GENETIC ANALYSIS OF INTRASPECIFIC VARIATION IN
PATHOVARS OF Xanthanonas campestris
GERARD RAYMOND LAZO
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA 1987
Gerard Raymond Lazo
I wish to express my gratitude to Dr. D. W. Gabriel, chairman of
the advisory committee, for his friendship, help and encouragement during the course of this research. Thanks are also expressed to all personnel associated with the laboratory, who were very generous and cooperative in supporting the research efforts which made this presentation possible. Special thanks are extended to A. Williams who
assisted with the race-specificity studies on cotton; G. V. Minsavage who assisted with the SDS-PAGE analysis; R. Pof fey, who assisted with the RFLP analysis; and A. Burges who assisted with various other aspects of this study. Thanks are also expressed to Drs. R. E. Stall,
D. R. Pring, and M. J. Bassett for serving on the advisory cormittee and for their support. Appreciation is also acknowledged for
financial support from the Oklahoma and Florida Agricultural Experiment Stations, and United States Department of Agriculture grant USD.A-58-7B30-3-465. Anid finally, I am grateful to my loving wife, Maria, for supporting and encouraging me throughout this degree program, for she tolerated my late hours and comforted me in the most difficult of times.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .................................................. iii
LIST OF TABLES ................................................... vi
LIST OF FIGURES .................................................. vii
ABSTRACT ......................................................... ix
ONE INTRODUCTION ........................................... 1
TWO LITERATURE REVIEW ...................................... 4
THREE HOST RANGE OF PATHOVARS OF XANTHOMONAS CAMPESTRIS ...... 9
Introduction ........................................... 9
Materials and Methods .................................. 10
Bacterial Strains and Host Plants ................. 10
Plant Inoculations ................................ 12
Physiological Differentiation ..................... 12
Polyacrylamide Gel Electrophoresis ................ 13
Results ................................................ 13
Pathogenicity Tests ............................... 13
Physiological Differentiation ..................... 20
Polyacrylamide Gel Electrophoresis ................ 22
Discussion ............................................. 22
FOUR CONSERVATION OF PLASMID DNA SEQUENCES AND PATHOVAR
IDENTIFICATION OF STRAINS OF XANTHOMONAS CAMPESTRIS .... 27
Introduction ........................................... 27
Materials and Methods .................................. 28
Bacterial Strains ................................. 28
Plasmid Extraction and Visualization .............. 31
Cloning of Plasmid
Restriction Endonuclease Fragments ................ 32
DNA/DNA Hybridization ............................. 32
Results ................................................ 33
Detection of Plasmid DNA .......................... 33
Restriction Endonuclease Profiles ................. 35
DNA/DNA Hybridization ............................. 40
Dot-blot Hybridization ............................ 42
Discussion ............................................. 45
FIVE ARE RACE-SPECIFIC AVIRULENCE GENES ENCODED BY PLASMIDS
OF XANTHOMONAS CAMPESTRIS PV. MALVACEARUM?...............50
Materials and Methods ....................................51
Bacterial Strains and Host Plants ...................51
Plant Inoculations ..................................51
Plasmid Analysis ....................................51
Plasmid Curing ......................................53
Race-Specificity Genes ..............................53
Plasrnid origin of Replication and Mobilization ...54
Plasmids and Race-Specificity .......................55
Plasmid Curing ......................................58
Race-Specificity Genes ..............................58
Plasmid Origin of Replication and Mobilization .... 61
six PATH~OVARS OF XANTHOMONAS CAMPESTRIS ARE DISTINGUISHABLE
BY RESTRICTION FRAGMENT LENGTH POLYMORPHISM ..............70
Materials and Methods ....................................72
Bacterial Strains ...................................72
DNA Extraction .....................................72
Agarose Gel Electrophoresis.........................75
DNA Probes ..........................................75
Restriction Fragment Patterns and Densitornetry ...77
Agarose Gel Electrophoresis .........................77
DNA Probes .........................................80
DNA Hybridization ...................................80
SEVEN SUMMARY .................................................92
Intraspecific Variation ..................................92
Restriction Fragment Length Polymorphism .................93
Conclusions ...........................o.......... &........95
LITERATURE CITED ..................................................97
BIOGRAPHICAL SKETCH .............................................. 106
LIST OF TABLES
3-1 Strains of X. campestris used in host range investigation ... 11 3-2 Legume plant reactions to inoculation with pathovars
of X. campestris ............................................ 14
3-3 Host range observed for pathovars of X. campestris
pathogenic to legume host plants ............................ 17
3-4 Malvaceous host plant reactions to inoculation
with strains of X. campestris pv. malvacearum ............... 19
3-5 Physiological reactions of strains of X. campestris ......... 21
4-1 Strains of X. campestris used for plasmid analysis .......... 29
4-2 Detection of plasmid DNA in strains of X. campestris ........ 34 4-3 Hybridization of radiolabeled plasmid probes to
total DNA of pathovars of X. campestris ..................... 44
5-1 Strains of X. campestris pv. malvacearum from cotton
used in race-specificity investigation ...................... 52
5-2 Plasmid groupings and cotton plant reactions to strains of
X. campestris pv. malvacearum ............................... 56
5-3 Pathogenicity of X. campestris pv. malvacearum
transconjugants on cotton host differentials ................ 59
5-4 Selection of plasmid replication genes in X. campestris
pv. malvacearum strain X .................................... 62
6-1 Strains of X. campestris used for restriction fragment
length polymorphism analysis ................................ 73
6-2 Sizes of DNA fragments from X. campestris genomic digests
which hybridized to the XCT1 DNA probe. See Figure 6-3 ..... 83 6-3 Sizes of DNA fragments from X. campestris genomic digests
which hybridized to the XCT1 DNA probe. See Figure 6-5 ..... 87
LIST OF FIGURES
3-1 SDS-Polyacrylamide gel electrophoresis of total proteins
from strains of X. campestris ............................... 23
4-1 Plasmnid DNAs from strains of X. campestris pv. malvacearum
digested with restriction endonucleases EcoRl and BamHI ..... 37
4-2 Graphic representation of plasmid EcoRI restriction fragment
profiles for pathovars of X. campestris ..................... 38
4-3 Plasmid DNAs from strains of X. campestris pv. citri
digested with restriction endonuclease EcoRI ................ 39
4-4 Plasmid DNAs from strains of X. campestris pv.
malvacearum digested with restriction endonuclease
EcoRI and hybridized to probe N4.5 .......................... 41
4-5 Plasmid DNAs from strains of X. campestris digested with
restriction endonuclease EcoRI and hybridized to probes
P2.0 and P2.3 ............................................... 43
5-1 Plasmid curing of X. campestris pv. malvacearum with
SDS treatment ............................................... 60
5-2 Transconjugants of X. campestris pv. malvacearum containing
pLXD and hybridized to the plasmid probe pSa4 ............... 64
5-3 Transformation of E. coli ED8767 with plasmid DNA
from X. campestris pv. malvacearum transconjugants
mated with pLXD ............................................. 66
5-4 Plasmid DNA of X. campestris pv. malvacearum strain X
digested with restriction endonucleasea and hybridized
to the plasmid probe pXD-1 .................................. 67
6-1 Genomic DNA of strains of X. campestris pvs. phaseoli,
alfalfae, and campestris digested with restriction
endonuclease EcoRI .......................................... 78
6-2 Genomic DNA of strains from different pathovars of
X. campestris digested with restriction endonuclease EcoRI .. 79
6-3 Genomic DNA of strains of X. campestris pvs. phaseoli,
alfalfae, and campestris digested with restriction
endonuclease EcoRI and hybridized with probe XCT1 ........... 81
6-4 Genomic DNA of strains of X. campestris pvs. phaseoli,
alfalfae, and campestris digested with restriction
endonuclease EcoRI and hybridized with probe XCT11 .......... 84
6-5 Genomic DNA of strains from different pathovars of
X. camoestris digested with restriction endonuclease EcoRI
and hybridized with probe XCTI ............................ 85
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
GENETIC ANALYSIS OF INTRASPECIFIC VARIATION IN
PATHOVARS OF Xanthomonas campestris
Gerard Raymond Lazo
Chairman: Dean W. Gabriel
Major Department: Plant Pathology
Xanthomonas campestris is always found in association with plants. Those strains of X. campestris which are known to be pathogenic are differentiated into over 125 pathovars (pathogenic variants), on the basis of limited pathogenicity tests. Some members of a pathovar may have a broader host range than others, since exhaustive pathogenicity testing is impractical. Other than these tests, there is no definitive means to classify an unknown X.
campestris isolate. Those strains of X. campestris which are not pathogenic are unclassifiable by the pathovar system, yet they may exhibit similar host range specificity to those which are pathogenic. This work was conducted in an effort to better understand the variability of X. campestris.
Over 115 strains of X. campestris were examined for plasmid content and restriction endonuclease profiles. Of the 26 pathovars examined, only 13 were found to contain plasmids. Restriction
endonuclease digested plasmid DNA from strains within a given plasmid-containing pathovar gave surprisingly similar, but not always identical, digestion profiles by agarose gel electrophoresis. All 60 strains tested of X. campestris pvs. glycines, malvacearum, phaseoli, and vignicola could be accurately identified by pathovar from
determination of the restriction fragment profile and/or by Southern hybridizations of that profile with a selected plasmid DNA probe. In
no instance was the same plasmid profile seen in more than one pathovar. The apparent stability of the plasmids provides a natural genetic marker that can be strain specific and perhaps useful in epidemiological investigations.
Cloned DNA fragments, derived from a cosmid library of a Florida isolate of X. campestris pv. citri strain 3041, were used to detect
restriction fragment length polymorphism (RFLP) of total DNA from 87 strains of X. campestris, comprising 23 different pathovars.
Autoradiographs of Southern transfers of genomic DNA probed with cosmid-sized clones revealed hybridization profiles which appeared to be highly conserved. The hybridization patterns observed between
different pathovars suggested that some DNA fragments were conserved at the species level, others were conserved at the pathovar level, and
still others were variable. The degree of variability appeared to depend on the DNA probe used. By using more than one DNA probe, or by digesting the genomic DNAs with different restriction endonucleases, all of the strains of X. campestris described as belonging to a given pathovar could be differentiated. All strains of X. campestris were readily grouped by RFLP phenotypes, and the classification based on
RFLP patterns correlated well with the classification based on pathogenicity. Although certain pathovars may need to be redefined, this work supports and helps validate the natural taxonomic groupings provided by the pathovar naming system.
The genus Xanthomonas Dowson consists of bacteria described as being gram-negative, obligately aerobic, non-fermentative rods which are motile by a single polar flagellum and can produce a brominated yellow pigment, called xanthomonadin (6). All reported strains of this genus have been described as plant associated, and most are reported as being pathogenic to a particular plant host. Based on
microbiological classification, this genus can be separated into at
least five separate species: X. campestris (Pammnel) Dowson, X. albilineans (Ashby) Dowson, X. ampelina Panagopoulos, X. axonopodis Starr and Garces, and X. fragariae Kennedy and King (and possibly X. populi (Ride) Ride and Ride). Over 125 variants, or pathovars, of X. campestris, are essentially indistinguishable from each other, except for their plant host range (6,89).
To maintain order in differentiating among strains of X. campestris by host range, the rank of pathovar has been assigned in
addition to the species name (24,98). The primary means by which
strains can be differentiated as pathovars is by inoculation of the plants which serve as susceptible hosts to a given pathovar. The
convention for naming pathovars of X. campestris has generally been to
name it for the host from which it was first isolated (89) A
particular pathovar designation suggests a limited host range for a
given strain. The actual host range may be much broader, but a
complete host range description of a given strain would require testing on all plants known to support growth of the genus. This
would be a monumental task, and therefore the potential host range of a given strain of X. campestris is unknown. Obviously, any
information regarding pathogenicity is of value to pathologists. However, without knowledge of the potential host range variation of strains within a pathovar, the taxonomic value of the pathovar system is questionable.
Attempts to devise alternative means to differentiate pathovars
of X. campestris have included serology (1,94), membrane protein profiles (70), phage-typing (37), rRNA-DNA and DNA-DNA hybridization (21,73), and gas chromatography of fatty acids (65,81). Although most of these methods have been useful to differentiate given sets of strains of X. campestris, there is little evidence that the groupings
distinguished by these methods have any correlation to the host range groupings observed for given pathovars of X. campestris. None of
these techniques have been successful in replacing pathogenicity tests to identify X. campestris to the pathovar level. Because differences in host range primarily determine the pathogenic variability observed
in X. campestris, it is of interest to know more about the nature of this microorganism and the mechanisms which determine host-range specificity.
Genetic analyses of pathogenicity can easily be done with bacteria, for they can be quickly cloned and assayed, and studied using current recombinant DNA technologies (50). X. campestris is
useful because it is a facultative parasite with a high level of natural variation and has been shown to follow similar gene-for-gene patterns of interactions with plants as found with fungal pathogens
The objective of this research was to develop a better understanding of the intraspecific variation and specificity present in the species X. campestris. This was conducted by first verifying
the pathogenicity of the working strains as pathovars and, where appropriate, as races within selected pathovars of X. campestris through pathogenicity testing. This work was followed by physical
characterization of the plasmids and chromosomes of X. campestris races and pathovars. These DNA analyses were designed to determine the extent of physical variation within the species, pathovars, and races of X. campestris. Finally, genetic variation was analyzed at the race level by using cloned genes and complementation analyses to identify specific avirulence genes which contribute to the race phenotypes. By conducting experiments in a multi-faceted approach, it was hoped that a better understanding of the interactions of strains of X. campestris with their host plant(s) could be appreciated.
At least 124 monocotyledonous and 268 dicotyledonous plants species are susceptible to infection by bacterial strains of the genus
Xanthomonas (57). Compiled listings indicate that the genus
Xanthomonas has a wide host range extending over nine monocotyledonous families comprising 66 genera, and 49 dicotyledonous families comprising 160 genera. Pathogenicity toward gymnosperms or lower plants has not been reported.
Although pathovars are generally assumed to have a restricted host range, some strains of X. campestris have extended host ranges to more than one plant species. In these instances, host range usually
extends toward other members of the same plant family (with few exceptions). For example, the plant families Compositae, Cruciferae,
Graminae, and Leguminosae each contain different species susceptible to the same X. campestris pathovar (X. c. pv.). An example of a broad
host range pathovar of graminaceous hosts is X. c. pv. translucens (Jon., John., and Redd.) Dye, which is pathogenic on barley, rye, and wheat. Other strains pathogenic on graminaceous hosts have been described, but their interrelationships among the other pathovars were
poorly defined (6). In another example, X. c. pv. alfalfae, is
pathogenic on the legumes alfalfa, bean, and pea. Another special relationship is that of X. c. pv. vesicatoria (Doidge) Dye, which is a
pathogen of Capsicurm annuum L. (pepper) and Lycopersicon esculentum (tomato). While some strains of X. campestris appear to have a host
range to more than one plant species, most others appear to have an extremely narrow host range (6,57). The host range characteristic is considered to be a stable trait, although there is one report of altered host range by artificial selection (23). This report has been
refuted, with no change in host-range specificity observed for X. campestris (84). Within at least some pathovars, variability among strains exists as race-specific plant interactions (8,17,20). These are recognized when resistant plants of an otherwise susceptible plant species are discovered. The ability to differentiate races of X. campestris pathovars is dependent on the discovery of plant hosts, called host differentials, which exhibit variability in their response to different pathogenic strains. Pathogenic races have been described for X. c. pv. malvacearum (Smith) Dye, X. c. pv. oryzae, X. c. pv. translucens, X. c. pv. vesicatoria, and possibly others. This racespecific variation in X. campestris was suggestive of gene-for-gene interactions between pathogenicity determinants in the pathogen with resistance genes in the host (8,20).
