Title: Characterization of the RNA-dependent RNA polymerase gene of citrus tristeza closterovirus
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00100735/00001
 Material Information
Title: Characterization of the RNA-dependent RNA polymerase gene of citrus tristeza closterovirus
Physical Description: Book
Language: English
Creator: Çevik, Bayram, 1970-
Publisher: University of Florida
Place of Publication: Gainesville Fla
Gainesville, Fla
Publication Date: 2001
Copyright Date: 2001
 Subjects
Subject: Plant Pathology thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Plant Pathology -- UF   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )
 Notes
Summary: ABSTRACT: Tristeza disease is the most destructive and economically important problem in most citrus producing regions of the world. The disease is caused by citrus tristeza closterovirus (CTV) which is a single-stranded positive-sense RNA virus classified in the genus Closterovirus. The development of efficient control measures for CTV requires the characterization of genes encoded by the viral genome and a better understanding of how the virus replicates and causes disease. To improve the molecular understanding of CTV, the RNA-dependent RNA polymerase (RdRp) gene, which is required for the replication of the virus, was characterized. To study the sequence variability in the RdRp gene of biologically and geographically different isolates of CTV, the RdRp genes from ten different strains were amplified by reverse transcriptase- polymerase chain reaction (RT-PCR) and sequenced. Analysis of the nucleotide and amino acid sequences showed that the RdRp gene was highly conserved among all isolates of CTV. Evolutionary relationships among different isolates was predicted using phylogenetic analysis of the RdRp sequences. The RdRp gene from the Florida quick decline (QD) isolate T36 of CTV, was cloned and expressed in E. coli and polyclonal antiserum specific to the expressed RdRp fusion protein was produced. Using this antiserum, the RdRp was detected in CTV-infected tissue and found to be localized mostly in the membrane fraction of infected citrus cells.
Summary: ABSTRACT (cont.): It has been proposed that the RdRp gene of CTV is expressed by a +1 translational frameshift, however, the occurrence of the +1 frameshift has not been demonstrated for CTV or any other closterovirus. The +1 translational frameshift was demonstrated by an in vitro transcription and translation method. An in vivo Agrobacterium-mediated transient expression assay was developed to investigate the involvement of the CTV sequences in the +1 frameshift. Deletion studies in the transient assay showed that 123-nucleotide overlapping sequence of the ORFs 1a and 1b was necessary and sufficient for the +1 frameshift. It was found that the 123 nt region contains a conserved stem loop structure, and that this conserved stem loop structure is required for the +1 frameshift. To explore the possibilities of replicase- mediated pathogen-derived resistance against CTV, five different constructs of the RdRp gene of CTV, including full length, untranslatable, antisense and two modified RdRp genes (with mutations and a deletion at the GDD motif) were produced. Following transformation of grapefruit (Citrus Paradisi cv. Duncan) seedling segments using Agrobacterium containing these constructs, a number of transgenic plants were regenerated and established in the greenhouse. The transgenic nature of most of these plants were confirmed by PCR amplification of the GUS and RdRp genes from their genomic DNA.
Summary: KEYWORDS: Citrus tristeza closterovirus, RdRp, RNA-dependent RNA polymerase, translational control, translational frameshift, transgenic plants, replicase mediated resistance, citrus
Thesis: Thesis (Ph. D.)--University of Florida, 2001.
Bibliography: Includes bibliographical references (p. 119-137).
System Details: System requirements: World Wide Web browser and PDF reader.
System Details: Mode of access: World Wide Web.
Statement of Responsibility: by Bayram Çevik.
General Note: Title from first page of PDF file.
General Note: Document formatted into pages; contains xiii, 138 p.; also contains graphics.
General Note: Vita.
 Record Information
Bibliographic ID: UF00100735
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 49241093
alephbibnum - 002766273
notis - ANP4312

Downloads

This item has the following downloads:

DissertationBC ( PDF )


Full Text











CHARACTERIZATION OF THE RNA-DEPENDENT RNA POLYMERASE GENE OF
CITRUS TRISTEZA CLOSTEROVIRUS












By

BAYRAM CEVIK


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001
































Copyright 2001

by

Bayram Cevik































To my wife Mehtap















ACKNOWLEDGMENTS


This dissertation has been made possible through the support and encouragement I have

received from many people. First, I would like to express my sincere gratitude and

appreciation to my mentor Dr. Charles Niblett for having me in his lab for many years. I thank

him for his advice, support and encouragement over the years.

I want to extend my gratitude to Dr. Richard Lee for serving as co-adviser for my

graduate studies and for his support, encouragement and for his help in preparation of this

manuscript all the way from Lake Alfred.

I would like to thank Drs. Ken Derrick and Gloria Moore for serving in my committee

and for their helpful suggestions and comments and for reviewing this manuscript. I would

also like to thank Dr. Gloria Moore for her help with the citrus transformation and use of her

lab facilities. I want to extend my thanks to Dr. Rachel Shireman for her support and

dedication to graduate students.

I want to give my special thanks to the former and current members of the Niblett lab

especially, Drs. S. S. Pappu, H. R. Pappu, V. J. Febres, R. Chandrika, K. L. Manjunath and

my fellow students J. Vazquez, R. Harakava, M. Rosales and Y. Petersen for the limitless

help, understanding and friendship they have provided to me in the past five years. I thank

Mrs. B. Garagorry and Clyde Graham for their technical assistance. I also want to thank my

friends in Gainesville and around the world for the support and friendship they provided to

me in my endeavor.









My deepest thanks and appreciation go to my parents for the constant support and

unconditional love they have provided throughout my life. I appreciate the endless love and

support I receive from my sisters Yeter and Muyesser, and my brothers Emrullah and

Kuddusi and their respective families. I want to thank my in-laws for their support and

encouragement.

I would like to give my most special thanks and love to my wonderful wife Mehtap

Sahin- evik, for her constant love, help, support, understanding, and encouragement since the

day we met.

Finally, I thank the Turkish Ministry of Education for their generous financial support

during my graduate studies at the University of Florida.















TABLE OF CONTENTS
Page



ACKNOW LEDGM ENTS ............................................... iv

LIST OF TABLES ......... ........................................... ix

LIST OF FIGURES ................ ............................ x

ABSTRACT ....... ........ ................................ xii

CHAPTERS

1. INTRODUCTION ................ ............................ 1

2. REVIEW OF LITERATURE ....................................... 5

Tristeza Disease .................................................. 5
Citrus Tristeza Virus . ................ ............................ 6
Taxonomy ................... .............. ............... 6
Morphological Characteristics of CTV .............................. 6
Biological Characteristics of CTV ................................. 7
Host range of CTV ............................... ......... 7
Cytopathic effects of CTV ....................... ......... 7
Symptoms of CTV ........................................ 8
Transmission of CTV ...................................... 9
Biochemical Properties of CTV ................................. 10
Genome Organization of CTV ................ ................... 11
Characteristics and Functions of Genes of CTV ........................ 12
Replication of CTV ................................ ......... 14
Defective RNAs Associated with CTV ............................ 15
CTV Gene Expression Strategies ................................ 16
Polyprotein processing ................. .................. 17
Subgenomic RNA ...... ................................. 17
Translational frameshifting ............................... 18
Sequence Variation Among CTV Isolates ............................. 19

Control of CTV .................................................... 22

vi









Mild Strain Cross Protection ................. ................... 22
Use of CTV Tolerant Rootstocks ................................ 22
Breeding for CTV Resistance ................ ................... 23
Genetic Engineering for Virus Resistance ............................ 24
Pathogen-derived resistance .............................. 25
Replicase-mediated resistance ............................ 26
Mechanisms of replicase-mediated resistance ................... 30
Protein-mediated resistance ......................... 31
RNA-mediated resistance ............................. 32
Genetic Engineering of Citrus ................ ................... 33

3. SEQUENCING AND ANALYSIS OF THE RNA-DEPENDENT
RNA POLYMERASE GENE OF CTV ISOLATES ........................ 37

Introduction................................................... 37
Materials and Methods ......... ................................. 38
Virus Isolates ......... ................................ 38
Oligonucleotide Primers ........................................ 38
RNA Isolation and Complementary DNA (cDNA) Synthesis ............. 40
Polymerase Chain Reaction (PCR) ............................... 40
Sequencing of the RdRp Genes of CTV ........................ 40
Sequence Analysis ................... ................. ....... 42
Results and Discussion ............................................ 42
C conclusions .. . ..... ....................................... 49

4. CLONING AND EXPRESSION OF THE RNA-DEPENDENT
RNA POLYMERASE OF CTV IN Escherichia coli ....................... 51

Introduction .......... ...................................... ......... 51
Materials and Methods ............................................. 53
Cloning of the CTV RdRp Gene ................................. 53
Expression of the CTV RdRp .................................... 53
Production of Polyclonal Antibodies Specific for the CTV RdRp .......... 54
W western B lot A analysis ........................................... 54
Cell Fractionation Assays ................ ...................... 54
Results and Discussion ............................................. 55
Expression ofCTV RdRp in E. coli ............................... 55
Production and Testing of Polyclonal Antibodies to the CTV RdRp ........ 57
Detection and Sub-Cellular Localization of the CTV RdRp in
CTV -Infected Citrus Tissue ....................................... 57
Conclusions ........................................ 60
5. CHARACTERIZATION OF THE +1 TRANSLATIONAL FRAMESHIFT
FOR THE RNA-DEPENDENT RNA POLYMERASE GENE OF CTV ......... 62

Introduction ......... ....................................... ......... 62









Material and Methods .............................. ............... 65
In vitro Transcription and Translation ............................ 65
Amplification, Mutagenesis and Cloning of Overlap Region of CTV ........ 67
Sequence Analysis of the Overlap Region of Different CTV Isolates ........ 72
Transient Expression Assay ................ ................... 72
Plant Materials ............... ............................. 73
Preparation of the Agrobacterium Suspension ........................ 73
Infiltrations of the Leaves with Agrobacterium Suspensions .............. 73
Histochemical GUS Staining .................................... 75
Results and Discussion ...................................... ... ..... 75
In vitro Transcription and Translation ............................ 75
Involvement of the Overlap Region of ORFla and lb in the +1 Frameshift ... 79
Sequence Comparison and Secondary Structure Analysis of
the Overlap Region ........................................... 82
The Role of the Secondary Structure and the Rare Arginine
Codon in the +1 Frameshift ................ ..................... 84
C conclusions .. . ..... ....................................... 86

6. GENETIC TRANSFORMATION OF Citrus paradisi WITH THE RNA
DEPENDENT RNA POLYMERASE GENE OF CTV ...................... 88

Introduction .......... ...................................... ......... 88
Materials and Methods ............................................. 92
Cloning and Mutagenesis of CTV RdRp for Transformation .............. 92
Genetic Transformation of Citrus with RdRp gene of CTV .............. 99
Seed germ nation ........................................ 99
Transformation of epicotyl segments ......................... 99
Selection and regeneration of transgenic shoots ................. 101
Rooting of transgenic shoots ............................ 101
Analysis of Transgenic Shoots and Plants ........................... 102
Flourescent microscopy ............................... 102
-glucuronidase (GUS) assay ............................ 102
Polymerase Chain Reaction Assay .......................... 103
Results and Discussion ..................................... .......... 105
Regeneration of Transgenic Plants ................................. 105
Analysis of Transgenic Plants by Polymerase Chain Reaction Analysis .... 109
Conclusions .............. ................... .................. 114

7. SUMMARY AND CONCLUSIONS ........................... ... 115

REFERENCES ................................................ 119

BIOGRAPHICAL SKETCH ......................................... 138















LIST OF TABLES


Table Page

3-1 Characteristics of the CTV isolates used in this study. .................. 39

3-2 Sequences of the oligonucleotide primers used in this study. .............. 41

3-3 Percentage of nucleotide sequence identity of the RdRp genes from
different isolates of CTV. ....................................... 43

3-4 Percentage of and amino acid sequence identity of the RdRp genes from
different isolates of CTV. ....................................... 44

5-1 The sequences of the primers used for amplification, mutagenesis and
cloning of the frameshift constructs. ............... ............ ... 66

5-2 The percentage of nucleotide sequence identity of the overlap region
ofORF la and lb among ten different isolates of CTV ................. 81

6-1 List of primers used for PCR amplification, mutagenesis and sequencing
of the RdRp constructs and the PCR amplification of the GUS gene. ........ 93

6-2 Summary of transformation experiments with five different constructs of
CTV-RdRp gene. .. ........................................ 106

6-3 Summary of the PCR analysis of the transgenic plants with different
constructs of CTV RdRp gene ................ .................. 113















LIST OF FIGURES
Figure Page

3-1 Schematic representation of the RdRp region of CTV and the strategy used
for sequencing the RdRp gene. ................. ................ 41

3-2 Homology and absolute complexity graphs of the nucleotide
sequences of the RdRp............................... 43

3-3 Homology and absolute complexity graphs of deduced amino acid sequences
of the RdRp gene ................................... ......... 44

3-4 Multiple sequence alignment of the 3' region of the RdRp genes of CTV
isolates showing the 18 base insertion in some isolates. ................. 46

3-5 Aphylogenetic tree showing relationships among CTVisolates based
on their RdRp genes sequences. ................ ................. 48

4-1 Induction of the expression of the CTV RdRp in E. coli by IPTG. .......... 56

4-2 Time course study of expression of the CTV RdRp in E. coli
after the induction by IPTG. ................. ................... 56

4-3 Accumulation of the expressed CTV RdRp in cellular fractions ............ 56

4-4 Western blot analysis of the CTV RdRp expressed in E. coli ............... 59

4-5 DetectionofCTVRdRp in CTV-infected citrus tissue by Westernblot ........ 59

4-6 Sub-cellular localization of the CTV RdRp by Western blot analysis using ... 59

5-1 The cloning of the CTV ORFla and lb for in vitro transcription and translation.66

5-2 Amplification mutagenesis and cloning strategy used for the constructs
of overlap region of CTV ORFla and lb. ............................ 68

5-3 The constructs prepared for the overlap region of CTV ORF la and
lb in pC 1303 plasm id vector ...................................... 70









5-4 The pC1303 plasmid vector containing the constructs of the overlap
region ofCTV ORFla and lb with mutations ......................... 71

5-5 The flowchart of the Agrobacterium-mediated transient expression assay .... 74

5-6 In vitro transcription and translation of HEL-RdRp construct i
n wheat germ extract ....................................... 76

5-7 Expression of the GUS and GFP reporter genes in citrus leaves
infiltrated Agrobacterium with the pC1303 ........................... 78

5-8 Multiple alignment and homology graph of nucleotide sequences of the
overlapping region of ORFla and lb ............................... 81

5-9 Possible folding patterns of the highly conserved nucleotide sequence around
the arginine codon in the overlapping region of the ORF la and lb ......... 83

5-10 Expression of the reporter genes GUS and/or GFP in citrus leaves infiltrated
with pC1303 plasmids. ......................................... 85

6-1 Description and flow chart of the site-directed mutagenesis of
the RdRp gene ofCTV .......................................... 95

6-2 Partial nucleotide and amino acid sequence alignments of CTV-RdRp
constructs showing the deletion and point mutations ................... 96

6-3 Cloning of the CTV-RdRp constructs for plant transformation. ............ 98

6-4 Production oftransgenic grapefruit plants from epicotyl segments using
Agrobacterium-mediated transformation method ..................... 100

6-5 Analysis of the transgenic shoots regenerated from epicotyl segments
transformed with Agrobacterium tumefaciences ................... .. 104

6-6 The transgenic plants with different constructs of the CTV-RdRp gene ..... 110

6-7 Analysis of representative putative transgenic plants by
polymerase chain reaction. ................. ................... 111















Abstract of Dissertation Presented to the Graduate School of the University of Florida in
Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

CHARACTERIZATION OF THE RNA-DEPENDENT RNA POLYMERASE GENE OF
CITRUS TRISTEZA CLOSTEROVIRUS

By

Bayram (evik

August 2001


Chairperson: Dr. Charles L. Niblett
Cochairperson: Dr. Richard F. Lee
Major Department: Plant Pathology


Tristeza disease is the most destructive and economically important problem in

most citrus producing regions of the world. The disease is caused by citrus tristeza

closterovirus (CTV) which is a single-stranded positive-sense RNA virus classified in the

genus Closterovirus. The development of efficient control measures for CTV requires the

characterization of genes encoded by the viral genome and a better understanding of how

the virus replicates and causes disease. To improve the molecular understanding of CTV,

the RNA-dependent RNA polymerase (RdRp) gene, which is required for the replication of

the virus, was characterized.

To study the sequence variability in the RdRp gene of biologically and

geographically different isolates of CTV, the RdRp genes from ten different strains were

amplified by reverse transcriptase- polymerase chain reaction (RT-PCR) and sequenced.

Analysis of the nucleotide and amino acid sequences showed thatthe RdRp gene was highly









conserved among all isolates of CTV. Evolutionary relationships among different isolates

was predicted using phylogenetic analysis of the RdRp sequences. The RdRp gene from the

Florida quick decline (QD) isolate T36 of CTV, was cloned and expressed in E. coli and

polyclonal antiserum specific to the expressed RdRp fusion protein was produced. Using

this antiserum, the RdRp was detected in CTV-infected tissue and found to be localized

mostly in the membrane fraction of infected citrus cells.

It has been proposed that the RdRp gene of CTV is expressed by a +ltranslational

frameshift, however, the occurrence of the +1 frameshift has not been demonstrated for CTV

or any other closterovirus. The +1 translational frameshift was demonstrated by an in vitro

transcription and translation method. An in vivo Agrobacterium -mediated transient

expression assay was developed to investigate the involvement ofthe CTV sequences in the

+1 frameshift. Deletion studies in the transient assay showed that 123-nucleotide

overlapping sequence of the ORFs la and lb was necessary and sufficient for the +1

frameshift. It was found that the 123 nt region contains a conserved stem loop structure, and

that this conserved stem loop structure is required for the +1 frameshift.

To explore the possibilities of replicase- mediated pathogen-derived resistance

against CTV, five different constructs of the RdRp gene of CTV, including full length,

untranslatable, antisense and two modified RdRp genes (with mutations and a deletion at

the GDD motif) were produced. Following transformation of grapefruit (Citrus Paradisi

cv. Duncan) seedling segments using Agrobacterium containing these constructs, a number

oftransgenic plants were regenerated and established in the greenhouse. The transgenic

nature of most of these plants were confirmed by PCR amplification of the GUS and RdRp

genes from their genomic DNA.














CHAPTER 1
INTRODUCTION

Citrus is one of the mostwidely grown and economically important fruit crops in the

world with an annual production of about 100 million metric tons. The world citrus

production has been increasing steadily to accommodate growing demands of domestic and

international markets. Citrus is not only economically important for generation of income

and foreign trade but also very important in terms of nutrition. The United States (US) is the

second largest citrus producer in the world, and the annual retail value of citrus grown in

the US alone is about $20 billion. Thus citrus production makes a significant contribution

to the economy of citrus growing states, especially the state of Florida where the majority

of the US citrus is produced.

