Molecular mechanisms of host specific virulence and avirulence caused by cotton blight and citrus canker pathogens

MISSING IMAGE

Material Information

Title:
Molecular mechanisms of host specific virulence and avirulence caused by cotton blight and citrus canker pathogens
Physical Description:
v, 129 leaves : ill., photos ; 29 cm.
Language:
English
Creator:
Yang, Yinong, 1962-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1994.
Bibliography:
Includes bibliographical references (leaves 115-128).
Statement of Responsibility:
by Yinong Yang.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002028118
notis - AKL5729
oclc - 33032693
System ID:
AA00003644:00001

Full Text








MOLECULAR MECHANISMS OF HOST SPECIFIC VIRULENCE
AND AVIRULENCE CAUSED BY COTTON BLIGHT
AND CITRUS CANKER PATHOGENS







By

YINONG YANG


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



UNIVERSITY OF FLORIDA


1994











ACKNOWLEDGMENTS


I wish to express my appreciation and gratitude to Dr. Dean W. Gabriel,

chairman of the supervisory committee, for his time, guidance and support during the

course of this study. Appreciation is also extended to Drs. D. R. Pring, H. C. Kistler,

L. C. Hannah, and R. A. Jensen for serving on the supervisory committee and critical

reading of this manuscript. Special thanks go to Dr. W. B. Gurley for reading my

dissertation and participating in my final examination.

I wish to thank Drs. Robert De Feyter, Mark T. Kingsley and Mr. Gary Marlow

for stimulating scientific discussions and technical assistance. I gratefully acknowledge

the help of Dr. Michael Paddy in fluorescence microscopy and Dr. Suzy Cocciolone in

microprojectile bombardment. I also greatly appreciate the technical assistance of

Qiaoping Yuan, Ruhui Li and Blanca Garagorry.

I would like to thank my parents for their understanding and encouragement.

Finally, I must thank my wife, Qin Wang, and daughter, Catherine, for their

companionship and strong support.












TABLE OF CONTENTS


ACKNOWLEDGMENTS ................................... ii


ABSTRACT

CHAPTERS


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


2. HOST-SPECIFIC SYMPTOMS AND INCREASED
RELEASE OF XANTHOMONAS CITRI AND
X. CAMPESTRIS PV. MALVACEARUM
FROM LEAVES ARE DETERMINED BY
THE 102 BP TANDEM REPEATS OF
pthA AND avrb6, RESPECTIVELY .....


. . 9


3. INACTIVATION OF MULTIPLE AVIRULENCE
GENES IN XANTHOMONAS CAMPESTRIS
PV. MALVACEARUM CREATED A
NONPATHOGENIC ENDOPHYTE
OF COTTON ........................... 37

4. INTRAGENIC RECOMBINATION OF A SINGLE
PLANT PATHOGEN GENE PROVIDES A
MECHANISM FOR THE EVOLUTION
OF NEW HOST SPECIFICITIES ............... 61

5. PLANT NUCLEAR TARGETING SIGNALS ENCODED
BY A FAMILY OF XANTHOMONAS
AVIRULENCE/PATHOGENICITY GENES ........ 87

6. SUMMARY AND CONCLUSIONS .................. 111

LITERATURE CITED ................................... 115

BIOGRAPHICAL SKETCH ................................... 129











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

MOLECULAR MECHANISMS OF HOST-SPECIFIC VIRULENCE
AND AVIRULENCE CAUSED BY COTTON BLIGHT
AND CITRUS CANKER PATHOGENS

By

Yinong Yang

August 1994


Chairman: Dean W. Gabriel
Major Department: Plant Molecular and Cellular Biology

Several members of a Xanthomonas avirulence (avr) gene family were shown to

play an important role in determining host-specific pathogenicity (pth) of cotton blight

and citrus canker pathogens. X. campestris pv. malvacearum strain XcmH1005, the

cotton blight pathogen, carries twelve members of what is now known as an avr/pth gene

family. Six plasmid-bome members were previously shown to confer avirulence on

cotton in a gene-for-genes manner. Four out of six newly cloned chromosomal members

also conferred avirulence on cotton. To investigate possible pleiotropic functions of these

avr genes in XcmH1005, marker exchange mutagenesis and complementation analyses

were carried out. Gene avrb6 was found to be a major pth gene of XcmH1005 and was

required for severe pathogenic symptoms on cotton. Other avr genes in XcmH1005 also

had an additive effect on pathogenicity. These avr genes did not contribute to bacterial

iv






growth in plant, but were essential for the induction of host-specific symptoms and

release of the pathogen onto the leaf surface, strongly indicating their role in pathogen

dispersal. Another member of the same gene family, pthA of X. citri, was previously

shown to be essential for symptoms of citrus canker disease and release of the pathogen

to the leaf surface. By constructing chimeric genes among pthA and six other avr/pth

genes, both avirulence and pathogenic specificities were found to be determined by the

102 bp tandem repeats present in the central region of all members of the gene family.

Furthermore, intragenic recombination was demonstrated to provide a mechanism for the

generation of new avr/pth genes with novel host specificities. Sequence analysis revealed

that pthA and avrb6 encoded proteins containing three nuclear localization sequences in

the C-terminal region. The predicted PthA and Avrb6 proteins were constitutively

expressed in Xanthomonas as detected by Western analysis using anti-PthA antibody.

When transiently expressed in plant cells, p-glucuronidase fusion proteins containing the

C-terminal region of PthA or Avrb6 were observed to localize specifically in the nucleus.

A working hypothesis is proposed that Avr/Pth proteins enter the plant cell nucleus and

directly affect gene expression, leading to the induction of host-specific symptoms.











CHAPTER 1
INTRODUCTION


Disease resistance of plants and virulence of microbial pathogens are governed

by their combined genotypes. During the process of reciprocal evolution, many plant

and microbial genes have coevolved to actively engage in pathogenic interactions between

host and pathogen, thus determine the final outcome of plant diseases. Successful

infection by microbial pathogens may require surface attachment, degradation of host

chemical and physical barriers, production of toxins, and inactivation of plant defenses

(Lamb et al. 1989). Overall, up to 100 genes may be required for bacterial pathogenesis

(Panopoulos and Peet 1985, Daniels et al. 1988), including those involved in bacterial

growth in plant, induction of pathogenic symptoms, host range and avirulence (Gabriel

1986). On the other hand, host plant resistance to infection involves preformed barriers,

inactivation of microbial toxins, recognition of pathogen signals and induction of active

plant defenses. In incompatible plant-bacteria interaction, induced resistance is usually

associated with a hypersensitive response (HR), characterized by the rapid, localized

death of plant cells at the site of infection. This process requires activation of plant

resistance genes involving signal perception and transduction and plant defense genes

involving biosynthesis of pathogenesis-related proteins, phytoalexins and other antibiotics

(Dixon and Harrison 1990, Dixon and Lamb 1990).






2

Bacterial plant pathogens require virulence genes to infect and to colonize plants.

A large group of genes, called hypersensitive response and pathogenicity (hrp) genes,

have been found generally conserved in all phytopathogenic bacteria examined (e. g.,

Erwinia, Pseudomonas and Xanthomonas) (Willis et al. 1991). Whatever the bacterial

pathogens studied, typical hrp- mutants are completely nonpathogenic and elicit neither

pathogenic symptoms on host plants nor hypersensitive response on nonhost plants. Most

hrp genes are clustered on 20-30 kb DNA fragments and are conserved at the genus or

family level. High levels of DNA homology are found among many hrp genes from

different genera of phytobacteria (Boucher et al. 1992). Several biological functions are

encoded by different hrp genes. Some hrp genes (e.g., hrpS of P. syringae pv.

phaseolicola) are clearly regulatory, responding to plant signals to induce transcription

of hrp operon in a manner that may be analogous to the two-component signaling systems

(Rahme et al. 1991). Other hrp genes are involved in the synthesis and export of

pathogenicity factors that also induce an HR on nonhost plants. DNA sequence analysis

reveals that several hrp genes of Pseudomonas and Xanthomonas are remarkably similar

to key virulence genes of animal bacteria such as Yersinia and Shigella, which are

involved in export of proteinaceous pathogenicity factors (Gough et al. 1992, 1993;

Fenselau et al. 1992). Some hrp genes of Erwinia and Pseudomonas were shown to

encode proteinaceous HR elicitors that are secreted via Hrp-dependent pathway (Wei et

al. 1992; He et al. 1993). Since HR on host plants (race specificity) conferred by

avirulence (avr) genes is dependent on hrp function, it was proposed that Hrp proteins






3
may also involve export of avirulence elicitors (Fenselau et al. 1992). Therefore, hrp

genes may control or condition the functions encoded by virulence and avirulence genes.

Bacteria belonging to the genus Xanthomonas are plant-associated and infect 392

monocot and dicot species in 68 families (Leyns et al. 1984). Although the genus as a

whole has wide host range, individual strains have host ranges limited to only certain

plant species. This phenomenon of host specialization indicates that establishment of

Xanthomonas infection involves host-specific recognition in addition to general virulence

function of pathogen conferred by hrp genes. Host-specific virulence (hsv) genes

encoding positive function have been shown to be a major determinant for host range and

pathovar status of Xanthomonas (Gabriel et al. 1993). These genes are required for

bacterial pathogens to grow and/or to induce pathogenic symptoms on specific host

plants, but not in other plant species. Several host-specific virulence genes have been

isolated from X. campestris pv. citrumelo (Kingsley et al. 1993), X. c. pv. translucens

(Waney et al. 1992) and X. citri (Swarup et al. 1991). For example, opsX of X. c. pv.

citrumelo is involved in biosynthesis of lipopolysaccharide and extracellular

polysaccharide and is required for virulence on citrus, but not on bean (Kingsley et al.

1993). These hsv genes are analogous to host-specific nodulation (hsn) genes of

Rhizobium. The hsn genes of Rhizobium is involved in host-specific modification of

lipooligosaccharide Nod factors and function only for nodulation on specific hosts (Fisher

and Long 1992; Denarie et al. 1992).

Besides hsv genes, avr genes have been shown to play a role in determining host

specificity at the species level (Keen 1990; Heath 1991). Single cloned avr genes from






4
a pathogen of one host species can cause an otherwise virulent pathogen of another host

species to become avirulent on its own host (Kobayashi et al. 1989; Whalen et al. 1988).

Therefore, host species specificity is conditioned by both host-specific virulence and

avirulence genes with the former as the major determinant.

Although avr genes may condition host range at the species level in some cases,

they act as negative factors to confer avirulence and to limit host range only if the host

plants have specific resistance (R) genes. The interaction of pathogens with avr genes

and host plants with specific R genes results in the plant defense response, often known

as hypersensitive response. This genetic requirement for specific plant R genes and

specific microbial avr genes is called a gene-for-gene interaction. The gene-for-gene

interactions were first demonstrated by Flor (1942, 1946, 1947) using flax and flax rust

and have since been observed in many plant-parasite interactions between fungi, bacteria,

virus, nematodes and their respective hosts (Thompson and Burdon 1992).

During the past decade, many avr genes have been isolated from different

bacterial and fungal plant pathogens (Gabriel et al. 1993; Keen 1990). In some

instances, avr genes encode proteinaceous elicitors that directly induce the HR on

resistance plants. For example, avr4 and avr9 of the fungal pathogen Cladosporium

fulvum encode race-specific peptides which are excreted into the apoplast of infected

leaves and induce the HR on tomato cultivars that carry resistance genes Cf4 and Cf9,

respectively (Joosten et al. 1994; van Kan et al. 1991). In another case, avrD of

Pseudomonas syringae pv. tomato appears to encode a enzyme involved in synthesis of

a race-specific glycolipid elicitor (Keen et al. 1990; Kobayashi et al. 1990, Midland et






5
al. 1993). In most cases, however, the biochemical identities of avr gene products are

unknown and the sequence analyses of avr genes are uninformative as to how they might

interact with R gene products to trigger the HR and specify gene-for-gene interaction.

Several models (including elicitor/receptor model, dimer model and ion channel

defense model) have been proposed to explain the molecular and biochemical bases of

gene-for-gene recognition between plant and pathogen (Gabriel and Rolfe 1990). But

very little experimental evidence is available to prove any of the hypotheses. Recently,

the first plant R gene, Pto, has been isolated from tomato by map-based positional

cloning (Martin et al. 1993). This gene encodes a serine/threonine protein kinase that

is homologous to the receptor protein kinase SRK6 involving Brassica pollen-stigma

recognition and to the Raf protein kinase involving the mammalian Ras signaling

pathway. The Ras pathway is the best known circuit of signal transduction found in

yeast, insect, as well as mammalian systems (Culotta and Koshland 1993). The Ras

pathway begins with extracellular signals (e. g. growth factors or cytokines) that interact

with receptor protein kinases, which in turn trigger phosphorylation cascades and finally

expression of targeted genes (Crews and Erikson 1993).

Although Ras-like signaling pathway has not been demonstrated in plants,

homologues of mammalian signal transduction components were found in many plants.

Those include homologues of G-protein subunits (e. g. auxin-regulated ArcA of tobacco,

Ishida et al. 1993), receptor protein kinases (e.g. pollen-specific SRK of Brassica,

Goring and Rothstein 1992), Ras-related proteins (e.g. Rgpl of rice, Kamada et al.

1992), Raf (e.g., Pto of tomato, Martin et al. 1993; and ethylene-regulated CTR1 of






6

Arabidoposis, Kieber et al. 1993) and mitogen-activated protein kinase (e.g., MsERK1

of alfalfa, Duerr et al. 1993). It is possible that virulence and avirulence gene products

may activate the similar signal transduction pathway in plants, leading to disease

symptoms or hypersensitive response. For example, transcriptional activation of plant

defense genes is modulated by phosphorylation (Felix et al. 1991; Yu et al. 1993) and

can be blocked by inhibitors of mammalian protein kinases (Raz and Fluhr 1993). It was

proposed that recognition of avr signals by R gene products would trigger a

phosphorylation cascade, leading to plant defense responses (Lamb 1994).

In this study, two economically important, world-wide diseases caused by

Xanthomonas have been used as model systems. Cotton bacterial blight is caused by X.

campestris pv. malvacearum; symptoms include angular, watersoaked spots on leaves,

black arm on stems and bollrot (Verma 1986). Citrus canker is caused by X. citri;

symptoms include erumpent, corky hyperplasia on leaves, stems and fruit (Schoulties et

al. 1987; Stall and Civerolo 1991). Recently, members of a Xanthomonas avirulence

(avr) and pathogenicity (pth) gene family were found to play an important role in

determining host-specific virulence and avirulence caused by these two pathogens (De

Feyter and Gabriel 1991a; De Feyter et al. 1993; Swarup et al. 1991, 1992). For

example, a pathogenicity gene, pthA, is required for X. citri to elicit hyperplastic lesions

on citrus (Swarup et al 1991). It also functions as an avr gene to confer the ability to

elicit an HR on cotton and bean (Swarup et al. 1992). Members of this gene family also

include avrB4, avrb6, avrb7, avrBIn, avrBlO1 and avrB102 of X. campestris pv.

malvacearum (De Feyter and Gabriel 1991a; De Feyter et al. 1993), avrBs3, avrBs3-2






7
and avrBsP of X. campestris pv. vesicatoria (Bonas et al. 1989, 1993; Canteros et al.

1991), and avrxa5, avrXa7 and avrXalO of X. oryzae (Hopkins et al. 1992). The most

conspicuous feature of this gene family is the nearly identical, tandemly arranged, 102bp

repeats in the central portion of genes (Fig. 1-1). Deletion analyses of avrBs3 have

shown that the 102bp repeated motifs determines the avirulence specificity of the gene

(Herbers et al. 1992). However, many questions remain regarding to their pleiotropic

functions, specificity, evolution, and biochemical mechanism of gene-for-gene

recognition.

The primary objectives of this work were to determine (1) why avirulence genes

are present in the cotton blight pathogen and whether they confer any pleiotropic

functions, (2) what determines avirulence and pathogenic specificity in cotton blight and

citrus canker pathogens, (3) how the specificity is created and evolved in this avr/pth

gene family, and (4) how the signals encoded by the avr/pth genes are perceived and

transduced in plant cells.























Gene Structure


# 102bp
Repeats


B P St H BS
I I I I I II

B P St H BS


BP St H BS
iI I 1 I II


H BS


17.5


13.5

17.5


15.5


Pleiotropic
Function


cankering


watersoaking


Figure 1-1. A Xanthomonas avr/pth gene family. Listed are four representive members:
pthA of X. citri, avrb6 of X. campestris pv. malvacearum, avrBs3 of X. campestris pv.
vesicatoria, and avrXalO of X. oryzae pv. oryzae. The tandemly arranged blocks
represent the 102-bp repeats in the central portion of genes. Restriction enzyme cleavage
sites, relevant to this work and found in most members of the gene family, are: B,
BamHI; P, PstI; St, StuI; H, HincII; S, Sstl.


Name


pthA


avrb6

avrBs3


avrXa 10


BP St


I I
1


s











CHAPTER 2
HOST-SPECIFIC SYMPTOMS AND INCREASED RELEASE OF XANTHOMONAS
CITRI AND X. CAMPESTRIS PV. MALVACEARUM FROM LEAVES
ARE DETERMINED BY THE 102BP TANDEM REPEATS
OF pthA AND avrb6, RESPECTIVELY


Introduction



Microbial genes involved in plant-microbe interactions may be functionally

classified into four broad categories: parasitic, pathogenic, host range and avirulence

(Gabriel 1986). Genes involved in parasitism are absolutely required for growth in

plant and are widely conserved at the family or genus level. Examples include most

hrp (hypersensitive response and pathogenicity) genes of Erwinia, Pseudomonas and

Xanthomonas, and the common nod nodulationn) genes of Rhizobium (Boucher et al.

1992, Denarie et al. 1992, Willis et al. 1991). Genes involved in pathogenicity are

required for induction of symptoms. Examples include pectate lyase, polygalacturonase

and endoglucanase genes (Collmer and Keen 1986, Schell et al. 1988, Roberts et al.

1988), dsp (disease specific, Arlat and Boucher 1991), wts (watersoaking, Coplin et al.

