. V .
u 5.. .. .
u 3.3") L :5;

 

1;... .(.

||

WWJéE hihuzfgf.” . , .. la. mama"

 

l‘l.‘

willMillil lilillllli

93 0177

This is to certify that the

thesis entitled

Genetic variation in strains of Clavibacter
michiganensis subspecies michiganensis and
the development of bird's eye fruit lesions on tomatoes

 

 

 

presented by

Carmen M. Medina-Mora

has been accepted towards fulfillment
of the requirements for

M. 8. degree in mm Plant Pathology

 

WWW

Major ’progssor

Date May 12, 1999

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY

Michigan State
University

 

 

 

 

PLACE IN RETURN BOX to remove this chedtout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

FEBE E {555239.11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/” WWW“

 

GENETIC VARIATION IN STRAINS OF CLA VIBACTER
MICHIGANENSIS SUBSP. MICHIGANENSIS AND THE DEVELOPMENT
OF BIRD’S EYE FRUIT LESIONS ON TOMATOES
By

Carmen M. Medina-Mora

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Botany and Plant Pathology

1 999

ABSTRACT
GENETIC VARIATION IN STRAINS OF CLA VIBACTER
MICHIGANENSIS SUBSP. MICHIGANENSIS AND THE DEVELOPMENT OF
BIRD’S EYE FRUIT LESIONS ON TOMATOES
By

Carmen M. Medina-Mora

Bacterial canker of tomato is a major concern worldwide because disease
occurrence is sporadic and may result in large yield reductions. Computer-assisted
analysis of repetitive sequence-based Polymerase Chain Reaction (rep-PCR) genomic
fingerprints generated with the universal primers BOX and (GTG)5 suggested genomic
variation in Clavibacter michiganensis subsp. michiganensis (Cmm) is relatively limited.
However, (GTG)5-PCR fingerprint patterns do differentiate strains within the previously
designated BOX-PCR types (A, B, C, D, and E). Intraspecific diversity using combined
(GTG)5 -PCR and BOX -PCR data is as follows; C > A 3 B > D > B. These data may
provide useful genetic makers to trace Cmm strains used in epidemiological studies.

Fruit spotting resulting from Cmm infection is one of the most unpredictable
tomato disease symptoms of bacterial canker in Michigan. Therefore, fruit spot
formation was studied by spraying flowers with Cmm at various developmental stages.
Maximum incidence of spotted fruit and maximum severity of fruit lesions resulted when
inoculum was applied twice, a possible indication that a bacterial population threshold is
required for the onset of fruit spots. Chemical control strategies could be formulated to

reduce the level of unmarketable tomato fi'uit.

Copyright by
Carmen M. Medina-Mora
1 999

 

To my parents, husband, and
specially to my grandfather, Ramén I. Mora, on his memory...
With a lot of love and respect this is for you...

iv

rm
Us“

'0”.

l

til

[if

":3
L:

‘0‘.

'r‘
.1 Li

bath

ACKNOWLEDGMENTS

I would like to acknowledge my committee members, Dr. Dennis W. Fulbright,
Dr. Mary K. Hausbeck, and Dr. Frans J. de Bruijn. Dennis, I am very grateful that on the
daily basis you nourished my scientific growth by asking challenging questions that
inspired constant critical thinking. Thank you for teaching me that there are more than
one route to get to the top, for letting me be part of your laboratory and consider my
opinion, and for all the side activities where I learn by observation what is to be a
scientist and even more a teacher. Also, for being near me through thin and thick without
even noticing it, and specially for becoming my FRIEND. Mary, thanks for believing in
me as a scientist, for planting the seed of professional excellence on me, for your special
guidance on being a female scientist in this field, and for financial support. Frans, thank
you for opening the doors for me at MSU, for recognizing me as a PCR expert and, for
providing me with an indispensable tool for my Master’s project.

Thanks to every past or present member of Dr. Fulbright’s laboratory, Mario
Mandujano, R. Scott Shaw, Virginia Rinkel, Steven Berkheimer, and Christa Chatfield
who made me feel right at home, and shared professional and personal moments. My
gratitude and deep respect to Dr. Julia Bell, for her support not only in research but also
for her words of advice when I needed a kind of mom’s talks. Also, for trusting me, and
sharing with me the bacterial canker project. Special thanks for the rides home every
now and then! Also, my special appreciation to Nicole Werner, for sharing with me the

bacterial canker project as well as Cmm strains, and for the hours in front of GelCompar,

and John Rogers, for your indescribable help with statistics, and for your special attention
in regards to my cultural background.

I would like to give special thanks to my Puertorican friends, Lycely Sepulveda,
Hector Ayala, and Marcos Morales, who always made me remember where I came from,
and to be proud of my culture. My deepest gratitude goes to all of you, for being near me
in good and bad times, and for being my family at MSU.

Also, I would like to acknowledge my family, specially to my Mom and Dad,
who taught me to fight for what I believe in and want from life, and who teach me the
right way of doing whatever your hands and mind are on. Furthermore, thank you for
showing me, through your own life, that in this life you have to give up some things to
become a better person. And, thanks for many years of a wonderful childhood and, for
giving me the pleasure and joy to be a sister, to be a teacher, and to be a friend of
Anarnaris.

And last but not least, I will like to give thanks to my left hand, my partner, my
family, my friend, my love, and my husband, Vladimir F errer, for being always there on
your own special way, day by day, providing me with strength when my faith was
limited. For believing in me every time I have doubts about my professional career, and
give me the liberty to grow as an individual. Furthermore, for just being YOU.

Thank you; I would not have reached this goal without all of you.

vi

GET
8H

CH:
DE\

WC

COX.

APP E

TABLE OF CONTENTS

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

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

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW
Introduction .................................................................................. 1
Bacterial canker of tomato ................................................................. 5
Symptoms .................................................................................... 6
Infection avenues ........................................................................... 9
Disease cycle- Source of inoculum ...................................................... 10
Dissemination of C. michiganensis subsp. michiganensis ........................... 11
Management strategies .................................................................... 13
Characterization and Identification ...................................................... 15
DNA fingerprinting ........................................................................ 23
Taxonomy .................................................................................. 26
Objectives .................................................................................. 30
References .................................................................................. 33

CHAPTER 2

GENETIC VARIATION IN STRAINS OF CLA VIBAC T ER MICHIGANENSIS
SUBSPECIES MICHIGANENSIS

Introduction ................................................................................. 47
Materials and Methods .................................................................... 51
Results ....................................................................................... 57
Discussion .................................................................................. 91
References .................................................................................. 99

CHAPTER 3

DEVELOPMENT OF BIRD’S EYE FRUIT LESIONS AFTER THE
APPLICATION OF CLA VIBAC T ER MICHIGANENSIS SUBSPECIES
MICHIGANENSIS TO PROCESSING TOMATO FLOWERS

Introduction ............................................................................... 104
Materials and Methods .................................................................. 107
Results ..................................................................................... 1 14
Discussion ................................................................................. 122
References ................................................................................. 129
CONCLUSIONS ................................................................................. 1 33
APPENDIX A
BACTERIAL CANKER: SYMPTOMS ...................................................... 137

Vii

APPENDIX B

BACTERIAL CANKER: DISEASE CYCLE ................................................ 139
APPENDIX C
BACTERIAL CANKER: DISEASE CYCLE COMPONENTS ........................... 141
APPENDIX D
BACTERIAL CANKER: CONTROL ......................................................... 143
APPENDIX E
CHARACTERIZATION AND IDENTIFICATION ........................................ 145

viii

la

LIST OF TABLES

Table 1 Strains of Clavibacter michiganensis subspecies used in this study ............... 52

Table 2 Fruit spot development afier spraying open corolla flowers with

strains of Clavibacter michiganensis subsp. michiganensis (Cmm68
and Cmm292). ........................................................................... 117

Table 3 Fruit spot development after direct application of Clavibacter

michiganensis subspecies michiganensis to the surface of young
tomato fruit .................................................................................. 119

Table 4 Fruit spotting afier spraying flowers at various developmental stage
with Clavibacter michiganensis subspecies michiganensis (Cmm68) . ....121

Table 5 Various components of the bacterial disease cycle caused by
Clavibacter michiganensis subsp. michiganensis ................................... 141

Table 6 Management strategies for control of bacterial canker caused by
Clavibacter michiganensis subsp. michiganensis ................................... 143

Table 7 Phenotypic and genotypic methods used for the characterization and
identification of C Iavibacter michiganensis subsp. michiganensis ............... 145

ix

Figu

Fian-

figui

Figure

“ng6

LIST OF FIGURES

Figure 1 Cluster analysis (UPGMA) of (GTG)5-PCR-generated genomic
fingerprints of strains of Clavibacter michiganensis subspecies .................. 59

Figure 2 Cluster analysis (UPGMA) using product moment correlation of
BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains used to define the
experimental variability baseline ....................................................... 62

Figure 3 Cluster analysis (U PGMA) using product moment correlation of
(GTG)5-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains used to define the
experimental variability baseline ....................................................... 64

Figure 4 Cluster analysis (U PGMA) using product moment correlation of
two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of Clavibacter michiganensis subsp.
michiganensis strains used to define the experimental variability
baseline .................................................................................... 66

Figure 5 Cluster analysis (UPGMA) using product moment correlation of
two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of Clavibacter michiganensis subsp.
michiganensis strains designated as type A ........................................... 70

Figure 6 Clavibacter michiganensis subsp. michiganensis strains
designated type A used for the cluster analysis (UPGMA) of the
linearly combined (GTG)5-PCR and BOX-PCR-generated genomic
fingerprints in Figure SC ............................................................... 71

Figure 7 Cluster analysis (U PGMA) using product moment correlation of
two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of Clavibacter michiganensis subsp.
michiganensis strains designated as type B .......................................... 74

Figure 8 Clavibacter michiganensis subsp. michiganensis strains
designated type B used for the cluster analysis (UPGMA) of the
linearly combined (GTG)5-PCR and BOX-PCR—generated genomic
fingerprints in Figure 7C ................................................................ 75

Figure 9 Cluster analysis (U PGMA) using product moment correlation of
two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of Clavibacter michiganensis subsp.
michiganensis strains designated as type C ........................................... 78

K'QWJ —

FY“!
.__. .
r4

'1

firm

FELL?

It,
I ‘7, P
1 Ashe

Figure 10 Clavibacter michiganensis subsp. michiganensis strains

designated type C used for the cluster analysis (U PGMA) of the
linearly combined (GTG)5-PCR and BOX-PCR-generated genomic
fingerprints in Figure 9C .............................................................. 79

Figure 11 Cluster analysis (UPGMA) using product moment correlation of

two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of Clavibacter michiganensis subsp.
michiganensis strains designated as type D ........................................ 83

Figure 12 Clavibacter michiganensis subsp. michiganensis strains

designated type D used for the cluster analysis (UPGMA) of the
linearly combined (GTG)5-PCR and BOX-PCR-generated genomic
fingerprints in Figure 11C ............................................................. 84

Figure 13 Cluster analysis (UPGMA) using product moment correlation of

two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of Clavibacter michiganensis subsp.
michiganensis strains designated as type B ........................................ 86

Figure 14 Clavibacter michiganensis subsp. michiganensis strains

designated type B used for the cluster analysis (U PGMA) of the
linearly combined (GTG)5-PCR and BOX-PCR-generated genomic
fingerprints in Figure 13C ............................................................ 87

Figure 15 Cluster analysis (U PGMA) using product moment correlation of

two linearly combined (GTG)5-PCR and BOX-PCR-generated
genomic fingerprints of eighty seven C Iavibacter michiganensis subsp.
michiganensis strains used in this study ............................................ 90

Figure 16 Spray inoculation method used to study fruit spot development

on processing tomatoes. A) tomato flowers representing open corolla

stages; B) use of plastic cup method to prevent spread of aerosol to

other flowers and plants; and C) incubation of flowers for 16-18 hours

with plastic bag used for third (1X Cmm + bag) and fourth (2X Cmm

+ bag) treatments ...................................................................... 110

Figure 17 Tomato flowers representing four of the five developmental stages

Figure 18

used in the experiment to determine the susceptibility of flowers to
infection with C. michiganensis subsp. michiganensis ........................... 113

Tomato fruit showing bird’s eye lesions eight weeks after C.

michiganensis subsp. michiganensis inoculum was sprayed on
flowers with an open corolla ......................................................... 116

xi

Fa,—

Figure 19 Bacterial canker symptoms on mature tomato plants and fruit ................. 137

Figure 20 Disease cycle of the bacterial canker disease of tomato caused by
Clavibacter michiganensis subsp. michiganensis ................................ 139

xii

 

Ron:

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

Heterotrophic bacteria associated with plants can be classified as parasites and
pathogens, when they invade and live in plant tissues; epiphytes, when they live on the
epidermis of the living plant tissue; or, saprophytes, when they degrade dead plant tissue.
Various combinations also exist, for example pathogens may exist as epiphytes before, or
as saprophytes after invading and killing tissues. Bacterial species may also demonstrate
a specialization in terms of the host species, as well as the type of host tissue colonized.
Epiphytic populations of bacteria can be found on leaves, stems, or buds, and may serve
as the source of primary inoculum of disease in epidemics; but, the actual pathogenic
phase may be limited to specific tissues on the plant. It would appear that the epiphytic
stage serves the bacterium best, in terms of growth, in that populations are maintained on
a living host absorbing nutrients without inciting disease and harming its host (Tsiantos,
1987; Goto, 1992). The epiphytic stage allows for population increases but it may not
provide the needed protection for dissemination or overwintering (Gitaitis, 1989; Atlas &
Bartha, 1998). Once a pathogenic stage is achieved and death of the host tissue follows,
the bacterial species must shift to the saprophytic stage, where bacteria must compete
against well adapted bacterial and fungal species specializing in colonization of dead
tissue and soil survival (Goto, 1992). Phytopathogenic bacterial species, those bacterial
species that cause plant diseases, frequently produce identifiable symptoms resulting
from the infection process. Bacteria that cause leaf blighting and spotting generally

infect through stomatal openings, during physiological conditions that promote high

wour
(Alia
1P6 ‘
in i:
1987;

can 51

Bari}:
subtrc

flan}

1981; .
Clucjfe
Wilt

SFSIEin:

epiphytic populations due to physical action of wind driven rain or irrigation (Agrios,
1997). Bacterial species that inducing wilt can invade xylem tissue through natural
openings, such as lenticels, or through wounds or insect bites. Still, other bacterial
species may cause galls and hypertrophies of stem and root tissue after colonizing
wounds, while other species specialize in post harvest decay of plant tubers and fruits
(Atlas & Bartha, 1998). Few species of phytopathogenic bacteria produce more than one
type of symptom while colonizing their host, and generally, have a particular host range,
that is, they are restricted in the species or variety of plant they can infect (Tsiantos,
1987; Goto, 1992). Those pathogens that do produce symptoms on various tissues and
can survive as epiphytes and saprophytes have received special attention due to their
ability to cause severe crop losses.

Xanthomonas campestris pv. campestris, causal agent of black rot of crucifers,
Burkholderia (Pseudomonas) solanacearum, causal agent of wilt of various tropical or
subtropical plants, Erwim'a amylovora, causal agent of fire blight of apple and pear and
Clavibacter michiganensis, are only a few examples of pathogens that have the ability to
cause large crop losses and show special adaptations in their pathogenic, epiphytic and
survival stages. Xanthomonas campestris pv. campestris, is known to overwinter as an
epiphyte on wild host species that live through the winter at the edges of fields (Schaad,
1981; Jones, 1991). The ability of this pathogen to disseminate from the crucifer weed to
crucifer crop species is well known, but not well understood. Once infected, the host
produces various symptoms including leaf necrosis, xylem colonization leading to

systemic infection and ultimately a rot or disintegration of the host plant (Jones, 1991;

ha

01 i
m FL
935?

diff-

 

 

Agrios, 1997). Alternatively, the pathogen can initiate the infection process through seed
dispersal and seedling infection.

Burkholderia solanacearum infects a large number of unrelated host species
including banana, solanaceous species, and various ornamental species (Goto, 1992).
This pathogen is not only unique because of its ability to infect many species, but also in
its ability to survive in tropical and subtropical soils. Due to these characteristics, it can,
for example, cause serious disease on newly planted banana plantations (Woods, 1984).
Erwim'a amylovora is different than B. solanacearum in that it is restricted in its host
range to closely related species in the Rosaceae family, including apple, pear and
pyrocantha (Agrios, 1997). It overwinters in flower buds and can cause leaf and flower
blight and stem cankers. This pathogen can be found infecting trees of established
orchards as well as on grafted nursery stock in nurseries (McManus & Jones, 1994).

Clavibacter michiganensis infects many different hosts and causes different types
of symptoms on these hosts. Currently five subspecies are recognized, including C.
michiganensis subpecies michiganensis, nebraskensis, sepedonicus, tessellarius, and
insidiosum, based on the analysis of total protein profiles, immunological and metabolic
differences such as pigment production and carbon source utilization patterns (Davis et
al., 1984). The subspecies are restricted in their host range, for example, in other
pathogenic species, the variants infecting different host are often called pathovars or
pathogenic variants (Schaads, 1987; Agrios, 1997). This nomenclature was not adopted
for Clavibacter because the characteristics of the subspecies are greater than just the

pathogenic variation.

, v
illicit

Lita;
only 'i
P???
Strut
19802
host i1
This 1
limit
at has
flitting
Mains
the sol

”WEI

{16mm

The Clavibacter subspecies that infects tomato, C. michiganensis subsp.
michiganensis, represents an example of a phytopathogenic bacterial species that shows
unique abilities to survive and produce a wide range of symptoms on its host,
Lycopersz'cum esculentum, the commercial tomato. This subspecies of C. michiganensis
only known to infect tomato, although there have been a couple of reports of infection of
pepper plants (reviewed by Strider, 1969a). This pathogen is unique due to its ability to
survive in soil (Strider, 1967; Gaititis, 1989) and on seeds (Thyr, 1969; Dhanvantari,
1989a & 1989b; Gitaitis, 1989), as well as the production of an array of symptoms on its
host including leaf blight, vascular infection, wilt, cankers and fruit spots (Gitaitis, 1993).
This bacterium is capable of overwintering for several years in plant debris in the
northern Midwest and tomato producers are warned to stay out of contaminated fields for
at least 3 years (Stephens & Fulbright, 1986; Gitaitis, 1989). The relationship of C.
michiganensis subspecies michiganensis to other subspecies of C. michiganensis is
unknown and the evolutionary linkage of C. michiganensis subsp. sepedont'cus infecting
the solanaceous potato plant and C. michiganesis infecting tomato remains an intriguing
mystery.

Tomato is one of the most important vegetable crops grown worldwide due to the
demand for the product and high cost of production (Gould, 1992). Bacterial diseases of
tomato are feared because of the ability of the bacterial diseases to rapidly reduce yield or
make tomato fruits unmarketable (or unprocessable), and the producer’s inability to
manage them with pesticides (Bryan, 1930; Strider, 1969a; Ricker & Riedel, 1993).
Tomato is host to at least six common species of phytopathogenic bacteria including

Pseudomonas syringae pv. syringae, P. syringae pv. tomato, P. corrugata, X campestris

pr. I‘r
specil
misfit
since

produ

Bacte

patiog
Rapids
tsetse
due to I
Riedel.
Studies .
Waste
and the l
genetics
ilOmOger
Wilts h

M Dir.

““5ng I

105335 earl

pv. vesciatoria, B. solanacearum and C. michiganensis subsp. michiganensis. Of these
species and pathovars, X. campestris pv. vesicatoria and C. michiganensis subsp.
michiganensis represent the greatest threat to Midwest tomato producers and processors
since both species can cause devastating epidemics when not diagnosed early in the

production cycle (Phoronezny & Volin, 1983; Gleason et al., 1993).

Bacterial Canker of Tomato

Clavibacter michiganensis subsp. michiganensis was one of the first bacterial
pathogens discovered on tomato in the early part of this century. It was first called Grand
Rapids disease and later changed to bacterial canker. It is considered a devastating
disease of worldwide importance because of its sporadic and unpredictable nature and
due to the severity of the disease in certain years (Stephens & Fulbright, 1986; Ricker &
Riedel, 1993). Since its identification by Erwin F. Smith in 1909 (Smith, 1910), many
studies have been conducted on C. michiganensis subsp. michiganensis to help
understand various aspects of the disease, including the genetics of the pathogen, the host
and the environmental and horticultural conditions that promote epidemics. Of these, the
genetics of the pathogen has received the least attention due to its apparently
homogeneous nature and its cell wall composition; Gram positive phytopathological
species have received must less attention than Gram negative species due to classical
work performed on Escherichia coli.

As destructive as this pathogen can be, and as well adapted to living on and
causing disease on all plant parts of the plant, this pathogen does not cause economic

losses each year (Emmatty & John, 1973; Hausbeck et al., 1999). Perhaps, the reason

why
afier
Oflh
epid:
rises
latte
Basu.

lQQZ;

radii
inany

lreriex
(Staph;
1966LI
\Ieren
field. It

196°a;n

 

why more research has not been done on this serious disease is its sudden disappearance
after a two or three year epidemic (Gleason et al., 1993). The determination of the origin
of the primary inoculum has always been of interest in attempts to understand the
epidemiological aspect of bacterial canker (Dhanvantari, 1989b). In trying to predict
disease outbreaks, research has focused on seed dispersal, southern-state transplant
infection, and soil and farm equipment overwintering and contamination (Strider, 1967;
Basu, 1970; Dhanvantari, 1989a; Chang et al., 1991; Chang et al., 1992b; Gitatitis et al.,
1992; Gleason et al., 1993; Carlton et al., 1994).

The disease affects plants in a wide range of geographical areas where processing
and fresh market tomato varieties are grown (Gitaitis, 1991). After Smith’s initial report,
many countries confirmed the presence of bacterial canker in their greenhouses and fields
(reviewed by Strider, 1969). Many outbreaks have been reported in the United States
(Stephens & Fulbright, 1986; Gitatitis, 1991; Gleason et al., 1993) and Canada (Basu,
1966; Dhanvantari, 1989a; Speranzini, 1995) as well as in Europe (Lépez et al., 1987;
Vaerenbergh & Chauveau, 1987). Depending on when symptoms are first observed in a
field, it can reduce yields by 50-80% due to smaller, weakened and killed plants (Strider,
1969a; Chang et al., 1992c; Gleason et al., 1993). In addition, profit can be reduced due
to fruit spots that appear as a result of the infection of fruit (Bryan, 1930; Gleason et al.,

1993).

Symptoms
Bacterial canker of tomato can be associated with diverse symptoms (Appendix

A). The symptoms are defined as marginal necrosis, unilateral wilt, curling of leaves and

Stem

Althc
cotjrl:
your;
condi

WO‘JII

antic
subsp
BXOp-o
“Ill \I
this ba
plants
siuthei
51153;),
10“ en
death 0
NeCrosi
his sin
lGllallis
1939). A

Elitx'pen'1

stem canker caused by the systemic infection of C. michiganensis subsp. michiganensis.
Although bacterial canker symptoms are often described for mature tomato plants,
cotyledons and seedlings can Show white blister and water-soaked lesions along the
young petioles, stems, and leaves (Basu, 1966; Strider, 1969b). Under laboratory
conditions, Chang et al. (1992a) were able to observed wilt and canker development on
wounded two-week old seedlings.

In a fully-grown tomato plant, wilt is frequently caused by the lack of water and
nutrient transport to the leaves due to the presence of high numbers of C. michiganensis
subsp. michiganensis located in the vascular system. It is thought that the production of
exopolysaccharides by the bacterium plays a role in the wilting symptom (Goto, 1992).
Wilt was one of the first symptoms recognized by Smith (1910) during the description of
this bacterium. He was able to demonstrate that the organism causing wilt to the tomato
plants in Grand Rapids was not the same organism causing wilt to solanaceous crops in
southern states, recognized as Pseudomonas solanacearum. When C. michiganensis
subsp. michiganensis is the infectious agent, wilt development occurs slowly from the
lower to the upper leaves while plants infected with P. solanacearum suffer a sudden
death of the entire plant since the wilting of leaves occurs relatively fast (Goto, 1992).
Necrosis of the foliage can be a confusing symptom in bacterial canker infection due to
this symptoms’s resemblance to symptoms caused by other bacteria and fungal infections
(Gitaitis, 1991) as well as leaf injuries caused by fungicide applications (Thompson et al.,
1989). Although, “leaf firing” or marginal necrosis of the leaves may be confusing to the

inexperienced, it can be a distinct symptom for bacterial canker diagnosis (Basu, 1966).

alvayi

plt‘SCl‘Ii
strong

result i
been It
is hour
tonal-0r
*lie lifii

carrier 1

filth X CI

The mail

The brown sunken areas, called cankers, along the stems and petioles are not
always the most common symptom associated with infection by C. michiganensis subsp.
michiganensis. The name bacterial canker was used since bacterial wilt had already been
used for Pseudomonas solanacearum on tomato. Cankers result due primarily to the
presence of the pathogen in the vascular system. The pathogen is known to produce
strong enzymes and these enzymes are involved in tissue destruction. Cankers can often
result in areas where a wound has been produced after a branch, leaf, or flower bract has
been removed due to natural or accidental injuries (Gitaitis, 1991; Carlton et al., 1994). It
is thought that the severity and incidence of the disease is more intense in staked
tomatoes than in tomatoes lefi growing on the ground due to the injuries obtained during
the lifting and tying. Nevertheless, wounds are not required for the development of
canker (Bryan, 1930).

C. michiganensis subsp. michiganensis may also produce spotted fruit showing
symptoms that usually appear as small necrotic spots surrounded by white-chlorotic
halos, known as bird’s-eye lesions. During early development of the fruit lesions, bird’s-
eye spots appear as small superficial round white lesions on green fruits (Bryan, 1930).
If bird’s eye lesions are observed in the field, it is highly indicative that infection with C.
michiganensis subsp. michiganensis has occurred due to the unique appearance of this
fruit lesion. However, a few reports (Bryan, 1930; Weebb et al., 1967; Gould, 1992)
have suggested the resemblance of the white bird’s-eye lesions with early fruit infection
with X. campestris pv. vesicatoria but both lesions show differences in their development.
The main difference between these two types of fruit infection is observed as the disease

progresses. Bacterial canker bird’s-eye lesions remain small and superficial with

116erI
camp"

halo Cl

lnfi"cti

the ing
Hillier
mclutll
u'omdf
mm '
Ln \‘EL’lO
m ate
applicat:
l963; L;
“as a “I
(1962') a
on Iomai
Michigan
RI
lllllalllOI’li
phjtopath

l985; Blai

 

necrotic lesions forming in the center of the white spot while fruit spots caused by X.
campestris pv. vesicatoria develop into large deep necrotic lesions lacking the chlorotic-

halo characteristic of bird’s eye lesions (Bryan, 1930).

Infection avenues

The systemic invasion of the host by phytopathogenic bacteria may result from
the ingress of bacterial cells through artificial or natural Openings in the host tissues
(Appendix C). It has been reported that the entrance of xylem-inhabitant bacteria,
including C. michiganensis subsp. michiganensis, is solely dependent on the presence of
wounds on target the host tissue (Pine et al., 1955) since no association with insects were
known that could transfer bacterial cells directly to the host vascular system (Ark, 1944).
In various laboratories, successful infection with C. michiganensis subsp. michiganensis
was accomplished when tomato stems and leaves were mechanically damaged prior the
application of bacteria (Bryan, 1930; Grogan & Kendrick, 1953; Thyr, 1968; Kontaxis,
1962; Layne, 1967; Farley, 1971). The theory that C. michiganensis subsp. michiganensis
was a wound-dependent pathogen prevailed until the studies conducted by Kontaxis
(1962) and Layne (1967). Independently, both studies provided evidence that trichomes
on tomato leaves could serve as infection sites for C. michiganensis subsp.
michiganensis.

Reports have confirmed the importance of natural openings such as stomata,
hydathodes, and trichome-base holes on leaves and fruits for infection of
phytopathogenic bacteria (Bashan et al., 1981; Getz et al 1983; Erhrig & Griesbach,

1985; Blanke 1986; Fankle et al., 1993; Carlton et al., 1998).

Disease c
In
associater
productio
rampany
for field p
tiring llt’
Stalin ant
(Hambccl
origin of I]
F ir

hing screc
includes C
Dham'anta
”llt‘ltiganw
l'tais becai
Where the l
infected Sec
alosg 0f .‘lfi
William ‘
M the gro
mic-higanm

Disease cycle—Source of inoculum

In North America, the appearance of the bacterial canker disease is primarily
associated with the tomato production cycle (Appendix B). In general, the tomato
production in the Midwest begins when a grower obtains seed from a seed-producing
company or transplants from a greenhouse grower who had planted seeds. Transplants
for field production are grown in a greenhouse for approximately two months. Generally
during field production, bacterial canker symptoms are observed from the mid-growing
season and later, however, symptoms may be observed in the greenhouse on transplants
(Hausbeck et al., 1995b). At least three sources may be considered responsible for the
origin of the primary inoculum in bacterial canker epidemics (Appendix C).

First, infected seed could be a source of the bacterial canker pathogen even after
being screened by seed-certification programs for various bacterial pathogens, which
includes C. michiganensis subsp. michiganensis (Grogan & Kendrick, 1953;
Dhanvantari, 1989a; Chang et al., 1991; Parker et al., 1995). The acceptance C.
michiganensis subsp. michiganensis as a seed-home pathogen has been controversial for
years because few studies have been conducted to establish the location on the seed
where the pathogen survives (Bryan, 1930). Gitaitis et al. (1991) reported that one
infected seed in 10,000 is capable of establishing 74-124 disease foci/hectare, resulting in
a loss of yield and value. AS Van Vaerenbergh & Chauveau (1987) stated “the
production of tomato seed free from the pathogens is of great economic importance for
both the growers and the seed industry.” Although, C. michiganensis subsp.
michiganensis is accepted as a seed-home pathogen, seed-producing companies deny that

they are responsible for the observed epidemics.