Tristeza disease of citrus, caused by citrus tristeza virus (CTV), is the most

destructive and economically important disease in most citrus producing regions of the

world. In the past, destructive epidemics oftristeza disease severely damaged or destroyed

the citrus industries on sour orange rootstock by killing millions of citrus trees grafted in

Brazil, Argentina, Venezuela, and Spain. In fact, the name tristeza, which means sadness in

Spanish and Portugese, was given after the destructive epidemics in South America. The

disease is now found in almost all citrus growing regions including Florida. It can

significantly reduce yield and fruit quality and has become a limiting factor for commercial

citrus production in some regions. The importance and destructiveness oftristeza disease









2

may increase and reach a new level in Florida due to the recent invasion of the brown citrus

aphid (Toxoptera citricida), the most efficient vector of CTV.

Even though several strategies have been developed for managing the tristeza

disease, there is no effective measure for controlling or eliminating CTV from citrus

orchards. The currently available management strategies of quarantine and eradication are

only effective for regions where CTVis not present or it has beenjust introduced. Tolerant

rootstocks and mild strain cross protection are used in some regions where CTVis endemic;

however, these strategies are not effective against all strains of CTV. As in other diseases,

genetic resistance is the most effective strategy for controlling tristeza disease of citrus.

Although a single dominant gene conferring resistance to CTV is found in Poncirus

trifoilata, which is a close relative of citrus, integration of this gene into commercially

important citrus scions has been hindered by the problems associated with conventional

citrus breeding. The resistance gene has been mapped using molecular markers, but the

isolation, characterization and use of this gene for developing CTV resistant citrus plants

will still require many years of research. In recent years, advances in molecular biology,

plant transformation and tissue culture techniques have provided new and more efficient

approaches, such as genetic engineering, to overcome the limitations of conventional

breeding. Using genetic engineering, transgenic plants with a defined trait can be developed

by incorporating a specific gene from a different source into the plant genome without

altering the other desirable characteristics. Genetically engineered resistance has been

developed by transforming plants with genes or sequences from pathogens, which is termed

pathogen-derived resistance (PDR). Different crop plants resistant to many viruses have

already been successfully developed by transforming with viral genes. Hence, the genetic









3

transformation of citrus with sequences from the CTV genome has great potential for

developing PDR against CTV.

As with other viruses, development of efficient control measures for CTV requires

the characterization of the genome and a better understanding of the biology of the virus.

Citrus tristeza virus is a positive-stranded RNA virus in the genus Closterovius. It is the

largest knownplant RNA virus with about 20 kb genome. Besides its large genome size, the

genome organization, number and functions of the genes encoded by the genome, mechanisms

of gene expression, and population complexity and unusual sequence variation among

isolates are all important aspects ofCTVbiology. Genetic analysis of the CTV genome and

understanding of the mechanisms of replication and pathogenesis of CTV were limited by

the large genome size, low titer in virus-infected tissue, restriction to phloemtissue and the

lack of a herbaceous experimental host. The genetic analysis of the CTV genome has

recently begun with the development of a full length infectious cDNA clone and a protoplast

system for replication of CTV.

Ininfected plants, CTV isolates consist of unusually complex populations of distinct

genotypes. The complexity of CTV populations creates problems for diagnosis and strain

identification, and also hinders the understanding of the mechanisms of pathogenesis.

Sequencing and molecular characterization of genomes and/or individual genes of different

isolates has improved our understanding of CTV biology, and this information should

provide new approaches to control the tristeza diseases caused by CTV. The RdRp genes

are highly conserved among all RNA viruses, and they are required for replication of these

viruses. In addition, it has been demonstrated that the functional and modified sequences of

the RdRp gene of many viruses confer resistance to virus infection in transgenic plants when









4

they are transformed with these genes. These general characteristics of the RdRp genes, in

general, as well as the translational regulation of their expression, which is controlled by

a +1 translational frameshift in closteroviruses, emphasize the significance of the RdRp

gene for CTV and its potential use in development of resistance to the tristeza disease.

The overall objective of this study is to characterize the RdRp gene and its

expression by the +1 translational frameshift, and to explore the possibilities of replicase-

mediated resistance to CTV. The specific objectives are the following:

1) To determine the sequences of the RdRp genes from geographically and
biologically different isolates and to study the sequence variation and relationships among
different isolates of CTV.

2) To clone and express the RdRp gene of CTV inEscherichia coli, and to produce
polyclonal antiserum against the expressed protein for the detection of the RdRp in CTV-
infected plants.

3) To demonstrate the occurrence of the +1 translational frameshift proposed to be
required for the expression of the RdRp gene of CTV and to characterize the sequences
involved in the +1 translational frameshift.

4) To transform Duncan grapefruit plants with different constructs of the RdRp gene
from CTV and to regenerate and produce transgenic plants potentially resistant to CTV.














CHAPTER 2
REVIEW OF LITERATURE

Tristeza Disease

Tristeza is one of the most destructive and economically important diseases of

commercial citrus worldwide (Bar-Joseph et al. 1989). Although the disease is believed

to have originated in southeastAsia (Wallace etal. 1956), it was first reported as a decline

of citrus scions grafted on sour orange rootstock in South Africa around 1910 (Weber,

1943). A similar decline disease in the 1930s causing epidemic losses of sweet orange

trees on sour orange rootstock was reported in Argentina and Brazil. Tristeza was first

thought to be a rootstock-scion incompatibility, or a nutritional problem or a root disease.

In 1946, Meneghini transmitted the disease by aphids and experimentally proved itto be of

viral origin (Bar-Joseph et al. 1989; Lee and Rocha-Pena, 1992). Earlier reports of

incompatibility, nutritional problems and quick decline are now all recognized as different

forms of the same viral disease caused by citrus tristeza virus (CTV) (Grant et al. 1951;

McClean, 1950). Some other disorders found later in citrus, such as stem pitting (Da Graca

et al. 1984) and seedling yellows (Roistacher, 1982), also are associated now with CTV.

In the United States, quick decline of citrus on sour orange rootstock was first

reported in California (Fawcett and Wallace, 1946) and later in Florida (Grant, 1952). It

was suggested that the disease was introduced into California with Meyer lemon

introductions from China in 1908, and then introduced into Texas and Florida by the

movement of Meyer lemon (Wallace and Drake, 1955). Tristeza is now present in most









6

world, with the exception of some Mediterranean and Central American countries, and some

isolated islands (Lee and Rocha-Pena, 1992).

Citrus Tristeza Virus

Taxonomy

Based on distinct morphological, biological and molecular characteristics and

phylogenetic analysis, CTV is classified in the genus Closterovirus in the Closteroviridae

family of the positive-stranded plant RNA viruses (Bar-Joseph et al. 1979; Koonin and

Dolja.,1993; Dolja et al. 1994). Citrus tristeza virus and the other members of the

closterovirus group are characterized by long flexuous rod-shaped particles, a large

undivided genome, association with phloem tissue of their hosts, semi-persistent mode of

transmission and vector specificity, and typical inclusion bodies produced in the infected

cells (Bar-Joseph et al. 1979).

Morphological Characteristics of CTV

Citrus tristeza virus has long thread-like, flexuous, filamentous particles about 2000

nm by 11 nm (Bar-Joseph et al. 1979). Virions of CTV consist of one single-stranded

positive-sense RNA molecule encapsidated with two capsid proteins (CPs). The 25 kDa

major CP encapsidates about 95% of the genome, and the remaining portion of the genome is

encapsidated by the 27 kDa minor CP on one end of the virion (Febres et al. 1996). The

virion has helical symmetry with a primary pitch of 3.5 to 3.7 nm and 8.5 tolO CP subunits

per helical turn, respectively (Bar-Joseph., 1972). Long flexuous particles of CTV can easily

be observed with an electron microscope in leaf-dip and partial preparations fromthe phloem

tissue of infected citrus plants (Bar-Joseph et al. 1972; Febres et al. 1996).











Biological Characteristics of CTV

Host range of CTV

Citrus tristeza virus has a narrow natural host range essentially limited to plants in a

single genus, the genus Citrus, in the family Rutaceae. Citrus tristeza virus is able to infect

most species, varieties and hybrids of Citrus (Muller and Gamsey, 1984) as well as some

close relatives of citrus including Aglopsis chevalieri, Afraaegle paniculata, Pamburus

missions (Knorr, 1956) and Aegle marmelos (Muller and Gamsey, 1984). On the other

hand, some citrus relatives, such as Severinia buxifolia, Swinglea glutinosa, Poncirus

trifoliata, and hybrids between P. trifoliata and sweet orange or grapefruit, are resistant or

immune to CTV infection. The virus is not able to replicate or cause symptoms in these hosts

(Gamsey et al. 1987a). To determine the host range and identify a herbaceous experimental

host, CTV has been inoculated to about 200 plant species outside of the family Rutaceae, but

the virus multiplied only in some species ofPassiflora, particularly in Passiflora gracilis

(Muller et al. 1974).

Cvtopathic effects of CTV

Plant viruses usually produce inclusion bodies which are distinctive intracellular

structures of aggregates of virus particles and related proteins found in the virus-infected

cells. Citrus tristeza virus produces two types of inclusion bodies characteristic of members

of the closterovirus group. Inclusion bodies appear as aggregated virus particles in cross-

banded arrays in the phloem and associated cells of the CTV-infected plants. They may be

seen by Azure A staining, in situ immunofluorescence, and light and electron microscopy

(Gamsey et al. 1980; Brlansky, 1987; Brlansky et al. 1988). Recently, it was demonstrated

that a 20 kDa protein, encoded by ORF10 of CTV, is the major component of the characteristic

amorphous inclusion bodies present in CTV infected cells (Gowda et al. 2000). Studies with









8

different isolates ofCTVindicated a positive relationship with the number ofinclusionbodies

to increasing strain severity and virus titer in different host plants. Thus detection of CTV

inclusions has been used as a rapid method for diagnosis and strain differentiation (Brlansky

and Lee, 1990). A second type of inclusion occurring in CTV-infected plants consists of

groups of fibril-containing vesicles that are surrounded by cytoplasmic membranes (Chen et

al. 1971). Similar vesicles are also produced in plants infected with other closteroviruses and

presumably contain double-stranded RNA; thus, they might be involved in virus replication

(Coffin and Coutts, 1993).

Symptoms of CTV

A variety of symptoms is produced in different hosts infected with different CTV

isolates. Symptoms caused by different isolates of CTV are divided into five major groups

including mild vein clearing, seedling yellows (SY), quick decline (QD), stem pitting on

grapefruit (SP-G) and on sweet orange (SP-O). Mild isolates of CTV express weak vein

clearing and flecking on leaves of Mexican lime and produce no noticeable symptoms on

commercial citrus. The SY is expressed as severe chlorosis and dwarfing of sour orange,

lemon or grapefruit. The SY symptoms are easily observed under greenhouse conditions, but

they are not usually seen in the field (Roistacher, 1982). The QD is more severe and can

occur on sweet orange, mandarin and grapefruit scions grafted on sour orange rootstock. The

QD is caused by a virus-induced phloem necrosis in the bark of the rootstock just below the

graft union that prevents the movement of carbohydrates from the canopy to the root system.

Lack of carbohydrates in the root system causes the roots to degenerate and inhibits formation

of new fibrous roots that result in the decline of trees (Garnsey et al. 1987). The stem pitting

symptoms are characterized as stunting, chlorosis, reduced fruit number and size, and typical









9

pitting on stems, major branches and twigs of most citrus scions, especially grapefruit and

sweet orange, regardless of their rootstock. The pits are caused by the collapse of the phloem

tissue and can be observed easily when the bark of twigs is peeled (Da Graca et al. 1984).

A standardized set of indicator plants has been established to determine the biological

characteristics ofCTVisolates. The indicator plants include sour orange for SY, sweet orange

grafted on sour orange for QD, Duncan grapefruit for SP-G, Madam Vinous for SP-O, and

Mexican lime as an universal indicator (Garnsey et al. 1987).

Transmission of CTV

Citrus tristeza virus is readily graft transmitted among the compatible Citrus spp. using

phloem-containing tissue (Roistacher, 1976). The virus was mechanically transmitted to

several hosts by stem-slashinoculationwith inoculum preparations of different levels of purity

(Muller and Garnsey 1984). In nature, CTV is transmitted by aphid vectors including

Toxoptera citricida, Aphis gossypii, A. spiraecola, and T. aurantii (Roistacher and Bar-

Joseph, 1987; Brunt et al. 1990).

Citrus tristeza virus is transmitted in a semi-persistent manner by its aphid vectors in

nature. Efficient transmission of CTV requires 30 min to 24 hrs of acquisition feeding

(Roistacher and Bar-Joseph, 1987), and the aphid retain the ability to transmit for one to three

days (Yokomi and Garsey, 1994). Toxoptera citricida, commonly called the brown citrus

aphid (BCA) was first recognized in 1946 in Brazil, and is the most efficient vector of CTV

(Lee and Rocha-Pena, 1992; Bar-Joseph, 1989). Comparative analysis of transmission

efficiency with different aphid species showed that T citricida transmits CTV with about 16

to 20% efficiency using a single aphid transmission, compared to 0.5 to 1.4% efficiency for

Aphis gossypii (Yokomi et al. 1994). Transmissibility of CTV by its aphid vectors,









10

particularly A. gossypii, is affected by the strain of the virus (Roistacher and Bar-Joseph,

1984), donor and receptor host species and environmental conditions, such as temperature

(Bar-Joseph et al. 1977). Over the years, the BCA has been responsible for natural spread

of CTV in most citrus growing areas including South America (Costa and Grant, 1951), South

Africa (McClean, 1975), Australia, and Asia (Tanaka, 1969). In the 1990s, the BCA moved

northward from South America invading Central America and the Caribbean Basin (Yokomi

et al. 1994; Rocha-Pena et al. 1995) and was reported in Florida in 1995 (Fasulo and

Halbert,. 2000).

Aphis gossypii is the principal vector of CTV in California (Roistacher and Bar-

Joseph, 1984; Roistacher and Bar-Joseph, 1987), Spain (De Mendoza et al. 1984), and Israel

(Raccha et al. 1976) where the BCA is not present. Although A. spiraecola and T. aurantii

are found in Florida (Norman and Grant, 1959), California (Dickson et al. 1951), Israel

(Raccha et al. 1976) and Spain (De Mendoza et al. 1984), they are less efficient vectors and

have little significant effect on the spread of CTV (Roistacher and Bar-Joseph, 1987).

Biochemical Properties of CTV

Citrus tristeza virus is the largest known plant virus, containing about 20 kb of single-

stranded positive-sense RNA genome (Karasev et al. 1995). It has been proposed that the

unusually large genome ofCTVand other closterovirus have evolved from ancestral tobamo-

like viruses by the shuffling ofgenome elements, gene duplication and divergence, acquisition

of novel genes from host cells and other viruses and the development of new gene expression

strategies (Dolja et al. 1994). Besides its large genome size, the genome organization,

number and functions of the genes encoded by the genome, mechanisms of gene expression,

and population complexity and unusual sequence variation among isolates are all important









11

molecular and biochemical aspects of CTV. Their role in the tristeza disease is only now

being investigated and determined (Satyanarayana et al. 1999).

Genome Organization of CTV

The complete genome sequence of four CTV isolates, eachwith a different biological

activity, were determined (Karasev et al. 1995; Mawassi et al,1996; Yang et al. 1999; Vives

et al, 1999). Based on sequence analysis, the genome of CTV is organized into 12 open

reading frames (ORF) potentially encoding 17 protein products, plus 3' and 5' untranslated

regions (UTR). The first ORF (ORF la) encodes a 369-kDa putative polyprotein containing

two papain-like proteases, methyltransferase and helicase domains. The ORFlb codes for

the putative RNA-dependent RNA polymerase (RdRp) which is thought to be expressed by

a +1 frameshift at the 3' terminus ofORFla. (Karasev et al. 1995). The remaining ten ORFs

are individually expressed through sub-genomic mRNAs and encode 33, 6, 65, 61, 27, 25,

18, 13, 20 and 23 kDa protein products from the 5' to 3' direction (Pappu et al. 1994; Karasev

et al. 1995).

The genome ofCTVcan be divided into four modules which are conserved among the

closteroviruses: the core module, the upstream accessory module, the chaperon module and

the CP module. The core module consists of the domains of methyltransferase, helicase and

RdRp, which are all associated with virus replication and conserved throughout the alphavirus

supergroup of the RNA viruses. The upstream accessory module includes two leader papain-

like protease domains probably involved in polyprotein processing. The chaperone module

includes a small protein with membrane-binding domains, one heat shock protein 70 homolog

(HSP70) and one protein distantly related to heat shock protein HSP90. The CP module

consists of the major and the minor CP genes and four additional 3' terminal ORFs.









12

Characteristics and Functions of Genes of CTV

The CTV polyprotein is about 369 kDa and contains two putative proteases, a

methyltransferase and a helicase domain. The protease domains are similar to papain-like thiol

proteases with a cysteine and histidine residue at the catalytic site which are probably

involved in processing of the polyprotein of CTV. Significant similarities in their size and

amino acid sequence suggests that the two proteases of CTV may have evolved by gene

duplication(Karasev et al. 1995). The methyltransferase and helicase domains are generally

required for virus replication and are associated with the RdRp. The methyltransferase and

helicase ofbeetyellows virus (BYV) were detected in infected tissue, and both proteins were

found to be associated with membrane compartments of the infected cells. Detection of these

proteins in the infected tissue demonstrates that they are processed from the polyprotein by a

host or virus encoded protease (Erokhinia et al. 2000).

The putative RdRp of CTVis encoded by ORFlb and is proposed to be expressed by

a +ltranslational ribosomal frameshift at the 3' terminus of ORFla (Karasev et al. 1995).

Deletion analysis of the genome of the infectious CTV clone demonstrated that only the gene

products of the ORFla and lb, which includes protease, methyltransferase, helicase and

RdRp, are required for replication of CTV in Nicotiana benthamiana protoplasts

(Satyanarayana et al. 1999). The p33 gene is not found in other members of the closterovirus

group and has no homology with any known sequences. The p6 gene encodes a small

hydrophobic protein containing sequences similar to membrane-spanning domains and maybe

associated with membranes (Pappu et al. 1994). The p65 gene contains a conserved ATPase

domain and shows high homology with the cellular heat shock protein HSP70. The p61 gene

of CTV and its analogs in some, but not all, closteroviruses has conserved sequences with









13

limited homology to another cellular chaperone HSP90 (Pappu et al. 1994). The HSP70s are

ubiquitous molecular chaperone-like proteins and participate in a wide range of cellular

processes including protein folding, assembly, translocation and intercellular transport (Feder

and Hofmann, 1999). Recent studies revealed the role and importance of the HSP homolog

proteins in CTV and other closteroviruses. The analog of CTV p65 gene in BYV was able to

complement the movement proteins of other viruses, and it is involved in cell-to-cell

movement of BYV (Agranovsky, 1999). In addition, analogs of CTV p65 and p61 proteins

were found to be physically associated with virion particles in other closteroviruses (Tian

et al. 1999; Napuli et al. 2000). Furthermore, itwas demonstrated for CTV that the p65 and

the p61 proteins are required for assembly of the CTV virions. Deletion of these genes

significantly reduced the formation and passage of full-length CTV virions (Satyanarayana,

2000).