1992) genes, phytohormone biosynthetic genes (Smidt and Kosuge 1978) and toxin

biosynthetic genes (Mitchell 1984). Genes involved in conditioning host range are host

specific and required for growth on specific hosts. Examples include hsv (host-specific

virulence), pth pathogenicityy), and hsn (host-specific nodulation) genes of Pseudomonas,






10
Xanthomonas and Rhizobiwn (Denarie et al. 1992, Kingsley et al. 1993, Ma et al. 1988,

Salch and Shaw 1988, Swarup et al. 1991, Waney et al. 1991, Gabriel et al., 1993).

Information on conservation of these genes within species, pathovars and biovars is

scarce, but these genes appear to determine biovar and pathovar status (Djordjevic et al,

1987; Gabriel et al, 1993). The fourth group are termed avirulence (avr) genes because

they negatively affect virulence. The avr genes are superimposed on basic compatibility

(Ellingboe 1976), are not highly conserved, and determine pathogenic races below the

species, biovar or pathovar levels. These four broad categories are not mutually

exclusive.

The interaction of microbes having avr genes and host plants having resistance

(R) genes can result in plant defense responses, often observed visually as a

hypersensitive reaction (HR) and characterized by the rapid necrosis of plant cells at the

site of infection and the accumulation of phytoalexins. Most avr genes do not appear to

confer selective advantage to the pathogen (Gabriel 1989, Keen and Staskawicz 1988).

Pleiotropic functions have been identified for only 3 of the 30 avr genes cloned to date

(Gabriel et al, 1993). Furthermore, the DNA sequences of the cloned avr genes have

been remarkably uninformative in terms of function (Keen, 1990). The presence of most

avr genes in plant pathogens therefore remains enigmatic.

Recently, an avr gene family has been discovered in many different

xanthomonads; members include avrBs3, avrBs3-2 and avrBsP (Bonas et al. 1989, 1993;

Canteros et al. 1991) of X. campestris pv. vesicatoria, avrB4, avrb6, avrb7, avrBIn,

avrBlOl and avrBl02 of X. campestris pv. malvacearum (De Feyter and Gabriel 1991a,






11
De Feyter et al. 1993) and avrxa5, avrXa7 and avrXalO of X. oryzae (Hopkins et al.

1992). Interestingly, this gene family includes a gene, pthA, that is required for

pathogenicity of X. citri on citrus. This gene is not known to function for avirulence in

X. citri (Swarup et al. 1992), but is required for X. citri to induce cell divisions in the

leaf mesophyll of citrus, leading to epidermal rupture and subsequent release of the

bacteria onto the leaf surface. The gene also confers the ability to induce cell divisions

on citrus to X. campestris strains from several different pathovars (Swarup et al. 1991,

1992).

De Feyter and Gabriel (1991a) observed that avrb6 and avrb7 enhanced the

watersoaking ability of several X. campestris pv. malvacearum strains on cotton, but the

role of these genes in pathogenicity was not determined. The family therefore consists

primarily of avr genes, but includes at least one, and perhaps more, host-specific

pathogenicity genes. The most conspicuous feature of this highly homologous gene

family is the presence of nearly identical, tandemly arranged, 102bp repeats in the central

region of the genes. These repeats are known to determine the gene-for-gene specificity

of avrBs3 (Herbers et al. 1992). The purpose of this study was to characterize the

watersoaking functions of avrb6 and to investigate the role of the 102bp repeats of pthA,

avrB4, avrb6, avrb7, avrBIn, avrBlO1 and avrB102 in pathogenicity and avirulence.








Materials and Methods



Bacterial strains, plasmids. and culture media. The bacterial strains and plasmids

used in this study are listed in Table 2-1. Strains of Escherichia coli were grown in

Luria-Bertani (LB) medium (Sambrook et al. 1989) at 37*C. Strains of Xanthomonas

were grown in PYGM (peptone-yeast extract-glycerol-MOPS) medium at 30*C (De

Feyter et al. 1990). For culture on solid medium, agar was added at 15g/L. Antibiotics

were used at the following final concentrations (mg/L): ampicillin (Ap), 25; kanamycin

(Km), 20; gentamycin (Gm), 3; spectinomycin (Spc), 50; tetracycline (Tc), 15; and

rifampicin (Rif), 75.

Recombinant DNA techniques. Total DNA isolation from Xanthomonas was as

described (Gabriel and De Feyter 1992). Plasmids were isolated from E. coli by alkaline

lysis methods (Sambrook et al. 1989). Restriction enzyme digestion, alkaline

phosphatase treatment, DNA ligation and random priming reactions were performed as

recommended by the manufacturers. Southern hybridization was performed by using

nylon membranes as described (Lazo et al. 1987). Otherwise, standard recombinant

DNA procedures were used (Sambrook et al. 1989).

Construction of chimeric genes. To construct BamHI fragment-swapped chimeric

genes among avrB4, avrb6, avrb7, avrBIn, avrBlO1, avrB102, the BamHI fragments

from these genes were cloned into an avrb7 shell (containing 5' and 3' ends of avrb7,

but deleted for its BamHI fragment, pUFR163ABam) and an avrBIn shell (containing 5'

and 3' ends of avrBIn, but deleted for its BamHI fragment, pUFR186ABam) on the









Table 2-1. Bacterial strains and plasmids used in this study.


Strain or Plasmid Relevant Characteristics Reference or Source


E. coil


DH5-


HBI01


ED8767


Xandwhownas citi


3213
B21.2


X phaseoUl


G27
G27Sp


X campesinos pv. citrumelo


3048
3048Sp


X coampestris pv. alfalfa


KX-ISp


F-, endAl. hsdI7(rk(mk), supE44,
ihi-1, recAl, gyrA, reLAl, 80dlacZ
AM15, A(acZYA-arwF)U169
supE44, hsdS20(rmk+), vcAI3,
ara-14, proA2, lacYl, galK2, rpsL20,
xyl-5, ma-I
supE44, supF8, sdS3(rk-mk+), recAS6,
galK2, galT22, metBl



ATCC49118, citrus canker type strain
pthA::Ta5-gusA, marker-exchange
mutant of 3213


ATCC49119, bean blight type strain
Spc' derivative of 027



ATCC49120, citrus leaf spot pathotype strain
Spc' derivative of 3048



Spc' derivative of KX-1, isolated from
alfalfa, causing citrus leaf spot


Gibco-BRL,Gaithersburg,MD


Boyer & Roulland-Dussoix 1969



Murray et al. 1977


Gabriel ea a. 1989
Swamp et at. 1991



Gabriel et al. 1989
Swarup et al. 1991



Gabriel et al. 1989
Swarup et al. 1991



Swarup et ai. 1991


X campesris pv. malvacearum


Natural isolate from cotton from
Oklahoma; carries six avr genes used in this
study on pXcmH, plus additional avr genes
Spontaneous Rif derivative of XcmH
avrb6::TnS-gutA, marker-exchange
mutant of XcmHIO05
avrb7::Tan-gusA, marker-exchange
mutant of XcmHI005
avrBln::TnS-gusA, marker-exchange
mutant of XcmHI005
Natural isolate from cotton from
Upper Volta, Africa
Spc' Rif derivative of XcmN



ColEl, Kn', Tra+, helper plasmid
pRK2013 derivative, npt::Tn7,Kmn, Sp',
Tra+, helper plasmid
IncW, Kin', Gmn, Mob+, acZa+, Par+,
IncW, Gmn, Apr, Mob+, lacZa+, Par+


De Feyter & Gabriel, 1991a


This study
This study

This study

This study

Gabriel et al. 1986

DeFeyter & Gabriel 1991a



Figurski & Helinski, 1979
Leong et al. 1982

DeFeyter & Gabriel 1991a
De Feyter et al. 1993


XcmH


XcmHI005
XcmHI407

XcmHl427

XcmHl431

XcmN

Xcml003


Plasmid


pRK2013
pRK2073

pUFRO42
pUFRO47








Table 2-1--continued.


Strain or Plasmid Relevant Characteristics Reference or Source


pUFRO49

pUFRO34

pUFR1 15
pUFR127
pUFR135
pUFR142
pUFRI56
pUFR1S?
pUFRl63
pUFRI63ABam
pUFR171
pUFRl72
pUFR173
pUFR174
pUFRI75
pUFRI76
pUFR17S

pUFR179

pUFRiSO
pUFR186

pUFRI86A~ain
pUFRI9O-194


pUFRI96-200


pUFR20S-209



pUFR211-215



pUFR2I7
pUFR220
pUFR227
PUFY019


pUFY020


pZit45
pGiEM7Zf(+)
pGEMI lZf(+)


RSF1010 replicon, Cmf, Smo, lacW+,
Mob' displacement vector
IncP, Tc', Mob', containing methylases
Xmal and Xmaml
7.5kb fragment containing avrB4 in pUFRO42
5kb fragment containing awrb6 in pUFRO42
lacZ::avrb6 fusion in pUFRO42
9kb fragment containing avrBOl in pUFRO47
12.9 kb fragment containing avrBIn in pUFRO42
11kb fragment containing avrB102 in pUFRO42
10kb fragment containing aw*bin pUFRO42
pUFRI63 deleted for 3.4 kb BamHI fragment
Internal BamHI fragment of avrB4, in pGem lZf(+)
Internal BamHI fragment of avrb6, in pGeml IZf(+)
Internal BamHI fragment of avrBlOI, in pGeml IZf(+)
Internal BamHI fragment of avrBIn, in pGeml lZf(+)
Internal BamHI fragment of avrBI02, in pGeml IZf(+)
Internal BamHI fragment of avrb7, in pGemllZf(+)
Internal BamHI fragment of avrBIn,
in pGemI lZf(+), Sal site filled
Internal BamHI fragment of avrb7, in pGemllZf(+),
Sall site filled
9.5 kb fragment containing avrb6 in pUFRO42
10.3 kb Bgll-EcoRI fragment containing avrBln
from pUFR156 in pUFRO47
pUFRI86 deleted for 3.7 kb BamHI fragment
BamHI-swapped chimeric genes with avrb7
5' and 3' ends plus BamHI fragment of avrB4,
avrb6, avrBIn, avrBlOl, or avrBI02 in pUFRO47
BaimHI-swapped chimeric genes with avrBIn
5' and 3' ends plus &amHI fragment of avrB4,
avrb6, avrb7, avrBlOl, or avrBI02 in pUFRO47
Saul/Hincl-swapped chimeric genes with avrBIn
5' and 3' regions plus Sal/Hincll fragment of
avrBlOl, avrB4, awrb6, awrBJ02, or avrb7
in pUFRO47
SAuI//incU-swapped chimeric genes with avrb7
5' and 3' regions plus Saul/Hinc fragment of
avrBlOI, avrB4, awvr6, avrBln, or avrBI02
in pUFRO47
pUFR156 derivative, avrMBn::Ta5-gu4sA
pUFR163 derivative, avrb7::Tn5-gusA
pUFRISO derivative, avrb6::Tn5-gusA
3.7 kb Saul/Hinc-swapped fragment with phAA
5' and 3' regions of pZit45 plus avrb6 internal
repeat region of pUFR135 in pUFRO47
3.8 kb Saul/Hincl-swapped fragment with avrb6
5' and 3' regions of pUFRI35 plus pthA internal
repeat region of pZit45 in pUFRO47
4.5kb fragment containing pthA in pUFRO47
ColEl, Ap', lacZa*
ColEl, Ap, lacZa


Swamp et at. 1991

DeFeyter & Gabriel, 199 1b

De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter et at. 1993
De Feyter et al. 1993
De Feyter et al. 1993
De Feyter et al. 1993
De Feyter el al. 1993
This study
De Feyter e: al. 1993
De Feyter et al. 1993
De Feyter et al. 1993
De Feyter et at. 1993
De Feyter et al. 1993
De Feyter el al. 1993
This study

This study

De Feyter et al. 1993
This study

This study
This study


This study


This study



This study



This study
This study
This study
This study


This study


Swamp et al. 1992
Promega Co., Madison, WI
Promega Co., Madison, WI






15
pUFRO47 vector. The resulting BamHI fragment-swapped chimeric genes formed

pUFR190-200. The avr genes from pXcmH and pthA from X. citri all have unique StuI

and HincI sites delimiting the 102bp repeated regions, which allow the swapping of the

internal repeated regions among these genes. To construct StuI/HinclI fragment-swapped

chimeras using the pXcmH avr genes, pUFR174 and pUFR176 were cut with Sall, blunt

ended using the Klenow fragment and religated to destroy the HindI sites on the

pGEM1 lZf(+) portions of the plasmids, forming pUFR178 and pUFR179, respectively.

These were then cut with StuI and HincH to delete the internal fragments, and used as

recipients for the Stul/HincdI internal fragments from all pXcmH avr genes on pUFR171-

176. The BamHI fragments from pUFR178 and its StuI/HinclI chimeras were recloned

into pUFR186ABam, forming pUFR205-209. The BamHI fragments from pUFR179 and

its Stul/Hincll chimeras were recloned into pUFR163ABam, forming pUFR211-215. To

facilitate construction of chimeric genes between pthA and avrb6, the 4.1 kb Sall

fragment containing pthA from pZit45 and the 3.0 kb EcoRIISall fragment containing

avrb6 from pUFR135 were inserted into pGEM7Zf(+). The Stul/HinclI internal

fragments were swapped between the two pGEM derivatives to create chimeric genes.

The chimeric genes were then recloned as single EcoRIIHindSl fragments into pUFRO47,

forming pUFY019 and pUFY020.

Bacterial conjugation. Triparental matings were carried out to transfer broad host

range plasmids from E.coli DH5a to various Xanthomonas strain by using pRK2013 or

pRK2073 as helper plasmids as described (De Feyter and Gabriel 1991a). To transfer






16
plasmids into Xcm, the modifier plasmid pUFR054 carrying XcmI and XcmlII methylase

genes was used to increase the transfer frequency (De Feyter and Gabriel 1991b).

Marker-exchange mutagenesis. Marker-exchange mutagenesis of strain

XcmH1005 was accomplished by introducing the displacement vector pUFRO49 into

XcmH1005 transconjugants harboring pUFR227 (avrb6::Tn5-gusA), pUFR220

(avrb7: :Tn-gusA) or pUFR217 (avrBIn::Tn5-gusA) derivatives. The procedure was the

same as described (Swarup et al. 1991), except that the modifier plasmid pUFR054 was

used to facilitate plasmid transfer into XcmH1005.

Plant inoculations. Cotton (Gossypiwn hirsutum L.) lines used were Acala-44

(Ac44) and its congenic resistance lines Bl, B2, B4, b6, b7, Bin and BIn3 as described

(Swarup et al. 1992, De Feyter et al. 1993). Cotton plants were grown in the

greenhouse, transferred to growth chambers before inoculation, and maintained under

conditions as described (De Feyter and Gabriel 1991a). Bacterial suspensions of X.

campestris pv. malvacearum (108 cfu/ml) in sterile tap water were gently pressure

infiltrated into leaves of 4-5 week old cotton plants. Pathogenic symptoms were

observed periodically 2-7 days after inoculation.

All citrus (Citrus paradise 'Duncan', grapefruit) and bean (Phaseolus vulgaris

'California Light Red') plants were grown under greenhouse conditions. Plant

inoculations involving X. citri or pthA or derivatives of pthA were carried out in BL-3P

level containment (refer to Federal Register Vol.52, No.154, 1987) at the Division of

Plant Industry, Florida Department of Agriculture, Gainesville. Bacterial suspensions

were standardized in sterile tap water to 108 cfu/ml and pressure infiltrated into the






17
abaxial leaf surface of the plants. All plant inoculations on cotton, citrus or bean were

repeated at least three times.

Bacterial growth in plant. To determine the growth of X. campestris pv.

malvacearum in the susceptible cotton line Ac44, bacterial suspensions were adjusted to

106 cfu/ml and inoculated into cotton leaves by pressure infiltration. Leaf discs (1 cm2)

were taken by using a sterilized cork borer, then macerated in 1 ml sterile tap water. At

least three samples were taken at each time point for each strain inoculated. Viable

counts were determined by serial dilutions on plates containing appropriate antibiotics.

Data shown in Fig. 2-2 are the mean and standard error from three experiments.

To quantify the amount of bacteria present on the surface of watersoaked leaf

spots, leaves were inoculated with 107 cfu/ml bacterial suspensions from overnight

cultures. Five days later, 100 p~ sterile tap water was dispensed onto the watersoaked

leaf surface and spread to an area of approximately 1 cm2. After mixing with the

bacteria and slime on the leaf surface, the bacterial suspension was collected with a pipet.

Each 1 cm2 watersoaked leaf area was washed ten times, and a total of 1 ml of bacterial

suspension was collected. To quantify the bacteria remaining inside the leaf, leaf discs

(1 cm2) were taken with a sterilized cork borer, then macerated in 1 ml sterile tap water.

Viable counts were determined by serial dilutions on plates containing appropriate

antibiotics. Data reported in the results (external and total counts) were mean and

standard error of six replicates from two experiments.








Results



Pleiotropic pathogenicity functions of avrb6. X. campestris pv. malvacearum

strain XcmH carries avrB4, avrb6, avrb7, avrBIn, avrBlO1 and avrB102 on a single

plasmid, pXcmH, and elicits an HR on cotton lines carrying any one of many different

resistance (R) genes (De Feyter et al., 1993). Strain XcmH and a spontaneous

rifamycin-resistant derivative of XcmH, XcmH1005, were virulent on susceptible cotton

line Acala-44 (Ac44), and both elicited severe watersoaking and necrosis associated with

growth in plant (XcmHl005 on Ac44 is shown in Figs. 2-1 & 2-2). Mutations of

avrb6, avrb7 and avrBIn were individually generated in XcmH1005 by marker-exchange

mutagenesis and each mutant was confirmed to carry a single TnS-gusA insertion in the

appropriate DNA fragment by Southern blot hybridization (refer Fig. 2-3; some data not

shown). As predicted by gene-for-gene theory, marker-exchange mutants XcmH1407

(avrb6::Tn5-gusA), XcmH1427 (avrb7: :Tn5-gusA) and XcmH1431 (avrBIn: :Tn-gusA)

gained virulence on cotton lines with the resistance genes b6, b7 and BIn, respectively

(Table 2-2). Plasmids pUFR127 (avrb6+), pUFR163 (avrb7+) and pUFR156 (avrBIn+)

were able to fully complement the specific avirulence defects of XcmH 1407, XcmH1427

and XcmH1431, respectively.