10

pmilUCll
trough
nggestc‘
al. 1991
er al. 19
al. (1999‘
colony it
as the for
conductor

ll
tomato dc
found to c
Sinder. 19
recOgnizi:

Re;
Canker dis:
palliOgen l.
D).

Disseminal

Use
tomato deb

ydllsp. ”ll-ch 1‘

A second possible source of primary inoculum is the transplants used to establish
production fields. Transplants could become infected through contaminated seed or
through contaminated greenhouse equipment and structures. Various studies have
suggested that symptomless transplants could be the source of bacterial canker (Gitaitis et
al., 1991; Chang et al., 1992a; Gleason et al., 1993; Hausbeck et al., 1995a & 1995b; Bell
et al., 1996; Haubeck et al., 1999) introduction into the field environment. Hausbeck et
al. (1999) suggested that symptomless transplants can harbor high population (106-107
colony forming units/gram of tissue) Of C. michiganensis subsp. michiganensis, and serve
as the foci for infection and diseased plants in the field. Further studies must be
conducted to confirm this theory.

The third most probable way that bacterial canker epidemics are initiated is from
tomato debris left from the previous years infected plants, since the pathogen has be
found to overwinter in infested tomato debris for 2-3 years (Gorgan & Kendrick, 1953;
Strider, 1967; Basu, 1970; Gleason et al., 1991; Chang et al., 1992b). Many growers
recognize this as a potential source of infection, and rotate accordingly.

Regardless of the primary source of inoculum, various aspects of the bacterial
canker disease cycle concern tomato growers such as dissemination and spread of the
pathogen (Appendix C) and management strategies to control bacterial canker (Appendix

D).

Dissemination of C. michiganensis subsp. michiganensis

Use of certified seed and transplants fiom certified seed, as well as the removal of
tomato debris has not stopped the dissemination in the Midwest of C. michiganensis

subsp. michiganensis. Regardless of the source of primary inoculum seed, infected

11

mrspla
building
19921:).
plants in
ungaiim
11'3ch d1
populatit
the epiph
invasion
W: C a
C1
human to
313.“ alm :
al. 1992b
cultural pl
plum helg
[Implant
ll.
anOlller _\’t
such as ni;
W891 Clio

allFmElIlVe

transplant or tomato debris, the pathogen most likely spreads as an epiphyte, slowing
building its population until a threshold is reached (Gleason et al., 1991; Chang et al.,
1992b). C. michiganensis subsp. michiganensis can spread from diseased to healthy
plants in the same field as a result of natural events such as rainsplash or overhead
irrigation (Ark, 1944; Strider, 1967; Strider, 1969a; Tsiantos, 1986; Gleason et al., 1993).
Water droplets can carry a large number of bacterial cells that can survive as an epiphytic
population on the host tissue. As favorable environmental conditions emerge, the cells of
the epiphytic population can multiply, and if infection avenues are available in the host,
invasion of susceptible sites could result in systemic infection (Stephens & Fulbright,
1986; Carlton et al., 1998).

Cultural practices such as pruning, staking, fumigation, and or any other type of
human contact with diseased and healthy plants during any stage of tomato production
may also serve as a mechanism for spread (Strider, 1967; Chang et al., 1991; Chang et
al., 1992b; Carlton et al., 1994). Gitaitis et al. (1991) provided evidence that commercial
cultural practices such as clipping the top of systemically infected transplants to control
plant height could result in the dissemination of the pathogen throughout healthy
transplant fields.

The dissemination of C. michiganensis subsp. michiganensis from one year to
another year could also occur through the epiphytic population of various weed species
such as nightshade, horsenettle, and jimsonweed (reviewed by Strider, 1969a; Ricker,
1989; Chang et al., 1992b). It has been reported that these weed species serve as

alternative hosts of high population numbers of the pathogen.

12

llanagen
C l

he dissen
managem
mategies
seed and l
on grower
lh

subsp. not
of non-inf}
Miment 1
treatments
chemicals
indicate tl
umwb
RlClitl’. 19‘
lornato pu]:
Romans

pamotto.
lhe

0r1381 of an
Wplams .‘

mzchigatwn‘

Management Strategies

Clearly, numerous sources of primary inoculum and as well as opportunities for
the dissemination of the inoculum exists in the tomato production cycle, thus various
management strategies have been integrated to reduce the spread of the pathogen. These
strategies can be accomplished through certification programs such as pathogen-free
seed and pathogen-free transplants. Other practices as well can be instituted dependent
on grower cultural practices and sanitation.

The primary control strategy for seed-borne pathogens like C. michiganensis
subsp. michiganensis is based on the use of pathogen-free, certified seed. The production
of non-infected seeds is dependent on the use of fruit from uninfected plants, and
treatment of seed with disinfectants to eliminate any bacterial cells present. Seed
treatments have included soaking the seeds in disinfectants including bleach, alcohol, and
chemicals such as and hydrochloric or acetic acids which have been shown to efficiently
eradicate the bacteria from the seed coat and seed-hairs (Bryan, 1930; Ark, 1944;
reviewed by Strider, 1969a; Weebb et al., 1967; Dhanvantari, 1989; Gleason et al., 1995;
Ricker, 1995). Another seed treatment process frequently used is the fermentation of the
tomato pulp before seed extraction, although this is not as efficient as the chemical
treatments since it reduces the number of bacterial cells but does not eradicate the
pathogen.

The use of certified transplants and cuttings, is another method to prevent the
onset of an epidemic in a disease-free location (Strider, 1969a) primarily if symptomless
transplants and cuttings are sampled for the presence of C. michiganensis subsp.

michiganensis prior to distribution (Gitaitis et a1, 1991). Both of these strategies are

13

presen
l930:

like ni
potent
sugges
gr enh

techniq

tomato .
mom
can be a.
lGlIaitis
uncontm

Pt
applicatio
Control pr,
Cells Whirl
Mi fnn'
application

mkhlguflen,

based on the sensitivity of detection and identification protocols. Fortunately, highly
sensitive techniques are now available to detect low number of phytopathogenic bacteria.

Since the bacterial canker pathogen can persist in soil where tomato debris is
present, crop and field rotations of 2-5 years have been routinely recommended (Bryan,
1930; Strider, 1969a; Gleason et al., 1993). Also, reducing the potential of alternate hosts
like nightshade and jimsonweed from both the greenhouse and fields will decrease the
potential of weeds as the source of inoculum (Ricker, 1989; Chan et al., 1992b). Other
suggestions have been reported such as the use of fumigation or steam sterilization of
greenhouse soils and fall plowing of fields (Weebb et al., 1967). Each of these
techniques has questionable environmental aspects as well as added costs.

A thorough sanitation program consisting of cleaning equipment used during
tomato production such as transplant trays, pruning tools, and stakes has been strongly
recommended as a management strategy to reduce bacterial canker outbreaks. Sanitation
can be achieved with the use of disinfectants such as chloride, bromide, or formalin
(Gitaitis et al., 1992). Since the pathogen can be disseminated through splashing water,
uncontrolled and excessive periods of overhead irrigation should be minimized.

Perhaps the biggest breakthrough in management schemes is the preventive
application of copper-based bactericides to transplants while in the greenhouse. The
control provided by bactericides is accomplished by reducing the number of bacterial
cells which will delay the onset of bacterial canker, thus an increase in yield and high
quality fruits can be achieved (Hausbeck et al., 1997; Hausbeck et al., 1999). The
application of streptomycin provides an effective control to C. michiganensis subsp.

michiganensis, unfortunately, the application of compounds with this antibiotic agent is

14

prohibit:

al. 1995

subsp. m
and van:
1997). l
at the de
decrease
the genet
character
dex'elOpi:

ll
bullet“
agillls COL

lll'llli‘lll st

Character

The
“ldth inch
lllil‘dCllOn a
Same Organ

grow] on C

prohibited in the greenhouse and in the field in some regions Of the country (Hausbeck et
al., 1995a).

Another mechanism of control is the use of varieties resistant to C. michiganensis
subsp. michiganensis. Some companies have reported the production of breeding lines
and varieties with tolerance or moderate resistance to bacterial canker (Ricker et al.,
1997). The resistance obtained by some of the tomato lines is based on characters such
as the delay of onset of symptoms, low incidence of bacterial canker symptoms, and a
decrease in severity of specific bacterial canker symptoms. Since resistance is based on
the genetic composition of the pathogen and the host, studies involving the
characterization of various genetic aspects will provide knowledge to improve the
development of fully resistant tomato lines.

The use of biological agents to control bacterial canker has not been reported and
however, earlier descriptions and characterization of naturally avirulent strains and their
ability to produce bacteriocins still seems promising (Echandi, 1976). The biological
agents could provide control through various mechanisms such that the population of

virulent strains is reduced to a safe level, or completely inhibited.

Characterization and Identification

The identification of phytopathogenic bacteria is based on Koch’s postulates
which include the isolation of a pure culture of the presumptive pathogenic organism, the
induction and evaluation of symptoms caused by this organism, and the reisolation of the
same organism from infected tissue (Goto, 1992; Agrios, 1997). If the isolation and

growth on culture media of the organism is feasible, and if pathogenicity tests and the

15

expres
reisolz
C mic
tissue

and: :
amend
inhibit
sereeni

selectii

0n culiu

bdtlerla
511?th
ml)‘ rec
prei'enu
amKate
llllfle 5)
hither la
QUSC slg

lHaugbe‘

expression of symptoms can be easily obtained under laboratory conditions, the
reisolation of the pathogenic bacteria could be a challenge. For example, distinguishing
C. michiganensis subsp. michiganensis from saprophytic bacteria found on diseased
tissue can be problematic if only colony morphology is used due to the similarity in color
and texture of the saprophytes and the pathogen (Gitaitis & Beaver, 1990). The use of
amended culture media with nalidixic acid has been a successful approach because it
inhibits the growth of a significant amount of Gram-negative bacteria which enhances the
screening process (F atrni & Schaad, 1983). Fatrni & Schaad (1983) developed a semi-
selective culture media (SCM) which has facilitated the isolation and recovery of C.
michiganensis subsp. michiganensis from seed extracts even when low number of
bacterial cells are present. SCM has presented an advantage over general media because
C. michiganensis subsp. michiganensis colonies develop a specific coloration and texture
on culture plates.

The identification of C. michiganensis subsp. michiganensis and diagnosis of
bacterial canker can be a problem when small populations of bacteria are living
epiphytically and only invading the host from time to time (Dreier et al., 1995). The
early recognition and confirmation of the presence of the pathogen are critical in
preventing and slowing bacterial canker epidemics. Therefore, there is a need for fast,
accurate and reliable detection techniques especially during the tomato seedling stage
where symptoms are rarely present (Gleason et al., 1993). Symptomless seedlings may
harbor large population numbers of C. michiganensis subsp. michiganensis, which can
cause significant yield reduction once the plants have been transferred to the field

(Hausbeck et al., 1995b).

16

the doc
spt’t‘lllC
identific
Gudmes
based tee
Irgeted ‘
aubodie
fuse-p05
specific n
haciena Il
ll'ietzorel
l99~ll (Is I
monoclon
allalhOgei
Pelhogen 2
Al
With "Zero
ml. Slagc
bacteria] 9.
lmmllnolo;
"II-{ligand

ofELISA ‘

As the number of samples increase due to the sampling and screening techniques,
the development of detection techniques based on the production of antibodies against
specific pathogenic bacteria have been employed to enhance the rate and accuracy of
identification and enumeration (de Boer & Wieczorek, 1984; van Vuurde, 1987; Baer &
Gudmestad, 1993; Drennan et al., 1993). The success and efficiency of immunology-
based techniques are often firmly established on the use of highly specific antibodies
targeted to unique bacterial cell wall components. The production of monoclonal
antibodies has an advantage over polyclonal antibodies established by the reduction of
false-positive results (de Boer & Wieczorek, 1984). As of today, a large number of
specific monoclonal antibodies have been produced for the detection of phytopathogenic
bacteria that retain a zero tolerance status. For instance, various studies (de Boer &
Wieczorek, 1984; Gudmestad et al., 1991; de Boer & Gudmestad, 1993; de Boer et al.,
1994; de Boer & Hall, 1996) have successfully achieved the production of highly specific
monoclonal antibodies against the potato pathogen, C. michiganensis subsp. sepedonicus,
a pathogen typically given “zero tolerance” status where any plants observed with this
pathogen are destroyed.

Although C. michiganensis subsp. michiganensis is not considered a pathogen
with “zero tolerance,” antibodies have been produced to accurately detect the pathogen in
early stages of tomato production. Screening early could possibly prevent the onset of a
bacterial canker epidemic. Immunology-based techniques such as enzyme linked
immunology assay (ELISA), and immunoisolation are commonly used for detection of C.
michiganensis subsp. michiganensis. Stephens et a1. (1988) evaluated the detection level

of ELISA wells pre-coated with antiserum against C. michiganensis subsp.

17

miehigar
units Ell
pathogen
miehigun
mirhigun
cannot be
Ghabi &
C michig
Pathogen

ln
technique
lfilhogeni
lfiununoll.
”lithigam
lExitilal c
colonies 0
Itldill'el}.
NSlllVe it
mutinel}. l
[maul a
Chan‘fau

As

d““"ilOn'xI

michiganensis. They showed that the lower level of detection was 1x1 03 colony forming
units/ELISA well. Gitatits et a1. (1991) analyzed seedlings for the bacterial canker
pathogen using ELISA plates consisting of antiserum against C. michiganensis subsp.
michiganensis and they were able to successfully detect C. michiganensis subsp.
michiganensis in symptomless seedlings. However, from this assay bacterial cultures
cannot be obtained to confirm the presence of the pathogen or its pathogenic nature.
Gharbi & Nameth (1992) developed a monoclonal antibody specific for the detection of
C. michiganensis subsp. michiganensis. They were able to detect the bacterial canker
pathogen from symptomless tomato plants through ELISA and dot immuno-blots.

Irnmunoisolation protocols can overcome the limitations of ELISA because the
technique combines the use of specific antibodies with the culture of presumptive
pathogenic cells in solid culture media. F ranken et a1. (1993) evaluated the potential of
immunoflourescence (IF) and subsequent plating on semiselective media to detect C.
michiganensis subsp. michiganensis from tomato seeds. They accurately identify the
bacterial canker pathogen after pathogenicity assays were conducted for the bacterial
colonies Obtained from positive seed lots. The limitation of this technique is the
relatively long time necessary to conduct pathogenicity tests for the confirmation of false
positive results. Although the efficiency and specificity between detection methods is
routinely under evaluation, immunology-based techniques have the potential to detect
naturally avirulent C. michiganensis subsp. michiganensis isolates (van Vaerenburgh &
Chauveau, 1987).

As the number of samples to be tested increases, various techniques have been

developed to reduce the time necessary for characterization and detection of C.

18

mrhgt
analysis
shaman
potentiz
a deer
Gitaitis
subsp. n
for spec
host tiss
sensitive

Mlt‘lllgw

' iochern.
lhmmu
ragents
BiolOg G
[OSUCCes
negative
.limihom
militia
for he pa

michiganensis subsp. michiganensis. For example, fatty acid methyl ester (FAME)
analysis and microplates (e.g. Biolog) based on the cell wall constituents and biochemical
characteristics, respectively, have accelerated the detection process but have limited
potential for characterization at the subspecies level. FAME uses fatty-acid composition,
as determined by gas chromatography to help differentiate unknown isolates of bacteria.
Gitaitis & Beaver (1990) constructed a library of FAME profiles of C. michiganensis
subsp. michiganensis. They found that the ratio between specific fatty acids was distinct
for species that showed similar colony morphologies which were recovered from various
host tissues. In comparison to semiseletive media and ELISA, FAME analysis was less
sensitive and more difficult to use for detecting latent, symptomless infections of C.
michiganensis subsp. michiganensis in tomato seedlings (Gitaitis et al., 1991).

Biolog is an identification system based on a database containing the results of
biochemical reactions obtained for each species tested with the system. The
identification process is dependent on microplates where its wells are coated with
reagents that will simultaneously perform a number of microbiological tests. Using
Biolog GN microplates designed for Gram-negative species, Jones et a1. (1993) were able
to successfully identify strains of the Gram-positive genus Clavibacter from Gram-
negative phytopathogenic genera, A grobacterium, Erwinia, Pseudomonas, and
Xanthomonas. The introduction of plates designed specifically for Gram-positive
bacteria, including C. michiganensis subspecies has improved the accuracy of this assay
for the pathogenic coryneform species, although certain carbon sources must be included

in the system to improve the identification at the subspecies level (Harris-Baldwin &

19

Gudme

chma

chin l‘
more re

technlql

primers
plij~10p3
um} lUI‘U
al- 1988
mirhegw

andCXJ

DNA of

later. ll
michigur
they Obit
l'sing ll]:
ai'uulent
tommerc
Michigan

Specified

Gudmestad, 1996). Clearly even more sensitive and specific assays for identification and
characterization are needed.

Since the acceptance of molecular biology, DNA hybridization and polymerase
chain reaction (PCR) techniques have revolutionized cell detection techniques making it
more reliable, specific and less-time consuming (Vandamme et al., 1996). PCR based
techniques are ten times more sensitive than microplate identification and immunology-
based procedures (Hu et al., 1995). Many reports on the use of DNA probes and PCR
primers have demonstrated the efficiency and success in the detection of important
phytopathogenic species such as Erwinia caratovora (de Boer & Ward, 1995), E.
amylovora (McManus & Jones, 1995) C. michiganensis subsp. sepedonicus (Verrault et
al., 1988; Johansen et al., 1989; Drennan et al., 1993; Rademaker & Janse, 1994), C.
michiganensis subsp. michiganensis (Thompson et al., 1989; Rademaker & Janse, 1994),
and C. xyli subsp. xyli (Pan et al., 1998).

For example, Verrault et a1. (1988) were able to detect up to 1 nanogram of the
DNA of C. michiganensis subsp. sepedonicus after the development of specific probes.
Later, Thompson et a1. (1989) constructed a chromosomal library from stains of C.
michiganensis subsp. michiganensis. After the digestion of clones with endonucleases,
they obtained a 5 kilobase pair (kbp) fragment unique to this pathogenic subspecies.
Using the 5 kb fragment as a probe, they were able to successfully detect virulent and
avirulent strains of the bacterial canker pathogen. Rademarker & Janse (1994) used
commercially available probes, MIC 1 and Diagen, to detect C. michiganensis subsp.
michiganensis and C. michiganensis subsp. sepedonicus, respectively. To confirm the

specificity of the two probes, they derived PCR primers to amplify chromosomal

20

{1'3ng
endonu
gecifir
deieeio
pnmers
mi elil g6
deteeiet

saproph

sequent:
Il'lirza e
ltl‘é‘l of s
iden ificz
technique
lhllm ct
Colombg

T
imitated
“ll? able
”51' deg.
Me 1 Ce 1.

{“0 Plasn

mill-£072.

fragment of the two C. michiganensis subspecies. After treating the PCR products with
endonuclesases, the restriction analysis of the amplified fragments supported the
specificity of the probes developed since other C. michiganensis subspecies were not
detected with this method. Recently, Sousa-Santos et al. (1997) developed two PCR
primers, CM3 and CM4, that confirmed the presence of C. michiganensis subsp.
michiganensis isolated from tomato seed extracts. A minimum of 40 cells/ ml were
detected with CM3 and CM4 probes, and no positive results were observed with the
saprophytic bacteria tested.

The use of other than entire chromosomal sequences such as intergenic spacer
sequences between ribosomal genes, and specific sequences on the 16S ribosomal gene
(Mirza et al., 1993; Mills et al., 1997; Daffonchio et al., 1998) have provided a higher
level of specificity to the identification process. Besides the advantages offered by new
identification protocols for identifying subspecies of C. michiganensis, some of these
techniques have increased the amount of information to conduct epidemiological studies
(Mirza et al., 1993; Rademaker & Janse; 1994; Rivera et al., 1995; Liu et al., 1995;
Colombo et al., 1997) and formulate novel management strategies (Pan et al., 1998).

The use of DNA-hybridization in combination with PCR protocols has also
increased our knowledge of pathogenicity characters. For instance, Dreier et al. (1995)
were able to distinguish pathogenicity genes on C. michiganensis subsp. michiganensis.
They described a pathogenicity gene (patl) with unknown function and an endocellulase
gene (celA). Both genes are found in C. michiganensis subsp. michiganensis isolates on
two plasmids, pCMl and pCM2, respectively. After digestion of total DNA from C.

michiganensis subspecies and subsequent treatment with endonucleases, the DNA was

21

subsp. l
ahybn'c
Chill-f
bacteria
They we
They ob
min. 1
Plant ext
obsen'ed
alnoxm
contain a
deltttion
pml’ldt‘ Si
detection
II
“351' that
illemam.e
[0 general,-

Ofgellt am

probed with celA and an RFLP pattern was obtained for C. michiganensis subsp.
michiganensis. A 3.2 kbp fragment was obtained specific to C. michiganensis subsp.
michiganensis since the fragment was lacking in other C. michiganensis subspecies
tested. The probe patlwas shown to be specific for virulent strains of C. michiganensis
subsp. michiganensis as none of the laboratory-generated avirulent strains tested showed
a hybridization signal for this fragment. They also developed a set of PCR primers,
CMM-S and CMM-6, derived from the pat] gene to test their ability to identify virulent
bacteria from infected plant extracts and seeds harvested from infected tomato plants.
They were able to detect virulent isolates from 50 seeds containing 1,000 bacterial cells.
They obtained negative results from healthy plants and plants infected with an avirulent
strain. The detection limit of the suggested PCR protocol was of 200 bacterial cells/ml of
plant extract without prior isolation and enrichment of the pathogen. The only limitation
observed in this technique was the use of plants infected artificially in the laboratory with
a known concentration of the pathogen and not from naturally infected plants that may
contain a higher population of saprophytes that serve as PCR inhibitors decreasing the
detection limit. Once more, this study proved the potential of PCR-based assays to
provide speed, sensitivity, and specificity that will reinforce the results from standard
detection methods.

The development of DNA probes require previous knowledge of DNA sequence,
a task that can be time-consuming as well as less attainable due to its high cost. An
alternative approach to avoid the extra time and costs is the use of universal PCR primers
to generate DNA fragments unique to bacterial species. DNA-fingerprinting, as this type

of gene amplification protocol is generally termed, has been intensively applied in studies

22

dealing ‘
Rodl'lgu
Appuhzu
l998;le
DNA-fir
microorg
al.. 1995
been an t
Smith et
ll'eingar

Rademal

dealing with topics such as genomic diversity and variability (Rivera et al., 1995;
Rodriguez-Barradas et at., 1995; Balerias Couto et al., 1996; van Belkurn et al., 1996;
Appuhamy et al., 1997; Murry et al., 1997; Sechi et al., 1998; Vasquez-Arroyo et al.,
1998; Jersek et al., 1999), and taxonomy (Balkvvill et al., 1997; Vinuesa et al., 1998).
DNA-fingerprinting techniques have been applied to a broad number of genera of
microorganisms from fungi to bacteria (Weising et al., 1991; de Bruijn, 1992; Weising et
al., 1995; Louws et al., 1996; Thanos et al., 1996), and plant pathogenic bacteria have not
been an exception (McManus & Jones, 1995, Louws et al., 1994; Louws et al., 1995;
Smith et al., 1995; Opgenorth et al., 1996; Pooler et al., 1996; Vera Cruz et al., 1996;
Weingart & VOlksch, 1997; Louws et al., 1998; Rademaker & de Bruijn, 1997;

Rademaker et al., 1997; Jaunet & Wang, 1999; Rademaker et al., 1999a).

DNA fingerprinting

The repetitive sequence-based PCR (rep-PCR) is one of the two DNA
fingerprinting techniques used to define genetic characters and polymorphic regions of
prokaryotic and eukaryotic genomes (Louws et al., 1996; Schneider & de Bruijn, 1996).
The fingerprints generated from rep-PCR technology originated from naturally occurring
repetitive sequences that are randomly arranged in high copy number on the circular
bacterial genomes (de Bruijn, 1992; Schneider & de Bruijn, 1996; Vera Cruz etal., 1996;
Versalovic et al., 1997). Three groups of repetitive sequences have been studied in detail
and have been used to generate primers for the amplification of DNA through PCR in
various bacterial species (Versalovic et al., 1991). The three primers used for the

production of rep-PCR genomic fingerprints consist of 18-22 base pair oligonucleotide

23

sequener
include I
repetitiyi
1996; Sc
deyelope
eoli and .
combinai
positive 1

used in c1

Which C01
thousand
1995). n
c“'llllllemr
repetitiye
important

degradarjc

sequences recognized via DNA hybridization studies (Versalovic et al., 1994). They
include the repetitive extragenic palindromic (REP) sequence, the enterobacterial
repetitive intergenic consensus (ERIC) sequence, and the BOX element (Louws et al.,
1996; Schneider & de Bruijn, 1996; Versalovic et al., 1997). These primers were
developed after their characterization in the Gram-negative enteric bacteria, Escherichia
coli and Salmonella typhimurium for the REP and ERIC primers, respectively. The
combination of boxA, boxB, and boxC elements was first characterized in the Gram-
positive bacterium, Streptococcus pneumonia and an unrelated Gram-negative species
used in characterizing the BOX primer (Versalovic et al., 1994; Koueth et al., 1995).

A different group of primers can be generated fi'om the minisatellite sequences,
which consist of 1060 base pairs arranged in a head to tail fashion with two to several
thousand motifs, most likely containing non-coding DNA (revised by Weising et al.,
1995). The polytrinucleotides, (GTG)5 and (GCC)5 are examples of primers
complementary to minisatellites (V ersalovic et al., 1994). The precise function of the
repetitive and minisatellite sequences is unknown, but it has been postulated to be
important in chromosomal organization, DNA replication and prevention of DNA
degradation, as well as in regulation of gene expression (de Bruijn, 1992; Weising et al.,
1995). Meyer et al. (1993) demonstrated the usefulness of (GTG)5 as rep-PCR primers
for the differentiation of eukaryotic species. Some of the first reports where (GTG)5 was
used to primer rep-PCR were realized with human pathogenic species such as the fungus
Cryptococcus neoformans (Meyer et al., 1993; Meyer & Mitchell, 1995). Various
scientists showed that minisatellite sequences are randomly spread through the entire

fungal and prokaryotic genomes (Meyer et al., 1993; Meyer & Mitchell, 1995; Weising et

24

(1'53

al., 1995; Warren et al., 1996). These studies were important in generating information
related to origin and geographic distribution of various strains; data necessary for
epidemiological studies. A recent report has demonstrated the usefulness of (GTG)5-PCR
genomic fingerprinting for assaying genetic variability in various fungal species (Balerias
Couto et al., 1996).

As in medical pathology, PCR-based techniques have revolutionized the detection
and differentiation of pathogens in plant pathology (Roberts et al., 1996; Sadowsky et al.,
1996; Lee et al., 1997a & 1997b; Rademaker & de Bruijn, 1997; Sousa Santos et al.,
1997; Rademarker et al., 1999a & 19%). Specifically, rep-PCR genomic fingerprinting
has provided useful information for the reevaluation of taxonomic names and
evolutionary relationships within phytopathogenic bacterial species. For example, Louws
et al. (1994 & 1995) were able to redefine phylogenetic groups and effectively
differentiate phytopathogenic strains within the pseudomonad and xanthomonad species
that can cause symptoms on a wide variety of hosts. More recently, Louws et al. (1998)
were able to rapidly and effectively differentiate five Clavibacter michiganensis
subspecies using rep-PCR. The C. michiganensis subspecies analyzed are among
characterized by their virulence on different hosts. They also reported a higher level of
genetic resolution at the subspecies level when a wide collection of strains of C.
michiganensis subsp. michiganensis was analyzed with BOX-PCR. They defined four
BOX-PCR (A, B, C, and D) types based on polymorphic DNA bands after agarose-based
electrophoresis. During the completion of Louws et al. (1998) manuscript a new BOX-
PCR genotype was described as type E (Bell et al., 1997). Although the biological

significance of the BOX-PCR types was not established, genetically typing the strains

25

Taron

011 11101
batten.
limited
named
Butteri
hxnh
Ware
dirneie
liming
Biochen
islatin C
Withigw
Will the
rimming
Mic}; '

' 1.9m.

(00718 5‘

“remained

provided useful groupings. For example, almost all of the C. michiganensis subsp.
michiganensis strains that are asymptomatic on tomato (avirulent strains) belong to BOX-

PCR type A, while highly virulent strains belong to BOX-PCR type C.

Taxonomy

The assignment of genus and species to bacterial isolates has been primarily based
on morphological, physiological and biochemical characteristics. The taxonomy of the
bacterial canker pathogen has been controversial since its discovery (Smith, 1910) due to
limited physiological and biochemical tests. The bacterial canker pathogen was first
named Bacterium michiganense due to the similarities in colony morphology with
Bacterium campestre and B. solanacearum. Bacterium michiganense colonies were
described as pale yellow with smooth round surface and edges, and with a wet shinny
appearance. Bacterial cells were described as short rods (0.35-0.4 x 0.8-1.0 microns in
diameter). This pathogen was recognized as a Gram-positive bacterium, non-spore
forming, and slow growing on solid media even at optimum temperatures (25°C)
Biochemical tests such as reduction of nitrogen source (nitrate and nitrite), and starch and
gelatin degradation were also analyzed. Later, B. michiganense was named Aplanobacter
michiganense due to the absence of flagella (non-motile) and morphological similarities
with the potato pathogen, Aplanobacter rathayi. Thereafter, additional attempts at
renaming the tomato pathogen included Pseudomonas michiganensis, Phytomonas
michiganensis, Erwinia michiganensis, Mycobacterium michiganense, and
Corynebacterium michiganense_(as indicated by Strider, 1969a). The latter nomenclature

remained until recent molecular-based analyses were included as taxonomic criteria.

