vector to generate N-terminal double hemagglutinin peptide-six histidine (HA2His6)-tagged proteins (Lindbo, 2007a, b). The plasmids were transformed into A. tumefaciens C58C1, and cultures of these

56 transformants were used to infiltrate N. benthamiana leaves as described (Wang et al.,

2005). Tissues were collected ~5 days post infiltration and ground in liquid nitrogen followed by the addition of 1.25 volumes of extraction buffer (50 mM HEPES, pH 7.5,

0.1% Triton X100, 10 mM MgCl2, 50 mM NaF, 1 mM EGTA, 1 mM benzamidine,

10uM MG132-proteasomal inhibitor, 1x plant protease inhibitor cocktail (Sigma), 5 mM

β-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride). Extract was filtered through miracloth and centrifuged at 12,000g for 20 min at 4°C. Clarified supernatant (20-25 ml) was then added to 0.5 ml of washed nickel nitrilotriacetic acid agarose beads (Invitrogen) and incubated at 4°C on a rocker for 2 to 3 h. Columns were washed in 6-12 column volumes of wash buffer (50 mM NaH2PO4 (pH 8.0), 300 mM NaCl, 0.1% (v/v) Tween

20, 5 mM β-mercaptoethanol, with or without 20 mM imidazole). Bound HA2His6- tagged proteins were eluted with wash buffer containing 250 mM imidazole at 4°C, and then dialyzed in buffer containing 40 mM Tris-HCl (pH 7.5), and 10% (v/v) glycerol.

Protein concentrations were estimated using the Bradford assay (BioRad) with bovine serum albumin (BSA) as a standard.

2.4.2 Protein expression in E. coli

GST-SAMS and GST-SAMA SnRK1 kinase substrates were prepared using synthetic oligonucleotides containing the SAMS or SAMA peptide sequence (relevant codons separated by periods): 5'–CATGGGTGGT.CAT.ATG.CGT.AGC.GCG.ATG.(AGC or

GCG).GGT. CTG.CAT.CTG.GTG.AAA.CGT.CGT.GGTCAATTGGAGCT (Davies et al., 1989) (Fig. 1A). These synthetic oligonucleotides were annealed to complementary

57 but shorter oligonucleotides to yield duplexes with four-base, single-strand overhangs

(underlined). Duplex oligonucleotides were ligated into NcoI-SacI digested pGEX-KG

(GE Healthcare Life Sciences), a bacterial expression vector containing glutathione-S- transferase (GST) open reading frame, to create C-terminal GST-SAMS and GST-SAMA fusion proteins. The proteins were expressed in E. coli BL21 (DE3) cells and purified by glutathione-agarose chromatography. GST-SnRK1-KD and GST-SnRK1-KD-K49R fusion proteins were expressed, and purified as described previously (Hao et al., 2003).

2.4.3 Kinase assays

SnRK1-KD autophosphorylation, GST-SAMS, and ADK phosphorylation (by SnRK1-

KD) assays were performed essentially as described by Celenza and Carlson (Celenza and Carlson, 1989). Reactions contained 10 to 15 ng of SnRK1-KD (or SnRK1-KD

K49R) with and without 3 µg of substrate proteins (ADK or SAMS peptide). Reactions were initiated by the addition of 0.5 mM unlabeled ATP and incubated at 30 oC for 20 min to activate SnRK1 and also in some cases to obscure SnRK1-KD autophosphorylation. -32P-ATP (3000 Ci/mmol; Perkin Elmer) was then added to a final concentration of 0.05 µM along with substrate protein, and the reaction mixture incubated a further 30 min at 30°C before electrophoresis on SDS–10% (w/v) polyacrylamide gels. In SnRK1 autophosphorylation assays, pre-incubation with unlabeled ATP was excluded. Labeled proteins were visualized and quantitated using a phosphorimager (BioRad). ADK assays were performed with -32P-ATP and employed

58 thin layer chromatography to separate the adenosine substrate from labeled AMP product, as described (Wang et al., 2003).

2.4.4 Yeast two-hybrid analysis

The yeast two-hybrid system was used to identify interactions between ADK and SnRK1 proteins (Durfee et al., 1993; Harper et al., 1993). Bait and prey constructs containing

ADK, SnRK1, and SnRK1-KD have been described (Hao et al., 2003; Wang et al.,

2003). Positive interaction between bait and prey was indicated by the growth of strain

Y190 cells on synthetic complete medium lacking histidine and containing 50 mM 3- aminotriazole. The medium also lacked leucine and tryptophan to ensure maintenance of expression plasmids. Additional confirmation of interactions was obtained by assessing

β-galactosidase activity using a filter lift assay.

2.4.5 BiFC analysis of interactions

The bimolecular fluorescence complementation (BiFC) protocol was based on the method of Hu et al. (Hu et al., 2002). BiFC expression vectors containing the N- or C-terminal portions of enhanced yellow fluorescent protein (YFP) for fusion to the N- or C-terminus of test proteins (pYN and pYC, or p2YN and p2YC, respectively) have been described

(Yang et al., 2007). Plasmids containing the ADK2 and SnRK1.2 genes from

Arabidopsis have also been described (Hao et al., 2003; Wang et al., 2003). ADK2 was amplified by PCR using the forward primer 5'-CCCTTAATTAACATG.GCT.-

TCT.TCT.TCT.AAC.TAC and the reverse primer 5'-GGGACTAGT.GTT.AAA.GTC.-

GGG.TTT.CTC.AGG. The SnRK1.2 gene was PCR-amplified using the forward primer

5'-CCCTTAATTAACATG.GAT.CAT.TCA.TCA.AAT.AG and the reverse primer 5'-

GGGACTAGT. GAT.CAC.ACG.AAG.CTC.TGT.AAG. In the primer sequences provided, periods separate codons of the amplified genes and recognition sequences for the restriction endonucleases PacI and SpeI are underlined. The resulting PCR products were digested with PacI and SpeI and ligated into similarly digested BiFC vectors.

For BiFC experiments, plasmids were transformed into A. tumefaciens C58C1, and cultures were used to infiltrate N. benthamiana leaves (Wang et al., 2005). Cultures containing YN- and YC-based plasmids were mixed 1:1 immediately prior to infiltration.

Histone 2b fused to red fluorescent protein (RFP-H2B) was used as a nuclear marker. A plasmid expressing RFP-H2B (Chakrabarty et al., 2007) was kindly provided by Michael

Goodin (University of Kentucky). Leaf tissue was analyzed by microscopy ~48 h postinfiltration using a Nikon PCM 2000 confocal laser scanning microscope equipped with argon and green helium neon lasers with excitation wavelengths of 488 nm and 544 nm, respectively. To record YFP fluorescence, a band-pass emission filter (EM515/30HQ) with a 450 to 490 nm excitation wavelength and 515 nm emission wavelength was used.

To record RFP fluorescence, a 565 nm long-pass filter (E565LP) was employed. Images were captured using Simple PCI Software and compiled with Adobe Photoshop.

2.4.6 Immunoprecipitation and immunodetection

Immunoprecipitation with anti-ADK antibody and anti-TMD1-antibody was performed using crude extracts from 4 week-old wild-type Arabidopsis (Col-0) plants grown under short day conditions. The methods followed for immunoprecipitations have been described previously (Kim et al., 2007). Immunoprecipitates were subjected to gel blot analysis using antibodies against ADK, human AMPK pT172 antibody (Santa Cruz

Biotechnology) (to detect SnRK1) and TMD1. The immunocomplexes were also used to carry out GST-SAMS assay (to detect SnRK1 activity) and ADK assay.

Figure 2.1 Gel-based SnRK1 assay.

A. Diagrams of GST-SAMA and GST-SAMA substrates are shown. These proteins were expressed and purified from E. coli BL21 (DE3) cells, and consist of GST fused to the SAMS peptide, a SNF1/AMPK/SnRK1 consensus site. The target serine residue is underlined in GST-SAMS and is replaced by alanine in the negative control substrate

GST-SAMA. The line connecting GST and the SAMS/A peptide indicates a flexible glycine-rich linker (GGGGG). Kinase assays employed either HA2-His6-SnRK1-KD or

SnRK1-KD-K49R (kinase inactive mutant, negative control), which were expressed and purified from N. benthamiana cells. Kinases were incubated with substrates in the presence of 32P-ATP, and reactions were terminated by addition of EDTA, as described in Materials and Methods. Aliquots of kinase assays were fractionated on SDS-PAGE gels and exposed to a phosphorimager to detect 32P incorporation. GST-SAMS and GST-

SAMA were monitored by Coomassie Blue staining, while western blots were probed with anti-HA to detect the kinases.

B. Autophosphorylation activity of SnRK1-KD and SnRK1-KD-K49R (negative control).

C. SnRK1-KD phosphorylates GST-SAMS in vitro. GST-SAMS or GST-SAMA (3 µg) were incubated with varying amounts of SnRK1-KD and SnRK1-KD-K49R, as indicated.

D. In the experiment shown in (C), a positive, linear correlation was observed between the intensity of labeled GST-SAMS signal and the amount of SnRK1-KD added

(R2=0.9689).

E. Crude extracts from N. benthamiana cells transiently expressing full-length SnRK1,

SnRK1-KD, or GUS (control for endogenous activity) were assayed for kinase activity following the addition of GST-SAMS.

Figure 2.1 Gel-based SnRK1assay.

Figure 2.2 SnRK1 and ADK interact in the cytoplasm. BiFC analysis of SnRK1 and

ADK proteins in epidermal cells was performed. Constructs expressing full-length

SnRK1 or ADK fused to the N- or C-terminal portion of YFP were delivered by agroinfiltration to N. benthamiana leaves. Shown here are irregularly shaped epidermal cells, in which most of the central portion of the cell is occupied by the vacuole. Different combinations of the constructs were used to test interaction between ADK and SnRK1.

Images were captured at 40x magnification 48 h postinfiltration using a confocal laser scanning microscope. Representative images are shown. Histone 2B fused to RFP (RFP-

H2B) was used as a marker for the nucleus. Note that SnRK1 and ADK form cytoplasm complexes but do not interact with DCL4 (negative control), while DCL4 interacts with

DRB4 in the cytoplasm and the nucleus.

Figure 2.2 SnRK1 and ADK interact in the cytoplasm.

Figure 2.3 SnRK1 co-immunoprecipitates with ADK.

A. Western blot analysis of immune complexes obtained with anti-ADK and anti-TMD1.

SnRK1 was detected using -pT172, a phospho-specific antibody raised against the conserved activation loop of AMPK. The proteins detected with these antibodies were

37- (ADK), 75- (TMD1), and 70 kD (SnRK1), respectively

B. ADK activity in immune complexes. Equal amounts of immunoprecipitates obtained with -ADK or -TMD1 were incubated with 1 µM adenosine and 5 µCi of γ-32P-ATP

(1500 Ci/mmol). Reaction mixtures were resolved on a TLC plate and labeled AMP was detected using a phosphorimager.

C. SnRK1 activity in immune complexes. Equal amounts of immunoprecipitates obtained from -ADK and -TMD1 were incubated with 3 µg of GST-SAMS or GST-

32 SAMA in a kinase reaction with P-labeled γ-ATP. Reactions with purified HA2His6-

SnRK1-KD were included as controls. Following the kinase reactions, samples were fractionated on an SDS-PAGE gel and exposed to a phosphorimager to detect 32P incorporation. GST-SAMS and GST-SAMA were monitored by Coomassie Blue Blue staining.

Figure 2.3 SnRK1 co-immunoprecipitates with ADK.

Figure 2.4 Endogenous SnRK1 activity copurifies with ADK in N. benthamiana.

HA2His6-tagged Arabidopsis ADK and HA2His6-GFP (control) were expressed in N. benthamiana using a TMV-based vector and purified using Nickel-NTA resin as described in Methods.

A. Proteins were fractionated on SDS-polyacrylamide gel, and a Coomassie Blue Blue stained sample of HA2His6-tagged ADK is shown.

B. GST-SAMS and 32P-ATP were added to ADK or GFP preparations (3 µg protein) to detect a co-purifying, endogenous SnRK1 activity. Following kinase reactions, samples were separated on SDS-PAGE and exposed to a phosphorimager. The Coomassie Blue stained panel is a loading control for GST-SAMS and GST-SAMA, while the Western blot (bottom panel, probed with anti-HA) is a loading control for SnRK1-KD and GFP.

Figure 2.4 Endogenous SnRK1 activity copurifies with ADK in N. benthamiana.

Figure 2.5 ADK is phosphorylated by SnRK1 in vitro. Kinase assays were conducted using -32P-ATP and SnRK1-KD or SnRK1-KD-K49R, either alone or with the substrates ADK, GST-SAMS or GST-SAMA, as indicated. With the exception of autophosphorylation assays, SnRK1-KD (and SnRK1-KD-K49R) were preincubated with unlabeled ATP to obscure autophosphorylation, as described in the text. Samples were fractioned by SDS-PAGE and exposed to a phosphorimager. The Coomassie Blue Blue stained panel is a loading control for GST-SAMS and GST-SAMA, while western blots were probed with anti-ADK, or with anti-HA to detect SnRK1-KD or SnRK1-KD-K49R.

A. SnRK1-KD autophosphorylation, with SnRK1-KD-K49R negative control.

B. SnRK1-KD phosphorylates ADK in vitro. ADK protein (3 µg) was incubated with 10 ng SnRK1-KD or SnRK1-KD-K49R. Note that SnRK1-KD autophosphorylation (in the lane lacking ADK) was rendered undetectable by pre-incubation with unlabeled ATP.

C. Control showing phosphorylation of GST-SAMS by SnRK1-KD (10 ng) following pre-incubation with unlabeled ATP.

Figure 2.5 ADK is phosphorylated by SnRK1 in vitro.

Figure 2.6 Both SnRK1-KD and kinase-inactive SnRK1-KD-K49R enhance ADK activity in vitro.

A. All proteins used in these experiments were expressed and purified from N. benthamiana as HA2His6-tagged fusion proteins. ADK (10 ng) was incubated alone, or with 15-20 ng of SnRK1-KD, SnRK1-KD-K49R, GFP, or APT1 (molar ratio of 2:1).

ADK activity was normalized to ADK+GFP.

B. All proteins used in these experiments were expressed and purified from E. coli as

GST fusion proteins. ADK (10 ng) was incubated alone, or with 13-35 ng of GST,

SnRK1-KD, or SnRK1-KD-K49R (molar ratio of 2:1). ADK activity was measured and normalized to ADK + GST. For both A and B, data obtained from 3 independent experiments, with two replicates each, were used to calculate standard error. Asterisks

(**) indicate significant differences in ADK activity at the 99% confidence level, as determined by Student‘s t test.

C. Stoichiometry of ADK stimulation. The graph shows relative ADK activity plotted against increasing SnRK1-KD:ADK ratio. Note that ADK stimulation approaches maximum stimulation at a 1:1 molar ratio with both HA2His6-SnRK1-KD expressed in plant cells (diamonds) and GST- SnRK1-KD expressed in E. coli (squares).

Figure 2.6 Both SnRK1-KD and kinase-inactive SnRK1-KD-K49R enhance ADK activity in vitro.

Figure 2.7 SnRK1 and ADK activities increase in parallel in vivo.

A. ADK activity increased in N. benthamiana suspension culture cells treated with metabolic inhibitors. Cells were grown in MS medium containing sucrose for seven days, after which 2-deoxyglucose and NaN3 were added to a concentration of 100 mM an 3 mM, respectively. After one hour, cell extracts were obtained and ADK activity was measured as described in methods. SnRK1 activity, as determined with the GST-SAMS assay, was extremely low owing to the sucrose-containing medium in which the cells were grown (data not shown).

B. SnRK1 activities increased in transgenic Arabidopsis plants ectopically expressing

ADK (dex-inducible). 4 week-old transgenic Arabidopsis plants were either dex or mock treated, and crude extracts were obtained. Equal amounts of total protein extracts from each sample were used to measure ADK (200 ng) and SnRK1 (20 µg) activities, as described in methods.

In A and B, data from 3 independent experiments with two replications each, were used to calculate standard error and to perform Student‘s t test. Asterisks indicate significant differences in ADK and SnRK1 activities between mock and dex-treated plants at 99%

(**) confidence level, as determined by Student‘s t test.

Figure 2.7 SnRK1 and ADK activities increase in parallel in vivo.

Figure 2.8 GST-SnRK1-KD expressed in E. coli is poorly active and does not phosphorylate ADK in vitro. GST-SnRK1-KD and GST-SnRK1-KDKR expressed in E. coli, and HA2His6-SnRK1-KD, expressed in N. benthamiana, were subjected to kinase reactions in vitro to compare their abilities to autophosphorylate (A), phosphorylate GST-

SAMS (B), and phosphorylate HA2His6-ADK (C).

Following the kinase reactions, the proteins were subjected to SDS-PAGE and exposed to a phosphorimager for 5 h to detect the levels of 32P incorporation. In A and B, the asterisk indicates longer exposure (72 h). After longer exposure, the HA2His6-SnRK1-KD autophosphorylation and GST-SAMS phosphorylation (by HA2His6-SnRK1-KD) signals were saturated (data not shown). In C., to obscure autophosphorylation and for activation, HA2His6-SnRK1-KD GST-SnRK1-KD were pre-incubated with 0.5 mM of cold ATP for 20 min in kinase buffer before incubating with ADK protein and 32P- labeled γ-ATP.

For the reasons that are not clear, major GST-SnRK1-KD and GST-SnRK1-KD-K49R species typically migrate at slightly different rates upon PAGE.

Figure 2.8 GST-SnRK1-KD expressed in E. coli has minimal autophosphorylation activity and does not phosphorylate ADK in vitro.

Table 2.1 Interactions between ADK, SnRK1, and SnRK1-KD in the yeast two-hybrid system. The indicated bait proteins were expressed as GAL4 DNA binding domain fusions, and the prey proteins as GAL4 activation domain fusions in yeast Y190 cells.

SnRK1 indicates the full-length protein (amino acids 1-512) and SnRK1-KD denotes the kinase domain (residues 1-343). CAT (chloramphenicol acetyl transferase) and p53 were used as negative controls.

aInteraction was indicated by the ability of cells transformed with both bait and prey plasmids to grow on medium lacking His and containing 50 mM aminotriazole. As an additional indicator of interaction, colonies were monitored for LacZ activity (blue color) using a filter-lift assay. Interaction symbols are as follows: +, interaction; -, no interaction.

Bait Prey Interactiona

SnRK1 ADK +

ADK SnRK1 +

SnRK1-KD ADK +

ADK SnRK1-KD +

ADK CAT -

ADK p53 -

SnRK1 CAT -

SnRK1 p53 -

Table 2.1 SnRK1 and ADK interact in the yeast-two hybrid assay.

Table 2.2 ADK2 of Arabidopsis thaliana has three evolutionarily conserved potential

SnRK1 sites. Protein sequences of ADK from different taxonomic groups were compared to determine the extent of conservation of SnRK1 sites. The amino acids in red font are critical for SnRK1 recognition, and Ser/Thr residues in red font and underlined are predicted to be phosphorylated by SnRK1. Hyphen (–) indicates not conserved.

% Identity, entire Organism protein T68 T105 Arabidopsis thaliana (ADK2) 100 IAGGATQNSI MKKDATAAGV Arabidopsis thaliana (ADK1) 92 IAGGATQNSI MKKDATAAGV Vitis vinifera 86 IAGGATQNSI - Populus trichocarpa 86 IAGGATQNSI MKKNSTEAGV Ricinus communis 84 IAGGATQNSI - Solanum tuberosum 85 IAGGATQNSI - Nicotiana tabacum 86 IAGGATQNSI - Gossypium hirsutum 85 IAGGATQNSI - Sorghum bicolor 82 IAGGATQNSI - Oryza sativa 82 IAGGATQNSI - Medicago truncatula 83 IAGGATQNSI - Glycine max 81 IAGGATQNSI - Zea mays 82 IAGGATQNSI - Picea sitchensis 76 IAGGATQNSI - Physcomitrella patens 67 IAGGATQNTI MFKLASEGGV Mus musculus 57 - - Danio rerio 59 - - Rattus norvegicus 57 - - Equus caballus 57 - - Macaca mulatta 56 - - Homo sapiens 56 - - Xenopus tropicalis 58 - - Drosophila willistoni 48 - - Caenorhabditis briggsae 51 - - Saccharomyces cerevisiae 40 - - Continued

Table 2.2 ADK2 of Arabidopsis thaliana has three evolutionarily conserved potential SnRK1 sites.

Table 2.2 Continued

% Identity, entire Organism protein S175 S196 Arabidopsis thaliana (ADK2) 100 FFLTVSPESI FTMNLSAPFI Arabidopsis thaliana (ADK1) 92 FFLTVSPESI FTMNLSAPFI Vitis vinifera 86 FFLTVSPESI FMMNLSAPFI Populus trichocarpa 86 FFLTVSPESI FTMNLSAPFI Ricinus communis 84 FFLTVSPESI FTMNLSAPFI Solanum tuberosum 85 FFLTVSPESI FSMNLSAPFI Nicotiana tabacum 86 FFLTVSPESI FSMNLSAPFI Gossypium hirsutum 85 FFLTVSPESI FSMNLSAPFI Sorghum bicolor 82 FFLTVSPDSI FMMNLSAPFI Oryza sativa 82 FFLTVSPDSI FLMNLSAPFI Medicago truncatula 83 FFLTVSPESI FMMNLSAPFI Glycine max 81 FFLTVSPDSI FSMNLSAPFI Zea mays 82 FFLTVSPDSI FLMNLSAPFI Picea sitchensis 76 FFLTVSPESI - Physcomitrella patens 67 FFLTVSPESM FMMNLSASFV Mus musculus 57 FFLTVSPESV FTLNLSAPFI Danio rerio 59 FFLTVSLESI FCLNLSAPFI Rattus norvegicus 57 FTMNLSAPFV FTLNLSAPFI Equus caballus 57 FFLTVSPESV FTLNLSAPFI Macaca mulatta 56 FFLTVSPESV FTLNLSAPFI Homo sapiens 56 FFLTVSPESV FTLNLSAPFI Xenopus tropicalis 58 FFLTVSPESI FCMNLSAPFI Drosophila willistoni 48 - FLMNLSAPFI Caenorhabditis briggsae 51 - FTLNLSAPFI Saccharomyces cerevisiae 40 FHLTVSPDAI FVLNFSAPFI

CHAPTER 3

ADENINE PHOSPHORIBOSYL TRANSFERASE 1 (APT1) INTERACTS WITH SnRK1 AND ADK

3.1 Introduction

1998; Polge and Thomas, 2007). These Ser/Thr kinases play a central role in the regulation of metabolism (Halford and Hardie, 1998; Hardie et al., 1998; Polge &

Thomas, 2007). Specifically, in response to nutritional and environmental stresses that deplete ATP and increase AMP (elevated AMP/ATP ratio), these kinases turn off energy- consuming biosynthetic pathways and turn on alternative ATP-generating systems. This is called cellular stress response (CSR). In plants, for example, SnRK1 kinases phosphorylate and inactivate key enzymes that control steroid and isoprenoid synthesis, nitrogen assimilation for amino acid and nucleotide synthesis, and sucrose biosynthesis

(Polge and Thomas, 2007; Sugden et al., 1999b). Remarkably, in addition to directly altering the activity of key enzymes, SnRK1 has also been shown to integrate diverse

81 energy and stress signals on a global scale by inducing extensive changes in the transcriptome that have the effect of promoting catabolism and inhibiting anabolic pathways (Baena-Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008).The CSR, mediated by SnRK1, also plays a key role in antiviral defense (Hao et al., 2003).