Neither XcmH1427 nor XcmH1431 showed any change in watersoaking ability

as compared with XcmH1005. However, marker-exchange mutant XcmH1407 not only

lost avirulence on a cotton line with the b6 resistance gene, but also elicited significantly

less watersoaking and necrosis on susceptible cotton line Ac44 (Table 2-2 and Fig. 2-1A
























Figure 2-1. Watersoaked lesions caused by X. campestris pv. malvacearum strains on
the susceptible cotton line Acala-44. A. Leaf inoculated five days previously with: 1,
XcmH1005; 2, XcmH1407; 3, XcmH1407/pUFR127; 4, XcmH1431. B. Leaf
inoculated seven days previously with: 1, XcmH1005; 2, XcmH1407. C. Leaf
inoculated five days previously with: 1, XcmH1005; 2, Xcml003/pUFR127; 3,
Xcml003/pUFR156; 4, Xcml003.






20




























a-4


0





































n-4











C4.
tj '0)









































































0n o r-


C


o o
N0

0
C0


c









to
(I)

10 0

Cl)

0


(N)


So D
^- uC
0
C



00








(0 LIo 4 rfl


wzu-bs/n;o 5o4


on o3 r Q0 LO .:










1 2 3 4


O
12kb-
0@



iee





3kb- 4



Figure 2-3. Southern blot hybridization of total DNA from XcmH1005 and marker-
exchanged mutants of XcmH1005 after EcoRIISstI digestion. The blot was probed with
the internal 2.9kb BamHI fragment from avrb6. The expected positions of fragments
carrying each plasmid-borne member of the avr/pth gene family were determined from
the restriction map of pXcmH [De Feyter and Gabriel, 1991]. The upper arrow indicates
the expected position of the DNA fragment carrying avrBIn; the lower arrow indicates
the position of avrb6. Lane 1, 1 kb DNA ladder; Lane 2, XcmH1005; Lane 3,
XcmH1407; Lane 4, XcmH1431.









Table 2-2. Phenotypes of marker-exchange mutants of Xanthomonas campestris pv.
malvacearum strain XcmH1005 on cotton cultivar Acala-44 (Ac44) and congenic lines.


Xcm strains Ac44 Acb6 Acb7 AcBIn


XcmH1005 ++ -
XcmH1407 (avrb6::Tn5-gusA) + + -
XcmH1427 (avrb7::Tn5-gusA) ++ ++ -
XcmH1431 (avrBIn::TnS-gusA) ++ ++


aAcb6, Acb7, and AcBIn are congenic lines of Ac44 containing resistance gene b6, b7,
and BIn, respectively. + + = strong watersoaking symptoms; + = weak watersoaking
symptoms; = hypersensitive response.






25
& 2-1B). Plasmid pUFR127 was able to fully complement the pathogenicity defects)

of XcmH1407 on susceptible cotton plants, in addition to complementing the specific

avirulence defect. Despite its reduced ability to elicit pathogenic symptoms on

susceptible cotton lines, mutant XcmH1407 exhibited the same growth rate and yield as

that of the wild type XcmH1005 and XcmH1431 on Ac44 (Fig. 2-2A).

X. campestris pv. malvacearum strain Xcm1003 carries no known avr genes (De

Feyter and Gabriel 1991a), caused less watersoaking on Ac44 than XcmH1005 and

exhibited a lower growth rate on Ac44 than XcmH1005. Introduction of pUFR127

(avrb6+) into Xcml003 conferred increased watersoaking ability (Fig. 2-1C), but growth

rate and yield inplanta of the transconjugant containing pUFR127 were not increased in

comparison with that of Xcml003 or the transconjugant Xcml003/pUFR156 (avrBln+)

(Fig. 2-2B). Therefore pUFR127 affected pathogenic symptoms in both XcmH1005 and

Xcml003, but not bacterial growth rate or yield in plant of either strain.

Strains containing avrb6 (e.g. XcmH1005 and Xcm1003/pUFR127) elicited more

severe watersoaking and necrosis and were associated with much more slime oozing from

the watersoaked areas in comparison with strains lacking avrb6 (e.g. XcmH1407 and

Xcm1003; Fig. 2-1). The peak number of total colony forming units (cfu) per square

centimeter of watersoaked leaf inoculated with XcmH1005 or XcmH1407 were basically

the same (1.41 0.09 x 109 cfu/cm2 vs. 1.48 0.17 x 109 cfu/cm2). However,

14.1% of the XcmH1005 bacteria present in the lesion (2.32 0.56 x 108 cfu/cm2)

were released onto the surface of the leaf, whereas only 0.06% of the XcmH1407

bacteria (9.64 4.42 x 105 cfu/cm2) were released onto the surface of the leaf.






26
Therefore, more than 240 times more bacteria were present on the external surface of

the watersoaked lesions caused by XcmH1005 than by XcmH1407. In experiments

where the leaves were moistened periodically after inoculation with a hand-held mist

sprayer, X. campestris pv. malvacearum strains carrying avrb6 always exhibited many

secondary infections around the original inoculation site. In contrast, those strains

lacking avrb6 rarely exhibited secondary infections.

The pathogenicity functions encoded by pthA and avrb6 are host specific and

determined by 102bp repeats. Single StuI and Hincdl sites are found in avrb6 at positions

1281 and 2860, respectively, that closely flank the 102 bp tandemly repeated region of

the gene (De Feyter et al. 1993). Unique StuI and HincH sites are also found at the

same relative positions in pthA, and the DNA sequences of avrb6 and pthA are identical

in the flanking regions from the 102 bp tandem repeats to these restriction sites (refer to

Chapter 4). Chimeric genes were constructed by swapping the Stul/HincII fragments

containing the internal 102bp repeated regions between pthA and avrb6. At least three

individual clones of both chimeric genes were introduced into a mutant X. citri strain

B21.2 (pthA::Tn5-gusA) and wild type strains of the following pathogens: X. phaseoli

strain G27 (host range on bean), X. campestris pv. citrumelo strain 3048 (host range on

bean and citrus), X. campestris pv. alfalfae strain KX-1 (host range on alfalfa, bean and

citrus) and Xcml003. Strains and transconjugants were inoculated on citrus, bean, and

cotton leaves. As shown in Figure 2-4, pUFY020, carrying a chimeric gene containing

5' and 3' ends of avrb6 and the internal repeats of pthA, complemented the mutant B21.2

to full virulence on citrus. The citrus canker symptoms caused by B21.2/pUFY020 were





































Figure 2-4. Phenotypes of X. citri wild type and mutant strains on a grapefruit (Citrus
paradisi cv.'Duncan') leaf. 1, wild-type strain 3213; 2, marker-exchange mutant of
3213, B21.2; 3, B21.2/pZit45; 4, B21.2/pUFY020; 5, B21.2/pUFR135; 6,
B21.2/pUFY019.






28
indistinguishable from the symptoms caused by B21.2/pZit45 (pthA +). When pUFY020

was present in 3048 or KX-1, both of which cause watersoaked leaf spots on citrus,

canker symptoms were induced by the transconjugants that were indistinguishable from

those conferred by pZit45 in the same strains. Also like pZit45 (Swarup et al, 1992),

pUFY020 conferred avirulence on bean to 3048, KX-1 and G27, and on cotton to

Xcml003.

A chimeric gene containing the 5' and 3' ends of pthA and the internal repeats of

avrb6 on pUFY019 enhanced the ability of Xcml003 and XcmH1407 to watersoak cotton

and cause necrosis. The watersoaking symptoms caused by Xcml003/pUFY019 were

indistinguishable from those caused by Xcml003/pUFR127. This chimeric gene on

pUFY019 behaved like avrb6 on pUFR127 and conferred no detectable symptoms to

B21.2, 3048 or G27 on citrus or bean.

Cultivar-specific avirulence is determined by 102bp repeats. Each of the seven

avr/pth genes (avrB4, avrb6, avrb7, avrBIn, avrBl01, avrB102, and pthA) exhibits

unique avirulence specificity in Xcml003 on cotton resistance lines differing by single

R genes (De Feyter et al. 1993, Swarup et al. 1992). Each of these seven genes

contains two BamHI sites, one near the 5' end and one near the 3' end of each gene, and

single StuI and Hincd sites flanking the 102 bp direct repeat region (De Feyter et al.

1993, Swarup et al. 1992). To localize the region that determines the specificity of the

reactions, a series of chimeric genes were constructed by swapping BamHI and

StuI/HinclI internal fragments among these seven members of the gene family. One to

three individual clones of each chimeric gene were introduced into Xcml003 and








Table 2-3. Avirulence specificity of BamHI-fragment swapped chimeric genes in
Xcm1003 on cotton.



Chimeric Gene
Plasmid Avirulence Specificity
5' and 3' end BamHI fragment


pUFR190 avrb7 avrB4 avrB4
pUFR191 avrb7 avrb6 avrb6
pUFR192 avrb7 avrBIn avrBIn
pUFR193 avrb7 avrB1O1 avrBlOl
pUFR194 avrb7 avrB102 avrB102
pUFR196 avrBIn avrB4 avrB4
pUFR197 avrBIn avrb6 avrb6
pUFR198 avrBIn avrb7 avrb7
pUFR199 avrBIn avrBlO1 avrBlOl
pUFR200 avrBIn avrB102 avrB102

Note: Chimeric genes were introduced into Xcml003 and tested on cotton cv. Ac44 and
its congenic resistance lines AcB1, AcB2, AcB4, Acb6, Acb7, AcBIn and AcBIn3 for
avirulence specificity.









Table 2-4. Avirulence specificity of Stul/HincdI-fragment swapped chimeric genes in
Xcml003 on cotton.



Chimeric Gene
Plasmid Avirulence Specificity
5' and 3' region Stul/HincII fragment


pUFR205 avrBIn avrB4 avrB4*
pUFR206 avrBIn avrb6 avrb6*
pUFR207 avrBIn avrb7 avrb7*
pUFR208 avrBIn avrBlOl1 avrBlOl
pUFR209 avrBIn avrB102 avrB102
pUFR211 avrb7 avrB4 avrB4
pUFR212 avrb7 avrb6 avrb6
pUFR213 avrb7 avrBIn avrBIn*
pUFR214 avrb7 avrBlO1 avrBlOl
pUFR215 avrb7 avrB102 ND
pUFY019 pthA avrb6 avrb6
pUFY020 lacZ::avrb6 pthA pthA +

Note: Chimeric genes were introduced into Xcml003 and tested on cotton cv. Ac44 and
its congenic resistance lines AcB1, AcB2, AcB4, Acb6, Acb7, AcBIn and AcBIn3 for
avirulence specificity. = weak avirulence; + = strong avirulence; ND = no
avirulence detected.






31
inoculated on cotton cultivar Ac44 and its congenic resistance lines, each differing by a

single R gene (Bl, B2, B4, b6, b7, Bin or BIn3). The avirulence specificities of the

seven avr/pth genes were first localized within the internal BamHI fragments (Table 2-3),

and then were further localized within the internal Stul/HinclI fragments (Table 2-4).

In all cases, the avirulence specificity of a given gene was determined inside the

StuV/HinclI (tandem repeat) region.



Discussion



Plant pathologists have long been puzzled by the presence of avirulence genes in

pathogens. These genes act as negative factors to limit virulence and in most cases do

not appear to provide selective advantage to the pathogens (Ellingboe, 1976; Gabriel,

1989; Keen and Staskawicz 1988). Rare exceptions have been reported. For example,

avrBs2 from X. campestris pv. vesicatoria is required for optimal growth in plant

(Kearney and Staskawicz 1990). Both avrb6 and avrb7 from X. campestris pv.

malvacearum strain XcmH are known to enhance the watersoaking ability of X.

campestris pv. malvacearum strain Xcml003 on cotton (De Feyter and Gabriel 1991a).

In this study we demonstrated that the ability of strain XcmH1005 to cause strong

watersoaking and necrosis on cotton requires the presence of avrb6 but not avrb7.

Although avrb6 increased symptom elicitation by both Xcm1003 and XcmH1005, neither

their growth rates nor their maximum bacterial counts per square centimeter of leaf were

affected by the presence or absence of avrb6. Therefore avrb6 functions as a






32
pathogenicity gene and increases symptoms of cotton blight, but without eliciting an HR

on cotton lines lacking b6.

The strong watersoaking ability conferred by avrb6 was correlated with much

higher (240-fold) levels of bacterial cells released from inside the plant leaf to the

surface. Since bacterial blight of cotton is usually spread by rain splash, the presence

of large numbers of bacteria on the leaf surface would undoubtedly contribute to the

dissemination of the population. Strains carrying avrb6 would thereby have a selective

advantage on cotton plants lacking the b6 gene. Similarly, pthA appears to aid in the

dissemination of X. citri by rupturing leaf epidermis and releasing bacteria, although it

does so by inducing tissue hyperplasia (Swarup et al. 1991). Therefore, pthA and avrb6

not only contribute to the amount of damage caused by their respective xanthomonads

to their hosts, but may also contribute to the ecological fitness of their respective

bacterial populations as pathogenicity genes.

Both pthA and avrb6 may help determine host range in a positive manner, and not

as avr genes. When pthA was transferred to X. campestris pv. malvacearum Xcml003

and X. phaseoli G27, it conferred avirulence on the respective hosts, and did not induce

tissue hyperplasia (Swarup et al, 1992). In the present study, when avrb6 was

transferred to X. citri B21.2, X. campestris pv. citrumelo 3048 and X. phaseoli G27, it

conferred no detectable effect when these strains were inoculated onto their respective

hosts. Therefore, the pathogenicity functions of avrb6 and pthA are host specific. If

release of the pathogen to the leaf surface is host specific and it contributes to the

ecological fitness of the pathogen as we propose, then avrb6 and pthA function to






33
determine host range. Furthermore, the avirulence conferred by pthA appeared to be

gratuitous in terms of restricting host range (Swarup et al., 1992) and in the present

study, avrb6 failed to confer avirulence to three other pathogens. Therefore if avrb6 and

pthA help determine host range, it is not because of their function as avr genes. In a

formal genetic sense, these pleiotropic avr/pth genes resemble some Rhizobium host-

specific nodulation (hsn) genes which are required for host range on some hosts, but

which can also confer avirulence when transferred to other Rhizobium strains with a

different host range (Debelle et al. 1988; Faucher et al. 1989; Lewis-Henderson and

Djordjevic 1991).

By swapping the tandemly repeated regions, the avirulence specificities of avrB4,

avrb6, avrb7, avrBIn, avrBlOl, avrB102 and pthA were shown to be determined by the

102bp tandem repeats. These results are consistent with and extend the findings of

Herbers et al. (1992), who showed that the avirulence specificity of avrBs3 is determined

by the 102bp repeats of that gene. In addition, the swapping experiments clearly

demonstrated that the pathogenicity functions of pthA and avrb6 were distinct (cankers

vs. watersoaking), host specific, and in both cases the pathogenic specificity was

determined by their 102bp tandemly repeated regions. Furthermore, the use of chimeric

genes also ruled out the possibility that the host-specific pathogenicity on cotton and

citrus were result of additional, unidentified pathogenicity genes encoded on the plasmids

used (pUFR127 and pZit45).

Extensive deletion analyses of both the 5' and 3' ends of the pXcmH avr genes

showed that all sections of these genes are required for avirulence activity (De Feyter et






34
al. 1993). From the present study, all sections of pthA and avrb6 were required to

confer pathogenicity functions. However, the 5' and 3' ends of the genes outside of the

repeats appeared to be isofunctional among members of the gene family.

Besides the members of this Xanthomonas avr/pth gene family, internal tandem

repeats are found in some pathogenicity genes of animal pathogens. Examples include

the outer membrane protein A gene (ompA) of some Rickettsia species (Anderson et al.

1990, Gilmore 1993), internalin gene (in)) of Listeria monocytogenes (Gaillard et al.

1991), toxin A gene of Clostridium difficile (Dove et al. 1990), and M protein genes

(emm) of Streptococcus (Hollingshead et al. 1987). Tandem repeats were also found in

many genes from protozoan and metazoan parasites such as Plasmodium (McConkey et

al. 1990), Trypanosoma (Hoft et al. 1989), Leishmania (Wallis and McMaster 1987), and

Meloidogyne (Okimoto et al. 1991). Genes from Trypanosoma cruzi (the protozoan

agent of American trypanosomiasis) and mitochondria of Meloidogyne javanica (plant

root knot nematode) contain 102 bp tandem repeats. Most of these genes encode surface

proteins and the distinctive arrangement of the tandem repeats in these genes are thought

to encode a protective, strain-specific conformational epitope for evasion of host

immunity (Gilmore 1993, Hoft et al. 1989, McConkey et al. 1990). By contrast,

AvrBs3 is mainly located in cytosol (Knoop et al. 1991, Brown et al. 1993) and it is not

clear how it may interact with the plant cell.

Southern hybridization has shown that potential members of the avr/pth gene

family exist, often in multiple copies, in 9 of 12 Xanthomonas species or pathovars

examined (Bonas et al. 1989, Swarup et al. 1992, De Feyter et al. 1993). In all X. citri







35
and X. campestris pv. malvacearum cotton strains tested to date, multiple DNA fragments

hybridizing to pthA and avrb6 are found. The pleiotropic pathogenicity functions of pthA

and avrb6 and their potential fitness value may explain why these genes are maintained

in plant pathogens. Based on these data and knowledge of the functions of avrb6 and

pthA, we assume that all strains of X. campestris pv. malvacearum capable of strongly

watersoaking cotton carry an avrb6 gene or homologue that functions for pathogenicity.

Similarly, we assume that all strains of X. citri capable of causing cankers on citrus carry

a pthA gene. A number of pathogenic strains and pathovars in the genus Xanthomonas

do not carry members of the gene family, and therefore these genes are not required for

Xanthomonas virulence generally. Within pathovars where members of the avr/pth gene

family are found in some, but not all strains tested (such as X. campestris pv.

vesicatoria), there may be no pleiotropic pathogenicity function; for example, there is no

evidence of a pathogenicity function of avrBs3, avrBs3-2 and avrBsP (Bonas et al. 1989,

1993; Canteros et al. 1991).