26

coryu
physii
anima
1973t
‘ell di

COIllfl

specks \
Eineiica]

0“; Cons

Based on classical taxonomy, the genus Corynebacterium consisted of
coryneform bacteria regardless of their host specificity. Morphological, biochemical, or
physiological characters used to discriminate between plant pathogenic and
animal/human pathogenic bacteria were lacking until Yomada & Kogarnata (1972a &
1972b) conducted a numerical analysis. In the numerical analysis, they used the type of
cell division, bacterial cell wall composition, and guanine plus cytosine (G+C) DNA
content as the primary characters for the differentiation of coryneform bacteria. They
analyzed 112 strains fi‘om the designated genera Corynebacterium, Microbacterium,
Cellulomonas, Arthrobacter, and Brevibacterium. They were able to differentiate the
strains into seven groups where the main distinction was the presence or absence of
diaminopirnelic acid (DAP). As a result, strains without DAP were assigned to a single
group. This single group consisted of two plant pathogenic species, C. michiganense and
C. insidiosum. These species contained diaminobutyric acid (DAB) as a principal cell
wall amino acid, a bending type cell division, and a GC DNA content range from 69-
78%. As a result plant pathogenic species were recognized to be different from the
animal/human pathogenic species.

After recognizing the differences between plant and animal pathogenic
coryneform bacteria, the next challenge to be accomplished was the differentiation
among phytopathogenic species. Starr et al. (1975) used DNA-DNA homology and GC
content to differentiate seven Corynebacterium species. Since the pathology of the
species was highly specific, they suspected that DNA will provide useful information to
genetically group the seven species. Indeed, they were able to established two groups;

one, consisting of Corynebacterium michiganense, C. insidiosum, and C. sepedonicum

27

gums
obsery
huge
eonfiir
byme
that
degree
nehrns.
unmn

that the

differei
Pilyaei
Species
Pfilhijgg‘
noon
subspeg
"2&1;ng
all-Cl SUlT!
bi{theta
PAGE1
mhmgh

strains, and the second of C. poinsettiae, C. betae, C. flaccumfaciens, and C. ilicis. They
observed a low level of DNA-homology (15%) between the species in each group, and
low genetic variability between species in the same group. Later, defer et a1 (1982)
confirmed the homology between C. michiganense, C. insidiosum, and C. sepedonicum
by the determination of the type of peptidoglycan layer and DNA homology. Their
results indicated a close relatedness between C. michiganensis and C. insidiosum, and a
degree of similarity between C. sepedonicum, C.michiganense and the corn pathogen C.
nebraskense. However, they suggested that the presence of type B27 peptidoglycan layer
is strong evidence to consider the strains as a single species. Therefore, they suggested
that these strains should not be considered subspecies.

As the controversy continued, techniques other than DNA homology were used to
differentiate closely related Corynebacterium species. Carlson & Vidaver (1982) used
polyacrylamide gel electrophoresis (PAGE) of cellular proteins of 13 Corynebaterium
species including C. michiganense, C. insidiosum, C. nebraskensis, and a new wheat
pathogen named as C. tessellarius. They concluded that small differences in PAGE-
protein profile dissimilarity observed between the species showed these strains to be
subspecies of C. michiganense. Thus, they proposed the recognition of the strains as C.
michiganense subspecies michiganense, subspecies insidiosum, subspecies nebraskense,
and subspecies tessellarius. Later, Davis et al. (1984) combined morphological,
biochemical, cell wall amino acid and sugar characteristics, GC-DNA content, and
PAGE-protein profile results to corroborate subspecies subscription on strains of C.
michiganense. Recently before Davis et al (1984) study concluded, a xylem inhabitant

fastidious organism with similar morphological, biochemical and physiological

28

chmt‘lr
from so
C {I'll C
Conrnu
reelessii
genus C

subspee

suggesre
iyllennin
example
stains re
Rhodocc
Oldie F .-‘
classificz
classifier
similarity
Oiihese !

in -
ntmgur;

5TUdies 51
Much he

'w'

Mora,

characteristics was isolated and described (Davis et al., 1980). The bacterium isolated
from stunted sugarcane plants was included in the genus Corynebacterium, and named as
C. xyli due to the morphological, biochemical, and physiological similarities with the
Corynebacterium phytopathogenic bacteria. Davis et al. (1984) proposed the
reclassification of all the Corynebacterium michiganense subspecies to be included in the
genus Clavibacter, and C. xyli and bennudagrass pathogenic coryneform bacteria as a
subspecies of Clavibacter xyli.

Although the genus Clavibacter has been redefined, various studies have
suggested the use of the subspecies level within the species C. michiganensis
(Henningson & Gumestad, 1991; de Bryne et al., 1992; Kampfer et al., 1993). For
example, Hennington & Gumestad (1991) subdivided phytopathogenic coryneform
strains representing the genera Arthrobacter, Clavibacter, Curtobacterium, and
Rhodococcus based on the analysis of fatty acids by gas chromatography. The analysis
of the FAME profiles suggested the division of four genera without the taxonomic
classification at the subspecies level. Kampfer et al. (1993) conducted a numerical
classification study based on 280 physiological characters. They observed a high
similarity between the genera Clavibacter and C urtobacterium, and a close relationship
of these genera to Microbacterium. Individually, these studies suggested that C.
michiganensis should be returned to a full species status.

As of today, Clavibacter michiganensis subspecies status prevails since the
studies suggesting the full species status lacked the analysis of more than one data source
which helps on the resolution of the small differences between the subspecies. The

pathovar designation for Clavibacter michiganensis has been avoided since genetic

29

dih‘erer
michigi
batten:
are 0er
in rheir
brunchc
Eridenl

padrooe

Objecti

Florida

differences have been reported besides their host specificity. Therefore, Clavibacter
michiganensis subspecies has been circumscribed to plant pathogenic xylem-inhabitant
bacteria with the following distinctive characters, pleomorphic gram-positive rods which
are often arranged in V or L formation, obligately aerobic, DAP as the major component
in their cell wall, MK-9 as the main respiratory molecule, with anteiso- and iso-methyl
branched fatty acids, and about 70 % CG DNA content (Collins & Bradbury, 1991).
Evidently, Clavibacter is still the most clearly taxonomically defined genus of the plant

pathogenic bacteria.

Objectives

In spite of the fact that many studies have provided answers to the differentiation
and characterization of the bacterial canker pathogen, C. michiganensis subsp.
michiganensis, many questions regarding the epidemiological basis of the pathogen still
exists. Defining genetic variability should help us understand more about the role of the
strains in the host/pathogen relation and may provide us with powerful tools in examining
the epidemiological picture of this disease. To this end, I have focused on two
objectives:

1) To determine if another rep-PCR universal primer, (GTG)5, supports the
subdivision of C. michiganensis subsp. michiganensis into the five genotypes established
with BOX-PCR; and, if so,

2) Determine if the measurable diversity allows for a better understanding of the
disease and the origin of the strains within the epidemic.

Although the application of copper-base chemicals in the greenhouse have

provided management strategies to reduce yield loss caused by bacterial canker pathogen

30

infect
fruit 5
money
under
dereli

ndeOI

and.

hunsr

infection, these control strategies do not appear to provide control for the occurrence of
fruit spotting in the field. It is reasonable to suggest that in order to formulate novel
management strategies to control bacterial canker at the fruit spot level, a better
understanding of C. michiganensis subsp. michiganensis infection and bird's-eye lesions
development is required. Therefore, two objectives were set to help me determine the
role of the pathogen in causing fruit spots:
1) To determine if bird’s eye lesions are the result of external flower infection;

and,

2) If it is, determine the flower stage which provides the maximum number of

fruit spots.

31

REFERENCES

32

Basu. P
C.

REFERENCES

Agrios, G. N. 1997. Plant diseases caused by prokaryotes: bacteria and mollicutes.
Pages 407-470 in: Plant Pathology. 4th edition. Academic Press, San Diego, CA.

Appuhamy, S., Parton, R., Coote, J. G., and Gibbs, H. A. 1997. Genomic fingerprinting
of Haemophilus somnus by a combination of PCR methods. Journal of Clinical
Microbiology 35:288-289.

Ark, P. A. 1944. Studies on bacterial canker of tomato. Phytopathology 34: 394-400.

Atlas, R. M., and Bartha, R. 1998. Interactions between microorganisms and plants.

Pages 116-140 in: Microbial Ecology. 4th edition. Benjamin/Cummings Science
Publishing Inc, CA.

Baer, D., and Gudmestad, N. C. 1993. Serological detection of nonmucoid strains of

Clavibacter michiganensis subsp. sepedonicus in potato. Phytopathology 83: 157-
163.

Balerias Couto, M. M., Eijsma, B., Hofstra, H., Huis in’t Veld, J. H. J., and van der
Vossen, J. M. B. M. 1996. Evaluation of molecular typing techniques to assign
genetic diversity among Saccharomyces cerevisiae strains. Applied and
Environmental Microbiology 62:41-46.

Balkwill, D. L., Drake, G. R., Reeves, R. H., Fredrickson, J. K., White, D. C.,
Ringelberg, D. B., Chandler, D. P., Romine, M. F., Kennedy, D. W., and Spadoni,
C. M. 1997. Taxonomic study of aromatic-degrading bacteria from deep-
terrestrial-subsurface sediments and description of Sphingomonas aromaticivorans
sp nov, Sphingomonas subterranea Sp nov, and Sphingomonas stygia Sp nov.
International Journal of Systematic Bacteriology 47: 191 -201.

Bashan, Y., Sharon, E., Okon, Y., and Henis, Y. 1981. Scanning electron and light
microscopy of infection and symptom development in tomato leaves infected with
Pseudomonas tomato. Physiological Plant Pathology 19: 139-144.

Basu, P. K. 1966. Conditions for symptomatological differentiation of bacterial canker,
spot, and speck on tomato seedlings. Canadian Journal of Plant Science 46:525-
530.

Basu, P. K. 1970. Temperature, an important factor determining survival of
Corynebacterium michiganense in soil. Phytopathology 60:825-827.

33

Bell, J., Medina, C., Fulbright, D., and Hausbeck, M. 1996. Epidemiology of
Clavibacter michiganensis subsp. michiganensis on tomato transplants in the
greenhouse and the field. Proceedings, Twelfth Annual Tomato Disease Workshop,
Ohio, Dec. 5-6.

Bell, J ., Medina-Mora, C., Hausbeck, M., and Fulbright, D. 1997. Prevalence of
different PCR fingerprint groups and virulence class of Clavibacter michiganensis
subsp. michiganensis in four intensively managed tomato fields-preliminary report.

Proceedings, Thirteenth Annual Tomato Disease Workshop. Indianapolis, Indiana
Dec. 4-5.

Blanke, M. M. 1986. Comparative SEM study of stomata on developing quince, apple,
grape, and tomato fi'uit. Angew. Botanik 60:209-214.

Bryan, M. K. 1930. Studies on Bacterial Canker of Tomato. Journal of Agricultural
Research 41:825-850.

Carlson, R. R., and Vidaver, A. K. 1982. Taxonomy of Corynebacterium plant
pathogens, including a new pathogen of wheat, based on polyacrylamide gel
electrophoresis of cellular proteins. International Journal of Systematic
Bacteriology 32:3 15-326.

Carlton, W. M., Gleason, M. L., and Braun, E. J. 1994. Effect of pruning on tomato
plants supporting epiphytic populations of Clavibacter michiganensis subsp.
michiganensis. Plant Disease 78:742-745.

Carlton, W. M., Braun, E. J ., and Gleason, M. L. 1998. Ingress of Clavibacter
michiganensis subsp. michiganensis into tomato leaves through hydathodes.
Phytopathology 88:525-529.

Chang, R. J., Ries, S. M., and Pataky, J. K. 1992a. Effects of temperature, plant, age,
inoculum concentration, and cultivar on the incubation period and severity of
Bacterial Canker of tomato. Plant Disease 76:1 150-1 155.

Chang, R. J ., Ries, S. M., and Pataky, J. K. 1992b. Local sources of Clavibacter

michiganensis ssp. michiganensis in the development of bacterial canker on
tomatoes. Phytopathology 82:553-560.

Chang, R. J ., Ries, S. M., and Pataky, J. K. 1992c. Reductions in yield of processing
tomatoes and incidence of bacterial canker. Plant Disease 76:805-809.

Chang, R. J ., Ries, S. M., and Pataky, J. K. 1991. Dissemination of Clavibacter

michiganensis subsp. michiganensis by practices used to produce tomato
transplants. Phytopathology 8 1 : 1 276-1 28 l .

34

Collin
P

Colom
IN
8:
t}

Daffon

Collins, M. D., and Bradbury, J. F. 1991. The genus Clavibacter. Pages 1358—1363 in:
Prokaryotes. USA.

Colombo, M. M., Mastrandrea, S., Leite, F., Santona, A., Uzzau, S., Rappelli, P., Pisano,
M., Rubino, S., and Cappuccinelli, P. 1997 . Tracking of clinical and
environmental Vibrio cholerae 01 strains by combined analysis of the presence of
the toxin cassette, plasmid content and ERIC PCR. FEMS 19:33-45.

Daffonchio, D., Borin, S., Frova, G., Manachini, P. L., and Sorlini, C. 1998. PCR
fingerprinting of whole genomes: the spacers between the 16S and 23S rRNA genes
and of intergenic tRNA gene regions reveal a different intraspecific genomic
variability of Bacillus cereus and Bacillus licheniformis. International Journal of
Systematic Bacteriology 48: 107-1 16.

Davis, M. J., Gillaspie, A. G. Jr., Harris, R. W., and Lawson, R. H. 1980. Ratoon

stunting disease of sugarcane: Isolation of the causal bacterium. Science 21021365-
1367.

Davis, M. J ., Gillaspie, A. G. Jr., Vidaver, A. K., and Harris, R. W. 1984. Clavibacter: a
new genus containing some phytopathogenic coryneform bacteria, including
Clavibacter xyli subsp. xyli sp. nov., subsp. nov. and Clavibacter xyli subsp.
cynodontis subsp. nov., pathogens that cause ratoon stunting disease of sugarcane
and bermudagrass stunting disease. International Journal of Systematic
Bacteriology 34:107-1 17.

de Boer, S. H., and Hall, J. W. 1996. The probability of detecting Clavibacter
michiganensis subsp. sepedonicus by indexing seed potato lots with serological
tests. Journal of Phytopathology 144:459-463.

de Boer, S. H., and Ward, L. J. 1995. PCR detection of Erwinia carotovora subsp.
atroseptica associated with potato tissue. Phytopathology 85:854-858.

de Boer, S. H., Stead, D. E., Alivizatos, A. S., Janse, J. D., van Vaerenbergh, J ., Dehaan,
T. L., and Mawhinney, J. 1994. Evaluation of serological tests for detection of

Clavibacter michiganensis subsp. sepedonicus in composite potato stem and tuber
samples. Plant Disease 78:725-729.

de Boer, S. H., and Wieczorek, A. 1984. Production of monoclonal antibodies to
Corynebacterium sepedonicum. Phytopathology 74: 143 1-1434.

de Bruijn, F. J. 1992. Use of repetitive (repetitive extragenic palindromic and
enterobacterial repetitive intergeneric consensus) sequences and the polymerase
chain reaction to fingerprint the genomes of Rhizobium meliloti isolates and other
soil bacteria. Applied and Environmental Microbiology 58:2180-2187.

35

— “'31. ' (if? ..

 

de Bryne, E., Swings, J ., and Kersters, K. 1992. Enzymatic relatedness amongst
phytopathogenic coryneform bacteria and its potential use for their identification.
Systematic and Applied Microbiology 15:393 -401.

Dhanvantari, B.N. 1989a. Effect of seed extraction methods and seed treatments on
control of tomato bacterial canker. Canadian Journal of Plant Pathology 1 1:400-
408.

Dhanvantari, B. N. 1989b. Seed treatment for eradication of tomato bacterial canker.
Proceedings, Fifih Annual Tomato Disease Workshop, Michigan, Dec. 7-8.

Dopfer, H., Stackebrandt, E., and Fiedler, F. 1982. Nucleic acid hybridization studies on

Microbacterium, C urtobacterium, Agromyces and related taxa. Journal of General
Microbiology 128: 1697-1708.

Dreier, J ., Bermpohl, A., and Eichenlaub, R. 1995. Southern hybridization and PCR for
specific detection of phytopathogenic Clavibacter michiganensis subsp.
michiganensis. Phytopathology 85:462-468.

Drennan, J. L., Westra, A. A. G., Slack, S. A., Delserone, L. M., Collmer, A.,
Gudmestad, N. C., and Oleson, A. E. 1993. Comparison of a DNA hybridization
probe to ELISA for the detection of Clavibacter michiganensis subsp. sepedonicus
is field-grown potatoes. Plant Disease 77: 143-1247.

Echandi, E. 1976. Bacteriocin produced by Corynebacterium michiganense.
Phytopathology 66:430-432.

Ehrig, F., and Griesbach, E.. 1985. Development of bird eye spots on tomato fruits-

Investigations by scanning electron microscopy. Arch. Phytopathol. Pfalnzenschutz
21:419-423.

Emmatty, D. A., and John, C. A. 1973. Comparison of yield loss to bacterial canker in
resistant and a susceptible variety. Plant Disease Reporter 57:787-788.

Farley, J. D. 1971. Recovery of Corynebacterium michiganense from overwintered stem
by excised-petiole inoculation method. Plant Disease Reporter 55:654-656.

Fatmi, M., and Schaad, N. W. 1988. Semiselective agar medium of isolation of
Clavibacter michiganensis subsp. michiganensis from tomato seed.
Phytopathology 78: 121-126.

F ranken, A. A. J. M., Kamminga, G. C., Snijders, W., Vanderzouwen, P. S., and
Bimbaum, Y.E. 1993. Detection of Clavibacter michiganensis ssp. michiganensis

in tomato seeds by immunofluorescence microscopy and dilution plating.
Netherlands Journal of Plant Pathology 99: 125-137.

36

 

 

Prank

Get;
1

I

Chub
l

l

Git-am

Gitaitii

Frankle, W. G., Hopkins, D. L., and Stall, R. E. 1993. Ingress of the watermelon fruit
blotch bacterium into fruit. Plant Disease 77 :1090-1092.

Getz, S., Fulbright, D. W., and Stephens, C. T. 1983a. Scanning Electron Microscopy of
infection sites and lesion development on tomato fruit infected with Pseudomonas
syringae pv. tomato. Phytopathology 73:39-43.

Getz, S., Stephens, C. T., and Fulbright, D. W. 1983b. Influence of developmental stage
on susceptibility of tomato fruit to Pseudomonas syringae pv. tomato.
Phytopathology 73 :3 6-3 8.

Gharbi, A. G., and Nameth, S. T. 1992 (Abstract). The use of monoclonal-antibody-
based serodiagnostic assay for detection of Clavibacter michiganensis subsp.
michiganensis in symptomless tomato seedlings. Phytopathology 82: 1076.

 

Gitaitis, R., McCarter, S., and Jones, J. 1992. Disease control in tomato transplants in S
Georgia and Florida. Plant Disease 76:651-656.

 

Gitaitis, R. 1991. Bacterial canker. Pages 24-26 in: Compendium of Tomato Diseases. T"
Jones J.B., Jones J.P., Stall RE, and litter, T.A., eds. 2"" edition. American
Phytopathological Society Press, St. Paul, MN.

Gitaitis, R. D., Beaver, R. W., and Voloudakis, A. E. 1991. Detection of Clavibacter
michiganensis subsp. michiganensis in symptomless tomato transplants. Plant
Disease 75:834-838.

Gitaitis, R. D., and Beaver, R. W. 1990. Characterization of fatty methyl ester content of
Clavibacter michignensis subsp. michiganensis. Phytopathology 80:318-321.

Gitaitis, R. D. 1989. Survival and mechanical transmission of Clavibacter michiganensis
subsp. michiganensis in southern-grown transplants in 1989. Proceedings, F ifih
Annual Tomato Disease Workshop, Michigan, Dec. 7-8.

Gleason, M. L., Parker, S. K., and Braun, E. J. 1995. Assaying tomato seed coat fiber

for Cmm. Proceedings, Eleventh Annual Tomato Disease Workshop, Ontario, Dec.
7-8.

Gleason, M. L., Gitaitis, R. D., and Ricker, M. D. 1993. Recent progress in
understanding and controlling bacterial canker of tomato in eastern North America.
Plant Disease 77:1069-1076.

Gleason, M. L., Braun, E. J., Carlton, W. M., and Peterson, R. H. 1991. Survival and

dissemination of Clavibacter michiganensis subsp. michiganensis in tomatoes.
Phytopathology 81 :1519-1523.

37

mi

Harris-E
c0:

Pit

real

Jaunet. T
Rut
89:

JEWEL; B
Her
SEq‘

Gould, W. A. 1992. Tomato Production, Processing, and Technology. 3lrd edition. CTI
publications, Inc. Baltimore, MD.

Goto, M. 1992. Fundamentals of Bacterial Plant Pathology. Academic Press, Inc. San
Diego, CA.

Gudmestad ,N. C., Baer, D., and Kurowski, C. J. 1991. Validating immunoassay test-
performance in the detection of Corynebacterium sepedonicum during the growing
season. Phytopathology 81:475-480.

Grogan, R. G., and Kendrick, J. B. 1953. (Abstract) Seed transmission, mode of
overwintering, and spread of bacterial canker of tomato caused by Corynebacterium
michiganense. Phytopathology 43:473.

Harris-Baldwin, A., and Gudmestad, N. C. 1996. Identification of phytopathogenic

coryneform bacteria using the Biolog automated microbial identification system.
Plant Disease 80:874-878.

Hausbeck, M. K., Bell, J ., Medina-Mora, C. M., and Fulbright, D. W. 1999. (In progress)
Impact of bactericides applied to tomato seedlings in the greenhouse on C.
michiganensis subsp. michiganensis populations and field productivity.

Hausbeck, M., Fulbright, D., Bell, J ., Medina, C., and de Bruijn, F. J. 1995a.
Epidemiology of bacterial diseases on vegetables transplants in the greenhouse.
Proceedings, Eleventh Annual Tomato Disease Workshop, Ontario, Dec. 7-8.

Hausbeck, M., Fulbright, D., Louws, F., Bell, J ., and de Bruijn, F. J. 1995b. (Abstract)
Bacterial canker among tomato transplants in the greenhouse and impact on yield.
Phytopathology 85:1119.

Henningson, P. J ., and Gudmestad, N. C. 1991. Fatty acid analysis of phytopathogenic
coryneform bacteria. Journal of General Microbiology 137:427-440.

Hu, X., Lai, F.-M., Reddy, A. S. N., and Ishimaru, C. A. 1995. Quantitative detection of
Clavibacter michiganensis subsp. sepedonicus by competitive polymerase chain
reaction. Phytopathology 85: 1 468- 1473 .

Jaunet, T. X., and Wang, J. F. 1999. Variation in genotype and aggressiveness of
Ralstonia solanacearum race 1 isolated from tomato in Taiwan. Phytopathology
89:320-327.

Jersek, B., Gilot, P., Gubina, M., Klun, N., Mehle, J., Tchemeva, E., Rijpens, N., and

Herman, L. 1999. Typing of Listeria monocytogenes strains by repetitive element
sequence-based PCR. Journal of Clinical Microbiology 37:103-109.

38

Johansen, I. E., Rasmussen, O. F ., and Heide, M. 1989. Specific identifiaction of

Clavibacter michiganense subsp. sepedonicum by DNA-hybridization probes.
Phytopathology 79: 101 9-1 023.

Jones, J. B., Chase, A. R., and Harris, G. K. 1993. Evaluation of the Biolog GN
microplate system for identification of some plant-pathogenic bacteria. Plant
Disease 77:553-558.

Jones, J. B. 1991. Bacterial Spot. Page 26 in: Compendium of Tomato Diseases. 2nd
edition. Jones J .B., Jones J.P., Stall RE, and litter ,T.A., eds. American
Phytopathological Society Press, USA.

Kampfer, P., Seiler, H., and Dott, W. 1993. Numerical classification of coryneform
bacteria and related taxa. Journal of General and Applied Microbiology 39: 135-
214.

Kontaxis, D. G. 1962. Leaf trichomes as avenues for infection by Corynebacterium
michiganense. Phytopathology 52: 1306-1307.

Koueth,T., Versalovic, J ., and Lupski, J. R. 1995. Differential subsequence conservation

of interspersed repetitive Streptococcus pneumoniae BOX elements in diverse
bacteria. Genome Research 5:408-418.

Layne, R. E. C. 1967. Foliar trichomes and their importance as infection sites for
Corynebacterium michiganense on tomato. Phytopathology 57:981-985.

Lee, I.-M., Bartoszyk, I. M., Gundersen, D. E., Mogen, B., and Davis, R. E. 1997a.
Nested PCR for ultrasensitive detection of the potato ring rot bacterium,

Clavibacter michiganensis subsp. sepedonicus. Applied and Environmental
Microbiology 63: 2625-2630.

Lee, I.-M., Bartoszyk, I. M., Gundersen, D. E., and Davis, R. E. 1997b. Phylogeny and
classification of bacteria in the genera Clavibacter and Rathayibacter on the basis

of 16S rRNA gene sequence analysis. Applied and Environmental Microbiology
63 :263 1-2636.

Liu, P. Y. F., Lau, Y. J., Hu, B. S., Shyr, J. M., Shi, Z. Y., Tsai, W. 8., Lin, Y. H., and
Tseng, C. Y. 1995. Analysis of clonal relationships among isolates of Shigella

sonnei by different molecular typing methods. Journal of Clinical Microbiology
33:1779-1783.

Lopez M. M., Noval C., and Palazon, I. 1987. Plant pathogenic bacteria identified in

Spain. Pages 645-652 in: Current Plant Science Biotechnology Agriculture.
Dordrecht, Martins Nijhoff Publishers.

39

Lou“

Louu'

louu'

louu‘

Louws, F. J., Bell, J ., Medina-Mora, C. M., Smart, C. D., Opgenorth, D., Ishimaru, C. A.,
Hausbeck, M. K., de Bruijn, F. J ., and Fulbright, D. W. 1998. Rep-PCR-Mediated
genomic fingerprinting: A rapid and effective method to identify Clavibacter
michiganensis. Phytopathology 88:862-868.

Louws, F. J ., Schnieder, M., and de Bruijn, F. J. 1996. Assessing genetic diversity of
microbes using repetitive sequence-based PCR (rep-PCR). Pages 63-94 in: Nuclei
Acid Amplification Methods of Environmental Samples. Torazos G., ed.
Technomic Publishing, Co. Lancaster, PA.

Louws, F. J., Fulbright, D. W., Stephens, C. T., and de Bruijn, F. J. 1995.
Differentiation of genomic structure by rep-PCR fingerprinting to rapid classify
Xanthomonas campestris pv. vesicatoria. Phytopathology 85:528-536.

Louws, F. J., Fulbright, D. W., Stephens, C. T., and de Bruijn, F. J. 1994. Specific
genomic fingerprints of phytopathogenic Xanthomonas and Psuedomonas pathovars
and strains generated with repetitive sequences and PCR. Phytopathology 60:2286-
2295.

McManus, P., and Jones, A. 1995. Genetic fingerprinting of Erwinia amylovora strains
isolated from tree-fruit crops and Robus sp. Phytopathology 85:1547-1553.

McManus, P., and Jones, A. 1994. Role of wind-driven rain, aerosols, and contaminated
budwood in incidence and spatial pattern fire blight in an apple nursery. Plant
Disease 78:1059—1066.

Meyer, W., and Mitchell, T. G. 1995. Polymerase chain reaction fingerprinting in fungi
using single primers specific to minisatellites and simple repetitive DNA sequences:
Strain variation in Cryptococcus neoformans. Electrophoresis 16: 1648-1656.

Meyer, W., Mitchell, T. G., Freedman, E. Z., and Vilgalys, R. 1993. Hybridization
probes for conventional DNA fingerprinting used as single primers in the
polymerase chain reaction to distinguish strains of Cryptococcus neoformans.
Journal of Clinical Microbiology 31 :2274-2280.

Mills, D., Russell, B. W., and Hanus, J. W. 1997. Specific detection of Clavibacter
michiganensis subsp. sepedonicus by amplification of three unique DNA sequences
isolated by subtraction hybridization. Phytopathology 87:853-861.

Mirza, M. S., Rademaker, J. L. W., Janse, J. D., and Akkermans, A. D. L. 1993. Specific
16S ribosomal RNA targeted oligonucleotide probe against Clavibacter
michiganensis subsp. sepedonicus. Canadian Journal of Microbiology 39:1029-
1034.

40

Pine. 1
c

POhl'Ol
0

Radar.

Murry, M. A., Konopka, A. S., Pratt, S. D., and Vandergon, T. L. 1997. The use of
PCR-based typing methods to assess the diversity of F rankia nodule endophytes of
the actinorhizal shrub Ceanothus. Physiologia Plantarum 99:714-721.

Opgenorth, D. C., Smart, C. D., Louws, F. J., de Bruijn, F. J., and Kirkpatrick, B. C.
1996. Identification of Xanthomonas fligariae field isolates by rep-PCR genomic
fingerprinting. Plant Disease 80:868-873.

Pan, Y.-B., Grisham, M. P., Burner, D. M., Darnann, K. E. Jr., and Wei, Q. 1998. A
polymerase chain reaction protocol for the detection of Clavibacter xyli subsp. xyli,

the causal bacterium of sugarcane ratoon sttmting disease. Plant Disease 82:285-
290.

Parker, S. K., Gleason, M. L., and Braun, E. J. 1995. Determining the probability of
detecting Cmm in tomato seed lots. Proceedings, Eleventh Annual Tomato Disease
Workshop, Ontario, Dec. 7-8.

Pine, T. S., Grogan, R. G., and Hewitt, Wm. B. 1955. Pathological anatomy of bacterial
canker of young tomato plants. Phytopathology 45:267-271.

Pohronezny, K., and Volin, R. B. 1983. The effect of bacterial spot on yield and quality
of fresh market tomatoes. HortScience 18:69-70.

Rademaker, J. W. L., Hoste, B., Louws, F. J ., Kersters, K., Swings, J ., Vauterin, L.,
Vauterin, P., and de Bruijn, F. J. 1999a. (In preparation) Comparison of AF LP and
rep-PCR genomic fingerprinting with DNA-DNA homology studies: Xanthomonas
as a model.