SNF1-AMPK family members function as a heterotrimeric complex consisting of an  catalytic subunit, a  subunit, and a  subunit that appears to bridge the  and  subunits. They are activated through phosphorylation of the  subunit at a conserved Thr residue residing in the activation loop. In mammals, 5'-AMP allosterically activates

AMPK and inhibits its deactivation due to dephosphorylation of the  subunit at Thr-172 of activation loop by protein phosphatase 2C (PP2C) (Davies et al., 1995; Suter et al.,

2006; Sanders et al., 2007). Recent evidence indicates that these allosteric effects are mediated by AMP binding to the  subunit (Sanders et al., 2007). In plants, direct allosteric stimulation of the SnRK1 complex by AMP has yet to be demonstrated.

However, AMP has been shown to suppress dephosphorylation by PP2C (Sugden et al.,

1999 a). Despite the apparent differences in mechanisms of activation, it is clear that 5'-

AMP stimulates kinase activity and that high concentrations of ATP have an antagonistic effect. For any of these kinases, the exact nature of the signals that lead to kinase activation is not yet known. Therefore, we attempted to study the relationship of SnRK1 with two 5'-AMP-generating enzymes, ADK and APT1.

Adenosine kinase (ADK; EC 2.7.1.20; adenosine 5'-phophotransferase) is an abundant purine kinase present in all eukaryotic cells examined to date. In the presence of magnesium, this enzyme catalyzes the transfer of -phosphate from ATP or GTP to adenosine to produce 5'-AMP. ADK plays critical roles in the adenine and adenosine

82 salvage pathways and is important for the synthesis of nucleic acids and nucleotide cofactors. It also plays major roles in sustaining the methyl cycle, the gravitropic response, and cytokinin (CK) regulation in plants (von Schwartzenberg et al., 1998;

Moffatt et al., 2000; Moffatt et al., 2002; Young et al., 2006). ADK‘s roles in sustaining the methyl cycle and in cytokinin regulation are components of host defense against geminiviruses (Wang et al., 2003; Wang et al., 2005; Raja et al., 2008; Baliji et al., 2010).

Arabidopsis encodes two ADK proteins, ADK1 and ADK2, which share 92% amino acid identity (Moffatt et al., 2000). The studies described here employed ADK2.

Adenine phosphoribosyltransferase (APT: pyrophosphate phospho-D- ribosyltransferase - EC 2.4.2.7) is a constitutively expressed enzyme that generates 5'-

AMP by a one-step adenine salvage. This reaction involves Mg2+-dependent transfer of the phosphoribosyl group of 5-phosphorylribosyl 1-pyrophosphate (PRPP) to adenine

(Allen et al., 2002). APT, like ADK, also contributes to 5'-AMP pools and we speculated that it might also play a role in activating the cellular stress response (CSR), which is triggered by an increased AMP/ATP ratio. Arabidopsis encodes five APT or APT-like genes, but only APT1, APT2, and APT3 are expressed (Allen et al., 2002). APT1 was selected for this study because it converts adenine to AMP 30- to 50-fold more rapidly than APT2 or APT3 and because APT1 deficiency generates unique phenotypes that are not complemented by APT2 or APT3 (Gaillard et al., 1998).

We hypothesized that due to their shared ability to generate 5'-AMP, both ADK and APT1 might play a role in SnRK1 activation and thus in CSR activation. In fact, there is evidence for increased expression (25-50 fold) as well as enzyme activities (2- fold) of APT1 and ADK during stress (wounding) in potato (Katahira and Ashihara,

2006). In Chapter 2, I discussed the SnRK1-ADK complex. Here, I focus on the relationship between SnRK1, ADK, and APT1 as a possible ternary complex.

3.2 Results

3.2.1 APT1 interacts with SnRK1 in the cytoplasm

We obtained convincing data in support of our hypothesis that SnRK1 and ADK form a complex (Chapter 2). Because both ADK and APT generate 5'-AMP, we sought to test a similar hypothesis with respect to SnRK1 and APT1. To examine this possibility, we first employed bimolecular fluorescence complementation (BiFC) to determine whether APT1 and SnRK1 interact. In this method, potentially interacting proteins are fused to the N- or

C-terminal portions of yellow fluorescent protein (YFP) and are introduced into living cells. Association of interacting partners reconstitutes YFP, resulting in fluorescence which indicates interaction and reveals where the interacting proteins accumulate in the cell. We adopted a transient system in which Nicotiana benthamiana leaves are coinfiltrated with Agrobacterium tumefaciens cultures to deliver disarmed Ti plasmids designed to express the fusion proteins to be tested. Cells were viewed under confocal microscope 48 h postinfiltration. The test proteins consisted of the N-terminal fragment of the YFP open reading frame (YN, encoding amino acids 1 to 158) or the C-terminal YFP fragment (YC, encoding amino acids 159 to 238) fused to either the N- or C-terminus of

APT1 or SnRK1 proteins. Cells expressing oppositely tagged proteins (i.e., YN + YC

84 fusion proteins potentially capable of reconstituting active yellow fluorescent protein;

YFP) were viewed under a confocal microscope 48 h later.

APT1 is a cytoplasmic enzyme while SnRK1 is found in the cytoplasm and the nucleus. We found that APT1 and SnRK1 interact almost entirely in the cytoplasm

(Figure 3.1). We also found that APT1 self-interacts predominantly in the cytoplasm

(Figure 3.1). However, it is not possible from these experiments to determine whether

SnRK1 interacts with APT1 monomers or homodimers. SnRK1 does not interact with dicer-like 4 (DCL4) (a negative control) while DCL4 interacts with double-stranded

RNA binding protein 4 (DRB4) (Figure 3.1). However, for ATP1 interaction, a negative control (a cytoplasmic protein that do not interact with APT1 in a BiFC assay) remains to be determined.

3.2.2 SnRK1-KD phosphorylates APT1 in vitro

The protein sequence of Arabidopsis APT1 contains two evolutionarily conserved potential SnRK1 sites (S16 and T35) (Table 3.1). The minimum recognition motif reported for SnRK1 and related kinases is: Hyd-(Basic or X)-X-X-Ser/Thr-X-X-X-Hyd

(Halford and Hardie, 1998). Arabidopsis has five APT or APT like genes of which only three (APT1, APT2 and APT3) are expressed (Allen et al., 2002). The two potential

SnRK1 sites present in APT1 (S16 and T35) are also found in ATP3 but APT2 has only one of them (T35). Both S16 and T35 are conserved in both monocotyledonous and dicotyledonous plants where as T35 also appears to be conserved in algae and mosses

(Table 3.1).

We sought to test if APT1 is a substrate of SnRK1. To perform an in vitro kinase assay, wildtype APT1 and APT1-S16A mutant proteins (Ser16 is predicted to be phosphorylated by SnRK1) were chosen as substrates. Because it proved difficult to express sufficient amounts of full-length SnRK1, the SnRK1-KD (amino acids 1-343) was used for in vitro experiments. All proteins were expressed as double hemagglutinin peptide, six histidine fusions (HA2His6 tagged) from a TMV-based TRBO vector

(Lindbo, 2007a) in N. benthamiana and purified using Nickel-NTA chromatography.

SnRK1 kinase is known to autophosphorylate (Figure 3.2 A, Hao et al., 2003; Shen et al.,

2009). To obscure autophosphorylation and for better activation, SnRK1-KD was first incubated with 0.5 mM of unlabeled ATP in kinase buffer for 20 min before adding substrate proteins and 32P-labeled -ATP.

We found that partially purified SnRK1-KD phosphorylates both APT1 and

APT1-S16A very strongly compared to some background phosphorylation by SnRK1-

KD-K49R (a kinase-inactive mutant) (Figure 3.2 B). This suggests that APT1 is likely a

SnRK1 substrate and that Ser16 is not phosphorylated or is not the only amino acid phosphorylated by SnRK1. These data establish that SnRK1-KD and APT1 directly interact in vitro. However, the consequences of APT1 phosphorylation by SnRK1 have yet to be investigated. Interestingly, we found that overexpressed APT1 in N. benthamiana migrated as two bands, and that the minor species was sensitive to phosphatase (Figure 3.2 C). This indicates that APT1 is phosphorylated in vivo, but additional experiments are needed to determine whether SnRK1 or some other kinase is responsible.

3.2.3 Endogenous SnRK1 copurifies with overexpressed APT1 in Nicotiana benthamiana

As an additional test of interaction, we asked whether an endogenous SnRK1 activity could be co-purified with Arabidopsis APT1 expressed in N. benthamiana. HA2His6-

APT1 was expressed in N. benthamiana leaves using TRBO, and tissue was harvested 5 days post-infiltration. Abundant APT1 was expressed as noted previously (Figure 3.2 C right panel). SnRK1 activity in partially purified APT1 preparations was then measured using GST-SAMS, which consists of a SnRK1 consensus site fused to glutathione-S- transferase (Figure 2.2 A). We observed strong SAMS phosphorylation by APT1 protein preparations compared to some background phosphorylation in extracts of similarly expressed and purified GFP (Figure 3.3). However, we failed to detect endogenous

SnRK1 protein on a western blot probed with -pT172, an antibody raised against the phosphorylated human AMPK-α subunit peptide (Santa Cruz Biotechnology). Although

-pT172 cross reacts with SnRK1 protein, the kinase assay is far more sensitive and could detect SnRK1 protein even at very low abundance. Copurification of an endogenous N. benthamiana SnRK1 activity with Arabidopsis APT1 suggests that APT1-

SnRK1 complex exists in vivo, and is conserved in diverse plant species.

3.2.4 SnRK1, APT1, and ADK interact with each other in the cytoplasm

After observing that SnRK1 interacts with APT1 and ADK in the cytoplasm, we hypothesized that SnRK1, ADK, and APT1 may exist as a ternary complex. Therefore, a

BiFC experiment to test interactions between ADK and APT1 was performed. The test

87 proteins consisted of the N-terminal fragment of the YFP open reading frame (YN, encoding amino acids 1 to 158) or the C-terminal YFP fragment (YC, encoding amino acids 159 to 238) fused to either the N- or C-terminus of APT1 or ADK proteins. Cells expressing oppositely tagged proteins (i.e., YN + YC fusion proteins potentially capable of reconstituting active yellow fluorescent protein; YFP) were viewed under a confocal microscope 48 h later.

We found that ADK and APT1 interact with each other predominantly in the cytoplasm (Figure 3.4). We also observed a protein corresponding to the size of APT1 only in immunocomplexes obtained using anti-ADK antibody but not with anti-TMD1

(Transmembrane Domain 1; Meier I, unpublished data) antibody (Figure 3.5). We speculate that the phosphorylated protein shown by a (*) in figure 3.5 might be APT1 based on the following observations

1. Its migration on a SDS-PAGE gel that is expected for APT1 (Allen et al., 2002)

2. BiFC analysis show interaction between ADK and APT1 (Figure 3.1)

3. SnRK1 phosphorylates APT1, in vitro (Figure 3.2 B)

4. Endogenous APT1 is phosphorylated (Figure 3.2 C)

5. SnRK1 activity is specific to anti-ADK IPs.

We also speculate that endogenous SnRK1 that is co-immunoprecipitated with

ADK is responsible for phosphorylation of this protein. However, to confirm these results, immunoprecipitations with antibodies against APT1 need to be performed. The

88 consequences of ADK:APT1 interaction, and whether the interaction is direct or indirect, remain to be investigated.

3.3 Discussion

We hypothesize that AMP generated through pathways involving both APT and ADK could play a vital role in enhancing the AMP/ATP ratio. In fact, there is evidence for increased expression (25-50 fold) as well as enzyme activities (2 fold) of APT1 and ADK during stress (wounding) in potato (Katahira and Ashihara, 2006). In addition, we previously observed about 1.5-3 fold increase in ADK activity in N. benthamiana infected with several RNA viruses or DNA-containing geminiviruses lacking a protein that inhibits ADK (Wang et al., 2003).

We obtained in vivo and in vitro evidence in support of our hypothesis that

SnRK1 forms a complex with APT1. Based on the consequences we observed of SnRK1-

ADK interaction, we speculate that SnRK1 may also increase APT1 activity. Our BiFC data also suggest that SnRK1, ADK, and APT1 might form a ternary complex in vivo

(Figure 3.4). Further investigations are needed to confirm this notion. One method would be to co-immunoprecipitate the members of the complex using antibodies against one of them. Currently, we lack an APT1 antibody, and are attempting to obtain some. Because

SnRK1 forms a complex with both APT1 and ADK, the two proteins that generate the potential SnRK1 activator AMP using different substrates, we propose that this complex might be acting as ‗sensosome‘ that mediates rapid activation of the CSR (Figure 3.6).

Determining the exact roles of this putative complex in sensing as well as mediating the

89 stress response merits further investigation. The members of this putative complex have been shown or implicated to play roles in multiple stresses-biotic as well as abiotic (Hao et al., 2003; Wang et al., 2003; Radchuk et al., 2006; Katahira and Ashihara, 2006;

Baena-Gonzalez et al., 2007; Baliji et al., 2010).

According to our model, 5'-AMP generated by APT1 and ADK play a role in activating SnRK1, and SnRK1 stimulates their activities forming a positive feedback loop. This raises an important question: how is this complex negatively regulated upon relief from stress? One possibility is that phosphorylation of APT1 and ADK by SnRK1 could lead to negative regulation of the complex. In support of this idea, we found that

SnRK1 phosphorylates APT1 and ADK in vitro, and that the potential SnRK1 sites in these proteins are evolutionarily conserved across diverse taxa (Figure 3.2 B; Tables 2.2 and 3.1). We also found that overexpressed APT1 is endogenously phosphorylated and it remains to be confirmed if SnRK1 or other kinase is responsible (Figure 3.2B). The consequences of APT1 phosphorylation by SnRK1 remain to be determined but data we obtained with ADK showed no apparent effect of phosphorylation by SnRK1 in vitro.

This may not be surprising, given the fact that inactivation of some known SnRK1 targets requires binding of 14-3-3 proteins to the phosphorylated sites of the target protein

(Bachmann et al., 1996a; Moorhead et al., 1996; Sugden et al., 1999b; Ikeda et al., 2000;

Huber et al., 2002). Therefore, it would be interesting to test if ADK and APT1 can interact with any of the plant 14-3-3 proteins upon phosphorylation and also to test consequences of such interactions, if found. Arabidopsis 14-3-3 gene family has 13 members (named general regulatory factor GRF 1 through GRF 13) that are classified into  and non- groups (Sehnke et al., 2002). Further investigations could determine if

90 any of these 14-3-3 proteins are involved in the regulation of SnRK1-ADK-APT1, a putative complex. It is also possible that other mechanisms in addition to or instead of 14-

3-3 protein-mediated negative regulation of the complex might exist. Whatever may be the mechanisms of SnRK1-ADK-APT1 regulation, our study has opened new avenues to further our understanding on stress sensing and stress responses.

Lastly, why is it advantageous for SnRK1 to form a complex with ADK and

APT1? Recent evidence with AMPK suggests that its upstream activating kinase LKB1 is constitutively active, and its activity is not changed by stimuli that activate AMPK (Suter et al., 2006; Sanders et al., 2007). Thus, the phosphorylation state of AMPK at the conserved Thr-172 depends on the relative rates of phosphorylation catalyzed by its upstream kinases (LKB1 and CaMKKb), and dephosphorylation catalyzed by protein phosphatases. So, according to the current model, increased ratio of phosphorylated to dephosphorylated AMPK would occur in response to decreased dephosphorylation, following a rise in AMP levels. In parallel, AMP would allosterically activate AMPK, and both these effects are mediated by  subunit of AMPK holoenzyme (Sanders et al.,

2007). In plants, AMP has been shown to prevent dephosphorylation of SnRK1 by PP2C but evidence is lacking in support of direct allosteric activation by AMP. In this context, generating 5'-AMP around the SnRK1 complex might rapidly enhance SnRK1 phosphorylation by increasing the ratio of activated SnRK1: inactive SnRK1 through prevention of SnRK1 dephosphorylation. It might be useful to have two proteins in this complex that make 5'-AMP because they use different substrates (ADK uses adenosine and ATP while APT1 uses adenine and PRPP to make 5'-AMP) and therefore AMP could be generated far more rapidly compared to either one of them alone. Also, such a

91 complex could ensure enough AMP production even during the deficiency of either ADK or APT1, or their substrates. Quite likely, ATP is limiting (one of the ADK substrates) under the stress conditions and APT1, which do not require ATP, could compensate for

ADK. Overall, it might be crucial for plants to have multi-pronged strategies to ensure

SnRK1 activation considering its roles in various kinds of biotic and abiotic responses.

3.4 Materials and Methods

3.4.1 Plant material, agroinfiltration and bimolecular fluorescence complementation

(BiFC)

The bimolecular fluorescence complementation (BiFC) protocol used was based on the method of Hu et al. (2002). The construction of BiFC expression vectors using enhanced yellow fluorescent protein (YFP) has been described previously (Yang et al., 2007). We used the same approach to construct BiFC constructs for Arabidopsis SnRK1, APT1 and,

ADK. APT1 cDNA was obtained from ABRC (stock# U15127) and the sources of

SnRK1 and ADK cDNA have been reported earlier (Wang et al., 2003; Hao et al., 2003).

The genes were amplified using PCR to incorporate either 5' PacI-3' SpeI or 5' NotI-3'

XbaI restriction sites. The resulting PCR products were either digested with PacI and

SpeI or NotI and XbaI restriction enzymes and ligated into PacI-SpeI-digested p2YN and p2YC vectors (to make C-terminal fusions to either the N- or C-terminal portion of YFP)

92 or NotI-XbaI-digested pYC1 and pYN1 vectors (to make N-terminal fusions to either the

N- or C-terminal portion of YFP).

Leaf tissue was analyzed by microscopy ~48 h post-infiltration using a Nikon PCM 2000 confocal laser scanning microscope equipped with argon and green helium neon lasers with excitation wavelengths of 488 nm and 544 nm, respectively. To record YFP fluorescence, a band-pass emission filter (EM515/30HQ) with a 450- to 490 nm excitation wavelength and 515 nm emission wavelength was used. Images were captured using Simple PCI Software and compiled with Adobe Photoshop.

3.4.2 Protein expression and purification

Expression and purification of proteins in N. benthamiana

The mutant APT1-S16A, was generated through site directed mutagenesis using a previously described protocol (Zheng et al., 2004). Plasmids containing SnRK1-KD,

SnRK1-KD-K49R, APT1 and ADK have been described above. PCR products of

Arabidopsis APT1, APT1-S16A, ADK, SnRK1-KD and SnRK1-KD-K49R genes were cloned into PacI and AvrII digested pJL-TRBO, a Tobacco Mosaic Virus RNA-Based

Over-expression vector to generate N-terminal double hemagglutinin peptide-six histidine (HA2His6) fusion proteins (Lindbo, 2007 a; Lindbo, 2007 b). The plasmids

93 were transformed into A. tumefaciens C58C1, and cultures were used to infiltrate N. benthamiana leaves as described (Wang et al, 2005). Tissues were collected ~5 days post infiltration and ground in liquid nitrogen followed by the addition of 1.25 volumes of

extraction buffer [50 mM HEPES (pH 7.5), 0.1% (v/v) Triton X-100, 10 mM MgCl2, 50 mM NaF, 1 mM EGTA, 1 mM benzamidine, 10 µM MG132-proteasomal inhibitor, 1x plant protease inhibitor cocktail (Sigma), 5 mM β-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride]. Extract was filtered through miracloth and centrifuged at

12,000g for 20 min at 4°C. Clarified supernatant (20-25 ml) was then added to 0.5 ml of washed nickel nitrilotriacetic acid agarose beads (Invitrogen) and incubated at 4°C on a rocker for 2 to 3 h. Columns were washed in 6 to 12 column volumes of wash buffer [50 mM NaH2PO4, pH 8.0, 300 mM NaCl, 0.1% (v/v) Tween 20, 5 mM β-mercaptoethanol, with or without 20 mM imidazole]. Bound HA2His6-tagged proteins were eluted with wash buffer containing 250 mM imidazole at 4°C, and then dialyzed in buffer containing

40 mM Tris-HCl, pH 7.5, and 10% glycerol. Protein concentrations were estimated using the Bradford assay (BioRad) with bovine serum albumin (BSA) as a standard. Purified

APT1 was subjected to phosphatase treatment using lambda phosphatase (New England

Buffers).

Expression and purification of proteins in E. coli

GCG).GGT. CTG.CAT.CTG.GTG.AAA.CGT.CGT.GGTCAATTGGAGCT (Davies et al., 1989) (Fig 2.2 A). These were annealed to complementary but shorter oligonucleotides to yield duplexes with four base, single-strand overhangs (underlined).