The 25 kDa protein encoded by the ORF7 is the major capsid protein (CP) (Sekiya et

al. 1991) which encapsidates most of the CTVgenome. The 27 kDa protein identified as minor

capsid protein is a diverged copy of the major CP (Febres et al. 1994) and coats only 5% of

the genome at one terminus (Febres et al. 1996). Identification of a conserved epitope in most

CPs of severe isolates of CTV, but not in mild isolates, indicated that CTV may have

pathogenicity determinants associated with the major CP gene (Pappu et al. 1993). The 20

kDa protein shows high affinity for itself and accumulates in the characteristic amorphous

inclusion bodies present in CTV infected phloem cells. Thus, this gene is involved in the

formation of inclusion bodies in the infected cells (Gowda et al. 2000). The 23 kDa protein,

encoded by the 3' terminal gene, is localized predominantly in the cytoplasm of CTV-infected

cells (Pappu et al. 1997). Recently the recombinant 23 kDa protein was demonstrated to be









14

able to bind both single and double-stranded RNA in a non-sequence specific manner,

suggesting that the 23 kDa protein of CTV is an RNA-binding protein. The RNA-binding

activity was mapped to a region containing a cluster of positively-charged amino acids and

sequences similar to putative zinc-finger domains (Lopez et al. 2000). The rest of the genome

has no homology with any known sequences in the databases (Pappu et al. 1994; Karasev et

al. 1995).

Replication of CTV

Replication of the genome is the fundamental aspect in the biology of all viruses. The

replication of positive-stranded RNA viruses takes place in two main stages. First, the

negative or complementary-strand RNA is synthesized from the genomic positive-strand RNA

template, and then the negative-strand of RNA is used as a template to produce positive-

strand RNA progeny. Studies of replication of a number of animal and plant RNA viruses

showed that virus replication requires several viral encoded proteins including RdRp,

helicase and MT, cis-acting viral sequences, such as the 3' and 5' terminal sequences and

internal sequences, as well as host proteins (Buck, 1996).

The large genome size, low titer and restriction to phloem tissue, and the lack of an

herbaceous experimental host have greatly hindred genetic analysis and understanding of the

replication of CTV. Recent development of a full-length infectious cDNA clone of CTV

(Satyanarayana et al.1999) and protoplast systems for CTV replication (Price et al. 1996;

Navas-Castillo et al. 1997) have facilitated the genetic analysis of the CTV genome. Deletion

analysis of the infectious cDNA clone of CTV demonstrated that only ORF a and lb, coding

for the replication-associated proteins, as well as the 3' and 5' UTRs of the genome were

essential for replication of CTV in N. benthamiana protoplasts. A smaller CTV replicon









15

with a 10 kb genome, which contains only the ORFla\b and the 3' and 5' UTRs generated by

deletion often 3' ORFs replicated more efficiently than the full-genomic RNA in tobacco

protoplasts (Satyanarayana et al. 1999). Analysis of the 3' and 5' UTR sequences from

different isolates showed that the 3' UTR is highly conserved, whereas the 5' UTR sequences

were highlyvariable among different isolates of CTV. However, the presence of a conserved

secondary structure was predicted in the 5' UTR of all CTV isolates. This structure consists

of two stem loops, and it may function as a cis-acting element during replication (Lopez et al.

1998). To determine the effect of UTRs for replication, the UTRs of the CTV replicon were

substituted with the 3' and 5' UTRs from different isolates. Different 3' UTRs still enabled

replication, but the replication efficiency of the replicon with heterologous 5' UTRs was

significantly reduced. The reduction was proportional to the extent of sequence variability

among the 5' UTRs indicating that, in addition to the secondary structure, the primary structure

ofthe 5' UTR is also important for CTV replication (Satyanarayana et al. 1999). During CTV

replication, different forms of double-stranded RNAs and a number ofsubgenomic (sg) and

defective RNAs are produced in addition to the full-length replicative genomic RNA.

Defective RNAs Associated with CTV

Defective RNAs (D-RNAs) are non-autonomous RNA molecules derived mainly or

entirely from genomic nucleotide sequences of the virus and are associated with certain

viruses. The D-RNAs have been found to be associated with a number of animal viruses

(Huang, 1993), and some positive-stranded RNA plant virus groups, such as tombusviruses,

carmoviruses and potexviruses (Simon and Bujarski, 1994). Recently, a number of defective

D-RNAs ranging from 1.5 tolO kb were detected in CTV infected plants. These D-RNAs

show significant sequence homology with different portions of both the 3' and 5' ends of the









16

CTV genome, indicating that D-RNAs were formed by extensive internal deletion from the

genomic RNA of CTV (Mawassi et al. 1995; Yang et al. 1997; Ayllon et al. 1999; Yang et

al. 1999). Two different mechanisms of recombination were proposed for the generation of

D-RNAs detected in CTV-infected samples. It was found that a substantial number of D-

RNAs had variable 5' ends, but a common 3' terminus containing the complete sequence of

the sgRNA for the ORF 11. Based on this observation and the presence of an extra cytosine

nucleotide at the junction site, it was proposed that these D-RNAs were generated by "minus

strand jumping" during virus replication (Yang et al. 1997). Finding of the tetra-nucleotide

AAGC direct repeat located in the flanking or in the vicinity of the junction sites of some D-

RNAs suggested that these D-RNAs were generated by replication-driven template switching

during virus replication (Ayllon et al. 1999). Since D-RNAs lack the genes necessary for

replication, they are non-autonomous and require a helper virus for their replication in the

cells. Homologous and heterologous D-RNAs of different CTV isolates were replicated in

trans by the helper virus in the naturally infected plants and in tobacco protoplasts (Mawassi

et al. 2000). The presence of D-RNAs in the virus-infected plant may interfere with virus

replication and increase or decrease symptom severity. It was suggested that D-RNAs of 4.5

and 5.1 kb were associated with reduction of SY symptoms in some CTV isolates (Yang et

al. 1999). However, there is no conclusive data for the involvement of D-RNA in symptom

modulation in the CTV infected plants.

CTV Gene Expression Strategies

Based on phylogenetic analysis of sequences of the RdRp and other replication-

associated proteins, the CTV strategy is similar to the alpha-like supergroup of positive-

stranded RNA viruses (Koonin and Dolja, 1993). On the other hand, the organization and

expression of the CTV genome are more similar to the coronaviruses. Both coronaviruses and









17

CTV use several different gene expression strategies including polyprotein processing,

translational frameshifting, and subgenomic RNAs for expression of their large genomes.

Polyprotein processing

The ORF la is expressed as a 349 kDa polyprotein containing two putative proteases,

a methyltransferase and a helicase domain. In addition, the RdRp of CTV encoded by ORF lb

is thought to be expressed by a +ltranslational ribosomal frameshift near the 3' terminus of

ORF 1 a resulting in a 400 kDa polyprotein (Karasev et al. 1995). Individual domains become

active proteins when they are released from the polyprotein by protease cleavage. Based on

sequence analysis, the CTV genome encodes two papain-like proteases presumably for

processing of the polyprotein. Two glycine residues located atposition484-485 and 976-977

from the N terminus of the polyprotein are predicted cleavage sites for these proteases. Thus

processing of the 349 kDa polyprotein would produce two leader proteins of 54 and 55 kDa

and a 240 kDa protein containing the MT and helicase domains. Detection of the product of

both the MT and helicase domains of BYV as individual proteins in the infected tissue

indicates thatthe polyprotein of BYV, and possibly CTV, are further processed by either viral

or host proteases (Erokhinia et al. 2000).

Subgenomic RNA

Some viruses produce a set of subgenomic RNAs for translation of internal ORFs in

their genome. Several sgRNAs were detected in CTV-infected tissue. Characterization of the

CTV sgRNAs by Northern blot hybridization with gene-specific cDNA probes revealed the

presence of a series of nine 3' co-terminal sgRNAs responsible for translation of ORFs 2 to

10, located at the 3' half of the genome. The pattern of 3' co-terminal sgRNAs was not affected

by different citrus hosts (Hilfet al. 1995) and an identical pattern of sgRNA also was reported









18

inN. bentamiana protoplasts infected with CTV (Navas-Castillo et al. 1997). The relative

amounts of the different sgRNAs were variable, with the sgRNAs for the p20 and p23 ORFs

being the most abundant in both infected citrus tissue and tobacco protoplasts (Hilfet al. 1995;

Navas-Castillo et al. 1997). Analysis of the 5' terminus of the two most abundant sgRNAs

demonstrated that the p20 and p23 specific sgRNA have 48 and 38 nucleotide long 5' UTRs,

respectively. The sequence of the 5' UTRs was colinear with the downstream sequence of

their respective ORFs in the genome, indicating thatthe structure of sgRNAs ofCTVis similar

to that of alpha-like viruses that produce sgRNAs with different 5' UTRs colinear to their

corresponding genomic sequences. Thus, they are different from coronoviruses, which

produce sgRNAs with identical 5' UTRs derived from the 5' end of their genomic RNAs

(Karasev et al. 1997).

Translational frameshifting

Ribosomal frameshifting is a directed change in the translational reading frame that

allows production of a single protein from two overlapping genes. The frameshift can occur

in either the 5' (+1 frameshift) or the 3' (-1 frameshift) direction (Brierley, 1995). A number

of viruses and mobile genetic elements use the ribosomal frameshifting mechanismto control

expression of their replicase gene at the translational level. This mechanism enables

controlled low-level synthesis of the polymerase, which is needed in only small quantities

(Brault and Miller, 1992). The -1 frameshift was demonstrated for animal retro- and

coronaviruses (Brierley, 1995), yeast double-stranded RNA viruses (Dinman and Wicker,

1994), as well as in the plant luteovirus, sobemovirus, carlavirus, enamovirus and

dianthovirus groups (Maia et al. 1996). The +1 frameshifting has been described in a few









19

organisms including yeast transposon TY, the copia-like element of Drosophila (Farabaugh,

1997), and it has been proposed for plant closteroviruses (Agranovsky et al. 1994).

It has been proposed that the putative CTVRdRp gene is expressed via +1 frameshift

in the 3' terminus of ORF a (Karasev et al. 1995). Based on sequence analysis of CTV and

BYV, the frameshift is predicted to take place in CTV at nucleotide 9405, located in the

overlap region of ORFla and lb. Although the overlap region of ORFla and lb is relatively

conserved, there are significant differences between the BYV and CTV sequences around the

predicted frameshift site. A UGA stop codon, a GGGUUU slippery sequence and a stem-loop

structure were found in the predicted frameshift site of BYV, and they are proposed to be

involved in the +1 frameshift of BYV (Agranovsky et al. 1994). The absence of these

elements in CTV sequences indicated that the mechanism of the +1 frameshift differs between

BYV and CTV. It was suggested that the frameshift occurs at the rare arginine codon CGG of

CTV, which aligns with the UAG stop codon in BYV and then serves as a stop codon to pause

the ribosome during translation, thereby inducing the +lframeshift in CTV (Karasev et al.

1995).

Sequence Variation Among CTV Isolates

Citrus tristeza virus isolates consist of unusually complex populations of distinct

genotypes possibly contributed by the perennial nature and vegetative propagation of host

plants, multiple aphid transmissions and genetic properties of the virus such as recombination

and formation of D-RNAs. The complexity of CTV populations in the infected plants

presents problems for diagnosis and strain identification, and also for understanding the

mechanisms of pathogenesis and symptom induction in different host plants.









20

The complete genomic sequences were published for four biologically and

geographically different isolates of CTV. These include T36, a quick decline isolate from

Florida (Karasev et al. 1995); VT, a stem pitting and seedling yellows isolate from Israel

(Mawassi et al. 1996); SY568, a stem pitting and seedling yellows isolate from California

(Yang et al.1999); and T385, a mild isolate from Spain (Vives et al. 1999). The genome

organization of all isolates was identical, and their genome lengths were very similar, ranging

from 19,226 to 19,296 nucleotides. However, these isolates showed varying degrees of and

unusual distribution of sequence diversity in their genomes.

Comparative analysis of genomic sequences of the CTV isolates revealed an unusual

asymmetric sequence similarity along the T36 genome. The 3' halves of the genomes were

highly conserved with over 90% sequence identity, but the 5' halves of the genomes had as

low as 70% identity (Mawassi et al. 1996; Dodds et al.1997; Vives et al. 1999 ). The

genomes of VT and T385 showed about 90% sequence identity evenly distributed throughout

the genome. The genome of SY568 showed about 90% sequence identity toward both the 3'

and 5' end of T385 genome, but the central parts of two genomes were more than 99%

identical. From this data, it was inferred that the central portion of the SY568 genome results

from RNA recombination between two CTV genomes (Vives et al. 1999). Analysis of the 3'

and 5' UTR sequences from the genomic and partial sequences showed that the 3' UTR is

highly conserved with above 97% identity among different isolates of CTV, but the 5' UTRs

were highly variable, with as low as 42% sequence identity (Lopez et al. 1998;Vives et al.

1999). Using the variable 5' UTR sequences, CTVisolates were classified into three different

groups containing mild isolates in one group (III) and severe isolates in the other two groups

(I and H) (Lopez et al. 1998). Based on comparisons of representative genomic sequences









21

from the 5' and 3' regions of CTV isolates, formation of two groups of CTV isolates

represented by VT and T36 was proposed. The proposed VT group contained isolates having

relatively constant and evenly distributed genomic sequence divergence, and the T36 group

consists of isolates having highly diverged 5' genomic sequences (Hilf et al. 1999).

Sequence variability among CTV isolates has been studied extensively using CP gene

sequences. Comparison ofthe CP sequences from number ofbiologically and geographically

different isolates showed that the CP gene was highly conserved. Phylogenetic analysis of CP

sequences generated distinct groups of CTV isolates with similar biological activities such

as mild, quick decline and stem-pitting isolates (Pappu et al. 1993; Cevik et al. 1995). In

addition, minor sequence differences were found in CP genes from different CTV isolates.

Since these minor differences were conserved in a group of isolates with a specific biological

activity, they may be associated with those biological characteristics of the CTV isolates

(Cevik et al. 1996). Beside sequencing, several other methods including reverse

transcriptase-polymerase chain reaction (RT-PCR) (Cevik et al. 1995), restriction fragment

length polymorphism (RFLP) (Gillings et al. 1993), single-stranded conformation

polymorphism (SSCP) (Febres, 1995; Rubio et al 1996), and hybridization with group or

strain-specific probes (SSP) (Cevik et al. 1996) have been used to study the sequence

variation in the CP and several other 3' genes of different CTV isolates.

Control of CTV

Several strategies have been developed for managing the tristeza disease. Different

strategies are available for use based on the absence or presence of CTV in different citrus

growing areas. Quarantine, budwood certification and/or clean stock programs are used to

prevent introduction of CTV into countries where CTV is absent. In some regions where CTV









22

incidence is low, the disease can be managed by eradication or suppression programs, when

combined with use of certification and clean stock programs. Use of CTV tolerant rootstocks,

mild strain cross protection, and, in the future, genetically engineered resistance, combined

with certification and clean stock programs are the potential control measures for those

regions where CTV has become endemic (Lee and Rocha-Pena, 1992).

Mild Strain Cross Protection

Cross protection is a phenomenon in which a plant systemically infected with a mild

strain of a virus is protected from subsequent infection or expression of symptoms of severe

strains of the same virus, or closely related viruses (Fulton, 1986). Mild strain cross

protection is the only available management strategy that canbe implemented immediately to

control CTV in Florida and other areas where such mild strains have been collected and

evaluated. Cross protection has been used to control CTV on a large scale in commercial

citrus plantations, particularly with Pera sweet orange in Brazil (Costa and Muller, 1980),

grapefruit in Australia (Barkley et al. 1990), South Africa, Japan, Citrus hystrix in Reunion

and limes in India (Lee and Rocha-Pena, 1992). Cross protection is now being applied in

other citrus growing areas where CTV has become endemic, such as Florida (Lee et al. 1987;

Lee and Brlansky, 1990). Mild strain cross protection is an effective control strategy for CTV

in situations where the disease is endemic and impossible to control by eradication or

suppression, and when cross protecting mild strains are available which are mild in all citrus

cultivars in the region (Costa and Muller, 1980; Lee et al. 1987).

Use of CTV Tolerant Rootstocks

Most commercial citrus varieties and Citrus relatives are susceptible to CTV.

However, some rootstocks are tolerant to CTV, meaning that the virus replicates in the host,









23

but no or minor disease symptoms are expressed in the infected plant. Some Citrus sp.

generally used as rootstocks including C. reticulata, C. volkameriana and C. jambhiri

(Rangpur lime) are naturally tolerant to QD-inducing isolates of CTV. Other hybrid

rootstocks including citranges (Citrus sinensis X Poncirus trifoliata) and citrumelos (C.

paradisi X P. trifoliata) have been developed and are used as CTV tolerant rootstocks to

control CTV in some citrus growing regions. The presence of other economically important

diseases, such as citrus blight, Phytopthora sp. and viroids, and undesirable horticultural

characteristics limit the usefulness of these rootstocks in some citrus growing areas (Garnsey

et al. 1987; Davies and Albrigo, 1994). In addition, some CTV isolates are able to induce

the SP symptoms in the scions regardless of the tolerance of their rootstocks (Bar-Joseph and

Lee, 1989). Thus, the CTV tolerant rootstocks do not give control against CTV-SP isolates

in citrus growing areas where these isolates are widespread.

Breeding for CTV Resistance

There is no genetic resistance in the genus Citrus available that's effective against all

CTV isolates. However, some citrus relatives, such as Hesperethusa, Luvunga, Merope

Oxantherea,, Severina, Swinglea and Poncirus, are immune to CTV, meaning that the virus

does not replicate in plants fromthese genera (Garnsey, 1987). All of these citrus relatives,

exceptPoncirus, are sexually incompatible with Citrus. Thus, it is not possible to incorporate

the CTVresistance into most commercially desirable citrus varieties by conventional breeding

techniques. Genetic studies revealed that resistance found in P. trifoliata, which is sexually

compatible with citrus, is controlled by a single dominant gene that prevents virus replication

in the plant by an undetermined mechanism (Yoshida, 1996). Poncirus trifoliata has been

crossed with different citrus species to introgress this CTV resistance into desirable citrus









24

cultivars. Although several CTV-tolerant rootstocks, such as citranges and citrumelos were

developed as a result of these crosses, no scion variety with the CTV resistance and

acceptable horticultural characteristics has been produced. The integration of the CTV

resistance gene into commercially important citrus by classical breeding is difficult because

of the problems associated with citrus breeding including large plant size, long juvenility

period, polyembryony, heterozygosity, sterility, self- and cross-incompatibility and inbreeding

depression (Soost and Roose, 1996). The CTV resistance gene in P. trifoliata, designated

Ctv, has been mapped recently using molecular markers (Gmitter et al. 1996; Deng et al.