Rademaker, J. W. L., Louws, F. J ., Rossback, U. Vinuesa, P., and de Bruijn, F.J. 1999b.
Computer-assisted pattern analysis of molecular fingerprints and database
construction. In: Molecular Ecology Manual. Akkerrnans A.D.L., van Elsas J .D.,
and de Bruijn, F .J ., eds. Kluwer Academic Publishers, Dordrecht.

Rademaker, J. W. L., and de Bruijn, F. J. 1997. Characterization and classification of
microbes by rep-PCR genomic fingerprinting and computer-assisted pattern
analysis. Pages 151-171 in: DNA markers: Protocols, Applications, and
Overviews. Caetano-Anollés, G., and Gresshoff, P.M., eds. Wiley and Sons, Inc.
New York, NY.

Rademaker, J. W. L., Louws, F. J ., Schultz, M. H., Vauterin, L., Swings, J., and de
Bruijn, F. J. 1997. (Abstract) Molecular systematics of xanthomonads by rep-PCR

genomic fingerprinting and computer-assisted pattern analysis. Phytopathology
87:88 1 .

41

Rob

 

Rademaker, J. W. L., and Janse, J. D.. 1994. Detection and identification of Clavibacter
michiganensis subsp. sepedonicus and Clavibacter michiganensis subsp.
michiganensis by nonradioactive hybridization, polymerase chain reaction, and
restriction enzyme analysis. Canadian Journal of Microbiology 40: 1007-101 8.

Ricker, M. D. 1997. Evaluation of four tomato genotypes for resistance to Clavibacter
michiganensis subsp. michiganensis. Proceedings, Thirteenth Annual Tomato
Disease Workshop, Indiana, Dec. 4-5.

Ricker, M. D. 1995. (Abstract) Isolation of Clavibacter michiganensis subsp.
michiganensis from detached tomato fibers. Phytopathology 8521153.

Ricker, M. D. and Riedel, R. M. 1993. Effect of secondary spread of Clavibacter
michiganensis subsp. michiganensis on yield of Northern processing tomatoes.
Plant Disease 77:364-366.

Ricker, M. D. 1989. Bacterial canker research 1985-89. Proceedings, Fifth Annual
Tomato Disease Workshop, Michigan, Dec. 7-8.

Rivera, I. G., Chowdhury, M. A. R., Huq, A., Jacobs, D., Martins, M. T., and Colwell, R.
R. 1995. Enterobacterial repetitive intergenic consensus sequences and the PCR to
generate fingerprints of genomic DNAs from Vibrio cholerae 01, 0139, and non-
01 strains. Applied and Environmental Microbiology 61 :2898-2904.

Roberts, P. D., Jones, J. B., Chandler, C. K., Stall, R. E., and Berger, R. D. 1996.
Survival of Xanthomonas fiagariae in summer nurseries in Florida detected by

specific primers and nested polymerase chain reaction. Plant Disease 80: 1283-
1288.

Rodriguez-Barradas, M. C., Hamill, R. J ., Houston, E. D., Georghiou, P. R., Clarrdige, J.
E., Regnery, R. L., and Koehler, J. E. 1995. Genomic fingerprinting of Bartonella

species by repetitive element PCR for distinguishing species and isolates. Journal
of Clinical Microbiology 33:1089-1093.

Sadowsky, M. J ., Kindel, L. L., Bowers, J. H., amd Schottel, J. L. 1996. Use of
repetitive intergenic DNA sequences to classify pathogenic and disease-suppressive
Streptomyces strains. Applied and Environmental Microbiology 62:3489-3493.

Schaad, N. W. 1987 . Problems with the pathovar concept. Pages 783-785 in: Current
Plant Science Biotechnology Agriculture. Dordrecht, Martius Nijhofi“ Publishers.

Schaad, N. W. 1981. Cruciferous weeds as sources of inoculum of Xanthomonas
campestris in black rot of crucifers. Phytopathology 71 : 1215-1220.

42

Sch:

Seth

Sand

Schnieder, M., and de Bruijn, F. J. 1996. Rep- PCR mediated genomic fingerprinting of
rhizobia and computer-assisted phylogenetic pattern analysis. World Journal of
Microbiology and Biotechnology 12: 163-174.

Sechi, L. A., Zanetti, S., Dupre, I., Delogu, G., and Fadda, G. 1998. Enterobacterial
repetitive intergenic consensus sequences as molecular targets for typing of
Mycobacterium tuberculosis strains. Journal of Clinical Microbiology 36:128-132.

Smith, E. F. 1910. A New Tomato Disease of Economic Importance. Science 31 :794-
796.

Smith, J. J., Offord, L. C., Holderness, M., and Saddler, G. S. 1995. Genetic diversity of
Burkholderia solanacearum (synonym Pseudomonas solanacearum) race 3 in
Kenya. Applied and Environmental Microbiology 61 :4263-4268.

Sousa Santos, M., Cruz, L., Norskov, P., and F rasmussen. 1997. A rapid and sensitive
detection of Clavibacter michiganensis subsp. michiganensis in tomato seeds by
polymerase chain reaction. Seed Science and Technology 25:581-584.

Speranzini, D. 1995. Incidence of tomato diseases in Ontario-1995. Proceedings,
Eleventh Annual Tomato Disease Workshop, Ontario, Dec. 7-8.

Starr, M.P., Mandel, M., and Murata, N. 1975. The phytopathogenic coryneform
bacteria in the light of DNA base composition and DNA-DNA segmental
homology. Journal of General and Applied Microbiology 21 :13-26.

Stephens, C. T., Stebbins, T., and Fulbright, D. W. 1988. Detection of Cmm using four
selective media and Patho-screen ELISA assay (Agdia). Proceedings, Fourth
Annual Tomato Disease Workshop.

Stephens, C. T., and Fulbright, D. W. 1986. Bacterial Canker of Tomato. Page 1-4 in:
MSU Ag Facts. Extension Bulletin E-1943.

Strider, D. L. 1969a. Bacterial Canker of Tomato caused by Corynebacterium
michiganense: A literature review and bibliography. North Carolina Experiment
Station, Technical Bulletin 193.

Strider, D. L. 1969b. Foliage blight phase of bacterial canker of tomato and survival of
Corynebacterium michiganense in toxicants and air-dried conditions. Plant
Disease Reporter 53:864-868.

Strider, D. L. 1967. Survival studies with tomato bacterial canker organism.
Phytopathology 57: 1067-1071 .

43

Than

lhorr.

lhu.

I

Thyr.
I

Tsianti
n

Tsianii
n

Vaeren
ll]

Vahdar
Pl

Thanos, M., Schonian, G., Meyer, W., Schweynoch, C., Graser, Y., Mitchell, T. G.,
Presber W., and Tietz, H.-J. 1996. Rapid identification of Candida species by
DNA fingerprinting with PCR. Journal of Clinical Microbiology 34:615-621.

Thompson, E., Leary, J. V., and Chun, W. W. 1989. Specific detection of Clavibacter
michiganense subsp. michiganense by a homologous DNA probe. Phytopathology
79:3 1 1-3 14.

Thyr, B. D. 1969. Assaying tomato seed from Corynebacterium michiganense. Plant
Disease Reporter 53:858-860.

Thyr, B. D., 1968. Resistance to bacterial canker in tomato, and its evaluation.
Phytopathology 58:279-281 .

Tsiantos, J. A. 1987. Epiphytic survival of Corynebacterium michiganense pv.
michiganense on tomato leaves. Microbios Letters 34:59-66.

Tsiantos, J. A. 1986. Survival of phytopathogen Corynebacterium michiganense pv
michiganense in water. Microbios Letters 32:69-74.

Vaerenbergh, J. P. C., and Chauveau, J. F. 1987. Detection of Corynebacterium
michiganense in tomato seed lots. Bulletin OEPP/EPPO 17:131-138.

Vandamme, P., Pot, B., Gillis, M., de Vos, R, Kersters, K., and Swings, J. 1996.
Polyphasic taxonomy, a consensus approach of bacterial systematics.
Microbiological Reviews 60:407-438.

van Belkum, A., Sluijter, M., de Groot, R., Verbrugh, H., and Hermans, P. W. M. 1996.
Novel BOX repeat PCR assay for high-resolution typing of Streptococcus
pneumoniae strains. Journal of Clinical Microbiology 34:1 176-1179.

van Vuurde, J. W .L. 1987. New approach in detecting phytopathogenic bacteria by
combined immunoisolation and immunoidentification assays. Bulletin
OEPP/EPPO 17:139-148.

Vera Cruz, C. M., Ardales, E. Y., Skinner, D. Z., Talag, J ., Nelson, R. J., Louws, F. J.,
Leung, H., Mew, T. W., and Leach, J. E. 1996. Measurment of halotypic variation
in Xanthomonas oryzae pv. oryzae within a single field by rep-PCR and RFLP
analyses. Phytopathology 86:1352-1359.

Verrault, H., Lafond, M., Asselin, A., Banville, G., and Bellemare, G. 1988.

Characterization of two DNA clones specific for identification of C orynebacterium
sepedonicum. Canadian Journal of Microbiology 34:993-997.

44

Versa

Yam

Vés-qu

Yinue:

(9‘13 (10 r-O. n

Waiter

al

Weebb

Versalovic, J ., de Bruijn, F. J ., and Lupski, J. R. 1997. Repetitive sequence-based PCR
(rep-PCR) DNA fingerprinting of bacterial genomes. Pages 43 7-454 in: Bacterial
Genomes: Physical Structure and Analysis. de Bruijn, F. J ., Lupski, J. R., and
Weinstock, G. M., eds. Chapman and Hall, NY.

Versalovic, J ., Schneider, M., de Bruijn, F. J ., and Lupski, J. R. 1994. Genomic
fingerprinting of bacteria using repetitive sequence-based Polymerase Chain
Reaction. Methods in Molecular and Cellular Biology 5:25-40.

Versalovic, J ., Koeuth, T., and Lupski, J. R.. 1991. Distribution of repetitive DNA
sequences in eubacteria and application to fingerprinting of bacterial genomes.
Nucleic Acids Research 19:6823-6831.

Vésquez-Arroyo, J ., Sessitsch, A., Martinez, E., and Pena-Cabriales, J. J. 1998. Nitrogen
fixation and nodule occupancy by native strains of Rhizobium on different cultivars
of common bean (Phaseolus vulgaris L.). Plant and Soil 204: 147-1 54.

Vinuesa, P., Rademaker, J. L. W., de Bruijn, F. J ., and Werner, D. 1998. Genotypic
characterization of Bradyrhizobium strains nodulating endemic woody legumes of
the Canary Islands by PCR-restriction Fragment Length Polymorphism analysis of
genes encoding 16S rRNA (16S rDNA) and 168-238 rDNA intergenic spaces, and
partial l6 rDNA sequencing. Applied and Environmental Microbiology 64:2096-
2104.

Warren, R., Richardson, M., Sampson, S., Hauman, J. H., Beyers, N., Donald, P. R., and
van Helden, P. D. 1996. Genotyping of Mycobacterium tuberculosis with

additional markers enhances accuracy in epidemiological studies. Journal of
Clinical Microbiology 341221 9-2224.

Weebb, R. B, Good, J. M., and Danielson, L. L. 1967. Tomato diseases and their
control. US. Department of Agriculture, Agriculture Handbook 203:9-13.

Weingart, H., and Volksch, B. 1997. Genetic fingerpriting of Pseudomonas syringae
pathovars using ERIC-, REP-, and ISSO-PCR. Journal of Phytopathology 145:339-
345.

Weising, K., Nybom, H., Wolf, K., and Meyer, W. 1995. DNA Fingerprinting in Plants
and Fungi. CRC press, F L.

Weising, K., Kaemmer, D., Epplen, J. T., Weigand, F ., Saxena, M., and Kahl, G. 1991.
DNA fingerprinting of Ascochyta rabiei with synthetic oligodeoxynucleotides.
Currents Genetics 19:483-489.

Woods, A. C. 1984. Moko disease: Atypical symptoms induced by afluidal variants of
Pseudomonas solanacearum in banana plans. Phytopathology 74:972-976.

45

Yamada, K., and Komagata, K. 1972a. Taxonomic studies on coryneform bacteria IV.
Morphological, cultural, biochemical, and physiological characteristics. Journal of
Genetics and Applied Microbiology 18:399-416.

Yamada, K., and Komagata, K. 1972b. Taxonomic studies on coryneform bacteria V.
Classification of coryneform bacteria. Journal of Genetics and Applied
Microbiology 18:417-431.

46

subspe
until it
ldemif
critical
dii'ersii
host'pa
manor
michigc
Pathogc
Similafi‘
lgila &
and the L

U56 0f p1

t
01 P r0100

CHAPTER 2

GENETIC VARIATION IN STRAINS OF
CLA VIBA CTER MICHIGANENSIS SUB SPECIES MICHIGANENSIS
The genetic characterization of the bacterium Clavibacter michiganensis
subspecies michiganensis, causal agent of bacterial canker on tomatoes, is limited and
until recently (Louws et al., 1998) no consistent genetic variability has been noted.
Identification and detection of the pathogen at various times of tomato production is
critical for development of control strategies and similarly assessment of the genetic
diversity of the pathogen in the field is necessary for epidemiological studies as well as
host/pathogen studies. Identification and assessment of diversity is the cornerstone of
taxonomy. Yet, over the years classification of Clavibacter michiganensis subspecies
michiganensis has been accomplished through physiological, chemical, serological, and
pathogenic analyses, which have proved controversial due to the high degree of genetic
similarity between the Clavibacter michiganensis subspecies (Yamada & Komagata,
1972a & 1972b; Carlson & Vidaver, 1982; Davis et al., 1984; Lee et al., 1997a & 1997b)
and the lack of diversity among strains of C. michiganensis subsp. michiganensis. The
use of protein measurable and DNA-based techniques such as protein profiles, restriction
fragment length polymorphism (RF LP), and polymerase chain reaction (PCR) technology
have provided needed insight and understanding of the genetic character of this
phytopathogenic species (Davis et al., 1984; Mogen et al., 1990; Rademaker & Janse,
1994; Drier et al., 1995; Louws et al., 1998). Since the discovery and identification of
DNA repetitive sequences found in prokaryotic and eukaryotic genomes, a wide variety

of protocols to amplify genes with PCR using these primers has been developed for the

47

provit
strain
Britijn
PCR)
this tel
knowlt
stains
appeali
and cos

1994; \

define g
genome
hOm re
3’? ranc
(Verse).
19%; \
detail h.
0that be
fingerp,

ht‘bn'dn

detection, characterization, and differentiation of microbes, including bacteria and fungi
of ecological, medical, and agricultural importance. Each PCR based technique designed
provides different levels of taxonomic resolution, ranging from genus to subspecies and
strain discrimination (Louws et al., 1996; Vandamme et al., 1996; Rademaker & de
Bruijn; 1997). However, the repetitive-sequence based polymerase chain reaction (rep-
PCR) has provided several advantages over other DNA techniques. For example, with
this technology it is possible to generate a genomic fingerprint without previous
knowledge of DNA sequences at a resolution that identifies species, subspecies, and
strains of various species (Louws et al., 1996; Rademaker & de Bruijn; 1997). The most
appealing attributes of the rep-PCR technique are its reproducibility, reliability, speed,
and cost when compare against other DNA fingerprinting techniques (V ersalovic et al.,
1994; Weising et al., 1995; Schneider & de Bruijn, 1996; Versalovic et al., 1997).

The rep-PCR protocol is one of the two DNA fingerprinting techniques used to
define genetic characters and polymorphic regions of prokaryotic and eukaryotic
genomes (Louws et al., 1996; Schneider & de Bruijn, 1996). The fingerprints generated
from rep-PCR technology originated from naturally occurring repetitive sequences that
are randomly arranged in high copy number on the circular bacterial genomes
(V ersalovic et al., 1991; de Bruijn, 1992; Koueth et al., 1995; Schneider & de Bruijn,
1996; Vera Cruz et al., 1996). Three groups of repetitive sequences have been studied in
detail have been used to generate primers for the amplification of DNA through PCR in
other bacterial species. The three primers used for the production of rep-PCR genomic
fingerprints consist of 18-22 oligonucleotides sequences recognized through DNA

hybridization studies (V ersalovic et al., 1991; Versalovic et al., 1994). They include the

48

eleme
(liner

the BI

consis
motif:
(M C
l 994 )
it has
WW]
1992;
as rap

Where

repetitive extragenic palindromic (REP) sequence, the enterobacterial repetitive
intergenic consensus (ERIC) sequence, and the BOX element (Louws et al., 1996;
Schneider & de Bruijn, 1996). These primers were developed after their characterization
in the Gram-negative enteric bacteria, Escherichia coli and Salmonella typhimurium for
the REP and ERIC primers, respectively. The combination of boxA, boxB, and boxC
elements were first characterized in Gram-positive bacterium, Streptococcus pneumonia
(Koeuth et al., 1995) and unrelated Gram-negative species were used in characterizing
the BOX primer (V ersalovic et al., 1994; Versalovic et al., 1997).

A different group of primers can be generated from minisatellite sequences, which
consist of 10-60 base pairs arranged in a head to tail fashion with two to several thousand
motifs, most likely containing non-coding DNA. The polytrinucleotides, (GTG)5 and
(GCC)5 are examples of primers complementary to minisatellites (V ersalovic et al.,
1994). The precise function of the repetitive and minisatellite sequences is unknown, but
it has been postulated to be important in chromosomal organization, DNA replication and
prevention of DNA degradation, as well as in regulation of gene expression (de Bruijn,
1992; Weising et al., 1995). Meyer et al. (1993) demonstrated the usefulness of (GTG)5
as rep-PCR primers for the differentiation of eukaryotic species. Some of the first reports
where (GTG)5 was used to prime rep-PCR were realized with human pathogenic species
such as the fungus Cryptococcus neoformans (Meyer et al., 1993; Meyer & Mitchell.,
1995). Various scientists showed that minisatellite sequences are randomly spread
through the entire fungal and prokaryotic genomes (Weising et al., 1991; Meyer et al.,
1993; Meyer & Mitchell., 1995; Weising et al., 1995; Warren et al., 1996). These studies

were important in generating information related to origin and geographic distribution of

49

various
demon:

t'atiabi

and dif
et al.. 1
1997: S
hngerp
and etc
serious
were :11
strains
“1dr? V';
chitin
C mic}
differs:
level \k‘
analyzc-
poll'mc
LOUWS
el al 1
from w

Types \k

various strains; data necessary for epidemiological studies. A recent report has
demonstrated the usefulness of (GTG)5-PCR genomic fingerprinting for assaying genetic
variability in various fungal species (Balerias Couto et al., 1996).

As in medical pathology, PCR-based techniques have revolutionized the detection
and differentiation of pathogens in plant pathology (Rademaker & J anse, 1994; Roberts
et al., 1996; Sadowsky et al., 1996; Lee et al., 1997a & 1997b; Rademaker & de Bruijn,
1997; Sousa Santos et al., 1997). Specifically, the use of rep-PCR genomic
fingerprinting has provided useful information for the reevaluation of taxonomic names
and evolutionary relationships within phytopathogenic bacterial species that are of
serious economical and agricultural concern. For example, Louws et al. (1994 & 1995)
were able to redefine phylogenetic groups and effectively differentiate phytopathogenic
strains within the pseudomonad and xanthomonad species that can cause symptoms on a
wide variety hosts. More recently, Louws et al. (1998) were able to rapidly and
effectively differentiate five Clavibacter michiganensis subspecies using rep-PCR. The
C. michiganensis subspecies analyzed are among characterized by their virulence on
different hosts. They also reported a higher level of genetic resolution at the subspecies
level when a wide collection of strains of C. michiganensis subsp. michiganensis was
analyzed with BOX-PCR. They defined four BOX-PCR (A, B, C, and D) types based on
polymorphic DNA bands after agarose-based electrophoresis. During the completion of
Louws et al. (1998) manuscript a new BOX-PCR type was described as the E-type (Bell
et al., 1997). The number of E-type strains is limited and these have been recovered only
from two geographical locations. Although the biological significance of the BOX-PCR

types was not established, genetically typing the strains provided useful groupings. For

50

example, almost all of the C. michiganensis subsp. michiganensis strains that are
asymptomatic on tomato (avirulent strains) belong to BOX-PCR type A, while highly
virulent strains belong to BOX-PCR type C.

The creation and analysis of dendrograms based on more than one DNA
fingerprint can help support the subdivision of strains into genotypes based on genetic
diversity (Schneider & de Bruijn, 1996; Louws et al., 1996; Rademaker et al., 1997;
Rademaker et al., 1998; Vinuesa et al., 1998; Jaunet & Wang, 1999; Rademaker et al.,
1999a; Rademaker et al., 1999b). That C. michiganensis subsp. michiganensis can be
divided into five genotypes based on BOX-PCR protocol suggests that this subspecies is
genetically narrow. Defining genetic variability should help us understand more about
the role of the strains in the host/pathogen relation and may provide us with powerful
tools in examining the epidemiological picture of this disease. To this end, I have
focused on two objectives; 1) to determine if other rep-PCR universal primers support the
subdivision of C. michiganensis subsp. michiganensis into the five genotypes established
with BOX-PCR; and if so 2) determine if the measurable diversity allows for a better

understanding of the bacterial canker disease cycle and the relationship of the strains.

MATERIALS AND METHODS
The 6 Clavibacter michiganensis strains representing different subspecies of C.
michiganensis and the 175 representing C. michiganensis subsp. michiganensis strains
used in this study were obtained from a culture collection and are described in Table l.
Clavibacter michiganensis subsp. michiganensis strains listed in Table 1 are categorized

by their previous BOX-PCR type as described by Louws et al. (1998). Subdivision of

51

Table 1

’

5mm!

\

{mutton

.,. -, 'A
Sandal}:

Table 1. Strains of subspecies of Clavibacter michiganensis used in this study.

 

 

Subspecies PCR typea Originb Varietyc Tissued Virulencec Strain(s)f
insidiosum CDA2(ATTCC33 14)
sepedonicus CIC4 (NCPPB2139)
nebraskensis CICl3, CIC 17
tessellarius C1C21, C1C22
michiganensis A Ml P seedling V 139, 210, 211, 297, 300, 301
l 129
AV 132, 133, 134, 136, 138, 207,
302
foliage V T3, T4, T5, T66, T67, 936
fruit T1, T8, T12, T13, T17, T18,
T19, T20, T23, T24, T27,
T35, T41, T42, T43, T45,
T56, T57, T63, T65, T70,
T80, 294
AV T2
nk V 229, 292, 294, 672, 676, T73
l 671
AV 209, 673, 674, 675
ND 622, 624, 625, 899, 900, 933,
934, 935, 950
FM fruit V 299
nk V 127
ND 357, 360
OH P fruit V 353, 355, 356
ND 363
nk nk AV 9, 29
B Ml P fruit V 30
FM fruit V 12
foliage V 35
nk nk ND 619, 620, 621, 623, 668
OH FM foliage V 69
fruit V 68, 77
nk nk V 226
CA FM foliage ND CA-Ol, CA-lOR, CA-Y
fruit ND CA-S, CA-SR
ON nk nk V R28

 

52

Table I

’—
SJPSWC
——-A—_

srnzgin

Table 1. (cont'd)

 

 

Subspecies PCR typea Originb Varietyc Tissued Virulencec Strain(s)f
michiganensis C Ml P foliage V 955, 958, 961, 962, 963
l 15, 33
FM foliage V 18, 954, 956
l 24
ND NW89, NW90, NW99,
NWIOO, NW107, NW108,
NW110
fruit V 8, 44, 53, 236
fruit 1 56
nk foliage V 237
nk fruit V 234
l 218
nk nk V 13
l 216
ND 81,83, 84, $6,812
F M foliage V 122
nk V 117, 230
nk nk V 27, R19
1 26
NC nk foliage V 21
fruit V 3
nk V 54, R4
ND 14
CA P foliage V GH- 1 97
seed V GH-7902
Chile P foliage V GH-213
seed V GH-202A, Gil-71033
I GH-ZOZB
China P seed V GH-Zl l, GH-212, GH-71127
FM seed V GH-7l68, GH-71182,

l

GH-71196, (EH-71235,
GH-71290, GH-7l335,
GH-71341,GH-71412,

GH-17427, GH-7l428,
Gl-l-71430, GH-7l431,
(EH-71472, GH-71564,
GH-71474, GH-713350

GH-7l 176

 

53

Table 1. (cont'd)

 

Subspecies PCR typeaI Originb Varietyc Tissued Virulencee Strain(s)f

 

michiganensis C nk nk nk V R13
Morroco FM seed ND GH-7904
D MI FM foliage ND Betl-5-1, Bet1-5-2, Bet3-5-1,

Bet3-5-2, FMl-4-l,
MFl-S-l, MFl-S-Z,

MF3-5-1
fruit I 286, 288
India F M seed ND GH-7137P
Morroco FM seed ND GH-7219, GH-7219P,
GH-7388, GH-7388C
E MI nk foliage I 2S, 10S, l4S, 30S, 31S
ND 488
fruit 1 IF, 9F, 10F
E OH nk foliage AV 118

 

 

IPCR= pfiymerase chain reaction designated type using B_O'X primer Gouws et al., 1998).
b Original source of C. michiganensis subspecies other than subspecies michiganensis were reported on
Louws et al., 1998. Michigan=MI, Ohio= OH, North Carolina=NC, Califomia=CA, Nebraska=NE,

and Ontario=ON.
° Processing variety=P, F resh-Market variety=FM, and not known=nk.

dnot known=nk.
°Virulence= producing canker and wilt; l=intermediate, producing canker only; AV=avirulent,producing

no symptoms; and ND= not determined (Louws et al., 1998).
f Strain source: T=F.J.Louws collection, CA and GH= California, CDA= California Department of

Agriculture, CIC= Carol Ishirnaru collection, R= Mark Ricker, and Bet,FM,MF,& NW= Nicole
Werner; Michigan State University, Dept. of Botany & Plant Pathology.

54

each PC F
plant sou
our labor
by Louis
stains o

glucose )

Versaloi-
161(3), 1
merlay I
NJ. LS;

fitting 1]

Smittur
E31 eleer
louws (
“11h eat
EatllOE

“'35 pre

gels \Va
Kortnj}

rEPOTIe.

each PCR-type (A, B, C, D, and E) is based on the geographical location, plant variety,
plant source, and virulence. Virulence of each strain used was previously determine by
our laboratory using the pathogenicity and hypersensitive response (HR) tests as reported
by Louws et a1 (1998). All bacteria were stored in glycerol at -20° C or -70° C. Bacterial
strains were grown on nutrient both yeast extract (N BY; Fatrni & Schaad, 1988) agar (no
glucose) at 27° C for 4 to 10 days depending of the subspecies.

The (GTG)5-PCR and BOX-PCR protocols were carried out as described by
Versalovic et al. (1994) with the following modifications. For the amplification with
(GTG)5 primer, the annealing time was lowered to 30 sec, and mineral oil was not used to
overlay each PCR reaction mixture as the thermal cycler (Genemate-Techne, Princeton,
NJ, USA) used consisted of a heated lid which prevented evaporation of the product
during the amplification procedure.

The (GTG)5 and BOX primers were synthesized by the Macromolecular
Structure, Sequence and Synthesis Facility at Michigan State University as the agarose
gel electrophoresis and gel photographic conditions were similar to those reported by
Louws et a1. (1998). Ninety five percent of the strains included in Table l were amplified
with each primer at least two times during separate (GTG)5-PCR and BOX-PCR
reactions. Genomic DNA from Cmm 936 used as positive control in every PCR reaction
was prepared as described by Louws et al. (1998).

Computer-image analysis of photographs of the ethidium bromide-stained agarose
gels was performed using the GelCompar gel analysis system program (Applied Math,
Kortrijk, Belgium) following the recommendations of the manufacturer and those

reported by Vauterin & Vauterin (1992), and Rademaker & de Bruijn (1997).

55

Photogra
images v
Kb) DN.
fngerpri
dcnsiton
generate
and iron
Similari
and Ulm
describe
Single fi
lGlGlg

Tfiomn

Photographs were scanned using a scanner (Hewlett Packard, Scan Jet 3C), and raw
images were stored. Each fingerprint pattern was standardized to the 1 kilobase pair (1
Kb) DNA maker (Gibco, USA) on three lanes in each gel. Normalized gel tracts for each
fingerprint pattern was digitized and stored for further analysis. An area of the
densitometric curves for each gel tract was selected before a similarity matrix was
generated; for (GTG)5-PCR and BOX-PCR the area corresponding from the 2.2 to 0.5 kb
and from the 2.0 to 0.4 kb range of fingerprint bands were selected, respectively.
Similarity matrices and dendrograms were generated using product-moment correlation
and unweighted pair group with arithmetic averages (UPGMA) clustering, as previously
described by Louws et al. (1994) and Rademaker et al. (1999b). Dendrograrns from
single fingerprint patterns (either (GTG)5-PCR or BOX-PCR) and from combined
(GTG)5-PCR and BOX-PCR fingerprint patterns were generated following the
recommendations of the manufacturer.

Cluster designation for each dendrogram was accomplished after the definition of
an arbitrary baseline. To define the baseline for the computer-generated similarity index
for each dendrogram, two reproducibility experiments were conducted. Ten independent
PCR master-mixes for each primer ((GTG)5 and BOX) were prepared on one day with
the same stock-aliquot for each PCR reagent. One Clavibacter michiganensis subsp.
michiganensis strain of each BOX-PCR type (Table 1) was used as DNA template;
strains Cmm 936 (type A), Cmm 68 (type B), Cmm 236 (type C), Cmm 286 (type D),
and Cmm 10F (type B), and genomic DNA was used as positive control for strain Cmm
936. From the same strain cultured on NBY agar, six colonies for each C. michiganensis

subsp. michiganensis strain was used as DNA templates for each PCR reaction. Five of

56

the color
colonies
protocol.
on the pi
conduct:
michigar
agarose ;

4‘0. E:

the colonies were used as template in five independent PCR reactions and one of the
colonies was repetitively used for the remaining five PCR reactions. PCR amplification
protocols and thermal cycler conditions were adjusted as previously indicated depending
on the primer used; therefore, (GTG)5-PCR and BOX-PCR amplifications were
conducted on consecutive days. Six microliters of each PCR product from the same C.
michiganensis subsp. michiganensis strain were resolved at the same time in a single
agarose gel (1.75% in 0.5 X TAE) under the same electrophoretic conditions (83 volts at
4°C). Each agarose gel was stained, photographed, and computer analyzed as mentioned,
above. Two experimental variability baselines were established; one for dendrograms of
fingerprint patterns from a single primer, and another for dendrograms of combined

fingerprint patterns.