Duplex oligonucleotides were ligated into NcoI-SacI digested pGEX-KG (GE Healthcare

Life Sciences), a bacterial expression vector containing glutathione-S-transferase (GST) open reading frame, to create C-terminal GST-SAMS and GST-SAMA fusion proteins.

The proteins were expressed in E. coli BL21 (DE3) cells and purified by glutathione- agarose chromatography. GST-SnRK1-KD and GST-SnRK1-KD-K49R fusion proteins were similarly prepared, expressed, and purified (Hao et al., 2003).

3.4.3 SnRK1 kinase assay

SnRK1-KD autophosphorylation, GST-SAMS, and APT1 (both WT and S16A mutant) phosphorylation (by SnRK1-KD) assays were performed essentially as described previously (Celenza and Carlson, 1989). Reactions contained 10 to 15 ng of SnRK1-KD

(or SnRK1-KD K49R) with or without 3 µg of substrate proteins (APT1, APT1-S16A or

SAMS peptide). Reactions were initiated by the addition of 0.5 mM unlabeled ATP and incubated at 30° C for 20 min to activate SnRK1 and also in some cases to obscure

SnRK1-KD autophosphorylation. -32P-ATP (3000 Ci/mmol; Perkin Elmer) was then added to a final concentration of 0.05 µM along with substrate protein, and the reaction mixture incubated a further 30 min at 30°C before electrophoresis on SDS–10% polyacrylamide gels. In SnRK1 autophosphorylation assays, pre-incubation with

95 unlabeled ATP was excluded. Labeled proteins were visualized and quantitated using a phosphorimager (BioRad).

3.4.4 Protein sequence comparison using BLAST analysis

Protein sequences of APT1 were retrieved from GenBank (NCBI) and subjected to

‗protein BLAST‘ to recover the related protein sequences. Using these protein sequences searches were made manually to identify SnRK1 consensus sites based on the features described by Halford and Hardie (1998), and also, to determine the extent of conservation of the SnRK1 sites.

Figure 3.1 APT1 interacts with SnRK1 in the cytoplasm. BiFC analysis of APT1 and

SnRK1 proteins in N. benthamiana epidermal cells was performed. Constructs expressing APT1 and SnRK1 fused to the N- or C-terminal portion of YFP were delivered by agroinfiltration to N. benthamiana leaves. Different combinations of the constructs were used to test interaction between APT1 and SnRK1, and APT1 self- interaction. Images were captured at 20x magnification 48 h post-infiltration using a confocal laser scanning microscope. Representative images are shown. Note that

APT1:SnRK1 and APT1:APT1 complexes accumulate mostly in the cytoplasm. SnRK1 does not interact with DCL4, but DCL4 can interact with DRB4 in the nucleus and the cytoplasm (Figure 2.1).

Figure 3.1 APT1 interacts with SnRK1 in the cytoplasm

Figure 3.2 SnRK1 phosphorylates APT1 in vitro.

A. SnRK1-KD autophosphorylation. 10 ng of partially purified HA2His6-tagged SnRK1-

KD or SnRK1-KD-K49R (kinase-inactive mutant) protein was incubated with 32P labeled

γ-ATP. Following the kinase reactions, the samples were separated on a SDS-PAGE gel and exposed to a phosphorimager to detect the levels of 32P incorporation. SnRK1-KD autophosphorylates but SnRK1-KD-K49R cannot.

B. SnRK1-KD phosphorylates APT1. Partially purified APT1-wt (wildtype) and APT1-

S16A mutant proteins (2.5 µg) were incubated with 10 ng of partially purified HA2His6 tagged SnRK1-KD or SnRK1-KD-K49R (kinase-inactive mutant) protein and 32P labeled

γ-ATP. Following the kinase reactions, the samples were separated on a SDS-PAGE gel and exposed to a phosphorimager to detect the levels of 32P incorporation. Note that the

SnRK1 autophosphorylation signal is not seen because the enzymes were preincubated with 0.5 mM of cold ATP for 20 min in a kinase buffer before incubating with the substrates and 32P-labeled γ-ATP.

C. APT1 is phosphorylated in vivo. HA2His6-tagged APT1 protein, expressed and purified from N. benthamiana, is endogenously phosphorylated. APT1 was treated with either lambda protein phosphatase or buffer and subjected to western blot analysis with anti-HA antibody, following SDS PAGE (left panel). The right panel shows a Coomassie

Blue stained gel depicting a shifted band for APT1, confirmed by western blot on the left side.

Figure 3.2 SnRK1 phosphorylates APT1 in vitro.

Figure 3.3 Endogenous SnRK1 activity copurifies with APT1 in N. benthamiana.

HA2His6-APT1 and HA2His6-GFP were expressed in N. benthamiana from the TRBO vector and purified using Nickel-NTA resin as described in methods. Aliquots of the partially purified proteins (3 µg) were then incubated with GST-SAMS (SnRK1 substrate) or GST-SAMA (negative control substrate) to detect co-purified SnRK1 activity.

Figure 3.3 Endogenous SnRK1 activity copurifies with APT1.

100

Figure 3.4 SnRK1, APT1, and ADK interact with each other in the cytoplasm. BiFC analyses of SnRK1, APT1 and ADK proteins were performed. Constructs expressing

SnRK1, APT1 and ADK fused to the N- or C-terminal portion of YFP were delivered by agroinfiltration to N. benthamiana leaves. Different combinations of the constructs were used to test interaction between APT1 and SnRK1, ADK and SnRK1, and ADK and

APT1. Representative images are shown. Cells were photographed 48 h postinfiltration using a confocal laser scanning microscope at 20x magnification.

Figure 3.4 SnRK1, APT1, and ADK interact with each other in the cytoplasm.

101

Figure 3.5 APT1 might coimmunoprecipitate with ADK. Immunoprecipitations (IPs) were performed with polyclonal antibodies raised against ADK and TMD-1 (control) proteins. Thus obtained IPs were subjected to SnRK1 assay using GST-SAMS peptide. A phosphorylated protein (*) was observed in IPs obtained from anti-ADK antibodies but not from anti-TMD1 antibodies. We speculate that this protein is APT1 based on four observations as described in the text.

(For methods please see Chapter 2, section 2.4)

Figure 3.5 APT1 might coimmunoprecipitate with ADK.

102

Figure 3.6 Overall hypothesis. We hypothesize that SnRK1, ADK, and APT1 form a complex and, under stress conditions that deplete ATP, this complex might act as a

‗sensosome‘ of stress and increase AMP levels leading to increased AMP:ATP ratio, subsequently stimulating SnRK1 activity and the cellular stress response (CSR). Also,

SnRK1 could form a positive feedback loop with both APT1 and ADK.

Figure 3.6 Overall hypothesis.

103

Table 3.1 APT1 of Arabidopsis thaliana has two evolutionarily conserved potential

SnRK1 sites. Ser16 and Thr35 (in red font and underlined) in APT1 are predicted to be phosphorylated by SnRK1. The protein sequences of APT from different taxonomic groups were compared. The amino acids in red font are crucial for SnRK1 recognition.

Hyphen (-) indicates not conserved.

% Identity, entire Organism protein S16 T35 Arabidopsis thaliana-APT1 100% IAKIASSIRV MFQDITTLLL Arabidopsis thaliana-APT2 69% - MFQDITTLLL Arabidopsis thaliana-APT3 71% IHGIKTKIRV MFQDITTVLL Brassica napus 92% IPKIASSIRV MFQDITTLLL Glycine max 80% LARIASAIRV LFQDITTLLL Populus trichocarpa 82% IARISSAIRV MFQDITTLLL Vitis vinifera 80% IARISSAIRV MFQDITTLLL Solanum tuberosum 81% IAGIASAIRV MFQDITTLLL Zea mays 80% LAGIASSIRV MFQDITTLLL Oryza sativa 79% VARIASTIRV MFQDITTMLL Triticum aestivum 77% VERIASSIRA LFQDITTLLL Physcomitrella patens 68% - MFRDVTTLLL Chlamydomonas reinhardtii 79% - LFWDVTTIML

Table 3.1 APT1 of Arabidopsis thaliana has two evolutionarily conserved potential

SnRK1 sites.

104

CHAPTER 4

GEMINIVIRUS PATHOGENICITY FACTORS INTERACT WITH SnRK1, ADK, AND APT1, A PUTATIVE COMPLEX THAT LIKELY PLAYS A ROLE IN ANTIVIRAL DEFENSE

4.1 Introduction

Geminiviruses are plant pathogens that cause crop diseases that result in large economic losses. In addition, because of their small genome size (few kilobases), they are also interesting models for fundamental research to study various host cellular events, including stress responses (Jeske, 2009). Geminivirus proteins interact with various host components either to make use of them for multiplication and spread, or to inactivate them in order to block host defense responses. Earlier, our group identified two host proteins, SnRK1 and ADK, that are targeted by geminiviruses for inactivation (Wang et al., 2003; Hao et al., 2003). The AL2 proteins encoded by TGMV and other begomoviruses, and L2 encoded by BCTV and other curtoviruses, have been specifically shown to be responsible for this inactivation.

SnRK1 belongs to the conserved SNF1-AMPK-SnRK1 family of kinases that are known to inactivate ATP-consuming biosynthetic pathways and alternatively activate

ATP-generating systems, as part of the cellular stress response (CSR). We have demonstrated that the CSR is an innate antiviral defense which is suppressed by

105 geminivirus AL2 and L2 proteins, which inactivate SnRK1 (Sunter et al., 2001; Hao et al., 2003). Transgenic N. benthamiana plants that constitutively express anti-sense

SnRK1 show enhanced susceptibility to geminiviruses and RNA viruses, while plants that over-express sense SnRK1 show enhanced resistance to infection (Hao et al., 2003).

Enhanced susceptibility or resistance was not observed in terms of disease symptoms, but with respect to virus infectivity, as measured by the concentration of virus inoculum required to cause disease in 50% of plants: ID50. Although we know that SnRK1 plays critical roles in innate antiviral defense, its downstream targets or interacting partners relevant to virus infection, have yet to be identified. In this study, we hypothesize that

SnRK1 down regulates protein synthesis in response to viral infection through phosphorylation-mediated inactivation of the translation initiation factors eIF-2 and/or eIF-(iso)4E. Similar events occur in animals as part of the interferon response (IFN) in which RNA-activated Protein kinase (PKR) kinase phosphorylates and inactivates eIF-2

(Gale and Katze, 1998). One of the purposes of this study is to ask whether a similar mechanism of antiviral defense exists in plants.

Geminivirus AL2 and L2 proteins also inactivate ADK, and this inactivation serves to inhibit host methylation of geminivirus genomes and suppress RNA silencing

(Wang et al., 2003; Wang et al., 2005; Raja et al., 2008; Buchmann et al., 2009). In an earlier study (Chapter 2), we established that ADK and SnRK1 form an in vivo complex and that this complex may promote the activities of both kinases. In support of this idea, we observed significantly increased SnRK1 activity in Arabidopsis plants expressing inducible ADK transgene and in turn, SnRK1 enhances ADK activity by ~4-6 fold in vitro, suggesting a novel role for SnRK1-ADK complex in antiviral defense (Gireesha et

106 al., unpublished data; chapter 2). Thus, because 5'-AMP generated by ADK promotes

SnRK1 activity, and the geminivirus pathogenicity proteins AL2 and L2 inhibit SnRK1 and ADK (Wang et al., 2003; Hao et al., 2003; Hardie, 2007), we speculated that inhibition of both kinases might be a dual strategy to suppress SnRK1.

ADK also plays a role in cytokinin metabolism, and this has been also recently shown to play a role in defense against DNA viruses (Baliji et al., 2010). Cytokinins are phytohormones that play a critical role in regulating plant cell proliferation and differentiation, and also control diverse processes in plant growth and development

(Sakakibara, 2006). These hormones are N6-substituted adenine derivates which exist in base, nucleoside and nucleotide forms (von Schwartzenberg et al., 1998; Sakakibara,

2006). The nucleotide form is inactive compared to other forms, and therefore the interconversion of these forms determines the level of bioactive cytokinin pool

(Sakakibara 2006; Baliji et al., 2010). ADK converts the active nucleoside form of cytokinin to the inactive nucleotide form. As geminiviruses have DNA genomes and depend on host machinery for their DNA replication, it is essential for them to induce host cell replication. Because cytokinins promote cell division and proliferation, geminiviruses act to keep them in an active state by inactivating ADK (Wang et al., 2003;

Baliji et al., 2010).

Interestingly, another cellular protein, adenine phosphoribosyl transferase

(APT), generates 5'-AMP. APT makes 5'-AMP using adenine and phosphoribosyl pyrophosphate as substrates and, like ADK, it also catalyzes the conversion of active base forms of cytokinin to inactive nucleotide forms (Sakakibara, 2006). Therefore, we initiated an investigation to test if APT plays a role in antiviral defense. Among the three

107 expressed APT proteins of Arabidopsis, we chose APT1 to study because it metabolizes adenine to 5'-AMP 30-50 times faster than APT2 and APT3 (Allen et al., 2002). Also,

APT1 deficiency leads to male sterility and slow growth, indicating that APT1 might play some unique roles for which APT2 and APT3 cannot compensate (Gaillard et al., 1998).

We found that APT1 interacts with both SnRK1 and ADK, suggesting it may be part of a ternary complex (Chapter 3). In this chapter, we report the interactions of four different geminivirus proteins AL2 (of TGMV), L2 (of BCTV), AV2 (of ToLCNDV), and C1 (of

DNA , a satellite virus) with APT1 and also with other members of this putative complex, SnRK1 and ADK. Further, we discuss the possible consequences of these interactions in the context of defense and counter-defense by hosts and viruses. We also show that SnRK1 phosphorylates translation initiation factors eIF-2 and eIF-(iso)4E in vitro. Possible consequences of these phosphorylation events are discussed.

4.2 Results

4.2.1 AL2, L2, AV2, and C1 interact with SnRK1, ADK, and APT1 in BiFC analyses

We have obtained strong evidence that SnRK1 interacts with APT1 (Chapter 3). We also know from earlier studies that geminivirus proteins AL2 and L2 interact with and inactivate ADK and SnRK1 (Wang et al., 2003; Hao et al., 2003). Therefore, we asked if

APT1 is also a target of the AL2 and L2 proteins. To test this, we first employed a BiFC approach as described in Chapter 2, to see if AL2 and L2 interact with APT1.

Construction of BiFC vectors containing Arabidopsis APT1, ADK, and SnRK1, and

108 geminivirus AL2 and L2 proteins has been described earlier (Chapters 2 and 3; Yang et al., 2007). Constructs expressing full-length Arabidopsis APT1, SnRK1, ADK or the geminivirus proteins AL2 and L2, fused to the N- or C-terminal portions of yellow fluorescent protein (YN and YC, respectively) were introduced into N. benthamiana leaf cells by agroinfiltration. Cells expressing oppositely tagged proteins (i.e., YN + YC fusion proteins potentially capable of reconstituting an active yellow fluorescent protein;

YFP) were viewed under a confocal microscope 48 h later. Histone 2B and Fibrillarin 1 fused to red fluorescent protein (RFP-H2B and RFP-Fibrillarin 1) served as a markers for the nucleus and nucleolus, respectively.

APT1 and ADK are cytoplasmic proteins, whereas SnRK1 is found in the cytoplasm and the nucleus. Geminivirus AL2 and L2 proteins localize to both the cytoplasm and the nucleus (Wang et al., 2003; Yang et al., 2007; Figure 4.1 B). When oppositely tagged APT1 and AL2 or L2 fusion proteins were tested, strong YFP fluorescence was observed both in the nucleus and the cytoplasm (Figure 4.1 A and B), indicating that AL2 and L2 interact with APT1 in vivo. For comparison, interactions of

AL2 and L2 with ADK and SnRK1 are also shown (Figure 4.1 A and B). However, it is intriguing not to observe any interaction between SnRK1 and L2 in repeated BiFC experiments despite previous evidence of their interaction (Hao et al., 2003). As discussed in Chapter 3, negative control for APT1 BiFC interactions remains to be determined.

Next, we tested two more geminivirus proteins, AV2 (of ToLCNDV) and C1

(protein encoded by satellite DNA ) that are known to act as pathogenicity factors and suppressors of RNA silencing, similar to AL2 and L2. But, they perform different

109 functions in their native viruses. BiFC constructs expressing full length AV2 and C1 were constructed using similar methods that were used to construct AL2 and L2. When oppositely tagged APT1 and AV2 or L2 fusion proteins were tested, strong YFP fluorescence was observed (Figure 4.1 C, D and E), indicating that AV2 and C1 also interact with APT1 in vivo. In similar experiments, we also observed that AV2 and C1 self-interact (Figure 4.1 D and E). Interaction of AV2 with SnRK1 appears as spots in the cytoplasm adjoining the plasma membrane while its interactions with ADK and APT1 are mostly in the cytoplasm (Figure 4.1 C). AV2 self-interaction is observed in the cytoplasm and also in the spots along the cytoplasm (Figure 4.1 D). We speculate that these spots are localized to the plasmodesmata because AV2 is required for virus cell-to-cell movement that takes place through plasmodesmata (Briddon et al., 1989; Rothenstein, et al., 2007). For the same reason we speculate that SnRK1-AV2 complexes are also localized to the plasmodesmata. Because, YFP fluorescence for SnRK1-AV2 complexes is limited only to the spots, we used chloroplast autofluorescence (red) to show that the spots are along the plasma membrane, similar to AV2-AV2 complexes. (Figure 4.1 D).

C1 interacts with ADK and APT1 mostly in the cytoplasm and some complexes appear to localize in the nucleus (Figure 4.1 E), and previous yeast-two-hybrid experiments indicate that C1 and SnRK1 also interact (X. Zhou, unpublished data). As negative controls, we found that L2 does not interact with itself (Figure 4.1 B), and C1 do not interact with DCL3 (Figure 4.1 E). However, DCL3 interacts with DRB3 (P. Raja, unpublished data). Negative control for AV2 interaction (a plant protein that does not interact with AV2 in a BiFC assay) needs to be determined. Further investigation is required to test whether the new interactions we discovered are direct or indirect, and to

110 determine the consequences of these interactions. However, interactions of four different viral proteins with SnRK1, ADK, and APT1 suggest that these proteins might be important to virus replication.

4.2.2 AL2 forms nuclear speckles with SnRK1 but not with ADK

Using the BiFC approach, we further characterized the subnuclear localization of AL2-

SnRK1 and AL2-ADK complexes. For this study, BiFC plasmids were coinfiltrated with plasmids containing red fluorescent protein fused to Histone2b (RFP-Histone 2b) or with plasmids containing RFP fused to Fibrillarin 1 (RFP-Fibrillarin 1). RFP-Histone 2b and

RFP-Fibrillarin 1 were used as markers for nucleus and nucleolus, respectively

(Chakrabarty et al., 2007). We found that AL2-SnRK1 complexes form nuclear speckles and none of these appear to localize in the nucleolus (Figure 4.2 A and B). On the contrary, AL2-ADK complexes localize uniformly within the nucleus except in the nucleolus (Figure 4.2 C).

4.2.3 AL2 interacts with APT1 in an in vitro GST-pull down assay

We next employed an in vitro GST-pull down assay to test whether the interaction between geminivirus AL2 and Arabidopsis APT1 proteins is direct or indirect. For this experiment, we used GST-tagged AL2 expressed in E. coli and HA2His6-tagged APT1 expressed in N. benthamiana. Two µg of GST-AL2 or GST (control protein) was incubated with two µg of APT1 or ADK in vitro in 1x PBS buffer with 10 µM of MG132

111

(a proteasomal inhibitor) and 1x plant protease inhibitor cocktail (Sigma-Aldrich) for 1 hr at 4oC. In one of the treatments, only APT1 was included to show that it does not interact non-specifically with GST-sepharose (negative control). GST pull-down was performed using GST-sepharose followed by western blot analyses with appropriate antibodies. We found that APT1 interacts with GST-AL2 but not with GST or GST-sepharose (negative controls) (Figure 4.3). As a positive control, we employed ADK (HA2His6-tagged) which is known to interact with AL2. Anti-GST antibody was used to confirm the presence of

GST and GST-AL2, and anti-HA antibody was used to detect ADK and APT1. We hypothesize that AL2 (and possibly other viral proteins tested above) interact with APT1 to inactivate it. Further experiments need to be done to test this hypothesis.

4.2.4 SnRK1 phosphorylates translation initiation factors eIF-2 and eIF-(iso)4E in vitro

In animals, viral infections induce host interferon (IFN) responses which results, among other effects, in the activation of PKR. Activated PKR phosphorylates and inactivates the eukaryotic translation initiation factor-2 (eIF-2) leading to decreased protein translation. This inhibits viral protein synthesis and reduces virus replication. Not surprisingly, most if not all successful animal viruses have developed a mechanism/s to combat PKR-mediated inactivation of translation initiation machinery (Gale and Katze,

1998). However, the functional equivalent of PKR in plants remains to be identified, although a PKR-like activity has been reported (Hiddinga et al., 1988; Crum et al., 1988;

Langland et al., 1996). We know from our earlier work that SnRK1 plays an important

112 role in innate antiviral defense not only against DNA viruses but also against RNA viruses, and that the geminivirus AL2 and L2 proteins directly interact with and inactivate SnRK1 (Hao et al 2003). Upon observing the interaction of additional geminivirus proteins (AV2 and C1) with SnRK1, we hypothesized that SnRK1 might play a role similar to PKR in plants. To examine this, we first manually scanned the sequences of two Arabidopsis thaliana eukaryotic translation initiation factors, AteIF-2 and AteIF-(iso)4E (from now on we would refer to these as eIF-2 and eIF-(iso)4E, respectively, for convenience) for the presence of SnRK1 consensus sites. Arabidopsis encodes one gene for eIF-(iso)4E and three genes for eIF4E (Robaglia and Caranta,

2006), and SnRK1 phosphorylation site is conserved in all of these proteins. However, eIF-(iso)4E was selected for this study because it is unique to plants. Arabidopsis has two eIF-2 genes, eIF-2 and eIF-22, and they share 84% amino acid identity. Predicted protein masses for these two proteins are 38 kD and 28 KD, respectively. However, both have conserved SnRK1 sites. To our knowledge, eIF-22 has not been reported yet to encode a fully functional protein, but eIF-2 has been (Browning, 1996; Zhang et al.,

2008; Lageix et al., 2008). Therefore, we studied eIF-2.