The mechanisms) by which leaf spotting pathogens elicit watersoaking and

necrosis is unknown, but appears to involve both damage to leaf cell membranes (without

eliciting an HR) and production of extracellular polysaccharide (EPS). The EPS does

not appear to be involved in suppressing a potential HR, since production levels of EPS

by X. campestris pv. malvacearum are similar in both susceptible and resistant cotton

lines (Pierce et al. 1993). Instead, mutational analyses of Xanthomonas EPS biosynthetic

genes and inoculations with purified EPS have shown that the EPS contributes to

watersoaking by trapping water and nutrients in the intercellular spaces after they are






36
released (reviewed by Leigh and Coplin 1992). Coplin et al. (1992) have proposed that

watersoaking and pathogenicity of Erwinia stewartii involves the EPS plus a cell leakage

factor encoded by wts (watersoaking) genes. This hypothesis may well apply to X.

campestris pv. malvacearum, with avrb6 encoding a cell leakage factor. Like some wts

genes of E. stewartii, avrb6 was not required for bacterial growth in plant, but strongly

affected watersoaking on its host.

The ion-channel defense model of the gene-for-gene hypothesis invokes a cell

leakage factor as the product of an avr gene (Gabriel et al. 1988, Gabriel and Rolfe

1990). In this model, avr genes produce a protein or compound that opens an ion

channel in the plant cell membrane that rapidly depolarizes the membrane, causing

electrolyte leakage and host cell death. This gene-for-gene model is not inconsistent with

the idea that the product of an avr gene might induce slower cell leakage on susceptible

hosts. The only difference might be the allelic form of the host R gene. If the leakage

were slow enough to avoid cascade amplification of a wound-response signal, changes

in the osmotic gradient could cause a net loss of water from the cell and redistribution

to the apoplast, thereby increasing the fluidity of the EPS (M. Essenberg and M. Pierce,

personal communication). Increased fluidity or amounts of EPS may increase bacteria

exuding onto the leaf surface through stomata (Thiers and Blank 1951). Another possible

function is that the increased levels of necrosis induced by avrb6 may serve to collapse

the palisade layer and physically squeeze more bacteria onto the leaf surface.











CHAPTER 3
INACTIVATION OF MULTIPLE AVIRULENCE GENES IN XANTHOMONAS
CAMPESTRIS PV. MALVACEARUM CREATED A NONPATHOGENIC
ENDOPHYTE OF COTTON



Introduction



Avirulence (avr) genes act as negative factors to confer avirulence and to limit

host range of the pathogen on host plants carrying specific resistance (R) gene. As a

result, most avr genes reported to date appear to be gratuitous and of no obvious

selective value to the microbial plant pathogens (Gabriel et al. 1993). However, a few

of them have demonstrated pleiotropic functionss. For example, avrBs2 is required for

full virulence of Xanthomonas campestris pv. vesicatoria on susceptible hosts (Kearney

and Staskawicz 1990). More recently, pthA of X. citri and avrb6 of X. campestris pv.

malvacearum were also shown to have host-specific pathogenicity function in addition to

avirulence function (Swarup et al. 1991, 1992; Chapter 2). Gene pthA is essential for

X. citri strain 3213 to induce hyperplastic cankers on citrus specifically and confers that

function to several other X. campestris pathovars. Similarly, avrb6 is important for X.

campestris pv. malvacearum strain XcmH to induce watersoaking on cotton specifically,

and confers that function to other X. campestris pv. malvacearum strains. Both pthA and

avrb6 appear to contribute to the fitness of their respective pathogens on their hosts by






38
increasing release of the pathogens from plant leaves, thus affecting host range through

increased dispersal.

Both pthA and avrb6 are members of a large Xanthomonas avirulence (avr) and

pathogenicity (pth) gene family, and both confer gene-for-gene avirulence to X.

campestris pv. malvacearum on cotton (Swarup et al. 1993; De Feyter et al. 1993).

Other members of the avr/pth gene family include avrB4, avrb7, avrBIn, avrBlOl and

avrB102 of X. campestris pv. malvacearum (De Feyter and Gabriel 1991a, De Feyter et

al. 1993), avrBs3, avrBs3-2 and avrBsP of X. campestris pv. vesicatoria (Bonas et al.

1989, 1993; Canteros et al. 1991) and avrXa5, avrXa7 and avrXalO of X. oryzae

(Hopkins et al. 1992). Family members sequenced to date are 95%-98% identical to

each other (Bonas et al. 1993; Hopkins et al. 1992; De Feyter et al. 1993). Most

members of this gene family were isolated as avirulence genes with no evidence of

pathogenicity function.

Xanthomonas campestris pv. malvacearum (Smith) Dye is the causal agent of

bacterial blight of cotton, an economically important, world-wide disease. Six members

of this gene family were previously isolated from pXcmH, a 90-kb plasmid of X.

campestris pv. malvacearum strain XcmH. Southern hybridization analysis reveals at

least five additional hybridizing fragments in XcmH besides the six plasmid-borne avr

genes (De Feyter et al. 1993). Furthermore, all strains examined of Xanthomonas

campestris pv. malvacearum that attack cotton carry at least 4-11 DNA fragments that

hybridize with members of the avrlpth gene family, including many strains which carry

no known avr genes (De Feyter et al. 1993). The purpose of this research was to clone






39
and to mutagenize all avr genes in X. campestris pv. malvacearum strain XcmH and to

identify potential pleiotropic functions.



Materials and Methods



Bacterial strains, plasmids, and matings. The bacterial strains and plasmids used

in this study are listed Table 3-1. Strains of Escherichia coli were grown in Luria-

Bertani (LB) medium (Sambrook et al. 1989) at 370C. Strains of Xanthomonas were

grown in PYGM (peptone-yeast extract-glycerol-MOPS) medium at 30*C (De Feyter et

al. 1990). For culture on solid media, agar was added at 15g/L. Antibiotics were used

as previously described (refer to Chapter 2). To transfer plasmids from E.coli strains

HB101 or DH5a to X. campestris pv. malvacearum, helper plasmid pRK2073 and

modifier plasmid pUFR054 were used as described (Feyter and Gabriel 1991a,b).

Recombinant DNA techniques. Total DNA isolation from Xanthomonas was as

described (Gabriel and De Feyter 1992). Plasmids were isolated from E. coli by alkaline

lysis methods (Sambrook et al. 1989). Restriction enzyme digestion, alkaline

phosphatase treatment, DNA ligation and random priming reactions were performed as

recommended by the manufacturers. Southern hybridization was performed by using

nylon membranes as described (Lazo et al. 1987). Otherwise, standard recombinant

DNA procedures were used (Sambrook et al. 1989).

Gene replacement and marker eviction. To carry out marker-exchange

mutagenesis, a 3.8 kb BamHI fragment containing a nptl-sac cartridge from pUM24








Table 3-1. Bacterial strains and plasmids used in this study


Strain or Plasmid Relevant Characteristics Reference or Source


F-, endA1, hsdRl 7 (r-mk+), supE44,
thi-1, recAl, gyrA, relAl, #8OdlacZ
AM15, A(lacZYA-argF)U169
supE44, hsdS20 (r-m ), recAl3,
ara-14, proA2, lacYl, galK2, rpsL20,
xyl-5, mdl-I
supE44, supF58, hsdS3 (r-mk+), recA56,
galK2, galT22, metBI


Gibco-BRL,
Gaithersburg, MD

Boyer &
Roulland-Dussoix 1969


Murray et al. 1977


X campestris pv. malvacearum


Natural isolate from cotton from
Oklahoma; carries six avr gene on pXcmH,
and additional six potential avr genes.
Spontaneous Rif derivative of XcmH
avrBIn::nptl-sac, marker exchange
mutant of XcmHl005
avrb6::nptl-sac, marker exchange
mutant of XcmHl005
avrB5::nptl-sac, awvrb6), marker
exchange mutant of XcmH1005
avrB5-, avrb6- marker-evicted
mutant derived from HM 1.20
Hc l: :nptl-sac, marker-exchange
mutant of XcmH1005
Hcl-, avrBIn', avrBl04- marker
exchange mutant of XcmH1005
avrBn: :nptl-sac, marker exchange
mutant of XcmH1005
avrBlO3::nptl-sac, marker exchange
mutant of XcmH1005
avrB4-, avrbT, avrBIn-, avrBlOl',
avrl02- marker exchange mutant of XcmHl005
avrHc6~, avrB4", avrb6-, avrbT,
avrBIn-, avrBlOl', avrl02"), marker exchange
mutant of HM1.20S.
XcmH1005 (avrHc6-, avrB4-, avrb6", avrbT,
avrBIn-, avrBlOl', avrl02"), marker-evicted
mutant derived from HM2.2S.
Natural isolate from cotton from
Upper Volta, Africa.
Spc' Rif derivative of XcmN


De Feyter & Gabriel, 1991a


Chapter 2
This study

This study

This study

This study

This study

This study

This study

This study

This study

This study


This study


Gabriel et al. 1986

De Feyter & Gabriel 1991a


E. coli

DH5-,


HB101


ED8767


XcmH


XcmHl005
HM1.10

HM1.15

HM1.20

HM1.20S

HM1.26

HM1.32

HM1.34

HMI.36

HMI.38

HM2.2


HM2.2S


XcmN

Xcml003








Table 3-1--continued.

Strain or Plasmid Relevant Characteristics Reference or Source


Plasmid


pRK2073

pUFR004
pUFR034
pUFR042
pUFRO47
pUFRO54

pUFR103
pUFR106
pUFR107
pUFR109
pUFRI11
pUFRI12
pUFR113
pUFRI115
pUFR127
pUFRI42
pUFR156
pUFRI57
pUFR163
pUFYI.48

pUFY10.1

pUFY31.46
pUFY33.19
pUFY36.8
pUFY36.26
pUFY37.62

pUFY38.1
pUM24
pXcml.12
pXcml.21
pXcm2.23
pXcm2.12
pXcm2.11
pXcml.22
pZit45


pRK2013 derivative, npt::Tn7,Knm, Sp',
Tra+, helper plasmid
ColEl, Mob*, Cm'
IncW, Km', Mob*, lacZa*, Par*, cos
IncW, Knm, Gm', Mob*, lacZa*, Par+
IncW, Gmn, Ap', Mob+, lacZa+, Par+
IncP, Tc', Mob*, containing methylases
Xmal and Xmanl
Cosmid clone carrying avrB4, avrb6.
Cosmid clone carrying avrB4, avrb6, avrBlO1
Cosmid clone carrying avrb6, avrBlO1, avrBIn
Cosmid clone carrying avrBIOl, avrBIn
Cosmid clone carrying avrBlOl, avrBIn, avrBl02
Cosmid clone carrying avrBIn, avrB102, avrb7
Cosmid clone carrying avrBI02, avrb7
7.5-kb fragment containing avrB4 in pUFRO42
5-kb fragment containing avrb6 in pUFRO42
9-kb fragment containing avrBlOl in pUFRO47
12.9-kb fragment containing avrBIn in pUFRO42
11-kb fragment containing avrB102 in pUFRO42
10-kb fragment containing avrb7 in pUFRO42
8.4-kb fragment carrying pthA::npt-sac
in pUFRO47
8.1-kb fragment carrying pthA: :npt-sac
in pUFR004
8.5-kb fragment carrying avrBI03 in pUFRO47
5-kb fragment carrying avrB5 in pUFRO47
8.6-kb fragment carrying Hcl in pUFRO47
5-kb fragment carrying avrBn in pUFRO47
11-kb fragment carrying avrBl04
and avrBS on pUFR047
8-kb fragment carrying avrB104 in pUFRO47
nptl-sac marker in a 3.8-kb BamHl fragment
Cosmid done carrying Hcl and avrBn
Cosmid clone carrying avrBl03
Cosmid clone carrying avrBn
Cosmid clone carrying avrB104 and avrB5
Cosmid clone carrying avrB5
Cosmid clone carrying Hc6
4.5-kb fragment containing pthA in pUFRO47


Leong et al. 1982

De Feyter et al. 1990
De Feyter et al. 1990
De Feyter & Gabriel 1991a
De Feyter et al. 1993
De Feyter & Gabriel 1991b

De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter & Gabriel 1991a
De Feyter et al. 1993
De Feyter et al. 1993
De Feyter et al. 1993
De Feyter et al. 1993
This study

This study

This study
This study
This study
This study
This study

This study
Ried and Collmer 1987
This study
This study
This study
This study
This study
This study
Swarup et al. 1992






42
(Ried and Collmer 1987) was randomly ligated into a Ball site of pthA on pZit45.

Recombinant plasmids were screened for an insertion in the middle of the tandemly

repeated region, and pUFYl.48 was selected, in which the nptl-sac cartridge was found

in the Ball site of repeat number 10 of pthA. An SstI fragment carrying the pthA::nptl-

sac fusion from pUFY1.48 was recloned into the suicide vector pUFR004 (De Feyter et

al. 1990), forming pUFY10.1. Marker-exchange mutants (avr::nptl-sac) were created

by transferring pUFR10. 1 to XcmH 1005 and selecting for colonies resistant to kanamycin

(15 pg/ml) and sensitive to chloramphenicol (35 ;tg/ml) and sucrose (5%). Some

marker-exchange mutants were plated on the PYGM medium containing 5 % sucrose and

nptl-sac marker evicted strains were selected to allow further rounds of marker exchange

mutagenesis.

Plant inoculations and bacterial growth in plant. Cotton (Gossypiwn hirsutum

L.) lines used were Acala-44 (Ac44) and its congenic resistance lines AcB1, AcB2,

AcB4, AcB5a, AcB5b, Acb6, Acb7, AcBIn and AcBIn3 as described (Swarup et al.

1992, De Feyter et al. 1993). Cotton plants were grown in the greenhouse, transferred

to growth chambers before inoculation, and maintained under conditions as described (De

Feyter and Gabriel 1991a). Bacterial suspensions of X. campestris pv. malvacearum (108

cfu/ml) in sterile tap water were gently pressure infiltrated into leaves of 4-6 week old

cotton plants. Pathogenic symptoms were observed periodically 2-7 days after

inoculation.

The growth of X. campestris pv. malvacearum in the susceptible cotton line Ac44

was determined as described (refer to Chapter 2). In plant growth experiments were






43

performed three times. Data points shown in Fig. 3-4 were the mean and standard error

of three samples from one experiment. Bacterial populations released onto the leaf

surface were quantified as described (refer to Chapter 2). These data are reported as the

mean and standard error of three samples. Similar results were obtained in two

independent experiments.



Results



Cloning of six potential avr genes from XcmH by hybridization. In addition to

the six members of the Xanthomonas avr/pth gene family previously cloned from plasmid

pXcmH of strain XcmH (De Feyter and Gabriel, 1991a), Southern blot analyses revealed

the presence of six more DNA fragments in XcmH that hybridized to an internal BamHI

fragment from avrb6 (Fig. 3-1). All of the fragments appeared to be large enough to

carry functional members of the avr/pth gene family. Based on the intensity of

hybridization and the fact that XcmH appears to carry only a single plasmid (pXcmH;

De Feyter and Gabriel, 1991a), the additional hybridizing fragments appeared to be

chromosomal. Colony hybridizations were carried out to isolate these chromosomal

DNA fragments from a genomic library of XcmH (De Feyter and Gabriel 1991a) by

probing with an internal BamHI fragment from avrb6. Twenty-three out of seventy-two

cosmid clones identified by colony hybridization were found to carry DNA fragments not

found on pXcmH. These new, presumably chromosomal DNA fragments were readily

distinguished from the DNA fragments that form pXcmH by restriction fragment length

















rd14







Co~0 to



Fo

I4-R. .













oP.- 0 ce0
OA ~ I-*









E E) E ~
,uU

&,- -..

m b


Ga~







45









* .










e 0-M A
S1.




ate-mal -a a- m


I I I I
mn r'I V* 0
M da 4 A4







46
polymorphisms (Fig. 3-1). Six different chromosomal DNA fragments hybridizing with

avrb6 had been cloned and identified. These DNA fragments were arbitrarily named

according to size as Hcl 6. Six cosmid clones were selected for further study:

pXcml.12 carried Hcl and Hc3, pXcml.21 carried Hc2, pXcm2.23 carried Hc3,

pXcm2.12 carried Hc4 and Hc5, pXcm2.11 carried Hc5 only, and pXcml.22 carried

Hc6 (Fig. 3-1). Restriction digests of these clones revealed that Hcl, Hc2 and Hc3 were

closely linked, and that Hc4 and Hc5 were closely linked.

Avirulence activity of hybridizing fragments. All cosmid clones were conjugally

transferred to the widely virulent X. campestris pv. malvacearum strain Xcml003 and

tested on cotton cv. Acala 44 (Ac44) and 9 cotton lines, congenic with Ac44 and each

carrying one of the following resistance genes: Bl, B2, B4, B5a, B5b, b6, b7, Bin and

BIn3. To localize members of the avr/pth gene family, all cosmid clones with avirulence

activity were directionally subcloned into pUFRO47 by EcoRI and Hindull digestion

(these enzymes are not known cut within members of the family published to date).

pXcml.12 was subcloned to separate fragments Hcl and Hc3, generating pUFY36.8

(Hcl) and pUFY36.26 (Hc3). Fragment Hc2 on pXcml.21 was subcloned to form

pUFY31.46. pXcm2.12 was subcloned to generate pUFY37.62, still carrying Hc4 and

Hc5 on a single, 11 kb EcoRI/HindII DNA fragment. Fragment Hc4 on pUFY37.62

was further subcloned to form pUFY38.1. Fragment Hc5 on pXcm2.11 was subcloned

to form pUFY33.19. Since Hc6 on pUFY1.22 conferred no avirulence to Xcml003, it

was not evaluated further.






47
The indicated subclones were transferred to Xcm 1003 and tested on the congenic

cotton lines and on cotton line 20-3, carrying resistance gene Bn. The Hcl and Hc6

DNA fragments conferred no detected phenotype to Xcml003. The Hc2 and Hc4 DNA

fragments conferred exactly the same avirulence specificity to Xcml003 as did clones

carrying avrBIO1 (De Feyter et at 1993): avirulence on AcB5a and AcBIn3. Therefore

we conclude that Hc2 and Hc4 carry two new members of the gene family, named

avrB103 and avrB104, respectively. The Hc3 DNA fragment on pXcm2.23 conferred

avirulence to Xcml003 on cotton lines containing resistance genes Bn or BIn3. Gene

avrBn had previously been cloned from XcmH on pUFA-HI and shown to confer

avirulence on cotton lines containing Bn (Gabriel et al, 1986), but avrBn was not known

to be a member of this avr/pth gene family. Restriction and Southern hybridization

analyses (not shown) revealed that avrBn is a member of this gene family, and is carried

on both pUFA-HI and pXcm2.23. The Hc5 DNA fragment conferred avirulence to

Xcml003 on AcB5b and AcBIn3. Therefore we conclude that Hc5 carried a new

member of the gene family, named avrB5.