Gene-for-gene relationships can only be established by genetic crosses of the host and the parasite. Because the bacterial genetics
can only be done with recombinant DNA techniques, gene-for-gene interactions could only be recently established. Proof of
gene-for-gene interactions with bacterial pathogens was documented for X. c. pv. malvacearum (33).
The high level of host-range and race specificity described above
for X. campestris is observed with some other plant pathogenic bacteria. For example, there are more than forty pathovars of Pseudncmonas syringae which are primarily distinguished by host-range specificity (75,86). Some P. syringae pathovars have host ranges which are narrowly restricted to specific plant species; others have a wider range of hosts. Strains of Erwinia chrysanthemi are currently
being considered for differentiation at the pathovar level (56). Strains of Agrobacterium spp. and Rhizobium spp. are not
differentiated using the pathovar system; however, host-specific interactions analogous to those observed in X. campestris exist with these bacteria. Species within each genus can be differentiated
biochemically and by plant inoculations. These two genera, which cause very different symptoms, are both in the family Rhizobiaceae and are genetically highly related (25,42). In Rhizobium spp. and
Agrobacterium spp., host range appears to be specified by the bacterium by genes which are expressed as a positive function (32). In several species of Rhizobium, host range function has been identified as being plasmid borne (4). The introduction of one of these plasmids may extend the host range of a limited host range species of
Similarly, A. tumefaciens tumor inducing (Ti) plasmids carry host range determining genes (64). Ti plasmids from wide-host-range
strains, when moved to A. tumefaciens strains with limited hostranges, confer a wide host range to the recipient (10,43,48). The pTAR plasmid, found in A. tumefaciens strains from grapevine, carries
a gene for tartaric acid catabolism (35). The expression of pTAR
genes appears to function more like a host range determinant rather than an oncogenic determinant, as found on tumor inducing plasmids.
The proof of plasmid involvement in virulence and host range with species of Agrobacterium and Rhizobium apparently stimulated research of plasmid involvement among other phytopathogenic bacteria. Plasmids
have been found to exist in many other phytopathogenic bacteria, but few have been associated with pathogenicity (76). Examples of plasmid functions in pathogenicity on plants include tumor induction in A. tumefaciens (97), nodulation in Rhizobium spp. (62), indoleacetic acid production by P. syringae pv. savastanoi (90), cytokinin production by
Corynebacterium fascians (syn. Rhodococcus fascians) (72), virulence of Pseudomonas solanacearum (5) and race-specific avirulence in X. c. pv. vesicatoria (88), among others. A general conclusion from these studies is that plasmids from a wide range of plant pathogenic genera can carry a variety of genes which determine the outcome of plantpathogen interactions.
Plasmids have been shown to be physically similar among strains within pathovars of P. syringae (77). The similarity is based on
homology studies using restriction endonuclease digestion patterns and
DNA-DNA hybridization experiments (15). Additionally, there are instances where similar plasmids, or even identical ones, may be found
in closely related pathovars of P. syringae (77). Although the
functional utility of these plasmids is unknown, they often cannot be eliminated or "cured" from the bacterial strain, possibly because such
an event would be lethal. For example, in P. syringae pv.
phaseolicola, a plasmid, pMC7105, can be forced to integrate into the host chromosome, but cannot be cured (79,92). The plasmid pMC7105 has some homology with other P. syringae pv. phaseolicola strains, and
even P. svringae pv. glycinea strains. Homology between these
plasmids may be due to transpositional events, the presence of
repetitive sequences, variability due to integration and excision of the plasmid, plasmid origin of replication, or other plasmid
maintenance function (79,92). Another possibility is that genes essential for pathogenicity, host-range, or race specificity may be encoded on them. The association of plasmid homology with host-range specificity may indicate the involvement of plasmids in host range determination in these cases.
HOST RANGE OF PATHOVARS OF XANTHOMONAS CAMPESTRIS
There are over 125 pathovars of X. campestris described, and the only means to differentiate between them is by pathogenicity testing
(6). Same strains of X. campestris have extended host ranges to more than one host, usually to other members of the same plant family, with few exceptions (57). The host range can sometimes overlap onto
another host for which a different X. campestris pathovar has been described. For instance, X. c. pv. alfalfae is a pathogen of the legumes alfalfa, bean, and pea, but strains exist of X. c. pv. phaseoli and X. c. pv. pisi which have host ranges limited to bean and pea, respectively. The relatedness between these strains at the
subspecific level is unknown, except for their host ranges. In
another instance, strains of X. c. pv. malvacearum have been isolated from the malvaceous hosts, cotton and hibiscus (7,12). Although given the same pathovar designation because they were each isolated from plants of the Malvaceae family, their relatedness is unknown. Similar
situations, where the relationship between strains is unclear, exist among other pathovars of X. campestris (6). The purpose of this
investigation was to verify the pathogenicity and host range of all
stock strains used in this dissertation, to examine the relationships between apparently related strains of X. campestris in terms of host9
range specificity, and to find a physiological test diagnostic for one or more pathovars used in these studies.
Materials and Methods
Bacterial Strains and Host Plants
Bacteria used in this investigation consisted of pathovars of X. campestris which were pathogenic to members of either the Leguminosae or Malvaceae plant families. The pathogens of leguminous host plants
consisted of X. campestris pvs. alfalfae, cyamopsidis, glycines, phaseoli, pisi, and vignicola isolated from alfalfa, guar, soybean, kidney bean, pea, and cowpea, respectively. The pathogens of
malvaceous host plants consisted of strains of X. c. pv. malvacearum isolated from cotton or hibiscus. The strains of X. campestris used in this study are shown in Table 3-1. The legume host plants used in this investigation included Medicago sativa cv. "FL-77", Glycines max cv. "Evans", Phaseolus vulgaris cv. "California Light Red", and Vigna unguiculata cv. "California Blackeye #5". The malvaceous hosts
included three cultivars of Hibiscus rosa-sinensis and eight cultivars of Gossypium hirsutum. One of the cultivars of H. rosa-sinensis was
cv. Brilliant Red; the other two were not named, but chosen to represent diversity in the species. Eight cotton cultivars were
selected for differentiating races of X. c. pv. malvacearum, two of which were cv. 101-102B, and cv. Gregg (45). The other cultivars were cv. Acala 44 (no resistance to cotton strains), and five cultivars derived from cv. Acala 44 crossed to obtain single gene resistance to the cotton strains of X. c. pv. malvacearum, the cultivars being Acala Bl, B2, B3, B5, and BIN (7).
Table 3-1. Strains of X. campestris used in host range investigation.
Pathovar Strain Plant Host Location
alfalfa KS Medicaao sativa Kansas
FL M. sativa Florida
cyamopsidis 13D5 Cyamopsis tetragonoloba
X002,XO05,XO16,XO17 C. tetragonoloba Arizona
glySines B-9-3 Glycine max Brazil
1717 G. max Africa
17915 -. max
S-9-8 G. max Wisconsin
malvacearum DMNOUVXYZTX84 osT Tium hirsutan Texas
ABEFGH G. hirsutum Oklahoma
OhlCh2 G. hirsutum Chad
HV25 -. hirsutum Upper Volta
Su2,Su3 G. hirsutum Sudan
FL79 G. hirsutum Florida
083-4244,M84-11 R-ibiscus rosa-sinensis Florida
XlOX27,X52,XlO2,XlO8 H. rosa-sinensis Florida
phaseoli EK11,Xph25,Xpfll Thaiie-olus vulgaris Nebraska
XpaXpll P. vulqaris Wisconsin
82-1,82-2 P. vulgaris Florida
LB-2,SC-3B P. vulgaris Nebraska
XP2 P. vulgaris New York
XP-JL P. vulgaris Kansas
XP-JF T. vulgaris Missouri
XP-DPIB5B P. vulgaris
RiL XP1 Pisum sativum Japan
vignicola A81-331,C-1,CB5-1, Vigna ungiuculata Georgia
Xvl9,SN2,432,82-38 V. unguiculata Georgia
Pathogenicity tests were conducted by pressure infiltrating bacterial suspensions into leaf tissue with a blunt-end syringe and incubating the plants in 28 C or 30 C growth chambers until symptoms were expressed. The leguminous host plants were incubated at 28 C and
symptoms were recorded after one week. The Inalvaceous hosts were incubated at 30 C and symptoms were recorded after a two week period. Pressure infiltration facilitated rapid screening of plant reactions, but was not representative of natural inoculation. Because alf alf a
leaves were small in size, infection was by spray inoculation. Bacterial suspensions were prepared by centrifuging overnight cultures of X. campestris, resuspending them in 0.7% NaCl to an approximate optical density (OD600nm) of 0.3 for plant inoculation. These
pathogenicity tests were repeated at least once. The results were recorded to determine compatible (pathogenic) or incompatible (nonpathogenic or hypersensitive) plant reactions to inoculations. Physiological Differentiation
As X. campestris is defined (6), the physiological characteristics for gelatin and starch hydrolysis may be variable for this species. Other physiological characteristics which are consistent
with this species includes xanthomonadin production, mucoid growth, and esculin and casein hydrolysis. These above mentioned
characteristics were determined using standard methods (36,82). Additionally, physiological tests for production of cellulase (96), lecithinase, lipase, and pectinases were included. Hildebrand's
medium prepared at three pH's (4.5, 7.0, and 8.5) was used to detect pectolytic enzyme activity. In each of these three media, sodium
polypectate was used as the carbon source. The pH 5 medium included only pectin as the sole carbon source. These tests were each repeated once.
Polyacrylamide Gel Electrophoresis
Bacteria were grown overnight at 30 C in a peptone-glycerol broth. Cells were adjusted to a 0.3 optical density (OD600rm), and 1.5 mls of cells were collected by centrifugation, washed once in water, and resuspended in 50 ul 10% sorbitol. Then, 50 ul
solubilization buffer (90 mM Tris (pH 6.8), 20% glycerol, 4% SDS, 10%
B-mercaptoethanol, and .002% bromophenol blue) was added to the suspension and the mixture was boiled for 5 minutes. Approximately 10 mg protein (10 ul solution) was added to a 10% resolving acrylamide gel (51), and samples were electrophoresed in a 20 cm x 17 cm x 1.5 mm vertical gel unit at 10 V/cm and 15 C until the dye reached the bottom of the gel. Biorad low-range SDS-PAGE standards were used as
molecular weight markers. Gels were stained with 0.1% Coomassie blue in 42% methanol and 17% glacial acetic acid. The gel was destained in
30% methanol and 10% glacial acetic acid. Electrophoresis was only performed once. A photograph of the gel was taken using Polaroid type .55 film.
All of the strains of X. c. pv. phaseoli, with one exception, gave the expected disease reactions on the legume hosts tested. The exception, strain B5B, was found to be non-pathogenic to all four
hosts tested. The remaining six strains were pathogenic on kidney bean (Table 3-2), giving watersoaked lesions on inoculated leaves
Table 3-2. Legume plant reactions to inoculation with pathovars of X. campestris.
Inoculation Strain bean cowpea soybean alfalfa
phaseoli 82-1 +B C D 0E
Xpa + 0
Xpfll + 0
XP-JF + 0
EK11 + 0
XP2 + _F 0
vignicola CB5-1 -H +I _J 0
Xvl9 + 0
432 + 0
A81-331 + 0
C-1 + 0
82-38 +/-K + 0
SN2 +/- + 0
glycines B-9-3 +L _C +M 0
1717 + + 0
17915 + + 0
alfalfae FL +L C _D +N
malvacearum N _F 0 _F 0
control 0 0 0 0
A = + is compatible, is incom- G = strain appeared nonpatible, +/- is intermediate, pathogenic.
and 0 is a null reaction. H = dry necrotic lesion with
B = compatible lesions were slight watersoaking at
watersoaked and appeared to periphery of inoculation site.
be spreading. I = dry necrotic lesion with
C = dry necrotic lesion with wine shothole effect.
red reaction. J = dry necrotic lesion.
D = dry necrotic lesion with K = as described for H, but
chlorosis. slight shothole effect present.
E = no reaction seen with spray L = watersoaked lesion.
inoculation. M = watersoaked chlorotic lesion.
F = slight tissue discoloration N = watersoaked leaf spots.
at inoculation site.
after four days. The lesions extended from the inoculation sites
giving the appearance that the bacterium was spreading throughout the leaf.
The strains of X. c. pv. phaseoli which were found pathogenic on kidney bean were not found to be pathogenic on cowpea, soybean, or alfalfa. In cowpea and soybean, a hypersensitive reaction was apparently elicited, resulting in dry necrotic lesions at the
inoculation sites. In soybean the lesion was slightly chlorotic, whereas for cowpea, a wine red lesion occurred. However, X. c. pv. phaseoli strain XP2 did not elicit the wine red color on cowpea (Table 3-2). No response was observed on alfalfa from the spray inoculations with X. c. pv. phaseoli.
The strains of X. c. pv. vignicola were pathogenic on cowpea, resulting in irregularly shaped lesions which had a tearing, or "shothole" appearance at the inoculation site. On soybean, the
incompatible response appeared dry and collapsed. Again, no response was observed from spray inoculation on alfalfa (Table 3-2).
On kidney bean, a range of different responses occurred from inoculations with X. c. pv. vignicola. For two of the strains, A81331 and Cl, an incompatible response appeared as a dry collapsed lesion. For the other strains of X. c. pv. vignicola, there appeared to be some slight water-soaking in the tissue about the periphery of the inoculation site. In two of these instances, for strains 8238 and SN2, the "shothole" effect as seen on cowpea was observed.
The strains of X. c. pv. glycines appeared to be pathogenic on both soybean and kidney bean, but not on cowpea or alfalfa (Table 32). The pathogenic responses appeared as watersoaked lesions, with
the lesion being slightly chlorotic in soybean. The incompatible
response in cowpea appeared as wine red in color like that seen with X. c. pv. phaseoli.
From spray inoculations on alfalfa, only X. c. pv. alfalfa, resulted in a pathogenic response. After 7 days, small water-soaked leaf spots which later turned necrotic appeared. X. c. pv. Jlfalfa was also pathogenic on kidney bean, and appeared so slightly on soybean. There were small water-soaked spots which occurred at the inoculation sites on soybean. The wine red incompatible response was observed on cowpea.
A strain of X. c. pv. malvacearum, which was not a pathogen of legumes, resulted in only a slight tissue discoloration at the inoculation site on bean and soybean. A null reaction, as observed with the water-only control, was observed on cowpea. The water-only control also resulted in a null reaction on the other hosts tested. A summary of the plant reactions observed for the inoculation experiments is given in Table 3-3.
In separate inoculation tests, a strain of X. c. pv. cyamopsidis"'''" was not found pathogenic on alfalfa, kidney bean, cowpea, or soybean and a X. c. pv. Disi strain was not found pathogenic on kidney bean. Overall, most of the reactions observed conformed to those reported in the literature, with some exceptions. Variation was evident among strains of a given pathovar in addition to that between pathovars. The X. c. pv. alfalfae strains appeared to have overlapping host ranges, which extended to bean and pea in addition to alfalfa. The kidney bean cultivar, California Light Red, appeared susceptible to a
Table 3-3. Host range observed for pathovars of Xanthomonas campestris pathogenic to legume host plants.
Inoculation reaction on host plantsa Pathovar G.max M.sativa P.sativum P.vulgaris V.unguiculata
alfalfaeb + +c + _d
glycines + nt +
phaseoli -e ntf +
vignicola -e nt +9 +
a + = compatible, = incompatible, nt = not tested. Host plants were
Glycines max (soybean) cv. Evans; Medicago sativa (alfalfa) cv.
FL-77; Pisum sativum (pea) cultivar not known; Phaseolus vulgaris (bean) cv. California Light Red; and Vigna unguiculata (cowpea) cv.