There are three SnRK1 sites present in eIF-2(S56, S65, and T201) and one in eIF-(iso)4E (T57) and interestingly, these sites are evolutionarily conserved (Tables 4.1 and 4.2). The SnRK1 sites of eIF-2and eIF-(iso)4E are conserved across photosynthetic eukaryotes (monocotyledonous and dicotyledonous plants, mosses, and algae). In addition, S56 and T201 of eIF-2 are conserved in non-photosynthetic organisms, including yeast, insects, mammals, etc. (Tables 4.1 and 4.2).

113

Next, we expressed and purified GST-tagged Arabidopsis eIF-2 and eIf-(iso)4E in E. coli to test if SnRK1-KD, expressed in N. benthamiana, can phosphorylate these proteins, in vitro. In separate in vitro kinase assays, 5-10 ng of SnRK1-KD or SnRK1-

KD-K49R (kinase-inactive mutant) was incubated with 2-3 µg of GST-eIF-2 or GST- eIF-(iso)4E and 32P-labeled -ATP. To obscure SnRK1-KD autophosphorylation

(because SnRK1-KD and GST-(iso)-4E migrate at similar positions during PAGE separation) and achieve better activation, SnRK1-KD was pre-incubated with 0.5 mM unlabeled ATP.

Both the translation initiation factors we tested are phosphorylated by SnRK1-KD in in vitro kinase experiments (Figure 4.4). However, we observed background levels of phosphorylation by SnRK1-KD-K49R preparations (the kinase-inactive mutant). More experiments need to be done to prove the in vivo relationships between SnRK1 and the above-mentioned translation initiation factors in regulating the protein synthesis.

4.3 Discussion

In this study, we show that four different geminivirus proteins; AL2 (of TGMV), L2 (of

BCTV), AV2 (of ToLCNDV), and C1 (of DNA , a DNA satellite) interact with SnRK1,

ADK, and APT1 in BiFC analyses (Figure 4.1 A to E). From earlier studies we knew that

AL2 and L2 interact with and inactivate ADK and SnRK1 (Wang et al., 2003; Hao et al.,

2003). Here we characterized the subcellular localizations of the complexes formed by

AL2 as well as L2 with ADK and SnRK1 using BiFC approach. AL2 and L2 interact with ADK and APT1 both in the nucleus and the cytoplasm (Figure 4.1 A and B). This is

114 interesting because ADK and APT1 are not known to localize to the nucleus (Moffat et al., 2002; Somers, personal communication). AL2 and L2 localize to both the cytoplasm and the nucleus on their own (Figure 4.1 B; Yang et al., 2007). Therefore, it is conceivable that AL2 and L2 might be carrying some of ADK and APT1 into the nucleus. We do not know the significance of ADK and APT1 relocalization. On the other hand, AL2 interacts with SnRK1 predominantly in the nucleus (Figure 4.1 A).

Interestingly, AL2-SnRK1 complexes form ‗nuclear speckles‘ and none of these appear to localize to the nucleolus (Figure 4.2 A and B). By contrast, AL2-ADK complexes do not form ‗nuclear speckles‘ rather they are uniformly distributed throughout the nucleus except in the nucleolus (Figure 4.2 C). We do not know the significance of ‗nuclear speckles‘ formed by AL2 and SnRK1 but speculate that they might be associated with transcriptional regulation of genes associated with stress responses (Baena-Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008; Kleinow et al., 2009). Our speculation is based on some recent findings and what we already knew about AL2. First, AL2 is a transcription activator and its homologs in Mung bean yellow mosaic virus (MYMV) and

African cassava mosaic virus (ACMV), have been shown to activate more than 30 host genes (Trinks et al., 2005). A recent finding also demonstrates that transactivation of late viral genes by AL2 requires its binding to a plant-specific transcription factor called

Peapod 2 (Lacatus and Sunter, 2009). SnRK1 has also been shown to integrate diverse energy and stress signals on a global scale by inducing extensive changes in the transcriptome that have the effect of promoting catabolism and inhibiting anabolic pathways (Baena-Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008). In addition,

SnRK1 has been shown to interact with ATAF1, a stress-responsive, NAC domain

115 transcription factor in Arabidopsis (Kleinow et al., 2009). Roles for yeast Snf1 (ortholog of SnRK1) in transcriptional regulation are well established. SNF1 acts as a histone kinase, and in concert with the histone acetyltransferase Gcn5 and RNA polymerase II, it regulates transcription (Kuchin et al., 2000; Lo et al., 2001). Therefore, it would be interesting to test if the ‗nuclear speckles‘ formed by AL2-SnRK1 complexes are associated with transcriptional regulation. However, we failed to see interaction between

L2 and SnRK1 in our BiFC experiments (tested in all possible combinations) although we know from earlier findings that they interact (Hao et al., 2003). Our speculation is that

YFP fusions made to SnRK1 and L2 might be causing some kind of hindrance to their interaction. However, the YFP fusions do not appear to hinder interactions of SnRK1 or

L2 with other proteins e.g. ADK (Figures 2.1 and 4.1).

In BiFC assays, we found that AL2, L2, AV2, and C1 interact with APT1

(Figure 4.1 A-E; Chapter 3). All of the viral proteins tested play different roles in their native viruses but share several features: all are pathogenicity factors and act as suppressors of RNA silencing (Saunders et al., 2000; Briddon and Stanley 2006; Wang et al., 2005; Mubin et al., unpublished data). The AL2-APT1 interaction appears to be direct, as demonstrated by a GST pull-down assay. However, the consequences of this interaction and the remaining interactions remain to be investigated. We hypothesize that the geminivirus pathogenicity factors inactivate SnRK1, ADK, and APT1 to block a range of host defense responses. Apparently, these proteins are ―high-value‖ targets for geminiviruses.

Why are SnRK1, ADK, and APT1 targeted by geminiviruses? First, ADK and

APT1 make 5'-AMP which is involved in mediating the activation of SnRK1, which

116 plays a key role in innate antiviral defense (Sugden et al., 1999a; Hao et al., 2003).

Second, ADK and APT1 are also involved in the metabolism of cytokinins (Allen et al.,

2002; Sakakibara H, 2006), phytohormones which play a crucial role in various phases of plant growth and development (Sakakibara, 2006). These hormones are N6-substituted adenine derivates which exist in base, nucleoside, and nucleotide forms (von

Schwartzenberg et al., 1998; Sakakibara, 2006). The nucleotide form is inactive compared to other forms, and therefore the interconversion of these forms regulates bioactive cytokinin pools (Sakakibara 2006; Baliji et al., 2010). Both ADK and APT can convert the active form (nucleoside or base) of cytokinin to the inactive form (nucleotide)

(Figure 4.5). In the case of ADK, this conversion has already been shown to be important for geminivirus infection (Baliji et al., 2010). Because cytokinins promote cell division, it would be advantageous for DNA-containing geminiviruses to keep these hormones in an active state to facilitate their replication and multiplication. In fact, exogenous application of cytokinin results in an increased susceptibility to geminivirus infection characterized by reductions in mean latent period and enhanced viral replication (Baliji et al., 2010).

So, by inactivating ADK and APT1 geminiviruses can increase the pool of active cytokinins to facilitate their own multiplication.

Third, like ADK, APT1 might also play a role in the maintenance of the ―methyl cycle‖. Phosphorylation of adenosine by ADK is an important step for methyl cycle to advance in forward direction to facilitate transmethylation reactions (Figure 4.6, modified from Moffat et al., 2002). Host methylation of geminivirus DNA is a significant host defense response and geminivirus proteins block this methylation at least in part by inactivating ADK (Wang et al., 2003; Raja et al., 2008; Buchmann et al., 2009).

117

However, salvage of adenosine can happen in two more ways in addition to that which requires ADK. One is by conversion of adenosine to inosine by adenosine deaminase, and the other one involves two steps. First, adenosine gets converted to adenine by adenosine nucleosidase followed by conversion of adenine to 5'-AMP by APT (Allen et al., 2002; Moffat et al., 2002). Although basal adenosine nucleosidase activity is very low

(Auer, 1999; Moffatt et al., 2002; Katahira and Ashihara, 2006), its activity has been shown to increase by 100 fold during wound-induced stress in potato (Katahira and

Ashihara, 2006). In comparison, APT activity was increased about two fold, and adenosine may therefore be readily converted to adenine by adenosine nucleosidase, followed by subsequent conversion of adenine to 5'-AMP by APT (Katahira and

Ashihara, 2006). Thus, inactivation of APT1 by geminiviruses might also contribute to blocking the methylation of their DNA.

Fourth, SnRK1 might negatively regulate protein synthesis by inactivating one or more translation initiation factors, akin to the inactivation of eIF-2 by PKR as part of antiviral interferon response in animals (Figure 4.7; Gale and Katze, 1998). Plants do not contain PKR gene as revealed by Arabidopsis and rice genome data, although some studies have demonstrated PKR-like activity (Hiddinga et al., 1985; Crum et al., 1988;

Langland et al., 1996; Browning 2004). We speculate that SnRK1 might play a role similar to PKR in plants. At least, two findings support this hypothesis. First, eIF-2 and eIF-4E/(iso)4E possess SnRK1 consensus sites that are conserved across photosynthetic and non-photosynthetic eukaryotes (Tables 4.1 and 4.2). Second, SnRK1 phosphorylates these translation initiation factors in vitro. In animals, PKR is not known to phosphorylate eIF-4E, and AMPK also may not phosphorylate as we failed to find any

118

AMPK (SnRK1) sites in model animal species. However, regulation of translation involving phosphorylation of eIF-4E and eIF-4E binding proteins (4E-BP) is well- established in mammals (Frederickson et al., 1991; Morley and Traugh, 1993; Fraser et al., 1999; Khan and Goss, 2004). Intriguingly, plants appear to lack 4E-BP proteins, and might employ slightly different strategies to regulate protein synthesis. We speculate that

SnRK1-mediated phosphorylation of eIF-(iso)4E might be one of them. In fact, a reduction in affinity of wheat eIF-(iso)4E (1.2-2.6 fold) to mRNA cap was observed upon phosphorylation by casein kinase II in in vitro studies (Khan and Goss, 2004). Similar changes might be happening to eIF-(iso)4E upon SnRK1 phosphorylation. Further investigations are required to confirm this hypothesis. However, should this proved to be the case, it would be a significant milestone in understanding plant viral pathogenesis.

It appears that the SnRK1 pathway is important for plant defenses against viruses.

In fact, the expression levels of SnRK1 and its upstream activating kinases GRIK1 and

GRIK2, increase in plants infected with geminiviruses (G Sunter, unpublished data; Shen and Hanley-Bowdoin, 2006). Also, for GRIK1 and GRIK2, higher levels of protein accumulation have been demonstrated in the infected plants. It remains to be seen if viral infection also leads to increased expression and/or protein accumulation of ADK and

APT1. Because different geminiviruses interact with SnRK1, ADK, and APT1, we speculate that a putative complex of these proteins might be central to anti-viral defense responses in plants.

119

4.4 Materials and Methods

4.4.1 Plant material, agroinfiltration, and bimolecular fluorescence complementation (BiFC)

ADK, and the viral proteins AL2, L2, AV2, and C1. cDNAs of AV2 and C1 were obtained from our collaborators (Shahid Mansoor, National Institute for Biotechnology and Genetic Engineering, Pakistan; Xueping Zhou, Zhejiang University, China) and the sources of L2, AL2, SnRK1, APT1, and ADK cDNAs have been mentioned earlier

(Chapter 3; Wang et al., 2003; Hao et al., 2003). The genes were amplified using PCR to incorporate either 5' PacI-3' SpeI or 5' NotI-3' XbaI restriction sites. The resulting PCR products were either digested with PacI and SpeI or NotI and XbaI restriction enzymes and ligated into PacI-SpeI-digested p2YN and p2YC vectors (to make C-terminal fusions to either the N- or C-terminal portion of YFP) or NotI-XbaI-digested pYC1 and pYN1 vectors (to make N-terminal fusions to either the N- or C-terminal portion of YFP).

For BiFC experiments, plasmids were transformed into A. tumefaciens C58C1 and cultures were used to infiltrate N. benthamiana leaves (Wang et al., 2005). Cultures containing YN- and YC-based plasmids were mixed 1:1 immediately prior to infiltration.

120

Histone 2b and Fibrillarin 1 fused to red fluorescent protein (RFP-H2B and RFP-

Fibrillarin 1), were used as nuclear and nucleolar markers, respectively. Plasmids expressing RFP-H2B and RFP-Fibrillarin 1 (Chakrabarty et al., 2007) were kindly provided by Michael Goodin (University of Kentucky). Leaf tissue was analyzed by microscopy ~48 h post-infiltration using a Nikon PCM 2000 confocal laser scanning microscope equipped with argon and green helium neon lasers with excitation wavelengths of 488 nm and 544 nm, respectively. To record YFP fluorescence, a band- pass emission filter (EM515/30HQ) with a 450- to 490 nm excitation wavelength and 515 nm emission wavelength was used. To record RFP fluorescence, a 565 nm long-pass filter (E565LP) was employed. Images were captured using Simple PCI Software and compiled with Adobe Photoshop.

4.4.2 Protein expression and purification

Expression and purification of proteins in N. benthamiana

SnRK1-KD and SnRK1-KD-K49R constructs have been described previously (Hao et al.,

2003). PCR products of Arabidopsis SnRK1-KD and SnRK1-KD-K49R genes were cloned into PacI and AvrII digested pJL-TRBO, a Tobacco mosaic virus RNA-based over-expression vector to generate N-terminal double hemagglutinin peptide-six histidine

(HA2His6) fusion proteins (Lindbo, 2007 a; Lindbo, 2007 b). The plasmids were transformed into A. tumefaciens C58C1, and cultures were used to infiltrate N. benthamiana leaves as described (Wang et al, 2005). Tissues were collected ~5 days post infiltration and ground in liquid nitrogen followed by the addition of 1.25 volumes of

121

extraction buffer [50 mM HEPES (pH 7.5), 0.1% (v/v) Triton X-100, 10 mM MgCl2, 50 mM NaF, 1 mM EGTA, 1 mM benzamidine, 10 uM MG132-proteasomal inhibitor, 1x plant protease inhibitor cocktail (Sigma), 5 mM β-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride]. Extract was filtered through miracloth and centrifuged at

12,000g for 20 min at 4°C. Clarified supernatant (20-25 ml) was then added to 0.5 ml of washed nickel nitrilotriacetic acid agarose beads (Invitrogen) and incubated at 4°C on a rocker for 2 to 3 h. Columns were washed in 6-12 column volumes of wash buffer [50 mM NaH2PO4 (pH 8.0), 300 mM NaCl, 0.1% (v/v) Tween 20, 5 mM β-mercaptoethanol, with or without 20 mM imidazole]. Bound HA2His6-tagged proteins were eluted with wash buffer containing 250 mM imidazole at 4°C, and then dialyzed in buffer containing

40 mM Tris-HCl (pH 7.5), and 10% (v/v) glycerol. Protein concentrations were estimated using the Bradford assay (BioRad) with bovine serum albumin (BSA) as the standard.

Expression and purification of proteins in E. coli

Arabidopsis AteIF-2 and AteIF-(iso)4E were amplified by PCR and cloned into NcoI-

SacI sites of pGEX-KG vector to make protein fusions to GST. Construction of GST-

AL2 has been described previously (Wang et al., 2003). cDNAs of AteIF-2 and AteIF-

(iso)4E were obtained from ABRC (stock# U82491 and U16070, respectively). GST-

AL2, GST, GST-eIF-2, and GST-eIF-(iso)4E proteins were expressed in E. coli strain

BL21 and purified by glutathione-agarose chromatography (Hartitz et al., 1999). All expressed proteins were confirmed by western blot analyses using appropriate antibodies.

122

4.4.3 SnRK1 kinase assay and GST pull down assay

SnRK1 Kinase assay.

GST-eIF-2 and GST-eIF-(iso)4E phosphorylation (by SnRK1-KD) assays were performed essentially as described previously (Celenza and Carlson, 1989). Reactions contained 10 to 15 ng of SnRK1-KD (or SnRK1-KD-K49R) with 2-4 µg of substrate proteins (GST-eIF-2 or GST-eIF-(iso)4E. Reactions were initiated by the addition of 0.5 mM unlabeled ATP and incubated at 30°C for 20 min to activate SnRK1 and also in some cases to obscure SnRK1-KD autophosphorylation. -32P-ATP (3000 Ci/mmol; Perkin

Elmer) was then added to a final concentration of 0.05 µM along with substrate protein, and reaction mixtures incubated a further 30 min at 30°C before electrophoresis on SDS–

10% (w/v) polyacrylamide gels. Labeled proteins were visualized and quantitated using a phosphorimager (BioRad).

GST pull down assay

Expression and purification of GST-AL2, GST, HA2His6-APT1 and HA2His6-ADK was carried out as described above. Partially purified GST-AL2 or GST (negative control) was incubated with HA2His6-APT1 or HA2His6-ADK (positive control) with GST-

Sepharose in 1x pull-down buffer (1x PBS, 10 uM MG132, a proteasomal inhibitor, 1x plant protease inhibitor cocktail (Sigma)) at 4oC for 1h. Following incubations, the reaction mixtures were centrifuged at 3000g for 1 min and then washed with 1x PBS buffer three times. The GST-sepharose beads with pull-down complexes were subjected

123 to SDS-PAGE and western blot analyses with anti-GST antibody and, anti-HA antibody.

To compare the migrations, the GST-AL2 and HA2His6-APT1 proteins were directly loaded onto the gel without incubating with GST-Sepharose.

4.4.4 Protein sequence comparison using BLAST analysis

Protein sequences of AteIF-2 and AteIF-(iso)4E were retrieved from GenBank (NCBI) and subjected to ‗protein BLAST‘ to recover the related protein sequences. Using these protein sequences, manual searches were performed to identify SnRK1 consensus sites based on the features described by Halford and Hardie (1998), and to determine the extent of the SnRK1 site conservation.

124

Figure 4.1 AL2, L2, AV2, and C1 interactions with SnRK1, ADK and APT1. BiFC analyses of viral proteins (AL2, L2, AV2 and C1) with Arabidopsis SnRK1, APT1, and

ADK proteins in N. benthamiana epidermal cells were performed. Constructs expressing

AL2, L2, AV2, C1, SnRK1, APT1, and ADK fused to the N- or C-terminal portion of

YFP were delivered by agroinfiltration to N. benthamiana leaves. Different combinations of the constructs were used to test interactions of:

A. AL2 with APT1, ADK, and SnRK1

B. L2 with APT1 and ADK

C. AV2 with APT1, ADK, and SnRK1

D. AV2 with AV2, and SnRK1 with or without chloroplast autofluorescence background

E. C1 with APT1, ADK, and itself. Although L2 is known to interact with SnRK1 (Hao et al., 2003), we failed to see their interaction in our BiFC analyses (data not shown). L2 fused to double GFP localizes both to the nucleus and the cytoplasm (B). L2 and APT1 do not interact with DCL3 (negative controls)

Cells were photographed 48 h postinfiltration using a confocal laser scanning microscope. The images were captured at 20- or 40x magnification, as indicated.

125

Continued

Figure 4.1 AL2, L2, AV2, and C1 interact with SnRK1, ADK and APT1 in BiFC analyses.

126

Figure 4.1 continued

127

Figure 4.2 AL2-SnRK1 complexes accumulate in nuclear speckles. BiFC analyses of viral protein AL2 with Arabidopsis SnRK1 and ADK proteins in N. benthamiana epidermal cells were performed. Constructs expressing AL2, SnRK1, and ADK fused to the N- or C-terminal portion of YFP were delivered by agroinfiltration to N. benthamiana leaves.

A. AL2 interacts with SnRK1 in the nucleus

B. AL2-SnRK1 complexes form nuclear speckles and none of these appear to localize in the nucleolus

C. AL2 interacts with ADK in the cytoplasm and in the nucleus. AL2-ADK complexes are uniformly distributed throughout the nucleus except in the nucleolus.

Cells were photographed 48h postinfiltration using a confocal laser scanning microscope. The images are captured at 20-, 40- or 100x magnification, as indicated.

RFP-Histone 2B and RFP-Fibrillarin 1 were used as nuclear and nucleolar markers, respectively.

128

Figure 4.2 AL2-SnRK1 complexes accumulate in nuclear speckles.

129

Figure 4.3 AL2 interacts with APT1 in a GST pull-down assay. Partially purified GST-

AL2 (expressed in E. coli) was incubated with HA2His6-APT1 (expressed in N. benthamiana) with GST-Sepharose in 1x pull-down buffer (1x PBS, 10 uM MG132, a proteasomal inhibitor, 1x plant protease inhibitor cocktail {Sigma} ) at 4oC for 1 hour.

Following the incubations, pull-down was performed as described in methods and subjected to SDS-PAGE and western blot analysis. GST (expressed in E. coli) was used as a negative control and HA2His6-ADK (expressed in N. benthamiana) was used as a positive control.

WB control: Western blot control. To compare migrations of proteins obtained from

GST-pull-down, GST-AL2 or HA2His6-APT1 proteins were directly loaded onto gel without subjecting them to GST-pull-down assay.

Figure 4.3 AL2 interacts with APT1 in a GST pull-down assay.

130

Figure 4.4 SnRK1 phosphorylates AteIF-2 and AteIF-(iso)4E in vitro. Partially purified

GST-AteIF-2 and GST-AteIF-(iso)4E proteins were incubated with partially purified

HA2His6-tagged SnRK1-KD or SnRK1-KD-K49R (kinase-inactive mutant). The recombinant eIF-2 and eIF-(iso)4E proteins were expressed in E. coli, while SnRK1-

KD and SnRK1-KD-K49R were expressed in N. benthamiana. Proteins were incubated in kinase reactions with 32P-labeled γ-ATP. Following the kinase reactions, samples were separated on a SDS-PAGE gel and exposed to a phosphorimager to detect 32P incorporation.

Note that SnRK1 autophosphorylation signal is not seen (left panel) because the enzymes were preincubated with 0.5 mM of cold ATP for 20min in a kinase buffer before incubating with the substrates and 32P-labeled γ-ATP. Western blot analysis was carried out for AteIF-2 and AteIF-(iso)4E proteins using anti-GST antibody (right panel).