Inactivation of avr genes in XcmH1005. To study the pleiotropic function of all

members of the avr/pth gene family in XcmH1005 (a rifamycin mutant of XcmH),

marker exchange-eviction mutagenesis was carried out to inactivate or delete each of the

avr genes, singly or in combination. More than 140 mutants were generated and 60 of

them were tested on 10 congenic cotton resistance lines and analyzed by Southern

hybridization. Marker exchange-eviction mutants affecting all 10 avr gene were

obtained. In Table 3-2 are listed selected mutants which exhibited altered phenotypes on











I I I


+
I +


I I I


+
I I I +


+
I I + + + +


+
I I I +1 +1 I 0


+
I I I I +


+ +
+ + + + +


r-r

C13
C0 i


'to (
I

o5 re o ,--
1~~ ~ 1100'i?!i
Co CU (0 0 ^ s I
I o i
cu M O Oa
M co cc
to a
I I ^ (U1 01


1 LO0 c (j) -


x XI I 3m I X mm t


0
C0
0
C

0








C
I 8
0
0






0
0


C





Co






0+
0









o.









S-



o4)
0:

0 +1



0 0







S-
O







49
cotton. Southern analysis of 11 mutants was shown in Fig. 3-2. As predicted by gene-

for-gene theory, mutations of avrBIn (HM1.10), avrb6 (HM1.15), or avrB4 plus avrb7

(HM1.38) in XcmH1005 resulted in the loss of avr specificity for cotton blight resistance

genes Bin, b6, B4, and b7, respectively. As reported in the previous chapter, a single

mutation of avrb6 reduces the watersoaking ability of XcmH1005 on susceptible cotton

lines. In contrast, mutation of avrB4, avrb7 or avrBIn did not affect watersoaking ability

of XcmH1005 on Ac44. A mutation in avrB5 (HM1.20) was shown to result in the same

phenotypic defects as the mutation in avrb6, i. e., a loss of avr specificity for resistance

gene b6 and reduction of watersoaking on the susceptible cotton line Ac44. However,

this mutation could not be complemented by the cloned avrB5 gene, instead, was

complemented by the cloned avrb6, indicating avrb6 in HM1.20 was likely to be

affected. Mutation of both avrB5 and avrb6 in HM1.20 and HM1.20S resulted a partial

loss of avirulence specificity for cotton blight resistance gene B5b. Mutation of avrBlO1,

avrB102, avrB103 (HM1.36), avrBIO4 (HM1.32), avrBn (HM1.34) or Hcl (HM1.26)

fragment resulted in no alteration of avirulence phenotype. This is consistent with the

fact that the cotton resistance genes that react with these avr genes also react with other

avr genes (gene-for-genes; De Feyter et al. 1993). No mutations affecting fragment Hc6

were recovered.

To inactivate a large number of avr genes in XcmH1005, the nptl-sac marker-

evicted mutant HM1.20S (avrB5-, avrb6-) was subjected to a second round of marker

exchange mutagenesis. The resulting mutant HM2.2 and marker evicted progeny

HM2.2S had at least 7 avr genes (avrB4", avrB5", avrb6', avrb7, avrBIn', avrBlOl',












1 2 3 4 5 6 7 8 9 10 11 12


a- -


Figure 3-2. Southern hybridization of EcoRI and SstI double digested genomic DNAs of
XcmH mutants. The blot was probed with 32-P labelled internal BamHI fragment of
avrb6. Lane 1, XcmH; lane 2, HM1.10; lane 3, HM1.15; lane 4, HM1.38; lane 5,
HM1.26; lane 6, HM1.36; lane 7, HM1.34; lane 8, HM1.32; lane 9, HM1.20; lane 10,
HM1.20S; lane 11, HM2.2; lane 12, HM2.2S. a, Hcl; b, avrB4; c, avrBl01; d,
avrB102, e, avrBl03; f, avrBn; g, avrBIn; h, avrB104; i, avrb7; j, avrBS; k, avrb6; 1,
Hc6.


t-5


!~Sw


b-
c,d-
e-
g-

I-
h-

jr


.0..






51
avrl02") inactivated. These two mutants not only lost avr specificity on most of cotton

resistance lines, but also lost nearly all watersoaking ability on the susceptible line Ac44

(Tab. 3-2, Fig. 3-3).

Despite repeated attempts to inactivate all members of the gene family in XcmH,

we were unable to reduce the number of members below five. In all cases, members of

the family appeared to rearrange to form new members, as evidenced by the appearance

of new DNA fragments of sufficient size to encode a member, and often accompanied

by the appearance of a new avirulence specificity phenotype (data not shown). Also in

all cases, at least one member remained plasmid-borne, as evidenced by comparatively

strong hybridization intensities characteristic of the plasmid-borne genes.

Effect of loss of watersoaking on bacterial growth in plant. Since avrb6'

mutants of XcmH1005 that had lost significant watersoaking ability on Ac44 were shown

to be unaffected in growth in plant (refer to Chapter 2), we evaluated growth of

HM2.2S on Ac44 in comparison with the parental strain XcmH1005. While inactivation

of 7 avr genes resulted in a loss of nearly all watersoaking ability on cotton, in plant

growth rate and yield of HM2.2S was not affected (Fig. 3-4). However, the amount

of HM2.2S released onto the leaf surface was 1600 times less than that of XcmH1005.

Only 0.007% (8.8 3.0 X 104 cfu/cm2 out of 1.18 0.24 X 109 cfu/cm2) of HM2.2S

bacteria present in the infected leaf tissue were released onto the leaf surface, whereas

9.2% (1.42 0.15 X 10' cfu/cm2 out of 1.55 0.23 X 109 cfu/cm2) of the XcmH1005

bacteria were released onto the surface. Therefore, the mutant HM2.2S was basically

a nonpathogenic endophyte of cotton.






































Figure 3-3. Effect of avr genes on watersoaking symptoms caused by X. campestris pv.
malvacearum on cotton susceptible line Ac44.































0 2 4 6 8 10 12

Day post-inoculation





Figure 3-4. Inactivation of multiple avr genes in X. campestris pv. malvacearum did not
affect bacterial growth in plant on cotton susceptible line Ac44. Filled circle,
XcmH1005; open circle, HM2.2S.






54

Additive effect of different avr genes on watersoaking ability of XcmH1005.

Complementation tests were carried out to further analyze the potential pleiotropic

pathogenicity function of 12 avr genes isolated from XcmH. As shown in Fig. 3-5,

when plasmids carrying single avr genes isolated from XcmH were introduced into the

mutant HM2.2S, only avrb6 was able to partially complement the watersoaking defect

of HM2.2S. None of the other 9 avr genes or Hcl or Hc6 fragments were able to even

partially complement the watersoaking defect in HM2.2S when introduced individually.

However, when plasmids carrying two or three of these same avr genes (not necessarily

including avrb6) were introduced into HM2.2S, many were able to partially complement

the watersoaking defect (Figs. 3-3 and 3-5). For example, plasmids containing avrBIn

or avrBl01 alone conferred no watersoaking activity, but plasmids containing both

avrBIn and avrBl01 together exhibited watersoaking activity. Cosmid clones (pUFR106

and pUFR107) carrying avrb6 and at least two other avr genes were able to fully

complement the watersoaking defects of HM2.2S (Fig. 3-3). Therefore, at least three

of the pXcmH avr/pth genes were redundant in terms of pathogenicity function.

To exclude the possibility that some watersoaking genes were linked to avr genes

on these complementing cosmid clones, subcloning of pUFR107 was carried out.

Activity assays of regions between the avr genes did not reveal any watersoaking activity

when introduced into HM2.2S.


















avrB4 avrb6 avrBl0O avrBIn

1 I II I I I


20.5


avrBI02

I I


12.9


avrB 102 avrb7


avrBlOt avrBIn


avrwBIO avrBIn avrB02


avrBIn avrBI02


avrwBO avrBIn


avrB1Of


Figure 3-5. Pleiotropic watersoaking function of cloned XcmH avr genes in mutant
HM2.2S. HM2.2S was nearly asymptomatic on susceptible cotton line Ac44; + weak
watersoaking activity; ++ intermediate level watersoaking activity; +++ full
watersoaking activity (as wild type XcmH).


pXcmH



pUFR127


11.6 5.0 4.5 4.8


avrb7


avrb6


28.4


pUFR103 avrB4 avrb6


Activity

in HM2.2S


pUFR113


pUFRI09


pUFR111


pUFR112


pUFR107


pUFR106


+


+


++


++


++


avrb7


avrb6


avrB4 avrb6








Discussion



Nearly all of the avr genes reported from the genus Xanthomonas are members

of a single, highly homologous gene family, and with the exception of pthA

pathogenicityy) of X. citri (Swarup et al. 1991, 1992), all of the other published members

were identified and cloned as avr genes. Southern analyses have shown that members

of this avr/pth gene family exist in multiple copies (from 4 to 11) in all tested X.

campestris pv. malvacearum strains isolated from cotton, but that some strains,

particularly those from Africa, appear to lack functional avirulence (De Feyter et al.

1993). For example, the African strain Xcml003 used in this study contains at least five

hybridizing fragments, but is virulent on all tested cotton resistance lines. Of course it

is possible that the hybridizing fragments in a single strain are all nonfunctional, but it

is more difficult to explain the presence of from 4 11 hybridizing DNA fragments

present in all known X. campestris pv. malvacearum strains if they encoded no function

beneficial to the pathogen.

All 12 DNA fragments that hybridize with an internal fragment from the gene

family have now been isolated from a single X. campestris pv. malvacearum strain,

XcmH. Six of them are clustered on a 90-kb indigenous plasmid, pXcmH, and were

originally isolated as functional avr genes, although two of them (avrb6 and avrb7) were

noted to confer pleiotropic pathogenicity (watersoaking) function (De Feyter and Gabriel

1991a; De Feyter et al. 1993). We report here the cloning of the other six hybridizing

fragments that appear to be from the chromosome. One of these fragments encodes the






57
previously described avrBn (Gabriel et al 1986). Three of these encode previously

undescribed avr genes (avrBS, avrB103 and avrB104), and the remaining two fragments

conferred no avirulence on cotton and were presumably non-functional.

Mutations of all ten avr genes in XcmH1005 were achieved by marker exchange-

eviction mutagensis. With the exception of avrb6, mutations of individual members of

the avr/pth gene family in XcmH1005 had no obvious effect on watersoaking on

susceptible cotton lines. Mutation of avrb6 resulted in a significant loss of watersoaking

ability of XcmH1005. Although avrb6 appeared to be the major pathogenicity gene

among 10 homologous avr genes in the strain XcmH, many XcmH1005 avr genes also

contributed to watersoaking function of XcmH1005 on cotton. Therefore these avr genes

also function pleiotropically as pth genes. Mutant HM2.2S had at least seven avr genes

(including avrb6) inactivated, and completely lost the ability to induce watersoaking on

mature cotton leaves. Since HM2.2S suffered deletions of avr genes and adjacent DNA

fragments, mutational analysis alone did not rule out the potential involvement of closely-

linked virulence genes. However, complementation tests confirmed that avr/pth family

members fully complemented the watersoaking defects) of HM2.2S. Individually, none

of the 10 avr/pth genes, other than avrb6, could even partially complement the

nonpathogenic mutant strain HM2.2S in visual assays. However, DNA fragments

carrying multiple avr/pth genes from plasmid pXcmH were able to partially complement

the pathogenic defect of HM2.2S, even if they did not carry avrb6, demonstrating that

other avr/pth genes from XcmH also contributed to its pathogenicity.






58
Some avr/pth genes appeared to have stronger pathogenicity function than others

in HM2.2S, and all six pXcmH genes appeared able to contribute to pathogenicity on

cotton, when tested in combinations. However, no individual avr/pth gene, except

avrb6, exhibited pth function in HM2.2S. Full restoration of the watersoaking defect of

HM2.2S required avrb6 plus two other pXcmH avr/pth genes. Thus, the cotton pathogen

XcmH contained one major and multiple minor genes from the same avr/pth gene family.

Some of the minor avr/pth genes appeared to be redundant in terms of watersoaking

function. Therefore mutation of one or few of minor avr/pth genes would not affect the

watersoaking and release of the pathogen, and presumably, the fitness of the pathogen.

In contrast to avrBs2 of X. campestris pv. vesicatoria (Kearney and Staskawicz

1990), there was no evidence that the avr/pth genes of XcmH are even slightly involved

in bacterial growth in plant. Instead, they were involved in the induction of

watersoaking symptoms and associated release of 9.2% of the total bacteria of the leaf

onto the leaf surface, presumably aiding in dispersal. Mutant HM2.2S, with at least

seven avr/pth genes destroyed, released 1,600 times less bacteria to the leaf surface than

the wild type. Although HM2.2S was asymptomatic on cotton, it maintained the same

level of bacterial growth in plant as the wild type strain. Therefore, inactivation of

multiple avr/pth genes in X. campestris pv. malvacearum created a nonpathogenic

endophyte of cotton. In the case of X. campestris pv. malvacearum, pathogenicity and

host tissue destruction appears to be beneficial for the reproductive fitness of the

pathogen. Since X. campestris pv. malvacearum spreads primarily by rain splash, it

seems doubtful that an endophyte could disseminate efficiently in natural field situations.






59
The lack of correlation between disease symptoms and growth in plant has been

observed with other plant pathogens. Indole acetic acid-deficient strains of Pseudomonas

syringae pv. savastanoi failed to elicit gall on oleander but exhibited similar growth

patterns as the gall-eliciting wild-type strain (Smidt and Kosuge 1978). The lemA gene,

which encodes a two-component regulator, is required by P. syringae pv. syringae for

disease lesion formation on bean plants, but not bacterial growth within or on the leave

of bean (Hrabak and Willis 1994; Willis et al. 1990). Mutation of the fungal pathogen

Colletotrichum magna, resulted in a strain which is not able to induce pathogenic

symptoms, but grows in host tissue as an endophyte and retains the wild-type host range

(Freeman and Rodriguez 1993).

A growing number of Xanthomonas virulence / pathogenicity genes are not hrp

(hypersensitive response and pathogenicity) genes, but instead are host-specific

determinants of disease and/or host range (Gabriel et al, 1993). For example, several

genes affecting host range were identified and complementing DNA clones isolated from

X. campestris pv. citrumelo (Kingsley et al, 1993) and X. camnpestris pv. translucens

(Waney et al, 1992). These genes are host specific and affect growth in some hosts, but

not in other hosts. Other non-hrp Xanthomonas pathogenicity genes, such as pthA of X.

citri and avrb6 of X. campestris pv. malvacearum are host specific and strongly affect

elicitation of pathogenic symptoms. From the present study, it appears that many

members of the avr/pth gene family of X. campestris pv. malvacearum are host-specific

determinants of disease and host range on cotton. Genetically, pthA of X. citri and the

avr/pth genes of X. campestris pv. malvacearum resemble some Rhizobiwn host-specific






60
nodulation genes which are required for host range on some hosts, but which can also

confer avirulence when transferred to other Rhizobium strains with a different host range

(Debelle et al. 1988; Faucher et al. 1989; Lewis-Henderson and Djordjevic 1991).

The pleiotropic function and apparent fitness value of some avr genes do not

imply that most avr genes have selective value for a specific plant pathogenic strain. In

fact, most avr genes reported to date appear to be dispensible and have no obvious

selective value to the pathogen (Gabriel et al. 1993). At least some members of the

Xanthomonas avr/pth gene family are found on self-mobilizing plasmids (for example,

Bonas et al. 1989), and their presence in some strains of a pathovar may be due to a

coincidental linkage with another (selected) factor, such as copper resistance, on the same

plasmid. Once horizontally transferred to a different strain, these genes may not remain

stable enough to retain pth function specific for some ancestral host. In both X.

campestris pv. malvacearum and X. citri, members of the avr/pth gene family mutate at

very high frequency (approximately 10-4) through intergenic or intragenic recombination

(De Feyter et al. 1993; Chapter 4). High frequency recombination provides a powerful

mechanism to generate variation and horizontal gene transfer provides a powerful

mechanism to disseminate that variation in the genus. It seems likely that many avr

genes carried by bacterial plant pathogens may result from variation in redundant and/or

unnecessary virulence/pathogenicity genes which are selectively neutral in natural

populations.











CHAPTER 4
INTRAGENIC RECOMBINATION OF A SINGLE PLANT PATHOGEN GENE
PROVIDES A MECHANISM FOR THE EVOLUTION
OF NEW HOST SPECIFICITIES



Introduction



In gene-for-gene interactions between plants and microbial plant pathogens, host

plant resistance results from the genetic recognition of resistance (R) genes in the plant

and avirulence (avr) genes in the pathogen (Keen 1990; Gabriel and Rolfe 1990). It is

unknown why microbial plant pathogens carry genes that function to limit virulence,

since the majority of these genes are dispensable (Gabriel et al. 1993). For example, to

avoid plant resistance, plant pathogens evade R gene recognition by deletion (De Feyter

et al. 1993), mutation (Kobayashi et al. 1990; Joosten et al. 1994) or transposon

insertion (Kearney et al. 1988) of avr genes. These reports all involve natural

mechanisms whereby avirulence may be lost. Nevertheless, mutational analyses reveal

that gains in avirulence, including new avirulence specificities, are generated in at least

some pathogens at approximately the same frequency as they are lost (Statler 1985).

Although presumably an avr gene may be reactivated by reverse mutation, there are no

reports of natural mechanisms whereby new avirulence specificities or new avr genes are

generated. The mechanisms by which new avr genes evolve are unknown.






62

Xanthomonas citri is the causal agent of citrus canker disease, which occurs

worldwide and is subject to eradication and quarantine regulations in many countries.