1997). Mapping the Ctv gene and identification of markers closely linked with the gene is

useful for marker-assisted rapid selection of hybrids resistant to CTV and map-based cloning

of the Ctv gene in the future. The cloned resistance gene can be introduced into the citrus

genome by genetic transformation, which now offers a more efficient approach for

development of CTV-resistant citrus cultivars.

Genetic Engineering for Virus Resistance

Recent developments in molecular biology and plant transformation techniques

provide new approaches and open new possibilities for the generation and evaluation of

sources of virus resistance outside of conventional breeding methods. Genetic engineering

allows the development oftransgenic plants with a defined trait by incorporating a specific

gene into the plants genome without altering the other desirable characteristics. Virus

resistance has been engineered by transforming plants with genes or sequences from viruses

and/or other sources (Fuchs and Gonsalves., 1997). Virus resistance was engineered by

expressing a number of genes from a variety of organisms including genes encoding anti-viral

proteins such as ribosome inactivating proteins from pokeweed Phytolacca americana









25

(Lodge et al. 1993), rat 2'-5' oligoadenylate synthetase (Truve et al. 1993), human 2-5A

synthetase and a 2-5A-dependent RNase L (Mitra et al. 1996; Ogawa et al. 1996), virus

specific antibodies "plantibodies" (Voss et al. 1995), synthetic peptides (Marcos et al. 1995)

defense related compounds (Herbers et al. 1996) and ribozymes (Nakamura et al. 1995).

Although expression of these genes resulted in resistance to a specific virus or in some cases

a diverse groups of plant viruses in the transgenic plants, the majority oftransgenic crop plants

engineered for virus resistance has been developed using sequences derived from plant viral

genomes.

Pathogen-derived resistance

The pathogen-derived resistance (PDR) concept was first developed as a strategy to

produce resistance against a specific or range of pathogens by transforming host cells with

entire genes or sequences derived fromthe pathogen genome. It was proposed that expression

of genes in an inappropriate amount, form, time, or location in the cell may interfere with the

normal life cycle of the pathogen in their host and thereby induce resistance against the

pathogen (Sanford and Johnson, 1985). Pathogen-derived resistance for a plant virus was first

developed in 1986 (Powell-Abel et al. 1986) by demonstrating that tobacco plants

transformed with the coat protein (CP) gene of tobacco mosaic virus (TMV) showed

resistance to TMV infection. Since then, full-length, truncated and untranslatable constructs

of CP gene sequences from a number of plant viruse have been transformed into many

different plant species to engineer CP-mediated (CPM) resistance against viruses. Coat

protein-mediated resistance has been successfully developed against a number of positive-

stranded RNA viruses in the genera Potyvirus, Cucumovirus, Ilarvirus, Tobravirus,

Potexvirus, Tobamovirus, Tobravirus, Carlavirus, and Luteovirus in many crop plants









26

including vegetables, fruits, cereals and forage crops (Baulcombe, 1994; Beachy, 1994;

Hackland et al. 1994; Pappu et al. 1995; Fuchs and Gonsalves, 1997). Morever, several

transgenic plants engineered for CPM resistance have already been commercialized, and many

more have been evaluated in field trials for release (Fuchs and Gonsalves, 1997).

Use of viral sequences other than CP genes has been explored to engineer pathogen-

derived resistance to viruses in plants. Plants were transformed with non-coding sequences

from the 5' and 3' UTR of viral genomes (Nelson et al. 1993; Zaccomer et al. 1993) as well

as satellite RNAs (Harrison et al. 1987) and D-RNAs (Kollar et al. 1993) to produce

transgenic plants resistant to viruses. In addition, nonstructural genes encoding the protease

(Maiti et al. 1993; Vardi et al. 1993), the cell-to-cell movement protein (Malyshenko et al.

1993) and the replicase (Carr and Zaitlin., 1993; Palukaitis and Zaitlin, 1997) have been used

for PDR against a number of viruses in many crops. Although the results have been somewhat

variable with different plant-virus systems, the use of non-structural genes, especially

movement protein and replication associated proteins such as RdRp, is a promising strategy

for developing virus resistance in transgenic plants (Beachy 1994; Pappu et al. 1995;

Palukaitis and Zaitlin, 1997).

Replicase-mediated resistance

Virus resistance induced by expression of native or modified forms of replication-

associated genes, such as the RdRp RNA polymerase of viruses in plants, is called replicase-

mediated resistance (RMR). The first RMR was developed against bacteriophage Q, where

expression of the modified replicase and a replicase binding site in the host generated bacteria

which were resistant to the bacteriophage Q (Inokuchi and Hirasima, 1987). Replicase

mediated resistance to plant viruses was first reported by Golemboski et al. (1990) who









27

transformed tobacco plants with the non-structural gene of TMV encoding the 54 kDa protein

to determine the function ofthis protein. Transgenic plants expressing the 54 kDa protein were

highly resistant to TMV infection. Since this first report of resistance to TMV in tobacco,

RMR has been extensively explored for a number other plant RNA viruses using full-length

and defective constructs of replicase genes. Defective constructs contain truncated replicase,

lacking either the 3' or 5' terminus, and the conserved GDD motif of the RdRp domain as well

as mutants constructed with one or two amino acid changes in the conserved GDD motif.

Replicase-mediated resistance has been successfully developed for a number of viruses from

ten different genera including Alfamovirus, Bromovirus, Comovirus, Cucumovirus,

Luteovirus, Potexvirus, Potyvirus, Tobamaovirus, Tombusvirus and Tobravirus (Palukaitis

and Zaitlin 1997).

In the first example ofreplicase-mediated resistance to TMV induced by the 54 kDa

protein, plants expressing the 54 kDa protein were highly resistant to infection with virions

and RNA from the Ul strain of TMV from which the sequence was derived (Golemboski et

al. 1990). The resistance to TMV infection was observed in both transgenic plants and

protoplasts derived from transgenic plants. Since the resistance was expressed atthe single

cell level in protoplasts, it was suggested that the resistance was mainly due to interference

with viral replication (Carr and Zaitlin, 1991). It was reported that the cell-to-cell movement

of the TMV RNA was also suppressed in the inoculated leaves of the resistant plants

(Nguyen et al. 1996). Tobacco plants were also transformed with full-length replicase genes

encoding the 126 and 183 kDa proteins and a mutant gene for the 126 kDa protein with a

bacterial transposable element to develop RMR to TMV. Plants expressing the full-length 126

and 183 kDa proteins did not show resistance to TMV. However, plants expressing the mutant









28

126 kDa protein were resistant to TMV and several other tobamoviruses (Donson et al. 1993).

N. benthamiana plants were transformed with both a wild-type and a truncated replicase gene

encoding the 54kDa of another tobamovirus, pepper mild mottle virus (PMMV). Plants

expressing wild-type 54 kDa showed two types of resistance response to PMMV infection.

Some of the plants were susceptible to PMMV infection, but they were able to recover from

the PMMV infection later, indicating a delayed induced resistance. Other transgenic plants

showed a complete resistance to PMMV from the beginning of the infection, and no symptoms

were observed onthem, indicating a pre-established resistance (Tenllado et al. 1995). Onthe

other hand, plants expressing the truncated 54 kDa protein were either highly resistant or

susceptible to virus infection, and they did not show delayed induced resistance (Tenllado et

al. 1996).

The replicase gene from cucumber mosaic virus (CMV) was mutated at the GDD motif

by a deletion of 94 nucleotides or a truncation at C- terminus by a frameshift mutation. N.

tabacum plants transformed with this replicase construct showed absolute resistance to high

concentrations ofvirions and RNA from the homologous Fny strains of CMV (Anderson et al.

1992). The transgenic plants show resistance to strains from subgroup I, which are closely

related to the Fny strain of CMV, and they were mostly susceptible or only partially resistant

to some strains from CMV subgroups II (Zaitlin et al. 1994). The resistance induced by

defective CMV replicase operated at the single cell-level (Carr et al. 1994) to inhibit virus

replication and restrict cell-to-cell and long distance movement of the virus in transgenic

tobacco (Carr et al. 1994; Hellwald and Palukaitis, 1995; Nguyen et al. 1996) and tomato

plants (Gal-On et al. 1998). Although transgenic plants showed very limited cell-to-cell virus

movement compared to nontransgenic plants, long distance movement of CMV was









29

completely inhibited by blocking the entry of the virus into the vascular system (Wintermantel

et al. 1997). Analysis of transgenic plants containing the translatable and non-translatable

replicase gene of CMV and examination of steady-state mRNA levels suggested that

translatability of the transgene increases the effectiveness ofreplicase-mediated resistance to

CMV (Wintermantel and Zaitlin, 2000).

Different forms of the alfalfa mosaic virus (A1MV) P2 replicase gene, including full-

length and N-terminally truncated and modified constructs in which the conserved GDD motif

was mutated to GGD, GVD VDD or DDD, were used to engineer resistance to A1MV in

tobacco (Brederode etal. 1995). No resistance was observed in transgenic plants expressing

the full length, truncated and VDD mutants, however, complete or partial resistance to A1MV

was achieved by the expression of the GGD, GVD, and DDD mutant replicase gene in

tobacco. Analysis of transgenic plants showed that resistance was associated with the high

level expression of mutant replicases (Brederode et al. 1995). In contrast, transgenic plants

expressing full-length functional replicase fromAlMVwere not only fully susceptible to virus

infection, but they were also capable of complementing replication of the mutant virus lacking

its replicase gene (Taschner et al. 1991).

Replicase-mediated resistance was developed to three different potyviruses using

different constructs of nuclear inclusion b (NIb) from potato virus Y (PVY), plum pox virus

(PPV) and pea seed-born mosaic virus (PsbMV). N. tabacum plants transformed with full-

length NIb gene and 3' and 5' truncted NIb genes conferred resistance to PVY infection, but

the plants expressing NIb gene lacking the GDD motif did not show resistance to PVY. The

resistance was specific to the strain fromwhichthe sequences were obtained, and plants were

susceptible to even very closely related stains of PVY (Audy et al. 1993). N. benthamiana









30

plants expressing mutant NIb genes in which the GDD motif was changed to ADD or VDD

were resistant to PPV infection. Plants carrying the NIb gene with the VDD mutation showed

a delayed resistance response characterized by total recovery from the initial infection with

PPV (Guo and Garcia., 1997). A similar type of resistance response was observed in

transgenic pea plants expressing full-length NIb from PsbMV (Jones et al. 1998).

Transgenic tobacco plants with full-length and 5' truncated 166 kDa replicase of

potato virus X (PVX) exhibited resistance to PVX, but plants expressing partial sequences of

this gene encoding the nucleotide binding domain or the GDD motifwere not resistant to PVX

( Braun and Hemenway, 1992). The mutant replicase genes of PVX with single amino acid

mutations in the GDD motif, GED, GAD or ADD as well as untranslatable sequence of the

replicase gene did induce resistance to PVX in transgenic tobacco plants (Longstaff et al.

1993; Mueller et al. 1995). The transgenic plants resistant to PVX showed resistance only to

a specific strain, and also low transgene RNA accumulation (Mueller et al. 1995).

A variable level of resistance has been achieved by expressing full-length replicase

genes of several viruses including pea early browning virus (PEBV) (MacFarlane and

Davies), cymbidiumringspot virus (CymRSV) (Rubino et al. 1993;Rubino and Russo, 1995),

cowpea mosaic virus (CPMV) (Sijen et al. 1995), bromo mosaic virus (BMV) (Kaido et al.

1995), potato leaf roll virus (PLRV) (Kaniewski et al. 1995), and rice tungro bacilliform

virus (RTBV) (Huet et al. 1999).

Mechanisms of replicase-mediated resistance

Virus resistance observed in transgenic plants expressing intact or modified viral

replicase genes can be grouped into two broad categories. In the first category, the resistance

is mediated by functional or mutant dysfunctional proteins which interfere with the replicase









31

enzyme complex and disrupt the viral replication. In the second category, expression of the

protein is not required for resistance, and the resistance is mediated by the transgene RNA.

Recent reports on two extensively studied RMR, TMV and CMV, showed that the resistance

was contributed by both protein and RNA-mediated resistance mechanisms (Goregaoker et

al. 2000; Wintermantel and Zaitlin, 2000). These findings showed that RMR may be complex

in some systems, and that more than one factor can be responsible for the resistance

phenotypes observed in transgenic plants.

Protein-mediated resistance

Protein mediated resistance (PMR) is characterized by a direct correlation between

transgene protein and the degree of resistance. This type of resistance is usually effective

against a broad spectrum of strains of the same virus, and even against some related viruses.

Functional replicase expressed in a transgenic plant induces resistance to that virus if the

level, time, and location of the expression in the cell interferes with replication, assembly, or

movement of the virus (Carr and Zaitlin, 1993). Plants expressing the 54 kDa protein

demonstrated a resistance response that is, in part, similar to this type of protein-mediated

mechanism (Golemboski et al. 1990; Goregaoker et al. 2000). Functional inactivation of a

gene by expression of a mutant form of the same gene is termed trans-dominant or negative

dominant mutation. This inactivation can be achieved by competing for a factor, substrate, or

a binding site or by formation of a non-functional complex of mutant and wildtype proteins

(Herskowitz, 1987). It has been suggested that defective replicase expressed in transgenic

plants may interfere with viral RNA replication by depleting host factors, or saturating the

specific binding sites, or by forming a non-functional replication complex with replication-

associated proteins of the virus (Carr and Zaitlin, 1993). Transgenic plants expressing









32

defective or mutant replicase of A1MV (Brederode et al. 1995), CMV (Carr et al. 1994;

Wintermantel and Zaitlin, 2000) and TMV (Donson et al. 1993) all displayed resistance

mediated by the mutant protein.

RNA-mediated resistance

Resistance responses determined by the transgene RNA transcript rather than the

expression of the transgene protein product are termed RNA-mediated resistance. This type

of resistance is characterized by detection of low steady-state levels of the transgene

transcript in the transgenic plants and was first reported in tobacco plants transformed with

the CP gene of tobacco etch virus (Lindbo and Dougherty, 1992). Similar RNA-mediated

resistance was later reported in transgenic plants expressing mutant, translatable, and

untranslatable replicase genes of PVX (Muller et al. 1995), PMMV (Tenllado et al. 1996),

PPV (Guo and Garcia., 1997), and PsbMV( (Jones et al. 1998). Expression of the protein is

not required for RNA-mediated resistance, and in most cases ofreplicase-mediated resistance,

the protein product of the transgene could not be detected in transgenic plants, thereby

indicating that most replicase-mediated resistances are RNA-mediated (Baulcombe, 1995;

Lommossoff, 1995; Baulcombe, 1996). Some transgenic plants display an instant resistance

to virus infection; however, some transgenic lines display a recovery phenotype, in which an

initial systemic viral infection was established, with the plants subsequently recovering from

the initial infection and becoming completely resistant to virus infection (Lindbo, et al. 1993,

Muller et al. 1995; Tenllado et al. 1995 Guo and Garcia., 1997).

The RNA-mediated and replicase-mediated resistances were shown to be very

specific and effective only against the specific strain of the virus from which the transgene

sequences were obtained or against closely related strains of the same virus with a high degree









33

of sequence homology (Audy et al. 1993; Zaitlin et al. 1994; Palukaitis and Zaitlin, 1997).

Since this resistance is dependent upon extreme sequence similarity between the mRNAof the

transgene and the inoculated virus, it is also called homology dependent resistance. Genetic

analysis of RNA-mediated resistance in transgenic plants demonstrated that RNA-mediated

resistance was related to post-transcriptional gene silencing (PTGS) in plants (Smith et al.

1994; English et al. 1996; Goodwin et al. 1996). Post-transcriptional gene silencing involves

the reduction or suppression of gene expression by sequence specific degradation of the

transgene mRNA (Stam et al. 1997). Although the actual mechanism of PTGS is not known,

several models have been proposed based on extensive studies of PTGS in many systems.

These include the RNA threshold model (Lindbo et al. 1993; Dougherty and Parks, 1995) and

the aberrant RNA model (English et al. 1996;Wassengger and Pelissier, 1998). The

observation of a virus induced resistance mechanism in non-transgenic plants which acts

similarly to PTGS (Ratcliffet al. 1997; Ratcliff et al. 1999) and suppression of PTGS by viral

pathogenicity determinants in silenced transgenic plants (Anandalakshmi et al. 1998;Brigneti

et al. 1998) have suggested that PTGS is a natural defense mechanism of plants which was

developed for protection against viruses (Ratcliff et al. 1999).

Genetic Engineering of Citrus

Genetic engineering is a promising approach for genetic improvement of woody plants,

such as citrus, because their improvement by conventional breeding has been considerably

limited by long juvenility period, heterozygosity, sterility, self- and cross-incompatibility.

Currently commercialized Citrus genotypes can be made even more desirable by transforming

them with specific genes from different sources to add new traits. As with other plants,

genetic engineering of citrus requires the availability of genes for agriculturally important









34

traits and an efficient genetic transformation method to integrate the desired genes into the

Citrus genome.

Development of an efficient genetic transformation technique has been attempted in

different Citrus species using a variety of plant transformation methods. The first report of

citrus transformation was the transformation of protoplasts of Trovita sweet orange by the

direct DNA uptake method using polyethylene glycol (PEG) (Kobayashi andUchimiya., 1989).

Later, genetic transformation of protoplasts from rough lemon was also achieved using PEG-

mediated direct DNA uptake (Vardi et al. 1990), and protoplasts from C. reticulata were

successfully transformed using electroporation (Hidaka and Omura, 1993). To develop

alternative methods for direct DNA uptake, suspension cell cultures from sweet orange calli

were transformed by co-cultivation with Agrobacterium tumefaciens (Hidaka et al. 1990),

and particle bombardment was used for transformation of embryogenic cells of tangelos,

resulting in transgenic embryos (Yao, 1997). Thus, transgenic citrus plants have been obtained

using PEG and electroporation-mediated direct DNA uptake, Agrobacterium-mediated and

biolistic transformation of protoplasts, and cell cultures from embryogenic callus. However,

these methods were very inefficient because of preparation of protoplasts or cell suspension

cultures (Gmitter et al. 1992), and the regeneration oftransgenic plants was difficult for most

citrus species.

An Agrobacterium-mediated transformation method for citrus was developed using

epicotyl segments. Inthis method, epicotyl segments from rootstock cultivars Carrizo citrange

or Swingle citrumelo seedlings were co-cultivated with A. tumefacience, and shoots were

produced from transformed segments by organogenic regeneration. Then, whole transgenic

plants were obtained by rooting the shoots (Moore et al. 1992; Moore et al. 1993). This









35

method was later adapted and used for more efficient transformation ofepicotyl segments from

P. trifoliata (Kaneyoshi et al. 1994). To improve the original Agrobacterium-mediated

transformation method for epicotyl segments, factors affecting the transformation and

regeneration efficiency were studied in detail, and the method was optimized and used for the

efficient transformation of citrange, lime and sour orange (Gutierrez et al. 1997).