Figure 4.4 SnRK1 phosphorylates AteIF-2 and AteIF-(iso)4E in vitro.

131

Figure 4.5 APT plays a key role in cytokinin metabolism. Cytokinins are phytohormones that are N6-substituted adenine derivates which exist either as bases, nucleosides or nucleotides. Base and nucleosides are active compared to nucleotide forms. APT and

ADK convert base and nucleoside forms of cytokinin to nucleotide form, respectively

(Sakakibara H, 2006).

Figure 4.5 APT plays a role in cytokinin metabolism.

132

Figure 4.6 Methyl cycle. SAM (S-adenosyl methionine) is the methyl donor for most transmethylation reactions. The product SAH (S-adenosyl homocysteine) is a methyltransferase inhibitor. SAH is converted to Hcy (homocysteine) and adenosine by

SAHH (S-adenosyl homocysteine hydrolase). Salvage of adenosine by ADK (adenosine kinase) is important because the SAHH reaction is reversible and the equilibrium lies in the direction of SAH synthesis. Adenosine can also be salvaged into 5‘AMP with the help of adenosine nucleosidase and APT (adenine phosphoribosyl transferase) (Modified from Moffat et al., 2002).

Figure 4.6 Methyl cycle.

133

Figure 4.7 Significance of SnRK1-ADK-APT1, a putative complex, in blocking protein synthesis as part of antiviral defense. We hypothesize that in response to virus infection

ADK and APT1 help stimulate SnRK1 which would then phosphorylate and inactivate translation initiation factors eIF-2 and/or eIF-(iso)4E to block initiation of protein synthesis, and thus deny the viruses protein translation machinery.

Figure 4.7 Significance of SnRK1-ADK-APT1, a putative complex, in blocking protein synthesis as part of antiviral defense.

134

Table 4.1 AteIF-2 has three evolutionarily conserved potential SnRK1 sites. Ser56,

Ser65, and Thr201 (in red font and underlined) are the amino acids predicted to be phosphorylated by SnRK1. The sequences of eIF-2 from diverse taxonomic groups were compared to determine the extent of conservation of SnRK1 sites. The residues in red bold font are crucial for SnRK1 recognition. Hyphen (-) indicates not conserved.

%Identity, entire Organism protein S56 S65 T201 Arabidopsis thaliana- 100% LFSELSRRRI IRSVSSLIKV IRR RMTPQPM eIF-2 Arabidopsis thaliana- 84% LFSELSRRRI IRSISSLIKV IRRRMTPQPM eIF-22 Glycine max 82% LFSELSRRRI IRSVSSLIKV IRRRMTPQPL Populus trichocarpa 81% LFSELSRRRI IRSVSSLIKV IRRRMTPQPL Ricinus communis 79% LFSELSRRRI IRSVSSLIKV IRRRMTPQPL Picea sitchensis 78% LFSELSRRRI IRSVSSLIKV IRRRMTPQPL Medicago truncatula 78% LFSELSRRRI IRSVSSLIKV IRRRMTPQPM Vitis vinifera 77% LFSELSRRRI IRSVSSLIKV IRRRMTPQPL Oryza sativa 74% LFSELSRRRI IRSISSLIKV IRRRMTPQPL Zea mays 74% LFSELSRRRI IRSISSLIKV IRRRMTPQPL Physcomitrella 70% LLSELSRRRI IRSISSLIKV IRRRMTPQPL Patens Chlamydomonas 60% LLSELSRRRI - IKRRMTPQPL reinhardtii Schizosaccharomyces 55% LLSELSRRRI - ISRRLTPQPL japonicus Saccharomyces 50% LLSELSRRRI - ISKRLTPQAV cerevisiae Drosophila 43% LLSELSRRRI - - Melanogaster Bombyx mori 47% LLSELSRRRI - IKRKLTSQAV Mus musculus 44% LLSELSRRRI - INRRLTPQAV Homo sapiens 44% LLSELSRRRI - INRRLTPQAV

Table 4.1 AteIF-2 has three evolutionarily conserved potential SnRK1 sites.

135

Table 4.2 AteIF-4E/AteIF-(iso)4E has an evolutionarily conserved potential SnRK1 site.

Thr57 (in red font and underlined) is the amino acid predicted to be phosphorylated by

SnRK1. The sequences of eIF-4E/eIF-(iso)4E from diverse taxonomic groups were compared to determine the extent of conservation of SnRK1 sites. The residues in red bold font are crucial for SnRK1 recognition.

%Identity, entire Organism protein T57 Arabidopsis thaliana-eIF-(iso)4E 100% LRKAYTFDTV Arabidopsis thaliana-eIF-4E1 51% LRPVFTFSTV Arabidopsis thaliana-eIF-4E2 43% LRSLYTFATI Arabidopsis thaliana-eIF-4E3 43% LRSLYTFGTI Solanum lycopersicum 100% LRKAYTFETV Nicotiana tabacum 91% LRKAYTFETV Capsicum annum 91% LKKAYTFDTV Populus trichocarpa 81% LRKIYSFETV Vitis vinifera 80% LRKAYTFETV Cucumis melo 80% LRKVYTFETV Lactuca sativa 74% LRNGYTFDTV Ricinus communis 73% LRKVYTFDTV Oryza sativa 71% LRKAYTFDTV Zea mays 70% LKKAYTFDTV Triticum aestivum 69% LKKGYTFDTV Carica papaya 69% LRKVFTFDTV Physocmitrella patens 57% LRAVYTFDTV Picea sitchensis 54% LRSVYTFKTV Chlamydomonas reinhardtii 46% LRSVYTFDTV

Table 4.2 AteIF-4E/AteIF-(iso)4E has an evolutionarily conserved potential SnRK1 site.

136

CHAPTER 5

DISCUSSION

5.1 SnRK1-ADK-APT1 Interactions and Their Consequences

5.1.1 SnRK1 and ADK form a complex in vivo that could rapidly activate CSR

SNF1-AMPK-SnRK1 family kinases function as a heterotrimeric complex consisting of an catalytic subunit, and  and  regulatory subunits (Hardie, 2007; Polge and Thomas,

2007). These kinases are known to regulate metabolism in response to stresses that deplete ATP. Under stress conditions, AMP levels rise leading to increased AMP/ATP ratios (Hardie et al., 1998; Hardie, 2007). Phosphorylation by upstream kinases activate

SNF1-AMPK kinases, and phosphatases are involved in deactivating these kinases

(Sugden et al., 1999a; Suter et al., 2006; Sanders et al., 2007). In at least some cases, the upstream kinases are constitutively active, and their activity is not increased by stress

(Suter et al., 2006; Sanders et al., 2007; Baena-Gonzalez et al., 2007). Thus, maintenance of activated (phosphorylated) SNF1-AMPK-SnRK1 is likely a major mechanism of kinase regulation. It is known that 5'-AMP inhibits the inactivation/dephosphorylation of

SNF1-AMPK-SnRK1 via AMP binding to the  regulatory subunit. Thus, increasing local concentrations of 5'-AMP could dramatically affect SNF1-AMPK-SnRK1 activity,

137 leading to its rapid stimulation. How 5'-AMP levels are increased is not completely understood, although adenylate kinase has been known to play a role in generating 5'-

AMP using ADP as substrate molecules (Hardie et al., 1998). The reaction catalyzed by adenylate kinase is: 2ADP ATP + AMP. We hypothesize that ADK, which also generates 5'-AMP, forms a complex with SnRK1, and is involved in elevating local 5'-

AMP levels under stress conditions, and thus rapidly activates the cellular stress response

(CSR). In support of our hypothesis, we found that ADK and SnRK1 form a complex in vivo, and that these two kinases mutually stimulate each other (Chapter 2). In vitro, we found that the SnRK1-KD (kinase domain) stimulates ADK activity by several fold

(Figure 2.6 A). Interestingly, this stimulation does not require SnRK1 kinase activity, as a similar ADK stimulation was observed with SnRK1-KD-K49R (kinase-inactive mutant)

(Figure 2.6 A). In addition, we observed comparable ADK stimulation by SnRK1-KD

(and SnRK1-KD-K49R) expressed in E. coli, where upstream activating kinases are not present and the expressed kinase is essentially inactive (Figure 2.6 B; Shen et al., 2009).

Further, ADK stimulation by SnRK1-KD expressed in plants or E. coli reaches a maximum at stoichiometric ratios of close to 1:1 (Figure 2.6 C). Together, these data clearly demonstrate that SnRK1 enhances ADK activity, and that enhancement does not require SnRK1 kinase activity. We speculate that physical interaction might alter the conformation of ADK in a manner that facilitates substrate binding and/or phosphoryl transfer. Thus enhanced ADK activity could lead to rapid increase in the local AMP concentrations around SnRK1 complex and thus, might help in the maintenance of activated SnRK1 heterotrimeric in vivo complexes. In support of this idea, we observed a significant increase (~ 1.8 fold) in SnRK1 activity in transgenic Arabidopsis plants

138 expressing ADK (Figure 2.7 A). Although this might be an indirect effect, we presume that overexpressed ADK generates more 5'-AMP which in turn might block dephosphorylation-mediated inactivation of SnRK1, thus leading to an increase in active

SnRK1 pools. Our speculation is based on the known effects of 5'-AMP in blocking dephosphorylation of SnRK1 (Sugden et al., 1999a) and AMPK (Suter et al., 2006;

Sanders et al., 2007). Therefore, we conclude that SnRK1-ADK complex might play a significant role in rapidly activating the CSR. Additional in vitro experiments with purified heterotrimeric SnRK1 complex would provide additional support for this conclusion. Other interesting questions include, SnRK1:ADK complex levels, and whether these are stable or change dynamically depending on the conditions (stress vs. non-stress).

If ADK and SnRK1 mutually stimulate each other, how is negative regulation accomplished? Phosphorylation of ADK by SnRK1 (Figure 2.5) might be involved in down-regulating ADK-SnRK1 complex, although we failed to see inhibition of ADK activity in vitro upon phosphorylation by SnRK1. This may not be surprising, given the fact that inactivation of some known SnRK1 and AMPK targets requires binding of 14-3-

3 proteins to the phosphorylated sites of the target protein (Bachmann et al., 1996a;

Moorhead et al., 1996; Sugden et al., 1999b; Ikeda et al., 2000; Huber et al., 2002; Gwinn et al., 2008). It would be interesting to test whether ADK can interact with any of the plant 14-3-3 proteins upon phosphorylation, and to test the consequences of such interactions, if found. Interestingly, 5'-AMP in maize has been shown to bind the 14-3-3 isoform GF14-6, and to disrupt its association with its target protein (Camoni et al.,

2001). The Arabidopsis 14-3-3 gene family has 13 members (named general regulatory

139 factor: GRF 1 through GRF 13) that are classified into  and non- groups (Sehnke et al.,

2002). Further investigation could determine if any of these 14-3-3 proteins are involved in the regulation of SnRK1-ADK complex, and if 5'-AMP is playing an additional regulatory role through 14-3-3 protein binding. It might also be worth investigating whether pleiotropic regulatory locus 1 (PRL1) and trehalose-6-phosphate (T6P), known negative regulators of SnRK1 (Bhalerao et al., 1999; Zhang et al., 2009), are involved in down-regulating the SnRK1-ADK complex. Phosphorylation of ADK appears to be conserved among different subfamilies of SnRK kinases, as demonstrated by in vitro and in vivo phosphorylation of ADK by SnRK2.8 (Shin et al., 2007). We studied SnRK1.2 that belongs to SnRK1 subfamily. Interestingly, SnRK2.8 also phosphorylates a 14-3-3 protein (Shin et al., 2007), although the consequences of this phosphorylation have yet to be determined.

5.1.2 Activation of ADK by SnRK1 is independent of its kinase activity: a novel role for SNF1-AMPK-SnRK1 family kinases

An interesting and a novel observation of this study is that SnRK1-KD and SnRK1-KD-

K49R (kinase-inactive mutant) expressed in N. benthamiana (HA2His6-tagged) or in E. coli (GST-tagged) were able to increase ADK activity (~ 4-7 fold) in vitro (Figure 2.6 A and B). This ADK stimulation does not require SnRK1 kinase activity. To our knowledge, such a role has not yet been reported for SNF1 or AMPK and suggests a novel mechanism of SnRK1 autostimulation. Stimulation of ADK by SnRK1 leads to an increase in 5'-AMP levels which in turn could lead to SnRK1 stimulation. Recent

140 evidence indicates that activation of AMPK by 5'-AMP requires the  subunit (Suter et al., 2006; Sanders et al., 2007; Hardie, 2007). This is likely also the case for SnRK1.

The novel role of SnRK1 in stimulating ADK, independent of its kinase activity, could be potentially useful in breeding crop plants for resistance against geminiviruses.

ADK has been shown to play critical roles in different antiviral defense responses. First,

ADK is required for host methylation of viral DNA and RNA silencing, two important defense mechanisms that inhibit viral multiplication (Wang et al., 2003; Wang et al.,

2005; Raja et al., 2008; Buchmann et al., 2009). Second, ADK also plays a role in cytokinin metabolism that has been shown to aid in antiviral defense (Baliji et al., 2010).

Third, ADK appears to be involved in the stimulating SnRK1 (Figure 2.7 A) which also plays a role in innate antiviral defense (Hao et al., 2003). Therefore, increasing ADK activity through SnRK1 could be exploited to increase the resistance of crop plants against viruses. Because SnRK1-KD-K49R (kinase-inactive mutant) is also capable of activating ADK, increased ADK activities could be achieved in a more controlled manner without introducing functional SnRK1, ectopic expression of which might lead to some undesirable phenotypes (Hao et al., 2003). It remains to be tested whether SnRK2 and

SnRK3 kinases, which are unique to plants, are also capable of stimulating ADK in a similar manner. At least one SnRK2 (SnRK2.8) has been shown to interact with ADK

(Shin et al., 2007), suggesting that this mechanism of ADK stimulation might be conserved in other SnRKs. Lastly, it would be interesting to investigate if such a mechanism also exists in yeast and animals, because like SnRK1, ADK is also conserved across eukaryotes (Moffat et al., 2002).

141

5.1.3 SnRK1-ADK-APT1, a putative complex that might play significant roles in responses to various kinds of stresses

Adenine phosphoribosyl transferase 1 (APT1) is one of the three functional APT proteins in Arabidopsis. Like ADK, it generates 5'-AMP although using different substrates (Allen et al., 2002). Therefore, we predicted that APT1 might have a similar relationship with

SnRK1. We obtained solid evidence in support of the prediction that SnRK1 and APT1 would interact, in vivo (Figures 3.1 and 3.2). Specifically, we have shown that SnRK1 and APT1 interact in a BiFC assay (Figure 3.1), and that SnRK1 phosphorylates APT1 in vitro (Figure 3.2). We speculate that SnRK1 also stimulates APT1 based on findings from SnRK1-ADK interaction (Chapter 2). However, this has yet to be investigated.

Surprisingly, we also observed that APT1 and ADK interact in BiFC experiments (Figure

3.4). We presume that the interaction between ADK and APT1 might be indirect because these proteins failed to copurify when either was overexpressed in N. benthamiana.

However, a protein with a mass similar to APT1 was found in immunocomplexes obtained from Arabidopsis using anti-ADK antibody (Figure 3.5). Interestingly, that protein is also phosphorylated in a kinase assay, presumably due to endogenous SnRK1 that is co-immunoprecipitated with ADK. In addition, endogenous SnRK1 activity was copurified when either ADK or APT1 was overexpressed (Figures 2.4 & 3.3). Based on these results, we speculate that SnRK1, ADK, and APT1 might form a ternary complex in vivo. However, further investigation will be required to test this hypothesis.

Our experiments with SnRK1, ADK, and APT1 have led us to pose some important questions. First, why is SnRK1 in complex with AMP-generating proteins?

Although 5'-AMP generated by ADK and APT1 could diffuse to reach SnRK1 to prevent

142 its deactivation by phosphatases, the presence of ADK and APT1 in a SnRK1 complex might rapidly increase the local 5'-AMP concentration which might lead to more robust

SnRK1 stimulation. Second, why are both ADK and APT1 present in a complex with

SnRK1, since both make 5'-AMP? That two different AMP-generating proteins might be in complex with SnRK1 strongly supports a role for 5'-AMP in SnRK1 stimulation in plants. As discussed above (section 5.1.1), 5'-AMP-mediated suppression of SnRK1 dephosphorylation could be a major mechanism for maintaining high SnRK1 activity.

Also, by using two different proteins that make 5'-AMP with different substrates, plants appear to ensure 5'-AMP production around SnRK1 even under conditions that might deplete the substrates required for one (ADK: adenosine and ATP; APT1: adenine and

PRPP). In fact, cells are likely to face a shortage of ATP as ATP depletion is a hallmark of the stresses that lead to the CSR. Third, could a putative SnRK1-ADK-APT complex play additional roles in the cellular stress response? Because ADK and APT1 can rapidly increase the AMP/ATP ratio by generating more AMP, this putative complex might play a role in sensing or amplifying a cellular stress signal. It would be interesting to further study the dynamics of this putative complex under stress conditions, to elucidate clearly its roles in sensing and responding to cellular stress. In support of our speculation, a wound-induced stress in potato has been shown to increase the expression (25-50 fold) as well as activities (2 fold) of both APT1 and ADK (Katahira and Ashihara, 2006). It would be significant to know if SnRK1 also affects the expression of APT1 and ADK under stress conditions, as it has already been shown to increase expression of other stress-responsive genes (Baena-Gonzalez et al., 2007).

143

5.2 SnRK1, ADK, and APT1 could be “High-Value” Targets for Geminiviruses and

Perhaps to All Plant Viruses

5.2.1 Different geminivirus proteins interact with SnRK1, ADK, and APT1

Geminivirus AL2 and L2 proteins are pathogenicity factors of begomoviruses (e.g.

TGMV, CaLcuV) and curtoviruses (e.g. BCTV), respectively (Sunter et al., 1990; Sunter and Bisaro, 1991; Hormuzdi and Bisaro, 1995). Both AL2 and L2 directly interact with and inactivate ADK and SnRK1 (Wang et al., 2003; Hao et al., 2003). We found that

AL2 and L2 also interact with APT1 in a BiFC experiment (Figure 4.1 A&B), and results from an in vitro GST pull-down assay suggest that the interaction of AL2 with APT1 might be direct (Figure 4.3). However, further studies are needed to confirm this, and also to determine the consequences of the interaction. We speculate that AL2 and L2 inactivate APT1 because it can be beneficial to viruses, as discussed below.

Interestingly, the subcellular localization patterns of AL2 complexes with SnRK1,

ADK, and APT1 are different. AL2 and SnRK1 on their own localize both to the nucleus and to the cytoplasm (Wang et al., 2003; Zhou et al., unpublished data), while ADK and

APT1 are cytoplasmic proteins (Moffat et al., 2000; Allen et al., 2002). AL2-APT1 and

AL2-ADK complexes accumulate in both the cytoplasm and the nucleus (Figure 4.1).

This suggests that AL2 might redirect ADK and APT1 proteins into the nucleus. By contrast, AL2 interacts with SnRK1 predominantly in the nucleus (Figure 4.1). What do these results mean? One possibility is that AL2 might disrupt interactions among SnRK1,

ADK, and APT1, and carry some of them to the nucleus (ADK and APT1). However, we

144 do not know the significance of ADK and APT1 relocalization. By contrast, AL2 may interact with SnRK1 in the nucleus directly, and complexes accumulate in ‗nuclear speckles‘ unlike AL2-ADK or AL2-APT1 complexes which appear to be uniformly distributed in the nucleus. We do not know the significance of the ‗nuclear speckles‘ formed by AL2 and SnRK1 but speculate, based on earlier findings, that they might be associated with transcriptional regulation of genes associated with stress responses

(Baena-Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008; Kleinow et al., 2009).

AL2 is a transcription activator and homologs from Mung bean yellow mosaic virus

(MYMV) and African cassava mosaic virus (ACMV), have been shown to activate more than 30 host genes (Sunter et al., 1990; Sunter and Bisaro 1991; Hormuzdi and Bisaro,

1995; Trinks et al., 2005). A recent study also demonstrated that transactivation of late viral genes by AL2 requires its binding to a plant-specific transcription factor called

Peapod 2 (Lacatus and Sunter, 2009). SnRK1 has been shown to integrate diverse energy and stress signals on a global scale by inducing extensive changes in the transcriptome that have the effect of promoting catabolism and inhibiting anabolic pathways (Baena-

Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008). In addition, SnRK1 has been shown to interact with ATAF1, a stress-responsive, NAC domain transcription factor in

Arabidopsis (Kleinow et al., 2009). Also, a role for yeast SNF1 in transcriptional regulation is well established. In particular, SNF1 also acts as a histone kinase, and in concert with the histone acetyltransferase Gcn5 and RNA polymerase II, it regulates transcription (Kuchin et al., 2000; Lo et al., 2001). In mammalian cells, nuclear speckles have been shown to be dynamic and often become active transcription sites (Lamond and

Spector, 2003). Therefore, it would be interesting to investigate if the ‗nuclear speckles‘

145 in which AL2-SnRK1 complexes accumulate are associated with transcriptional regulation.