RESULTS
(GTG)5-PCR differentiates subspecies of Clavibacter michiganensis. To determine if
the genetic diversity of Clavibacter michiganensis subsp. michiganensis resolved by
BOX-PCR into genotypes A, B. C, D, and E, can be further resolved using another
primer, known as (GTG)5, a collection of 175 C. michiganensis subsp. michiganensis
strains was studied. The genomic fingerprint obtained with (GTG)5-PCR consisted of
series of amplified bands with a more complex pattern when compared to genomic
fingerprints generated with BOX-, ERIC- and REP-PCR. The number of strongly
amplified bands obtained with BOX-PCR was approximately 15, whereas 11 were
obtained when using (GTG)5-PCR; however, less strongly amplified bands were

generated with the (GTG)5 primer than the BOX primer.

57

To demo
differentiated in
gains from a C(
characterization
Egroups as dist
group. The Olht
compared to C
19981 and are 1i
100.5. lb region
fingerprints t F i;
michiganensis l
ubspecies (F i g
the cluster anal
diderentiated f1
lherefore, (GT
fmm other sub:

BOX'PCR due

ReProducibili
fingerprint pan
SITaims of C m
tithe PCR rea

experimental V

To demonstrate that Clavibacter michiganensis subsp. michiganensis could be
differentiated from other subspecies of Clavibacter michiganensis using (GTG)5-PCR, 12
strains from a collection of 175 strains were selected for (GTG)5-PCR amplification and
characterization (Table 1). Of the 12 strains, two each represented the type A, B, D, and
E groups as distinguished by BOX-PCR, and four represented the larger type C type
group. The other subspecies of C. michiganensis represented strains previously
compared to C. michiganensis subsp. michiganensis using BOX-PCR (Louws et al.,
1998) and are listed in Table 1. Approximately tvventy-five visible bands between the 2.2
to 0.5 kb region were included in the cluster analysis of the (GTG)5-PCR genomic
fingerprints (Figure 1). One half of the visible bands from the C. michiganensis subsp.
michiganensis fingerprints co-migrated with bands from the four C. michiganensis
subspecies (Figure 1). Based on the four C. michiganensis subspecies used in this study,
the cluster analysis indicated that all of the C. michiganensis subspecies could be easily
differentiated from each other based solely on the fingerprint pattern obtained.

Therefore, (GTG)5 was capable of distinguishing C. michiganensis subsp. michiganensis
from other subspecies, however, it is more intrinsically difficult to use and interpret than

BOX-PCR due to the large number of weakly amplified bands.

Reproducibility of (GTG)5-PCR fingerprints and definition of clusters. If the
fingerprint patterns obtained from the (GTG)5 primer are to be useful in differentiating
strains of C. michiganensis subsp. michiganensis the amount of variability due to artifacts
of the PCR reaction must be gauged. Two experiments were performed to test this

experimental variability. In the first experiment, the reaction mixtures were held

58

sid
sid
stu

Similarity Index

22
20 60 100 0.5 Kb
I t I 1| [.1 I”! O .1 1 C.m.wbsp.nebraskensis
r l Illa I ... 2
D an O 0 r . .
’ “I , . . . t l i C. m. subsp. tessellarius
.4 p a On. I! ‘ . O 5 -C.m.mbsp.s¢pedomcus
pn- au I I u an 0- O n 6 -C.m.subsp.insidiomm
II t I. 0'. "~0* ' 7 D
I 0. I“. It ‘00 =0 8
ect.!” t 0 ~ 9 C] 0
0° h H O t 10 ‘3
l .0 .. 0 '0 0‘ C 11 D g
tutti-DINO“! 't' 12D 3'
. Mt. “I 13
r l .00.. '~ ' 14 E E
D m 0'0... N ‘ ‘ 15 a
1 ~' in H O“ C 0 16 a S
r .0 0H. 49* 17 E
. I .0 H at G 13 E
I II C 19
0 0' HO '0 O NEE;

 

C. m. subsp. michiganensis strains

Box-PCR type E =3 HEEEEEB =E

Figure l. Cluster analysis (UPGMA) of (GTG)5-PCR-generated genomic
fingerprints of strains of C lavibacter michiganensis subspecies. The gray bar
above the fingerprint patterns ranging from 2.2 to 0.5 kilobase pairs represents

the area of the banding pattern used to generate the dendrogram shown on the left
side of the figure. The columns of non-pattemed and patterned boxes on the right
side of the figure represent the results from the (GTG)5-PCR obtained from this
study and the BOX-PCR genotypes previously designated by Louws et al. (1998),
respectively. Lanes 1 to 6, (GTG)5- PCR generated fingerprint patterns for C.
michiganensis subspecies including subsp. nebraskensis strains CIC 12 (lane 1)
and CIC 16 (lane 2); subsp. tessellarius strains CIC 21 (lane 3) and CIC 22

(lane 4); subsp. sepedonicus strain CIC 4 (lane 5); and subsp. insidiosum strain
CDA 4 (lane 6). Lanes 7 to 20, (GTG)5- PCR generated fingerprint patterns for
C. michiganensis subsp. michiganensis including representatives of the designated
as type C strains on the basis of BOX-PCR (Gil-7904 (lane 7), GH-71033 (lane 8),
Cmm56 (lane 9), Cmm122 (lane 10), GH-71182 (lane 11), and le4 (lane 12)),
type B strains (Cmm118 (lane13) and Cmm14S (lanel4)), type D strains (Cmm288
(lane 15), and GH-7137P (lane 16)), type B strains (CA-5 (lane 17) and Cmm30
(lane 18)), and type A strains (Cmm936 (lane 19), and CmmT13 (lane 20)).

59

constant i
primers.
were Obit]
erperime:
Tl
clonal str:
PC R reac
the cluste
alone cou
t'anabilit;

taken into

Genetic r
MICillgap

constant and five individual colonies of one strain were amplified with BOX and (GTG)5
primers. In the second experiment five different reactions with BOX and (GTG)5 primers
were obtained from the same colony of the strain assayed in the first experiment. These
experiments were performed on a strain representing BOX-PCR types A, B, C, D and E.
The overall amount of variability observed in these reactions indicated that each
clonal strain showed a rather large amount of experimental variability based solely on the
PCR reaction protocol, and not due to the genetics of the strain (Figures 2 and 3). When
the cluster analysis (UPGMA) was made for single primers, the experimental protocol
alone could account for up to 25% of the observed variability and up to 20% for the
variability observed in combined primers (Figure 4). This experimental variability was

taken into consideration when strains were compared for genetic diversity.

Genetic variability measured by (GTG)5-PCR fingerprint analysis in strains of C.
michiganensis subspecies michiganensis designated as BOX-PCR Type A.

In a preliminary study, 23 type A strains with greatest overall diversity, including
BOX-PCR fingerprint type, origin of recovery, etc. were selected to represent the 73
strains listed as type A strains in Table 1. These 23 strains were amplified with BOX and
(GTG)5 primers and the resulting banding patterns were analyzed for differences through
cluster analysis (UPGMA; Figure 5).

As expected with the BOX primer, all BOX-PCR type A strains clustered as one
large group when the experimental variability baseline was subtracted (25%). Because
the strain appeared so similar, an outgroup consisting of 6 type C strains was included in

the analysis to provide a similarity index less than 75% (Figure 5A).

60

Figure 2. Cluster analysis (UPGMA) using product moment correlation of BOX-PCR-
generated genomic fingerprints of Clavibacter michiganensis subsp. michiganensis
strains used to define the experimental variability baseline. The gray bar above the
fingerprint patterns ranging from 2.0 to 0.4 kilobase pairs represents the area of the
banding pattern used to generate the dendrogram shown on the left side of the figure.

The percent value next to each BOX-PCR type (patterned boxes) corresponds to the
similarity value observed for the BOX-PCR—generated fingerprints using the same culture
as the DNA template. The average of the similarity values observed among the strains
used was considered to be the experimental variability baseline (approximately 75.0%)
for the analysis of the following dendrograms.

61

%
0.

780%
76.0%
750%

Kb

 

4
.W.onaaan. 2:222:22
IIIOIIIOIOIIa so. . p e . nigoooalilllio . Ian-IIICIIIIIIII. o

M Muicutnoonio‘oooeaflont dutIuIOO Olliuooain no»... a.
unuuuuuuuuumuu otourounuuuunnuu mm u unnuuuuuunumuuuu muuunu
aan...“nun-«tumuuuuuu.unun.u..-.~uuo...u-.....-u- --nunn»
until! . IMIIHU'I-.I.IHIII lil'.‘...l-l-
“MI MN“. tflm lilo: roll
a IIHHOHHJ . aNIOHHHMmh t III
0 I.-ltl-.ltil...- 1! -rllI I-‘.Illllfl t .t.“”'. I l.'rhvhlll
Z t
i I t 5””.t “ . r
Fullbbl'flu.bfliihhhhrn~ it DI.D.IDIL3IID‘HDIPDLU
OI” 0.0m - OHM! .. “00.1...-I t Hr-

 

:I.’

 

nan . - .

t..- 0 iv MOM.UCUIII-O-H- .. 3.... ...... U
new“..n.....n... -------n.- mama.
N uh

«v.1
noun-aun-oqnuuu-uueu "mum

‘ IMII

WMWWWWWW WSWWNWWWWWWW.WWWW. WWW.WWWWMW--WWWWWWWWWW WWWWWWWW
m t r at to- are a .-
. - p. » err . a-.. .- . - ..,.O o -.

 

Similarity Index
80 100

 

 

 

C. michiganensis subsp. michiganensis BOX-PCR types

[IIIIZF‘3

-E

'=A

D

62

Figure 3. Cluster analysis (U PGMA) using product moment correlation of (GTG)5-
PCR- generated genomic fingerprints of Clavibacter michiganensis subsp. michiganensis
strains used to define the experimental variability baseline. The gray bar above the
fingerprint patterns ranging from 2.2 to 0.5 kilobase pairs represents the area of the
banding pattern used to generate the dendrogram shown on the left side of the figure.
The percent value next to each BOX-PCR type (patterned boxes) corresponds to the
similarity value observed for the (GTG)5-PCR-generated fingerprints using the same
culture as the DNA template. The average of the similarity values observed among the
strains used was considered to be the experimental variability baseline (approximately
75.0%) for the analysis of the following dendrograms.

63

Similarity Index

0.5

2.2

100

80

 

% % % %
5 5 0. 0
1 I

7 8 m m

      

II I...‘ C....."l..... ...n.
.- i. r . at... :,

 

 

-u» u.n-uuummnmm"unuut.e-t-t , .

a. «nun-unuuouuumuuouuuuuuuuummmnuuu.uuuunuuuuouunun"nunnuoon
ll 1-00.010Nn0M0 I." h000u IflIIrrfllI 000! ”0000000010
”Mumim.flmtti000OOIIO d It. 00kt Ivltt HHWHHHMMiH.IM “NIH H

u . .tmtt . t of

IIU>IIINDrlnrrFPDDIDBmunhu~:WFDIIvDD OI in m

0 0.) n. 00.4%. In..- 1.0!-IWI...’ ”HH)‘ 4}. t”

I .t: r0r|vvll|l0tnce15000a1. l.erte
«mama-noo-uvcamnnnmn-mnumuumn unuuummmmmn muumumwwuunummmmmuo
it . t .r n

to .4”! in I - 0. .un OOH-t t.tu In N t.t t u

WWWWWWWWWWWWW.aW.. WWWWWWWWz moemom....m -wwmmmmmmnuwonm.W.WW

‘ 0 I '0}..- H

I x .01 000ri
, . t u p -1 . In“ to n.. o

 

L

EEE

 

C. michiganensis subsp. michiganensis BOX-PCR types

E§§§§§§=B
=E

=A

D

Figure 4. Cluster analysis (U PGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains used to define the experimental variability
baseline. The gray bars above the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to
0.4 kilobase pairs represent the area of the banding patterns used to generate the
dendrogram shown on the left side of the figure. The percent value next to each BOX-
PCR type (patterned boxes) corresponds to the similarity value observed for the linearly
combined (GTG)5-PCR and BOX-PCR-generated fingerprints using the same culture as
the DNA template. The average of the similarity values observed among the strains used
was considered to be the experimental variability baseline (approximately 80.0%) for the
analysis of the following dendrograms.

65

finflafiwlmhx

0A Kb

10

05

22

100

80

 

"unnunuunu wumuuuonumg
.HIHIII.I”IIWNNI. -IIH 00"0000-0

" '..-... 0..-..Ih

r ”.0500 hit
Ile.Mlunflh.|ltCOCI”
"Mm-ummMuMIIMMmmmuum
mwmwmuummumw a...

unnuuuuuu uuuu

 

IIIIIIII IIIIIIII”-

.00 II. .hba. ‘ .

Lorna- uuuuo,- ..... "nun”,
II.

2. -.pal-untrufluuz3.3.233 .1".

0.00l000l00I0100 Ix

I
a

0.-" . ll. I
yW.’.~|h.f'WWUflI.Il”rIIkF..F0II : n
o, 00 0 I s -1 .I It 0

I

nounnumwmnauuanunnnammmnnou-ouau m.-wmwm

DWI-n
”W

I Hm
.0“
Hm
I!" n
“00

\ “fig.
“”0!“

' .000
IN"
”to I
Ibhn
M
“u-
”MOI
“all
I000” v
DIM
NM
NM

”.mwwwmmmmmmm

 

C. michiganensis subsp. michiganensis BOX-PCR types

A

=E

D

66

With (GTG)5 , the BOX-PCR type A strains clustered into three major group afier
the experimental variability (25%) was taken into consideration (Figure 5B). The same
type C strains were used as an outgroup to define the branch of a cluster. When both
primers were analyzed simultaneously, four groups were differentiated after the
experimental variability (20%) was taken into consideration and each subgroup was
designated as types A-l, A-2, A-3, and A-4 (Figure 5C).

Looking at these four groups in more detail offers some interesting insights into
the genetic diversity of A type C. michiganensis subsp. michiganensis strains (Figure 6).
The A-l group was composed of strains primarily from Michigan isolated from
processing tomato plants in the 1990’s. The A-2 group was composed of 2 strains
(Cmm355 and Cmm360). In the BOX-PCR reaction, these two strains were found to be
similar to each other, and were the most distinct strains (lanes 18 and 19) in that single
large group. While both of these strains were found in the same group when (GTG)5
reactions were analyzed, other than year of isolation from diseased tomato plants, no
other obvious pathological or epidemiological characteristic united these two strains or
set them apart from other strains. Based on (GTG)5 and BOX-PCR fingerprint patterns,
these two strains would be seen as genetically similar, yet, each was recovered from
completely distinct types of tomato, one a fresh market variety and the other from a
processing variety, from fields 220 km apart.

The A-3 group is composed of a single representative strain, Cmm302, which was
also part of (GTG)5 group 2. Looking at the BOX-PCR analysis, this strain grouped with
the single large group but was placed in that group with strains isolated in 1993 from

processing varieties recovered from the southwestern part of Michigan. Cmm3 02 was

67

isolated in 1994 from a diseased processing tomato plant recovered from southeast
Michigan. Another representative strain, Cmm134, isolated from plants recovered from
that location in the same year grouped in the large A-l group in the combined PCR
analysis.

The A-4 group consisted of 3 representative strains that were considered (GTG)5
PCR group 1. These strains were also contained in the single large group in BOX-PCR
analysis and grouped together with another isolate CmmT56. CmmT56 had an identical
BOX-PCR pattern with the A-4 isolates, but when (GTG)5—PCR fingerprints alone were
analyzed, CmmTS 6 was not grouped with the A4 group. It is easy to see why they were
not grouped together (Figure 5). A prominent large band approximately 2 kb in size was
present in the A-4 isolates and lacking in the CmmT56 isolate. Therefore, where BOX-
PCR failed to pick up a distinction, the (GTG)5-PCR reaction was able to demonstrate a
strong difference in genotype. When taking into consideration the epidemiological
differences, both the strains making up the A-4 group and CmmT56 were all isolated
from in 1993 from southwest Michigan farms planted to processing tomato varieties.
CmmT56 was isolated from a different farm than the other strains.
Genetic variability measured by (GTG)5-PCR fingerprint analysis in strains of C.
michiganensis subspecies michiganensis designated as BOX-PCR Type B.

Seventeen type B C. michiganensis subsp. michiganensis strains were amplified
with BOX and (GTG)5 resulting in banding patterns analyzed for differences through
cluster analysis (U PGMA; Figure 7). The results of the B type strains were different than
those with A type strains as BOX-PCR resolved the B type strains into at least three

subgroups. With (GTG)5, alone, all BOX-PCR type B strains clustered into one large

68

Figure 5. Cluster analysis (UPGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains designated as type A. The gray bars above
the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to 0.4 kilobase pairs for (GTG)5-
PCR and BOX-PCR, respectively, represent the area of the banding patterns used to
generate the dendrograms shown on the left side of each panel. Panel A, analysis for
BOX-PCR fingerprints; panel B, analysis for (GTG)5-PCR fingerprints; and panel C,
analysis for the two linearly combined (GTG)5-PCR and BOX-PCR fingerprints. The
vertical lines bisecting each dendrogram represent the experimental variability baseline
used to define a cluster: 75.0 % for the dendrograms in panels A and B, and 80.0% for
the dendrogram in panel C. The designation of groups for each analysis is represented on
the right side of the panel

69

imilarity Index

20

 

 

_J_ _

 

 

 

 

IOII. 0.
..un-..

00“00
N“HIWHHNN000fl0IIWNIO0~0000”00

3 CI 0'

00... ........ It htneauill In
octoulbuenbebuuencn

I..-

III 1

«c.» u. 00.nn .5

.r ........n. . ..

n... _. a. ..
u-....xnr.§..---.:.:§. an

. .IJI‘

.... mo. .-

.. .mu- .muuom
. .i.. 000

I 1. Ibhnln .II..

:fifie

 

w
L

Kb

IIoIIOIOOCI¢U.oe.

.mmammmuuunmauauawr....t

III. III 00000000000

 

00000I0I~000000II00III0n.0000
III ill

.1. IHIIulii.
.I’r” ' 0 I

. .IODOIOIIOOI000II000. .IMI

 

5 Oudlm I. . in I
o. I. '00-: Iv: IaIIoIIIIIII.0I.0I
333.332.33.13-£3.
- -... . . - ..H.. .- um.
m.nmm..., .. u‘.uuuu..muuuuw
2 -8.-. -- T- -
n4

 

 

20 04 Kb

05

25

nzmuumwvnuo w

 

 

 

mum.

 

II
0000IIIIO0III000000D: 0v IMI
I

it u n. ‘I II
.n. .a ..... trim .v .. I
I 0 . I u0l0-0 I Os”

t.IvII00000II 000-00.

33.32.33.333}?-5.

.«4 .v «ltd-..

 

. 0 o H!
.c.. Itllxtt U. MMIHII
-. aw.41 on-c.~ann “Ola
w. i .0 I m

II

no . III um
...u.. .7... u.” in.

 

70

Similarity Index

60 80 100
«WW

NN
—-N

O

 

—— I" N
—~4“0~a~a

I4
15

 

l

 

 

 

 

i

mbwn—c‘

 

 

 

 

 

Sli 9.. I I I' V'| I’

Cmm T2
Cmm T1
Cmm T24
Cmm 936
Cmm 671
Cmm T56
Cmm9
Cmm 357
Cmm 294
Cmm 299
Cmm T12
Cmm T35
Cmm T45
Cmm T20
Cmm T66
Cmm 134
Cmm 672
Cmm 355
Cmm 360
Cmm 302
Cmm T74
Cmm T13
Cmm T63

93
93
93
93
91
93
88

ouu§amm§mwwwujaafimmmmuu

peppgpppvpptpgpgppppppp

fruit
fruit
stem
nk
fruit
nk
nk
fruit A-l
fruit
fruit
fruit
fruit
fruit
stem
seedling
nk
fruit A-2
nk
seedling A-3
fruit

 

fruit A-4

fruit
9..
:3
o
I—
29
:3
o

 

Figure 6. Clavibacter michiganensis subsp. michiganensis strains designated type A
used for the cluster analysis (UPGMA) of the linearly combined (GTG)5-PCR and
BOX-PCR-generated genomic fingerprints in Figure 5C. The vertical line bisecting
the dendrogram represents the experimental baseline (80%) used to define the groups.
The designation of groups for the analysis of the linearly combined (GTG)5-PCR and
BOX-PCR-generated fingerprints, and six strains used toform an outgroup ( Cmm12
(lane 6), GH-71182 (lane 1), Cmm56 (lane 2), GH-71033(lane 3), Gil-7904 (lane 4),
and Cmm 14 (lane5)) are indicated on the right side of the figure.

71

group. Because the relationship was so similar, the same 5 type C strains used in the type
A strain analysis were used for the cluster analysis (Figure 7B).

When both primers and an 20% experimental variability baseline were used in the
analysis, four subgroups with the B-type strains were resolved and are referred to as B-l ,
B-2, B-3, and B-4 (Figure 7C).

These subgroups, based on the combination of primers, appear to have some
epidemiological merit in regard to the time and type of tomato plants grown (Figure 8).
For example, two strains recovered from diseased fresh market tomato varieties on a farm
in Michigan in 1986 were placed in the B-1 group along with three strains isolated from
diseased fresh market tomato varieties growing in northwestern Ohio in 1987 and 1988.
These strains were part of a large continuing epidemic (1984-88) that initiated the tomato
bacterial canker workshop in 1984 (now referred to as the Tomato Disease Workshop).
All 5 strains grouped as B-2 were isolated from California and sent for analysis indicating
a geographic relationship. Similarly, 4 strains collected fi'om southwest Michigan in
1991 were placed in group B-3 along with a fifth strain isolated from fresh market tomato
in southwest Michigan earlier in 1987. The strain representing B-4 was isolated from the
same location in the same year as the 3 strains in B-3 and the genetic distinction of this
strain is in doubt as experimental variability should be taken into consideration due to the
weak differences separating B-3 from B-4.

Many of these same groups can be seen in the BOX-PCR analysis if the
experimental variability baseline is ignored. For example, the California isolates group
together in the BOX-PCR analysis but do not fall out as a unique group if the

experimental variability is taken in consideration. Interestingly, the 1991 strains isolated

72

Figure 7. Cluster analysis (UPGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains designated as type B. The gray bars above
the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to 0.4 kilobase pairs for (GTG)5-
PCR and BOX-PCR, respectively, represent the area of the banding patterns used to
generate the dendrograms shown on the left side of each panel. Panel A, analysis for
BOX-PCR fingerprints; panel B, analysis for (GTG)5-PCR fingerprints, and panel C,
analysis for the two linearly combined (GTG)5-PCR and BOX-PCR fingerprints. The
vertical lines bisecting each dendrogram represent the experimental variability baseline
used to define a cluster: 7 5.0 % for the dendrograms in panels A and B, and 80.0% for
the dendrogram in panel C. The designation of groups for each analysis is represented on
the right side of the panel.

73

Similarity Index

20

80

02 5 0.
l23456789mllBMIMWwwmzn

00000000 0000000o0e000

nu.uuu-o..o-u0n. ones
IIII--IIIIIIIO .

II 000000000I0000 00"
I

II -IIII ...... II.-IIIH

00II000III000000000000
000000Innu00l0lln00U0I
II.

I”.
0

In. . 0..
D I
l u n. 0 00‘
..I . MIIM
IIIIIOOOIIQIOIOOC MI. .
_

H0

.-

0-

 

00-000 I 0 00------ 0
..I I II oI-I
,III I...III-Iv-I0I I
”00-000000000000000 0
000i .0 y I III W
. .I I10 I
-Iun-.n-.. . I-

 

 

Kb

 

 

52
lémn89753l21 MM
00000 OIOIIIOI II
0.0.0 00n-000000
III VIII II0III00 00
00000000 I0000000
I

00ll0000 000000000000-
000000000000000000000.
I I I
00a4In000000000000000.
0-000-00Ih0000000
00IIII00 .00000

I“.
0

5

lord

Ii: IIIII

...-I IN
0000IIIIIIICQI.‘III *
I

 

0000! «00000000000 H

I I0 I IIQOIII 0

0 IIIII I II0IIII|00"I00

IIIIIIIIIIIIIIIIII III

F u. n u I

I I IIII *
I II I I0

I-II UI I

 

m4Kb

20
l

05
|

 

 

0 1746523

91867 llllllll 2354mm

.lllOOOIIIOIDIDIIIOIII
n00000~00000m0

000000

0000

00000000000 I

I I0IIII00000II

a.”
0
00

III

I
wnIn
II0
.. I
I000

00000000000000I000000-

 

 

IIM0I0H00000000000000I
IIIIIIIIIIIIHH-IH II»
Il0000I00000 I I”.
I
III: III 0
. l
m u I u 0 0!
I -..-I II Ih 00
IMIMI 10101.. M I .“.
I000II II 0
I 000.100-000-00000 0"II
I 0 IIIIIII II I 00
I I 00000000100 0“I000
000 IIIII 0"0.IMI.IWM0 I000
0 I00
muumwummm :0. noun-I MI
I I” M

74

Similarity Index

60 80 100

sum mmYmr—W

 

  

 

9 Cmm 69 OH 87 17 FM stem—‘
0 Cmm 12 MI 87 25 FM fruit
Cmm 35 MI 86 1 FM stem B-l
Cmm 77 OH 87 18 FM fruit
Cmm 68 OH 87 17 FM fruit
Cmm 30 MI 86 1 FM fruit
Cmm 226 CH 88 21 nk nk
CA-01 CA nk nk FM st
CA-Y CA nk nk FM stem [34
CA-10R CA nk nk FM stem
CA-5 CA nk nk FM fruit
CA—5R CA nk nk FM fruit
Cmm 621 MI 91 4 nk nk
Cmm 623 MI 91 4 nk nk I B-3
Cmm 620 M1 91 4 nk nk
Cmm 668 MI 87 3 nk nk
4 Cmm 619 MI 91 4 nk nk = B-4
8
9 2%
$3
1 29
8

Figure 8. Clavibacter michiganensis subsp. michiganensis strains designated type B
used for the cluster analysis (UPGMA) of the linearly combined (GTG)5-PCR and
BOX-PCR-generated genomic fingerprints in Figure 7C. The vertical line bisecting
the dendrogram represents the experimental baseline (80%) used to define the groups.
The designation of groups for the analysis of the linearly combined (GTG)5-PCR and
BOX-PCR-generated fingerprints, and five strains used toform an outgroup (GB-71182
(lane 18), Cmm56 (lane 19), GH-71033 (lane 20), GH-7904 (lane 21), and Cmml4
(lane22)) are indicated on the right side of the figure.

75

from Michigan cluster into BOX-PCR group 1 and the Michigan strain isolated in 1987,
that the combination of primers grouped with the 1991 strains, was separated and placed
with strains from Michigan recovered in 1986-87. This appears to be a more logical
association, but it also indicates a potential hazard in interpreting similarities and

differences among strains.

Genetic variability measured by (GTG)5-PCR fingerprint analysis in strains of C.
michiganensis subspecies michiganensis designated as BOX-PCR Type C.

In a preliminary study, 22 type C strains showing the greatest overall diversity,
based on BOX-PCR fingerprint patterns, origin of isolation, etc. were selected to
represent the 73 strains listed as type C in Table 1. These 22 strains were amplified with
BOX and (GTG)5 primers and the resulting bands were analyzed through cluster analysis
(UPGMA; Figure 9).

When the experimental variability baseline was subtracted (25%) from the BOX-
PCR banding pattern, five subgroups were obtained (Figure 9A). With (GTG)5, the
BOX-PCR type C strains cluster into 10 subgroups when 25% experimental variability
baseline was taken into consideration (Figure 98).

With both primers, 15 subgroups were observed with the type C strains using an
20% experimental variability baseline and each subgroup will be referred to as 01, C-2,
C-3, C-4, C-5, C-6, C-7, C-8, C-9, 010, C-1 1, C-12, C-13, C-14, and C-15 (Figure 9C).
The majority of the subgroups are represented by a single strain. From the subgroups that
consist of more than one strain, these strains usually have some interesting relationships

when characteristics other than PCR banding patterns are used to group the isolates. For

76

Figure 9. Cluster analysis (UPGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains designated as type C. The gray bars above
the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to 0.4 kilobase pairs for (GTG)5-
PCR and BOX-PCR, respectively, represent the area of the banding patterns used to
generate the dendrograms shown on the lefi side of each panel. Panel A, analysis for
BOX-PCR fingerprints; panel B, analysis for (GTG)5-PCR fingerprints; and panel C,
analysis for the two linearly combined (GTG)5-PCR and BOX-PCR fingerprints. The
vertical lines bisecting each dendrogram represent the experimental variability baseline
used to define a cluster: 75 .0 % for the dendrograms in panels A and B, and 80.0% for
the dendrogram in panel C. The designation of groups for each analysis is represented on
the right side of the panel.