California Blackeye #5.
b Other susceptible hosts reported are Trigonella and Melilotus (6).
c Reported as susceptible, but not tested.
d Reported as positive in the literature (6). e Reaction appears negative for most strains, but occasional limited
water soaking is evident.
f There is a report of pathovar phaseoli on Pisum lunatus (6).
g Symptoms are similar to those seen in V. unguiculata. h The natural host belongs to the genus Cyamopsis.
representative strain from each of the X. campestris pathovars of legumes tested, with the exception of X. c. pv. cyamopsidis. Various phenotypic responses were observed on the cotton and hibiscus cultivars with the X. c. pv. malvacearum strains tested. In general, it appeared that those strains of X. c. pv. malvacearum derived as pathogens from cotton were pathogenic to one of the hibiscus cultivars tested. None of the X. c. pv. malvacearum strains derived as pathogens from hibiscus were pathogenic on the cotton cultivars in these tests.
From the inoculation tests on the cotton host differentials with
the cotton derived strains of X. c. pv. malvacearum, it appeared that three races of the pathogen were being used. All of these strains were pathogenic on Acala 44 (susceptible host), but could be differentiated into races by the other cotton lines containing different resistance gene backgrounds. The X. c. pv. malvacearum
strain N was pathogenic on all the cotton lines tested, while strain H was only pathogenic on Acala 44. Strains FL79 and TX84 appeared to be
the same race because they were pathogenic on the same six out of eight cotton lines tested (Table 3-4). However, these two strains
differed in reactions on the three hibiscus cultivars inoculated. Strain FL79 appeared pathogenic on the three cultivars, but TX84 gave a null response similar to the water-only control. The two other X. c. pv. malvacearum strains (N and H) appeared identical in reaction with the three hibiscus cultivars, giving pathogenic responses to 2 cultivars, and a null response on the remaining cultivar (Table 3-4).
Table 3-4. Malvaceous host plant reactions to inoculation with strains of X. campestris pv. malvacearuma
Strain 44 B1 B2 B3 B5 BIN 101 Gregg Hlb H2 H3
N + + + + + + + + + 0 +
H + + 0 +
FL79 + + + + + + + + +
TX84 + + + + + + 0 0 0
Xl0 +/- + +
X27 0 0 0 0 0 0 0 0 0 0 0
X52 + +
X102 + +
X103 + +
X108 + + +
83-4244 + +
M84-ll 0 0 0 0 0 0 0 0 0 0 0
84-1093c0 0 0 0 0 0 0 +/- 0
a + = compatible reaction, = incompatible reaction, 0 = null
response, and +/- = intermediate reaction.
b H. rosa-sinensis cv. "Brilliant Red".
c Strain 84-1093 is X. c. pv. esculenti.
With the eight X. c. pv. malvacearum strains derived from hibiscus, two gave null reactions on all of the cotton and hibiscus cultivars tested (Table 3-4). The other strains derived from hibiscus
were pathogenic on at least two of the three hibiscus cultivars tested, but were not pathogenic on any of the cotton cultivars. In a
separate experiment, one of the strains (M84-11) did appear to give pathogenic symptons on Acala 44, while another strain (83-4244) gave a hypersensitive response. In this instance the mean temperature was at 25 C, rather than 30 C.
An additional strain tested, X. c. pv. esculenti strain 84-1093, which was a pathogen derived from okra, gave a null reaction on seven of eight cotton cultivars and one of three hibiscus cultivars. The
other cotton or hibiscus cultivars gave hypersensitive responses (Table 3-4).
All of the strains tested were mucoid on NYGA medium, produced xanthomonadin, and hydrolyzed esculin and casein. All of the strains tested also appeared to be positive for starch and gelatin hydrolysis, and production of lipase (Table 3-5). Six of the seven X. c. pv.
vignicola strains, and one strains of X. c. pv. phaseoli appeared to hydrolyze gelatin slower than the other X. campestris strains. Cellulase activity, indicated by pitting of the solid medium, was observed for all seven of the pathovars tested. However, not all
strains of X. c. pv. phaseoli appeared to have cellulase activity, and X. c. pv. alfalfae strains were weaker in this activity. None of the strains tested exhibited pectolytic activity at pH 4.5, suggesting a
Table 3-5. Physiological reactions of strains of X. campestris.
Physiological Test Mediuma
Na polypectate Pectin
Pathovar Starch Gelatin Cellulose 4.5 7.0 8.5 5.0 Lecithinb
KS + + w +
FL + + w +
13D5 + + + + + +
B-9-3 + + + + + +
1717 + + + + + +
17915 + + + + + +
S-9-8 + + + + + +
H + + +
N + + +
EK11 + + +
Xph25 + + + +
Xpfll + + +
Xpa + w +
Xpll + + +
82-1 + + +
82-2 + + +
LB-2 + + +
SC-3B + + + + +
XP2 + + +
XP-JL + + +
XP-JF + + +
XP-DPI + + +
XP1 + + + + + +
A81-331 + + + + + +
C-1 + w + + + +
CB5-1 + w + + + +
Xvl9 + w + + + +
SN2 + w + + + +
432 + w + + + +
82-38 + w + + + +
a All strains were mucoid, esculin positive, milk proteolytic, and
lipase positive. + = positive, = negative, and w = weak reaction.
b A lecithinase positive reaction was observed as a clearing zone around
the colony plus precipitation in the medium.
lack of polygalacturonase activity. At pH 7.0 and 8.5, pectolytic activity was observed for all strains of X. c. pv. cyamopsidis, X. c. pv. glycines, X. c. pv. _isi, and X. c. pv. vignicola. Only one of the
13 X. c. pv. phaseoli strains tested (SC-3B) had pectolytic activity. None of the strains degraded pectin as a carbon source, only sodium
Polyacrylamide Gel Electrophoresis
Total protein was extracted from 18 strains, representing seven different pathovars of X. campestris. The pathovars included were X. c. pv. alfalfae, X. c. pv. cyamopsidis, X. c. pv. glycines, X. c. pv.
malvacearum, X. c. pv. phaseoli, X. c. pv. pisi, and X. c. pv. vignicola. All of the pathovars, except X. c. pv. malvacearum (Fig. 31, lane P), are known pathogens of legumes. One of the X. c. pv.
phaseoli strains consisted of a biochemical variant of this pathovar, X. c. pv. phaseoli var. fuscans (Fig. 3-1, lane J). One of the X. c. pv. malvacearum strains was derived from hibiscus, rather than from cotton (not shown).
In general, there were few observable differences in the protein banding patterns for the X. campestris strains. Some minor differences in banding patterns between different pathovars were observed, but variation to the same extent was also present within a given pathovar.
The plant reactions observed from the X. campestris host range study conformed to expectations, in most cases. However, it was evident
that pathogenic variation in strains associated with a given pathovar could occur, as shown by the race-specific interactions with the X. c.
A B C D E F G iJ K L M N 0 P
Figure 3-1. Polyacrylamide gel electrophoresis of total proteins from strains of X. campestris. Lanes A and B, X. c. pv. alfalfae; lanes CF, X. c. pv. glycines; lanes G-I, X. c. pv. phaseoli; lane J, X. c. pv. phaseoli var. fuscans; lane K, X. c. pv. pisi, lane L, X. c. pv. cyamopsidis; lanes M-O, X. c. pv. vignicola; and lane P, X. c. pv. malvacearun (cotton). Protein size is labeled in kilodaltons.
pv. malvacearum strains. In the legume study, it was interesting to note that different hosts gave different phenotypic responses to different incompatible pathovars of X. camnpestris. These different incompatible phenotypes suggest that different host resistance genes may be involved in each of these incompatible interactions. Although the hypersensitive reaction is commonly considered to be the general resistance mechanism of plants against bacteria, these data suggest a more specific response, unique to each incompatible reaction (33,49). No formal studies on the hypersensitive reaction were implemented, so
the significance of this observation will require further investigation. Differences in the phenotypes which may account for this are time of recognition, cell number, electrolyte leakage, or other bacteria- or plant-conditioned responses. Likewise, a
hypersensitive reaction may have occurred for instances where a null response was observed (27).
The results from inoculation tests on the malvaceous hosts suggested that those strains of X. c. pv. malvacearum derived from cotton are potential pathogens of hibiscus, whereas those derived from
hibiscus are generally restricted to hibiscus. However, in one
experiment, it was found that one of the hibiscus strains caused watersoaked lesions on Acala 44 when the night temperatures were lowered. This observation indicates that some of the interactions may be temperature sensitive. Some resistance genes in cotton are known to be temperature sensitive (9).
Race-specific incompatibility may account for the varied X. c. pv. malvacearum (cotton strains) reactions on hibiscus. I t was
interesting to note that strains FL79 and TX84, both race 16 of X. c. pv. rnalvacearum cotton strains, each differed in reactions on the hibiscus differential lines. Since Acala 44 elicited a hypersensitive
response from 6 of 8 hibiscus strains, possibly Acala 44 contains an undiscovered resistance gene to the hibiscus derived strains. Those
strains for which no hypersensitive reaction was induced may have escaped host recognition, but lack the necessary pathogenicity determinants to incite disease. Avirulence genes do not give
phenotypic expression in non-parasites (88). Bacterial growth
kinetics would help determine if the bacteria are multiplying on these
hosts. This would be useful for the case of X. c. pv. esculenti, for which 1 of 8 cotton hosts were incompatible, and 2 of 3 hibiscus hosts were incompatible.
The data from the protein and physiological tests demonstrated differences between strains, but were not useful for typing strains by pathovar. There was significant variation among strains within a given pathovar. Although the use of SDS-PAGE to study total protein profiles did not differentiate the pathovars, some work has suggested limiting the analysis to just outer membrane proteins to improve resolution of strain differences (70).
These experiments confirmed pathogenicity of the stock strains and clarified inconsistencies in the literature concerning host range. The host range studies indicated that conclusions reached may often be
strain-dependent, and that examination of a large number of strains will reveal a different overall picture than an examination of a few strains. For example, the plant source of X. c. pv. malvacearum
determines whether a strain will have the host range listed for the pathovar (hibiscus and cotton) or a more limited host range (hibiscus). However, the plant source of X. c. pv. alfalfae, does not determine whether a strain will have the host range listed. A strain which attacks both alfalfa and bean is X. c. pv. alfalfae. A strain with a host range limited to bean is X. c. pv. phaseoli. Finally,
several biochemical methods were evaluated for their potential to classify a strain at the pathovar level. Although standard
microbiological tests and protein gels were not adequate for such classifications, an examination of plasmid DNAs (Chapter Four) and chromosomal DNAs (Chapter Six) yielded more rewarding results.
CONSERVATION OF PLASMID DNA SEQUENCES AND
PATHOVAR IDENTIFICATION OF STRAINS OF XANTHOMONAS CAMPESTRIS
More than 125 different pathovars of Xanthomonas campestris (Pammel 1895) Dowson 1939 are currently recognized (6,24), and the primary means for differentiating them is by inoculation of the plant host(s) of that pathovar. It would be a difficult task to inoculate every plant that could serve as a host to an X. campestris isolate, and therefore the potential host range of a given isolate is unknown. Most often, the pathovar name assigned to a strain of X. campestris is determined by the host it was isolated from. Such designations may be artifactual since the primary host may be different from the one the strain was isolated from; some X. campestris pathovars are known to be pathogenic on more than one host. Epiphytes cannot be classified in this pathovar identification system. Possible taxonomic relationships among pathovars are also elusive. It would be helpful if alternative
means to differentiate among X. campestris pathovars were available. Some suggested approaches for differentiating X. campestris pathovars
have included serology (1,94), membrane protein profiles (70), phage-typing (37), and gas chromatography of fatty acids (65,81). These approaches suffer because they are often strain specific,
This chapter contains copyrighted material from the journal Phytopathology. It is reprinted here with permission of the publisher.
dependent on constant environmental parameters, and/or so cumbersome that no extensive evaluative tests have been performed.
Plasmid DNA has been identified in several pathovars of X.
campestris (13,16,30,39,46,52,53,60,80). It is relatively simple to extract large numbers of strains and visualize their plasmids with
standardized alkaline lysis procedures (46,66). To characterize the plasmids, restriction endonucleases are used to digest the DNA into distinct fragments that can be separated by size, resulting in fragment patterns visualized by agarose gel electrophoresis. To date
there have been no systematic attempts to examine the extent of plasmid variation among a large number of strains involving a large number of X. campestris pathovars. Our preliminary studies on
selected pathovars of X. campestris suggested that there was a surprisingly high degree of plasmid sequence conservation within some pathovars (31,52,53). These studies further suggested that plasmids of purified strains were quite stable and hence useful in epidemiological studies (31), similar to those used to monitor the
spread of selected human pathogens (28). The purpose of this study was to survey the extent of variation of plasmid DNAs within a large number of X. campestris pathovars and to determine whether the plasmid
content of strains based on restriction fragment polymorphism and Southern hybridization could be used in differentiating the pathovars of X. campestris.
Materials and Methods
The X. campestris strains used in this study, their pathovar designations, geographic origin, and sources are listed in Table 4-1.
Table 4-1. Strains of X. campestris used for plasmid analysis.
Bacterium Strain designation
(number of strains) (geographic origin) Sourcea
Xanthomonas campestris pathovar
alfalfae (2) KS (Kansas); D.L.Stuteville
FL (Florida); R.E.Stall
argemones (1) 084-1052 (Florida); DPI
begoniae (1) 084-155 (Florida); DPI
campestris (4) XCl (Oklahoma); this study
084-1318 (Florida); DPI
carotae (3) G1 (Idaho),G5,G7 (California); R.E.Stall
citri (5+) X59 (Brazil),X62 (Japan),
X69 (Argentina), X70 (Brazil); E.L.Civerolo Fil (Florida); DPI
cyamopsidis (5) 13D5; C.I.Kado
X0017 (Arizona); J.Mihail
dieffenbachiae (2) 084-729,084-1373 (Florida); DPI
esculenti (1) 084-1093 (Florida); DPI
glycines (3) B-9-3 (Brazil),1717 (Africa),
hederae (3) 084-1789,084-3928,
251G (Florida); DPI
holcicola (2) Xh66 (Kansas); L.Claflin
XH1; this study
maculifoliigardeniae(2) 084-6006,084-6166 (Florida); DPI malvacearum-cotton(32) A,B,E,F,G,H (Oklahoma); M.Essenberg
HV25 (Upper Volta),
Su2,Su3 (Sudan); L.S.Bird
FL79 (Florida); DPI
C,J,K,L (Upper Volta); this study
malvacearum-hibiscus(8) X10 ,X27,X52,X102, X103,X108 (Florida); A.R.Chase
083-4344,M84-ll (Florida); DPI mangiferaeindicae (1) 084-166 (Florida); DPI
pelargonii (1) 084-190 (Florida); DPI
phaseoli (13) EKll,Xph25,Xpfll (Nebraska); M.Schuster
Xpa,Xpll (Wisconsin); A.W.Saettler
82-1,82-2 (Florida); R.E.Stall
LB-2,SC-3B (Nebraska); A.K.Vidaver
XP2 (New York); J.A.Laurence
XP-JL (Kansas); J.L.Leach
XP-JF (Missouri),XP-DPI; this study
(Table 4-1 continued.)
Bacterium Strain designation
(number of strains) (geographic origin) Sourcea
si (1) XPl (Japan); M.Goto
poinsettiicola (1) 083-6248 (Florida); DPI
pruni (3) 068-1008,084-1793 (Florida); DPI
82-1 (Florida); R.E.Stall
translucens (2) 82-1 (Florida); R.E.Stall
XT1; this study
vesicatoria (5) E-3,69-13,71-21,
82-8,82-23 (Florida); R.E.Stall
vignicola (7) A81-331,C-1,CB5-1,
Xvl9,SN2,432,82-38 (Georgia); R.D.Gitaitis vitians (3) 084-2057,084-2848,
084-4348 (Florida); DPI
zinniae (1) 084-1944 (Florida); DPI
unknown (1) G65 this study
X. albilineans (1) Xalb (Florida); M.J.Davis
X. fragariae (1) Xfra (Florida); R.E.Stall
a DPI = Florida Department of Agricultural and Consumer Services,
Division of Plant Industry, Gainesville, Florida.