Because AL2 and L2 are pathogenicity factors, studies were extended to two additional geminivirus pathogenicity factors, AV2 (precoat protein of ToLCNDV, a monopartite begomovirus) andC1. C1 is encoded by DNA , a satellite DNA that associates with monopartite begomoviruses such as Tomato yellow leaf curl China virus

(TYLCCNV) (Brough et al., 1988; Pascal et al., 1993; Brigneti et al., 1998; Saunders et al., 2000; Briddon and Stanley 2006). We tested whether interaction with SnRK1, ADK, and APT1, is a common feature of these pathogenicity factors. Using BiFC, it was observed that AV2 interacts with SnRK1, ADK, and APT1. C1 was found to interact with ADK and APT1 in BiFC experiments, and previous yeast-two-hybrid experiments indicate C1 and SnRK1 also interact (X. Zhou, unpublished data) (Figure 4.1). Further studies are needed to confirm these interactions through other methods, and to establish their consequences. All four viral proteins (AL2, L2, AV2, and C1) are pathogenicity factors and RNA silencing suppressors (Wang et al., 2005; Saunders et al., 2000; Briddon and Stanley 2006; Mubin et al., unpublished data). Inactivation of ADK and SnRK1 by

AL2 and L2 has been shown to be important for suppressing RNA silencing targeted against geminiviruses, and to combat host innate antiviral defense (Wang et al., 2003;

Hao et al., 2003; Wang et al., 2005;Raja et al., 2008; Buchmann et al., 2009). Therefore, we speculate that AV2 and C1 also inactivate one or all the members of SnRK1-ADK-

APT1 putative complex, but this remains to be tested. It might be relevant to note here that ToLCNDV and TYLCCNV have their own AL2 genes and similarly, TYLCCNV has its own AV2 gene. Do the proteins encoded by these genes perform similar functions

146 compared to their counterparts? We do not know the answer, and it remains to be investigated. However, according to the existing literature, it appears that each protein in geminiviruses is not necessarily evolved to perform the same function, and to the same extent if performing the same function. For example, AL2 (AC2) and AL4 (AC4) proteins have been shown to play differential roles in cassava geminiviruses (Vanitharani et al., 2004).

Based on our protein-protein interaction studies, we propose that the putative

SnRK1-ADK-APT1 complex might be central to host antiviral defense, and that many geminiviruses appear to have evolved proteins to counter this defense. It would be significant to know the consequences of these new interactions we discovered, and to identify common mechanisms of counter-defense that might have been conserved across geminiviruses. Discovering such mechanisms could contribute immensely to combating geminiviruses that are known to cause crop diseases resulting in significant yield losses.

Why might SnRK1, ADK, and APT1 be important to virus infection? The benefits viruses derive by inactivating ADK and SnRK1 are fairly well established, and these are discussed here. Our observation that ADK contributes to SnRK1 stimulation suggests one more antiviral role for ADK as SnRK1 is also involved in antiviral defense responses (Hao et al., 2003). ADK plays a significant role in the maintenance of the methyl cycle by phosphorylating adenosine, and this is crucial for host methylation of viral DNA as a defense strategy (Moffat et al., 2002; Raja et al., 2008; Buchmann et al.,

2009). We hypothesize that APT1 might also play a role in the maintenance of methyl cycle, as it contributes adenosine salvage through a different pathway (Figure 4.6).

Mainly, two pathways are involved in adenosine salvage. In one, adenosine is converted

147 to 5'-AMP by ADK while in the other, adenosine is first converted to adenine by adenosine nucleosidase, and adenine is subsequently converted to 5'-AMP by APT

(Sakakibara, 2006; Katahira and Ashihara, 2006). Interestingly, enzyme activities of

ADK, APT, and adenosine nucleosidase have been shown to increase significantly during wound-induced stress in potato (Katahira and Ashihara, 2006). Notably, adenosine nucleosidase activity increased by about 100 fold compared to a 2 fold increase in APT activity. Adenosine may therefore be readily converted to adenine, and subsequently salvaged by APT to 5'-AMP (Katahira and Ashihara, 2006). Similarly, ADK activity

(Wang et al., 2003) and likely SnRK1 activity (G. Sunter, unpublished data) also increase during viral infections (a biotic stress). Whether the activities of APT and adenosine nucleosidase also increase during viral infections remains to be determined.

ADK also plays a role in cytokinin metabolism. Cytokinins are phytohormones that play crucial roles in various phases of plant growth and development, including cell division (Sakakibara, 2006). These hormones are N6-substituted adenine derivates which exist both in bioactive (base and nucleoside) and bioinactive (nucleotide) forms (von

Schwartzenberg et al., 1998; Sakakibara, 2006). ADK converts the active nucleoside form to the inactive nucleotide form, and this has been shown to be detrimental to geminiviruses, likely because they use cytokinin-mediated cell cycle progression to enable their own DNA replication (Baliji et al., 2010). In fact, the mere exogenous application of cytokinins has been shown to render plants more susceptible for geminiviruses (Baliji et al., 2010). Because APT1 also converts the active (base) form of cytokinin to the inactive (nucleotide) form, it might also contribute to antiviral defense.

Thus, APT1 inactivation, similar to ADK inactivation, could be beneficial to

148 geminiviruses. SnRK1 stimulates ADK activity (and possibly APT1) and thus, could play a role in deterring geminivirus multiplication by indirectly increasing the inactive pool of cytokinins. Roles for SnRK1 in antiviral defense might also include blocking viral protein synthesis, as discussed in the following section.

5.2.2 One of the SnRK1 antiviral roles could be to block protein synthesis

In this study, we made an attempt to identify downstream targets or interacting partners of SnRK1 that could be relevant to virus infection. In animals, viral infections induce host interferon (IFN) responses which result, among other things, in the activation of

PKR (Gale and Katze, 1989). Activated PKR phosphorylates (at S51) and inactivates the eukaryotic translation initiation factor-2 (eIF-2), leading to decreased protein translation (Gale and Katze, 1989, for details on mechanisms of inactivation please refer

Chapter 1, section 1.5.1). This inhibits viral protein synthesis and reduces virus replication. Not surprisingly, most, if not all, successful animal viruses have developed a mechanism/s to combat PKR-mediated inactivation of translation initiation machinery

(Gale and Katze, 1998). However, the functional equivalent of PKR in plants remains to be identified, although a PKR-like activity has been reported (Hiddinga et al., 1988;

Crum et al., 1988; Langland et al., 1996). We know from our earlier work that SnRK1 plays an important role in innate antiviral defense not only against DNA viruses but also against RNA viruses, and that the geminivirus AL2 and L2 proteins interact with and inactivate SnRK1 (Hao et al 2003). Upon discovering interactions between SnRK1 and additional geminivirus proteins (Chapter 4), we speculated that SnRK1 might be playing a role similar to PKR in plants.

149

Manual searches of protein sequences of two Arabidopsis translation initiation factors, eIF-2 and eIF-(iso)4E, revealed three and one evolutionarily conserved potential SnRK1 sites, respectively (Tables 4.1 and 4.2). Using in vitro kinase assays, we found that SnRK1 phosphorylates both eIF-2 and eIF(iso)4E (Figure 4.4). What might be the relevance of these findings? We hypothesize that SnRK1 inhibits protein synthesis in response to viral infection by inactivating one or more translation initiation factors through phosphorylation. Clearly, additional studies are needed to test this hypothesis.

We describe below some of the needed experiments, and also discuss the potential impact of these findings.

SnRK1 phosphorylates eIF-2 in vitro, but we do not yet know the amino acids that are phosphorylated. However, we propose three sites (S56, S65, and T201) as potential SnRK1 targets based on their evolutionary conservation across diverse taxa of photosynthetic and non-photosynthetic eukaryotes (Table 4.1). The underlying premise is that conserved sites are likely to indicate physiologically relevant targets. This approach has been successfully employed to identify relevant targets of cAMP-dependent protein kinase (PKA) in yeast (Budovskaya et al., 2005). Site-directed mutagenesis can be employed to map the phosphorylation sites on eIF-2, and further confirmation can be obtained by subjecting endogenous eIF-2 to mass spectrophotometry.

Immunoprecipitation of endogenous eIF-2 appears to be a feasible option as demonstrated by previous studies (Zhang et al., 2008; Lageix et al., 2008). Evaluating the levels of eIF-2 phosphorylation under different levels of SnRK1 expression and during viral infection (with wild type BCTV and mutant virus lacking L2 protein that inactivates

SnRK1) would be crucial to evaluate our hypothesis. Interestingly, Arabidopsis GCN2

150 kinase, encoded by the single copy gene, has been shown to phosphorylate eIF-2 when plants were treated with herbicides that affect amino acid biosynthesis (Zhang et al.,

2008). However, eIF-2 was not phosphorylated when plants were infected with Turnip yellow mosaic virus (TYMV) or Turnip crinkle virus (TCV) (Zhang et al., 2008). There are several possible explanations for this result. First, TYMV and TCV might encode a protein that inhibits GCN2, and therefore these viruses might not have allowed GCN2- mediated eIF-2 phosphorylation. Second, at least in Arabidopsis, GCN2-mediated phosphorylation of eIF-2 may not act as a (defense) response to viral infections. Third, should GCN2-mediated eIF-2 phosphorylation may act as a defense response, but

TYMV and TCV may not elicit the response. Whatever might be the case, these results do not refute our hypothesis because the authors did not test phosphorylation at any of the potential SnRK1 sites (S56, S65, and T201) (Zhang et al., 2008).

Of the three conserved SnRK1 (and AMPK) sites noted above, two of them (S56 and T201) are also conserved across non-photosynthetic eukaryotes. Therefore, we speculate that AMPK might also phosphorylate eIF-2, presumably, in response to viral infection. In fact, some RNA viruses in animals inhibit AMPK to block various host defense responses including inhibition of protein synthesis (Kudchodkar et al., 2007;

Zhang and Wu, 2009; Mankouri et al., 2010). However, it is not yet known if AMPK- mediated phosphorylation of eIF-2 is responsible. Therefore, findings from further investigations could potentially be of broad significance in combating both plant and animal viruses.

151

We found that SnRK1 also phosphorylates eIF-(iso-4E) in vitro (Figure 4.4). The extent of conservation of SnRK1 site (T57) in eIF(iso)4E (across algae, mosses, and flowering plants) (Table 4.2) suggests that eIF-(iso)4E might be a relevant SnRK1 target.

The eIF-4E family in Arabidopsis comprises three eIF-4E proteins and one eIF-(iso)4E that is unique to plants (Browning, 2004; Robaglia and Caranta, 2006). All of these proteins contain the conserved SnRK1 site (Table 4.2). eIF-4E interacts with eIF-4G to form eIF-4F complex which is essential for translation initiation (Figure 1.5). In

Arabidopsis, experimentally proven interactions occur between eIF-4E1 and eIF-4G, and between eIF-(iso)4E and eIF-(iso)4G1 (Robaglia and Caranta, 2006; Figure 1.5).

Although eIF-4E proteins are also present in non-photosynthetic eukaryotes, this particular SnRK1 site (T57) is not conserved in them. Basically, eIF-4E protein binds 5'- cap of mRNA and this binding is essential for translation initiation. Geminiviruses depend on host translation machinery and because their mRNA also contains 5'-cap, we assume that cap binding by eIF-4E or eIF-(iso)4E also takes place. Such a binding is not yet shown for geminivirus mRNA, experimentally. We hypothesize that SnRK1 might disrupt eIF-4E/eIF-(iso)4E binding to viral mRNAs and thus inhibit their translation. In fact, in vitro phosphorylation of wheat eIF-(iso)4E by casein kinase II reduced its mRNA cap-binding affinity by 1.2 to 2.6 fold (Khan and Goss, 2004). Phosphorylation-mediated changes in mRNA cap-binding affinities of eIF-4E proteins are well documented in animals (Frederickson et al., 1991; Morley and Traugh, 1993; Pause et al., 1994).

However, further investigation is needed to confirm our hypothesis. A relevant experiment would be to test if SnRK1 phosphorylation reduces eIF-(iso)4E/eIF-4e mRNA binding affinity. Other studies will investigate the role of translation factor

152 phosphorylation in the slow growth as well as antiviral phenotypes of transgenic N. benthamiana plants constitutively expressing 35S-driven SnRK1 (Hao et al., 2003). In addition to or alternatively, other SnRK1-mediated mechanisms might be operating in plants to down-regulate protein synthesis, as found in yeast (Snf1) and animals (AMPK)

(Browne et al., 2004; Gwinn et al., 2008; Zhang et al., 2008; Cherkasova et al., 2010).

For example, AMPK phosphorylates Raptor to downregulate the mammalian target of rapamycin complex 1 (mTORC1)-mediated cell growth (Gwinn et al., 2008). mTORC1 modulates growth primarily by regulating protein synthesis (Wang and Proud, 2006;

Wullschleger et al., 2006). Overall, SnRK1 might regulate protein synthesis through phosphorylation of multiple targets. Therefore, geminiviruses could be valuable models to broaden our understanding of anti-viral defense responses, and presumably other stress responses mediated by SnRK1.

5.2.3 SnRK1 pathway plays significant role in antiviral defense and geminiviruses counteract this pathway

Earlier studies have shown that overexpressing SnRK1 can increase resistance to both

DNA and RNA virus infection in transgenic N. benthamiana plants (Hao et al., 2003).

However, the mechanisms involved in SnRK1-mediated antiviral defense have yet to be established. We propose that SnRK1 might mediate an antiviral defense response, at least in part, through down-regulation of protein synthesis (discussed in section 5.2.2), and through regulation of transcription of several stress-responsive genes (Baena-Gonzalez et al., 2007; Baena-Gonzalez and Sheen, 2008; Halford and Hey, 2009). SnRK1 kinases in

153 plants, akin to their counterparts in yeast and animals, are also activated through phosphorylation by upstream kinases. In Arabidopsis, Geminivirus Rep Interacting

Kinase 1 (GRIK 1) and GRIK 2 have been identified as SnRK1 upstream kinases (Shen et al., 2009). GRIK1 and GRIK2 are related to kinases that activate SNF1 in yeast, and

AMPK in animals (Shen and Hanley-Bowdoin, 2006). Genes encoding GRIK kinases have also been found in other plant species belonging to monocotyledonous and dicotyledonous plants (Shen and Hanley-Bowdoin, 2006). Geminivirus infections lead to upregulation of SnRK1 (G. Sunter, unpublished data) and GRIK kinases (Shen and

Hanley-Bowdoin, 2006). These findings emphasize the importance of SnRK1 pathway in antiviral defense, and geminiviruses appear to have evolved multiple strategies to block this pathway. First, geminivirus proteins AL2 and L2 inactivate SnRK1 (Hao et al.,

2003), and based on our studies, we speculate that inactivation of SnRK1 might be a widely conserved mechanism among many geminiviruses (Chapter 4). Second, geminivirus protein AL1 interacts with GRIK1 and GRIK2 (SnRK1 upstream kinases), which is believed to contribute to viral counter-defense, although this remains to be confirmed (Shen and Hanley-Bowdoin, 2006; Shen et al., 2009). Overall, this study provides many clues regarding SnRK1-mediated stress responses. More investigations are required to consolidate our findings, and to clearly understand SnRK1-mediated defense responses.

154

BIBLIOGRAPHY

Ach, R.A., Durfee, T., Miller, A.B., Taranto, P., Hanley-Bowdoin, L., Zambriski, P.C., and Gruissem, W. (1997). RRB1 and RRB2 encode maize retinoblastoma-related proteins that interact with a plant D-type cyclin and geminivirus replication protein. Mol. Cell. Biol. 17, 5077–5086. Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., and Walter, P. (2002). Molecular Biology of the Cell, Garland publishing 4th Edition. Allen, M., Qin, W., Moreau, F., and Moffatt, B. (2002). Adenine phosphoribosyltransferase isoforms of Arabidopsis and their potential contributions to adenine and cytokinin metabolism. Plant Physiol. 115(1), 56-68. Auer, C. (1999). The Arabidopsis mutation cym changes cytokinin metabolism, adenosine nucleosidase activity and plant phenotype. Biol Plant 42: S-3 Bachmann, M., Huber, J.L., Liao, P.C., Gage, D.A., and Huber, S.C. (1996a). The inhibitor protein of phosphorylated nitrate reductase from spinach (Spinacia oleracea) leaves is a 14-3-3 protein. FEBS Lett. 387(2-3), 127-31. Bachmann, M., Shiraishi, N., Campbell, W.H., Yoo, B.C., Harmon, A.C., and Huber, S.C. (1996b). Identification of Ser-543 as the major regulatory phosphorylation site in spinach leaf nitrate reductase. Plant Cell (3), 505-17. Baena-Gonzalez, E., and Sheen, J. (2008). Convergent energy and stress signaling. Trends Plant Sci. 13, 474-482. Baena-Gonzalez, E., Rolland, F., Thevelian, J.M., and Sheen, J. (2007). A central integrator of transcription networks in plant stress and energy signalling. Nature 448, 938-943. Baier, L. J., Shors, T., Shors, S. T. and Jacobs, B. L. (1993). The mouse antiphosphotyrosine immunoreactive kinase, TIK, is indistinguishable from the double- stranded RNA-dependent, interferon-induced protein kinase, PKR. Nucl. Acids Res. 21, 4830–4835. Baliji, S., Lacatus, G., and Sunter, G. (2010). The Interaction between geminivirus pathogenicity proteins and adenosine kinase leads to increased expression of primary cytokinin-responsive genes. Virology. [Epub ahead of print]

155

Baliji, S., Sunter, J., and Sunter, G. (2007). Transcriptional analysis of complementary sense genes in Spinach curly top virus and functional role of C2 in pathogenesis. Molecular Plant-Microbe Interactions 20, 194-206. Berne, R. M. (1993). Adenosine--a cardioprotective and therapeutic agent. Cardiovasc Res 27(1), 2. Bhalerao, R.P., Salchert, K., Bakó, L., Okrész, L., Szabados, L., Muranaka, T., Machida, Y., Schell, J., and Koncz, C. (1999). Regulatory interaction of PRL1 WD protein with Arabidopsis SNF1-like protein kinases. Proc Natl Acad Sci U S A. 96(9), 5322-7. Bilgin, D.D., Liu, Y., Schiff, M., and Dinesh-Kumar, S.P. (2003). P58(IPK), a plant ortholog of double-stranded RNA-dependent protein kinase PKR inhibitor, functions in viral pathogenesis. Dev Cell. 4(5), 651-61. Bisaro, D. M. (1996). Geminivirus replication. In "DNA Replication in Eukaryotic Cells" (M. DePamphilis, Ed.), pp. 833-854. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Bisaro, D. M. (2006). Silencing suppression by geminivirus proteins. Virology 344(1), 158-68. Bisaro, D.M., Hamilton, W.D., Coutts, R.H., and Buck, K.W. (1982) Molecular cloning and characterisation of the two DNA components of tomato golden mosaic virus. Nucleic Acids Res. 10(16), 4913-22. Briddon, R. W. and Stanley, J. (2006). Subviral agents associated withplant single- stranded DNA viruses. Virology 344, 198–210. Briddon, R.W., Bull, S.E., Amin, I., Idris, A.M., Mansoor, S., Bedford, I.D., Dhawan, P., Rishi, N., Siwatch, S.S., Abdel-Salam, A.M., Brown, J.K., Zafar, Y., and Markham, P.G. (2003). Diversity of DNA beta, a satellite molecule associated with some monopartite begomoviruses. Virology 312(1), 106-21. Briddon, R.W., Pinner, M.S., and Stanley, J. (1989). Geminivirus coat protein gene replacement alters insect specificity. Virology 177:85–94 Briddon RW, Watts J, Markham PG, Stanley J. The coat protein of beet curly top virus is essential for infectivity. Virology. 172(2), 628-33. Brigneti, G., Voinnet, O., Li, W.-X., Ji, L.-H., Ding, S.-W., and Baulcombe, D. C. (1998). Viral pathogenicity determinants are suppressors of transgene silencing in Nicotiana benthamiana. EMBO J. 17, 6739-6746. Brough, C.L., Hayes, R.J., and Coutts, R.H.A. (1988). Effect of mutagenesis in vitro on the ability of cloned tomato golden mosaic virus DNA to infect Nicotiana benthamiana plants. J Gen Virol 69, 481–492. Browne, G.J., Finn, S.G., and Proud, C.G. (2004). Stimulation of the AMP-activated protein kinase leads to activation of eukaryotic elongation factor 2 kinase and to its phosphorylation at a novel site, serine 398. J Biol Chem 279, 12220–12231.