77

0.4 Kb

 

 

 

 

2.0

 

80

0.] 2 3 4 5
0 123456 .8544444
2 345 123 4 567891 CCCCCCCCCCCCCCC
11J1||:|.=.=J b _i___IJ_.J_|_:I._:_ b E: Er: =___ __n_ _
1234se7s9munnmumnmw 22 K memnmowwsnnnsanmuzavmu K enmwlzasamun7sn9muumuw
I000. IIIIQIIIIIIII0-, IIIIIIIIIIIICI IIIOIII A] IIIIIIOCIII CHIIIIIIIII
m: magnum. .....1. $2.283; udfiflm o ....u. ”am $3.32.
I IIIIIIIIII0000 ....NI 0.100000'0000 I00! . I"0~0.I .IMIHIIIIIWIII
ii- .413 n nsmumciucuct--o -.. oniiiiiiiciuum
nunnuu..- -....umuun-nuu «tau..- ...wiihuuiu ...uu-mu-o--..-.... 3mm”
'I-..""...'.'hhh“ 0 I000. I0 0. '....I0.. 0.”-I..'...

I I [I0 I I I I “I 0 0 I I00 I I II

I I IIIIII IIII H.011I0IIII II 000 14.1". l 00 00 IIIIII
m: -IMH bunnmIMIunII IIOIIIInIIQII IIMm,u.I 0L Olauna - ,IINIIIMIQmIH

II . I
2

up a. . u. y. I: . up- or
. H113... .3» - 5. ,mm.mu:9...mfl-fi H- 5 23.; ... . 1.21.3.
- IIhmIoohuu Mn--IIII 0 II IIIIMIMMIMHIHMI II“ 0.1 ”it- -IIIMWOIMIIMI In

IIIW” I I” I ION I II III I I I‘IIII I I“
--m.....:.m... 3.3.2:.-- _ we..- Esai
mm- qnm- ”um..." 3. Hanna. ..... ”Jun «w .W .. aminuu."
mmmunmuwuunuuuma . wmu mmmmumm muuummu .mmmunw nmemmnwmuummummuummu
«upon. «ammummmum- "pm on uuuuuiouooo ummio w uummuuuuuhmww- nmmu
I'MIIII. I 01000 2 00II0I I 000000 ‘0“I 2 II 00-nII II I 00000

2 I b 2|.
I l I

 

 

Similarity Index

 

78

Similarity Index

 

 

 

   
 
 
 
 
  
 
 

60 30 100
$9319 Origin Yea: mum; Emmy m g:
Cmm 33 MI 87 2 P stem - C-l
Cmm 14 NC 84 23 nk nk — (3.2
Cmm 236 MI 88 3 FM fruit — (3.3
Cmm 218 MI 88 3 nk fruit — c_4
Cmm 21 NC 84 22 nk stem —
Cmm 237 MI 87 3 nk stem C-S
Cmm 56 MI 87 2 FM fruit 7‘06
GH-71428 China nk nk FM seed :1 C-7
GH-71182 China nk nk FM seed
Cmm 954 MI 92 5 FM stem'—‘ C -8
Cmm122 OH 88 20 FM stem_
Cmm 117 CH 88 20 FM nk - C-9
GH-7904 Morocco nk nk FM seed - C-lo
GH-71033 Chile nk nk P seed - C-ll
l GH-211 China nk nk P
GH-197 CA nk nk P stem C-12
0 GH-7902 CA nk nk P stem
3 GH-212 China nk nk P seed (:43
2 GH-202A Chile nk nk P seed :1
4 Cmm 27 OH nk nk nk nk :] C-l4
5 Cmm 15 MI 87 2 P stem

:1
x
:I
x
I
‘9
~
In

6 Cmm 26 OH 87 19

Figure 10. Clavibacter michiganensis subsp. michiganensis strains designated type C
used for the cluster analysis (UPGMA) of the linearly combined (GTG)5-PCR and
BOX—PCR-generated genomic fingerprints in Figure 9C. The vertical line bisecting the
dendrogram represents the experimental baseline (80%) used to define the groups. The
designation of groups for the analysis of the linearly combined (GTG)5-PCR and BOX-
PCR-generated fingerprints are indicated on the right side of the figure.

79

example, two strains from California are grouped together and two strains from China are
grouped together although another Chinese isolate is not grouped with the other two
Chinese isolates (Figure 10). Not much is known about these strains however, the two
grouped together are from fresh market varieties and the single isolate is from processing
tomatoes. The strains from North Carolina were separated by fingerprints generated by
all PCR primers, even though they were isolated in the same year, these strains were not
from the same field as they came from different counties.

Genetic variability measured by (GTG)5-PCR fingerprint analysis in strains of C.
michiganensis subspecies michiganensis designated as BOX-PCR Type D.

F ifieen D type C. michiganensis subsp. michiganensis strains described in Table l
were amplified with BOX and (GTG)5 and resulting bands were analyzed through the
cluster analysis (UPGMA; Figure 11). When the experimental variability baseline was
subtracted (25%), the BOX-PCR fingerprints broke into 2 subgroups, where isolates were
found to be clustered in correlation to their geographic origin. One group was from two
farms in central Michigan and northeast Michigan and the other group was from foreign
locations such as India and Morocco (Figure 11A). Within the large BOX-PCR group,
isolates from Michigan were clustered into smaller subgroups. One of these subgroups
contained D types from the two different farms mentioned above, both isolated in 1994.
The other Michigan strains subgrouped away from the 1994 isolates were isolated in
1997 from the northeast farm.

With (GTG)5, 3 groups were obtained afier a 25% experimental variability
baseline was established. Isolates from Michigan where represented in two of the three

groups. The strains isolated in 1994 grouped together and with two other Michigan

80

strains isolated in 1997. The foreign strains split into two similar groups with one 1997
Michigan strain included in one of the two groups (Figure 11B).

With both primers taken into consideration and after the experimental variability
was subtracted (20%), 3 subgroups were obtained and each subgroup will be referred to
as D-l, D-2, and D-3 (Figure 11C).

The foreign strains fall into D-3 and no Michigan strains are positioned with
them. The Michigan strains isolated in 1994 are found in D-2 alone. All the other
Michigan isolates isolated in 1997 can be found in D-l. All the D type strains have only

been found on fresh market tomatoes north and northeast of East (Figure 12).

Genetic variability measured by (GTG)5-PCR fingerprint analysis in strains of C.
michiganensis subspecies michiganensis designated as BOX-PCR Type E.

Ten type B C. michiganensis subsp. michiganensis strains described in Table 1
were amplified with BOX and (GTG)5, and resulted in bands that were analyzed through
the cluster analysis (UPGMA; Figure 13).

When all primers were used and the experimental variability was subtracted
(25%), all type B BOX-PCR strains formed one large group. Because the genetic
similarity was so similar, an outgroup group consisting of 6 type C strains was added to
the analysis to provide a similarity index of less than 75% (Figure 13).

At a higher similarity index (88%), two subgroups can be defined, although any
relationships can be made because every isolate was collected at location #4 from
Michigan with one exception. The second group contain two isolates from location #4
and the isolate recovered at location #20 from Ohio (Figure 14). This Ohio strain was

isolated in 1988 and no other type B strains were recovered until 1996 in southwest

81

Figure 11. Cluster analysis (U PGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains designated as type D. The gray bars above
the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to 0.4 kilobase pairs for (GTG)5-
PCR and BOX-PCR, respectively, represent the area of the banding patterns used to
generate the dendrograms shown on the lefi side of each panel. Panel A, analysis for
BOX-PCR fingerprints; panel B, analysis for (GTG)5-PCR fingerprints; and panel C,
analysis for the two linearly combined (GTG)5-PCR and BOX-PCR fingerprints. The
vertical lines bisecting each dendrogram represent the experimental variability baseline
used to define a cluster: 75.0 % for the dendrograms in panels A and B, and 80.0% for
the dendrogram in panel C. The designation of groups for each analysis is represented on
the right side of the panel.

82

Similarity Index

0.4 Kb

2.0

100

80

 

 

0

1.2345

123456789 111111

0

H“

.... . ..II "0
..

110000 000--'

NHNNflhHI

00

”HIMNN
*0..0

00000000000...

0014-01.

0000000000 I ..I 0.. I. z

“H“H'U PH-“-----

III
II

mee

 

-0- 0
0 “00-
..“.~
.N .
I-.. a
“HUN"
0-.
00-00
-0 0-

 

 

 

 

 

h
K

0.5

 

l 2

 

_
451 3
ll

2

0

1511342816 79 II
000 III
..fl '0“.........
“MMIMOII IIIII MW

Nflflflflflflflm

0 I1 I. .. 0 ..
------000000000.

0 II-

nu"

---“--“N "flu" ...... 5"

000

HI“ I

IIHmflI
who I
!II:
II

.‘

 

 

 

 

 

 

 

 

 

 

 

 

 

l

D

2

 

0.4 Kb

=l_

3

_III_

045123

ER“

0000

324816759 111111

.....'....." I n W

----

unwuu--IIm~IIII

flflfluflfluflm

00
-0

----
-.0***

00000.0 0... 0 0

00000

00000000000 1 ..I I. 0

n. t u'“-0--0

"H“.

2.0

 

0.5

2.2

0"

0..

 

 

000-

 

 

lemewmmmm

 

 

 

 

 

83

Similarity Index

 
  
 

 

 

 

 

 

 

60 80 100
3 Bet1-5-1 MI 97 14 FM stem —
2 Bet3-5—1 MI 97 14 FM stem
4 Bet3-5-2 MI 97 14 FM stem
8 MF1-5-1 MI 97 14 FM stem D-l
l Bet1-5-2 MI 97 14 FM stem
6 MF1-5-2 MI 97 14 FM stem
[7 MF3-5-1 MI 97 14 FM stem
x————5 FM1-4-1 MI 97 14 FM stem _
{9 Cmm 288 MI 94 14 FM fruit 3 0,2
10 Cmm 284 MI 94 13 FM fruit
14 GH-7137P India nk nk FM seed —‘
FE” GH-7219 I Morocco nk nk FM seed
II GH-7388 l Morocco nk nk FM seed 9'3
12 GH-73880 Morocco nk nk FM seed
{13GH-7219Pi Morocco nk nk FM seed _

 

 

Figure 12. Clavibacter michiganensis subsp. michiganensis strains designated type D
used for the cluster analysis (UPGMA) of the linearly combined (GTG)5- PCR and
BOX-PCR-generated genomic fingerprints in Figure 11C. The vertical line bisecting
the dendrogram represents the experimental baseline (80%) used to define the groups.
The designation of groups for the analysis of the linearly combined (GTG)5-PCR and
BOX-PCR— generated fingerprints are indicated on the right side of the figure.

84

Figure 13. Cluster analysis (U PGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR-generated genomic fingerprints of Clavibacter
michiganensis subsp. michiganensis strains designated as type E. The gray bars above
the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to 0.4 kilobase pairs for (GTG)5-
PCR and BOX-PCR, respectively, represent the area of the banding patterns used to
generate the dendrograms shown on the left side of each panel. Panel A, analysis for
BOX-PCR fingerprints; panel B, analysis for (GTG)5-PCR fingerprints; and panel C,
analysis for the two linearly combined (GTG)5-PCR and BOX-PCR fingerprints. The
vertical lines bisecting each dendrogram represent the experimental variability baseline
used to define a cluster: 75.0 % for the dendrograms in panels A and B, and 80.0% for
the dendrogram in panel C. The designation of groups for each analysis is represented on
the right side of the panel

85

 

Similarity Index

Q4Kb

20

um

80

lllllLJllll

a)

 

9.29:0

_ _ _ _

123456789WHHBM5M

.....I.... .1 .. ...
...IOOCCICOQIIOM

I .I a .... ‘ I ‘ .r I -

IIIIIIIIIII“ WI
IIIIIIII -III III
IIIIIIII IIIIIII

I
IIIIIII I I I .--.I
III---IIIIIIIII

 

IIIIII l-.r --
"h” ””IIIfiI. ...."
II I

ltUhIIII P i.
I III I.
i.II I... ~. “.
h llllll IIII WAN-”
mmuuum-m... H .m
i .1. 1H
1 no .. uiuiun
«amen-omw11 I11
-.-
fiMHHHHNNN “WIMNW
III ..IIiI I“ I
u.-u..... a
IIIIII” H l I.

 

JlLLLIIL

 

 

 

 

 

 

 

 

b
K

05

22

ND

80

 

l @353

_ ‘fiIIIIJ

613W254987HBUBMM

III'IIII....1..I
I...IIII.I...O.I

I I ‘ I I I I - I - I --- 1|
IIIIIIIIII“ I"-

IIIIIII --IIIIII
------- --- I I I -I
-II ... I-I I. I .l

IIIIIIIIII.I-.II
---------.I--III

- I -

..I 1! 1 i. i H i. III
IlflflvaflhIluiII. In“:
I - -

H.111
I.Hll
1'1
H.I
1.1
1.1
,4-,
Itv
I
I
a
I
N1uuluvo

I
”I
I
- i- u
I’ -
- I
I- h CI,-”” “I
QMUNMnmmIoi 1 1H
II
' “n “auflufinnf'.4m
IIIII iII I I“
mu ”.3“

 

 

 

 

 

 

 

 

 

 

 

 

20 WIKb

05

12

Km

LLLLllllLLillllll 111111111]

80

 

"b mafia

_ A]

2
123W4657891HEMBM

IIIIIIICIICIIIII
OICCIOICIIHCOCIO
IIIII I

I Ir ' I . . --- 1 I
IIIIIIIIIIuIWI
III---I-- IIIII I
----'---- ....II
I I I l I.-- I. I l

IIIIIIIIII. ”III
----.-‘iI‘II

 

lllllll IIII
”IN” "NII1'qu i
II. I
IF”&I..II .

I I

-... I31... ....
I.

-
. ‘
I I II I -1.
I... I III .-.
I.
I

nun-Ina... . Ii
IIIIII I I
III I III

I I

1 ”I.
mmmmmmmimomiuunm
u uuuuuwuuuvom
iIlI . 10: “II
n. ”use
IIIHII--u I.

 

 

 

 

 

 

 

 

86

Similarity Index

60 8
l I l SI . gun j: I I. 1: il 1.

uluultmluu.

00

 

 

Cmm1 OF
Cmm1F
Cmm 118

”QM

 

0 l

l CmmZS

2 Cmm14S

3 Cmm 308

10 Cmm 488

4 Cmm 31$

6 Cmm1OS
Cmm9F

\O

12

 

ll

 

 

———15

l6
l3

 

 

——[

l4

 

Ml
Ml
Ml
Ml
MI
Ml
Ml
Ml
MI
OH

8888888888

####A#¥Ju&

N
0

PM
PM
FM
FM
FM
FM
FM
FM
FM
nk

stem—i

stem

stem

stem

stem

stem

fruit

fruit

fruit
stem .—

 

outgroup

Figure 14. Clavibacter michiganensis subsp. michiganensis strains designated type B

used for the cluster analysis (UPGMA) of the linearly combined (GTG)5-PCR and

BOX-PCR-generated genomic fingerprints in Figure 13C. The vertical line bisecting
the dendrogram represents the experimental baseline (80%) used to define the groups.
The designation of groups for the analysis of the linearly combined (GTG)5-PCR and
BOX-PCR-generated fingerprints, and six strains used to form an outgroup (Cmml4
(lane12), Cmm12 (lane 11), Gil-71033 (lane 15), GH-7904 (lane 16), GH-7l 182
(lane 13), and Cmm56 (lane 14)) are indicated on the right side of the figure.

87

Michigan. This isolate always groups with two strains from Michigan, but these two
isolates cannot be distinguished from other E type Michigan strains when the

experimental variability baseline of 25% is used.

(GTG)5-PCR fingerprints provides higher subdivision of C. michiganensis
subspecies michiganensis strains.

Twenty-nine subgroups were resolved for the five BOX—PCR types defined for C.
michiganensis subspecies michiganensis strains; five for type A, three for type B, twelve
for type C, three for type D, and one for type B strains (Figure 15). The genetic diversity
with (GTG)5-PCR and BOX-PCR fingerprint patterns of the designated BOX-PCR types
is as follows; C > A 3 B > D > E. Type C strains represent the most distinct group of the
C. michiganensis subsp. michiganensis strains used in this study where these strains
show 50% similarity to the other strains. Type D strains divided in two subgroups are
58% similar. The five subgroups of type A stains are 60% similar, while the type B
strains are 62% similar to the other strains. Type E strains formed a single cluster, where
the strains making up this group were 86% similar most likely because the majority of the
strains were isolated from one location. The cluster analysis of (GTG)5-PCR and BOX-
PCR fingerprints indicate that type D, two subgroups of types C, B, and type E strains
have closer similarity to the type A strains, while thirteen subgroups of the type C strains
have more similar fingerprints among themselves. Overall, the cluster analysis of C.
michiganensis subsp. michiganensis strains using both primers provided a more confident
subdivision of the strains designated into the five BOX-PCR types probably because the
combination of amplified bands consisted of information from more areas in the C.

michiganensis subsp. michiganensis genome.

88

Figure 15. Cluster analysis (U PGMA) using product moment correlation of two linearly
combined (GTG)5-PCR and BOX-PCR—generated genomic fingerprints of eighty seven
Clavibacter michiganensis subsp. michiganensis strains used in this study. The gray bars
above the fingerprint patterns ranging from 2.2 to 0.5 and 2.0 to 0.4 kilobase pairs for
(GTG)5-PCR and BOX-PCR, respectively, represent the area of the banding patterns used
to generate the dendrogram shown on the left side of the figure. The vertical line
bisecting the dendrogram represents the experimental variability baseline (80.0%) used to
designate the genotypes resulting from the linearly combined (GTG)5-PCR and BOX-
PCR—generated genomic fingerprints. The columns of non-pattemed and patterned boxes
on the right side of the figure represent the groups defined by the (GTG)5-PCR and BOX-
PCR-generated fingerprints in this study and the BOX-PCR genotypes previously
designated by Louws et al. (1998), respectively.

89

hnflafitylndex

Kb

0.4

2.0

0.5

2.2

   
 

 

ruluill. __= = __ =

conga-on... .. on... coonlooolnooI-uoonoooooo-Io-.

mmm~MMImm mommmmmmmmnwmn :ooooo. “mum"....wmnanoooonoglonaolooumu con-wurn-ummmw

...... am... 5?...

 

 

E IL—lLEELFELCEEEE

..a.l0lll1OCCOQHI-IIIIIIIOQI

*uununumoon

.....D.‘

 

 

 

H II ‘1." IIIIII I 4 .....a 5...... CI‘IIIIIIIIII
u 110.. 1 1 .11“ al. 1MM101I~1.10.III
. I . . 1 1 IOI--.‘ IIIOI .1H1.
IOQCII1t1 unoqotthIHM-n 1 1 no 1“. “innill I I
In»... 1 1 $1111 .
1 1-111 I '1
IIIIIIIIIIIIHH1IH wnuul 0-0-0-1... 001...! III. 1|..Iu1'1
00.00.001.00300000. 1.1.0.01 In... . u1muuwuun11-uuuqun.uuln..l.DI00MuIMmm.u nuunluzlm
II l...1 I
11 1 up. not. 11: n 1.0 i
. . ,
“a. .,».....-.- .... . . 1w...u--. ...... -.. . u. p. a-
u 1 1 .h11. .r . 1 .11.
0 do . I I. a u .l u .. .. h .. . Q.» . O. 0 Ola I. I
2 . u a . . . .. . .... "a--. ...“. ...;
1-1111OIIOIII.OICI «1.... 1h. .I1l101INMMMNIM‘ 1.00.“..CM .11 .MMMIION= .IN

.nluuo1q

1.1.11.1!

ngna-¢¢......1a¢.-.¢.......ooaaoacn...m-uunu-unguuu.............«muummu1.-..m...m-mm~m

. hr.Wurmmmmmmmmmm ... Tam”
urmmflfl..wwwwmmwfl$5.8. mu.” drawn”. mm mm "an--.

m. a . «. “ma.. - --. -.. N . .
1...... .thhc v. .m..-.-.. 1H: 111Hu1-mh 1hn1I§n DUI I 1

 

III...
...

C. michiganensis subsp. michiganensis BOX-PCR types

 

=B

 

=E

=D

90

DISCUSSION

Most taxonomic studies of Clavibacter michiganensis have primarily focused on
techniques that differentiated the various species of Clavibacter and the subspecies of
Clavibacter michiganensis . With these studies and the recent reevaluation of the genus
Corynebacterium, in which most of these species previously resided, Clavibacter
michiganenesis is now arguably one of the best taxonomically resolved species of
phytopathogenic bacteria (Davis et al., 1984). These prior studies, however, did not
report on the diversity of strains making up the various subspecies, including Clavibacter
michiganensis subsp. michiganensis. If diversity is present, recognizing this diversity
could be useful in terms of understanding the dissemination and epidemiology of the
pathogen around the world or in geographically localized regions.

Until this study was undertaken, our knowledge on the genetic diversity of C.
michiganensis subsp. michiganensis strains was based on genomic fingerprints obtained
with three rep-PCR primers; REP, ERIC, and BOX. With rep-PCR technology, Louws et
al. (1998) grouped C. michiganensis subsp. michiganensis strains into four types based
on a polymorphic region of the BOX-PCR fingerprint pattern and ignoring other
polymorphic regions of the rep-PCR fingerprints. (Five types have since been recognized;
Bell et al., 1997). They focused on the combination of bands obtained with BOX-PCR
that migrated to the 1 kilobase pair (kbp) region. Strains described as type A fingerprints
displayed three bands in this region, while type B and type E had only two bands each
with different migration patterns. Louws et al. (1998) did not report on the newly

recognized type E strains (Bell et al., 1997). Type C and D fingerprints also showed

91

banding pattern differences at the same 1 kbp region with both groups showing only one
band but not at the same migration distance.

Using this narrow strain identification criterion, Louws et al. (1998) suggested
that certain trends in strain recovery might prove useful in the understanding various
aspects of bacterial canker epidemics that have occurred in Michigan and the Midwest
since 1980. For example, they suggested type A strains were found mostly associated
with processing tomatoes varieties; type B and C strains have been associated with both
fresh market and processing tomatoes; and type D strains have, so far, only been
associated with fresh market varieties. Although the biological significance of these
trends is still unclear, a substantial number of strains (30%) in the type A group were
avirulent in terms of wilt and canker symptom production and negative in the
hypersensitive reaction assay. None of the strains in the other fingerprint groups were
reported to be avirulent.

These studies were initiated to determine if a fourth primer, (GTG)5 could
provide more genotypic information or a greater resolution of the genomic fingerprint
than had already been established using the BOX-PCR reaction. Our primary interest in
this question stems from attempting to determine information about the primary source of
infection in commercial Midwest tomato fields. Determining which strains may be found
on certain farms in particular years may help determine if the inoculum is coming from
previous infections on the farm or if the inoculum is newly introduced each year. For
example, if the strains causing bacterial canker at a particular location were primarily
BOX-PCR type C one year, and primarily of type A the next, it would be difficult to

make any other conclusion than the inoculum was rapidly altered by a new source of

92

inoculum. But, if the fingerprint pattern remained the same year afier year, then an
argument could be made that the grower had an endemic source of inoculum on the farm
or in the region that initiated the disease on the farm.

BOX-PCR provides opportunities to pursue these questions but it is important to
determine if the genomic distinctions could be resolved to a greater level. For example,
in work by Louws et al. (1998), they presented evidence that type C strains could be
further resolved based on a few scattered polymorphic bands observed on the gels in
areas other than the 1 kbp region. These polymorphisms were used in a study where
strains were released in a greenhouse and several months later strains recovered from the
field plot could be traced back to the greenhouse based on these bands and antibiotic
resistance markers. These polymorphisms proved useful, and suggested that other, more
useful and prevalent, fingerprint patterns still could be resolved.

My first attempt to answer questions involving phenotypic diversity with (GTG)5-
PCR was attempted on the BOX-PCR type A strains. Since many of these strains were
recovered as avirulent as well as virulent, I wanted to determine if the resulting (GTG)5-
PCR banding patterns could be correlated to the virulence phenotype. The (GTG)5-PCR
reaction provided many more amplified bands, but the fingerprint patterns obtained with
(GTG)5 alone and in combination with BOX-PCR did not support any noticeable
genomic differences based on virulence.

This was not too surprising as other research groups interested in the pathogenic
response of C. michiganensis subsp. michiganensis to tomato have suggested that
virulence genes are primarily harbored on two plasmids (Meletzus et al., 1993). Since it

has been found that non-coding DNA regions are the main hybridization site for rep-PCR

93

primers and that plasmids are limited in the amount of non-coding sequences, it follows
that the rep-PCR technique would not separate strains based solely on plasmid
differences.

The question remained as to whether the computer-assisted analysis of BOX-PCR
type strains (A-E) may be better delineated using genetic diversity of C. michiganensis
subsp. michiganensis provided by (GTG)5-PCR analysis. Other studies have
demonstrated that the use of several primers provided more reliable information
regarding the genetic diversity of a species because more areas in the genome are
amplified resulting in more robust dendrograms (Louws et al., 1996; Schneider & de
Bruijn, 1996; Rademaker & de Bruijn, 1997; Rademaker et al. 1999b). My studies using
(GTG)5-PCR and BOX-PCR fingerprint patterns agrees with previous studies in that
clusters and strain identification were easily differentiated within each BOX-PCR
genotype. Although C. michiganensis subsp. michiganensis strains are overall highly
similar, more genetic variability usually resulted from the combined (GTG)5-PCR and
BOX-PCR fingerprint patterns than from using only one primer.

Perhaps the best example of this is with the BOX-PCR type A strains. Strains
CmmT74, CmmT13, and CmmT63 were collectively grouped with all of the other type A
strains when the BOX-PCR fingerprint was analyzed; however, when (GTG)5-PCR
fingerprint was analyzed these three type A strains were clearly differentiated based on a
very prominent band; this band was not observed in any other A strain. Since there were
other type A strains isolated from the same field in the same year that did not have the
predominate large band, it can be clearly stated that tomato plants in these fields were

infected by two different type A strains. Therefore, using just the BOX primer made the

94

 

infection of southwest Michigan’s processing tomatoes appear like a clonal infection
while (GTG)5-PCR differentiated the strains involved in this infection providing a better
genotypic resolution for multiple-year studies.

Strains placed in the A-l group could be subdivided into two groups depending
on where the experimental variability baseline (18 %) is placed. Placement of the
experimental variability baseline was a concern I faced with all BOX-PCR types when

analyzed with each primer alone or in combination. Type E strains remained the only

 

exception as these strains consistently remained a single cluster. The experimental

variability baseline for all reactions was calculated based on my finding that strains

 

analyzed colony by colony and within the same colony showed up to a 75% variability L-
when they, theoretically, should have been 100%. This variability could only be
accounted for by experimental reaction error such as slightly different amounts of DNA,
enzyme, or buffers added to each PCR reaction or by the placement of the reaction tubes
in the heating block. These experimental reaction errors are only a concern if the strains
are inconsistently placed within the different groups. If would be a major hindrance to
this research effort if a known strain fell into group A-l in one experiment and A-2 at
another time. Therefore, until all strains have been analyzed several times for their
position within each genotype, these dendrograms should be considered preliminary.

As stated above, my results suggested that type B strains represented a clonal
population. Only if the experimental variability baseline is moved considerably, two
strains, one from Michigan and the other from Ohio isolated several years earlier could

be separated. However, there is no experimental support for doing this and the group, so

95

far, remains clonal. Recovering more type E strains and the use of other universal rep-
PCR primers may prove useful in discovering diversity in the type E strains.

The type D strains required the use of a second primer to differentiate strains
collected from the same location in different years. At first, the analysis of the strains
recovered at the northeast Michigan farm was performed using only BOX-PCR and all
strains appeared clonal to each other recovered from Michigan but distinct from some
foreign strains sent to us from various locations throughout the world. When the isolates
were analyzed with only (GTG)5-PCR the strains showed genomic variability but nothing

could account for the variability as Michigan isolates were mixed with the foreign

 

isolates. After fingerprint patterns of the combined primers were analyzed, the type D
strains collected from Michigan were separated from the foreign strains. Surprisingly,
the type D strains isolated in 1994 were separated from the strains isolated in 1997. This
indicated that the type D strains on the northeast Michigan farm in 1994 were not the
same as those recovered on the farm in 1997 as each grouped into their own cluster. The
differentiation observed between these strains provided a useful resolution that may
contribute for a better understanding of bacterial canker epidemics because reproducible
polymorphic DNA bands could be used as genetic marker to trace these strains to their
primary source of inoculum.

Before too much emphasis is placed on the combined PCR genomic fingerprints
in trying to explain the primary source of inoculum on a specific farm, in a specific
region, during a given year or on a particular type of tomato, the nature of the PCR
reaction itself should be analyzed and if possible confirmed using other molecular

techniques such as restriction fragment length polymorphisms (RFLP) or monoclonal

96

immunological assays. The idea that one or two farms can harbor a specific pathogen,
and that pathogen appears endemic in one PCR primer but appears genetically distinct
when another PCR primer is used clouds the issue of genetic origin. How do these
specific types come about? Is there some kind of genomic rearrangement within
established strains? Could plarnsids, by their presence, including integration or excision
disrupt the genome such that genomic changes from year to year are observed in the F‘s
population? Certainly this has not been observed readily in the subspecific grouping of
pathovars of the xanthomonads, pseudomonads (Louws et al., 1995; Smith et al., 1995;

Opgenorth et al., 1996; Pooler et al., 1996;and Weingart & Volksch, 1997) and erwinias

 

TJH‘ .‘IS‘.’M“.Z"'.' ‘ .1 -.

(McManus & Jones, 1995). In those systems, clonality is commonly observed below the
pathovar level. With Clavibacter michiganensis subsp. michiganensis diversity could be
generated by small genomal rearrangements of the established strains. However, this
argument is hard to explain in field environments, although an experiment conducted in
closed environments have shown that over thousands of generations some changes of the
genome can be observed using rep-PCR technology (N akatsu et al., 1998)

Taking into considerations the cautions outlined above, it appears that we now
have a powerful method to distinguish genomic variation among strains at a high level.
Using both BOX-PCR and (GTG)5-PCR reactions and GelCompar analysis, we should be
able to determine if strains found on specific tomato farms change from year to year or
season to season as would be expected with pathogen populations arriving by seed or
seedling or whether they stay the same indicating clonal populations providing
continuous reinfection on the farm. These studies will require long-term surveys of the

farms and sampling procedures for adequate enumeration of genotypes.