A pathogenicity gene, pthA, is essential for the pathogen to cause hyperplastic canker

symptoms on citrus (Swarup et al. 1991). Furthermore, when transferred to other

xanthomonads, pthA confers ability to induce hyperplastic cankers on citrus and a

hypersensitive response (HR) on other hosts (Swarup et al. 1992). Therefore, pthA

exhibits pleiotropic pathogenicity and avirulence functions. Southern hybridization,

restriction analysis and partial DNA sequencing has shown that pthA belongs to a major

avr gene family widespread in the genus Xanthomonas (Swarup et al. 1992). Members

of this gene family include avrBs3, avrBs3-2, and avrBsP (a trancated form of avrBs3-2)

of X. c. pv. vesicatoria (Bonas et al. 1989, 1993; Canteros et al. 1991), avrB4, avrb6,

avrb7, avrBIn, avrBlO1 and avrB102 of X. c. pv. malvacearum (De Feyter and Gabriel

1991; De Feyter et al. 1993), and avrXalO and avrXa7 ofX. oryzae pv. oryzae (Hopkins

et al. 1992). Among these, only avrb6 and pthA are known to have pleiotropic

pathogenicity functions (Swarup et al. 1992; De Feyter and Gabriel 1991; also refer to

Chapter 2). The most striking feature of this gene family is the nearly identical, 102bp

tandem repeats in the central portion of the genes. Deletion analyses have shown that

the avirulence specificity of avrBs3 is determined by the 102bp repetitive motifs of that

gene (Herbers et al. 1992). By swapping the internal repeated regions between pthA and

avrb6, the 102 bp tandem repeats are found to determine both avirulence and pathogenic

specificities (refer to Chapter 2).






63
In most organisms, repetitive regions of DNA are known to be active sites for

homologous recombination (Albertini et al. 1982; Petes and Hill 1988). The high

frequency of such events causes them to be major genome modifying forces in evolution.

Intergenic recombination among members of this avr/pth gene family was proposed to

explain the high frequency of race change mutation found in X. c. pv. malvacearum (De

Feyter et al. 1993). Since both pathogenic and avirulence specificities are determined

by the 102bp tandem repeats, it seemed reasonable to suppose that homologous

recombination between genes and among repeats of the same gene might lead to an

acceleration in the evolution of this Xanthomonas avr/pth gene family, resulting in new

host specificities for the pathogen. In this study, we analyzed the nucleotide sequence

of pthA and demonstrate that intragenic recombination of a single gene can provide a

genetic mechanism for the creation of new avr/pth genes.



Materials and Methods



Bacterial strains. plasmids and matings. Escherichia coli strains DH5a (Gibco-

BRL, Gaithersburg, MD), JM83 (Yanisch-Perron et al. 1985) and derivatives were

grown in Luria-Bertani (LB) medium (Sambrook et al. 1989) at 370C. Xanthomonas

citri strain 3213 (causes hyperplastic cankers on citrus; Swarup et al. 1991, 1992), X.

campestris pv. citrumelo strain 3048 (causes water-soaked leaf spots on citrus and

common bean; Swarup et al. 1991, 1992), X. campestris pv. malvacearum strain

Xcml003 (causes water-soaked leaf spots on cotton; De Feyter and Gabriel 1991) and






64
derivatives were grown in PYGM (peptone-yeast extract-glycerol-MOPS) medium at

300C (De Feyter et al. 1990). Antibiotics were used as described in previous chapters.

Triparental matings were carried out to transfer plasmids from E.coli DH5a to various

Xanthomonas strains by using pRK2013 or pRK2073 as helper plasmids as described

(Swamp et al. 1991; De Feyter and Gabriel 1991a). To transfer plasmids into Xcml003,

the modifier plasmid pUFR054 carrying XcmI and XcmnIII methylase genes was used to

increase the transfer frequency (De Feyter and Gabriel 1991b).

DNA sequence analysis. The 4.1 kb Sall fragment carrying pthA in pZit45

(Swarup et al. 1992) was recloned into the pGEM 7Zf(+) vector (Promega, Madison,

WI) to yield pUFY14.5. Sets of overlapping, unidirectional deletion subclones were

generated in pUFY14.5 using exonuclease HI and mung bean nuclease as described by

Promega (Madison, WI). DNA sequencing was performed by the DNA Sequencing Core

Laboratory of the Interdisplinary Center for Biotechnology Research, University of

Florida, Gainesville. Some sequencing was determined by using the dideoxy nucleotide

chain termination technique with the Amersham (Arlington Heights, IL) multiwell

microtiter plate DNA sequencing system RPN1590. The sequence was analyzed using

the GCG package (version 7.00) by Genetics Computer Group, Inc., Madison, WI.

Gene replacement and selection of marker-evicted strains. Gene pthA on pZit45

was mutated by lighting a 3.9 kb BamHI fragment containing a nptl-sac cartridge (Ried

and Collmer 1987) into a Ball site in one of the 102bp tandem repeats of pthA. One of

the resulting transcriptional fusions, pUFY1.48, carried the nptl-sac cartridge inserted

into repeat #10 of pthA and was selected for further study. An SstI fragment carrying






65
the pthA::nptI-sac fusion from pUFYl.48 was recloned into "suicide" vector pUFR004

(colEl, mobP+, cat+; De Feyter et al. 1990), forming pUFY10.1. Marker-exchange

mutant Xcl.2 (pthA::nptI-sac) was created by transferring pUFR10. 1 to X. citri 3213 and

selecting for colonies resistant to kanamycin (15 pg/ml) and sensitive to chloramphenicol

(35 /g/ml) and sucrose (5%). Marker-evicted strains of X. citri were selected by plating

Xcl.2 on PYGM medium containing 5% sucrose and selecting colonies resistant to

sucrose. Colonies were then screened for sensitivity to kanamycin. Recombinant,

marker-evicted derivatives of pUFY1.48 were selected in X. campestris pv. citrumelo

3048, E. coli DH5a (recA~) and E. coli JM83 (recA+) by plating on PYGM or LB

medium containing 5% sucrose. Sucrose-resistant colonies were screened for sensitivity

to kanamycin.

To calculate recombination frequencies, XcMI.2, 3048/pUFY1.48,

DH5a/pUFY1.48 and JM83/pUFY1.48 were grown in liquid media in the presence of

appropriate antibiotics to densities of 109 cells/ml. Serial dilutions were then plated on

media with and without 5% sucrose. Recombination frequencies were calculated by

dividing the number of sucrose-resistant (also kanamycin-sensitive) colonies by the

number of colonies grown on the medium without sucrose selection (Xu et al. 1988).

Data shown in the results were the mean and standard error of four replicates obtained

from two independent experiments.

Plant inoculations. Citrus (Citrus paradise 'Duncan', grapefruit), common bean

(Phaseolus vulgaris 'California Light Red') and cotton (Gossypium hirsutum L.) plants

were grown under greenhouse conditions. Cotton lines used were Acala-44 (Ac44) and






66
its congenic resistance lines AcB1, AcB2, AcB4, AcB5, Acb6, Acb7, AcBIn and AcBIn3

as described (De Feyter et al. 1993, Swarup et al. 1992). Plant inoculations involving

X. citri or pthA or derivatives ofpthA were carried out in BL-3P level containment (refer

to Federal Register Vol.52, No.154, 1987) at the Division of Plant Industry, Florida

Department of Agriculture, Gainesville. Bacterial suspensions were standardized in

sterile tap water to 108 cfu/ml and pressure infiltrated into the abaxial leaf surface of the

plants. Plant inoculations were repeated at least three times on citrus, and two times on

bean and cotton.



Results



Sequence analysis of pt/A. The complete DNA sequence of pthA was determined

and shown in Fig. 4-1. DNA sequence comparisons revealed that pthA of X. citri is 98%

identical to avrBs3 and avrBs3-2 of X. campestris pv. vesicatoria (Bonas et al. 1989,

1993), 97% identical to avrb6 of X. campestris pv. malvacearum (De Feyter et al. 1993),

and 95% identical to avrXalO of X. oryzae (Hopkins et al. 1992) along its entire length.

As observed with all previously sequenced members of the gene family (Bonas et al.

1993; De Feyter et al. 1993), pthA is flanked by nearly identical 62bp terminal inverted

repeats that precisely define the limits of homology. The predicted amino acid sequence

encoded by pthA is presented in Fig. 4-2. PthA has a calculated molecular weight of

122kD and an isoelectric point of 7.72. The primary region which differentiates the

predicted PthA sequence from the predicted sequences of other members of the family















.- y)
~




0


~o 0
.I~ .~



'b-bc,,
0
.~ 0





1*
o ~





00

0






If~ **~









~


















0 0




> 0




















OK E-
40E
E- 0
13 <
0 0


00[-
@000

00 [
1) C
0 UE-.
000








E5E-'E


u-'Co
000
[-00




00
sa

L)
0 'C
0 E.
U L) E-
U U U





0 0
E. a
0 0

0'C








00-



VE-C
00
00E-
U F <












0E-0
U U E-

0< E0
0 0 E
<-00











0 0
0 0 1
U <


















HE-'E
3 U E-
u0 rl
404























Q 'E-4 Q
U00



E- N



[U E-

Ot- U
O t-'
0'00






0 5-'0
000
[-'00

00E-
0'00
000
000





O- 000
0t-0

[-00

'OBO
000
0 <

S00




'1 N
00 U
<0 EU
U U U E
u <



0 U

< U

U 0U


400
040 M


E-U< 6-E0UU( 00
E- 0 EE-0 0 O 0 <0 0
uu 0 )Cu u 9U L)
0 0 U O UE 2 U O 00 uE-
80~ 00 000F

E-'00 E U-'E.'0'C0
000 00000'C00

OOE- [-000 000 '
'1C' 0000 0'C't-
O6-0 U UE 0oE- 0
OU E 0 00 0U
OE- ED) Sr) 00000 Q 0


O 00 000 0
[- O000'O 0'
00 '0'C00 0 0
t- U0 000 0 0 E
H UE-L)6U< U-O
< 0 QE E- 0 E-[U E
[-00 OI- '0000 E
Du<
000 [-'O E 0000
I 0f- 0: U U U a E- 0
g U -()U E <5 u U [-00 UU0000 0-
E- UUOE E E-0
ou u 0u u E-o3!
t0 0 C- 0'UU0 0 <
EOU 0 0 0 E-E' U0 t- 0
00 000 00EE-U -0<
F. F. 0 DuUO->O 00 aU u 0 UUE.UU uU
00 0E- E- E'UUU- t 0
E-E-' -- E- U 'C0U C0
00 '0'C0 0000t-
00 00000'000
C


000 'C 'CCt''00'0 0
ZL')U< OUOOU u-E- 00000-E--UU U
OE- U < 6E- 0 < < <0 U
L'CO 000 .000 U 00000 E-00 0




'C0'C U 'C u U -E-tE-oo'a o
-[-0U E-'0
0 E- OUt-' E O 0 'C
4000 0t--00U0
[[0 00 00000
000 0t-oo -'Co <() U U L .U 9c ) EDL) L
U U .* <<(- 6uO<6U u
00 E- E- 0'-00-60
0 0 0 0 U-U E-







0[[- OO -'0'-00 0
0-' U'0 LL) C E- 0U00
I., UU00H U0000










UUU UUUUCft-OUU
0000>0'C0000
40000 000'CE''
O'CO [ '0 O 'C0040
000 [-0000 'C000
40 000000'000

E.' t-. 000 00 00000
LOUE U U a U 000
< U E- O<< *UsCo{U<0


0OK uEiOU E- H 0 0
0 o <-E 0 U c- 00
L)E-O u -3 U 0(u 0u L)U< <0 U

H0 U )0 L00 U 90 u
U U0 H U 0 E-

ou u f J Euou 06
E.0u 0oD 00 < 0
V0J U 0 60 U E .0E U

UE-U UU-O EUU UE-U EU 0 0 <

0 6E0 00 000 P
0 U E"U a 0
E- 00 U 0 L)E-UaE

SE-i 0 0UE OU <-000
HE-U6E < 0t 0t< UE-O0
. 0 .UU) UU 0E.0


E- E-; U E E 0 E- 9U 0E
00 U )9 0 000
4 E m OUE 000


E-U<~ < frU-E' E"'U
0QtU ) 0 0 UU 0 u
0 0 000 F 0 EUC u


U t< Ut0 0uu OuU
0 Ll I-'. L)4 u


L UUi, 6 o q ( U <
0 o < E- 0E-'UL) 0 0U E-


000 000000000
0 ~~ 0 0 N0 m
mwiU Wf'O c o0 0C0f
00t O *bU 000 *~


00 E-E- 00U 00-UOE-00 000 U 000UE-E-0
0 0 U00ED EOg- < ~00 <) 0 U0 < -
0 -<- 00 gOE-0 -UO 000 0LE U00<0
06 00 Q0.OE-U OO 0 Qf E QO 0
E 0 0 U 00DE-. 0 0 uN 0 u-0 E- C0 0 '
0 ) 0000 0 000000 u
<6 00Ot3400E0ttS ( U C >U E OUU<(OaDB
000 000~- ~00 00 O 0-
3 00UO 0E0UU' 0004000 00E-
000 .00U0 E S0'C<-.00E-.000E-'C 000
000 000EO00'U0000000'C 000000
UU E-0'C 0 0 C OOOOOOE-.E-E.000O0E-'CO
-0000E
t g 000 ooooC-o00 '0000 0U< OE<
u< uU 300 u00 0000'U 0 0 0E -'
UUOE-U0 0 0< Q0 000UU o u UE00 o 000 B00 0 0 0Uo00 C'0< o0
SOOE-0 E- 000 U E-U E00- ) 0 3E- (u0 0
pE UOUE-g(

1 8o0 C000 00'00000 0 00
o00-. 00 Ua40 UUU 0u-00
0000E.0 0 0000000 0'C00040
FOOEHOUO U u [U o UUU< e E- [" u

HEC 0 0<0 E.0 UE u 0 0 -E- 0 .00uUOE-UO E-
OEOE000 00 < 0000E-E-40004
g 000 0 000oUmE'OU O 0U U00U30U
000000 00000UE-E-E-00 U *u0'CEE00 00 u 0u0000-00000t'C0E
3C0000 000UE- U oO UE0 E00 E-'CE000 E-<



0000-000 0000~00E-000 0000 OE-
000 'C 0 'C E-E.O OOEE-00< 'c 000 ve0
00-000 O E-E-.0 a ooo -o 0e-0 00
S OUU E-UOU 0000E-00-u00 000
cUu UU0 -uo OO OE-UE-UE- U O 00U

uuuouo 6U600U0CHO Co0


080 0800000 00
U00 UOOUOOEuHE-OOEUUH 90E-. E
000U00-00E00000UE-0U 0u-0U00
E-O DO E-'E- 00U OU CE-O0 OE u E00
UE-uo Uo 000000-0O00000'U 00U-

uO 00'UU00000E-OUUE-D0' OE <0
Uo OMO-UE00E-H0' 00 -DOe<00C00

S0 0 U 00 EOU3 0t < E-00


OUU HUO 6-0EUUU00 E 00000 .00 0 E' E-U E-U E-00
E0-O0000 E-0 -0


00 08000@00000080 O
U0E-U00'U00000 -E0~00 0 000E-E
O UE UE-OUE E0000'CUU0000- -0OHE-

0UUO EEOuOCEC 00o0000' 0E-0OU



00&O0000uu0o0000 000E-000
UUUUUIUOU CONEOHOU 0 H H-U UU<<
HHOOUEgH O U UEO
S 0 EO00 0060O' 0<- U-0000
u EEO Ou' E- a 00 0 000 <0u0u





0'C00000 'C000 00 '0 0000 0000'00
U 00 O' < 0'C uu- 00 O E'C 0 uE-E-iE00 00



O0000H0 O 00 0 o0H000'00<00
OU6E-.0OU OU00 O6-. C< O -OUEO04E-*U U0<






00000000 0 00000000400
E COEEOE-.


0 'C0E- 'F E- 'U-O ' u00 00000 H 0o00U< 00 E-E-000


010000000 000 OE-.0 O0000
E D O :04 0ro(-) UQ g

19O-E0-00 00 OE--00 U00'U00
E- 0 UU E- 0 0 0< < E< 0 0 00 00
U 00 00 0'C00 04E-E00'<


0 E U- 0 000u U 0 0E-40 OOE- 3U E-0 'U0
0 0 00E- U 00 O EE- 0 0 'C0 E-E- 00
UE0 0 UO 0 0 0'U00 < I-00 E- 000<< 0 00E-.<00'C0' E-U0E0-
Q0 1- 0 0 L O OL)


000 0000 000 0 00000000
E-OU UU OE 000 O 000' 80U E O
UU^O <:U(OOU








0f 6D E- W 0 0 4 ED 0~ H 0Dr 0 6 e0 () (D r- .' 09 0
E-40UUOH>HONNU<0 HUENN N<-NNN NNt( UUUC m
UE-00 O 0 )U0 < U0 000 0 *104 < u uu L E uE-04
< 0U006 00E0 E 0'U
0 u 0 0 0E.0 U U r000 u E L) 0 E-UE
0 00 0 Q U 0 U 0 0< Q C U U U 0 0 < E- 0 L) U U
< UO L 006 Ca-) 000E. 86UU < a
U< 00 0 u U()0 E0U00E-HUU UE-OU6
u 0 E- 0 0 00 00U
LI ) 0E0 E4 0 uL)0 0 0 E- E- E- E4 E-E- 0
00 0<006 0 0 E- 90 No 000uO EU
1-0 0 E0- u0 u4 40 04 0E0 S 0 -0 400
<< OO U< < U tE- U 0kE r. 0E m 0




00000000000000000000000000
tnOW O ~ -tM ^l^r QDOOowMM TtwOr mQDIT'mem O


























Figure 4-2. Predicted amino acid sequence of PthA. The 17.5 internal tandem repeats
are aligned and numbered for comparison. Letters A through K represent different types
of repeats. indicates the repeats unique for PthA protein. Bold and underlined letters
indicate an amino acid deletion, substitution or specific combinations of amino acid
residues within a given repeat that is unique to PthA (and the combination is not found
in any repeat in any previously published member of the gene family). Three nuclear
localization consensus sequences are double underlined.