156

Browning, K.S. (1996). The plant translational apparatus. Plant Mol Biol. 32(1-2):107- 44. Review. Browning, K.S. (2004). Plant translation initiation factors, it is not easy to be green. Biochemical Society Transactions 32, 589–591. Browning, K.S., Webster, C., Roberts, J.K., and Ravel, J.M. (1992). Identification of an isozyme form of protein synthesis initiation factor 4F in plants. J. Biol. Chem. 267, 10096–10100. Buchmann, R.C., Asad ,S., Wolf, J.N., Mohannath, G., and Bisaro, D.M. (2009). Geminivirus AL2 and L2 proteins suppress transcriptional gene silencing and cause genome-wide reductions in cytosine methylation. J Virol. 83(10), 5005-13. Budovskaya, Y.V., Stephan, J.S., Deminoff, S.J., and Herman, P.K. (2005). An evolutionary proteomics approach identifies substrates of the cAMP-dependent protein kinase. Proc Natl Acad Sci USA 102(39), 13933-8. Camoni, L, Visconti, S, Marra, M, Aducci, P. (2001). Adenosine 5'-monophosphate inhibits the association of 14-3-3 proteins with the plant plasma membrane H(+)-ATPase. J Biol Chem. 276(34):31709-12. Castillo, A.G., Collinet, D., Deret, S., Kashoggi, A., and Bejarano, E.R. (2003). Dual interaction of plant PCNA with geminivirus replication accessory protein (REn) and viral replication protein (Rep), Virology 312, 381-394. Celenza, J.L., and Carlson, M. (1989). Mutational analysis of the Saccharomyces cerevisiae SNF1 protein kinase and evidence for functional interaction with the SNF4 protein. Mol. Cell. Biol. 9, 5034–5044. Chakrabarty, R., Banerjee, R., Chung, S.M., Farman, M., Citovsky, V., Hogenhout, S.A., Tzfira, T., and 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(7), 740-50. Chang, C. H., Cha, S., Brockman, R. W., and Bennett, L. L., Jr. (1983). Kinetic studies of adenosine kinase from L1210 cells: a model enzyme with a two-site ping-pong mechanism. Biochemistry 22(3), 600-11. Chang, L.Y., Yang, W.Y., Browning, K., and Roth, D. (1999). Specific in vitro phosphorylation of plant eIF2alpha by eukaryotic eIF2alpha kinases. Plant Mol Biol. 41(3), 363-70. Chellappan, P., Vanitharani, R., and Fauquet, C. M. (2005). MicroRNA-binding viral protein interferes with Arabidopsis development. Proc. Natl. Acad. Sci. USA 102, 10381- 10386. Chen, J., Pal, J., Throop, M. S., Gehrke, L., Kuo, I., Pal, J. K., Brodsky, M., and London, I. M. (1991). Cloning of the cDNA of the heme-regulated eukaryotic initiation factor 2a (eIF-2a) kinase of rabbit reticulocytes: homology to yeast GCN2 protein kinase and

157 human double-stranded-RNA-dependent eIF-2a kinase. Proc. Natl. Acad. Sci. USA 88, 7729–7733. Cherkasova, V., Qiu, H., and Hinnebusch, A.G. (2010). Snf1 promotes phosphorylation of the alpha subunit of eukaryotic translation initiation factor 2 by activating Gcn2 and inhibiting phosphatases Glc7 and Sit4. Mol Cell Biol. 30(12), 2862-73. Choudhury, N.R., Malik, P.S., Singh, D.K., Islam, M.N., Kaliappan, K., and Mukherjee, S.K. (2006). The oligomeric Rep protein of Mungbean yellow mosaic India virus (MYMIV) is a likely replicative helicase. Nucl. Acids Res. 34, 6362-6377. Clemens, M. J. and Elia, A. (1997). The double-stranded RNAdependent protein kinase PKR—structure and function. J. Interferon Cytokine Res. 17(9), 503-24. Clemens, M. J., Laing, K. G., Jeffrey, I. W., Schofield, A., Sharp, T. V., Elia, A., Matys, V., James, M. C., and Tilleray, V. J. (1994). Regulation of the interferon-inducible eIF-2 alpha protein kinase by small RNAs. Biochimie 76, 770–778. Clemens, M.J., Hershey, J.W., Hovanessian, A.C., Jacobs, B.C., Katze, M.G., Kaufman, R.J., Lengyel, P., Samuel, C.E., Sen, G.C., and Williams, B.R. (1993). PKR: proposed nomenclature for the RNA-dependent protein kinase induced by interferon. J Interferon Res. 13(3), 241. Clerot, D. and Bernardi, F. (2006). DNA helicase activity is associated with the replication initiator protein Rep of Tomato yellow leaf curl geminivirus. J. Virol. 80, 11322-11330. Crum, C.J., Hu, J., Hiddinga, H.J., and Roth, D.A. (1988). Tobacco mosaic virus infection stimulates the phosphorylation of a plant protein associated with double- stranded RNA-dependent protein kinase activity. Journal of Biological Chemistry 263, 13440–13443. Cui, X., Tao, X., Xie, Y., Fauquet, C.M., and Zhou, X. A. (2004). DNA beta associated with Tomato yellow leaf curl China virus is required for symptom induction. J Virol. 78(24), 13966-74. Dale, S., Arró M., Becerra, B., Morrice, N.G., Boronat, A., Hardie, D.G., and Ferrer, A. (1995a). Bacterial expression of the catalytic domain of 3-hydroxy-3-methylglutaryl-CoA reductase (isoform HMGR1) from Arabidopsis thaliana, and its inactivation by phosphorylation at Ser577 by Brassica oleracea 3-hydroxy-3-methylglutaryl-CoA reductase kinase. Eur J Biochem. 233(2), 506-13. Dale, S., Wilson, W.A., Edelman, A.M., Hardie, D.G. (1995b). Similar substrate recognition motifs for mammalian AMP-activated protein kinase, higher plant HMG- CoA reductase kinase-A, yeast SNF1, and mammalian calmodulin-dependent protein kinase I. FEBS Lett. 361(2-3), 191-5. Davies, S.P., Carling, D., and Hardie, D.G. (1989). Tissue distribution of the AMP- activated protein kinase, and lack of activation by cyclic-AMP-dependent protein

158 kinase,studied using a specific and sensitive peptide assay. Eur J Biochem. 186(1-2), 123- 8. Davies, S.P., Helps, N.R., Cohen, P.T., and Hardie, D.G. (1995). 5'-AMP inhibits dephosphorylation, as well as promoting phosphorylation, of the AMP-activated protein kinase. Studies using bacterially expressed human protein phosphatase-2C alpha and native bovine protein phosphatase-2AC. FEBS Lett. 377(3), 421-5. Dever, T. E., Feng, L., Wek, R. C., Cigan, A. M., Donahue, T. F. and Hinnebusch, A. G. (1992). Phosphorylation of initiation factor 2 alpha by protein kinase GCN2 mediates gene specific translational control of GCN4 in yeast. Cell 68, 585–596. Dong, X., van Wezel R., Stanley, J., and Hong, Y. (2003). Functional characterization of the nuclear localization signal for a suppressor of posttranscriptional gene silencing. J. Virol. 77, 7026-7033. Douglas, P., Morrice, N., and MacKintosh, C. (1995). Identification of a regulatory phosphorylation site in the hinge 1 region of nitrate reductase from spinach (Spinacea oleracea) leaves. FEBS Lett. 377(2), 113-7. Douglas, P., Pigaglio, E., Ferrer, A., Halfords, N.G., and MacKintosh, C. (1997). Three spinach leaf nitrate reductase-3-hydroxy-3-methylglutaryl-CoA reductase kinases that are required by reversible phosphorylation and/or Ca2+ ions. Biochem J.1;325 ( Pt 1), 101-9. Duprat, A., Caranta, C., Revers, F., Menand, B., Browning, K.S., and Robaglia, C. (2002). The Arabidopsis eukaryotic initiation factor (iso)4E is dispensable for plant growth but required for susceptibility to potyviruses. Plant J. 32, 927–934. Durfee, T., Becherer, K., Chen, P.L., Yeh, S.H., Yang, Y., Kilburn, A.E., Lee, W.H., and Elledge, S.J. (1993). The retinoblastoma protein associates with the protein phosphatase type 1 catalytic subunit. Genes Dev. 7(4), 555-69. Egelkrout, E.M., Robertson, D., and Hanley-Bowdoin, L. (2001). Proliferating cell nuclear antigen transcription is repressed through an E2F consensus element and activated by geminivirus infection in mature leaves. Plant Cell 13, 1437–1452. Fauquet, C. M., Bisaro, D. M., Briddon, R. W., Brown, J. K., Harrison, B. D., Rybicki, E. P., Stenger, D. C., and Stanley, J. (2003). Revision of taxonomic criteria for species demarcation in the family Geminiviridae, and an updated list of begomovirus species. Arch Virol 148(2), 405-21. Fontes, E.P.B., Eagle, P.A., Sipe, P.S., Luckow, V.A., and Hanley-Bowdoin, L. (1994). Interaction between a geminivirus replication protein and origin DNA is essential for viral replication, J. Biol. Chem. 269, 8459-8465. Fox, H. I., and Kelley, N. W. (1978). The role of adenosine and 2'-deoxyadenosine in mammalian cells. Ann. Rev. Biochem. 47, 655-686. Fragoso, S., Espíndola, L., Páez-Valencia, J., Gamboa, A., Camacho, Y., Martínez- Barajas, E., and Coello, P. (2009). SnRK1 isoforms AKIN10 and AKIN11 are

159 differentially regulated in Arabidopsis plants under phosphate starvation. Plant Physiol. 149(4), 1906-16. Fraser, R.S.S. (1986). Genes for resistance to plant viruses. CRC Crit.Rev. Plant Sci. 3, 257–294. Fraser, C. S., Pain, V. W., and Morley, S. J. (1999). The association of initiation factor 4F with poly(A)-binding protein is enhanced in serum-stimulated Xenopus kidney cells, J. Biol. Chem. 274, 196-204. Frederickson, R. M., Montine, K. S., and Sonenberg, N. (1991). Phosphorylation of eukaryotic translation initiation factor 4E is increased in Src-transformed cell lines, Mol. Cell. Biol. 11, 2896-2900. Gaillard, C., Moffatt, B.A., Blacker, M., and Laloue, M. (1998). Male sterility associated with APRT deficiency in Arabidopsis thaliana results from a mutation in the gene APT1. Mol Gen Genet. 257(3), 348-53. Gale, M. and Katze, M.G. (1998). Molecular Mechanisms of Interferon Resistance Mediated by Viral-Directed Inhibition of PKR, the Interferon-Induced Protein Kinase. Pharmacol. Ther. 78, 29–46. Gao, Z., Johansen, E., Eyers, S., Thomas, C.L., Noel Ellis, T.H., and Maule, A.J. (2004). The potyvirus recessive resistance gene, sbm1, identifies a novel role for translation initiation factor eIF4E in cell-tocell trafficking. Plant J. 40, 376–385 Gardiner, W. E., Sunter, G., Brand, L., Elmer, J. S., Rogers, S. G., and Bisaro, D. M. (1988). Genetic analysis of tomato golden mosaic virus: The coat protein is not required for systemic spread or symptom development. EMBO J. 7(4), 899-904. Goldman, N., Chen, M., Fujita, T., Xu, Q., Peng, W., Liu, W., Jensen, T.K., Pei, Y., Wang, F., Han, X., Chen, J.F., Schnermann, J., Takano, T., Bekar, L., Tieu, K., and Nedergaard, M. (2010). Adenosine A1 receptors mediate local anti-nociceptive effects of acupuncture. Nat Neurosci. [Epub ahead of print] Green, S. R., and Mathews, M. B. (1992). Two RNA bindingmotifs in the double- stranded RNA activated protein kinase, DAI. Genes Dev. 6, 2478–2490. Gutierrez, C. (1999). Geminivirus DNA replication. Cell Mol Life Sci 56(3-4), 313-29. Gwinn, D.M., Shackelford, D.B., Egan, D.F., Mihaylova, M.M., Mery, A., Vasquez, D.S., Turk, B.E., and Shaw, R.J. (2008). AMPK phosphorylation of raptor mediates a metabolic checkpoint. Mol Cell. 30(2), 214-26. Halford, N., and Hey, S.J. (2009). Snf1-related protein kinases (SnRKs) act within an intricate network that links metabolic and stress signalling in plants. Biochem. J. 419, 247-259. Halford, N.G., and Hardie, D.G. (1998). SNF1-related protein kinases: global regulators of carbon metabolism in plants? Plant Mol Biol. 37(5), 735-48.

160

Halford, N.G., Hey, S., Jhurreea, D., Laurie, S., McKibbin, R.S., Paul, M., and Zhang, Y. (2003). Metabolic signalling and carbon partitioning: role of Snf1-related (SnRK1) protein kinase. J Exp Bot. 54(382), 467-75. Hamilton, W.D., Bisaro, D.M., Coutts, R.H., and Buck, K.W. (1983). Demonstration of the bipartite nature of the genome of a single-stranded DNA plant virus by infection with the cloned DNA components. Nucleic Acids Res. 11(21),7387-96. Hanley-Bowdoin, L., Settlage, S. B., Orozco, B. M., Nagar, S., and Robertson, D. (2000). Geminiviruses: models for plant DNA replication, transcription, and cell cycle regulation. Crit Rev Biochem Mol Biol 35(2), 105-40. Hao, L., Wang, H., Sunter, G., and Bisaro, D.M. (2003). Geminivirus AL2 and L2 proteins interact with and inactivate SNF1 kinase. Plant Cell 15, 1034-1048. Hardie, D.G. (2007). AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy. Nature Rev. 8, 774-785. Hardie, D.G., Carling, D., and Carlson, M. (1998). The AMP-activated/SNF1 protein kinase subfamily: metabolic sensors of the eukaryotic cell? Annu Rev Biochem. 67, 821- 55. Harper, J.W., Adami, G.R., Wei, N., Keyomarsi, K., and. Elledge, S. J. (1993). The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinase. Cell 75, 805-816. Harthill, J.E., Meek, S.E., Morrice, N., Peggie, M.W., Borch, J., Wong, B.H., and Mackintosh, C. (2006). Phosphorylation and 14-3-3 binding of Arabidopsis trehalose- phosphate synthase 5 in response to 2-deoxyglucose. Plant J. 47(2), 211-23. Hartitz, M. D., Sunter, G., and Bisaro, D. M. (1999). The tomato golden mosaic virus transactivator (TrAP) is a single-stranded DNA and zinc-binding phosphoprotein with an acidic activation domain. Virology 263(1), 1-14. Hartitz, M.D., Sunter, G., and Bisaro, D.M. (1999). The tomato golden mosaic virus transactivator(TrAP) is a single-stranded DNA and zinc-binding phosphoprotein with an acidic activation domain. Virology 263(1), 1-14. Hawley, S.A., Davison, M., Woods, A., Davies, S.P., Beri, R.K., Carling, D., and Hardie, D.G. (1996). Characterization of the AMP-activated protein kinase kinase from rat liver and identification of threonine 172 as the major site at which it phosphorylates AMP activated protein kinase. J. Biol. Chem. 271, 27879-27887. Hawley, S.A., Pan, D.A., Mustard, K.J., Ross, L., Bain, J., Edelman, A.M., Frenguelli, B.G., and Hardie, D.G. (2005). Calmodulin-dependent protein kinase kinase-b is an alternative upstream kinase for AMP-activated protein kinase. Cell Metab. 2, 9-19. Hiddinga, H.J., Crum, C.J., and Roth, D.A. (1988). Viroid-induced phosphorylation of a host protein related to a dsRNA-dependent protein kinase. Science 241, 451–453.

161

Hinnebusch, A. G. (1994). The eIF-2a kinases: regulators of protein synthesis in starvation and stress. Semin. Cell Biol. 5, 417–426. Höhnle, M., Höfer, P., Bedford, I.D., Briddon, R.W., Markham, P.G., and Frischmuth T. (2001) Exchange of three amino acids in the coat protein results in efficient whitefly transmission of a nontransmissible Abutilon mosaic virus isolate. Virology 290, 164–171. Hormuzdi, S. G., and Bisaro, D. M. (1993). Genetic analysis of beet curly top virus: evidence for three virion sense genes involved in movement and regulation of single- and double-stranded DNA levels. Virology 193(2), 900-9. Hormuzdi, S. G., and Bisaro, D. M. (1995). Genetic analysis of beet curly top virus: Examination of the roles of L2 and L3 genes in viral pathogenesis. Virology 206, 1044- 1054. Hrabak, E.M., Chan, C.W., Gribskov, M., Harper, J.F., Choi, J.H., Halford, N., Kudla, J., Luan. S., Nimmo, H.G., Sussman, M.R., Thomas, M., Walker-Simmons, K,, Zhu, J.K., and Harmon, A.C. (2003). The Arabidopsis CDPK-SnRK superfamily of protein kinases. Plant Physiol. 132(2), 666-80. Hu, C.D., Chinenov, Y., and. Kerppola, T. K. (2002). Visualization of interactions among bZIP and Rel family proteins using bimolecular fluorescence complementation. Mol. Cell 9, 789-798. Huber, S.C., MacKintosh, C., and Kaiser, W.M. (2002). Metabolic enzymes as targets for 14-3-3 proteins. Plant Mol Biol. 50(6), 1053-63. Review. Hull, R. (2002). Matthews‘ plant virology, 4th edn. Academic Press, San Diego. Ikeda, Y., Koizumi, N., Kusano, T., and Sano, H. (2000). Specific binding of a 14-3-3 protein to autophosphorylated WPK4, an SNF1-related wheat protein kinase, and to WPK4-phosphorylated nitrate reductase. J Biol Chem. 275(41), 31695-700. Jaramillo, M. L., Abraham, N. and Bell, J. C. (1995). The interferon system: a review with emphasis on the role of PKR in growth control. Cancer Invest. 13, 327–338. Jeske, H. (2009). Geminiviruses, Curr. Top. Immunol. Microbiol. 331, 185-226. Jiang, R., and Carlson, M. (1997). The Snf1 protein kinase and its activating subunit, Snf4, interact with distinct domains of the Sip1/Sip2/Gal83 component in the kinase complex. Mol Cell Biol 17(4), 2099-106. Jossier, M., Bouly, J.-P., Meimoun, P., Arjmand, A., Lessard, P., Hawley, S., Hardie, D.G., and Thomas, M. (2009). SnRK1 (SNF1-related kinase 1) has a central role in sugar and ABA signalling in Arabidopsis thaliana. Plant J. 59, 316-328. Katahira, R., and Ashihara, H. (2006). Role of adenosine salvage in wound-induced adenylate biosynthesis in potato tuber slices. Plant Physiol Biochem. 44(10), 551-5.

162

Khan, M.A., Goss, D.J. (2004). Phosphorylation states of translational initiation factors affect mRNA cap binding in wheat. Biochemistry. 43(28), 9092-7. Kim, W.Y., Fujiwara, S., Suh, S.S., Kim, J., Kim, Y., Han, L., David, K., Putterill, J., Nam, H.G., and Somers, D.E. (2007). ZEITLUPE is a circadian photoreceptor stabilized by GIGANTEA in blue light. Nature. 449(7160), 356-60. Kimball, S.R. (1999). Eukaryotic initiation factor eIF2. Int J Biochem Cell Biol. Jan;31(1), 25-9. Review. Kleinow, T,. Himbert, S., Krenz, B., Jeske, H., and Koncz, C. (2009). NAC domain transcription factor ATAF1 interacts with SNF1-related kinases and silencing of its subfamily causes severe developmental defects in Arabidopsis. Plant Science. 177, 36- 370. Kong, L.J., Orozco, B.M., Roe, J.L., Nagar, S., Ou, S., Feiler, H.S., Durfee, T., Miller, A.B., Gruissem, W., Robertson, D., and Hanley-Bowdoin, L. (2000). A geminivirus replication protein interacts with retinoblastoma through a novel domain to determine symptoms and tissue-specificity of infection in plants. EMBO J. 19, 3485–3495. Kowaluk, E. A., Bhagwat, S. S., and Jarvis, M. F. (1998). Adenosine kinase inhibitors. Curr Pharm Des 4(5), 403-16. Kuchin, S., Treich, I., and Carlson, M. A. (2000). Regulatory shortcut between the Snf1 protein kinase and RNA polymerase II holoenzyme. Proc Natl Acad Sci U S A. 97(14), 7916-20. Kudchodkar, S.B., Del Prete, G.Q., Maguire, T.G., and Alwine, J.C. (2007). AMPK- mediated inhibition of mTOR kinase is circumvented during immediate-early times of human cytomegalovirus infection. J. Virol. 81, 3649-3651. Kumar, S.H., and Rangarajan, A. (2009). Simian virus 40 small T antigen activates AMPK and triggers autophagy to protect cancer cells from nutrient deprivation. J. Virol. 83, 8565-8574. Lacatus, G., and Sunter, G. (2008). Functional analysis of bipartite begomovirus coat protein sequences, Virology 376, 79-89. Lacatus, G., and Sunter, G. (2009). The Arabidopsis PEAPOD2 transcription factor interacts with geminivirus AL2 protein and the coat protein promoter. Virology. 392(2), 196-202. Lageix, S., Lanet, E., Pouch-Pélissier, M.N., Espagnol, M.C., Robaglia, C., Deragon, J.M., and Pélissier, T. (2008). Arabidopsis eIF2alpha kinase GCN2 is essential for growth in stress conditions and is activated by wounding. BMC Plant Biol. 8, 134. Lamond, A I., Spector, D.L. (2003). Nuclear speckles: a model for nuclear organelles. Nat Rev Mol Cell Biol. 4(8), 605-12.

163

Langland, J.O., Langland, L.A., Browning, K.S., and Roth, D.A. (1996). Phosphorylation of plant eukaryotic initiation factor-2 by theplant encoded double-stranded RNA- dependent protein kinase, pPKR, and inhibition of protein synthesis in vitro. Journal of Biological Chemistry 271, 4539–4544. Langland, J.O., Song, J. Jacobs, B., and Roth, D.A. (1995). Identification of a plant- encoded analog of PKR, the mammalian double-stranded RNA-dependent protein kinase. Plant Physiology 108, 1259–1267. Laufs, J., Traut, W., Heyraud, F., Matzeit, V.Rogers, S.G., Schell, J., and Gronenborn, B. (1995). In vitro cleavage and ligation at the viral origin of replication by the replication protein of tomato yellow leaf curl virus. Proc. Natl. Acad. Sci. USA. 92, 3879-3883. Laurie, S., McKibbin, R.S., and Halford, N.G. (2003). Antisense SNF1-related (SnRK1) protein kinase gene represses transient activity of an alpha-amylase (alpha-Amy2) gene promoter in cultured wheat embryos. J Exp Bot. 54(383), 739-47. Lazarowitz, S. G. (1987). The Molecular Characterization of Geminiviruses. Plant Molecular Biology Reporter 4(4), 177-192. Lecoq, K., Belloc, I., Desgranges, C., and Daignan-Fornier, B. (2001). Role of adenosine kinase in Saccharomyces cerevisiae: identification of the ADO1 gene and study of the mutant phenotypes. Yeast 18, 335-342. Lellis, A.D., Kasschau, K.D., Whitham, S.A., and Carrington, J.C. (2002). Loss-of- susceptibility mutants of Arabidopsis thaliana reveal an essential role for eIF(iso)4E during potyvirus infection. Curr. Biol. 12, 1046–1051. Leonard, S., Plante, D., Wittmann, S., Daigneault, N., Fortin, M. G., Laliberte, J.-F. (2000). Complex Formation between Potyvirus VPg and Translation Eukaryotic Initiation Factor 4E Correlates with Virus Infectivity. J. Virol. 74, 7730-7737. Lindbo, J.A. (2007a). TRBO: a high-efficiency tobacco mosaic virus RNA-Based overexpression vector. Plant Physiol 145, 1232–1240. Lindbo, J.A. (2007b). High-efficiency protein expression in plants from agroinfection- compatible Tobacco mosaic virus expression vectors. BMC Biotechnology 7:52. Liu, H., Boulton, M.I., and Davies, J.W. (1997). Maize streak virus coat protein binds single- and double-stranded DNA in vitro. J Gen Virol. 78 ( Pt 6), 1265-70. Lo, W.S., Duggan, L., Emre, N.C., Belotserkovskya, R., Lane, W.S., Shiekhattar, R., and Berger, S.L. (2001). Snf1--a histone kinase that works in concert with the histone acetyltransferase Gcn5 to regulate transcription. Science 293(5532), 1142-6. Luque, A., Sanz-Burgos, A.P., Ramirez-Parra, E., Castellano, M.M., and Gutierrez, C. (2002). Interaction of geminivirus Rep protein with replication factor C and its potential role during geminivirus replication. Virology 302, 83-94.

164

Man, A.L., Purcell, P.C., Hannappel, U., and Halford, N.G. (1997). Potato SNF1-related protein kinase: molecular cloning, expression analysis and peptide kinase activity measurements. Plant Mol Biol. 34(1), 31-43. Mankouri, J., Tedbury, P.R., Gretton, S., Hughes, M.E., Griffen, S.D.C., Dallas, M.L., Green, K.A., Hardie, D.G., Peers, C., and Harris, M. (2010). Enhanced hepatitis C virus genome replication and lipid accumulation mediated by inhibition of AMP-activated protein kinase. Proc. Natl. Acad. Sci. USA 107, 11549-11554. Mansoor, S., Briddon, R.W., Zafar, Y., and Stanley, J. (2003) Geminivirus disease complexes: an emerging threat. Trends Plant Sci 8, 128–134. Mansoor, S., Zafar, Y., and Briddon, R. W. (2006). Geminivirus disease complexes: the threat is spreading. Trends Plant Sci 11, 209–212. Martin, B. J., McClanahan, T. B., Van Wylen, D. G., and Gallagher, K. P. (1997). Effects of ischemia, preconditioning, and adenosine deaminase inhibition on interstitial adenosine levels and infarct size. Basic Res Cardiol 92(4), 240-51. McCormack, S. J., and Samuel, C. E. (1995). Mechanism of interferon action: RNA- binding activity of full-length and R-domain forms of the RNA-dependent protein kinase PKR—determination of KD values for VAl and TAR RNAs. Virology 206, 511–519. McGaraughty, S., Cowart, M., and Jarvis, M. F. (2001). Recent developments in the discovery of novel adenosine kinase inhibitors: mechanism of action and therapeutic potential. CNS Drug Rev 7(4), 415-32. McMichael, R.W. Jr, Bachmann, M., and Huber, S.C. (1995). Spinach Leaf Sucrose- Phosphate Synthase and Nitrate Reductase Are Phosphorylated/Inactivated by Multiple Protein Kinases in Vitro. Plant Physiol. 108(3),1077-1082. McNellis, T.W., Mudgett, M.B., Li, K., Aoyama, T., Horvath, D., Chua, N.-H., and Staskawicz, B.J. (1998). Glucocorticoid-inducible expression of bacterial avirulence gene in transgenic Arabidopsis induced hypersensitive cell death. Plant J. 14, 247-257. Mellits, K. H., Kostura, M. and Mathews, M. B. (1990). Interaction of adenovirus VA RNA1 with the protein kinase DAI: nonequivalence of binding and function. Cell 61, 843–852. Moffat, A. (1999). Geminiviruses emerge as serious crop threat. Science 286, 1835. Moffatt, B.A., Stevens, Y.Y., Allen, M.S., Snider, J.D., Pereira, L.A., Todorova, M.I., Summers, P.S., Weretilnyk, E.A., Martin-McCaffrey, L., and Wagner, C. (2002). Adenosine kinase deficiency is associated with developmental abnormalities and reduced transmethylation. Plant Physiol. 128(3), 812-21. Moffatt, B.A., Wang, L., Allen, M.S., Stevens, Y.Y., Qin, W., Snider, J., von Schwartzenberg, K. (2000). Adenosine kinase of Arabidopsis. Kinetic properties and gene expression. Plant Physiol. 124(4), 1775-85.

165

Moorhead, G,. Douglas, P., Morrice, N., Scarabel, M., Aitken, A., and MacKintosh, C. (1996). Phosphorylated nitrate reductase from spinach leaves is inhibited by 14-3-3 proteins and activated by fusicoccin. Curr Biol. 6(9), 1104-13. Morley, S. J., and Traugh, J. A. (1993) Stimulation of translation in 3T3-L1 cells in response to insulin and phorbol ester is directly correlated with increased phosphate labelling of initiation factor (eIF-) 4F and ribosomal protein S6, Biochimie (Paris) 75, 985-989. Mubin, M., Amin, I., Amrao, L., Briddon, R.W., and Mansoor, S. (2010). The hypersensitive response induced by the V2 protein of a monopartite begomovirus is countered by the C2 protein. Mol Plant Pathol. 11(2), 245-54. Muller, U., Steinhoff, U., Reis, L. F. L., Hemmi, S., Pavlovic, J., Zinkernagel, R. M. and Aguet, M. (1994). Functional role of type I and type II interferons in antiviral defense. Science 264, 1918–1924. Nagar, S., Pedersen, T.J., Carrick, K., Hanley-Bowdoin, L. and Robertson, D. (1995). A geminivirus induces expression of a host DNA synthesis protein in terminally differentiated plant cells. Plant Cell, 7, 705–719. Nakazawa, Y., Hiraguri, A., Moriyama, H., and Fukuhara, T. (2007). The dsRNA- binding protein DRB4 interacts with the Dicer-like protein DCL4 in vivo and functions in the trans-acting siRNA pathway. Plant Mol. Biol. 63, 777-785. Nicaise, V., German-Retana, S., Sanjuán, R., Dubrana, M.P., Mazier, M., Maisonneuve, B., Candresse, T., Caranta, C., and LeGall, O. (2003). The eukaryotic translation initiation factor 4E controls lettuce susceptibility to the Potyvirus Lettuce mosaic virus. Plant Physiol. 132, 1272–1282. Palmer, K.E., and Rybicki, E.P. (1998). The molecular biology of mastreviruses. Adv Virus Res. 50, 183-234. Review. Pascal, E., Goodlove, P.E., Wu, L.C., and Lazarowitz, S.G. (1993). Transgenic tobacco plants expressing the geminivirus BL1 protein exhibit symptoms of viral disease. Plant Cell. 5(7), 795-807. Pause, A., Belsham, G. J., Gingras, A. C., Donze, O., Lin, T. A., Lawrence, J., and Sonenberg, N. (1994). Insulin-dependent stimulation of protein synthesis by -cap function. Nature 371, 762-767. Pilartz, M., and Jeske, H. (1992). Abutilon mosaic virus double-stranded DNA is packed into minichromosomes. Virology 189, 800–802. Pilartz, M., and Jeske, H. (2003). Mapping of Abutilon mosaic geminivirus minichromosomes. J. Virol. 77, 10808–10818. Piroux, N., Saunders, K., Page, A., and Stanley, J. (2007). Geminivirus pathogenicity protein C4 interacts with Arabidopsis thaliana shaggy-related protein kinase AtSKeta, a component of the brassinosteroid signalling pathway. Virology 362(2), 428-40.

166

Piroux, N., Saunders, K., Page, A., and Stanley, J. (2007). Geminivirus pathogenicity protein C4 interacts with Arabidopsis thaliana shaggy-related protein kinase AtSKeta, a component of the brassinosteroid signalling pathway. Virology. 362(2), 428-40. Polge, C., and Thomas, M. (2007). SNF1/AMPK/SnRK1 kinases, global regulators at the heart of energy control? Trends Plant Sci.12(1), 20-8. Pringle, C.R. (1999a). Virus nomenclature. Arch Virol. 144(7), 1463-6. Pringle, C.R. (1999b). Virus taxonomy—1999. The universal system of virus taxonomy, updated to include the new proposals ratified by the International Committee on Taxonomy of Viruses during 1998. Arch Virol. 144(2), 421-9. Erratum in: Arch Virol 1999;144(8):1667. Purcell, P.C., Smith, A.M., and Halford, M.G. (1998). Antisense expression of a sucrose non-fermenting 1 related protein kinase sequence in potato results in decreased expression of sucrose synthase in tubers and loss of sucrose-inducibility of sucrose synthase transcripts in leaves. Plant J 14, 195–202. Qin, S., Ward, B. M., and Lazarowitz, S. G. (1998). The bipartite geminivirus coat protein aids BR1 function in viral movement by affecting the accumulation of viral single-stranded DNA. Journal of Virology 72, 9247-9256. Radchuk, R., Radchuk, V., Weschke, W., Borisjuk, L.,and Weber, H. (2006). Repressing the expression of the SUCROSE NONFERMENTING-1-RELATED PROTEIN KINASE gene in pea embryo causes pleiotropic defects of maturation similar to an abscisic acid-insensitive phenotype. Plant Physiol. 140(1), 263-78. Raja, P,, Sanville, B.C., Buchmann, R.C., and Bisaro, D.M. (2008). Viral genome methylation as an epigenetic defense against geminiviruses. J Virol. 82(18), 8997-9007. Raja, P., Wolf, J.N., and Bisaro, D.M. (2010). RNA silencing directed against geminiviruses: post-transcriptional and epigenetic components. Biochim. Biophys. Acta 1799, 337-351. Robaglia, C., and Caranta, C. (2006). Translation initiation factors: a weak link in plant RNA virus infection. Trends Plant Sci. 11(1), 40-5. Rogers, G.W., Jr., Komar, A.A., and Merrick, W.C. (2002). eIF4A: the godfather of the DEAD box helicases. Prog Nucleic Acid Res Mol Biol 72, 307-331. Rojas, M. R., Hagen, C., Lucas, W. J., and Gilbertson, R. L. (2005). Exploiting chinks in the plant's armor: Evolution and emergence of geminiviruses. Annu. Rev. Phytopathol. 43, 361-394. Rothenstein, D., Krenz, B., Selchow, O., and Jeske, H. (2007). Tissue and cell tropism of Indian cassava mosaic virus (ICMV) and its AV2 (precoat) gene product. Virology. 359(1), 137-45.

167

Ruffel, S., Gallois, J.L., Lesage, M.L., and Caranta, C. (2005). The recessive potyvirus resistance gene pot-1 is the tomato orthologue of the pepper pvr2-eIF4E gene. Mol. Gen. Genomics 274(4), 346-53. Rushing, A.E., Sunter, G., Gardiner, W.E., Dute, R.R. and Bisaro, D.M. (1987). Ultrastructural aspects of tomato golden mosaic virus infection in tobacco. Phytopathology 77, 1231-1236. Saeed, M., Zafa, Y., Randles, J.W., and Rezaian, M.A. (2007). A monopartite begomovirus-associated DNA beta satellite substitutes for the DNA B of a bipartite begomovirus to permit systemic infection. J Gen Virol. 88(10), 2881-9. Sakakibara, H. (2006). Cytokinins: activity, biosynthesis, and translocation. Annu Rev Plant Biol. 57, 431-49. Review. Sanders, M.J., Grondin, P.O., Hegarty, B.D., Snowden, M.A., and Carling, D. (2007). Investigating the mechanism for AMP activation of the AMP-activated protein kinase cascade. Biochem. J. 403, 139-148. Saunders, K., Bedford, I. D., Briddon, R. W., Markham, P. G., Wong, S. M. & Stanley, J. (2000). A unique virus complex causes Ageratum yellow vein disease. Proc Natl Acad Sci U S A 97, 6890–6895. Saunders, K., Bedford, I. D., Yahara, T., and Stanley, J. (2003). Aetiology: The earliest recorded plant virus disease. Nature 422 (6934), 831. Scott, J.W., Hawley, S.A., Green, S.A., Anis, M., Stewart, G., Scullion, G.A., Norman, D.G., and Hardie, D.G. (2004). CBS domains form energy-sensing modules whose binding of adenosine ligands is disrupted by disease mutations. J. Clin. Invest. 113, 274- 284. Sehnke, P.C., Rosenquist, M., Alsterfjord, M., DeLille, J., Sommarin, M., Larsson, C., and Ferl, R.J. (2002). Evolution and isoform specificity of plant 14-3-3 proteins. Plant Mol Biol. 50(6), 1011-8. Selth, L.A., Dogra, S.C., Rasheed, M.S., Healy, H., Randles, J.W., and Rezaian, M.A. 92005). A NAC domain protein interacts with tomato leaf curl virus replication accessory protein and enhances viral replication. Plant Cell. 17(1), 311-25. Sen, G. C., and Ransohoff, R. M. (1993). Interferon-induced antiviral actions and their regulation. Adv. Virus Res. 42, 57. Shen, W., and Hanley-Bowdoin, L. (2006). Geminivirus infection up-regulates the expression of two Arabidopsis protein kinases related to yeast SNF1- and mammalian AMPK-activating kinases. Plant Physiol. 142(4), 1642-55. Shen, W., Reyes, M.I., and Hanley-Bowdoin, L. (2009). Arabidopsis protein kinases GRIK1 and GRIK2 specifically activate SnRK1 by phosphorylating its activation loop. Plant Physiol. 150 (2), 996-1005.

168

Shin, R., Alvarez, S., Burch, A.Y., Jez, J.M., and Schachtman, D.P. (2007). Phosphoproteomic identification of targets of the Arabidopsis sucrose nonfermenting-like kinase SnRK2.8 reveals a connection to metabolic processes. Proc Natl Acad Sci U S A. 104(15), 6460-5. Singh, D.K., Islam, M.N., Choudhury, N.R., Karjee, S., and Mukherjee, S.K. (2007). The 32 kDa subunit of replication protein A (RPA) participates in the DNA replication of Mung bean yellow mosaic India virus (MYMIV) by interacting with the viral Rep protein, Nucl. Acids Res. 35(3), 755-770. Slocombe, S.P., Beaudoin, F., Donaghy, P.G., Hardie, D.G., Dickinson, J.R., Halford, N.G. (2004). SNF1-related protein kinase (snRK1) phosphorylates class I heat shock protein. Plant Physiol Biochem. 42(2), 111-6. Smith, R.D., and Walker, J.C. (1996) Plant protein phosphatases. Annu Rev Plant Physiol Plant Mo1 Biol 47, 101-125 Stanley, J. (1983). Infectivity of the cloned geminivirus genome requires sequences from both. DNAs. Nature 305, 643-645. Stanley, J., Latham, J. R., Pinner, M. S., Bedford, I., and Markham, P. G. (1992). Mutational analysis of the monopartite geminivirus beet curly top virus. Virology 191(1), 396-405. Stenger, D. C., Revington, G. N., Stevenson, M. C., and Bisaro, D. M. (1991). Replicational release of geminivirus genomes from tandemly repeated copies: Evidence for rolling circle replication of a plant viral DNA. Proc. Natl. Acad. Sci. USA 88, 8029- 8033. Sugden, C., Donaghy, P.G., Halford, N.G., and Hardie, D.G. (1999a). Two SNF1-related protein kinases from spinach leaf phosphorylate and inactivate 3-hydroxy-3- methylglutaryl-coenzyme A reductase, nitrate reductase, and sucrose phosphate synthase in vitro. Plant Physiol. 120(1), 257-74. Sugden, C., Crawford, R.M., Halford, N.G., and Hardie, D.G. (1999b). Regulation of spinach SNF1-related (SnRK1) kinases by protein kinases and phosphatases is associated with phosphorylation of the T loop and is regulated by 5'-AMP. Plant J. 19(4), 433-9. Sunter, G., and Bisaro, D. M. (1991). Transactivation in a geminivirus: AL2 gene product is needed for coat protein expression. Virology 180(1), 416-9. Sunter, G., and Bisaro, D. M. (1992). Transactivation of geminivirus AR1 and BR1 gene expression by the viral AL2 gene product occurs at the level of transcription. Plant Cell 4(10), 1321-31. Sunter, G., and Bisaro, D.M. (2003). Identification of a minimal sequence required for activation of the tomato golden mosaic virus coat protein promoter in protoplasts. Virology 305, 452-462.

169

Sunter, G., Hartitz, M. D., Hormuzdi, S. G., Brough, C. L., and Bisaro, D. M. (1990). Genetic analysis of tomato golden mosaic virus: ORF AL2 is required for coat protein accumulation while ORF AL3 is necessary for efficient DNA replication. Virology 179(1), 69-77. Sunter, G., Stenger, D. C., and Bisaro, D. M. (1994). Heterologous complementation by geminivirus AL2 and AL3 genes. Virology 203(2), 203-10. Sunter, G., Sunter, J.L., and Bisaro, D.M. (2001). Plants expressing tomato golden mosaic virus AL2 or beet curly top virus L2 transgenes show enhanced susceptibility to infection by DNA and RNA viruses. Virology 285(1), 59-70. Suter, M., Riek, U., Tuerk, R., Schlattner, U., Wallimann, T., and Neumann, D. (2006). Dissecting the role of 5'-AMP for allosteric stimulation, activation, and deactivation of AMP-activated protein kinase. J. Biol. Chem. 281, 32207-32216. Sutherland, C.M., Hawley, S.A., McCartney, R.R., Leech, A., Stark, M.J.R., Schmidt, M.C., and Hardie, D.G. (2003). Elm1p is one of three upstream kinases for the Saccharomyces cerevisiae SNF1 complex. Curr. Biol. 13, 1299-1305. Tatlisumak, T., Takano, K., Carano, R. A., Miller, L. P., Foster, A. C., and Fisher, M. (1998). Delayed treatment with an adenosine kinase inhibitor, GP683, attenuates infarct size in rats with temporary middle cerebral artery occlusion. Stroke 29(9), 1952-8. Trinks, D., Rajeswaran, R., Shivaprasad, P. V., Akbergenov, R., Oakeley, E. J., Veluthambi, K., Hohn, T., and Pooggin, M. M. (2005). Suppression of RNA silencing by a geminivirus nuclear protein, AC2, correlates with transactivation of host genes. J Virol 79(4), 2517-27. Vanitharani, R., Chellappan, P., Pita, J.S., and Fauquet, C.M. (2004). Differential roles of AC2 and AC4 of cassava geminiviruses in mediating synergism and suppression of posttranscriptional gene silencing. J Virol. 78(17), 9487-98. van Wezel, R., Liu, H., Tien, P., Stanley, J., and Hong, Y. (2002). Mutation of three cysteine residues in tomato yellow leaf curl virus-China C2 protein causes dysfunction in pathogenesis and posttranscriptional gene silencing-suppression. Mol. Plant Microbe Interact. 15, 203-208. van Wezel, R., Liu, H., Wu, Z., Stanley, J., and Hong, Y. (2003). Contribution of the zinc finger to zinc and DNA binding by a suppressor of posttrancriptional gene silencing. J. Virol. 77, 696-700. Voinnet, O., Pinto, Y. M., and Baulcombe, D. C. (1999). Suppression of gene silencing: A general strategy used by diverse DNA and RNA viruses of plants. Proc. Natl. Acad. Sci. USA 96, 14147-14152. von Schwartzenberg, K., Kruse, S., Reski, R., Moffatt, B., and Laloue, M. (1998). Cloning and characterization of an adenosine kinase from Physcomitrella involved in cytokinin metabolism. Plant J. 13, 249-257.

170

Wang, H., Buckley, K. J., Yang, X., Buchmann, R. C., and Bisaro, D. M. (2005). Adenosine kinase inhibition and suppression of RNA silencing by geminivirus AL2 and L2 proteins. J. Virol. 79, 7410-7418. Wang, H., Hao, L., Shung, C.-Y., Sunter, G., and Bisaro, D. M. (2003). Adenosine kinase is inactivated by geminivirus AL2 and L2 proteins. Plant Cell 15, 3020-3032. Wang, X., and Proud, C.G. (2006). The mTOR pathway in the control of protein synthesis. Physiology (Bethesda) 21, 362–369. Wiesner, J. B., Ugarkar, B. G., Castellino, A. J., Barankiewicz, J., Dumas, D. P., Gruber, H. E., Foster, A. C., and Erion, M. D. (1999). Adenosine kinase inhibitors as a novel approach to anticonvulsant therapy. J Pharmacol Exp Ther 289(3), 1669-77. Wilson, W.A., Hawley, S.A., and Hardie, D.G. (1996). The mechanism of glucose repression/derepression in yeast: SNF1 protein kinase is activated by phosphorylation under derepressing conditions, and this correlates with a high AMP:ATP ratio. Curr. Biol. 6, 1426-1434. Woods, A., Dickerson, K., Heath, R., Hong, S.-P., Momcilovic, M., Johnstone, S.R., Carlson, M., and Carling, D. (2005). Ca2+/calmodulin-dependent protein kinase kinase b acts upstream of AMP-activated protein kinase in mammalian cells. Cell Metab. 2, 21-33. Wullschleger, S., Loewith, R., and Hall, M.N. (2006). TOR signaling in growth and metabolism. Cell 124, 471–484. Xiao, B., Heath, R., Saiu, P., Leiper, F.C., Leone, P., Jing, C., Walker, P.A., Haire, L., Eccleston, J.F., Davis, C.T., Martin, S.R., Carling, D., and Gamblin, S.J. (2007). structural basis for AMP binding to mammalian AMP-activated protein kinase. Nature 449, 496-500. Yang, X., Baliji, S., Buchmann, R.C., Wang, H., Lindbo, J.A., Sunter, G., and Bisaro, D.M. (2007). Functional modulation of the geminivirus AL2 transcription factor and silencing suppressor by self-interaction. J Virol. 81(21), 11972-81. Young, L.S., Harrison, B.R., Narayana Murthy, U.M., Moffatt, B.A., Gilroy, S., and Masson, P.H. (2006). Adenosine kinase modulates root gravitropism and cap morphogenesis in Arabidopsis. Plant Physiol. 142(2), 564-73. Zhang, H.-S., and Wu, M.-R. (2009). SIRT1 regulates Tat-induced HIV-1 transcription through activating AMP-activated protein kinase. Virus Res. 146, 51-57. Zhang, Y., Wang, Y., Kanyuka, K., Parry, M.A., Powers, S.J., and Halford, N.G. (2008). GCN2-dependent phosphorylation of eukaryotic translation initiation factor-2alpha in Arabidopsis. Journal of experimental botany 59(11), 3131-3141. Zheng, L., Baumann, U., and Reymond, J.L. (2004). An efficient one-step site-directed and site-saturation mutagenesis protocol. Nucleic Acids Res. 32(14), e115.

171

Zhu, C. Z., Mikusa, J., Chu, K. L., Cowart, M., Kowaluk, E. A., Jarvis, M. F., and McGaraughty, S. (2001). A-134974: a novel adenosine kinase inhibitor, relieves tactile allodynia via spinal sites of action in peripheral nerve injured rats. Brain Res. 905(1-2), 104-10. Zrachya A, Kumar PP, Ramakrishnan U, Levy Y, Loyter A, Arazi T, Lapidot M, Gafni Y. Production of siRNA targeted against TYLCV coat protein transcripts leads to silencing of its expression and resistance to the virus. Transgenic Res. 16(3), 385-98.

172