A similar method having relatively higher transformation and regeneration efficiency

was developed to transform pineapple sweet orange. Transgenic shoots regenerated from

epicotyl segments were not rooted; instead they were grafted on in vitro grown young

rootstock seedlings using a micro-shoot tip grafting technique and then re-grafted onto

vigorous seedlings of Rough lemon in the greenhouse (Pena et al. 1995a; Pena et al. 1995b).

This method was later optimized and used for the genetic transformation of lime (C.

aurantifolia) (Pena et al. 1997) and citrange (Cervera et al. 1998). It also was reported that

Washington navel sweet orange was transformed using this method, with a modification in the

grafting stage in which transgenic shoots were directly micro-grafted to greenhouse grown

seedlings of Carrizo citrange (Bond and Roose., 1998). Even though several citrus species

have been successfully transformed and transgenic plants regenerated using these methods, the

recovery of whole transgenic plants still requires use of complicated shoot tip or micro-

grafting techniques.

Recently, a transformation protocol overcoming the difficulties associated with rooting

of transgenic shoots in the original protocol was described and used for the first successful

transformation and recovery of Duncan grapefruit plants (Luth and Moore., 1999). This

method not only increased the efficiency of Agrobacterium-mediated transformation, but it









36

also provided a simple and efficient protocol for rooting transgenic shoots in a relatively short

time of one to three weeks, (Luth and Moore., 1999).

Genetic transformation and regeneration of mature transgenic citrus plants was

achieved using an Agrobacterium-mediated transformation of internodal segments from

mature tissue of sweet orange, followed by shoot tip grafting. Several mature transgenic

Pineapple sweet orange plants were obtained, and these plants flowered and bore fruit in 14

months (Cervera et al. 1998). An Agrobacterium rhizogenes-mediated transformation

protocol was developed as an alternative method to A. tumefaciences and used to transform

Mexican lime (Perez-Molphe-Balch and Ochoa-Alejo, 1998).

Currently, the Agrobacterium-mediated transformation methods using the epicotyl

segments of seedling orjuvenile plants is the most efficient method for producing transgenic

citrus plants. This method has been useful for expressing reporter genes in different Citrus

species, and it is now being used for transferring agriculturally important genes into citrus

cultivars and relatives. Production of transgenic P. trifoliata plants expressing a synthetic

gene encoding a human epidermal growth factor was reported in 1996 ( Kobayashi et al.

1996). More recently, Carrizo citrange was transformed with the halotolerance gen, HAL2,

isolated from yeast to improve salt-tolerance in citrus (Cervera et al. 2000). Furthermore,

the CP gene from different isolates of CTV has been introduced into sour orange (Gutierrez

et al. 1997) and Mexican lime (Dominguez et al. 2000) in attempts to develop pathogen-

derived resistance to CTV. In addition, transgenic Duncan grapefruit plants expressing CP and

other sequences from the CTV genome were produced, and they are now being evaluated for

resistance to CTV (Febres et al. 2000).














CHAPTER 3
SEQUENCING AND ANALYSIS OF THE RNA-DEPENDENT RNA POLYMERASE
GENE OF CTV ISOLATES

Introduction

Despite wide variation in morphology, genome organization and sequences in their

structural proteins, all positive-stranded RNA viruses encode an RNA-dependent RNA

polymerase (RdRp) (Goldbach., 1988). The RdRp functions as a catalytic subunit of viral

replicase and is required for replication of the viral genome (Buck 1996). Comparative

analysis of the amino acid sequences of the putative RdRp from positive-stranded RNA

viruses reveals the presence of conserved sequence motifs (Kamer and Agros, 1984).

Currently, eight conserved motifs have been identified in RdRps of RNA viruses, some of

which also are present in other polymerases (Poch et al. 1989; Koonin, 1991). The conserved

motifs of the RdRp correspond to the catalytic site for RNA polymerization (GDD motif),

nucleoside triphosphate binding site, and the template and product binding sites (LKR motif)

(Ishihama and Barbier, 1994). Inhibition of virus replication by specific point mutations in

some of these conserved motifs demonstrated thatthe motifs are functionally important (Peters

et al. 1994; Jablonsky and Marrow, 1995; Davenport and Baulcombe, 1997). Since the RdRp

is common in all positive-stranded RNA viruses and its amino acid sequence is conserved

among different groups of viruses, the RdRp genes have been used for phylogenetic analyses

and classification of positive stranded RNA viruses (Koonin 1991; Koonin and Dolj a, 1993;

Zanotto et al. 1996).









38

Sequence analysis of the CTV genome showed that a 56-kDa protein encoded by

ORFIb is expressed using a +1 translational frameshift at the carboxy terminus of the

polyptotein. This protein contained sequences with conserved motifs typical of RdRp of

positive- stranded RNA viruses (Karasev et al. 1995). Comparison of the complete genome

sequences of three CTV isolates revealed an unusual asymmetric sequence similarity along

the genome in which the 3' half of the genome is more conserved with over 90% sequence

identity, than the 5' half of the genome, where as low as 70% identity occurs (Mawassi et al.

1996; Yang et al.1999). Unexpectedly, these isolates showed as much as 25% sequence

variation in their RdRp genes. Since the RdRp gene sequences from only three isolates were

used for this analysis, better understanding of sequence variability in the RdRp gene of CTV

requires more sequence information. In this study, the RdRp gene sequences from biologically

and geographically different isolates of CTVwere determined using a direct DNA sequencing

approach, and the sequence variation and the phylogenetic relationships among CTV isolates

were analyzed.

Materials and Methods

Virus Isolates

Citrus tristeza virus isolates T3, T30, T36, T66, SY568, B53, B165, B185, B249, and

VT were obtained fromthe Collection ofExotic Citrus Diseases, inBeltsville, MD, USA. The

biological characteristics of these isolates are summarized in Table 3-1.

Oligonucleotide Primers

Primers for PCR amplification were designed based on conserved regions located up-

and down-stream of the RdRp genes of the three CTV genomic sequences in the GenBank

(Accession No: U16304, U56902 and AF001623) (Figure 3-1). Based on the sequence














Table 3-1. Characteristics of the CTV isolates used in this study.


Isolates Origin Biological Characteristics
VC SY OD SP-G SP-O
T3 Florida + + +
T30 Florida +

T36 Florida + + +

T66 Florida + + +

3800 Florida + ND +

B53 Japan + + + + +

B185 Japan + + + + +

B249 Venezuela + + + + +

B165 India + + + + +

T385 Spain +

VT Israel + + +

SY568 California + + + +


VC= vein clearing on Mexican lime only. SY= Seedling yellows on sour orange seedlings.
QD= Quick decline of scions grafted on sour orange rootstock. SP-G= Stem pitting on
grapefruit scions. SP-O= Stem pitting on sweet orange scions. ND= Not determined.









40

information obtained by using these external primers, other internal primers were designed

to complete the sequencing.

RNA Isolation and Complementary DNA (cDNA) Synthesis

Double-stranded RNA (dsRNA) was isolated from CTV-infected bark tissue using a

CF11 cellulose-based protocol reported by Valverde et al. 1990. The cDNA was synthesized

from dsRNA templates using primers CN306 or CN308 with Superscript II reverse

transcriptase (GIBCO/BRL) according to the manufacturer's instructions.

Polymerase Chain Reaction (PCR)

Amplification ofthe RdRp regionbyPCRwas performed using Taq DNApolymerase

(Promega) in 50 1 reactions containing 50 mM KC1, 10 mM Tris-HCl pH 9.0, 0.1% Triton

X-100, 2.5 mM MgCl2, 0.1 mM of each deoxyribonucleotide triphosphate (dNTP), 100 pmol

of each primer, and 2-10 1 of cDNA template. The mixture was incubated at 94 OC for 2.5

min for initial denaturation, and 94 C for 1 min, 45 C for 1 min, and 72 C for 2 min for 40

cycles followed by one cycle at 72 C for 10 min.

Sequencing of the RdRp Genes of CTV

PCR amplified RdRp genes were sequenced at the University of Florida DNA

Sequencing Core laboratory using ABI Prism Dye terminator cycle sequencing protocols

developed by Applied Biosystems (Perkin-Elmer) and using CTV RdRp specific primers. The

flourescent-labeled extension products were analyzed on a Applied Biosystems Model 373

Stretch DNA Sequencer (Perkin-Elmer). Sequencing strategy and the primers used for

sequencing are shown Figure 3-2 and Table 3-2, respectively.











Table 3-2. Sequences of the oligonucleotide primers used in this study.


Primers Sequence Orientation Positiona
CN305 5' GAATATAAGGGTAGTAAAGC 3' Sense 9198-9218
CN306 5' GCAAACATCTCGACTCAACTACC 3' Anti-sense 10881-10913
CN307 5' TTTACTGAGATGACGAACGCTG 3' Sense 9898-9119
CN308 5' CGCGTCGAATTTTATGAGTCCC 3' Anti-sense 11146-11167
CN309 5' TGTTTTGTACCGGACCCTTA 3' Sense 10405-10424
CN310 5' GTACTCGCCTTCCATCCA 3' Anti-sense 10081-10098

aThe positions of the primers are indicated according to the genomic sequence of CTV strain
T36.


CTV RdRp


5' 1 3'


59
UN30!3


110 CN
CN307 )0


CN366 or 308


Figure 3-1. Schematic representation of the RdRp region of CTV and the strategy used for
sequencing the RdRp gene. The black bars indicate the location of the primers and the lines
with arrowheads indicate the length and direction of the sequence obtained from the specific
primers.


1 31











Sequence Analysis

The RdRp nucleotide sequences were assembled and translated using the Genetics Computer

Group (GCG). Multiple sequence alignments were generated in the Align X program of the

Vector NTI suite. Phylogenetic analyses were done using the Diverge module of GCG which

makes codon-by-codon comparison of aligned protein coding sequences and estimates the

number of synonymous and non synonymous substitutions per site (Li, 1993). The phylogenetic

tree was generated in the GrowTree module of GCG usingKiamura two-parameter algorithms

with the neighbour-joining and UPGMA methods.

Results and Discussion

Primers designed based on conserved sequences located up- and downstream of the

RdRp gene of three different isolates of CTV were used to amplify the RdRp genes from a

number of biologically and geographically different isolates of CTV. The RdRp sequence of

CTV isolates T30, B53, B165, B185, B249 and 3800 were determined using direct DNA

sequencing of the ORF lb region. Two different RdRp sequences, designated as T2K and

T38K, were obtained from CTV isolate 3800, which possibly contains two distinct virus

populations (Manjunath et al. 2000). In addition to these sequences, the RdRp sequences of

CTV isolates T36 (Karasev et al. 1995), SY568 (Yang et. al.,1997), VT (Mawassi et al.
















101 2)1 31 401 5D1 601 701 8D1 9D1 1OD1 1101 1201 1301 1401
WEIII k LU k I J 11 1 J. 1 411.1.11L I IL I. I I II I 1M .I 1311,41


Figure 3-2. Homology (Top) and absolute complexity (Bottom) graphs of the nucleotide
sequences of the RdRp gene from 10 biologically and geographically different isolates of
CTV. The graphs were generated using the Align X program of Vector NTI Suite. The
homology graph displays alignment quality based on the total similarity values (1,0.5 and 0.2
for identical, similar or weakly similar, respectively) for the residues at a given position. The
absolute complexity graph displays a statistical significance profile of the alignment calculated
as the sum of all pair-wise substitution scores at a given alignment position. The 5' end of the
RdRp gene is to the left.


Table 3-3. Percentage of nucleotide sequence identity of the RdRp genes from different
isolates of CTV. The calculations were made using Clustal W. The most similar isolates are
shown in blue and the most different isolates are shown in red
Isolates B53 B249 VT SY568 T30 B165 T36 3800 T385 T2K T38K

B185 99 98 96 91 91 85 79 83 91 82 78

B53 98 96 91 91 85 79 83 91 82 85

B249 96 90 91 85 79 83 91 82 84

VT 89 89 84 79 82 90 82 84

SY568 91 81 75 79 91 79 81

T30 85 78 83 99 82 85

B165 79 83 85 81 86

T36 85 78 89 78

3800 83 90 89

T385 82 85

T2K 80











1.0




0 1 '41 81 121 161 '21 241 281 321 361 '401 '441 496






0

Figure 3-3. Homology (Top) and absolute complexity (Bottom) graphs of deduced amino acid
sequences of the RdRp gene from 10 biologically and geographically different isolates of CTV.
The graphs were generated using the Align X program of Vector NTI Suite. The homology
graph displays alignment quality based on the total similarity values (1, 0.5 and 0.2 for
identical, similar or weakly similar, respectively) for the residues at a given position. The
absolute complexity graph displays a statistical significance profile ofthe alignment calculated
as the sum of all pair wise substitution scores at a given alignment position. The 5' end of the
RdRp gene is to the left.


Table 3-4. Percentage of and amino acid sequence identity of the RdRp genes from different
isolates of CTV. The calculations were made using Clustal W. The most similar isolates are
shown in blue and the most different isolates are shown in red.
Isolates B53 B249 VT T30 B165 T36 3800 T385 T2K T38K

B185 99 98 97 96 94 90 90 96 93 92

B53 99 96 95 93 89 90 95 92 92

B249 96 95 94 90 90 95 93 92

VT 94 91 87 88 94 90 90

T30 93 89 90 100 92 92

B165 88 90 93 92 92

T36 91 89 94 88

3800 90 95 93

T385 92 92

T2K 90









45

including B53, B185, B249 and VT exhibited 96 to 99% sequence identity to each other

(Table 3-3). Comparative analyses of the deduced amino acid sequences of the RdRp genes

confirmed that T36 was the most diverse isolate with as much as 13% or 65 amino acid

differences in the RdRp gene (Table 3-4). On the other hand, the RdRp of two mild

isolates,T30 fromFlorida and T385 from Spain, were 100% identical and only a few amino

acid differences were identified among the RdRps of B53, B185 and B249, all SP-inducing

isolates from Spain, Japan and Venezuela, respectively. The B53 isolate reportedly was

originally found in an early satsuma mandarin illegally imported to Spain from Japan

(Ballester-Olmos et al.1988). This implies that B53 and B185 might have the same origin

(Table 3- 4). The remaining isolates had 5 tol0% variation in the amino acid sequences of

their RdRp genes. The isolates sequenced in this study (B53, B165, B185, B249 and T30)

were more similar to each other thanthe previously sequenced isolates (T36, VT and SY568).

This may be due to differences in the sequencing methods used. While T36, VT and SY568

were sequenced from clones, a direct DNA sequencing method was used in this study. One of

the major advantages of direct sequencing compared with sequencing of cloned DNA is that

it allows rapid determination of the major sequence of heterogenous virus populations

(Odeberg et al. 1995). The nucleotide reported at a specific position can be the additive result

of several minor isolates present in the population, identical at a specific position but different

at other positions. The major sequences obtained by direct sequencing is the consensus

sequence ofheterogenous virus populations present, but is not necessarily the sequence of the

most abundant strain (Odeberg et al. 1995). In contrast, sequences obtained from cloned DNA

sequencing do not represent the population, but individual













1 72
B249 TTT CTAAACCCGACGCTAGCGATGGT------------------CAAGCGGACGACTTAGCGACAGGCTGA
B185 TTT CTAAACCCGACGCTAGCGATGGT------------------CAAGCGGATGACTTAGCGACAGGTTGA
B165 TTTACTAAACCCGATGCTAGCGATGGT------------------CAAGCGGACGACTTAGCGACAGGCTGA
SY568 TTTRCTAAACCCGAYRCTAGCGATGGT------------------CAARYGGAYGACTTRRYGACWGRTTGA
VT TTT CTAAACCCGACGCTAGCGATGGT------------------CAAGCCGATGACTTAGCGACAGGCTGA
B53 TTT CTAAACCCGACGCTAGCGATGGT------------------CAAGCGGATGACTTAGCGACAGGTTGA
T30 TTTACTAAACCCGATGCTAGCGATGGT------ ---------CAA TGGACGACTT TGACTGGTTGA
T385 TTTACTAAACCCGATGCTAGCGATGGT------ ---------CAA TGGACGACTT TGACTGGTTGA
T3 TTTACTAAACCCGATGCTAGCGATGGT------------------CAAGCGGACGACTTAGCGACAGGCTGA
3800 TTC ATAAACCCGATGCTAGCGACGGT------------------CAAGCGGACGACTTAGCGACCGGCTGA
T38K TTC ATAAACCCGATGCTAGCGACGGT------------------CAAGCGGACGACTTAGCGACCGGCTGA
T2K TTTACTAAACCCGACTCTA CGATGGT ----------------CAAGCGGACGACTTAGCGACCGGCTGA
T66 TTCACTAAACCTGACGCTA CGACG TAACGTGGACGACCTCGGACAAGCGG TGAATTAGC ACCGGCTGA
T36 TTCACTAAACCTGACGCTA CGACG TAACGTGGACGACCTCGGACAAGCGG TGAATTAGC ACCGGCTGA


Figure 3-4. Multiple sequence alignment of the 3' region of the RdRp genes of CTV isolates
showing the 18 base insertion in some isolates. The alignments was produced using AlignX
module of Vector NTI. The asterisks indicates identical nucleotide in all isolates and dashes
indicate missing nucleotide sequence. Number shows the position of the nucleotide in the
genomic sequence of CTV strain T36.



strains, possibly the most abundant strain in a heterologous virus population. In order to get

a representation of a virus population, a large number of clones would have to be sequenced

(Odeberg et al. 1995). Multiple alignment of the RdRp genes also revealed an 18-nucleotide

in frame insertion close to the 3' end of RdRp gene of T36, a QD isolate from Florida (Figure

3-4). This insertion was not found in the RdRp gene of any of the ten isolates previously

sequenced or determined in this study. To identify other isolates having the 18 nucleotide

insertion in their the RdRp gene, the 3' half of the RdRp gene of two other CTV isolates from

Florida, T66 and T3, were amplified and directly sequenced from the PCR products. The

insertion sequence was found in T66, which is a QD inducing isolate similar to T36, but itwas

not present in T3 which has a different biological activity (Figure 3- 4). In order to screen

more CTV isolates for the presence of the insertion sequence, a PCR primer specific to the

insertion sequence was designed. More than 20 CTVisolates from the Collection of Exotic

Citrus Diseases and 10 isolates from Spain were tested by PCR using the insertion specific









47

primer and an internal primer specific to a region conserved in all CTVRdRp genes. Among

these isolates, only one QD inducing isolate from the collection contained the insertion

sequence. Although all three isolates with the 18-nucleotide insertion sequence in their RdRp

gene do induce QD, other QD inducing isolates tested by PCR did not have the insertion, thus

indicating that the insertion may not be related to the biological characteristics of CTV

isolates.