Some of the stock cultures of X. c. malvacearum were mixtures of different strains, maintained and described as "races" for purposes of screening cotton host differentials and breeding lines for disease
resistance. Described "races" of X. c. malvacearum may or may not be purified strains; as many as seven different strains have been derived from a single race 1 isolate (7). All strains used in this study were repeatedly purified from single colonies and confirmed to be pathogenic on their designated hosts. Broth cultures of bacteria were
grown in a peptone-glycerol medium (per L: 10.0 ml glycerol, 20.0 g peptone, and 1.5 g K2HPO4). Bacterial strains were co only stored at
-80 C in the same medium containing 15% glycerol. Plasmid Extraction and Visualization
Cultures were grown to mid- to late-logarithmic growth phase and extracted by either of two small-scale alkaline lysis extraction procedures (46,66). Extracted DNA was resuspended in TE (10 mM
Tris(hydroxymethyl)aminomethane (Tris), 1 mM sodium
ethylenediaminetetraacetate (Na2EDTA), and 20 ug/ml DNase-free pancreatic RNase; pH 7.6) and digested with either of two restriction
endonucleases, EcoRI or BamHI, using manufacturer (Bethesda Research Laboratories, Gaithersburg, MD) specifications. Plasmid DNA fragments were separated by size using agarose gel electrophoresis (0.6% agarose
(Sigma Type I:low EEO), 2-5 V/cm) in Tris-acetate buffer (40 mM Tris, 1 mM Na2EDTA, adjusted to pH 7.6 with glacial acetic acid). Fragments were visualized by ultraviolet irradiation (302 rrn) after staining
agarose gels in ethidium bromide (0.5 ug/ml). Photographs were taken with Polaroid Type 55 (or Type 57) film using a yellow filter (Tiffen
no. 12). All restriction fragment size estimates were based on the relative mobilities of linear DNA fragments using lambda phage DNA digested with HindIII as molecular size standards. Plasmid sizes were
estimated by addition of the sized restriction digested plasmid DNA fragments. All plasmid experiments were repeated at least twice for each strain examined.
Cloning of Plasmid Restriction Endonuclease Fragments
Plasmid DNA isolation from X. campestris was by a modification of either of two alkaline lysis extraction procedures (46,68). The
extracted plasmids were purified on CsCl-ethidium bromide gradients by
centrifugation at 55,000 rpm in a Beckman VTi65.2 rotor for 17 hr at 20 C. The purified plasmids were digested with the restriction enzyme EcoRI. The cloning vector, pUCD5 (14), was digested with EcoRI, treated with calf intestinal alkaline phosphatase (BoehringerMannheim, Indianapolis, IN), and ligated to the EcoRI digested X. campestris plasmid fragments with T4 DNA ligase (Bethesda Research Laboratories). The ligation products were transformed into E. coli strain ED8767, selecting transformed colonies on Luria-Bertani medium
containing ampicillin (50 ug/ml) or kanamycin (30 ug/ml). Selected colonies were analyzed for the vector containing desired cloned DNA fragments. These general cloning procedures are outlined in Maniatis et al. (66).
Plasmid DNAs were transferred from agarose gels to nitrocellulose
membranes by the method of Southern as described by Maniatis et al.
(66) and hybridized against radioactively-labeled DNA probes. The DNA
probes derived from plasmid DNA of X. campestris pathovars were either cloned restriction digested DNA fragments of plasmid DNA in the cosmid vector pUCDS, or of the complete X. campestris plasmid. DNA probes were labeled in vitro with use of a nick translation kit (Bethesda Research Laboratories) using 32P-deoxycytidine triphosphate and
hybridized against DNA bound to nitrocellulose membranes. The
membranes were pre-hybridized and hybridized in plastic bags at 68 C. The pre-hybridization and hybridization solutions were those described by Maniatis et al. Following hybridization, membranes were washed once
in 2X SSC, 0.5% SDS and washed once in 2X SSC, 0.1% SDS at ambient temperature, and washed two times in 0.1X SSC, 0.5% SDS at 68 C as described by Maniatis et al. (66) for stringent conditions. The
membranes were then air dried and exposed at -80 C to X-ray film (Kodak X-0mat AR) using intensifier screens. Similar methods were used to probe DNA transferred to nitrocellulose using a dot-blot manifold (Schleicher and Schuell Inc., Keene, NH). All hybridization experiments were repeated at least once.
Detection of Plasmid DNA
Indigenous, cryptic plasmids were detected in all strains of the following X. campestris pathovars: cyamopsidis, dieffenbachiae, glycines, malvacearum (cotton), pelargonii, phaseoli, pruni,
vesicatoria, vignicola, and vitians (Table 4-2). Plasmids were not detected in any strains of X. campestris pathovars alfalfae, argemones, begoniae, carotae, esculenti, holcicola, maculifoliigardeniae, malvacearum (hibiscus), mangiferaeindicae, pisi,
Table 4-2. Detection of plasmid DNA in strains of X. campestris.
No. of strains
Bacterium No. of strains tested Pathogenicitya
X. campestris pathovar
alfalfae 0/ 2 P
argemones 0/ 1 I
begoniae 0/1 I
campestris 2/ 4 R,I
carotae 0/ 3 R
citri 17/44 P,R,I
cyamopsidis 5/ 5 R
dieffenbachiae 2/ 2 I
esculenti 0/ 1 I
glycines 3/ 3 P
hederae 2/ 3 I
holcicola 0/ 2 R
maculifoliigardeniae 0/ 2 I
malvacearum cotton 32/32 P
malvacearum hibiscus 0/ 8 P
mangiferaeindicae 0/ 1 I
pelargonii 1/ 1 R
phaseoli 13/13 P
pisi 0/ 1 R
poinsettiicola 0/ 1 I
pruni 3/ 3 I
translucens 0/ 2 R
vesicatoria 5/ 5 P
vignicola 7/ 7 P
vitians 3/ 3 I
zinniae 0/ 1 I
X. albilineans 0/ 1 R
X. fragariae 0/ 1 I
a p = Pathogenicity of strains confirmed on appropriate host, conforms
to current available information for particular pathovar.
R = Received as named pathogen, appropriate host specificity and pathovar designation of strain assumed.
I = Isolated as a pathogen on host appropriate for designated pathovar; characterization of strain(s) incomplete.
poinsettiicola, translucens, and zinniae. Plasmids were found insome,
but not all strains of X. campestris pathovars campestris and
hederae. Similarly, plasmids were found in all type strains of X. c. pv. citri (A, B, and C types); this is in agreement with a previous plasmid study of this pathovar (13). However, not in all strains of X. campestris isolated from leaf spots of citrus in Florida contained
It appeared that plasmid-containing strains of X. campestris carried from one to three plasmids based on electrophoresis of extracted plasmid DNA. For example, a majority of the
plasmid-containing X. c. malvacearum strains contained only one plasmid, but some carried two or more. When plasmid-containing
strains of X. campestris were purified by repeated single colony isolation, the plasmid DNA content appeared to be stable. However, variation was present in bacterial stocks which were known to have been serially transferred in agar medium over a period of years. Restriction Endonuclease Profiles
Plasmid profiles for X. campestris were variable in over 60 different strains tested. Plasmids were placed into classes based on the variability of restriction endonuclease (EcoRI) digestion profiles on agarose gels. When a different restriction endonuclease (BamHI) was used, the plasmid profiles were placed into the same classes. In all cases, strains which belonged to the same plasmid class also belonged to the same pathovar. There was obvious variability within
classes, but there also appeared to be conservation of some DNA fragments of identical sizes (Fig. 4-1). Undigested plasmids were not
reliable for strain classification because several strains had plasmids of apparently identical size, but they were clearly different
after digesting the plasmid DNAs with restriction enzymes. By adding up the DNA fragment sizes yielded by restriction digests, plasmids in X. campestris were estimated as ranging from about 3 to 200 kb (kilobase pairs) in size. Estimation of some of these sizes were
difficult for some strains due to the presence of more than one plasmid.
Plasmid profiles of strains of X. c. cyamopsidis, X. c. glycines,
X. c. malvacearum, X. c. phaseoli, and X. c. vignicola were compared. Restriction fragment length polymorphism was evident within each of these pathovars. Although more than one plasmid was present in some of these strains, a subset of restriction fragments of similar length and overall pattern appeared to be consistent for strains within a given pathovar (Figure 4-2).
Four highly polymorphic plasmid variants were found in 17 out of 44 Florida strains tested, but Southern hybridization revealed no homology between some of the plasmids (31). Furthermore, there were no similarities in plasmid digestion patterns between X. c. pv. citri type A, type B and any of the Florida citrus leaf spot strains that carried plasmids (Figure 4-3). These Florida isolates are presumed to be X. c. citri, but they grow well and also cause disease symptoms on
kidney bean and alfalfa, thus making their pathovar status questionable (34).
A B C D E F G H I J K L M N OP Q R
Figure 4-1. Plasmid DNAs from strains of Xanthomonas campestris pv. malvacearum (Xcm) digested with restriction endonucleases EcoRl (lanes
B-I) and BamHI (lanes J-Q). Lanes shown above contain: A and R, lambda HindIII; B and J; Xcm J; C and K, Xcm N; D and L, Xcm H; E and M, Xcmn V; F and N, Xcm Z; G and O, Xcm Q; H and P, Xcm X; and I and Q, Xcm D.
111111 1 11122 2 2 2 22233333333334444444
123456789012345 6789012345678901 234567890123456
f I I I I E I I I I I I I,7T7 7 I
________I___I____,_,________i__F J I II I I I i J
Figure 4-2. Graphic representation of plasmid restriction fragment profiles for Xanthomonas campestris pvs. cyamopsidis (Xcc; lane 1), vignicola (Xcv; lanes 2-8), phaseoli (Xcp; lanes 9-17), phaseoli var. fuscans (Xcpf; lanes 18-19), glycines (Xcg; lanes 20-22), and
malvacearum (Xcm; lanes 23-46) digested with EcoRI. Lanes shown above contain 1, Xcc 13D5; 2, Xcv SN2; 3, Xcv A81-331; 4, Xcv C-1; 5, Xcv CB5-1; 6, Xcv Xv19; 7, Xcv 82-38; 8, Xcv 432; 9, Xc EK11; 10, Xcp 82-2; 11, Xcp Xpa; 12, Xcp Xpll; 13, Xcp XP2; 14, Xcp XP-JF; 15, Xcp XP-DPI; 16, Xcp Xph25; 17, Xcp 82-1; 18, Xcpf Xpfll; 19, Xcpf SC-3B; 20, Xcg B-9-3; 21, Xcg 17915; 22, Xcg 1717; 23, Xcm J; 24, Xcm L; 25,
Xcm C; 26, Xcm O; 27, Xcm K; 28, Xcm N; 29, Xcm A; 30, Xcm B; 31, Xcm E; 32, Xcm F; 33, Xcm G; 34, Xcm H; 35, Xcm S; 36, Xcm W; 37, Xcm I; 38, Xcm V; 39, Xcm D; 40, Xcm M; 41, Xcm X; 42, Xcm U; 43, Xcm Q; 44, Xcm R; 45, Xcmn Y; and 46, Xcm Z. Calculated DNA fragment sizes are represented here by a linear scale whereas migration of DNA fragments under electrophoresis conditions approximates a logarithmic scale which is inversely proportional to the molecular weight of the DNA fragment.
39 A B C D
Figure 4-3. Plasmid DNAs from strains of X. campestris pv. citri digested with restriction endonuclease EcoRI. Lane A, strain F11 (Florida type); lane B, strain X59 (A type); lane C; strain X62 (A type); and lane D, strain X69 (B type). Sizes shown in kilobase pairs.
Initial plasmid comparisons were done on strains of X. c. malvacearumn. Whole purified plasmid DNA from X. c. malvacearum strain
X, which contains only one plasmid, was hybridized against EcoRI digested plasmid DNAs of other strains of the same pathovar (not shown) This initial comparison demonstrated that the plasmids,although differing slightly in digestion patterns, were quite
homologous as the radiolabeled plasmid hybridized to almost all EcoRI fragments of the other strains.
A 4.5 kb EcoRI plasmid fragment of X. c. malvacearum strain N was cloned into the vector pUCD5 and used as a hybridization probe against
plasmid DNA from other strains of the same pathovar (Fig. 4-4). This plasmid DNA fragment hybridized to plasmids from all but one strain of X. c. malvacearum. This probe hybridized to EcoRI fragments of 4.5 kb
size (lanes B-J, M-N, Fig. 4-4) in several other strains of X. c. malvacearumn. Additionally, the probe hybridized to more than one of the EcoRI plasmid fragments in these strains, suggesting that some sequences on the cloned DNA are repeated in other parts of the plasmid DNA. The pUCD5 vector alone did not hybridize to any X. c. malvacearum plasmid fragments. Because pUCD5 is a cosmid vector and contains the cos site of lambda phage DNA, hybridization of the vector
to the corresponding DNA fragment of the molecular weight marker containing the cos site was observed. Plasmid DNA from X. c.
malvacearum strain Su2 did not hybridize to the 4.5 kb probe, and did not hybridize to any other plasmid fragments from X. c. malvacearum strain N. Another X. c. malvacearum strain (Ch2), which did not have
A BC D E F G H I J K L M N A B C D E F G H I J K L M N
Figure 4-4. Plasmid DNAs from strains of Xanthomonas campestris pv. malvacearum (Xcm) digested with restriction endonuclease EcoRI (left) and autoradiograph of plasmid DNAs probed with a clone containing a 4.5 kb EcoRI plasmid fragment from Xcm strain N (right). Lanes shown above contain: A, lambda HindIII B, Xcm H; C, Xcm W; D, Xcm V; E, Xcm X; F, Xcm Q; G, Xcm Y; H, Xcm L; I, Xcm N; J, Xcm HV25; K, Xcm Ch2; L, Xcm Su2; M, Xcm FL79; and N, Xcm TX84.
a 4.5 kb EcoRI plasmid DNA fragment, did have two other fragments (ca. 8 and 10 kb) which hybridized strongly to the 4.5 kb probe. In two other X. c. malvacearum strains (FL79 and TX84), which appeared to have multiple plasmids, the hybridization signal was weak for the 4.5
kb fragments as compared to larger EcoRI fragments (ca. 23 kb). Similar hybridization studies were done with cloned plasmid fragments from other pathovars. When the plasmid fragments selected were
smaller, they were much more specific. Two different probes were
constructed from 2.0 kb and 2.3 kb EcoRI plasmid fragments derived from X. c. phaseoli (strain XP2). When hybridized against X. c.
cyamopsidis, X. c. glycines, X. c. phaseoli, and X. c. vignicola, each of these probes only hybridized to X. c. phaseoli (Fig. 4-5). The 2.3
kb plasmid probe hybridized to similar sized fragments in other X. c. phaseoli strains, including X. c. phaseoli var. fuscans (strains SC3-B (not shown) and Xpf11). These X. c. phaseoli var. fuscans strains differ from typical X. c. phaseoli strains in that they produce an extracellular dark brown, melanin-like pigment in culture; otherwise,
they are considered similar. However, the 2.0 kb plasmid probe did not hybridize to the X. c. phaseoli var. fuscans strains, which did not have the corresponding 2.0 kb EcoRI fragment in their plasmid profile (lane C, Fig. 4-5, and strain SC3-B (not shown)). Repeated hybridizations with Southern transfers containing these strains had the same results.