97

REFERENCES

98

 

REFERENCES

Balerias Couto, M. M., Eijsma, B., Hofstra H., Huis in’t Veld, J .,H.,J ., and van der
Vossen, J. M. B. M. 1996. Evaluation of molecular typing techniques to assign
genetic diversity among Saccharomyces cerevisiae strains. Applied and
Environmental Microbiologl 62241-46.

Bell, J ., Medina-Mora, C., Hausbeck, M., Fulbright, D. 1997. Prevalence of different
PCR fingerprint groups and virulence class of Clavibacter michiganensis subsp.
michiganensis in four intensively managed tomato fields-preliminary report.
Proceedings, Thirteenth Annual Tomato Disease Workshop. Indianapolis, Indiana

Dec. 4-5.

Carlson, R. R., and Vidaver, A. K. 1982. Taxonomy of Corynebacterium plant
pathogens, including a new pathogen of wheat, based on polyacrylamide gel
electrophoresis of cellular proteins. International Journal of Systematic
Bacteriology 32:3 1 5-326.

Davis, M. J., Gillaspie, A. G. Jr., Vidaver, A. K., and Harris, R. W. 1984. Clavibacter: a
new genus containing some phytopathogenic coryneform bacteria, including
Clavibacter xyli subsp. xyli sp. nov., subsp. nov. and Clavibacter xyli subsp.
cynodontis subsp. nov., pathogens that cause ratoon stunting disease of sugarcane

and bermudagrass stunting disease. International Journal of Systematic
Bacteriology 34: 1 07-1 17.

de Bruijn, F. J. 1992. Use of repetitive (repetitive extragenic palindromic and
enterobacterial repetitive intergeneric consensus) sequences and the Polymerase
Chain Reaction to fingerprint the genomes of Rhizobium meliloti isolates and other
soil bacteria. Applied and Environmental Microbiology 58:2180-2187.

Dreier, J ., Bermpohl, A., and Eichenlaub, R. 1995. Southern hybridization and PCR for
specific detection of phytopathogenic Clavibacter michiganensis subsp.
michiganensis. Phytopathology 85:462-468.

Fatmi, M., and Schaad, NW. 1988. Semiselective agar medium of isolation of
Clavibacter michiganensis subsp. michiganensis from tomato seed.
Phytopathology 78:121-126.

Jaunet, T. X., and Wang, J .F . 1999. Variation in genotype and aggressiveness of
Ralstonia solanacearum race 1 isolated from tomato in Taiwan. Phytopathology
89:320-327.

99

 

Koueth, T., Versalovic, J ., and Lupski, J. R. 1995. Differential subsequence
conservation of interspersed repetitive Streptococcus pneumoniae BOX elements in
diverse bacteria. Genome Research 5:408-418.

Lee, I.-M., Bartoszyk, I.,M., Gundersen, D.,E., and Davis, R.,E.. 1997a. Phylogeny and
classification of bacteria in the genera Clavibacter and Rathayibacter on the basis

of 16S rRNA gene sequence analysis. Applied and Environmental Microbiology
63 :263 1 -2636.

Lee, I.-M., Bartoszyk, I. M., Gundersen, D.,E., Mogen, B., and Davis, R. E. 1997b.
Nested PCR for ultrasensitive detection of the potato ring rot bacterium,

Clavibacter michiganensis subsp. sepedonicus. Applied and Environmental
Microbiology 63: 2625-2630.

-1!
.1

Louws, F. J., Bell, J ., Medina-Mora, C. M., Smart, C. D., Opgenorth, D., Ishimaru, C. A.,
Hausbeck, M. K., de Bruijn, F. J ., and Fulbright, D.W. 1998. rep-PCR-mediated
genomic fingerprinting: A rapid and effective method to identify Clavibacter
michiganensis. Phytopathology 88:862-868.

 

Louws, F. J ., Schnieder, M., and de Bruijn, F.J. 1996. Assessing genetic diversity of
microbes using repetitive sequence-based PCR (rep-PCR). Pages 63-91 in: Nucleic
Acid Amplification Methods of Environmental Samples. Toranzos, G., ed.
Technomic Publishing, Co. Lancaster, PA.

Louws, F. J ., Fulbright, D. W., Stephens, C. T., and de Bruijn, F. J. 1995.
Differentiation of genomic structure by rep-PCR fingerprinting to rapid classify
Xanthomonas campestris pv. vesicatoria. Phytopathology 85:528-536.

Louws, F. J ., Fulbright, D. W., Stephens, C. T., and de Bruijn, F.J. 1994. Specific
genomic fingerprints of phytopathogenic Xanthomonas and Pseudomonas pathovars

ans strains generated with repetitive sequences and PCR. Phytopathology 60:2286-
2295.

Meletzus, D., Bermpohl, A., Dreier, J ., and Eichenlaub, R. 1993. Evidence for plasmid-
encoded virulence factors in the phytopathogenic bacterium Clavibacter

michiganensis subsp. michiganensis NCPPBB 82. Journal of Bacteriology
17522131-2136.

Meyer, W., Koch, A., Niemann, C., Beyerrnann, B., Epplen, J. T., and Bomer, T. 1991.
Differentiation of species and strains among filamentous fungi by DNA
fingerprinting. Current Genetics 19:239-242.

Meyer, W., and Mitchell, T. G. 1995. Polymerase chain reaction fingerprinting in fungi

using single primers specific to minisatellites and simple repetitive DNA sequences:
Strain variation in Cryptococcus neoformans. Electrophoresis 16:1648-1656.

100

Meyer, W., Mitchell, T. G., Freedman, E. Z., and Vilgalys, R. 1993. Hybridization
probes for conventional DNA fingerprinting used as single primers in the

polymerase chain reaction to distinguish strains of Cryptococcus neoformans.
Journal of Clinical Microbiology 31 :2274-2280.

McManus, P. S., and Jones, A. L.. 1995. Genetic fingerprinting of Erwinia amylovora
strains isolated from tree-fruit crops and Rubus spp. Phytopathology 85:1547-
1553.

Mogen, B. D., Olson, H. R., Sparks, R. B., Gudmestad, N. C., and Oleson, A. E.. 1990.
Genetic variation in strains of Clavibacter michiganense subsp. sepedonicum:

Polymorphisms in restriction fragments containing a highly repeated sequence.
Phytopatholog) 80:90-96.

Nakatsu, C. H., Korona, R., Lenski, R. E., de Bruijn, F. J., Marsh, T. L., and Fomey, L. J.
1998. Parallel and divergent genotypic evolution in experimental populations of
Ralstonia sp. Journal of Bacteriology 180:4325-4331.

Opgenorth, D. C., Smart, C. D., Louws, F. J., de Bruijn, F . J ., and Kirkpatrick, B. C.
1996. Identification of Xanthomonas fiigariae field isolates by rep-PCR genomic
fingerprinting. Plant Disease 80:868-873.

Pooler, M. R., Ritchie, D. F ., and Hartung, J .,S. 1996. Genetic relationships among
strains of Xanthomonas fiagariae based on random amplified polymorphic DNA
polymerase chain reaction, and enterobacterialrepetitive intergenic consensus PCR
data and generation of multiplexed PCR primers useful for the identification of this
phytopathogen. Applied and Environmental Microbiology 62:3121-3127.

Rademaker, J. W. L., Hoste, B., Louws, F. J ., Kersters, K., Swings, J ., Vauterin, L.,
Vauterin, P., and de Bruijn, F. J. 1999a. (In preparation) Comparison of AF LP
and rep-PCR genomic fingerprinting with DNA-DNA homology studies:
Xanthomonas as a model.

Rademaker, J. W. L., Louws, F. J ., Rossback, U. Vinuesa, P., and de Bruijn, F. J. 1999b.
Computer-assisted pattern analysis of molecular fingerprints and database
construction. In: Molecular Ecology Manual. Akkermans, A. D. L., van Elsas, J.
D., and de Bruijn, F. J ., eds. Kluwer Academic Publishers, Dordrecht.

Rademaker , J. W. L., and de Bruijn, F.J. 1997. Characterization and classification of
microbes by rep-PCR genomic fingerprinting and computer-assisted pattern
analysis. Pages 151-171 in: DNA markers: Protocols, Applications, and
Overviews. Caetano-Anollés, G., and Gresshoff, P. M., eds. Wiley and Sons, Inc.
New York, NY.

101

 

Rademaker, J. W. L., and Janse, J. D. 1994. Detection and identification of Clavibacter
michiganensis subsp. sepedonicus and Clavibacter michiganensis subsp.
michiganensis by nonradioactive hybridization, polymerase chain reaction, and
restriction enzyme analysis. Canadian Journal of Microbiology 40: 1007-1018.

Rademaker, J. W. L., Louws, F. J ., Schultz, M. H., Vauterin, L., Swings, J ., and de
Bruijn, F. J. 1997. (Abstract) Molecular systematics of xanthomonads by rep-PCR

genomic fingerprinting and computer-assisted pattern analysis. Phytopathology
87: S8 1 .

Roberts, P. D., Jones, J. B., Chandler, C. K., Stall, R. E., and Berger, R. D. 1996.
Survival of Xanthomonas fragariae in summer nurseries in Florida detected by

specific primers and nested polymerase chain reaction. Plant Disease 80:1283-
1288.

Sadowsky, M. J., Kindel, L. L., Bowers, J. H., and Schottel, J. L. 1996. Use of repetitive
intergenic DNA sequences to classify pathogenic and disease-suppressive
Streptomyces strains. Applied and Environmental Microbiology 62:3489-3493.

Sousa Santos, M., Cruz, L., Norskov, P., and Frasmussen. 1997. A rapid and sensitive
detection of Clavibacter michiganensis subsp. michiganensis in tomato seeds by
polymerase chain reaction. Seed Science and Technology 25:581-584.

Schnieder, M., and de Bruijn, F. J. 1996. Rep- PCR mediated genomic fingerprinting of
rhizobia and computer-assisted phylogenetic pattern analysis. World Journal of
Microbiology and Biotechnology 12:163-174.

Smith, J. J. Offord, L. C., Holderness, M., and Saddler, G. S. 1995. Genetic Diversity of
Burkholderia solanacearum (Synonym Pseudomonas solanacearum) Race 3 in
Kenya. Applied and Environmental Microbiology 61:4263-4268.

Vandamme, P., Pot, B., Gillis, M., de Vos, R, Kersters, K., and Swings, J. 1996.
Polyphasic taxonomy, a consensus approach of bacterial systematics.
Microbiological Reviews 60:407-438.

Vauterin, L., and Vauterin, P. 1992. Computer-aided comparison of electrophoresis

patterns of grouping and identification of microorganisms. European Microbiology
1:37-41.

Vera Cruz, C. M., Ardales, E. Y., Skinner, D. Z., Talag, J., Nelson, R. J., Louws, F. J .,
Leung, H., Mew, T. W., and Leach, J. E. 1996. Measurment of halotypic variation
in Xanthomonas oryzae pv. oryzae within a single field by rep-PCR and RFLP
analyses. Phytopathology 86: 1352-1359.

102

D 1““;

 

Versalovic J ., de Bruijn, F. J ., and Lupski, J. R. 1997. Repetitive sequence-based PCR
(rep-PCR) DNA fingerprinting of bacterial genomes. Pages 437-454 in: Bacterial
Genomes: Physical Structure and Analysis. de Bruijn, F.J., Lupski, J .R., and
Weinstock, G. M., eds. Chapman and Hall, NY..

Versalovic, J., Schneider, M., de Bruijn, F. J ., and Lupski, J. R. 1994. Genomic
fingerprinting of bacteria using repetitive sequence-based Polymerase Chain
Reaction. Methods in Molecular and Cellular Biology 5:25-40.

Versalovic, J ., Koeuth, T., and Lupski, J. R. 1991. Distribution of repetitive DNA

sequences in eubacteria and application to fingerprinting of bacterial genomes. .
Nucleic Acids Research 19:6823-6831. r-

Vinuesa, P., Rademaker, J. L. W., de Bruijn, F. J ., and Werner, D. 1998. Genotypic l]
characterization of Bradyrhizobium strains nodulating endemic woody legumes of '
the Canary Islands by PCR-restriction Fragment Length Polymorphism analysis of
genes encoding 16S rRNA (16S rDNA) and 168-238 rDNA intergenic spaces, and
partial 16 rDNA sequencing. Applied and Environmental Microbiology 64:2096-
2104. .;

 

Warren, R., Richardson, M., Sampson, S., Hauman, J. H., Beyers, N., Donald, P. R., and
van Helden, P. D. 1996. Genotyping of Mycobacterium tuberculosis with
additional markers enhances accuracy in epidemiological studies. Journal of
Clinical Microbiology 34:2219-2224.

Weising, K., Nybom, H., Wolf, K., and Meyer, W. 1995. DNA Fingerprinting in Plants
and Fungi. CRC press, Florida.

Weising, K., Kaemmer, D., Epplen, J. T., Weigand, F., Saxena, M., and Kahl, G. 1991.
DNA fingerprinting of Ascochyta rabiei with synthetic oligodeoxynucleotides.
Currents Genetics 19:483-489.

Weingart, H., and Volksch, B. 1997. Genetic fmgerpriting of Pseudomonas syringae
pathovars using ERIC-, REP-, and ISSO-PCR. Journal of Phytopathology 145:339-
345.

Yamada, K., and Komagata, K. 1972a. Taxonomic studies on coryneform bacteria IV.
Morphological, cultural, biochemical, and physiological characteristics. Journal of
Genetics and Applied Microbiology 18:399-416.

Yamada, K., and Komagata, K. 1972b. Taxonomic studies on coryneform bacteria V.
Classification of coryneform bacteria. Journal of Genetics and Applied
Microbiology 18:417-431.

103

CHAPTER 3

DEVELOPMENT OF BIRD'S EYE FRUIT LESIONS AFTER THE APPLICATION OF
CLA VIBA CTER MICHIGANENSIS SUBSPECIES MICHIGANENSIS
TO PROCESSING TOMATO FLOWERS
Bacterial canker of tomato, caused by Clavibacter michiganensis subspecies
michiganensis (Smith, 1910) is a concern for tomato growers because the disease in
appearance is unpredictable and symptoms may have devastating consequences on

tomato yield and quality (Bryan, 1930; Strider, 1969; Stephens & Fulbright, 1986; and

Gitaitis, 1991). The systemic infection of tomato plants is primarily associated with

 

'5‘1'. - -‘. .’

marginal necrosis, wilt, curling of the leaves, and stem canker symptoms. Seedling
infection may include white blister-like lesions on cotyledons and leaves (Basu, 1966).
Infection of the fruit may result in the development of superficial small necrotic spots
surrounded by white halos; these are called bird's eye lesions (Bryan, 1930; Strider,
1969). The presence of bacterial populations in the greenhouse as well as in the field
may cause serious economic losses because infections can reduce the amount and the
quality of the fruit produced from both processing and fresh market cultivars
(Pohronezny & Volin, 1983; Chang et al., 1992b; Gleason et al., 1993; Ricker & Riedel,
1993). Populations of C. m. subspecies michiganensis may be introduced to a disease
free area through the introduction of contaminated seeds (Bryan, 1930; Strider, 1969) and
symptomless transplants (Gitaitis et al., 1992). Dhavantari (1989) reported that the use of
389 transplants from infected seeds could result in 74.9% disease incidence with a
reduction of 53% in yield during a single year. Infected transplants may cause yield

losses up to 80 percent for individual commercial growers (Gleason et al., 1993).

104

Cultural practices during tomato production may enhance secondary infection and
dissemination in the field (Strider, 1969; Chang et al., 1991). The use of contaminated
equipment such as pruning tools, wood stakes, transplanters, cultivators, and tillers may
enhance the spread of the pathogen (Stephens & Fulbright, 1986; Carlton et al., 1994). In
a more natural manner, splashing rain and overhead irrigation systems (Gitaitis et al.,
1992) may also contribute to the dissemination of the epiphytic population harbored on
the leaves, stems and fruits.

The production of tomatoes in a previously contaminated field may serve as an

alternative source of inoculum since C. m. subsp. michiganensis can survive the winter

 

season for at least one year in plant debris mixed within the soil (Strider, 1967; Basu, 3;.
1970). A study conducted in a California field demonstrated that more than 50% of the
mature plants in the field showed canker symptoms when healthy transplants were sown
on soil beds containing contaminated plant material from the previous year (Gorgan &
Kendrick, 1953). Populations of C. m. subspecies michiganensis on plant debris may
result in yield reduction on the following year’s crop as shown in the studies conducted in
Illinois and Iowa (Gleason et al., 1991; Chang et al., 1992a). Weeds within fields may
also aid as a source of inoculum since nightshade, and jimsonweed species have been
reported as epiphytic hosts (Weebb et al., 1967; Ricker, 1989; Chang et al., 1992a).
Symptoms such as spotting, wilting, and vascular discoloration of cotyledons in
alternative hosts such as pepper and eggplant have been reported (reviewed by Strider,
1969)

Prevention of the disease and control of the pathogen include the use of certified

seed, healthy transplants, disinfected greenhouse materials and facilities, disinfected field

105

equipment, and fields rotated from tomatoes for at least two years (Bryan, 1930; Weebb
et al., 1967; Dhanvantari, 1989; Gitaitis, 1991). Several reports conducted in Michigan
have suggested the use of copper hydroxide at the greenhouse transplant stage to reduce
the populations of C. m. subspecies michiganensis on transplants taken to the field. The
suggested chemical regime benefits tomato production since there has been reported
significant reductions in yield loss (Hausbeck et al., 1995; Bell et al., 1996; Hausbeck et
al., 1999). Unfortunately, the application of these chemicals in the greenhouse do not
appear to provide control for the occurrence of fi'uit spotting in the field (Hausbeck et al., .
1995a; Hausbeck et al., 1999). *'

Some have concluded that the appearance and the development of fruit lesions are

 

primarily dependent on systemic or superficial infections of injured fi'uits caused by
abiotic events such as sandblasting (Pohronezny et al., 1992; Pohronezny et al., 1993).
However, it has been well accepted that foliar infections do not require wounded tissue
because tomato leaves contain trichomes which can harbor high populations of C. m.
subspecies michiganensis (Kontaxis, 1962; Layne, 1967). Other reports (Bashan et al.,
1981; Getz et al., 1983a; Ehrig & Griesbach, 1985) suggest that tomato fruit contain
trichomes hairs similar to those found on leaves, suggesting that the trichomes may be
involved in the fruit infection process. Blanke (1986) reported that the stomata present
on the epidermis of tomato fruits may serve as portals of entrance during bacterial
infection. Higgins (1922) suggested that there is only a short period of time early in the
development of the pepper fruit when it is susceptible to infection with Xanthomonas
campestris pv. vesicatoria. Getz et al. (1983a & 1983b) demonstrated that flower

inoculation resulted in fruit infection especially during and after anthesis (pollen

106

maturity) as the ovary is already under development during these stages. Getz et al.
(1983a) sprayed various flower developmental stages with Pseudomonas syringae pv.
tomato with the resulting fruit developing speck lesions. Fankle et al. (1993) studied the
ingress of watermelon fruit blotch after application of the bacterium Acidovorax
(Pseudomonas) avenae subsp. citrulli to the surface of young fruit. They agree with the

conclusion drawn by Getz et al. (19833) in that young fruits are more susceptible to

 

F‘
bacterial infection than older fruits. Such methods could be used to help understand C.
michiganensis. subspecies michiganensis infection and bird's-eye lesions development.
Therefore, the objectives of this work were to 1) determine if bird’s eye lesions are the
result of external flower infection and, 2) if it is, determine the flower stage which

provides the maximum number of fi'uit spots.

MATERIALS AND METHODS

Strains used. The Clavibacter michiganensis subspecies michiganensis strains used in
each experiment represent distinctive BOX-PCR fingerprint types designated as B type
(Cmm 68) or A type (Cmm 292) (Louws et al., 1998). Cmm 68 and Cmm 292 were
isolated fi'om fruit lesions from Ohio (1987) and Michigan (1993) fields, respectively.
Both strains were recovered from glycerol stocks (1 ml broth culture/0.5 ml 40%
glycerol) kept at -80° C and streaked onto nutrient broth yeast extract agar without

glucose (N BY; Fatmi & Schaad, 1988) plates. Pathogenicity tests on tomato plants and

107

the hypersensitive response (HR) (Gitaitis, 1990) on four-o’clock leaves (Mirabilis
jalapa) were conducted with both strains to determine their virulence.

Inoculum was prepared by inoculating 5ml of NBY broth (no agar) with either
bacterial strain and placing it on a shaker (New Brunswick Scientific NJ, USA) at 200
rpm at room temperature for 48 hours. A 100 ml of NBY broth was inoculated with the

fresh broth culture and cultured for an additional 48 hours as described. A 10 ml dilution

7'

was prepared from each broth culture. Spectrophotometer (Baush and Laurnb Spec 20)
readings were performed at 600nm for each dilution to calculate the culture volume

necessary to prepare a bacterial suspension of approximately 1 x 108 colony forming

 

Fffl‘ raw-3'4”}!n t ‘_'.- 4' . "_'

units (cfu)/ml (50% transmittance) with a final volume of 100 ml. The diluted culture
was transferred to Oakridge tubes and centrifiigcd in a desktop clinical centrifuge (IEC
Clinical Centrifuge MA, USA) for 15 min. at maximum speed. The pellet was
resuspended in 20ml of sterile distilled water (de), recentrifuged, and the cells
resuspended in 100ml of de. One milliliter of every bacterial suspension was stored in
0.5m] of glycerol (40%) in a micro-centrifuge tube (DOT Scientific MI, USA) at -20 C.
One to three days after the inoculum preparation, a 10-fold serial dilution of each
suspension was made by placing aliquots on NBY to enumerate the exact concentration
of the inoculum used in each experiment. The final concentration of the inoculum used
for all tomato flowers inoculations ranged from 3 x 107 - 3 x 108 cfu/ml. Prior to flower
inoculation, HR of each bacterial suspension prepared was tested for its ability to induce
an HR, a condition assumed to relate directly with the virulence of the pathogen.

Two tomato transplants of the susceptible processing variety (Heinz 8704), were

sown in 12 inch clay pots amended with commercially prepared potting media (Bacto).

108

Tomato plants were grown in the greenhouse for approximately eight weeks under
sodium lights with a photoperiod of 18 hours. Tomato plants were fertilized twice each
week with a diluted solution (1/ 16) of Peter's fertilizer (20-20-20) at 20 ppm.

Tomato flowers were tagged at the petiole with colored yarn to differentiate
various developmental stages. Inoculations were performed by applying approximately
0.7 ml of the washed C. michiganensis subspecies michiganensis suspension per flower
using an air pressurized sprayer (Preval, Precision Valve Co. NY, USA); sterile distilled

water (de) was used as control. A small circle, approximately 25 mm in diameter, was

 

cut in the bottom of a clear plastic cup (2 ounces) large enough to cover a single flower

 

with a wide open corolla. Individual flowers were covered with the plastic cup while
applying the inoculum or de to direct the fine mist toward each flower and to limit the

amount of aerosol landing on flowers and tissues of the immediate plants (Figure 163).

Fruit spot development. To develop a repeatable method with which to obtain fruit
spots similar to bird's-eye lesions, four treatments were initially evaluated. The first
treatment consisted of a one-time spraying of 2-4 where the corollas were open and wide
open (Figure 16A) per bract (1x Cmm). The second treatment consisted of a similar
inoculation except three days following the initial spray, a second spray was made (2x
Cmm). The third and fourth treatments were sprayed as in the treatments one and two
above, and following each inoculation, individual bracts were covered with clear plastic
bags (7in x 7in x 2.5in) for 16-18 hours (1x Cmm + bag and 2x Cmm + bag, respectively)
(Figure 16C). For every combination mentioned above, de was substituted as a control

inoculation (1x H20 and 2x H20 for one or two water applications, respectively). Each

109

 

Figure 16. Spray inoculation method used to study fruit spot development on
processing tomatoes. A) tomato flowers representing open corolla stages; B) use
of plastic cup method to prevent spread of aerosol to other flowers and plants; and
C) incubation of flowers for 16-18 hours with plastic bag used for third (1X Cmm
+ bag) and fourth (2X Cmm + bag) treatments.

110

 

treatment was conducted once in the spring and once in the fall. For the four treatments
indicated above, an average of 66 and 30 flowers were used for each experiment in the
spring and in fall, respectively. Three weeks after the second inoculation date for the
second (2x Cmm) and fourth treatments (2x Cmm + bag), all of the tomato fi'uits were
inspected for bird’s eye-lesions; the number of fruit infected and the number of spots per
fruit were recorded. All inoculations and incubations were performed during early
evening when relatively cooler temperatures were recorded in the greenhouse. Although
the two C. michiganensis subsp. michiganensis strains, Cmm68 and Cmm292, used in the
spring and in fall experiments belong to different BOX-PCR genotypes, these
experiments were considered as replicates because Cmm68 and Cmm292 shared
phenotypic characters such as virulence, plant tissue from where these were isolated, and
development of bird’s eye fruit lesions with typical appearance.

To obtain a higher incidence of diseased fruit an alternative method of inoculation
was attempted. Bacterial suspensions were prepared as described above and using a
small artist paintbrush (camel’s hair; #2) inoculum was directly applied to the surface of
young green fruits 22 mm in diameter. Eight C. michiganensis subspecies michiganensis
strains were selected based on their difference in BOX-PCR type and virulence rating
(Louws et al., 1998). Strains Cmm 299 (A-type), Cmm 68R (B-type), and Cmm 285 (D-
type) are classified as virulent; Cmm T33 (A-type), Cmm 236 (C-type), and Cmm 56 (C-
type) represent strains with intermediate virulence; and Cmm 208 (A-type), and Cmm
133 (A-type) are classified as avirulent. Each bacterial suspension was applied to a total
of five fruits; one fruit per plant, and tested for HR on four-o’clock leaves. Five fi'uits

were inoculated with SDW to serve as the control. This experiment was conducted once

111

.

 

during spring. Four weeks after fruit inoculation, the number of infected and noninfected
fruit and the number of spots per fruit were recorded.

The average number of spots per fruit for each treatment from both experiments
were analyzed by using Kruskal-Wallis ANOVA by ranks (P< 0.0001) (SIGMASTAT).
Differences between treatments were identified using the pairwise multiple comparison

by Dunn’s method.

Flower development stage and fruit spots. To determine the flower stage most
susceptible to the development of fruit spots, the bacterial suspension was applied to five
flower developmental stages by following the protocol for the third treatment (1x Cmm +
bag) described above. The five flower developmental stages were defined using anthesis
(pollen maturity) as the reference point. Bhadula and Sawhney (1987) established that
calyces of Lycopersicon esculentum (Mill) showing 11.0 -12.0 mm in length with an
open corolla corresponded to flowers at anthesis; therefore, the first stage of flower
development, considered to be about five days prior to anthesis consisted of buds with
closed calyces about 11-12 mm in length. The stages included in this experiment were
five (-5) and three (-3) days pre-anthesis, anthesis (0), two (+2), and four (+4) days post-
anthesis and each of these stages corresponded to the following flower appearance,
closed calyx, open calyx, open corolla, wide open corolla, and half-closed corolla,
respectively (Figure 17). For each flower developmental stage, de was used for control
inoculations (H20). Three replicates (I, II, and III) were conducted in the spring, and an
average of 57, 65, and 66 flowers were used for each replicate, respectively. When the

youngest fruit reached 27 mm in diameter, the number of infected and noninfected fi'uit

112

 

   

Figure 17. Tomato flowers representing four of the five developmental stages used
in the experiment to determine the susceptibility of flowers to infection with C.
michiganensis subsp. michiganensis.

113

and the number of spots per fruit were recorded. All inoculations and incubations were
preformed during early evening. The average number of spots per fruit for each replicate
was statistically analyzed as mentioned above.

To confirm that the observed fruit spots were caused by the inoculum used for
each experiment and for each replicate, randomly selected fi'uit spots were removed using
a sterile scalpel. Each fruit spot was placed in a droplet of de for 5 seconds to let
bacteria ooze from the plant material. Each droplet was streaked onto NBY plates and
incubated for 96 hours at room temperature. After pure culture of single-colonies,

suspected of being C. michiganensis subspecies michiganensis were subjected to rep- E

 

PCR fingerprinting protocol (Louws et al., 1998) to demonstrate that the BOX-PCR

genotype was the same as the inoculum used in each experiment.

RESULTS

Fruit spot development. To determine the best technique with which to obtain fi'uit
spots, flowers with open corollas were inoculated by spraying with 1 X108 C.
michiganensis subsp. michiganensis. Both C. michiganensis subsp. michiganensis strains
used in this study, Cmm68 and Cmm292, were considered virulent strains and produced
strong hypersensitive reactions in the HR assay conducted in parallel with this study.
When they appeared, fruit spots could be found as typical or atypical spots.
Typical fi'uit spots were observed as white lesions on small green fi'uit (10-12 mm in
diameter), 6-8 days after inoculation. Three to five days after the white lesions first

appeared, typical bird’s-eye lesions developed, as the center of the lesion became necrotic

114

(Figure 18). Atypical fruit spots were necrotic spots similar in size to the bird's-eye
lesions but lacking the characteristic white halo. The development of the necrotic area
appeared about the same time after inoculation in both types of lesions. In spite of this
difference, both the typical and atypical fi'uit spots were small, circular, and superficial
with a raised blister-like appearance approximately 2 mm in diameter. Upon reisolation,
both typical and atypical fruit spot lesions yielded the C. michiganensis subsp.
michiganensis strains used as inoculum in each experiment, as determined by rep-PCR
assays.