The RdRp is conserved in all positive-stranded RNA viruses, and it has been used

to determine the evolutionary relationships among the positive- stranded RNA viruses, and

also intheir classification (Koonin, 1991; Koonin and Dolja, 1993; Zanotto et al. 1996; Hong

et al. 1998). The RdRp gene also has been used for phylogenetic analyses of isolates of a

number of animal and plant RNA viruses (Arankalle et al. 1999; Hitomoto et al, 2000;

Bousalem et al. 2000). In this study deduced amino acid sequences of the RdRps were

analyzed to determine the evolutionary relationships among CTV isolates with diverse

biological properties and different geographical origin.

On the basis of phylogenetic analysis of the RdRp genes from different isolates of

CTV, two distinctly branched clusters could be identified. One of the main clusters (I)

consisted of only two isolates, T36 and T2K, from Florida. The second main cluster (II)

contained all other isolates of CTV and was divided into two sub-clusters (Figure 3-5). The

isolates 3800, T38K and B165 were grouped together in one sub-cluster (IIA) and the

remaining isolates grouped into sub-cluster IIB, which is further divided in to two different

branches (IIB-1 and IIB-2) (Figure 3-5). The clusters clearly are not related by the










48

B135,


B53B


249


VT


T30 H


SY568 2





T3S00

A
138K


B165,


T36





11
BYV-U



Figure 3-5. A phylogenetic tree showing relationships among CTV isolates based on their
RdRp genes sequences. The phylogenetic tree was generated by GCG's Diverge and
GrowTree programs using sequence alignments and matrix of evolutionary distances
determined by the number of synonymous and non synonymous substitutions per site in the
protein coding sequences. Biological characteristics and geographical origins of the isolates
are shown in Table 3-1. The RdRp gene of BYV was used as the out group.









49

geographical origin of the isolates. However, there is some association with the biological

properties of the isolates. Cluster I contains the known QD isolate from Florida, T36, and

also T2K, a sub-population of isolate 3800, which induces SP on grapefruit but has not been

tested for QD. It is possible that T2K may be the QD inducing component of isolate 3800.

Cluster IIA consists of the known SP isolates B165 and 3800 from India and Florida,

respectively. The presence of T38K, the other sub-population of isolate 3800, in this group

implies that T38K may be the SP-inducing component of the isolate 3800. The cluster IIB-1

contains two well characterized mild isolates of CTV, T30 fromFlorida and T385 from Spain,

and a known seedling yellows isolate SY568. Comparative analyses of genome sequences

of SY568 and other isolates of CTV suggested that the SY568 genome may have resulted from

RNA recombination between two CTVgenomes, one of which was almost identical to T385

(Vives et al. 1999). Thus, it is possible that the cluster IIB-1 actually contains only the mild

isolates of CTV. The cluster IIB-2 may be considered as the second SP group, since it

consists of well characterized SP isolates from different regions of the world. It can be

further inferred from this phylogenetic information that there are two lineages of CTV: the QD

lineage, and the second lineage which later diverged to mild and SP lineages. Although

isolates with similar biological activities grouped together in the phylogenetic analyses, the

results are not conclusive because of the limited number of isolates analyzed in this study.

Conclusions

Highly conserved regions of the CTV genome were used to design PCR primers for

non-selective amplification of the RdRp gene from very distinct isolates of CTV. These

would also be used for characterization of even more variable regions of the CTV genome.

Direct sequencing of PCR products is an effective method for identification of consensus









50

sequences found in a virus population. It also is easier, faster and more economical than

sequencing from the cloned cDNAs. The RdRp genes of biologically and geographically

different isolates of CTV are conserved and related in their nucleotide and amino acid

sequence. The most obvious difference among the RdRp genes of CTV isolates is the presence

of an 18-nucleotide in-frame insertion sequence close to the 3' end of the RdRp genes of

isolates T36 and T66, both QD isolates from Florida. At this point, the insertion sequence

appears to be rare and is not associated with all QD isolates.

On the basis of comparative sequence analyses, the QD isolate T36 fromFlorida was

the most distinct isolate, and the RdRp genes of some CTV isolates with similar biological

activity have a significantly higher level (98 to 100%) of sequence identity, indicating thatthey

are more closely related at the molecular level. Since CTV isolates show unusual and

asymmetric sequence relationships along their genomes, comparison of the RdRp genes

sequences may provide the most straight forward method for phylogenetic analysis. Based on

these phylogenetic analyses of their RdRp genes, CTV isolates can be divided into four

distinct groups. Three of these groups are interrelated, suggesting that they diverge from the

same lineage. However, the other group is more distantly related to these groups indicating

that it may be from a different lineage. Although it is not conclusive, the groups appear to

contain biologically similar isolates.















CHAPTER 4
CLONING AND EXPRESSION OF THE RNA-DEPENDENT RNA POLYMERASE OF
CTV IN Escherichia coli

Introduction

All positive-stranded RNA viruses use an RNA dependent RNA polymerase (RdRp)

for their replication. The RdRp is the catalytic subunit of the viral replicase complex which

includes other viral and host proteins. Therefore, they are absolutely required for the

replication of the virus in the host cell (Buck, 1996). Although the RdRps are universally

conserved among different groups of RNA viruses, they express their RdRps using a variety

of expression mechanisms. In some viruses, the RdRp is produced as a single peptide, but in

other viruses, it is expressed as a polyprotein containing RdRp and one or more other viral

proteins. The expression of the RdRp is controlled and regulated by different mechanisms in

different viruses by proteolytic processing, translational readthrough or frameshifting (Gallie,

1996). The RdRp of CTV is encoded by ORF Ib. It overlaps with ORF la which encodes two

proteases and the other replication associated proteins, a methyltransfrase and a helicase. It

is proposed that the CTVRdRp is expressed by a + 1 translational frameshifting and possibly

requires proteolytic processing to liberate it from the polyprotein encoded by ORFla\lb

(Karasev et al. 1995).

Isolation of viral RdRps from infected host cells is very difficult because they are

expressed and accumulated at very low levels. Consequently, most viral RdRs have been

identified based on sequence analysis of conserved domains (O'Reilly and Kao., 1998).









52

However, the catalytic activity of RdRp has been demonstrated for a number of animal and

plant viruses in partially purified extracts isolated frominfected host cells (Graaff and Jaspar.

1994). Among the plant viruses, the partially purified extract of RdRps of turnip yellow

mosaic virus (TYMV) and brome mosaic virus (BMV) have been well-characterized and used

in in vitro studies of viral replication. Development of these in vitro replication systems

using their RdRps revealed information on the mechanism and regulation for the replication

of these and similar viruses (Singh and Dreher, 1997; Sun and Kao, 1997; Siegel et al. 1998).

On the other hand, for most viruses the isolation of partially purified active RdRp from the

infected plants was unsuccessful due to host characteristics, low expression level, and/or

poor solubility of the purified RdRp itself (Graaff and Jaspar. 1994).

Difficulties associated with the purification of the RdRp from the infected host cell

may be overcome by expression of the RdRps in E. coli. Enzymatically active RdRps of

poliovirus (Rothstein et al. 1998), hepatitis C virus (Behrens et al. 1996; Lohmann et al.

1997), tobacco vein mottling virus (Hong and Hunt, 1996) and bamboo mosaic virus (BaMV)

(Li et al. 1998) were expressed in E. coli, and their function and catalytic activity were

characterized. Some other viral RdRp were expressed inE. coli and the products were used

for production of antibodies specific to viral RdRps. These antibodies were later used for

detection, identification and functional dissection of RdRps of RNA viruses in the infected

host tissue (Hayes et al. 1994; Scholthof et al. 1995).

In this study, the RdRp gene of CTV isolate T36 was cloned and expressed in E. coli

and polyclonal antibodies to the expressed protein were produced in order to detect and

localize the RdRp of CTV in the infected citrus tissue.









53

Materials and Methods

Cloning of the CTV RdRp Gene

The cDNA specific to the RdRp region of CTV isolate T36 was synthesized from

dsRNA templates using primer CN257 (5' CTACTCGAGATCTATCAATCGATCAGCC

GGTT 3') with Superscript II reverse transcriptase (GIBCO/BRL). The RdRp gene was

amplified from cDNA by PCR using Taq DNA polymerase with the primers CN256 (5'

TGTAAGCTTA TGGAGACACCGCCCCTCCT 3') and CN257. These primers were designed

based on published sequence information for isolate T36 (Karasev et al. 1995), and since the

RdRp gene does not have an initiation codon, an ATG codon was incorporated into the plus-

sense primer CN256. In addition, the specific recognition sequences for the restriction

enzymes HindIm and Bgl II (underlined sequences) were incorporated into the primers CN256

and CN257, respectively, to facilitate the cloning into the expression vector. The amplified

RdRp gene was first cloned into the pGEM-T cloning vector (Promega) using the T-A cloning

method. The HindI/Bgl II fragment ofthe pGEM-T clone containing the RdRp gene was sub-

cloned into the pETh-3c expression vector (Invitrogen), and the resulting plasmid was

designated as pETh3c-CTVRdRp. The presence and the integrity of the CTV RdRp ORF in

the pETh3c vector was confirmed by sequencing.

Expression of the CTV RdRp

Escherichia coli strain BL21 was transformed with the pETh3c-CTVRdRp, and

protein expression was induced by 0.4 M isopropyl thio- -D-galactoside (IPTG) at 37 C for

3 hrs. The expressed proteinwas separated fromother bacterial proteins ona 10 or 12% SDS-

polyacrylamide gel by electrophoresis (SDS-PAGE) and stained with Coomassie brilliant

blue G-250 (Sambrook et al. 1989)









54

Production of Polyclonal Antibodies Specific for the CTV RdRp

The recombinant protein was produced in large scale and separated from other

bacterial proteins by SDS-PAGE and stained with 0.3 M CuC12 to visualize the bands The

protein band was excised from the gel, and the gel slice destined in 0.25 M EDTA/ Tris-

HC1, pH 9.0 (Lee et al. 1987). The protein in the gel slice ( about 5-7.5 mg ) was used for

production of polyclonal antibodies in a rabbit by injecting about 1.0 -1.5 mg of expressed

protein into the rabbit weekly for four weeks (Cocalico Biologicals Inc.). Antiserum was

collected after the forth injection and tested for specificity to the expressed protein by Western

blot analysis. To increase the antiserum titer, one more injection with (about 2.5 mg) more

protein was performed after the fourth collection of antiserum.

Western Blot Analysis

Total protein was extracted from E. coli strain BL21 with or without the CTVRdRp

gene and from CTV infected and healthy citrus tissue. The extracts were separated on 10-12

% SDS-PAGE and transferred to nitrocellulose membranes using a semi-dry blotter (Bio-

Rad). The membranes were probed with different dilutions of antiserum raised against the

expressed CTV RdRp followed by detection with alkaline phosphatase (AP) or horseradish

peroxidase (HRP) conjugated anti-rabbit antibody. The membrane blots were developed

using the BCIP and NBT colorimetric system for the AP and the Super SignalTM

chemiluminescent substrate system (Pierce) for location of HRP.

Cell Fractionation Assays

Cell fractions of healthy and CTV- infected citrus tissue were prepared by differential

centrifugation using a previously described protocol (Godefroy-Colburn et al. 1986). The cell








55

fractions were analyzed on 10-12% SDS-PAGE followed by Western blot and

chemiluminescent detection system.

Results and Discussion

Expression of CTV RdRp in E. coli

Expression ofCTVRdRp was induced in E. coli strain BL21 transformed with the

pETh3c-CTVRdRp plasmid by addition of isopropyl thio- -D-galactoside (IPTG). CTV

RdRp was expressed as a fusion protein with an estimated molecular weight of about 60 kDa,

which contains about 2 kDa from the vector and 57 kDa from the CTVRdRp ORF. Separation

and analysis of bacterial proteins by SDS-PAGE revealed a high level of expression of the

60 kDa protein only in the induced bacterial cells containing the pETh3c-CTVRdRp plasmid.

The protein was not expressed in non-induced cells with the same plasmid or in induced cells

containing the pETh3c vector alone (Figure 4-1). Since the expression of the protein was

specifically induced and its size corresponded to the expected size, it was concluded that this

protein was the CTV RdRp.

The induction of the fusion protein was evaluated periodically starting 30 min after

addition ofIPTG. The expression of the protein increased continuously overtime and reached

maximum at 3 hr after induction at 37 C (Figure 4-2). The purification of the proteins from

bacterial cells indicated that the expressed RdRp was not soluble, and was found only in the

insoluble fractions of the E. coli cells. (Figure 4-3). To produce soluble CTV RdRp,

expression was induced at lower temperatures such as 30 and 35 O C, and several different

solvents such as urea were used to solubilize the recombinant CTV RdRp.








56
1 2 3 4 5 6


7 8


Figure 4-1. Induction of the expression of the CTV RdRp in E. coli by IPTG. Lanes land 7
induced and non-induced E. coli BL21 carrying pETh3c vector alone, respectively. Lane 2 is
non-induced E.coli BL21 with the pET3c-CTVRdRp. Lanes 3-7 induced E.coli BL21 with
pETh3-CTVRdRp; and lane 8 is molecular weight markers.
1 2 3 4 5 6 7 H


-- -- o


.. -


Figure 4-2. Time course study of expression of the CTV RdRp in E. coli after the induction
by IPTG. Lanes land 2 are induced and non-inducedE. coli BL21 carrying the pET3c vector
alone, respectively. Lane 3 is non-induced E.coli BL21 with the pETh3c-CTVRdRp. Lanes
4-7 are induced E. coli BL21 with pETh3c-CTVRdRp at 30 min, 1, 2, and 3 hours after the
induction by IPTG.
1 2 3 4 5 6 7 8 9


0~


4-0


Figure 4-3. Accumulation of the expressed CTV RdRp in cellular fractions of E.coli BL21.
Lanes 1 and 2 are insoluble and fractions, respectively, ofE. coli BL21 carrying the pETh3c
vector alone. Lanes 3, 6 and 8 are the insoluble fractions and lanes 4, 6 and 9 are soluble
fractions ofE. coli BL21with the pETh3c-CTVRdRp. Lanes 3-4, 6-7 and 8-9 were grownat
35, 37 and 30 C, respectively.


" L _'- L


,,, I T .


,


"' )I:LIB








57

However, the protein could notbe solubilized and always remained in the insoluble fractions

of the cell (Figure 4-3).

Production and Testing of Polyclonal Antibodies to the CTV RdRp

Polyclonal antibodies to the CTVRdRp were raised by injecting a rabbit with a total

of 7.5-10 mg recombinant RdRp. After four weekly injections, antiserum was collected and

tested for specificity by Western blot analysis. The antiserum reacted strongly with the 60 kDa

protein expressed in bacterial cells containing the pETh3c-CTVRdRp plasmid, but not with

proteins in cells carrying the pETh3c plasmid alone (Figure 4; lanes 1-4). A weak reaction

also occurred with two other smaller proteins in bacterial cells containing the pETh3c plasmid

with or without the CTV RdRp gene (Figure 4; Lanes 1-4). These results indicated that the

antiserum was specific to the expressed CTV RdRp but that it also cross reacts with two

abundant bacterial proteins. Different dilutions of the expressed CTV RdRp tested by Western

blot analysis to determine the sensitivity of the antiserum showed that a 1:10,000 dilution of

the antiserum could detect up to a 1:1000 dilution (about 10-100 ng) of the expressed protein

(Figure 4; Lanesl-4). However, best results were obtained using 1:1000 dilution of the

antiserum.

Detection and Sub-Cellular Localization of the CTV RdRp in CTV-Infected Citrus Tissue

To detect the expression of RdRp in the CTV-infected tissue, total protein from

infected and healthy citrus bark tissue was separated by 10-12% SDS-PAGE, transferred to

a nitocellulose membrane and probed with 1:1000 dilution of the antiserum to the 60 kDa

recombinant protein. A non-specific reaction with a protein of about 55 kDa was detected in

both infected and healthy tissue (Data not shown). This non-specific reaction was eliminated

by cross-absorbing the antiserum first with a healthy tissue extract before using it to probe

the blot. No specific reaction with the antiserum was detected in the healthy or infected tissue








58

using anti-rabbit antibody conjugated with AP followed by colorimetric detection with BCIP

and NBT. However, a reaction with a 50 kDa protein in CTV-infected tissue extract, which

was not present in the healthy tissue extarct, was observed using the more sensitive detection

method involving anti-rabbit antibodies conjugated with HRP followed by chemiluminescent

detention system (Figure 5;Lanes 3 and 4). The size of the protein does not correspond to the

size of the putative protein product of the CTV ORF b, expected to be 57 kDa, or the CTV

polyprotein which is about 400 kDa. Although the detected protein band was discreet and

solid, the reaction was weak and could only be detected when the gel was overloaded (using

25 1 of extract compared to of 10 1) with the freshly prepared extract from the infected

tissue. The difficulty detecting RdRp in the infected tissue could be due to a low level of

expression of the CTV RdRp in vivo. In general, since the RdRp is a catalytic protein for RNA

viruses their RdRps are expressed at low levels and could not even usually be detected in the

infected tissue.

To determine the sub-cellular localization of CTV RdRp in the infected citrus tissue,

fractionated preparations from infected and uninfected tissue extracts were analyzed by

Western blot analysis using antibodies specific to recombinant RdRp of CTV. A band of

about 50 kDa was detected in the membrane and cytoplasmic fractions of the CTV-infected

tissue (Figure 6; lanes 4 and 8). However, no reaction was observed with any fractions of

healthy tissue (Figure 6; Lanes 1, 3, 5, and 9). The band reacting with the antiserum was

present in the cytoplasmic fraction, but it accumulated predominantly in the membrane

fractions of the infected cells. Although the size of the protein detected in the infected tissue

and some of its fraction is smaller than the expected size, of the specificity of the reaction










1 2 3 4 5 6 7
60 kDa








Figure 4-4. Western blot analysis of the CTV RdRp expressed inE. coli. Lane 1 is the pETH-
3c vector without the RdRp as the negative control. Lanes 2, 3, 4, 5 and 6 are 1:10, 1:100,
1:1,000, 1:5,000 and 1:10,000 dilutions of the expressed protein, following purification from
E.coli, respectively. Lane 7 is pre-stained protein markers.
1 2 3 4


N


6.0 k a.


Figure 4-5. Detection of CTV RdRp in CTV-infected citrus tissue by Western blot analysis
using antiserum prepared to the CTV RdRp expressed in E. coli. Lane 1 is CTV RdRp
expressed in E. coli as the positive control. Lanes 2 and 4 are the negative controls from the
extract of E. coli containing the pETH-3c without the RdRp and extracts of healthy citrus
tissue, respectively. Lane 3 is the extract from citrus tissue infected with T36 isolate of CTV.