DNA probes were also hybridized against total DNA of other X. campestris pathovars fixed onto a nitrocellulose membrane by use of a dot-blot manifold apparatus. Radiolabeled total plasmid DNA from
A B C D E F G H I J K L A B C D E F G HI J K L ABC D E F G H I J K L
2.3 MAW 4-- bI4
Figure 4-5. Plasmid DNAs from strains of Xanthomonas campestris pvs. phaseoli (Xcp; lanes B, D-I), phaseoli var. fuscans (Xcpf; lane C), and cyamopsidis (Xcc; lane K) digested with restriction endonuclease EcoRI (left) and autoradiographs of plasmid DNAs probed with a clones containing a 2.0 kb EcoRI fragment (center) and a 2.3 kb EcoRI plasmid fragment (right) from Xcp strain XP2. Lane J contains chromosomal DNA of X. campestris pv. alfalfae (Xca) digested with EcoRI. Lanes shown above contains: A and L, lambda HindIII; B, Xcp Xph25; C, Xcpf Xpfll; D, Xcp EK11; E, Xcp 82-1; F, Xcp Xpa; G, Xc G65; H, Xcp Xpll; I, Xcp XP-JF; J, Xca FL; and K, Xcc 13D5. Xc G65 is a strain isolated from bean which was determined to be nonpathogenic and contains no plasmid.
Table 4-3. Hybridization of radiolabeled plasmid probes to total DNA of pathovars of Xanthomonas campestris and one other Xanthomonas species.
Bacterium (No. tested) N80 N4.5 V2.3 P2.3
X. campestris pv.
alfalfae (1) +
carotae (1) +
citri (3) + +/_b +/-b +/_b
cyamopsidis (1) + + +
glycines (1) + +
hederae (1) +
holcicola (1) +
malvacearum cotton (6) +/-c +/_c +
malvacearum hibiscus (2)
phaseoli (1) + + +
pruni (1) + +
translucens (1) +
vesicatoria (1) + + +
vignicola (1) + + +
X. albilineans (1)
a + = hybridization observed, = no hybridization observed.
N80 = plasmid DNA derived from X. c. pv. malvacearum strain N
(about 80 kb); N4.5 = cloned EcoRI plasmid fragment (4.5 kb) from strain N; V2.3 = cloned EcoRI plasmid fragment (2.3 kb) from X. c.
pv. vignicola strain SN2; P2.3 = cloned EcoRI plasmid fragment
(2.3 kb) from X. c. pv. phaseoli strain XP2
b Strain FL-11 was negative, only plasmid DNA used.
c Strain S2 was negative.
X. c. malvacearum strain N, which carries two plasmids, hybridized to DNA of thirteen out of twenty-three X. campestris pathovars tested (Table 4-3). Of these thirteen, seven cross-hybridized to the 4.5 kb
subcloned fragment of X. c. malvacearum strain N. Plasmids were not detected in some of the pathovars which hybridized to the probe. A 2.3 kb subcloned plasmid fragment from X. c. vignicola hybridized to only six of the 23 pathovars tested. Plasmids were present in all six
of those pathovars which hybridized to the probe. A 2.3 kb cloned plasmid fragment of X. c. phaseoli hybridized strongly to other strains of the same pathovar, and weakly to 2 of 3 different X. c. citri strains tested. The X. c. phaseoli probe appears to have
hybridized to chromosomal DNA of X. c. citri in this case, as the probe did not hybridize against Southern transfers of EcoRI digested plasmid fragments of X. c. citri strains (not shown).
It is suggested that there is extensive conservation of plasmid
DNA sequences (as represented by conserved restriction fragments) within, but not usually among, pathovars of X. campestris. Plasmid DNA
fragment patterns will be identical if there is no rearrangement of the DNA sequence at the restriction enzyme recognition site (a six base pair sequence for EcoRI and BamHI), if no new restriction fragments are created within the fragment, and if there are no major additions or deletions causing a change in fragment size. Given these possibilities, it was surprising to find so little restriction fragment length polymorphism of plasmids within a pathovar, especially when strains obtained from different continents were compared.
Southern hybridization analyses confirmed that plasmid DNA fragments of similar size were in fact highly homologous. For example, X. c. malvacearum strain N, isolated in North America, had a restriction digest pattern identical to one of the African strains (K) (Fig. 4-2).
All North American strains except strain N form a distinctive pattern subgroup and the African strains form a somewhat different subgroup. It is possible that either all the African strains are derived from
strain N, or that strain N was introduced to the North America from Africa. The latter possibility seems more likely, since the African strains were isolated from more than one location in Africa. The
conservation of overall restriction fragment profiles of plasmids from geographically isolated populations and the extent of homology seen in Southern hybridizations strongly suggests that plasmid sequences are both highly conserved and stable.
Conversely, it was surprising to find so much polymorphism of plasmids between pathovars. Homology among plasmids in different strains of one pathovar of Pseudomonas syringae (pv. glycinea) has
been reported (15). However, identical plasmid profiles were found present in more than one pathovar of P. syringae (77). In the present study, similar plasmid profiles were not found in more than one pathovar. In addition, cloned plasmid fragments were identified which failed to hybridize to plasmids of different pathovars. For example,
a cloned 2.3 kb plasmid fragment of X. c. phaseoli, which hybridized to all strains of that pathovar, failed to hybridize to total DNAs of
22 other pathovars. This DNA fragment is highly conserved and is apparently pathovar-specific, with one exception (Table 4-3).
Plasmid homology between X. c. phaseoli and X. c. phaseoli var. fuscans was revealed by hybridization that was not apparent by restriction digest patterns. There were distinct differences in
plasmid digestion patterns of X. c. phaseoli var. fuscans in comparison to typical X. c. phaseoli strains. The fact that both X. c. phaseoli and X. c. phaseoli var. fuscans have an identical host range suggests that the homologous plasmid DNA regions that are conserved within some pathovars may encode host range specificity functions. Such functions have been described on plasmids within the
Rhizobiaceae (62,64) This is a testable hypothesis that needs further experimental support.
In X. c. malvacearum, plasmids were found in all 32 strains isolated from cotton, and no plasmids were found in the 8 strains isolated from hibiscus. Cotton and hibiscus each belong to the same plant family (Malvaceae), hence the X. c. malvacearum designation. Atypical symptoms of cotton blight could be artificially produced in cotton using syringe inoculations with hibiscus strains, and in hibiscus using cotton strains. The ability of these strains to be pathogenic on both hosts under natural conditions has not been clearly established. Plasmid DNAs from cotton strains cross-hybridized with one another in Southern analyses, indicating extensive homology was evident between the plasmids. As with X. c. phaseoli, the cotton strains of X. c. malvacearum appeared to carry highly conserved
plasmid DNA sequences (similar in restriction digest sizes and by Southern analyses) which are unique to strains which have a host range on cotton. Plasmid DNA from the cotton strain (N), when radiolabeled,
did not hybridize to chromosomal DNA derived from the hibiscus strain.
Based on our limited pathogenicity tests and on the absence of plasmid DNA sequences in the hibiscus strains, there may be justification to differentiate the cotton and hibiscus strains into different pathovars.
Although some cross-hybridization between plasmids of strains
front different pathovars was detected by dot-blot analyses (Table 43), strains were readily differentiated by restriction digest profiles
of the plasmid DNAs and by hybridization with selected DNA probes to identifiable restriction digested DNA fragments. Cross-hybridization
may be the result of repetitive DNA sequences, insertion elements, or of conserved DNA sequences that are important for the stable
maintenance of plasmids in X. campestris. Examples of plasmid
functions which might be conserved are those involved with replication, incompatibility, or other host dependent factors. in
some instances plasmid DNAs hybridized to total DNA of pathovars in which no plasmids were detected (Tables 4-2 and 4-3). It is suggested
that some sequences encoded on plasmids in some pathovars are located on chromosomes in other pathovars. It is also possible that those plasmids may integrate into the bacterial chromosome and excise again, in a manner similar to those in P. syringae pv. phaseolicola (79,92).
With interest in developing rapid diagnostic methods to identify bacterial pathogens, it is possible that the combined usage of plasmid restriction digest profiles and of plasmid DNA probes may be
sufficient for the identification of some pathovars of X. campestris. This would require that a plasmid be stably associated with a given
pathovar, that plasmid profiles for specific pathovars were known, and
that an appropriate DNA probe consisting of conserved and unique DNA sequences were available. Of the few DNA probes constructed, it was apparent that plasmid DNA sequences were highly conserved within the pathovars studied. In some instances the DNA probes may prove sufficient for pathovar identification, provided they are extensively tested. This apparent stability of the plasmids provides a natural
genetic marker that can be strain specific and perhaps useful in epidemiological investigations. In addition to aiding the
identification for some pathovars of X. campestris, these observations may have taxonomic significance to differentiate these pathogens.
ARE RACE-SPECIFIC AVIRULENCE GENES ENCODED BY PLASMIDS
OF XANTHOMONAS CAMPESTRIS PV. MALVACEARUM?
Bacterial blight of cotton, caused by the pathogen X. c. pv. malvacearum (Snith) Dye, is a destructive disease of cotton, which occurs throughout the cotton-producing areas of the world. The
development of cotton cultivars resistant to bacterial blight is the primary means by which this pathogen is controlled. At least 16 major genes have been identified which condition resistance to bacterial
blight (8). A gene-for-gene pattern of interaction between the plant resistance genes and bacterial avirulence genes has been demonstrated in this host-parasite interaction (33). In screening for resistance to bacterial blight in breeding lines, races of the pathogen have been maintained as mixed cultures, originally derived from infected leaf samples. These mixed cultures have been mass transferred on agar media to prevent the appearance of new race phenotypes. The stability of the race phenotypes of these mixed cultures is periodically checked by inoculating the host differentials.
Plasmids have been detected in several phytopathogenic prokaryotes and have been associated with functions such as pathogenicity (5), resistance to antibiotics (3,88), avirulence (88), and host range (64). With the recent discovery of plasmids in several
pathovars in X. campestris (13,16,30,39,46,52,53,80), it is of interest to understand the function of these autonomously replicating,
covalently closed circular DNAs of the bacterial genome. The purpose of this investigation was to analyze the plasmids in X. c. pv. malvacearum and determine if they influenced race-specificity in cotton.
Materials and Methods
Bacterial Strains and Host Plants
The strains of X. c. pv. malvacearum used for this investigation
are listed in Table 5-1. All of the strains were from single colony isolation of either mixed, or pure cultures. The cotton differentials used included those recommended for differentiating the races of X. c. pv. malvacearum (45); also included were six congenic cotton lines of Gossypium hirsutum cv. Acala 44, which contained single known resistance genes (B genes) to X. c. pv. malvacearum (7). The congenic Acala series included cotton with resistance genes B2, B3, B5, B6, B7,
and BN. The other host differentials were cvs. Acala 44, 1-10B, 101102B, 20-3, Mebane Bl, Stoneville 20, 2B-S9, and Gregg. Plant Inoculations
Pathogenicity tests were conducted by pressure infiltration of abaxial leaf surfaces with bacterial suspensions using a blunt end of
a syringe, and incubating the plants in 30 C growth chambers until symptom were expressed. Symptoms development was monitored over a two week period.
Cultures were grown to mid- to late-logarithmic growth phase and extracted by either of two small-scale alkaline lysis extraction
Table 5-1. Strains of X. campestris pv. malvacearxn from cotton used in race-specificity investigation.
Strain Race Location Source
K HVlb tipper Volta L.S.Bird
N 2a Texas L.S.Bird
J HV3 Upper Volta L.S.Bird
C HVla Upper Volta L.S.Bird
0 2b Texas L.S.Bird
L HV7 Upper Volta L.S.Bird
A 2bSr Oklahoma M.Essenberg
B 2aSr Oklahoma M.Essenberg
E 4-1 Oklahoma M.Essenberg
F laSr Oklahoma M. Essenberg
G lbSr Oklahoma M. Essenberg
5 3-22 Oklahoma M.Essenberg
H 4-2 Oklahoma M.Essenberg
W 6 Texas L.S.Bird
1 3 Oklahoma M.Essenberg
V 3 Texas L.S.Bird
D 2 Texas L.S.]Bird
M 15 Texas L.S.Bird
X 18 Texas L.S.Bird
U 1 Texas L.S.B3ird
Q la Oklahoma M. Essenberg
R lb Oklahoma M. Essenberg
Z 7b Texas L.S.Bird
y 7a Texas L.S.Bird
T 18 Oklahoma M. Essenberg
Chl -Chad L.S.Bird
Ch2 -Chad L.S.Bird
H-V25 -Upper Volta L.S.Bird
Su2 -Sudan L.S.Bird
Su3 -Sudan L.S.Bird
FL79 16 Florida this study
procedures (46,66) as described in Chapter Four. Plasmids were
digested with either EcoRI or BamHI and DNA fragment patterns
visualized by agarose gel electrophoresis. All plasmid experiments were repeated at least twice for each strain examined. Additionally, EcoRI plasmid fragments of X. c. pv. malvacearum strains were cloned into the vector pUCD5 (14) and maintained in E. coli ED8767 as described in Chapter Four. These cloned fragments were used for subsequent plasmid studies.
To determine if plasmid encoded functions were involved in pathogenicity, attempts were made to expel, or cure, plasmids from the
wild-type strains by including ethidium-bromide or sodium lauryl sulfate (SDS) in the broth culture. Cultures were grown in peptoneglycerol broth for two 48 hr cycles, being adjusted to an optical density (OD600nm) of 0.03 at the beginning of each cycle. A dilution
series from the broth cultures were plated onto peptone-glycerol agar and single colonies were selected for inoculation on cotton host differentials.
Cloned genes encoding race-specific avirulence in X. c. pv. malvacearum have been identified and isolated (33), and it was of interest to determine if these genes were plasmid-borne. Using E. coli ED8767 containing the cloned plasmid fragments in pUCD5 as donors, a tri-parental mating scheme was used to move the cloning
vector into strains of X. c. pv. malvacearum by conjugation (22,69). Into the recipient, X. c. pv. malvacearum strain N (virulent strain),
were mated cloned plasmid fragments (in pUCD5) from X. c. pv.
malvacearum strain H (avirulent strain). As a control, plasmids
fragments from strain N were also moved into strain H. The helper
plasmid used for mobilization of the plasmids from E. coli into X. campestris were pSa322 or pRK2013 (22,93). The donor and helper
strains were grown overnight at 37 C in LB broth with appropriate antibiotic selection pressure. The antibiotic was washed from the cells prior to mating by centrifuging and resuspending the cells in 0.7% NaCl. The X. campestris recipients were grown overnight at 30 C in MOPS minimal medium broth (69). Donors and helpers at a
concentration of 107 cells and recipients at 107 cells were mixed in a small volume onto sterile 25 mm cellulose-acetate filter membranes and placed onto peptone-glycerol medium for 6 hours. The membranes were then transferred to MOPS minimal medium with 25 ug/ml kanamycin and incubated for 72 hours at 30 C. Bacterial growth from filters was
diluted in 0.7% NaCl and plated for single colony isolation on MOPS minimal medium with 25 ug/ml kanamycin. Colonies were selected and inoculated on the Acala cotton differentials containing known resistance genes.
Plasmid Origin of Replication and Mobilization
An assay was included to determine if X. c. pv. malvacearum contained self-mobilizing plasmids. The E. coli clones containing
cloned plasmid DNA fragments from X. c. pv. malvacearum strain X were mated into spectinomycin resistant X. c. pv. malvacearum strain N in the presence, or absence, of the E. coli helper plasmid strain with pRK2013. Transconjugants were selected on MOPS minimal medium with kanamycin (25 ug/ml), or directly on POPS (33) containing kanamycin (30 ug/ml) and spectinomycin (100 ug/ml).
Plasmids and Race-Specificity
All of the strains of X. c. pv. malvacearum tested contained at least one plasmid. The plasmid sizes ranged front about 50 to 100 kb. A majority of the American strains of X. c. pv. malvacearum contained
4.8, 4.7, and 4.5 kb EcoRI DNA fragments which appeared to be conserved (Fig. 4-2). The entire 4.5 kb EcoRI DNA fragment appeared to be conserved in all but 1 of the 25 strains tested, regardless of geographic origin.
Comparisons were made between the physical restriction fragment patterns (Fig. 4-2), the original race designations, and reactions on the 14 cotton differentials tested (Table 5-2). The strains of X. c.
pv. malvacearum were grouped into seven plasmid types based on the restriction profiles and the conservation of related restriction fragments. In the plasmid groupings I, III, and IV, a limited level of polymorphism in the restriction fragment profiles for each group were evident, but a majority of the fragments appeared to be
conserved. In plasmid groupings II, V, VI, and VII, no polymorphisms were observed.
The strains which contained plasmids of group I appeared to be pathogenic on a majority of the cotton host differentials. X. c. pv. malvacearum strain J was an exception for it was pathogenic to only 4
of the cotton differentials, whereas the other group I strains were pathogenic on at least 11 of the cotton lines. Analysis of the EcoRI restriction fragment pattern of strain J revealed it was missing a 4.5 kb fragment, and possibly had rearrangements which resulted in a shift of two other EcoRI fragments. With few exceptions, the 4.5 kb EcoRI
Table 5-2. Plasmid groupings and cotton plant reactions to strains of X. campestris pv. malvacearum.
Inoculation reaction of cottonV
A B C D E F G H I J K L M N 0 P Q R S T U
K HVlb I 80.7 2 + + + + + + + + + + + + + + 14 14
N 2a I 84.8 2 + + + + v + v + + + + + + + 12 14
J HV3 Ia 59.5 1 + - + v - + - 3 4
C HVla Ib 59.0 1 + + + v v v + + v v + v v + 7 14
0 2b Ib 59.0 1 + v v + v + v v + v v v 4 12
L HV7 Ic 57.9 1 + - v + + v v v + v v v 4 11
A 2bSr II 82.2 1 + + - + - - v - 3 4
B 2aSr II 83.3 1 + + - + - - v - 3 4
E 4-1 II 83.3 1+- - - v - - - 1 2
F laSr II 83.3 1 + + - + - - v - 3 4
G lbSr II 83.3 1 + + - - - v - 3 4
S 3-22 11 83.3 1+- - - - - - - 1 1
H 4-2 II 80.4 1 + - - - - - - 1 1
W 6 IIIa 88.7 1 + - v - - - - - 1 2
I 3 IIIa 88.4 1 + - v v - - - v 1 4
V 3 IIIb 87.0 1 + - v v v - v - 1 5
D 2 IVa 94.2 1 + + - + - - v + - 4 5
M 15 IVb 93.2 1 + + - + v - - v - 3 5
X 18 V 57.5 1 + v v v + - v v v v 2 9
U 1 V 62.9 1 + v v + v v v v v v 2 10
Q la VI 100.4 2+- - - - - - - 1 1
R ib VI 91.6 2 + - v v - - v - 1 4
Z 7b VII 93.0 1 + v + v v v - v v v v 2 10
Y 7a VII 98.0 1 + - v v - - - 1 3
T 18 1 + + - - - +- 4 4
A = Strain designation. P = Cultivar Mebane B1 (B2 +
B = Original strain, or race, polygenes).
designation. Q = Cultivar Stoneville 20 (B7
C = EcoRI plasmid group. + polygenes).
D = Kilobase pairs of plasmid DNA. R = Cultivar Polygenes (2B-S9) E = Number of plasmids. S = Cultivar Gregg (resistance
F = Cultivar Acala 44. genes unknown).
G = Cultivar Acala B2. T = Sum of cultivars which gave 4
H = Cultivar Acala B3. out of 4 compatible reactions
I = Cultivar Acala B5. (+).
J = Cultivar Acala B6. U = Total of cultivars which gave
K = Cultivar Acala B7. at least 1 compatible reaction
L = Cultivar Acala BN. (H).
M = Cultivar 1-10B BIN V = Plant reactions: +, compatible
(+ polygenes). reaction; -, incompatible
N = Cultivar 101-102B (B2B3 reaction; v, reaction variable
+ polygenes). over 4 replications.
0 = Cultivar 20-3 (BN + polygenes).
fragment appeared to be conserved in all group I strains of X. c. pv. malvacearun. The group I strains K and N each contained two plasmids and also appeared to be the most virulent of all the strains tested in any group, each being pathogenic on all of the cotton differentials tested (Table 5-2). Strain K was isolated in Upper Volta, Africa and strain N was isolated from Texas. Although the plasmids in group I were highly conserved and apparently widespread in the world, there was no correlation between the presence of these plasmids with race specificity.
The strains which were classified into plasrnid group II could be divided into two basically different pathogenicity patterns on the 14 cotton differentials. The strains which were derived from X. c. pv. malvacearun races 1 or 2, and strains derived from races 3 and 4 appeared to have related pathogenicity patterns. The plasmid EcORI
restriction fragments patterns for all seven of these strains appeared identical. Similar variability was observed in the pathogenicity patterns for the other strains placed within groups III, IV, V, VI, and VII. Again, no correlation between the pathogenicity pattern and the plasmid group was apparent. The original race designations of the
strain mixtures of X. c. pv. malvacearu~m did not correspond to the race phenotypes exhibited by the strains purified from the mixtures. For instance, of 5 strains derived from nominal race 1 (F, G, Q, R, and U), there were basically three different pathogenicity patterns observed, none of which corresponded to the reported pattern for race 1. The expected race 1 reaction would have been pathogenic on Acala 44 and Stoneville 2B-S9 (45).
In preliminary tests with the curing agents, it was determined that
0.3% SDS and 20 ug/ml ethidium bromide were the concentrations which allowed growth of X. campestris. Strains did not grow at concentrations of 0.5% SDS or 30 ug/ml ethidium bromide. The
bacterial colonies isolated from ethidium bromide treatments appeared phenotypically identical to the wild-type strains. Several colorless colonies were observed with the SDS treatment; however, over a 48 hr period, color appeared to be restored. Of 11 strains selected from the SDS treatment, all were found to be pathogenic on Acala 44. From plasmid analyses it appeared that 7 of 11 strains were missing the indigenous plasmids (Fig. 5-1).
Several clones containing cloned X. campestris plasmid fragments in
the vector pUCD5 were obtained and transformed into E. coli ED8767. Based upon restriction analyses, only 3 of the 8 EcoRI fragments of strain H, and 6 of 9 EcoRI fragments of strain N were cloned. These
clones carried 30 kb (36%) of the 84 kb strain H plasmid, and 70 kb (88%) of 85 kb strain N plasmids. Each of the E. coli clones were mated into X. c pv. malvacearum. The transconjugants were then selected and inoculated onto the 14 cotton differentials. The results of the inoculations are seen in Table 5-3.
None of the clones derived from plasmid DNA of X. c. pv. malvacearum changed the expected disease reactions of X. c. pv. malvacearum strain H. One clone of X. c. pv. malvacearum strain H, originally thought to carry a large plasmid insert, did carry an
Table 5-3. Pathogenicity of X. campestris pv. malvacearum (Xcm) transconjugantsa on cotton host differentials.Plant reEC-dic&b
N + + + + + + + + + + + + +
Plasid fragamts of X=n H in Xan N: FEA-H + + + + + + + + + + + + +
g~',5 + + + + + + + + + + + +- +
Plasriid fragrmts of Xan N in Xcm H:
UEA-* +- --
O-ransnal fra~et of Xcm H in Xan N: gJF.-1. + + +- + + + + + + + + +
a Cloned EcORI fragment sizes within each clone are: ptJFA-Hl, 4.9 kb;
pUFA-H2, 19.2 and 4.6 kb; pUFA-H5, 22.4 and 4.0 kb; pUFA-NI, 2.3 kb; pUFA-N3, 14.5; ptJFA-N4, 23.5 and 13.4 kb; pUFA-N5, 15.6 kb; pUFA-N9,
13.4 and 9.6 kb.
b + =compatible (pathogenic) and = in-cmpatible (hypersensitive)
A B C D E F G H I J K L
Figure 5-1. Plasmid curing of X. campestris pv. malvacearum with SDS treatment. Lanes A-L are plasmId extractions of strains which were
exposed to 0.3% SDS. Lanes A, E, I, and J appear to contain the indigenous plasmid. Lanes B-D, F-H, and K-L appear to be missing the indigenous plasmid.
avirulence gene. Transconjugants of this clone, pUFA-H1, in strain N
caused incompatibility on the Acala BN line (Table 5-3). Upon further examination of the pUFA-H1 clone by Southern hybridization analysis, it appeared that the cloned insert was not of plasmid origin. The
radiolabeled clone did not hybridize to Southern transfers of plasmid DNA derived from strain H, or 16 other strains of X. c. pv. malvacearum. However, this 4.9 kb cloned insert of pUFA-H1 did hybridize to a 23 kb EcoRI fragment of the plasmid found in strain W. Additionally, the radiolabeled pUFA-H1 clone hybridized to a 4.9 kb EcoRI chromosomal fragment of 18 strains of X. c. pv. malvacearum, and also to another clone, pUFA-702 (33). pUFA-702 was independently
derived and appeared to encode gene-for-gene avirulence on Acala BN. Thus, it was concluded that the pUFA-H1 clone was of chromosomal origin.
Plasmid Origin of Replication and Mobilization
Seven of the eight EcoRI fragments of the single 59.5 kb plasmid found in X. c. pv. malvacearum strain X were cloned into pUCD5 (Table 5-4). All of the fragments were successfully mobilized into a spectinomycin resistant strain of X. c. pv. malvacearum strain N (NSpR) in the presence of the helper plasmid, pRK2013, using MOPS minimal medium selection. In general, it appeared that the larger cloned fragments transferred at a slightly lower rate than the smaller cloned fragments, however the transfer rate was not quantified. One of these clones, pLXF, carrying a 4.7 kb fragment, was found to encode
plasmid transfer function (tra), missing on the vector pUCD5 (14). That is, pXLF could be mated into X. campestris without aid of the
Table 5-4. Selection of plasmid replication genes in X. campestris pv. malvacearum strain X.
Size MOPS K30 POPS K30SP100
Plasmid Insert no helper pRK2013 pRK2013
pLXA 17.0 kb few +
pLXB 14.0 kb NT NT NT
pLXC 9.0 kb + few
pLXD 5.0 kb ++ ++ (ori)
pLXE 4.8 kb +
pLXF 4.7 kb ++ +++ (tra)
pLXG 4.5 kb +++
pLXH 0.5 kb +++ few
a is no growth, + = about 100 colonies, ++ = 100-300 colonies, +++ =
swamped plate, NT = not tested. MOPS K30 is the MOPS minimal medium containing kanamycin (30 ug/ml) and POPS K30SP100 is the POPS complete medium containing kanamycin (30 ug/ml) and spectinomycin (100 ug/ml) pRK2013 is a helper plasmid used in tri-parental matings (22). pLXD contains the putative plasmid origin of
replication (ori) and pLXF contains a putative plasmid transfer (tra) function of the X. campestris pv. malvacearum strain X
helper plasmid. Since the vector pUCD5 is known to carry a
mobilization function, it could not be concluded that the strain X plasmid is self-mobilizable, but it must carry at least some functions required for self-mobilization.
pUCD5 is a broad host range vector with an origin of replication (ori) that functions poorly in X. c. pv. malvacearum. Matings on MOPS minimal medium provide a lengthy period of selection in which plasmid integration can take place, provided a homologous X. c. pv.
malvacearum DNA fragment is cloned on the vector to allow recombination to occur. Rapid selection procedures using complete media containing antibiotics only rarely allows this integration to occur before the recipients die. When transconjugants with X. c. pv.
malvacearum strain X plasmid DNA were selected on POPS complete medium with antibiotics only the clone containing a 5.0 kb EcoRI fragment, pLXD, mated at a high frequency. All other clones either did not grow on POPS with antibiotics, or grew poorly. Because pLXD grew rapidly
on selective medium, whereas other clones did not, it was proposed that the pXLD clone contained a functional origin of replication from the strain X plasmid.
Transconjugants containing pXLD were further analyzed because their growth rate was highly variable. Eleven of the pLXD
transconjugants were selected by colony size and grouped as large, medium or small. Plasmids from the transconjugants were extracted and
analyzed for plasmid content. All, but possibly one transconjugant, were missing a 4.5 kb EcoRI fragment associated with the indigenous plasmid of X. c. pv. malvacearum strain N-SPR (Fig 5-2). It was
ABC D F 1HI J K L M
Figure 5-2. Transconjugants of X. campestris pv. malvacearum N-SpR containing pLXD (top) and hybridized to the plasmid probe pSa4 (bottom) The transconjugants were digested with the restriction endonuclease EcoRI (top). Lanes A and M, Lamda HindII; lanes B-E, large colony-size; lanes F-I, medium colony-size, and lanes J-L, small size transconjugants. All of the transconjugants, except in lane J, appear to have lost the 4.5 kb plasmid fragment seen in wild-type strains. Plasmids in lanes F-I and K-L appear to have had other plasmid rearrangements occur. All strains had an additional 8 kb EcoRI fragment present. The autoradiograph (bottom) shows the
transconjugants probed with a pUCD5 related plasmid, pSa4 (93). The 8 kb fragment hybridized to the probe, but so did other plasmid fragments corresponding to those of X. campestris pv. malvacearum.
presumed that the homologous recombination occurred between the pLXD clone and the indigenous plasmid. As a result, there was no EcoRI fragment which corresponded to the 13 kb pUCD5 vector, or the 5 kb pXLD insert. However, a new 8 kb EcoRI fragment was present.
The plasmid DNA from the X. campestris transconjugants with pXLD was subsequently extracted and transformed into E. coli ED8767. The transformants were selected on Luria-Bertani medium containing kanamycin (25 ug/ml), or ampicillin (30 ug/ml). Resistance genes for
these antibiotics border the EcoRI cloning site of the pUCD5 vector
(14). All of the transformed plasmids were 8 kb in size, which indicated that the new band seen in the X. c. pv. malvacearum
transconjugants was, in fact, a new autonomously replicating plasmid (Fig. 5-3). There was obvious rearrangement in the original pXLD clone which led to a 10 kb deletion. Southern hybridization analyses
revealed the presence of DNA fragments from both pUCD5 and the insert of pLXD in this new vector, named pXD-1 (Figures 5-2 and 5-4).
Of the twenty-five clonal isolates of X. c. pv. malvacearum derived from 11 different nominal races, none of the clonal isolates corresponded to the expected disease reactions on the host differentials for that race. Likewise, there was no correlation between a given plasmid group and the observed pathogenic reactions on the 14 cotton differential lines. Although avirulence genes have been
found on plasmids in other phytopathogenic bacteria (88), no evidence has been found to link plasmids in X. c. pv. malvacearum with racespecific interactions. In these studies, it appears the plasmids
A B C D E F G H I J K L M
Figure 5-3. Transformation of E. coli ED8767 with plasmid DNA from X. campestris pv. malvacearum transconjugants mated with pLXD. Lane A, lamda HindIII; lanes B-M, E. coli ED8767 transformed with DNA extracted from transconjugants mated with pLXD and selected on kanamycin Luria-Bertani plates (see Fig. 5-2). The transformants were digested with restriction endonuclease EcoRI. Only the 8 kb EcoRI fragment was present; the smaller band was undigested DNA.
ABC D E F G H I A BC D E F G HI
MAP "M mmam
Figure 5-4. Plasmid DNA of X. camrpestris pv. malvacearum strain X digested with restriction endonucleases (left) and hybridized to the plasmid probe pXD-1 (right). Lanes A and I, lamda HindIII; lanes B-H, plasmid DNA from X. campestris pv. malvacearum strain X digested with
BamHI (lane B), BglI (lane C), EcoRI (lane D), HindIII (lane E), KpnI (lane F), PstI (lane G), and SalI (lane H). The 5.0 kb EcoRI fragment seen in lane D is the fragment thought to contain the origin of replication (ori).
served as useful markers for the genomic background. Plasmids in X. campestris generally appear to be stable and have in some instances been used as epidemiological markers, such as with the recent Florida outbreak of X. c. pv. citri (31).
However, five of the six strains with group I plasmids were more virulent than the other strains. All of the Upper Volta strains fell into this class, including some North American strains. The Upper
Volta strains were derived from recently discovered races which were
found to be virulent against all the known major genes for bacterial blight resistance (29). The X. c. pv. malvacearum strain J, although
an Upper Volta strain, appeared to be the exception and varied from the other plasmids by the loss, or rearrangement, of a 4.5 kb EcoRI fragment found in all of the other 24 strains tested. Since the other group I strains are highly virulent strains, and known to be lacking a number of cultivar specific avirulence genes (33), it may be possible that this 4.5 kb EcoRI fragment may serve some positive function in pathogenicity.
Nine avirulence genes have been reported cloned from X. c. pv. malvacearum, at least some of which are chromosomally encoded (33). With further testing, it is possible that plasmid-borne avirulence genes may be uncovered. In this study, approximately 50% of the
plasmid sequences derived from an avirulent strain were tested, leaving it still possible to uncover avirulence genes in X. campestris pv. malvacearum.
A high level of race-specific variability has been described for X. c. pv. malvacearum (8). Plasmids may be involved in genome
rearrangement as plasmids in P. syringae pv. phaseolicola have been found to integrate and excise from the chromosome causing rearrangement (79,92). In these studies, pUFA-H!, carrying a
chromosomally encoded avirulence gene, was found to hybridize with a plasmid DNA from X. c. pv. malvacearum strain W. The evidence that plasmids in X. c. pv. malvacearum may be self-mobilizing points to their potential role in variability of the X. campestris species. It is believed that the self-mobilizing, copper resistance plasmid identified in X. c. pv. vesicatoria facilitated the rapid build up of X. campestris strains resistant to field applied copper-containing compounds (88). If avirulence genes are on self-mobilizing plasmids,
rapid shifts in pathogenic races may occur, and the appearance of new races would be facilitated.
The cloning of the X. c. pv. malvacearum plasmid origin of replication may allow for the future construction of other cloning vectors designed specifically for the study of genes in X. campestris. Some of the currently used broad-host-range vectors appear unstable in
X. campestris and are useful because cloned inserts integrate into the chromosome. Although preliminary experiments with the X. campestris ori resulted in vector rearrangement, possibly the resolved form may be applied to the construction of future vectors.
PATHOVARS OF XANTHOMONAS CAMPESTRIS ARE DISTINGUISHABLE
BY RESTRICTION FRAGMENT LENGTH POLYMORPHISM*
The species Xanthomonas campestris (Parmnel) Dowson, a member of
the family Pseudomonadaceae which is always found in association with living plants, comprises over 125 different pathovars (6). The
pathovars have been named by "the plant from which first isolated"
convention. Unfortunately, strains of some pathovars can be
pathogenic on host species in different plant families (57) ; thus possible relationships between pathovars may go undetected, since there is no convenient way to determine pathogenicity for all possible host plants, nor is there a useful systematic method for narrowing the possibilities. Further, and despite a widely held view that all xanthanonads are plant pathogens (85), non-pathogenic members of X.
campestris are often found as ectoparasitic leaf colonizers (epiphytes) (26) and occasionally as endoparasites (65,71). Their inability to provoke a pathogenic plant response renders them unclassifiable in this system. The attention given to damaging plant pathogenic strains may not be taxonomically or ecologically warranted. Genes which function to provoke a pathogenic response may be entirely
different from those which confer host selectivity and parasitic ability (32). Epiphytes and nonpathogenic endoparasites are known to
* This chapter contains copyrighted material from International Journal of Systematic Bacteriology. It is reprinted here with permission of the publisher.
be host selective (19,65). Host selectivity is thought to be a stable
characteristic (84) Since xanthomonads are always found in association with living plants, host selectivity should result in genetic isolation of a pathovar population. over time such isolation and random genetic drift should produce distinctive and stable phenotypic characteristics and therefore, establish conserved genetic markers.
Attempts to differentiate pathovars by methods other than pathogenicity have included rRNA-DNA and DNA-DNA hybridization (21,73), serology (1,94,95), phage-typing (37,59), arnd comparisons of profiles resulting from plasmid and chromosomal DNA restriction enzyme
digests (40,54), protein electrophoresis (70), and gas-chromatography of fatty acids (65). Although all of these methods are useful for specific purposes, few have demonstrated utility to replace, clarify,
or indicate further pathogenicity tests.
We report here the use of restriction fragment length
polymorphism (RFLP) analysis to differentiate pathovars of X. campestris. RFLP analyses have been widely used in the medical field to identify DNA fingerprints specific for inherited diseases or other
genetic loci (2,11,61). These analyses have also been applied to plants (41), bacteria (38), and eukaryotic organelles (67,91). RFLP analyses allow the observance of genetic variation in organisms within defined regions of the genome due to DNA rear rangemen ts or mutations which affect recognition sites for restriction endonucleases. Variation may be examined over a small (a few base pairs) or large (30-40 kb) stretch of DNA depending on the desired level of
polymorphism detection. This variation can be observed for random or selected DNA target sequences within a species against a background of other genomic DNA fragments. In this study, RFLP analyses were applied to investigate the degree of genetic variation among 87 strains of X. campestris, comprising 23 different pathovars. Some of these results have appeared in abstracts (34,55).
Materials and Methods
The bacterial strains used in this study are shown in Table 6-1. These strains were isolated as plant pathogens, identified as X. campestris, and classified into pathovars according to the susceptible host plants involved. Strains were tested for pathogenicity after single colony purification. For some strains, no pathovar assignments were made due to the lack of known plant pathogenic responses. Broth cultures of bacteria were grown in a peptone-glycerol medium (per L: 10.0 ml glycerol, 20.0 g peptone, and 1.5 g K2HP04; pH 7.2). The strains were commonly stored and maintained at -80 C in the same medium containing 15% glycerol.
DNA was extracted from bacterial cultures at mid- to late-logarithmic growth phase. Extraction of bacterial DNA for cosmid library construction was by a modification of the method of Silhavy et al. (87). A critical modification was to wash cells in 50 mM TRIS, 50
mM Na2EDTA, 150 mM NaCl, resuspend in the same buffer containing 150 ug/ml proteinase K, add SDS (sodium dodecyl sulfate) to 1% (w/v), and
Table 6-1. Strains of X. campestris used for RFrAJP analysis.
Pathamr Strain Host Location Sourcea
Xant=mas cmpestris pabiwar
alfalfa ES madicago sativa Famas D.L.Stiteville
EL M. sativa Florida R.E.Stal I
084-1052 Azamm-- mexicEra Florida DPI
begonia 084-155 a3gonia sp. Florida DPI
LT& L Oes txis "Cl Brassica oleracea (cabbage) CkLahma t-ds sb--dy
084-809,084-1236 B. oleracea (cabbage) Florida t-ds sbxly
084-720 TT. cleracea (br. sprouts) Florida tIds sbaly
084--131-8 B. cleracea (broccoli) Florida DPI
carotae 13 Cai-= carota California R.E.Stal I
citri Y59,X70 Citrus Sp. B=il E.L.Civerolo
)62 t= S P. Japan E.L.Civerolo
X69 Citrus so' Argentina E.L.Civerclo
084-3401 citrus SP. Florida DPI
'dis 13D5 tetrapomIcba C.I.Kw3o
X002,XO05, C. t t Arizcna J.Mihail
X016,XD17 C. tet olcim ArizCr3a J.-Mihail
dieffenbachiae 084-729 Anthurim sp. Florida DPI
068-1163 Dieffenbachia sp. Florida DPI
esm1enti 084-1093 Abah s esa. lentas Florica DPI
gl cinas B-9-3 Glycine nax Brazil W.F.Rett
G. max Africa W.F.Flatt
17915 G. max W.F.Flett
S-9-8 E. Wisca-isin W.F.Fett
hs5arz e 084-1789 Ti Inlix Florida DPI
nlcioala &a Sort= vulgare Kansas L.Claflin
maculifohigardeniae 084-6166 (3ardmia sp. Florida DPI
malvacearm DMNOUV, Goss piun 1-drsutLm TMOS ii stpy
XYZUS4 G. hirsitin Tbos tiis sbjJy
ABE,,FGH G. drsqalan Od*Xra M.Es
ChlCh2 G. hdrsatn Chad L.S.Bird
M725 G. himitin Upper Volta L.S.Bird
U2,9-2 G. hir-Rib-m L.S.Bird
FL79 G'. hdrs-l'= Florida ti stxly
mm-eibmaeindicae 084 -116 Fki lfem in ica Florida DPI
miaraTecalans 084-1984 Arctim lappe Florida DPI
pelargonil 084-190,084-1370 u sp. Florida r1pi
lf i EICI,)Zh25,) :M P eolus w1garis Nebraska Mdtister
P. vulgaris Vdsconsin A.W.Saettler
82-1,82-2 P. vulqaris Florida R.E.Stall
LB-2,SC-M P. vulgaris Nebraska A.K.Vidmer
XP2 P. vugaris New York J.A.Lamerre
XP-JL P. vulgaris Kmsas J.L.Leach
)GLJF P. vulgaris llissazi this sb-xty
XP-DPT P. vulgaris this sb--dy
Table 6-1 (continued).
pathaVar Strain Host Lccation SaZCea
pisi Xpi Pisan sativun J--PM nGDto
poirret:tiicc1a 083--6248 El#x)rbia palcherrima Florida CpI
pard 084-1793 Prunis sp. Florida DPI
tonsluoa-z M05 Hcrdan Ep. Mat:zna D.Smds
vesicatcria F,3 C.-psic= arn-in Florida R.E.Stall
75-3 Lw-persiccn esculenbn Florida R.E.Stall
vigniccla A81-331,0-1, Vigna =uicilata Gxxgia R.D.Gitaitis
C25-11 V. Lrxjiumlata Gt=gia R.D.Gitaitis
XV19 qv, V. =juiculata Gam-gia R.D.Gitaitis
432,82-38 V. UTuiculata GB=gia R.D.Gitaitis
vitia-Z ICVa64 Lab-ra Sp. R.E.StaLl
zinnias 084-1944 Zinnia elegant Florida DPI
uial)iln 084-1373 Aiiloder&cn sp. (dieffen.) Florida DPI
084--3928 Fatsia sp. Cha3arae) Florida DPI
084-4348 Alocasia. sp. (vitians) Florida DPI
083-2057 Swrjoniu-n sp. (vitians) Florida DPI
084-2B48 CiSS.3S Sp. Florida DPI
084-1590 an us sp. Florida DPI
251G,084-40 hTpatiem S O- Elcrida DPI
064-6006 Jasrdnium sp. Florida ME
a-agariae xfrai Fragaria sp. Florida R.E.Stall
a DPI = Florida Department of Agricultural and Consumer Services,
Division of Plant Industry, Gainesville, Florida.
heat for 1 hr at 50 C. The sample was then extracted 2 times with phenol/chloroform/isoamyl alcohol (25:24:1; phenol equilibrated to pH 7.8 with 0.1 M TRIS), and the DNA spooled out from the aqueous phase after adding sodium acetate at 30 mM and 2 volumes 95% ethanol. This DNA was then washed in 70% ethanol. Extracted DNA was resuspended in TE (10 mM Tris(hydroxymethyl)aminomethane (TRIS), 1 mM disodium ethylenediaminetetraacetate (Na2EDTA) containing 20 ug/ml DNase-free pancreatic RNase; pH 7.6).
Agarose Gel Electrophoresis
Approximately 5 ug of DNA was digested with EcoRI or BamHI as specified by the manufacturer (Bethesda Research Laboratories). Digested DNA samples were run in a 20x25 can agarose gel (0.6%; type II, Sigma) in TRIS-acetate buffer (40 mM TRIS-acetate, 1 mM Na2EDTA; pH 7.6) with electrophoresis of gels set at 35 V for 14-15 hrs. Fragments were visualized by ultraviolet irradiation (302 nm) after staining agarose gels in ethidium bromide (0.5 ug/ml). Photographs were taken using Polaroid Type 55 (or Type 57) film and a yellow filter (Tiffen no. 12). After the gels were photographed, the DNA was transferred to nitrocellulose by the method of Southern as described by Maniatis et al. (66). Restriction fragment sizes were estimated
using DNA molecular size standards of lambda phage DNA digested with HindIII.
The DNA probes used in this study were derived from a genomic library of X. c. pv. citri strain 3401 constructed into the modified cosmnid cloning vector pUCD5B, which was the vector pUCD5 (14) with a 2
kb Bam I fragment deleted from it. DNA of strain 3401 was partially digested with the restriction enzyme Mbol and fractionated by size on a 10-40% (w/v) sucrose step gradient (66). The pUCD5B vector was
digested with the restriction enzyme BamHI and treated with calf intestinal alkaline phosphatase (Boehringer-Mannheim). Vector DNA and
large fragment size fractions (25-40 kb) from strain 3401 were mixed and treated with T4 DNA ligase (Bethesda Research Laboratories) and recombinant cosmids were packaged in vitro with extracts prepared and utilized according to Scalenghe and Hohn's protocol II as described by
Maniatis et al. (66). Transduction with Escherichia coli HB101 was performed using top agar on Luria-Bertani medium containing kanamycin
(35 ug/ml). Cloned DNA fragments of strain 3401 in the vector averaged 27-38 kb.
DNA clones used as probes were radiolabeled using 30-60 uCi of 32P-deoxycytidine (Du Pont NEN) with use of a nick translation kit (Bethesda Research Laboratories) and separated from low-molecular weight nucleotides on a mini-column of Sephadex G50-100. Southern
blots were pre-hybridized in plastic bags for 4 hr and hybridized with
the DNA probe for 16-18 hrs at 68 C. The pre-hybridization and
hybridization solutions were those described in Maniatis et al. (66).
Following hybridization, membranes were washed once in 2X SSC, 0.5% SDS (lX SSC is 0.15 M NaCl, 0.15 M sodium citrate, pH 7.0) and washed once in 2X SSC, 0.1% SDS at ambient temperature, and washed two times in 0.!X SSC, 0.5% SDS at 68 C as described by Maniatis et al. (66) for 'stringent' conditions. Nitrocellulose membranes were then air-dried
and exposed to Kodak X-Omat AR film at -80 C in cassettes with intensifying screens. Hybridization of the probes to individual strains of X. campestris was repeated at least three times. Restriction Fracment Patterns and Densitometr
From developed autoradiographs, DNA fragment sizes and profiles were determined and hybridization signals measured using a Gilford
Response spectrophotometer equipped with an autoradiograph scanner. Routine scans were done at 600 nm wavelength with 0.5 m aperture and 0.5 nm bandwidth setting. Scanning data was stored by computer. Using above information, comparison were made among strains of a given
pathovar and also strains derived from different pathovars of X. campest ris.
Agarose Gel Electrophoresis
For a given pathovar, the banding patterns appeared very similar
and polymorphism in these strains was apparently limited (Fig. 6-1). In all, DNA from 87 strains of X. campestris, representing 23 different pathovars and including some with no known pathovar status were digested with restriction enzymes EcoRI and BamHI and separated by size on agarose gels. Based on restriction fragment patterns
alone, it was possible to visualize DNA variability among the strains,
but recognition of particular DNA fragment banding patterns was difficult due to the numerous DNA fragments involved (Fig. 6-2). Some of the apparent variation was due to the presence of plasmid bands, which appear brighter against the chromosomal bands due to their relatively higher copy number within each cell.
AB C D E F G H I JK LM NO P Q R S T
Figure 6-1. Genomic DNA of strains of X. campestris digested with the restriction endonuclease EcoRI. Lane A, probe XCT1; lanes B-K, X. c. pv. phaseoli; lane B, JL; lane C, LB-2; lane D, Xph25; lane E, EKll; lane F, 82-1; lane G, 82-2; lane H, Xpa; lane I, XP2; lane J, XP2-1; lane K, JF; lanes L-M, X. c. pv. phaseoli var. fuscans; lane L, SC-3B; lane M, Xpfll; lanes N-O, X. c. pv. alfalfae; lane N, KS; lane O, FL;lanes P-R, X. c. pv. campestris; lane P, XCl (cabbage); lane Q, 084-1318 (broccoli); lane R, 084-720 (brussels sprouts); lane S, unknown X. c. G65; lane T, probe XCT11.