MDPIRSRTPS

SPAFSAGSFS


121 MRVAVTAARP

181 EALVGHGFTH

241 LLTVAGELRG


PARELLPGPQ

DLLRQFDPSL

PRAKPAPRRR

AHIVALSQHP

PPLQLDTGQL


PDGVQPTADR

FNTSLFDSLP

AAQPSDASPA

AALGTVAVKY

LKIAKRGGVT


GVSPPAGGPL

PFGAHHTEAA

AQVDLRTLGY

QDMIAALPEA

AVEAVHAWRN


DGLPARRTMS

TGEWDEVQSG

SQQQQEKIKP

THEAIVGVGK

ALTGAPLN


RTRLPSPPAP

LRAADAPPPT

KVRSTVAQHH

QWSGARALEA


LTPEQVVAIA

LTPEQVVAIA

LTPEQWAIA

LTPEQVVAIA

LTPEQVVAIA

LTPEQVVAIA

LTPDQVVAIA

LTPQQVVAIA

LTPEQVVAIA

LTPEQVVAIA

LTPEQVVAIA

LTPEQVVAIA

LTLDQVVAIA

LTPEQVVAIA

LTPDQVVAIA

LTPEQVVAIA

LTPEQVVAIA

LTPEQVVAIA


SNIGGKQALE

SN GGKQALE

SNIGGKQALE

SNIGGKQALE

SNIGGKQALE

SNGGGKQALE

SHDGGKQALE

SNGGGKQALE

SHDGGKQALE

SNGGGKQALE

SNGGGKQALE

SNGGGKQALE

SNGGGKQALE

SNSGGKQALE

SHDGGKQALE

SHDGGKQALE

CNGGGKQALE

SNGGGRPALE


TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQALLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

TVQRLLPVLC

SIVAQLSRPD


A LTNDHLVALA CLGGRPALDA VKKGLPHAPA

AQVVRVLGFF QCDSHPAQAF DDAMTQFGMS RHGLLQLFRR

DRILQASGMK RAKPSPTSTQ TPDQASLHAF ADSLERDLDA

DRAVTGPSAQ QSFEVRAPEQ RDALHLPLSW RVKRPRTSIG

MREQDEDPFA GAADDFPAFN EEELAWLMEL LPQ*


LIKRTNRRIP ERTSHRVADH

VGVTELEARS GTLPPASQRW

PSPTHEGDQR RASSRKRSRS

GGLPDPGTPT AADLAASSTV


QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

QAHG

PALA


900

951

1011

1071

1131






71
is in the central portion of the protein, characterized by 17.5 nearly identical, tandemly

arranged, 34 amino acid repeats. As shown in Fig. 4-2, these repeats may be classified

into 11 types (A-K) depending on slight differences in sequence. Among them, repeat

types B, E, H, J are only found in PthA, and not in the predicted peptide sequences of

any other published family member. Repeat 2 (type B) resulted from an amino acid

deletion and repeat 17 (type J) resulted from an amino acid substitution; neither this

deletion nor this substition has been observed in any other member of the gene family.

Repeat types E (repeats 7 and 15) and H (repeat 13) resulted from unique combinations

of amino acids in specific positions. In addition to the presence of unique repeats, PthA

differs from its homologues in the arrangement of repeat types within the region.

Gene replacement and selection of pthA recombinants in X. citri. X. citri mutant

Xcl.2 (pthA::nptI-sac) was generated from wild type strain 3213 by marker exchange,

using pUFY10.1. Marker exchange was confirmed by Southern hybridization analyses

(Fig. 4-3, compare lanes 1 and 2). When inoculated on plants, Xcl.2 was unable to

induce cankers on citrus or to induce an HR on bean (Table 4-1). By plating Xcl.2 on

a sucrose-containing medium and screening for sucrose resistant and kanamycin sensitive

colonies, many marker-evicted strains of X. citri were obtained at a frequency of 5.22

1.8 x 10-5.

To examine the recombination events, total DNAs were extracted from twenty-

four marker-evicted derivatives of Xcl.2, digested with BamHI and probed with the 35P-

labelled internal Stul/HinclI fragment (containing the tandemly repeated region) ofpthA.


















1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

4Kb-






2Kb-


Figure 4-3. Southern hybridization of BamHI digested, total DNA from X. citri strain
3213 and its pthA mutant derivatives. The blot was probed with the internal 2kb
Stul/HincH fragment (containing the tandemly repeated region) of pthA. Strain 3213
carries four BamHI fragments which hybridize to pthA; the largest fragment, indicated
by the arrow, encodes pthA. Not shown in lane 2 are two additional hybridizing
fragments (1.8 and 1.6 kb in size), resulting from the additional BamHI site in nptl-sac.
Lane 1, 3213; lane 2, Xcl.2 (pthA::nptl-sac); lanes 3-18, nptl-sac-evicted derivatives of
Xcl.2: XcS1, XcS15, XcS19, XcS24, XcS27, XcS30, XcS12, XcS13, XcS17, XcS18,
XcS26, XcS29, XcS32, XcS23, XcS34, XcS26 and XcS21, respectively.










Table 4-1. Phenotypes of Xanthomonas citri marker-evicted strains on citrus and bean.



Host
Strains -----------------------------------------------------
citrus bean


Xc3213 C HR
Xcl.2 0 0
XcS1, XcS9, XcS14 & XcS15 C HR
XcS5 C 0
XcS17 & XcS20 wC wHR
XcS12, XcS13 & XcS24 HR wHR
XcS7 wHR wHR
XcS6 & XcS11 wHR 0
XcS2* 0 0

*: Additional 11 strains (XcS3, XcS4, XcS8, XcS10O, XcS16, Xcsl8, XcS19, XcS20,
XcS21, XcS22 and XcS23); C, canker lesions; wC, weak canker; HR, hypersensitive
reaction; wHR, weak hypersensitive reaction; 0, no symptoms.






74
As shown in Fig. 4-3, the different derivatives of Xcl.2 exhibited different sized bands

(including those similar in size to pthA), indicating that new pthA homologues were

regenerated by homologous recombination among the tandem repeats of pthA. There was

no evidence of recombination between pthA and the other three DNA fragments that

hybridize with pthA, but the presence of the other three fragments raised the possibility

of intergenic recombination.

The twenty-four marker-evicted derivatives of Xcl.2 were inoculated on citrus

and bean plants. They exhibited different plant reaction phenotypes, including some

entirely new phenotypes, distinct from the wild type 3213 strain (Table 4-1). Five

derivatives reverted back to the parental 3213 phenotype, inducing strong canker lesions

on citrus (XcS1, XcS9, XcS14 and XcS15). Two derivatives caused weak canker (XcS17

and XcS20). Six derivatives elicited a previously unreported HR on grapefruit (XcS12,

XcS13 and XcS24). Eleven derivatives caused no pathogenic symptoms. In contrast to

wild type strain 3213, XcS5 retained the ability to elicit hyperplastic cankers on citrus,

but lost the ability to elicit HR on bean, indicating that these two pleiotropic functions

are separately encoded on pthA.

Intragenic recombination of pthA in X. campestris and E. coli. To demonstrate

the occurrence of intragenic recombination, plasmid pUFY 1.48 was introduced into three

bacterial strains which carry no DNA that hybridizes with pthA: X. campestris pv.

citrumelo strain 3048, E. coli strain DH5a (recA-) and E. coli strain JM83 (recA+).

The transconjugants 3048/pUFY1.48, DH5a/pUFY 1.48 or JM83/pUFY1.48 were plated

on sucrose-containing medium and screened for marker-eviction. Again many marker-






75
evicted derivatives were obtained. The frequencies of marker-eviction of pUFY1.48 in

3048, DH5a and JM83 were 4.05 0.44 x 10-4, 1.06 0.13 x 10-2 and 0.97 0.14

x 10-2, respectively.

To examine the recombination events, plasmid DNAs were analyzed from 24

marker-evicted strains from 3048 and 14 strains from DH5a. As shown in Fig. 4-4,

different sized BamHI fragments were observed, demonstrating the generation of new

pthA homologues by intragenic recombination. Southern hybridizations of total DNAs

extracted from these strains probed with the internal Stul/HincIL fragment ofpthA showed

only the presence of the same bands as those generated from the plasmid digestions (data

not shown).

The plasmids derived from the 24 3048 (pCS series) and 14 DH5a (pDS series)

marker evicted strains were individually introduced into X. citri strain B21.2 (a

nonpathogenic, pthA::TnS-gusA derivative of 3213, which is virulent only on citrus),

3048 (virulent on citrus and bean) and X. campestris pv. malvacearum Xcml003 (virulent

only on cotton). The resulting transconjugants were inoculated onto citrus, bean and

cotton plants (Fig. 4-5). Similar to the results observed with derivatives of Xcl.2, these

pthA recombinants exhibited altered phenotypes and/or host specificities. For example,

pCS23 conferred to B21.2 and to 3048 the ability to cause stronger canker lesions on

citrus than that conferred by pZit45 (pthA), and a weak HR often proceeded the canker

symptoms. Unlike pthA on pZit45, however, the recombinant gene on pCS13 conferred

to B21.2 or 3048 the ability to elicit canker lesions on citrus, but not HR on bean.

These results confirmed those indicated by recombinant X. citri mutant XcS5 (refer

























Figure 4-4. A. Intragenic recombinants of pthA generated in X. campestris pv. citrumelo
3048. Plasmid DNAs were digested with BamHl and fractionated in 0.7% agarose gel.
Lane 1, pZit45 (pth,4); lane 2, pUFY1.48 (pthA,::npt-sac); lane 3-19, pCS23, pCS13,
pCS16, pCS2, pCS17, pCS6, pCS1, pCS14, pCS5, pCS7, pCS11, pCS12, pCS15,
pCS19, pCS21, pCS18 and pCS22, respectively. B. Intragenic recombinants of pthA
generated in E. coli. Plasmid DNAs were digested with BamHI and fractionated in 0.7%
agarose gel. Lane 1, pZit45; lane 2, pUFY1.48; lanes 3-16, pDS2, pDS5, pDS4,
pDS13, pDS1, pDS10, pDS14, pDS9, pDS3, pDS6, pDS7, pDS8, pDSll and pDS12,
respectively.














M I 2 3 4 5 6 7 8 9 10111213141516171819


9 kb-


2 kb.


I kb.


M 1 2 3 4 5 6 7 8 9 1011 12 1314 1516


9 kb


4kb-

3 kb-


2 kb-,


1 kb-

















.44-


>bU g4*








00









4- 00
o C4 C









0 lz As



. 0 4) Or












C4.




00
.4J
*0 ac
>,. N
C4 .

042) 0


















0o

8 o



co


n 0 0












C












U'
C1 U







E 8
S> 0 0o CO (0 C) 0 0 0 0 on
a. a. a a. Q. a. a. a. a Q. Q. aQ a.







80
above): the ability to confer hyperplastic cankers on citrus and the HR on bean are

independently encoded on pthA.

Plasmids pCS1 and pCS6 conferred ability to elicit an HR on citrus and bean to

B21.2 and 3048, and ability to elicit an HR on 9 tested cotton lines (Ac44, AcB1, AcB2,

AcB4, AcB5, Acb6, Acb7, AcBIn and AcBIn3) to Xcml003. All pthA derivatives tested

which conferred the ability to induce HR to Xcml003 on normally susceptible cotton line

Ac44, also elicited HR on all tested cotton resistance lines. All pthA derivatives tested

were able to confer ability to induce an HR to Xcm1003 on one or more resistance cotton

lines, indicating that all pthA derivatives tested were functional genes.

Most of the variation within the predicted 34 amino acid repeat sequence of pthA

occurred at positions 12 and 13, and most repeats encoded either asparagine/glycine,

asparagine/isoleucine or histidine/aspartate at these positions. The latter two amino acid

combinations in the pthA repeats could be detected at the DNA level by presence of SspI

or BglI restriction sites, respectively. The SspI restriction enzyme cut specifically after

the 36th nucleotide of the 1st, 3rd, 4th and 5th repeats of pthA, while the BglI enzyme

cut specifically after the 38th nucleotide of the 7th, 9th, 15th, 16th repeats. By

comparing the BgII and SspI fragments of intragenic recombinants with those of pthA,

the specific repeats deleted or affected in 12 recombinants were determined (Fig. 4-5).

Deletions involving the 7th, 8th, and 9th repeats always reduced or abolished the canker-

eliciting ability of pthA (pCS16, pCS17, pDS5, pCS1, pCS6, pCS11, pCS22). Deletions

involving the 4th, 5th, and 6th repeats abolished all avirulence specificity; these

derivatives conferred ability to induce an HR on all citrus, bean and cotton lines tested.






81
Although recombinants with deletions in the 1st, 2nd, 16th or 17th repeats were not

obtained, these results indicated that the specific order of the first nine repeats of pthA

was important for ability to elicit cankers on citrus and for specificity of avirulence.

Deletions of the 10th through the 15th repeats appeared to alter avirulence specificity.



Discussion



The complete DNA sequence of pthA revealed that pthA is highly homologous to

avrb6, avrBs3, avrBs3-2 and avrXalO along its entire length. The pthA gene encoded

a protein of 1163 amino acids with calculated molecular weight of 122 kD. Western blot

analyses similar to those previously published for AvrBs3 (Knoop et al. 1991), but using

anti-PthA serum, revealed that a 122 kD protein was constitutively expressed in X. citri

and X. campestris pv. citrumelo transconjugants containing pthA (refer to Chapter 5).

Despite few differences in the predicted amino acid sequences of all members of the

avr/pth gene family, pthA has a unique ability to induce cell divisions on citrus leaf

mesophyll in addition to its pleiotropic ability to confer avirulence to pathogens of cotton

and bean (Swarup et al. 1992). Both functions were previously found to be localized to

the repeated region of pthA (refer to Chapter 2).

DNA sequence analysis revealed that there were five specific repeats only found

in pthA, but not in other published members of the avr/pth gene family sequenced to date

(avrb6, avrBs3, avrBs3-2 and avrXalO). By analyzing 12 intragenic recombinants with

BglI and SspI restriction enzyme digestion, we were able to approximately locate the






82
regions important for avirulence and pathogenic specificities. The first six repeats of

pthA appear to be essential for the pathogenicity function and any specificity (avirulence

or pathogenicity) at all. Deletions involving repeats 7 to 9 reduced or abolished canker-

eliciting ability of pthA on citrus. Deletions involving repeats 10 to 16 determined

avirulence specificity. These findings may also help to explain previously published data

on avrBs3 deletion derivatives (Herbers et al. 1992; Bonas et al. 1993). From the

published data, all deletions affecting repeats 1 through 6 of avrBs3 abolished all

specificity, although not avirulence per se.

The role of the tandem repeats in determining pathogenic or avirulence specificity

is striking, but it is not known how the repeats determine these functions. One of the

important biological functions for repetitive domains in prokaryotic and eukaryotic

proteins is ligand binding. In a family of clostridial and streptococcal ligand-binding

protein, for example, conserved C-terminal repeat sequences function as binding sites

(Wren 1991). Similarly, 34 amino acid repetitive motifs called tetratricopeptides (TPR)

found in proteins encoded by many mitotic genes (e.g., CDC16, CDC23, nuc2+ of

yeast, BimA of Aspergillus) are implicated to pair with WD-40 repeats found in many

proteins including the P-subunit of G-proteins and some transcriptional factors (Goebl

and Yanagida 1991; van der Voom and Ploegh 1992). It is possible that the repetitive

domain of Pth or Avr proteins may also function as a ligand binding site, specifically

interacting with corresponding plant receptors.

In this study, a nptl-sac cartridge was inserted into the tandemly repeated region

of pthA to serve as a selectable marker to detect homologous recombination between






83
intragenic repeats. Since the production of levansucrase encoded by sac is lethal to gram-

negative bacteria in the presence of sucrose, marker evicted strains can be selected by

the nptl-sac marker (Ried and Collmer 1987). The regeneration of functional genes from

the pthA::nptl-sac fragment on pUFY1.48 in 3048 or DH5a, which contains no DNA

hybridizing to pthA, demonstrated that intragenic recombination occurred. This was

confirmed by restriction analysis of internal tandem repeats of recombinants.

Intragenic recombinants of pthA were generated by two possible mechanisms.

The first mechanism involves a single crossover event among intragenic repeats on the

same plasmid, resulting in the deletion of nptl-sac marker and restoration of a functional

gene. The second mechanism requires a double crossover between plasmids. The

frequency of pthA recombination in E. coli was observed to be recA independent. Strain

DH5a (recA) yielded recombinants at a frequency of 1.06 0.13 x 10-2, while strain

JM83 (recA+) yielded recombinants at a frequency of 0.97 0.14 x 10-2. Since

recombination involving small repeated sequences (less than 1kb) is affected only slightly

by recA mutations (Petes and Hill 1988), recombination of pthA was expected to be recA

independent.

The intragenic recombination frequencies of pthA observed in the E. coli were

higher than those observed in X. campestris pv. citrumelo (4.05 0.44 x 10-4) and X.

citri (5.22 1.8 x 10-5). This may be due to differences in the copy number of

plasmids in which intragenic recombination occurred. Recombination of pthA in X. citri

strain Xcl.2 involved a native plasmid, while recombination in E. coli and X. campestris

pv. citrumelo involved pUFY1.48, which appeared (based on our plasmid DNA






84
extractions) to replicate at a higher copy number in E. coli than in Xanthomonas.

However, since we did not determine plasmid copy number, these results may be due to

different genetic backgrounds in each strain.

In animal pathogens, some pathogenicity genes encode surface proteins with a

distinctive arrangement of tandem repeats which may provide a protective, strain-specific

conformational epitope for evasion of host immunity (Hoft et al. 1989; McConkey et al.

1990; Hollingshead et al. 1987). Intragenic recombination between homologous repeats

has been found to be responsible for antigenic variations in the pathogens such as

Streptococcus, Neisseria, Salmonella, Plasmodiwn and Trypanosoma (McConkey et al.

1990; Hollingshead et al. 1987; Hagblom et al. 1985; Frankel et al. 1989). The

resulting antigenic variability allows pathogens to escape the host immune response. In

this study, we demonstrated that intragenic recombination of a single plant bacterial gene

can: 1) alter host specificity, 2) generate new pathogenic phenotypes, and 3) evade host

plant defenses. The evolutionary success of many plant pathogens may rely on their

ability to avoid host recognition. Tandemly repeated motifs in pthA and other members

of the gene family provide hot spots for recombination, and thus may accelerate evolution

of this gene family. Intragenic recombination between inexact repeats not only

introduces genetic variation, but also has the potential to amplify mutagenic effects. As

shown in the results, homologous recombination can lead to either duplication or deletion

of repeat blocks. A mechanism to generate genetic variation at high frequency may also

have the negative side effect of generating unnecessary, gratuitous avr genes.






85
Previously reported natural mechanisms for race change included deletions (De

Feyter et al. 1993), point mutations (Kobayashi et al. 1990; Joosten et al. 1994) or

transposon inactivation (Kearney et al. 1988) of avr genes and always resulted in a loss

of avirulence on a specific host plant, but not gains of specific avirulence. Because of

the unique 102bp tandem repeats and their role in determining both pathogenic and

avirulence specificities, intragenic recombination of pthA not only caused the loss of

pathogenic or avirulence phenotypes on a specific host, but also simultaneously resulted

in the gain of new pathogenic phenotypes and avirulence specificities. Many intragenic

recombinants of pthA lost the ability to confer canker-like lesions on citrus to 3048, but

conferred a new phenotype by reducing the watersoaking symptoms of 3048. Some

intragenic recombinants lost the ability to confer cankers on citrus and instead elicited

an HR on citrus, indicating that normsensitive (canker) and hypersensitive reactions may

share a very similar signal transduction pathway as initially suggested by Klement (1982).

Other recombinants lost the ability to elicit HR on bean, but retained the ability to induce

cankers on citrus. Therefore, the pleiotropic pathogenic and avirulence functions ofpthA

were separated by the rearrangement of tandem repeats through intragenic recombination.

Members of this avr/pth gene family have been found in many Xanthomonas

species and pathvars. When present in appropriate xanthomonads, many are able to elicit

an HR on diverse plant species including cotton, citrus, bean, tomato, pepper and rice.

As shown in Fig. 4-5, one intragenic recombinant of pthA (pCS6) has the ability to elicit

a strong HR on plants from taxonomically distinct plant families such as Rutaceae

(citrus), Legwuminosae (bean) and Malvaceae (cotton), suggesting that the corresponding






86
resistance genes may be widespread in different plant species. Tomato disease resistance

gene Pto, which encodes a protein kinase, hybridizes with multiple DNA fragments

found in a variety of different plants, indicating the presence of Pto homologues (Martin

et al. 1993). During the evolutionary process, members of the avr/pth gene family might

have moved horizontally among diverse xanthomonads pathogenic to a wide range of

plants (Gabriel et al. 1993; De Feyter et al. 1993), thereby enabling interactions with

members of the same plant gene family in a wide range of plants. As a result, members

of this avr/pth gene family may have accelerated the evolution and diversity of the

corresponding plant R gene family.

The gene-for-gene concept is a major model for research on the reciprocal

evolution between plants and parasites (Thompson and Burdon 1992). In the pathogen,

avirulence genes may be mutated to overcome resistance by evading host recognition.

In the plant, resistance genes may also evolve to recognize new avirulence genes,

resulting in new resistance genes. Recent genetic mapping data have shown that disease

resistance loci (especially those corresponding to different races of the same pathogen)

are often clustered within small genetic intervals, suggesting that intralocus recombination

may be responsible for generating resistance specific for new races of plant pathogens

(Dickinson et al. 1993; Martin et al. 1993; Pryor 1987). Intragenic recombination of

pthA (a member of an avr gene family) demonstrated in this research provides compelling

evidence that intragenic and intralocus recombination may play an important role during

the reciprocal evolution of gene-for-gene interactions.











CHAPTER 5
PLANT NUCLEAR TARGETING SIGNALS ENCODED BY A FAMILY
XANTHOMONAS AVIRULENCE/PATHOGENICITY GENES



Introduction



Interactions of microbial plant pathogens carrying avirulence (avr) genes with host

plants carrying specific resistance (R) genes results in plant defense responses, often

observed as a hypersensitive reaction (HR) that is characterized by rapid plant cell death

at the site of infection. Several models have been proposed to explain the genetically

specific recognition between microbial avr genes and plant R genes in these gene-for-gene

interaction (Gabriel and Rolfe 1990), and different models may be applicable to different

systems. In the elicitor/receptor model, avr genes encode a low molecular weight signal

molecule elicitorr) which is perceived by a sensor encoded by the corresponding R genes

and results in an HR. In Pseudomonas syringae pv. tomato, avrD encodes an enzyme

involved in the synthesis of a low molecular weight glycolipid elicitor (Midland et al.

1993; Smith et al. 1993). This elicitor induces an HR specifically on soybean plants

containing the resistance gene Rpg4 (Keen et al. 1990; Keen and Buzzel 1991; Kobayashi

et al. 1990). An alternative model is the dimer hypothesis in which the protein product

of an avr gene interacts directly with the protein product of a plant R gene or with the

R gene itself (Ellingboe 1982). The extracellular peptide encoded by avr4 and avr9 of

87






88
the fungal pathogen Cladosporium fulvum can directly induce a hypersensitive response

(HR) on tomato cultivars carrying the resistance gene Cf4 or CJ9 (Joosten et al. 1994;

van Kan et al. 1991). However, other than these three examples, the microbial signal

molecules that determine avirulence remain unidentified. In all cases, it is unknown how

the signals encoded by avr genes are specifically recognized by host cells with R genes,

triggering plant defense responses.

The first plant R gene, isolated from tomato, was shown to encode a

serine/threonine protein kinase that is homologous to the receptor protein kinase involved

in Brassica pollen-stigma recognition and to the Raf protein kinase involved in the

mammalian Ras signaling pathway (Martin et al. 1993). Since transcriptional activation

of plant defense genes is modulated by phosphorylation (Felix et al. 1991; Yu et al.

1993), and can be blocked by inhibitors of mammalian protein kinases (Raz and Fluhr

1993), it was proposed that recognition of avr signals by R gene products triggers

phosphorylation cascades, leading to plant defense responses (Lamb 1994).

In mammalian systems, two nearly complete signal transduction pathways used

by growth factors and/or cytokines have been elucidated (Culotta and Koshland 1993).

The best known is the Ras pathway, centered around the protein product of the prototype

oncogene ras (Hall 1993, Schlessinger 1993). Upon binding to extracellular signals, the

activated receptor protein kinase transduces signals to the GRB2, SOS, Ras and Raf

proteins. Raf (a Pto homolog) then initiates the mitogen-activated protein (MAP) kinase

cascade (Crews and Erikson 1993; Nishida and Gotoh 1993), which brings protein

kinases into the nucleus, activating nuclear transcriptional factors. The other route from






89
membrane to nucleus is more direct. In these cases, tyrosine kinases pick up signals

from receptors in the membrane and activate transcription factor subunits in the

cytoplasm, which are transported into the nucleus and induce transcription (Hunter 1993,

Kishimoto et al. 1994). In both cases, activated protein kinases and/or transcriptional

factors need to enter the nucleus in order to induce gene expression which may lead to

different physiological outcomes, including oncogenesis and apoptosis.

Viral pathogens encode some transcriptional factors that can directly enter the

nucleus, modulate gene expression of host cells. The bacterial plant pathogen

Agrobacteriwn tumafaciens also encodes several proteins targeted to the plant cell nucleus

(Citovsky and Zambryski 1993). Transport of proteins into the nucleus is an active

process and requires that proteins contain suitable nuclear localization sequences (NLSs,

Nigg et al 1991). NLSs are usually short stretches of 8-10 amino acids characterized by

basic amino acid residues and proline. NLSs are retained in the mature protein, may

be located at any position as long as they exposed on the protein surface, and can be

present in multiple copies. NLSs are recognized by the receptor, called NLS-binding

proteins, which direct NLS-containing proteins into nucleus through nuclear pores (Silver

1991). NLSs from different organisms, including the prokaryote Agrobacteriwn, are

functional throughout the eukaryotic kingdom. Most of the characterized NLS-containing

proteins were reported from mammalian virus and animal cells. Many are oncogene

products, hormone receptors and transcriptional factors, including those involving later

steps of the Ras signaling pathway such as the transcriptional factor AP-1 (Jun/Fos) (Kerr

et al. 1992). A few NLS-containing proteins from Agrobacterium, plant and plant virus






90
have also been characterized (Citovsky and Zambryski 1993; Restrepo et al. 1990; van

der Krol and Chua 1991; Varagona et al., 1991, 1992). To the best of my knowledge,

however, NLSs encoded by avr genes of plant pathogens have not been reported.

Recently, a large family of avirulence (avr)/pathogenicity (pth) genes has been

identified in many species and pathovars of Xanthomonas, a major group of plant

bacterial pathogens. Members includes avrBs3, avrBs3-2 and avrBsP of X. campestris

pv. vesicatoria (Bonas et al. 1989, 1993; Canteros et al. 1991), avrB4,avrb6, avrb7,

avrBIn, avrBlOl, and avrB102 of X. c. pv. malvacearum (De Feyter and Gabriel 1991;

De Feyter et al. 1993), pthA of X. citri (Swarup et al 1991, 1992), and avrxa5, avrXa7

and avrXalO of X. oryzae pv. oryzae (Hopkin et al. 1992). This avr/pth gene family

comprises the majority of described Xanthomonas avr genes and constitutes a large

portion of all avr genes cloned to date (Gabriel et al 1993). The pthA gene of X. citri

and avrb6 of X. c. pv. malvacearum are particularly intriguing because of their

pathogenicity functions (Swarup et al. 1992, Chapter 2). Gene pthA is essential for X.

citri to induce cell division specifically on citrus plants leading to hyperplastic canker

lesions. Gene avrb6 is important for X. c. pv. malvacearum to induce watersoaking

symptoms specifically on cotton. These two genes also function pleiotropically as avr

genes.

Despite the diverse avirulence and pathogenic specificities encoded by these

avr/pth genes, their DNA sequences are nearly identical (95-98%) (De Feyter et al,

1993, Chapter 4). Both pathogenic and avirulence specificities of the avr/pth genes are

determined by the nearly identical, 34 amino acid tandem repeats encoded in the central






91
region of the genes (Herbers et al. 1992; Chapter 2). To account for the precise

recognition and diverse specificities, it has been speculated that the avr/pth protein

product may directly mediate the plant hypersensitive response (Herbers et al. 1992), as

initially suggested in the dimer hypothesis.

In this study, we demonstrated that pthA conferred to X. campestris pv.

malvacearum (which is nonpathogenic on citrus) the ability to elicit hyperplastic canker

symptoms on citrus, independently of bacterial growth in plant. Although the

proteinaceous elicitors encoded by pthA or avrb6 were not detected three nuclear

localization sequences were identified to be encoded by all members of the Xanthomonas

avr/pth gene family sequenced to date. Furthermore, the C-termini of PthA and Avrb6,

containing the NLSs were shown to be capable of directing P-glucuronidase into the

nuclei of plant cells.



Materials and Methods



Bacterial strains and plasmids. Escherichia coli strains used were DH5a (Gibco-

BRL, Gaithersburg, MD) and BL21(DE3)pLysS (Novagen, Madison, WI). Xanthomonas

strains used were X. citri 3213 (causes hyperplastic cankers on citrus, Swarup et al.

1991, 1992), its marker exchange mutant Xcl.2 (pthA::npt-sac, causes no symptom on

citrus), X. campestris pv. citrumelo 3048 (causes water-soaked leaf spots on citrus and

common bean, Swarup et al. 1991, 1992), X. campestris pv. malvacearum XcmH1005

(causes water-soaked leaf spots on cotton) and its mutants XcmH1407 (avrb6::Tn5-gusA,






92
refer to Chapter 2), HM2.2S (seven avr genes deleted, refer to Chapter 3). E. coli

strains were grown in Luria-Bertani (LB) medium (Sambrook et al. 1989) at 370C or

30C. Xanthomonas strains were grown in PYGM (peptone-yeast extract-glycerol-

MOPS) medium at 300C as described (De Feyter et al. 1990). Plasmids pUC19

(Yanisch-Perron et al. 1985) and pUFRO47 (De Feyter et al. 1993) were used as cloning

vectors. Plasmid pUFR127 (De Feyter and Gabriel 1991a) contains avrb6 on a 5-kb

DNA fragment. Plasmid pZit45 (Swarup et al. 1992) contains pthA on a 4.5-kb DNA

fragment. Plasmid pUFY20 contains a chimeric gene consisting of the 5' and 3' regions

of avrb6 and internal repeated region of pthA (refer to Chapter 2). Plasmids pCS6 and

pCS23 contain intragenic recombinants of pthA (refer to Chapter 4). Triparental

matings were carried out to transfer plasmids from E. coli DH5a to Xanthomonas strains

by using pRK2013 or pRK2073 as helper plasmids (Swarup et al. 1991, De Feyter and

Gabriel 1991a). To transfer plasmids into X. campestris pv. malvacearum strains, the

modifier plasmid pUFY054 carrying XcmI and XcmIII methylase genes was used to

increase the transfer frequency (De Feyter and Gabriel 1991b).

Preparation of potential elicitors and plant inoculations. To detect elicitor

activity, crude lysates were prepared as described (He et al. 1993) from Xanthomonas

and E. coli cells carrying pthA or its derivatives. Basically, bacterial cultures were

grown overnight, harvested by centrifugation and resuspended in 10mM Tris-HCl

(pH8.0) solution at an ODr00 of 0.5-1. Bacterial cells were lysed by lysozyme

treatment (2mg/ml lysozyme in 10mM Tris-HCl,pH8.0) or disrupted by sonication in the

presence of 1 mM PMSF (phenylmethylsulfonyl fluoride). PthA protein was also






93
overexpressed in E. coli cells and purified by affinity chromatography (described later).

Cell-free suspensions (in 10mM Tris-HCl, pH8.0) or affinity column purified PthA were

immediately infiltrated into leaves of citrus and cotton plants.

Citrus (Citrus paradise 'Duncan', grapefruit) and cotton (Gossypium hirsutum)

plants were grown under greenhouse conditions. Cotton lines used were Acala-44 (Ac44)

and its congenic resistance lines Acb5b, Acb6, AcBIn3 as described (De Feyter et al.

1993; Swamp et al. 1992). X. campestris pv. malvacearum transconjugants carrying

pthA or its derivatives were inoculated into leaves of citrus plants in BL-3P level

containment (refer to Federal Register Vol.52, No. 154,1987) at the Division of Plant

Industry, Florida Department of Agriculture, Gainesville.

Preparation of antisera and immunoblotting. A 3.7kb BamHI fragment carrying

intact pthA was generated from partial digestion of pZit45 (Swarup et al. 1992), and

cloned into expression vector pET-19b (Novagen, Madison, WI). The fusion sites of the

resulting construct pUFY50.13 were determined by DNA sequence analysis. PthA protein

was overexpressed in E. coli strain BL21(DE3)/pLysS containing pUFY50.13 in the

presence of 1mM IPTG (isopropyl-P-D-thiogalactopyranoside). Since proteins expressed

on pET-19b are designed to contain 10 histidine residues at the amino terminus as an

affinity handle, overexpressed PthA was readily purified by His-tag affinity

chromatography (Novagen, Madison, WI). For immunization of rabbits, PthA was

further purified by SDS-polyacrylamide gel electrophoresis. Antiserum against PthA was

prepared from a New Zealand rabbit by Cocalico Biologicals, Inc. (Reamstown, PA).

For immunoblotting, E. coli or Xanthomonas cells were harvested by centrifugation and






94
resuspended in sample buffer (50 mM Tris-HC1, pH6.8, 10% glycerol, 2% SDS, 0.1%

bromophenol blue). After samples were boiled for 2 min and centrifuged for 5 min, they

were separated on 8% SDS-polyacrylamide gels and transferred to nitrocellulose by

electroblotting (Towbin et al. 1979). Western blots were probed with anti-PthA antibody

and visualized with goat anti-rabbit antibody conjugated with alkaline phosphatase.

DNA constructs for GUS fusion proteins. The DNA coding sequences for the C-

terminal region (191 amino acids) of PthA and Avrb6 were amplified by PCR from the

plasmid pZit45 (Swarup et al. 1992) by using the following primers:

1. CTCTAGAGCCATGACGCAGTI'C; 2. CAGATCTCTGAGGCAATAGCTC.

The amplified fragments were first cloned into pGEM-T (Promega Inc. Madison,

WI), then cut out with Xbal/BglII and recloned into the XbaI/BamHI site of plant

expression vector pBI221 (Clonetech, CA). The resulting plasmids pUFY082 and

pUFY083 contain the chimeric gene with the C-terminal coding sequence of PthA and

Avrb6, respectively, translationally fused with 5' end of GUS gene. Both pUFY082 and

pUFY083 were checked by sequence analysis at the fusion site.

Onion transformation system. The Helium Biolistic gene transformation system

(Du Pont) was used to transform epidermal cell layers of white onion as described

(Varagona et al. 1992). Inner epidermal layers were peeled and placed inside up on

Petri dishes containing MS (Murashige and Skoog 1962) basal media (4.4 g/L MS salt

mixture, Sigma, M5519, 30 g/L sucrose, pH 5.7) with 2.5 mg/L amphotericin B (Sigma)

and 6% agar. DNA samples were prepared as described (Taylor and Vasil 1991). Five

jug of column-purified (Magic Prep, Promega) plasmid DNA was precipitated onto 1 mg






95
of 1.6 pim gold particles using 50 4L of 2.5 M CaC!2 and 10 ,L of 100 mM spermidine.

DNA-coated gold particles were washed with 50 pd of 70% and 100% ethanol and finally

resuspended in 60 /L of 100% ethanol. The samples were dispersed by submerging in

a waterbath sonicator for a few seconds. Five pL samples were pipetted onto particle

delivery discs and delivered into onion epidermal cell layers at pressures of 1100 or 1300

p.s.i.. After microprojectile bombardment, Petri dishes were sealed with parafilm and

incubated overnight at 280C in the dark. The assays for nuclear localization activity of

the fusion protein were repeated in three independent transformation tests.

Histochemical analysis. The histochemical GUS assay was used to determine the

location of GUS fusion proteins in onion cells as described previously (Jefferson, 1987,

Varagona et al 1992). Onion epidermal layers were incubated at room temperature in

X-glu solution (50 mM phosphate buffer, pH7.0, ImM EDTA, 0.001% Triton X-100.

0.05mM sodium ferricyanide and ferrocyanide, 2mM X-glu). After detection of blue

color, onion epidermal layers were mounted on a glass slide with a solution containing

20/tg/ml of the nucleus-specific dye 4',6-diamidino-2-phenylindole (DAPI, 0.1 X

phosphate buffer, 10mM sodium azide, 90% glycerol). Cellular location of the blue

indigo dye produced by oxidative dimerization of the GUS product was determined under

bright-field optics using a Zeiss Axiophot microscope and compared with the location of

DAPI-stained nuclei under fluorescence optics.