1 2 3 4 5 6 7 S 9


60 kDa

5 kDa
4-

50 kDa


Figure 4-6. Sub-cellular localization of the CTV RdRp by Western blot analysis using
antiserum prepared to the CTV RdRp expressed in E. coli. in citrus tissue infected with the
T36 isolate of CTV. Lanes 1, 3, 5, and 9 are nuclear, cytoplasmic, cell wall and membrane
fractions of tissue, respectively. Lanes 2, 4, 7 and 8 are nuclear cytoplasmic, cell wall and
membrane fractions of citrus tissue infected with the T36 isolate of CTV, respectively. Lane
6 is the expressed protein as a positive control.












with the antibody and sub-cellular location where the protein was detected, indicate that the

detected protein is likely the RdRp of CTV.

The accumulation of the CTV RdRp in the membrane fractions of the infected

cells is consistent with previous reports that RdRps of most RNA viruses are membrane

associated (Graaff and Jaspars, 1994). In addition, the detection of the BYV helicase, which

interacts with RdRp during viral replication, in the membrane fractions also supports the

suggestion that the CTV RdRp is predominantly associated with membrane fractions of the

infected tissue.

It is possible that the RdRp could be processed and cleaved fromthe CTVpolyprotein

by a protease encoded by the virus or the host. Recently, it was reported that the helicase and

methyltransferase domains were detected in BYV infected tissue as individual proteins

indicating that these proteins were cleaved by the processing of the polyprotein (Erokhna et

al. 2000). A host cystein-like protease involving the processing of the RdRp ofLeishamania

virus from its polyprotein precursor has already been identified (Young-Tae., 1997). Even

though this virus has a dsRNA genome that is taxonomically unrelated to CTV, the expression

of the RdRp gene using the + 1 translational frameshifting mechanism is common to both

viruses.

Conclusions

The RdRp gene of T36, a QD isolate of CTV, was successfully expressed in E.coli,

and a polyclonal antiserum specific to the RdRp was produced. A 50 kDa protein was

detected in CTV-infected tissue extract using the RdRp specific antiserum. However, the

detection of this protein was difficult and required the use of highly sensitive detection









61

methods. The detection of a protein with anti serum specific to the recombinant RdRp suggests

that the RdRp gene of CTV is expressed in the infected tissue, and implies that the proposed

+1 translational frameshift takes place in the CTV infected tissue. The size of the detected

protein is smaller than the expected size of the CTV RdRp. This indicates that after the

expression of RdRp as a fusion protein with the CTV polyprotein, it is processed and

cleaved from the rest of the polyprotein encoded by CTV ORF a. The 50 kDa protein was

detected in the cytoplasmic and membrane enriched fractions of infected tissue, and mostly

accumulated in the membrane fraction and not in healthy tissue extracts. This is consistent

with the sub-cellular localization of RdRp and replication-associated proteins of other RNA

viruses.














CHAPTER 5
CHARACTERIZATION OF THE +1 TRANSLATIONAL FRAMESHIFT FOR THE
RNA-DEPENDENT RNA POLYMERASE GENE OF CTV

Introduction

Ribosomal frameshifting is a directed change in the translational reading frame, which

allows the production of a single protein from two overlapping genes. The frameshift can

occur in either the 5' (-1 frameshift) or 3' (+1 frameshift) direction (Brierley, 1995). A

number of viruses and mobile genetic elements use the ribosomal frameshifting mechanism to

control expression oftheir repli case gene at the translational level. This mechanismfacilitates

controlled low-level synthesis of the polymerase, which is needed in only small quantities.

(Brault and Miller, 1992). The -1 programmed ribosomal frameshifting has been

demonstrated for animal retroviruses, coronaviruses, toroviruses, arteriviruses and

astroviruses (Brierley, 1995; Farabaugh, 1996), a yeast double-stranded RNA virus (Dinman

and Wicker, 1994) and also for the plant luteoviruses, sobemoviruses, carlaviruses,

enamoviruses and dianthoviruses (Maia et al. 1996).

Sequence motifs responsible for inducing the -1 ribosomal frameshift were identified

by extensive sequence comparison and site-directed mutagenesis of the frameshift region of

these viruses. The -1 frameshift signal consists of two cis-acting sequence elements including

a shifty heptamer with the consensus sequence ofXXXY YYZ (the triplets represent 0 frame,

X=A, G or U; Y=A or U; Z=A, C or Y), called the "slippery" sequence and a down stream

RNA secondary structure in the form of a stem-loop or a more complex configuration, the

pseudoknot (Brierley, 1995; Farabaugh, 1996).









63

On the other hand, +1 ribosomal frameshifting is less common, and it has been

described in only a few systems including the yeast retrotransposon TY (Clare and Farabaugh,

1985; Clare et al. 1988), the copia-like element of Drosophila (Farabaugh, 1997) Leishmania

dsRNA virus (Lee et al. 1996) and cellular genes such as rat ornithine decarboxlyase

antienzyme gene (Matsufuji et al. 1995) and peptide release factor 2 (prfB) gene of E. coli

(Craigen and Caskey, 1986). The +1 frameshifting has been proposed for plant closteroviruses

(Agronowsky et al. 1994), including citrus tristeza virus (CTV) (Karasev et al. 1995).

The +1 frameshift was first reported for the yeast retrotransposon Ty-1 in which TYA

ORF coding for the gag analog overlaps with the first 38 nucleotide of the TYB ORF coding

for the pol analog of the retroviruses. The +1 translational frameshift occurred within the

conserved sequences of the overlap between TYA and TYB ORFs (Clare and Farabaugh,

1985). Based on the analyses of the 38 bp overlap region, itwas first suggested that31 nt in

that region were required for the frameshift (Wilson et al 1986) and later the length of the

required sequences was reduced to 14 nt (Clare et al. 1988). Further analysis of this sequence

demonstrated that only 7 nt (CUU-AGG- C) were necessary and sufficient to promote the +1

frameshift in Ty-1 (Belcourt and Farabough,1990). The heptomeric minimal signal required

for the +1 frameshift contained two overlapping leucine codons (CUU and UUA) and an

arginine codon (AGG) decoded by a low-abundance tRNA which is encoded by a single copy

gene, called rare arginine. Based on this information, the tRNA slippage model was proposed

for the mechanism of +1 frameshift in which the ribosome pauses at the rare arginine codon,

tRNA slips from the cognate to the near cognate codon and the frameshift takes place with

translation continuing in the +1 frame (Farabough,1996; 1997). This model was supported by

the report that over expression of arginine CCU tRNA suppressed the +1 frameshift in yeast









64

(Kawakami et al. 1993). More recently it was reported that the pokeweed antiviral protein

also specifically inhibited the Ty-1 directed +1 ribosomal frameshift in yeast (Tumer et al.

1998).

Genomic sequence analysis of CTV revealed that ORF la and lb overlap by 123

nucleotides indicating that the RNA-dependent RNA polymerase (RdRp) of CTV encoded by

the ORF lb is possibly expressed via a +1 frameshift in the 5' end ofORFla (Karasev et al.

1995). The overlap region had no homology to the characterized +1 frameshift signals of other

systems but showed some similarity with the overlap region of beet yellows virus (BYV), the

type member of the closterovirus grup. Although the overlap region of ORFla and lb is

relatively conserved, there were significant differences between the BYV and CTV sequences

around the predicted frameshift site. The overlap region of BYV contained a UGA stop codon,

a GGGUUU slippery sequence and a stem-loop structure (Agranovsky etal. 1994) whichwere

similar to -1 frameshift signals and absent in the CTV sequence (Karasev et al. 1995). The

absence of these elements in CTV indicated that the mechanism of the +1 frameshift used by

BYV and CTV, two closely related viruses, may differ. An arginine codon CGG in the

overlap region was one of the least frequent codons used in the CTV genome, and it aligned

with the UAG stop codon in the overlap region of BYV. Based on this information, it was

suggested that this rare arginine codon maybe important for the +1 frameshift of CTV, and the

mechanismof the frameshift may be similar to thatofTy-1. The rare arginine codon may serve

as a stop codon to pause the ribosome during translation and shift the reading frame in +1

direction (Karasev et al. 1995). These proposals have never been tested experimentally and

the occurrence of the frameshift has not been demonstrated for CTV, BYV and other









65

closteroviruses in vivo or in vitro, and the mechanism of the frameshift still remains to be

determined.

Inthis study, expression of the RdRp gene by the +1 frameshiftwas demonstrated using

an in vitro translation system and anAgrobacterium-mediated transient expression assay. The

involvement of the overlapping region of ORF la and lb and the possible sequence elements

required in the +1 frameshift were analyzed by the Agrobacterium-mediated transient

expression assay using the $-glucronidase (GUS) and the green fluorescent protein (GFP) as

fusion reporter genes.

Material and Methods

In vitro Transcription and Translation

To determine if the proposed +1 frameshift for the expression of the RdRp actually

occurs in CTV, the 3' half of the ORFla containing the helicase domain and the complete

ORF b was amplified from a clone containing both regions of the CTV sequences by PCR

using primers CN437 and CN257 (Table 5-1) (Figure 5-1). An ATG translation initiation

codon and Ndel restriction site at the 5' end and Bgl II restriction site at the 3' end were

incorporated by primers CN437 and CN357, respectively, during PCR amplification. The

amplified DNA fragment was cloned between Nde I and Bgl II sites in the pGEM-T vector

under the SP6 promoter. The clone was sequenced, and the presence and the integrity of the

sequences for the helicase, the overlap region of ORF a and ORF lb and the whole ORFlb

were confirmed. The CTV sequence in the clone was transcribed and translated using the

TNT in vitro transcription and translation kit (Promega) with wheat germ extract and 3H

labeled leucine according to manufacturer's instructions. The translation products were

separated by 12 % SDS-PAGE and detected by autoradiography.












Table 5-1. The sequences of the primers used for amplification, mutagenesis and cloning of
the frameshift constructs.
Primer Sequencea(5' to 3') Orientation

CN257 ctactcgagatctATCAATCGATCAGCCGGTT Anti-sense

CN356 aaagcggccgcaccagGAGACACTGCCCCTCCCGACTCC Sense

CN365 atgaccatGAATATAAGGGTAGTAAAGC Anti-sense

CN367 ctttactagtCGACTCTCGTACGACAGAAGG Anti-sense


CN368

CN416
CN417

CN424

CN425
CN426

CN427

CN428

CN429

CN437


ctttactagtaataTTACGACTCTCGTACGACAGAAGG

ctttactagtaaatattACGACTCTCGTACGACAGAAGG
ctttactagtaatattATATGGTAACATTATCACACCC

GCACGCGTCCGGAGATCTAAAGTTACAAGCAATTCCTCCAA

TTTAGATCTCCGGACGCGTGCAATTTCATGTAAGTTACCGG
AAACCATGGTAGATCTGACGGTGTGAGCAAGGGCGAGGAG

TCGCCCTIGCTCACACTAGCCACAGATCTACCATGGTTT

GGCTCGTGTTAGGCGTAGTAAGG

CCTTACTACTCCTAACACGAGCC

agagctcatatgGTGTCCTATAGGTGTCCTTG


Anti-sense

Anti-sense
Anti-sense

Sense

Anti-sense
Sense

Anti-sense

Sense

Anti-sense

Sense


aLower case letters in the 5'ends of the primers indicate the non-viral sequences such as restriction sites and
initiation codon andextrasequences incorporatedat the ends of the CTVsequences. The bold upercase letters
show the specific mutation introduced in to the overlap region.


HEL ORFIa


RdRp ORF1b


HEL

HEL-RdRp



Figure 5-1. The cloning of the CTV ORFla and lb for in vitro transcription and translation.
The CTV sequence was cloned betweenNdel and BglII sites in the pGEM-T plasmid under
the SP6 promoter. The ORFs and expected translation products are shown by the different
bars.









67

Amplification, Mutagenesis and Cloning of Overlap Region of CTV

Three fragments, called FS-1, -2 and -3 were amplified using primer pairs CN365-

367, CN365-416 and CN365-368, respectively (Figure 5-2). An ATG translation initiation

codon and Ncol restriction site at the 5' end, and a Spe I restriction site at the 3' end were

incorporated into all three fragments during PCR. In addition, the native stop codon of CTV

ORFla was removed from the 3' end of FS-1, and one and two extra nucleotides (T and TT)

were introduced after the stop codon in 3' end FS-3 and FS-2, respectively during PCR

amplification. The amplified fragments FS-1, -2 and -3 were cloned into pCambial303

(pC1303) plasmid between the cauliflower mosaic virus (CaMV) 35S promoter and the

reporter GUS-GFP gene at the NcoI and Spel sites so that the GUS-GFP gene was in the 0,

-1 and +1 frames in relation to the first ATG codon, respectively. The resulting plasmids

were designated as pC1303FS-1, pC1303FS-2 and pC1303FS- 3 (Figure 5-2 and 5-3). In

the pC1303F S-1 plasmid, the GUS/GFP gene was fused to the CTV sequences without the stop

codon in 0 frame to test the expression of GFP in the ORF la. This construct was used as the

control for the expression of the GUS-GFP. In the pC 13 03F S-2, the GUS-GFP gene was fused

to the CTV sequence (with the stop codon and added TT) in the -1 frame to test if the cloned

sequence could function as a -1 frameshift signal and induce expression of the GUS-GFP

cloned in the -1 frame. In the pC1303FS-3, the GUS-GFP gene was fused to the CTV

sequence (with the stop codon and one added T) at the +1 frame to test if the sequence of the

cloned region contained the frameshift signal and could induce expression of GUS-GFP in the

+1 frame (Figure 5-2 and 5-3).









68



(her-LAT; L )


f'1f E


F. p-


'tFN 1.


\.: I


4-


1 kiiI


-1








.pc


Gcus


in('FP5' WPis poiy-A


Figure 5-2. Amplification mutagenesis and cloning strategy used for the constructs of overlap
region of CTV ORFla and lb. The arrows with the numbers 1, 2, 3, 4, 5, and 6 indicates the
primers CN365, CN367, CN368, CN416, CN417 and CN356, respectively. The sequences
of these primers are shown in Table 5-1.


C a iMIV Z
promotq-r


I
-----c~-









69

Since the constructs pC1303FS-1, -2 and -3 contained sequences in additionto the 123

nt overlap region, it also was necessary to make individual constructs containing only the 123

nt overlap region and the upstreamhelicase sequence. Therefore, the upstream and the overlap

sequences were individually amplified by PCR using primer pairs CN365-CN417, and

CN356-CN368 (Table 5-1) and designated as FS-4 and FS-5, respectively. An ATG

translation initiation codon and Nco I restriction site at the 5' end, and an Spe I restriction site

and one extra nucleotide T at the 3' end were incorporated into both FS-4 and FS-5 during

PCR amplification. These fragments were cloned into the pC1303 between the CaMV 35S

promoter and the reporter gene at the Nco I and Spe I sites to make the GUS-GFP reporter gene

in the +1 frame in relation to the first ATG codon, and the resultant plasmids were called

pC1303FS-4 and pC1303FS- 5, respectively (Figure 5-2 and 5-3).

To study the importance ofthe conserved secondary structure identified in the overlap

region and the significance of the rare arginine codon proposed to be involved in the +1

frameshift, more constructs with specific mutations were generated by the overlap extension

PCR method (Ho et al. 1989). First, sense and anti-sense internal oligonucleotide primers

(CN424-5, CN426-7 and CN428-9) with the desired mutation were designed for each

construct. They were used with an external primer (CN356 or CN368) specific to the 3' and

5' ends of the overlap sequence for PCR amplification of two overlapping DNA fragments.

Then, these DNA fragments were mixed and used as template for a second round of PCR

amplification using the external primers (CN356 and CN368) to produce the complete (123

nt) overlap sequence with the desired mutations. Three mutant overlap sequences, FS-6, FS-7

and FS-8, were generated using this method. They were cloned into the pC1303 plasmid




















C @iih'irnK


CaMV 35S





*: ," '35S
i" ......i


of |imr CmTTS CFP


F: I i


Id 13g131\i 1


CaMV 353
I.;1 ;-1 1 ;[I


CaMV 35S
:,,', ,, ,- I


-I: .. -


pi..- j:I'(I*[\


II


I htS I


Figure 5-3. The constructs prepared for the overlap region of CTV ORF la and lb in pC1303
plasmid vector. The orange boxes indicate the 123 nt overlap sequences, and the yellow boxes
indicate the helicase sequences just upstream of the overlap region. The sequence
modifications and the translational frame for the reporter genes in relation to the first ATG
initiation codon are shown for each construct.


I i


,.., 1


1' F !: -. 1 k .,-
















'.Inu~ir $jj4 rr~ j$ I fl I r


' ,'-', i i, ,:,i A


p.. t~t'i FC-I


.'.-f 1- "' ;,, : i....; -A


in .;Y t;


II1' -


pC1303FS-8




Figure 5-4. The pC1303 plasmid vector containing the constructs of the overlap region of
CTVORF la and lb with mutations in the predicted stem-loop structure and the proposed rare
arginine codon. The orange boxes indicate the 123 nt overlap sequences, and the mutations in
this region are shownby green color in the construct and by boxes in the structures on the left.


f 4'6b t tl 1if


I!


CaMV 35S
I- I _ I


, II -


,. .ll I - -
i" ,',,,,,7,r,-








72

as were the previous constructs in which the GUS-GFP reporter gene was in +1 frame in

relation to the first ATG codon. The resulting plasmids were designated as pC1303FS-6,

pC1303FS-7 and pC1303FS-8, respectively. In the pC1303FS-6 plasmid, the rare arginine

codon CGG was changed to AGG the most frequent arginine codon in CTV, by a single

mutation. In the pC1303FS-7, the stem-loop structure was resolved by changing 12 of the 37

nt involved in the structure which resulted in only a single change in the amino acid sequence.

The pC1303FS-8 contained 30 mutations in the the 37 nt secondary structure, but this

completely different and unrelated sequence still formed an identical stem-loop structure in

the same region (Figure 5-4). All plasmids were maintained in E. coli and later introduced

into Agrobacterium strain AGL1 by triparental mating or transformation of Agrobacterium

with cold shock.

Sequence Analysis of the Overlap Region of Different CTV Isolates

The sequences of the overlapping region between the ORF a and lb from the CTV

isolates T30, T36, T66, B53, B165, B185, B249, SY568, T385,VT, and 3800 were obtained

as described in the Chapter 3. The sequences were assembled, aligned and analyzed using

GCG, and Vector NTI suite. The secondary structures for the sequences were detected and

analyzed using the RNA Draw program.

Transient Expression Assay

A previously reported Agrobacterium-mediated transient gene expression system

(Kapila et al 1997) was modified and used to study the effect of CTV sequences in the

expression of the GUS-GFP reporter genes (Figure 5-5).











Plant Materials

Initially, fully expanded leaves taken from 4-6 week old Nicotiana benthamiana and

2-3 months old Citrus sinensis var. Madam Vinus seedlings maintained at 25 C, were used

for infiltration. Later, only leaves from two months old in vitro grown Citrus paradisi var.

Duncan was used in all infiltration experiments.

Preparation of the Agrobacterium Suspension

Agrobacterium strainLBA4404 orAGL1 containing the pC1303 control plasmid, and

pC1303FS-1 to -8 frameshift test constructs were inoculated into YEP medium (10 g/1

Bactopeptone, 10 g/1 yeast extract and 5 g/1 NaCl pH 7.0) containing the appropriate

antibiotics. They were grown overnight to log phase (OD600n= 0.7-0.8) at 280 rpm and 28 C.