The number of inoculations and the treatment of the flower after inoculation were

 

Iran |‘.I_ K: o‘.l' .'.1 \‘u
I

varied in these experiments. Fruit spots were observed in all treatments, but the number
of spotted fruit obtained was affected primarily by the number of inoculations each
flower received (Table 2). Fruit developing from flowers that were inoculated twice with
either bacterial strain (2x Cmm treatments), showed the highest number of spotted fruit
and the highest number of mu spots per fruit. The maximum number of spotted fruit
ranged from 78 to 80 %, and 53 to 70 % was observed on fruit that developed fiom
flowers sprayed twice with Cmm68 and Cmm292, respectively. As with the results
observed for the number of spotted fruit, the maximum number of spots per fi'uit was
obtained from the flowers inoculated twice. The maximum number of spots per fruit
ranged from 7.1 to 12.0, and 7.6 spots per fruit when Cmm68 and Cmm292, respectively,
was applied to open flowers two times three days apart. The least amount of diseased
fruit was obtained when the flowers were inoculated only once (1x Cmm) and these fruit
averaged less than one spot per fruit. The minimum number of spotted fruit ranged from

14 to 32 %, and 9 to 28 % was observed on fruit that developed from flowers sprayed

115

£73”

   

Figure 18. Tomato fruit showing bird’s eye lesions eight weeks afier C.
michiganensis subsp. michiganensis inoculum was sprayed on flowers with an
open corolla.

116

Table 2. Fruit spotting development after spraying open corolla flowers with
strains of Clavibacter michiganensis subsp. michiganensis (Cmm68 and

 

 

 

 

 

 

 

 

 

 

 

 

Cmm292).

Treatment Total number of 1 Incidence 2 Average 3

flowers fruit fl (% diseased fruit) # sp_ots/ fruit
1x Cmm68 31 29 14 0.4 a 5
2x Cmm68 31 30 80 12.0 b
1x Cmm68 + bag 4 36 31 32 2.1 a
2x Cmm68 + [fig 29 23 78 7.1 b
1x Cmm292 81 46 9 0.2 a
2x Cmm292 128 61 70 7.6 b
11: Cmm292 + bag 64 29 28 1.2 a
2x Cmm292 + bag 122 72 53 7.6 b
I Number of flowers (open corolla) sprayed with Cmm versus the number of fruit that
developed from those flowers.

 

2 Percent of fiuit showing fruit spotting.

3 Total number of spots observed per total number of fruit that developed
per treatment.

4 Incubation of Cmm on flowers after covering individual bracts with clear plastic bags for
16-18 hours afier inoculation.

5 Statistical analysis using Kruskal-Wallis ANOVA by ranks and pairwise multiple
comparison by Dunn's method. (P= <0.0001)

117

once with Cmm68 and Cmm292, respectively. As with the results observed for the
number of spotted fruit, the minimum number of spots per fruit was obtained from the
flowers inoculated once. The minimum number of spots per fruit ranged from 0.4 to 2.1,
and 0.2 to] .2 spots per fruit when Cmm68 and Cmm292, respectively, was applied to
open flowers once regardless of the presence or absence of the bag. Bagging the flowers
to hold in moisture after inoculation did not provide an obvious advantage as similar
number of spots appeared on flowers inoculated twice regardless of the bagging regime
used. Bagging may have increased the number of spots obtained with the single
inoculation treatment, but the results were not statistically significant when compared to
the flowers without bags. No fruit spots developed on uninoculated control flowers. The
largest number of spotted fi'uit occurred when the surface of young fruit was inoculated
using a paintbrush to apply the bacterial suspensions (Table 3). Strains determined to be
virulent in previous studies and causing a hypersensitive reaction in the HR assay in this
study resulted in 100% of the inoculated fruit developing fi'uit spots. Strains classified as
intermediate in virulence and also causing positive hypersensitive reactions in HR assays
in this study resulted in 67-100% of the inoculated fruit developing fruit spots.
Surprisingly, strains determined to be avirulent in previous studies and not producing
hypersensitive reactions in HR assays in this study resulted in 75-80% of the inoculated
fruit developing fruit spots. The number of spots developing per fruit was not
statistically related to the strains used, however, avirulent strains produced the lowest
number of spots, and intermediate and virulent strains produced the highest number of

spots with one exception. Strain Cmm236, previously determined to be intermediate in

118

 

Table 3. Fruit spot development after direct application of Clavibacter
michiganensis subsp. michiganensis to the surface of young tomato fruit.

 

 

 

 

 

 

Virulence 1 Strain(s) used Incidence 2 Average3
(°/o diseased fruit) # spots/fruit
v Cmm 299 100 60.8 a 4
Cmm 68R 100 66.8 a
I Cmm T33 67 45.0 a
Cmm 236 100 109.8 a
Cmm 56 100 24.2 a
AV Cmm 208 80 15.4 a
Cmm 133 75 27.3 a

 

 

 

 

1Virulent= producing canker and wilt; l=intermediate, producing canker only; and AV=
avinllent, producing no symptoms (Louws et al., 1998).

2 Percent of film showing fi'uit spotting.

3 Total number of spots observed per total number of fruit that developed
per treatment.

4 Statistical analysis using Kruskal-Wallis ANOVA by ranks and pairwise multiple
comparison by Dunn's method. (P= 0.03)

119

 

virulence produced the largest number of spots per fruit than any other strain used in the

study.

Flower development stage and fruit spot development. To determine the flower
developmental stage most susceptible to infection, five flower developmental stages were
established and flowers were inoculated with a strain of C. michiganensis subsp.
michiganensis at each stage. The majority of the flowers inoculated post-anthesis (+2
and +4) developed into fruit with fruit spots (Table 4). The highest numbers of spots per
fruit were also obtained from flowers inoculated during post-anthesis (+2 and +4). Those
flowers inoculated prior to anthesis rarely developed fruit spots. No fnlit spot lesions
were observed when flowers were inoculated with water.

The number spots per fruit in this experiment was similar to those observed in the
fruit spot development experiment described above where the bacteria were delivered by
aerosol inoculation. For example, strain Cmm68 (Table 2) in the fruit spot development
experiment produced approximately 2 fi'uit spots per fruit when inoculated once and
bagged. In this experiment, strain Cmm68 was also inoculated once and bagged, and
again, produced approximately 2 fruit spots per fruit. Although a higher incidence of
fruit spots may have been achieved using two inoculations instead of one, in the flower
development stage experiment, I decided to use only one inoculation since the flowers
would have been at a different development stage when inoculated the second time, 3

days later.

120

 

Table 4. Fruit spotting after spraying flowers at various developmental stage wit
Clavibacter michiganensis subsp. michiganensis strain Cmm68.

 

 

 

 

 

 

 

Flower Stage Total number Incidence 2 Average 3
Anthesis of fruits' (% diseased fruit) # smts/fruit

(days) I “ I 11 111 1 11 111 1 11 I 111
-5 30/56 12/71 22/89 0 0 0 0.0 a 0.0 a 0.0 a
-3 23/54 14/74 20/121 0 0 5 0.0 a 0.0 a 0.05 a
0 48/56 29/70 16/79 6 0 0 0.1 a 0.0 a 0.0 a
+2 54/55 47/72 57/102 30 15 42 1.7 b 0.8 a 2.8 b F
+4 46/58 59/77 35/65 30 20 37 1.4 b 0.5 a 1.9 b .

 

1 Number of fruit that developed from flowers sprayed with Cmm68.

2 Percent of film showing fruit spotting.

3 Total number of spots observed per total number of fruit that developed for each treatment.
4 I, II, and III = three independent replicates conducted during spring.

5 Statistical analysis by Kruskal-Wallis ANOVA by ranks and pairwise multiple comparison E
by Dunn's method. (P= <0.0001) ..

 

121

DISCUSSION

Symptom expression in plants infected with bacteria is dependent on many factors
including the genetics of the pathogen, age of the tissue infected, availability of infection
sites, as well as environmental conditions (Strider, 1969). Previous studies with C.
michiganensis subsp. michiganensis have shown that young leaves are more susceptible
to infection by leaf blighting phytopathogenic bacteria than older leaves (Kontaxis, 1962;
Layne, 1967). Leaf trichomes have been implicated in bacterial canker infection
(Kontaxis, 1962) and Layne (1967), demonstrated that young tomato leaves contained

higher numbers of trichomes per unit area than older leaves suggesting that the leaf

 

trichomes of young leaves may provide an opportunity for growth and a more conducive
environment for subsequent infection.

Fruit age also may influence the expression of the fruit spot symptoms in bacterial
infection of plants (Gardner & Kendrick, 1923; Bryan, 1930; Getz et al., 1983; and
Pohronezny et al., 1993). Getz et al. (1983b) reported that young fi'uit were highly
susceptible to external infection with Pseudomonas syringae pv. tomato and showed the
presence of trichomes on the surface of tomato ovaries soon afier anthesis (Getz 1983a).
They suggested that these fruit trichomes may provide an avenue for infection as natural
openings occur through the process of trichome breakage. Scanning electron microscopy
revealed the presence of P. syringae pv. tomato cells inside holes resulting from broken
trichome-bases.

Similar processes with the bacterial canker bacterium may be involved with fruit
infection. One confounding aspect of bacterial canker infection, that is not an issue with

other tomato bacterial diseases that produce fruit spots, is that the disease is often

122

systemic, rapidly moving through the xylem. This has led some researchers to speculate
that the fruit spots develop from internal systemic infection. All of my studies strongly
suggest that flowers externally infected with C. michiganensis subsp. michiganensis
early after anthesis would produce fi'uit with bird’s eye lesions. It can be concluded that
flowers at post anthesis are more susceptible to C. michiganensis subsp. michiganensis
infection than flowers inoculated prior to anthesis. While the studies presented here do
not rule out the possible role of internal infection causing fi'uit spots, they agree with the
conclusions drawn by Getz et al. (1983) that spotted fi'uit can result from nonwounding,

external infection of ovaries.

 

The large number of spots obtained when small green fi'uit were inoculated with l3
bacterial suspensions may be a result of trichome injury, physical disruption of the
epiderrrris and waxes positioned to prevent pathogen invasion or simply a more direct
method of delivery of the pathogen to infection courts. This method of inoculation
should allow a simple, more consistent and objective method to enumerate the effect of
the pathogen on the host. It should be useful in selecting tolerant or resistant varieties
judging efficacy of bactericides, or determining the virulence of C. michiganensis subsp.
michiganensis strains.

Establishing the earliest developmental stage when flowers become susceptible to
bacterial pathogens might be usefill to formulate new, better targeted, control strategies.
For example, the application of bactericide(s) at flowering or before flowering for the
determinant tomato varieties may decrease the number of spotted fi'uit because
populations of the pathogen may decrease and remain in low numbers during critical

times for infection. Nontraditional control strategies such as biological control or

123

genetically engineered hosts could also benefit from knowledge of fruit spot
development. Effective control agents could be selected that might specifically colonize
flowers and ovaries out-competing or otherwise inhibiting the pathogen. Genetically
engineered tomatoes could also deliver inhibiting compounds in targeted cells, such as
trichomes on the ovary.

Surprisingly, few spotted fruit developed from flowers inoculated at anthesis or
pre-anthesis stages after inoculation with high levels of inoculum. It seems that if a large
population of the pathogen is in place a few days before the optimum infection point,
those cells making up the population should still be in place when the appropriate time
for infection arrives. Apparently, C. michiganensis subsp. michiganensis was not a
competent epiphyte under conditions provided during this experiment. Several factors
may be responsible for this observation. It is possible that bacterial multiplication may
have been hampered due to the relatively low humidity found in the research greenhouses
used in this study. Keeping the plants moist by placing a plastic bag around the flowers
after inoculation did increase the number of spotted fruit, although not statistically
significant (Table 2). The bagging treatment may have provided the moisture needed for
continued colonization of the bacterial canker pathogen but the bacteria may not have
been engaged in processes of active infection such as moving off of the petals to the
trichomes and surface of the fruit. If bacterial multiplication continued and if the
pathogen successfillly colonized the early flower developmental stages, it could be
possible that significant numbers of the bacterial population were lost during petal-fall, a
common event during tomato fruit development. As a result, epiphytic cells were not

efficiently delivered to the susceptible stage or tissue for fruit spot development. It is

124

 

n‘ ~I: - n

also possible that the low number of fruit spots obtained when inoculating at anthesis or
earlier is an artifact of the greenhouse inoculation protocol due to the lack of wind, rain
and water splashing that would normally occur in the field. Pohronezny et al. (1992)
reported that a high number of fruit spots resulted from infection with Xanthomonas
campestris pv. vesicatoria if the pathogen is at high concentrations at natural openings.
They suggested that, although the ntunber of natural openings in pepper fi'uits is limited,
the bacterial population is critical for the development of fruit spots in peppers. In their
study, they showed that environmental factors, such as wind and rain splashing, could
promote the production of natural openings on fi'uits.

It is also possible that certain C. michiganensis subsp. michiganensis strains are
more genetically capable of dissemination than others and that the strains used in this
study were not as capable of moving from petals to ovaries as others. I used at least two
strains with different histories, as indicated in the Materials and Methods, in the fruit spot
development experiments. More definitive answers on the effect of strains versus early
flower inoculation could be obtained by doing similar experiments under controlled
environments in growth chambers with other strains of the pathogen.

The most likely reason for delivering the inoculum prior to the most susceptible
stage and not seeing significant infection is a combination of two events. The first event
may be due to the bacterial population being delivered after it is already in the stationary
phase of growth. This means that for active growth to occur on the new substrate (the
plant), cells must go through a lag phase before growth can occur again. In the meantime,
death phase has been reached in the older cells of the population while the flower is

dramatically altering its tissues and shedding petals. Those flowers inoculated twice

125

 

T3757. -.'.
l.

probably had the population re-enforced during the most susceptible time for infection to
occur.

The higher number of spotted fruit developing from flowers sprayed twice with or
without a bag, supports the proposition that bacterial numbers are important for fruit
infection. While obtaining large numbers of spotted fruit were convenient for purposes
of these experiments, just a couple of fruit spots per tomato fruit can reduce the quality
and acceptability of the fresh market tomato.

During this study, high levels of flower abscission occurred resulting in low

number of fruit on several occasions. Getz et al. (1983b) and Pohronezny et al. (1993)

 

also observed the abortion of a high number of non-mature tomato and pepper fruit
during their studies. Various reasons may cause flower and young fruit abscission such
as poor pollination, nutrient deficiency, inappropriate light intensity and soil media, as
well as poor ventilation (personal communication with Dr. Ivin Widers- Michigan State
University). The paintbrush method of applying the inoculum avoided the problem of
flower abortion as only the fruit to survive flower abortion were inoculated. I observed
no significant fruit abortion afier inoculation of the fruit with the pathogen (data not
shown).

A normal distribution of the number of spots per fruit was not observed in any of
the experiments conducted in this study. The source of variability could have emerged
from various factors with biological implications such as inconsistent environmental
conditions, diverse localization of plants in the greenhouse as well as the localization of
the flowers infected on each plant, and variability in the interaction between the host and

the pathogen.

126

In conclusion, two protocols to obtain spotted fruit through external inoculation of
tomato flowers and fruit under greenhouse conditions have been developed. Studies,
such as this one, are necessary if we are to attempt further experiments involving the
management of bacterial canker symptoms on fruit as well as increase our knowledge of

the host-pathogen interaction of this important pathogen.

 

127

 

REFERENCES

128

 

REFERENCES

Bashan, Y., Sharon, E., Okon, Y., and Henis, Y. 1981. Scanning electron and light
microscopy of infection and symptom development in tomato leaves infected with
Pseudomonas tomato. Physiological Plant Pathology 19:139-144.

Basu, P. K. 1966. Conditions for symptomatological differentiation of bacterial canker,
spot, and speck on tomato seedlings. Canadian Journal of Plant Science 46:525-
530.

Bell, J ., Medina, C., Fulbright, D., and Hausbeck, M. 1996. Epidemiology of Clavibacter
michiganensis subsp. michiganensis on tomato transplants in the greenhouse and
the field. Proceedings, Twelfth Annual Tomato Disease Workshop, Ohio, Dec.5-6.

Bhadula, S. K., and Sawhney, V. K. 1987. Esterase activity and isozymes during the
ontogeny of stamens of male fertile Lycopersicon esculentum Mill., A male sterile E
starnenless-2 mutant and the low temperature-reverted mutant. Plant Science
52:187-194.

 

Blanke, M. M. 1986. Comparative SEM study of stomata on developing quince, apple,
grape, and tomato fruit. Angew. Botanik 60:209-214.

Bryan, M. K. 1930. Studies on Bacterial Canker of Tomato. Journal of Agricultural
Research 41 :825-85.

Carlton, W. M., Gleason, M. L., and Braun, E. J. 1994. Effect of pruning on tomato
plants supporting epiphytic populations of Clavibacter michiganensis subsp.
michiganensis. Plant Disease 78:742-745.

Chang, R. J ., Ries, S. M., and Pataky, J. K. 1992a. Local sources of Clavibacter

michiganensis ssp. michiganensis in the development of bacterial canker on
tomatoes. Phytopathology 82:553-560.

Chang, R.J., Ries, S. M., and Pataky J. K. 1992b. Reductions in yield of processing
tomatoes and incidence of bacterial canker. Plant Disease 76:805-809.

Chang, R. J ., Ries, S. M., and Pataky, J. K.. 1991. Dissemination of Clavibacter
michiganensis subsp. michiganensis by practices used to produce tomato

transplants. Phytopathology 81 :1276-1281.
Dhanvantari, B. N. 1989. Effect of seed extraction methods and seed treatments on

control of tomato bacterial canker. Canadian Journal of Plant Pathology 11:400-
408.

129

Ehrig, F., and Griesbach, E. 1985. Development of bird eye spots on tomato fi'uits-

Investigations by scanning electron microscopy. Arch. Phytopathol. Pfalnzenschutz
2 1 :41 9-423.

Fatmi, M., and Schaad, N. W. 1988. Semiselective agar medium of isolation of
Clavibacter michiganensis subsp. michiganensis from tomato seed.
Phytopathology 78: 1 21 -126.

Frankle, W. G., Hopkins, D. L., and Stall, R. E. 1993. Ingress of the watermelon fruit
blotch bacterium into fruit. Plant Disease 77:1090-1092.

Gardner, M. W. and Kendrick, J. B. 1953. Bacterial spot of tomato and pepper.
Phytopathology 13 2307-3 15.

Getz, S., Fulbright, D. W., and Stephens, C. T. 1983a. Scanning Electron Microscopy of
infection sites and lesion development on tomato fruit infected with Pseudomonas
syringae pv. tomato. Phytopathology 73:39-43.

Getz, S., Stephens, C. T., and Fulbright, D. W. 1983b. Influence of developmental stage
on susceptibility of tomato fruit to Pseudomonas syringae pv. tomato.
Phytopathology 73:36-38.

Gitaitis R. 1991. Bacterial canker. Pages 25-26 in: Compendium of Tomato Diseases.
Jones, J. B., Jones, J. P., Stall, R. E., and litter, T. A., eds. 2"d edition. American
Phytopathological Society Press, St. Paul, MN.

Gitaitis, R. 1990. Induction of a hypersensitive-like reaction in four-o'clock by
Clavibacter michiganensis subsp. michiganensis. Plant Disease 74:58-60.

Gitaitis, R., McCarter, S., and Jones, J. 1992. Disease control in tomato transplants in
Georgia and Florida. Plant Disease 76:651-656.

Gleason, M. L., Braun, E. J ., Carlton, W. M., and Peterson, R. H. 1991. Survival and
dissemination of Clavibacter michiganensis subsp. michiganensis in tomatoes.
Phytopathology 81 :1519-1523.

Gleason, M. L. Gitaitis, R. D., and Ricker, M. D. 1993. Recent progress in
understanding and controlling bacterial canker of tomato in eastern North America.
Plant Disease 77: 1069-1076.

Grogan, R. G., and Kendrick, J. B. 1953. (Abstract) Seed transmission, mode of

overwintering, and spread of bacterial canker of tomato caused by Corynebacterium
michiganense. Phytopathology 43:473.

130

 

Hausbeck, M. K., Bell, J ., Medina-Mora, C. M., and Fulbright, D. W. 1999. (In progress)
Impact of bactericides applied to tomato seedlings in the greenhouse on C.
michiganensis subsp. michiganensis populations and field productivity.

Hausbeck, M., Fulbright, D., Bell, J ., Medina, C., and de Bruijn, F. J. 1995a.
Epidemiology of bacterial diseases on vegetables transplants in the greenhouse.
Proceedings, Eleventh Annual Tomato Disease Workshop, Ontario, Dec. 7-8.

Higgins, B. B. 1992. The bacterial spot of pepper. Phytopathology 12:501-516.

Kontaxis, D. G. 1962. Leaf trichomes as avenues for infection by Corynebacterium
michiganense. Phytopathology 52:1306-1307. F

Layne, R. E. C. 1967. Foliar trichomes and their importance as infection sites for
Corynebacterium michiganense on tomato. Phytopathology 572981-985.

(' 0'." I. E‘ 'r ‘

 

Louws, F. J ., Bell, J ., Medina-Mora, C. M., Smart, C. D., Opgenorth, D., Ishimaru, C.A.,
Hausbeck, M. K., de Bruijn, F. J ., and Fulbright, D.W. 1998. Rep-PCR-Mediated
genomic fingerprinting: A rapid and effective method to identify Clavibacter
michiganensis. Phytopathology 88:862-868.

 

Pohronezny, K., Hewitt, M., and Infante, J. 1993. Susceptibility of pepper fruit to

infection by Xanthomonas campestris pv. vesicatoria. Canadian Journal of Plant
Pathology 15:107-109.

Pohronezny, K., Hewitt, M., Infante, J ., and Datnoff, L. 1992. Wind and wind-generated
sand injury as factors in infection of pepper by Xanthomonas campetris pv.
vesicatoria. Plant Disease 76:1036-1039.

Pohronezny, K., and Volin, R. B. 1983. The effect of bacterial spot on yield and quality
of fresh market tomatoes. HortScience 18:69-70.

Ricker, M. D., and Riedel, R. M. 1993. Effect of secondary spread of Clavibacter
michiganensis subsp. michiganensis on yield of Northern processing tomatoes.
Plant Disease 772364-366.

Ricker, M. D. 1989. Bacterial canker research 1985-89. Proceedings, F ifih Annual
Tomato Disease Workshop, Michigan, Dec. 7-8.

Smith, E. F. 1910. A New Tomato Disease of Economic Importance. Science 31 :794-
796.

Stephens, C. T., and Fulbright, D. W. 1986. Bacterial Canker of Tomato. MSU Ag
Facts, Extension Bulletin E-1943 Januaryzl-4.

131

Strider, D. L. 1967. Survival studies with tomato Bacterial Canker organism.
Phytopathology 57:1067-1071.

Strider, D. L. 1969. Bacterial Canker of Tomato caused by Corynebacterium
michiganense: A literature review and bibliography. North Carolina Experiment

Station, Technical Bulletin 193: 110p.

Weebb, R. E., Good, J. M., and Danielson, L. L. 1967. Tomato diseases and their
control. US. Department of Agriculture, Agriculture Handbook 203:9—13.

132

it
.1

 

CONCLUSIONS

Rapid and highly discriminatory strain identification has been accomplished by
DNA-based fingerprinting techniques such as the repetitive sequence-based Polymerase
Chain Reaction (rep-PCR). Rep-PCR protocols have been instrumental in differentiating
and classifying strains of microorganisms of medical, agricultural, and economic
importance. This technique has also provided new insights into the concepts of diversity
and speciation. The use of universal rep-PCR primer such as REP, ERIC, and BOX has
provided genetic information of microorganisms where genomic characteristics of the
strains were limited or lacking. The accurate identification of genetic variants of
pathogenic microbes is critical for understanding epidemics in regards to animal, human,
or plant pathogens.

The study of genomic diversity in the phytopathogenic bacterium Clavibacter
michiganensis subspecies michiganensis (Cmm), the causal agent of bacterial canker of
tomato has proven to be an excellent example in which to demonstrate the discriminatory
potential of rep-PCR. This is in part, due to the unique characteristics of the disease
cycle in relationship with the tomato production in northern states. The presence of
polymophic DNA fragments of Cmm strains below the subspecies level may provide
useful tools for finding the primary source of inoculum. The results of this study have
demonstrated the use of another rep-PCR primer, (GTG)5, for the differentiation of Cmm
strains. (GTG)5- PCR can be used alone or in combination with BOX-PCR to provide
even greater resolution of the genotypic diversity of Cmm. Various questions still remain
unanswered in regards to the relationship between the sporadic appearance of the disease

and the predominant genotype(s) responsible for the severity of the diverse symptom

133

 

 

 

expression in each field, in each region of Michigan and North America. Cmm strains of
different pathogenic abilities, ie. avirulent and virulent, can be present in the same field
during an epidemic. Unfortunately, the differentiation of these pathogenic variants was
not possible using the combined data generated from the BOX-PCR and (GTG)5-PCR
fingerprinting patterns. The potential of other DNA fingerprinting techniques such as
restriction fragment length polymorphism (RFLP) in conjunction with rep-PCR protocols
to differentiate pathogenic types needs to be further explored.

Until this study was conducted, the origin of inoculum responsible for the fi'uit
spots was unknown. In this study two protocols were developed to obtain spotted fruit.

Using these techniques, I was able to conclude that epiphytic populations of Cmm was

 

responsible for the primary source of epidermal fruit infection leading to the development
of bird’s eye lesions. The development of bird’s eye lesions through application of Cmm
through aerosols probably mimicked the natural mode of dissemination of the pathogen
but it was time-consuming and tedious. The application of Cmm using an artist’s
paintbrush was more reliable and reproducible than aerosols but it was artificial in terms
of inoculum application. Surprisingly, the paintbrush technique demonstrated that a
strain recognized as avirulent in stem inoculation assays could still insight fruit spots,
although at lower level than virulent strains. The paintbrush technique should prove
valuable in enumerating studies involving virulence, resistance, chemical treatments, and

biocontrol.

134

 

APPENDICES

135

 

 

 

APPENDIX A

BACTERIAL CANKER: SYMPTOMS

136

 

.990va E5 so $87.2
SE? 28 A208: 26me swarm H8? mm 565 £952 co $880: 3533: .m $55 was Aachev £862 :3 was
88m 06 mac? mmmouooadoxcao .< ESE 48¢ 28 Sam—m 92:9 0538 no ”c.8038? Hod—:8 356nm .3 25me

 

137

APPENDIX B

BACTERIAL CANKER: DISEASE CYCLE

138

 

 

......Eorcwiog .mmnsm £33335? $83330 3 .0823 0.9.8. we 038:. .838 3.363 0:. ..o 228 3.3me .om 3:3,...

 

    

 

 

 

 

.833 £3.82. 2.38.
......» «=3 1.53:... amour—......
M/ . Human...
. ... .YM.‘
T, Q
c.5898?
...—:3 688...:
309$

350—. 3.2.3.

 

 

 

.8359. 28.5.3
..etafioufio. bun?»

532:5... 8.5.1.3..» 8.3.3.3... 5:933...— Soc
\1/ \I/ \I/
ll'

41\

 

 

 

 

«magnum...» 20m

.8225, 5:8 .53
notes—......— team

.33.

0 D
000

 

139

APPENDIX C

BACTERIAL CANKER: DISEASE CYCLE COMPONENTS

140

 

 

 

 

208 $88.... .w.0
808.850 3.00....

:03me .80....08

0.00: .0000, 30:08:80
0030...... .3535

23.9.88 6:3, ...00m
.....Ea Z

5.3.3.8000... .8 00.82

 

05023.9. -2?

wfib

ngéouéaq
him:— 0:00.;

3.8.30.3...
0.080%

889.03..
3:80.... .9587.

00:88 8300.:—

 

.:0:........0 60.00.:—

8003 0.00.33... .wd
08.. 0>3..:..0..<

0.80.. 08.8...

300m

00.88 82:8:—

 

6.308388

...mnsm 0,3088%308 883.530 3 .0038 0.9.0 030m... .8380 330.03 0.... .8 8:089:00 msotm> .m 0331

141

APPENDIX D

BACTERIAL CANKER: CONTROL

142

 

0.00..0>0 05:00.80 .0: :5 6.0.0000
0:03.30 0:30.03.

:0000000: 80000000.. 0.00.8000 000004080008 .w.0v 80880.00 0:0 0.00. w:000..:.0.0

08080 80.330 :000000 0.0...
.0088 8080.05 0.00.. 03.0800? 0:0 0.8.00 0.080: .0 .0>080:
. . A.
30.03.00. 00.08.00. 8:08 H00. 0:00 000 0000 00:88
.... 0.. :
00.0.0300: 0003-39.00 000.000 . E. U

 

 

0030.005 80803—005.

 

.0.0=0§ME0.E
5.00:0 0.0:08ME0E 000000.085 0.0. 00008 0.080. .0 000.800 8.00.000 ..0 .8800 :0. 0080.050 8080980.). .0 030%

143

APPENDIX E

CHARACTERIZATION AND IDENTIFICATION

144

 

 

 

000-0000. .

£50.. 080 0 020 .08.
”9.00.000. «on. 00000 00:00.0000300003. .

05.880000 <20

<20. 00. 0 0.: .
<75 .0800009.

0.220 a. 0.220 .
020 0 0.20 .
0008.8. 9.0% 8.0003. 8000 000.0850“.

.22 .

:58. 802908030
5800 .:08w0:.. 5.00.803. \00008 <ZQ

0.9.0059

 

w0.0.m .
$00508 .00.80000.m.v 800000.03.

GEE. 0.0022 .300 .0002 22 00”. .
w:..:...&0w:.. 0.00 0000.. 00.2.00

...... 08000802008088. .
00.0 000 0030-30.08.88. .
<m.‘.m .
00.000.80 3:200:08 0:0 .0:0.0>.0& 30.00%

9208 0.000203088 U 00.. 0.008 0>..00.00.80m .

$96 .0008”. 0000> 0.00m. 80.8.2 .
30.00808 80.00

0.9.0082...

080.00 0.020830%: 0.0000

0.0:020ME0E 02000.83 (.0 :0..00.m.800. 0:0 :0..0N...0.00:000 000 00‘. 000: 000.008 0.058% 0:0 0.9.0050“. ... 0.00 H

145

"WEYILTIfifl'LFHflTLLWflWE/ZW