F11r.2.enncDNoftrisoX.cmetsdietdwhte resritin ndnulese coI.Lae roe X~l lneBX.C.pZ
FLigule LX. mc.DN pvf srarnsi 084-.90; lane M, Xigcstpd wphastol vla.fuan SL3; lane X. . 04-1X55; lane D, X. C. pv. tanpsuces 080; lane X. c. pv. vscataa 73; lane Cv. ..pv. igca S--; lane X. c. pv. itiansl ICP6; lane X. C pv. zinniae 084-1944; lane T, probe XOT11.
Two cosmid DNA clones, XCTl and XCTl1 were randomly selected from a genomic library of strain 3401 for use as DNA probes. XCT1 carried
a 30 kb insert, and XCTl1 carried a 37 kb insert, as determined by adding up the cloned DNA inserts from restriction endonuclease digests. There were no detectable plasmids in strain 3401, thus the cosmid clones are assumed to contain chromosomal fragments. Assuming a 3333 kb genome for X. campestris (47), each cosmid probe represents approximately 1% of the total genome. DNA Hybridization
Autoradiographs of Southern blots hybridized against either of the two DNA probes revealed conserved DNA fragments within each pathovar. For example, in figure 6-3 (lanes B K) are shown the RFLP pattern from 10 different strains of X. c. pv. phaseoli from different geographic locations. At least five EcoRI cut DNA fragments which hybridized to XCT1 appeared to be conserved in this pathovar. In
lanes L and M are the RFLP patterns of two strains of X. c. pv. phaseoli var. fuscans. These X. c. pv. phaseoli var. fuscans strains also attack beans (and hence have the pv. phaseoli designation), but are biochemically distinct (6). They produce a dark, melanin-like pigment in nutrient media and in these tests are clearly genetically distinguishable from the other X. c. pv. phaseoli strains. Conserved DNA fragments were also seen for X. c. pv. alfalfae (Fig. 6-3, lanes N and 0) in which six of the up to 15 fragments appeared identical. Likewise, in four different strains of X. c. pv. campestris (Fig. 6-3, lanes P Q, Fig. 6-5, lane D) isolated from three different crucifer
A B C D E F G H I J K LM NO PQ RST
Figure 6-3. Genomic DNA of strains of X. campestris digested with the restriction endonuclease EcoRI and probed with cosmid clone XCT1. Lane A, probe XCTl; lanes B-K, X. c. pv. phaseoli; lane B, JL; lane C, LB2; lane D, Xph25; lane E, EK11; lane F, 82-1; lane G, 82-2; lane H, Xpa; lane I, XP2; lane J, XP2-1; lane K, JF; lanes L-M, X. c. pv.
phaseoli var. fuscans; lane L, SC-3B; lane M, Xpfll; lanes N-O, X. c.
pv. alfalfae; lane N, KS; lane O, FL;lanes P-R, X. c. pv. campestris; lane P, XCl (cabbage); lane Q, 084-1318 (broccoli); lane R, 084-720
(brussels sprouts); lane S, unknown X. c. G65; lane T, probe XCT11.
sources (cabbage, broccoli, and brussels sprouts) at least four of over 10 fragments appeared to be identical. Sizes of hybridizing DNA
fragments are given in Table 6-2. In figure 6-4 is shown the RFLP patterns revealed using clone XCT11. In each case, a different RFLP
pattern is shown, but the same genetic relationship is revealed. Other sets of conserved DNA fragments within a pathovar were seen among seven strains of X. c. pv. vignicola (identical pattern), seven strains of X. c. pv. cyamopsidis (nearly identical), and at least three strains each for X. campestris pvs. citri, glycines, malvacearum, pelargonii, and vesicatoria (not shown).
DNA of strains representing other pathovars of X. campestris were also digested with either EcoRI or BamHI, and hybridized with either XCT1 or XCT11. The pathovars of X. campestris included in these
experiments were X. campestris pvs. alfalfae, begoniae, campestris, carotae, citri, cyamopsidis, dieffenbachiae, esculenti, glycines,
hederae, holcicola, maculifoliigardeniae, malvacearum, mangiferaeindicae, nigromaculans, pelargonii, phaseoli, ,
translucens, vesicatoria, vignicola, vitians, and zinniae. Also
included were some strains isolated as pathogens of an Alocasia sp., Cissus sp., Impatiens sp., Fatsia sp., Jasmine sp., Argemone sp., and a Euonymus species for which no pathovar designations were available.
The results of hybridization of probe XCT1 to DNA of different pathovars of X. campestris digested with EcoRI are shown in Figure 65. The clone XCTI appeared to contain DNA fragments of strain 3401 which were able to demonstrate variability of restriction fragment patterns among different pathovars of X. campestris. Variability in
Table 6-2. Sizes of DNA fragments from. Xanthanonas campestris genomic digests (EcoRI) which hybridized to the XCT1 DNA probe. See Figure 63.
A 12.2, 11.2, 7.9, 5.6, 5.1, 3.2, 2.2, 1.6, 1.0 D 14.1, 13.3, 10.4, 7.6, 6.4, 4.6, 4.3, 4.0, 3.7, 3.3, 3.1
C 14.4, 13.3, 10.4, 7.6, 6.4, 4.6, 4.3, 4.0, 3.7, 3.3, 3.1
D 14.1, 13.1, 10.4, 7.5, 6.4, 5.2, 4.6, 4.3, 4.0, 3.7, 3.1
E 13.9, 11.6, 10.4, 8.9, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1
F 13.7, 13.1, 11.6, 10.4, 8.9, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1,
G 13.7, 13.1, 11.6, 10.4, 8.9, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1,
H 13.1, 10.4, 7.5, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1, 1.9
1 13.1, 10.4, 7.5, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1, 1.9
J 13.1, 10.4, 7.5, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1, 1.9
K 13.1, 11.6, 10.4, 8.9, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1
L 11.9, 8.4, 7.1, 6.7, 4.5
M 17.0, 15.3, 10.0, 8.4, 4.5 N 19.7, 19.1, 15.5, 14.9, 12.9, 12.1, 8.0, 4.6, 3.8, 3.1, 2.7, 1.7
0 19.4, 18.0, 14.7, 13.7, 13.1, 11.7, 9.7, 7.3, 6.2, 5.2, 4.6, 3.7,
P 20.3, 15.3, 10.9, 10.5, 10.2, 9.3, 8.4, 6.9, 5.9, 5.2, 4.4, 3.9,
2.6, 2.2, 1.9
Q 12.9, 10.8, 10.5, 9.1, 8.4, 7.7, 7.2, 6.3, 4.6, 4.4, 3.9, 3.3,
2.4, 2.2, 1.9
R 18.6, 15.8, 12.7, 10.8, 10.2, 9.1, 8.4, 7.7, 6.7, 5.6, 4.4, 3.3,
S 11.3, 8.0, 6.9, 3.4, 2.5, 1.6 T 17.2, 15.3, 14.3, 10.5, 8.6, 7.2, 6.8, 5.8
A B C D E F G H I J K LM NO P 0 RST
Figure 6-4. Genomic DNA of strains of X. campestris digested with the restriction endonuclease EcoRI and probed with cosmid clone XCT11.
Lane A, probe XCT1; lanes B-K, X. c. pv. phaseoli; lane B, JL; lane C, LB-2; lane D, Xph25; lane E, EKII; lane F, 82-1; lane G, 82-2; lane H, Xpa; lane I, XP2; lane J, XP2-1; lane K, JF; lanes L-M, X. c. pv.
phaseoli var. fuscans; lane L, SC-3B; lane M, Xpfll; lanes N-O, X. c.
pv. alfalfae; lane N, KS; lane O, FL;lanes P-R, X. c. pv. campestris; lane P, XCl (cabbage); lane Q, 084-1318 (broccoli); lane R, 084-720
(brussels sprouts); lane S, unknown X. c. G65; lane T, probe XCT11.
AB C DE FG H I J K LM NOP Q R ST
Figure 6-5. Genornic DNA of strains of X. campestris digested with the restriction endonuclease EcoRI and probed with cosmid clone XCT1. Lane
A, probe XCTl; lane B, 7. c. pv. alfalfae FL; lane C, X. c. pv. begoniae 084-155; lane D, X. c. pv. caxnpestris 084-809; lane E, X. c. pv. carotae 13; lane F, X. c. pv. cyamopsidis X002; lane G, X. c. pv. dieffenbachiae 068-1163; lane H, X. c. pv. glycines S-9-3; lane I, X. c. pv. holcicola Xh6G; lane J, T. c. pv. maculifoliigardeniae 0846-166; lane K, X. c. pv. rnalvacearum FL79; lane L, X. c. pv. pelargonii 084-190; lane MX.c. pv. phaseoli var. fuscans S-C-3; lane N, X. c. pv. pisi XP1; lane 05, X. c. pv. translucens X1105; lane P, X. c. p. vesicatoria 75-3; lane Q, X. c. pv. vignicola, SN2; lane R, T. c. pv.* vitians ICPBl64; lane S, X. c. pv. zinniae 084-1944; lane T, probe XCT11.
RFLP pattern appeared as qualitative and quantitative differences among hybridization profiles of various sized DNA fragments in the
strains. For instance, the weak hybridization signal seen from X. c. pv. translucens (fig. 6-5, lane 0) suggests that this strain is distinct from the other pathovars compared, at least over the 1% of the genome contained on the clone XCTI. In other pathovars such as X. c_. pv. alfalfae (fig. 6-5, lane B), X. c. pv. begoniae (lane C), X. c.
pv. cyamopsidis (lane F), X. c. pv. malvacearum (lane K), X. c. pv. vignicola (lane Q), and X. c. pv. vitians (lane R) where some DNA fragments had strong hybridization signals, a close relatedness with portions of DNA in the cosmid clone is suggested. Some of the DNA fragments from different pathovars which hybridized most strongly to
the probe were also of identical size with the EcoRI fragments of the XCT1 clone. As expected, DNA of X. c. pv. citri strain 3401 digested with EcoRI, or BamHI (34,55) contained fragments which corresponded to
DNA fragments of the XCT1 probe (Fig. 6-5, lane A). Sizes of
hybridizing DNA fragments are given in Table 6-3.
From observations using different enzymes and different probes,
the same patterns emerged, indicating that the basic chromosome structure of X. campestris can be used to differentiate strains into specific RFLP types. Although a pathovar may contain more than one type of variant (e.g. X. c. pv. phaseoli), all strains of a given type appear to exhibit the same host selectivity. The RFLP pattern is highly distinctive for each type, and can be used to unambiguously assign an unknown sample to a pathovar, using simple visual
comparisons of pattern against a set of known samples, such as those given in Figure 6-5.
Table 6-3. Sizes of DNA fragnents from Xanthomonas campestris genomic digests (EcoRI) which hybridized to the XCTl DNA probe. See Figure 65.
A 12.2, 11.2, 7.9, 5.6, 5.1, 3.2, 2.2, 1.6, 1.0 B 20.5, 18.8, 15.3, 14.2, 13.5, 12.0, 7.3, 6.2, 5.1, 4.6, 3.7, 3.4,
C 18.7, 12.1, 9.5, 8.6, 8.2, 7.7, 6.9, 6.7, 6.2, 6.0, 4.8, 4.2,
4.0, 3.4, 3.0, 2.9, 2.5, 2.3
D 19.4, 16.4, 14.4, 13.1, 11.1, 10.5, 9.2, 8.5, 7.7, 6.7, 5.6, 4.3,
4.0, 2.2, 1.9
E 17.9, 12.9, 4.1, 3.5, 3.0, 2.3, 1.5 F 16.2, 11.5, 9.5, 8.7, 5.7, 4.0, 3.6, 3.1, 3.0, 1.4 G 14.2, 11.1, 4.5, 3.9, 3.5, 3.0 H 12.2, 10.2, 8.3, 7.91 6.5, 5.5, 4.2, 3.1 1 12.9, 12.0, 7.6, 7.1, 6.5, 4.7, 3.0, 1.7 J 15.6, 14.2, 9.4, 7.9, 6.3, 5.0, 4.6, 4.0, 3.1, 2.5, 2.2
K 20 .9, 17.1, 16.2, 15.5, 14.2, 11.0, 10.4, 9.2, 8.2, 7.5, 6.7,
4.2, 4.0, 3.3, 3.1, 3.0
L 19.4, 12.7, 10.4, 9.0, 7.9, 7.7, 6.8, 5.7, 4.2, 3.9, 3.0, 2.7,
2.3, 1.9, 1.5, 1.4
M 12.0, 8.4, 7.0, 6.8, 4.5 N 15.9, 8.4, 4.9, 3.4, 2.1, 1.9
0 9.2, 4.4
P 13.3, 9.2, 7.2, 6.9, 6.2, 4.6, 3.8, 3.1, 2.1
Q 11.0, 8.5, 7.0, 6.3, 4.2, 3.2 R 16.7, 14.2, 12.9, 9.9, 7.4, 6.9, 6.3, 4.8, 3.8, 3.5, 3.2, 2.6,
2.4, 2.2, 1.6
S 12.9, 12.5, 9.5, 7.7, 6.5, 4.4, 4.0, 3.8
T 17.6, 15.6, 14.7, 10.5, 8.6, 7.2, 1.4
The only widely accepted and most practical method for differentiating pathovars of X. campestris is to inoculate a plant suspected as the host for that pathovar. This practice can sometimes be tedious, time consuming, subjective and subject to a surprising number of artifactual influences. It is not known whether host selectivity is unstable as suggested by Dye (23); stable as suggested
by Schnathorst (84) or even taxonomically significant. Although the classification is thought to be useful, it can be highly misleading since emphasis is placed upon one characteristic pathogenicity. If
only one or few gene differences were involved in host selection, then differentiation by pathovar could be highly misleading, at least in the sense that a given group of strains might be capable of attacking more than one host in some cases. Alternatively, two or more strains of relatively unrelated groups could be cataloged together because they happen to attack the same host. This latter situation appears to be the case with X. c. pv. phaseoli and X. c. pv. phaseoli var. fuscans (see Fig. 6-3), and with Florida strains of X. c. pv. citri
To address these problems, we evaluated the potential of using DNA sequence variation of X. campestris for strain classification purposes. DNA sequence variation was first examined by digesting DNA
with restriction enzymes and visualizing directly the resulting fragments on ethidium bromide stained gels. These stained gels were useful for side-by-side comparisons of restriction fragments for samples run on the same gel, but comparisons between different gels
were more difficult. Subsequently, visualizing and comparing
variation in bacterial genomes with cosmid clones carrying 30-38 kb cloned X. campestris genome fragments was simplified. In both cases variation was revealed by alterations in the sizes of visualized DNA restriction fragments (RFLPs). The use of DN4A probes derived from chromosomal DNA fragments also alleviated some of the difficulty in comparing stained gels with the sometimes present polymorphism of the higher copy number plasmid DNA fragments which can occur in the background.
Some of the selected clones tested as DNA probes appeared to be useful for identifying DNA sequences conserved within a given pathovar, while others appeared to identify DNA sequences which were highly conserved at the species level. The DNA which was considered conserved at the species level was represented as banding patterns which were nearly identical among all strains over the several pathovars tested. Examples of DNA sequences which are known to be highly conserved are rP1NA encoding genes (38,83). If smaller DNA probes containing known genetic loci such as those associated with genes for pectate lyase and protease (18), or avirulence activity (33,88) were tested, these smaller, more defined DNA probes could be used to determine if such genetic loci were highly conserved or variable. With larger, randomly selected probes, the likelihood was increased for detecting strain- or pathovar-specific variation. The large (greater than 30 kb) size of the probes used in this study allowed detection of both highly conserved and variable regions.