The cultures were centrifuged at 4 OC and 5000 rpm for 5 min, and the pellets were re-

suspended to a final concentration of5x 108 cfu/ml with MS medium containing 4.3 g/l MS salt,

100 mg/1 myo-inoistol, 30 g/1 sucrose and 150 pM acetosyringone.

Infiltrations of the Leaves with Agrobacterium Suspensions

The leaves were detached from citrus seedlings or tobacco plants and placed onto wet

and sterile Whatman filter papers fitted into petri plates. One milliliter of the Agrobacterium

suspension was infiltrated into each leaf using a 1-ml syringe without a needle (Figure 5-5).

Plates were sealed and incubated in a growth room at 25 OC and a 16 hr light/8 hr dark photo-

period for 4-5 days.

Fluorescent Microscopy

The expression of GFP in the citrus and tobacco leaves infiltrated with Agrobacterium was

first analyzed using a dissecting microscope (Zeiss) with a fluorescent light source with

















alpha



lr'


.I


i t, ir; ti !ji :,a 28 fi-ir' 2-" =4 hr


S,'-a'uob- i -rt arnt Znfesnrf,-c-'trn
4i. .t,..l. A,.g i


Figure 5-5. The flowchart of the Agrobacterium-mediated transient expression assay used to
study the involvement of the overlap region of CTVORFla and lb in the putative CTV + 1
frameshift.


PFP"









75

a 515 nmlong pass emission filter transmitting red and green light and a 450-490 nm excitation

filter. The fluorescent images were photographed using a 35 mm camera attached to the

fluorescent microscope.

Histochemical GUS Staining

The leaves were placed in small petri dishes or 2-ml eppendorf tubes containing 50

mM NaPO4 pH 7.0 10 mMNa2EDTA solutions and 5 mg/ml 5-bromo-4-chloro-3-indolyl- -

D-glucuronide (X-Gluc). A mild vaccum was applied for 5 min to distribute the substrate

equally in the cells of the leaves. The leaves were incubated in the solution overnight at 37

OC, and then they were cleared using a solution of 70% ethanol and 30% acetic acid. The

reactions were photographed with a digital camera.

Results and Discussion

In vitro Transcription and Translation

Coupled in vitro transcription and translation of the pGEMT-HEL-RdRp construct

produced two expected bands with estimated molecular weight of 82 and 30 kDa, but also a

47 kDa unexpected band (Figure 5-6). No translation product was produced with the pGEM-T

plasmid alone (data not shown). The size of the 30 kDa protein corresponds to the protein

product of the helicase domain, and the size 82 kDa protein corresponds to the fusion protein

product of both helicase (ORFla) and the RdRp (ORFlb) of CTV. Since the translation

without frameshift could only produce the 30 kDa protein, and the 82 kDa protein could not

be produced unless the frameshift occurred, the detection of the 82 kDa band demonstrates

that the +1 frameshift took place during the in vitro translation. This result clearly shows that

the RdRp gene of CTV is in fact expressed by the +1 frameshift. The amount of the 30 kDa

protein was significantly higher thenthe 82 kDa protein because the frameshift usually occurs









76

infrequently. Quantification of these proteins showed that the amount of the 82 kDa protein

was about 1-5% of the amount of 30 kDa protein, indicating the efficiency of CTV frameshift

was about 1-5%. This result is in agreement with the translation efficiency reported in other

systems (Brierley, 1995; Farabaugh, 1996), implying that CTV has a relatively efficient + 1

translational frameshift. The unexpected 47 kDa band detected in the in vitro translation

system may be the product of alternative translations of the clone, or it might be result of the

processing of the 82 kDa protein by the proteases in the wheat germ system.




1 2 82 kDa
*---


47 kDa



30 kDa








Figure 5-6. In vitro transcription and translation ofHEL-RdRp construct in wheat germ extract
demonstrating the +1 frameshift for the expression of the CTV RdRp gene. Lanes 1 and 2
contain 2% and 10% of the total protein produced by the in vitro translation of the HEL-RdRp
construct.


Development of Transient Expression Assay

Four citrus and tobacco leaves were first infiltrated with Agrobacterium tumefaciens

strain LBA4404 or AGL1 with and without plasmid pC 1303 containing the GUS/GFP fusion









77

gene in the inoculation medium. The expression of the GUS and GFP in the infiltrated leaves

was monitored daily for four days by histochemical staining and using a fluorescence

microscope. Bright green fluorescent spots and blue histochemical staining were observed,

both in citrus and tobacco leaves infiltrated with plasmid pC1303, indicating that both GFP

and GUS were transiently expressed in the inoculated leaves. The expression of both the

GUS and GFP were first detected one day after inoculation (DAI), and maximum expression

of GFP was observed two DAI. The expression of both GUS and GFP could still be detected

three DAI, but a significant reduction in the level of expression was observed three and four

DAI. No GFP expression was detected in uninoculated leaves or those infiltrated with

inoculation medium only or LBA4404 or AGL1 without the plasmid. The detection of the both

GUS and GFP only in the pC 1303-inoculated leaves indicates that green spots were not auto-

fluorescence or artifacts due to wounding of the leaves during the infiltration. These results

show that an Agrobacterium-mediated transient expression assay can be used as an efficient

system to analyze involvement of the overlapping region of CTV in the + 1 frameshift. This

systemcanbe used also to test plant transformation constructs and to study the function or the

effect of other genes and/or regulatory elements in citrus. No significant differences were

observed between citrus and tobacco leaves or betweenAgrobacterium strains LBA4404 and

AGL1. Therefore, theAgrobacterium strain AGL1 and leaves from in vitro growngrapefruit

were used in following transient assays because they were readily available and commonly

used in our laboratory.















Constructs


ni


: 1 r I.. 1 1 I il ,


CaMV 35SSPeI
S NcoI GUS mGFP5* Nos poly-A

SpC1303
CaMV 35S
promoter GUS mGFP*Nospoly-A

S pC1303FS-1

CaMV 35S
-P mn Mo I n, T GUS mGFP5* Nos POlv-A

I pC1303FS-2


CaMV 35S
prni NcoI


i I GUS

SpC1303FS-3


CaMV 35S
Eit o I e GUS

pC1303FS-4

CaMV 35S
pC 13 F GUS
I pC1303FS-5


mGFP5* Nos poly-A
1


mGFP5* N olv-A



mGFP5*Noolv-A


Figure 5-7. Expression of the GUS and GFP reporter genes in citrus leaves infiltrated with
the pC1303 containing different constructs of the overlap region of the CTV ORF a and lb.









79

Involvement of the Overlap Region of ORFla and lb in the +1 Frameshift

Citrus leaves were infiltrated with individual suspensions of Agrobacterium strain

AGL1 harboring the control and frameshift plasmids pC 1303, pC 1303F S-1, pC1303F S-2 and

pC1303FS-3. The expression of the GUS and GFP was detected in citrus leaves inoculated

with pC 1303 and pC 13 03F S-1 as well as with pC 13 03F S-3, the +1 frameshift tester plasmid.

However, no GUS or GFP expression was detected in uninoculated leaves or leaves

infiltrated with inoculation medium or AGL1 without plasmid or with pC1303FS-2, the -1

frameshift tester plasmid (Figure 5-6). In the pC1303FS-3 construct, the GUS-GFP gene was

fused to the overlapping sequence at +1 frame in relation to the ATG translation initiation

codon (Figure 5-5), and the GUS and GFP could not be expressed unless the +1 frameshift

took place. Therefore, the expression of the GUS and the GFP in citrus leaves infiltrated

with pC1303FS-3 plasmid demonstrates that the +1 frameshift was induced by the CTV

sequence fromthe overlap and its upstreamregion. Since the pC1303FS-2 plasmid containing

essentially the same sequence did not show any expression of the GUS or GFP in 1 frame,

itwas shownthatthe CTV sequence did not induce the -1 frameshift, and the signals contained

in the sequence were specific to the +1 ribosomal frameshifting (Figure 5-7). High levels of

GUS and GFP expression were observed in citrus leaves infiltrated with the pC1303FS-1.

This was the 0 frame test construct in which the GUS/GFP gene was fused to CTV sequences

without the stop codon; thus, expression was expected regardless of the +1 frameshift (Figure

5-7). Even though a variation was observed in the number of cells expressing the GUS and

GFP with this construct and the control plasmid pC1303, the level of expression in the

individual cells was similar indicating that reporter gene expression was not dramatically

affected by the insertion of the CTV sequences. Since frameshifting generally occurs at low









80

frequency, the expression the GUS and GFP in the leaves infiltrated with the pC1303FS-3

plasmid was expected to be lower thanin leaves infiltrated with the pC1303 and pC1303FS-1

control plasmids (Figure 5-7). Although the number of blue stained and green fluorescence

spots were significantly lower in the pC1303FS-3 than the control plasmids, the expression

of the reporter genes could not be analyzed quantitatively because their expression levels

depends on and is affected by variables such as efficiency of infiltration and transformation.

Nevertheless, the assay is very simple and efficient for qualitative analysis of whether of not

the +1 frameshift occurs with specific constructs.

Since the constructs pC1303FS-1, 2 and 3 contained about 280 nt of CTV sequence,

including the 123 nt overlap region and about 160 nt upstream sequences, it was not

demonstrated whether the 123 ntoverlap sequence by itself is able to induce the +1 frameshift.

To determine that, citrus leaves were infiltrated with Agrobacterium carrying pC1303FS-4

and pC1303FS-5 as well as control plasmids pC1303 and pC1303FS-1. The expression of

the GUS and GFP was observed in leaves infiltrated with control plasmids pC1303 and

pC1303FS-1, and the pC1303FS-5 plasmid containing only the 123 nt overlap sequence of

CTV ORF la and lb (Figure 5-7). On the other hand, no expression of either reporter gene was

detected in leaves infiltrated with the pC1303FS-4 plasmid containing the about 160 nt

sequence fromthe helicase domainjustupstreamof the overlap region of CTVORFla and lb

(Figure 5-7). This result shows thatthe 123 nt overlap region is necessary and sufficient for

the induction of +1 frameshift in the transient assay. This result is consistent with the

information in the literature that the signals or sequence motifs required for






















KflRT1


.Pla I h


KFR4iR


. -
T T

T 7. T ,-




T..


H IIml ft1 ?vI


Figure 5-8. Multiple alignment and homology graph of nucleotide sequences of the overlapping
region ofORF a and lb of 10 biologically and geographically different isolates of CTV. The
alignments and the graphs were generated using the Align X program of Vector NTI Suite. The
letters W, Y and R indicates T or A, T or C and G or A, respectively. The homology graph
displays alignment quality based on the total similarity values (1, 0.5 and 0.2 for identical,
similar or weakly similar, respectively) for the residues at a given position..




Table 5-2. The percentage ofnucleotide sequence identity of the overlap regionofORF la and
lb among ten different isolates of CTV.








'* .: -
Ir tJ r "


:., i. I'.'. .%


11. :;;, i .'; :;') 01.1
"H : 'il> :-: |J1 ..- .,. ".

"^ Sol 4. ;! P :;; '


I7 I
I i. I r
:* ".... . ,.. .. i* .. ..,... i.- .. , .. .. i '.: .... L. .,.. i' ... ..,...: .. ....,^ .


. 'T
** r .1 *111 ."l~lli ? i i :Y
*,, T '"I '"* "."."." T- ' I ,, ,'. ',,, ,.,,,.,,.,,. I









82

frameshifts in most systems are located in the overlap region of two ORFs (Brierley, 1995;

Farabaugh, 1996). It also has been reported that signals or motifs required for a+1 frameshift

can be composed of a complex secondary structure (Lee et al. 1996) or simple sequences as

few as seven nt (Belcourt and Farabough, 1990) located in the overlap region of two ORFs.

Further analysis of the overlap region of CTV ORFla and lb was necessary to determine if

specific sequence elements or motifs of the overlap region were involved in the +1 frameshift

and required for the expression of the RdRp gene of CTV.

Sequence Comparison and Secondary Structure Analysis of the Overlap Region

In order to identify possible sequence elements or motifs that may be present in the

overlap region and required for the +1 frameshift, sequences of the overlapping region from

six different CTV isolates (T30, B53, B165, B185, B249 and 3800) were determined and

compared with each other and with the four sequences available in the GenBank, T36, SY568,

VT and T385. Multiple alignment and comparison of the sequences indicated that the

sequences of overlap regions from biologically and geographically different isolates of CTV

are highly conserved (Figure 5-7) and show 84-100% sequence identity among different

isolates (Table 5- 3). The most highly conserved sequences were surrounding the rare

arginine codon proposed to be involved in pausing the ribosome during the frameshift and at

the 3' end of the overlapping region towards the stop codon of the ORF la. These regions

showed the highest homology and scores (Figure 5-8).

Secondary structures such as pseudoknots and stem-loops have been shown to be

involved in frameshifting in other systems. To investigate the presence of secondary structures

inthe overlap region, the possible folding patterns of the highly conserved overlap sequences

were analyzed in the RNA Draw program. The secondary structure analyses of the











33 I. -
T
T385 .
B185

VT






1' I
T
3800 T . .,'.
Consensus CA C GAA C C GC TC GC C-'TT ',1:1: ,' ,TT .-_'..- ;: TI .












Consensus 3801) B165 B185 B249 B53 S1563 T30 B36 T385 VT

Figure 5-9. Possible folding patterns of the highly conserved nucleotide sequence around the
arginine codon in the overlapping region of the ORF 1 a and lb fromdifferent isolates of CTV.
The secondary structures were generated using the RNA Draw program the letters Y and R
indicates T or C and G or A, respectively.


overlap region ofisolate 3800 revealed the presence of a stem-loop structure composed of37

highly conserved nucleotides between positions 9387 and 9423 which is around the arginine

codon proposed to be involved in pausing the ribosome during frameshifting. An almost

identical stem-loop structure was also found in the same region of the ten biologically and

geographically different isolates of CTV (Figure 5-9). Since stem-loop structures were

reported to be involved in frameshifting in other viruses, and the structure was conserved in

all isolates of CTV, it was possible that the stem-loop structure might also be involved in the

+1 frameshift for the expression of the RdRp of CTV. Although the + 1 frameshift

characterized in Ty-1 retro transposon does not require any secondary structure, the presence









84

and requirement of a stem loop structure was reported for Leishmania dsRNA virus (Lee et

al. 1996). Therefore, it was worthwhile to study the possible role of the conserved secondary

structure found in the overlap region of CTV ORF1 an lb.

The Role of the Secondary Structure and the Rare Arginine Codon in the +1 Frameshift

To study the requirement of the conserved stem-loop structure and the rare arginine

codon in the +1 frameshift, citrus leaves were infiltrated with Agrobacterium containing

plasmids pC1303FS-6, pC1303FS-7 and pC1303FS-8 as well as control plasmid pC1303

and pC1303FS-1. The expression of GUS and GFP were observed in leaves infiltrated with

control plasmids pC1303, pC1303FS-1 as well as with plasmids pC1303FS-6 and

pC1303FS-8. The detection of the GUS and GFP expression in leaves infiltrated with

pC 13 03F S-6 demonstrated that the mutation in the rare arginine codon (CGG to AGG) did not

eliminate the frameshift which indicates thatthe rare arginine codon by itself i s neither the only

signal nor the major component of the sequence required for the +1 frameshift in this assay.

The expression of the reporter genes in the leaves infiltrated with plasmid pC1303FS-8

containing a modified overlap region with a stem-loop structure composed of a completely

different sequence than the native overlap region, but not with pC1303FS-7 containing the

overlap sequence without the conserved stem-loop structure demonstrates that the secondary

structure is important for the frameshift. The elimination of the +1 frameshift with mutation in

12 of the 37 ntinvolved inthe secondary structure indicates that either the secondary structure

itself or the mutated 12 nt sequence or the combination ofboth the sequences and the stem-loop

structure is required for the +1 frameshift. Additional




















Constructs


*-)ir !.r.: : : /..r I...


- .,I


. I; ., 1 ,


,I':


1 Io. f


CaMV: -. -
I -,,;,-,t


. -s ._- i -i


Figure 5-10. Expression of the reporter genes GUS and/or GFP in citrus leaves infiltrated
with pC1303 plasmids containing the native and mutant overlap region sequences of ORF a
and lb. The orange color indicates the native sequences and the mutations are shown by green
color over the orange.


I I i


.......
.............


( -1 1 I -


--_! i ,


I p F -1 -, P -









86

constructs with more specific mutations in the stem-loop structure or with specific deletions

or more specific mutations in the sequence could provide a better understanding of the role

of the overlap sequences in the +1 frameshift.

Conclusions

The requirement of the +1 frameshift for the expression of RdRp gene of CTV was

demonstrated using a coupled in vitro transcription and translation assay. This assay

demonstrated that the efficiency of the +1 transactional frame shift was about 1-5% An

Agrobacterium-mediated transient expression assay was adopted and used to analyze the

effect of the overlapping region of ORFla and lb of CTV in the +1 frameshift. The assay is

a rapid and efficient way to analyze CTV sequences fused to the GUS-GFP bi-functional

reporter gene in citrus leaves. Using the transient expression assay, it was demonstrated that

the overlap region of ORFla and lb of CTV promotes the +1 frameshift required for the

expression of the CTV RdRp gene. The 123 nt overlap sequence itself is necessary and

sufficient to induce the +1 frameshift, as indicated by the expression of both the GUS and GFP

reporter gene cloned in the +1 frame position in relation to the ATG translation initiation

codon. It was shownthat the upstream sequence from the helicase domain was not necessary

for the + 1 frameshifting. Involvement of the overlap region of the CTV ORFla and lb is

consistent with the other frameshifting systems where signals or the sequences required for the

frameshifting are located in the overlap region ofthe two ORF where one ofthem is expressed

by frameshifting.

Sequence analyses of the overlap region of the CTV ORFla and lb revealed that the

region is highly conserved among geographically and biologically different isolates of CTV.

A conserved secondary structure is present in the overlap region of all the isolates of CTV









87

examined. Sequence modification of the secondary structure eliminated the +1 frameshift,

indicating that the stem-loop structure is required for the frameshift. On the other hand, the

rare arginine codon was notrequired and even may notbe involved in the frameshift, because

the mutation of the rare arginine (CGG) to a more frequent arginine codon (AGG) did not have

any effect on the frameshift in the transient assay. Although the individual elements of the

frameshift signal remain to be determined, it was clearly shown by both in vitro translation

and the Agrobacterium-mediated transient assay that the RdRp gene of CTV is expressed by

the +1 frameshift induced by the overlap sequences ofCTVORFla and lb. Since it has been

reported that the +1 frameshift of Ty-1 in yeast is specifically inhibited by pokeweed

antiviral protein (Tumer et al. 1998), the inhibition of the +1 frameshift with this protein or

other molecules or genes provides a specific target for engineering resistance to CTV.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs