|Table of Contents|
Table of Contents
List of Tables
List of Figures
Chapter 1. Introduction
Chapter 2. Fusarium oxypsorum f. sp. cubense consists of a small number of divergent and globally distributed clonal lineages
Chapter 3. Genetic variation in two Honduran field populations of fusarium oxysporum f. sp. cubense
Chapter 4. Summary and conclusions
Appendix A. Exploring the potential of using polymerase chain reaction methodologies in phylogenetic studies of fusarium oxysporum f. sp. cubense
Appendix B. Additional information from phylogenetic study
Appendix C. Allelic data scored as presence or absence for each probe enzyme combination; isolate name and vegetative compatibility group included as a representative for each unique restriction fragment length polymorphism haplotype
Appendix D. List of isolates of f. oxysporum f. sp. cubense, their vegetative compatibility group, haplotype and lineage
PHYLOGENETIC AND POPULATION GENETIC ANALYSES OF FUSARIUM OXYSPORUM F. SP. CUBENSE,
THE CAUSAL AGENT OF FUSARIUM WILT ON BANANA
ROSALIE LYNN KOENIG
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 1997
Rosalie Lynn Koenig
I dedicate this research to my father and mother who always stressed the
importance of education, and to my husband Tom and daughter Amaleah, who put up with me during this stressful period of my life.
I would like to thank Dr. Corby Kistler for his support, patience, understanding and encouragement through all stages of my graduate studies.
I would like to thank Dr. Dave Mitchell for his guidance, counsel, encouragement and personal commitment to excellence and graduate students. Many thanks to my committee members Drs. Randy Ploetz and Eduardo Vallejos for their input and cooperation. Additional thanks to Ulla Benny, Gerald Benny and Lyndel Meinhardt who always lended their technical expertise and support.
Special thanks to Dr. Brian Bowen for his expertise in genetic analysis; he always made time for me and provided clear explanations of complex subject matters. I thank the NSEP for funding the field work in Honduras. I thank Drs. Maricio Rivera and Phil Rowe for their support in Honduras. Thanks to my many friends and classmates who shared their dreams and aspirations.
TABLE OF CONTENTS
LIST OF TABLES............... ....................... .......... vii
LIST OF FIGURES................................................ viii
ABSTRACT ........................... ........................... ix
1 INTRODUCTION ...................................1
The Host..................... .............................. 4
Genetic Characterization of Fusarium oxysporum f sp. cubense.............8
Clonality and its Consequences in the- Study of Population Genetics of
Asexually Reproducing Organisms ............................... 11
2 FUSARI9M OXYPSORUM F. SP. CUBENSE CONSISTS OF A SMALL NUMBER OF DIVERGENT AND GLOBALLY
DISTRIBUTED CLONAL LINEAGES ............... 15
Materials And Methods ......................................... 17
Results .................................................... 30
3 GENETIC VARIATION IN TWO HONDURAN FIELD POPULATIONS OF FUSARJUA'IOKYSPORUM F. SP. CUBENSE ...... 47
Introduction ................................................ 47
Materials and Methods ............................ ............. 50
Results. .................................................... 60
Discussion ................................................. 65
4 SUMMARY AND CONCLUSIONS ........................ .......73
A EXPLORING THE POTENTIAL OF USING POLYMERASE CHAIN REACTION METHODOLOGIES IN PHYLOGENETIC STUDIES
OF FUSARIUM OXYSPORUMF. SP. CUBENSE .................... 81
In tro d u ctio n . . . . . . . . . . . . . . . . . . . . . . . . . 8 1
R esults and D iscussion ........................................... 82
B ADDITIONAL INFORMATION FROM PHYLOGENETIC STUDY ....... 91
C ALLELIC DATA SCORED AS PRESENCE OR ABSENCE FOR EACH PROBE ENZYME COMBINATION; ISOLATE NAME AND
VEGETATIVE COMPATIBILITY GROUP INCLUDED AS A
REPRESENTATIVE FOR EACH UNIQUE RESTRICTION
FRAGMENT LENGTH POLYMORPHISM HAPLOTYPE ........... 94
D LIST OF ISOLATES OF F. OXYSPORUMF. SP. CUBENSE, THEIR VEGETATIVE COMPATIBILITY GROUP, HAPLOTYPE AND
L IN E A G E ................................................. 100
REFERENCES .................................................. 105
BIOGRAPHICAL SKETCH ................................. ......... 113
LIST OF TABLES
2-1: List of isolates, their vegetative compatibility groups (VCG), clonal lineages,
the cultivar, and geographical regions from which they were collected ....... 18 2-2: Similarity matrix of simple matching coefficients based on restriction fragment
length polymorphisms for selected isolates of Fusarium oxysporum f. sp.
cubense, F. oxysporum f sp. lycopersici and F. oxysporum f. sp. niveun .... 35 2-3: Clonal lineages of Fusarium oxysporum f sp. cubense isolates, their
geographical distributions and corresponding vegetative compatibility
groups (V C G ) ................................................. 36
2-4: A comparison of allelic data for isolates in FOC I, FOC II and FOC VII ........ 3,9 2-5: Clone-corrected measurements of gametic disequilibrium among pairs of
alleles in a world-wide collection of Fusarium oxysporum f. sp. cubense ...... 41 3-1. Genotypes of banana planted in Field I ................................. 51
3-3. Field design of banana plants in rows of the race 2 disease screening plot ....... 55 3-4. Isolates of Fusarium oxysporum f. sp. cubense collected in Field 1 and their
incidence on different host genotypes ................................ 63
LIST OF FIGURES
2-1: DNA of isolates was digested with Eco RV restriction enzyme and probed
w ith clo ne 187 . .. .. . .. .. . .. . .. . . . . . . . . . . . . 26
2-2: Frequency distribution of RFLP haplotypes among the 165 isolates
representing a world collection of Fusarium oxysporum f. sp. cubense ....... 31 2-3: Midpoint rooted 50% majority rule consensus tree representing 500
bootstrap replicates ............................. .... ......... 32
2-4: Mitochondrial DNA haplotypes of selected Fusarium oxysporum f sp.
cubense isolates ................................................ 40
3-1: mtDNA haplotype of selected Honduran field isolates ..................... 61
3-2: HindIII digested DNA of selected Honduran field isolates that were probed
with clone pEY10 to generate DNA fingerprints ...................... 64
3-3: Incidence and distribution of mtDNA haplotypes of Fusarium oxysporum f.
sp. cubense on host genotype ...................................... 66
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
PHYLOGENETIC AND POPULATION GENETIC ANALYSES
OF FUSARI1UM OXYSPORUM F. SP. CUBENSE,
THE CAUSAL AGENT OF FUSARIUM WILT ON BANANA By
Rosalie Lynn Koenig
Chairperson: Dr. H. C. Kistler
Major Department: Department of Plant Pathology
A worldwide collection of Fusarium oxysporum f. sp. cubense was analyzed using anonymous, single-copy, restriction fragment length polymorphism (RFLP) loci. Of the 165 isolates examined, 72 RFLP haplotypes were identified. Consistent with strict clonal reproduction, individuals with identical haplotypes were geographically dispersed, and clone-corrected tests of gametic disequilibrium indicated significant nonrandom association among pairs of alleles. Parsimony analysis divided haplotypes into two major branches. Ten clonal lineages were identified based on coefficients of similarity between 0.94-1.00. The two largest lineages had pantropical distribution. Isolates comprising one lineage (FOC VII) may represent either an ancient genetic exchange between individuals in the two largest lineages or an ancestral group. The two largest lineages (FOC I and II) and a lineage from East Africa (FOC V) are genetically distinct; each may have acquired the ability to be pathogenic on banana independently.
Populations of the pathogen were sampled from two disease screening fields in
Honduras. Field 1 was planted to 10 genotypes of banana and 97 isolates were recovered from host tissue. Isolates recovered in Field I comprised three mtDNA haplotypes, and their incidences and distributions varied with the host genotype sampled. DNA fingerprints indicated that the isolates represented five clones. Field 2 was planted to a single genotype of banana. The 59 isolates collected from host tissue represented a single mtDNA haplotype and clone. The frequencies with which the pathogen clones infected the host were dependent upon host genotype. Two races of the pathogen (race 1 and 2) have been reported in Honduras. Since the cultivar Highgate is susceptible to race I but resistant to race 2, isolates recovered from the cultivar Highgate were assumed to represent race 1. Isolates within a lineage associated with race 1 isolates were preferentially recovered from Highgate and the Fundaci6n Hondurefia de Investigacio'n Agricola (FHIA) hybrids. Isolates representing a second lineage were recovered at a low frequency from Highgate, suggesting that these two lineages differ in their ability to infect bananas susceptible to race 1. Assessing the pathogen populations and determining the lineages of isolates infecting particular banana breeding lines could be useful for banana breeding programs.
The genus Fusarium (Link) is comprised of fungi which have evolved from an ancient ancestor whose geological age has been postulated to date back at least 200 million years (Snyder, 1981). Whether or not this is an accurate estimate of the age of this important fungal group, the genetic diversity present in members of the genus Fusarium is the result of evolutionary processes which have developed over a long time span.
Fusarium oxysporum (Schlecht.) Synd. & Hans. is an ubiquitous, soil borne fungus that is comprised of strains that are ecologically adapted to diverse environments and which have the ability to utilize an array of substrates in the soil (Burgress, 1981). The reproductive success of strains within this species depends on the ability to produce a large number of propagules, in the form of microconidia or macroconidia, as well as resistant resting spores, referred to as chlamydospores, that can ensure long-term survival during adverse environmental conditions. The morphological characteristics of these propagules also have been used by Wollenweber and Reinking (1935) to include the species in section Elegans. No teleomorph have been described for the species.
Plant pathogenic isolates ofF oxysporum cause diseases known as vascular wilts and constitute a small portion of the strains within this species. However, because of their importance to agriculture, Snyder and Hansen (1940) developed the concept offormae speciales to recognize plant pathogenic strains ofF. oxysporum that were morphologically
indistinguishable from saprophytic strains of the species but had the ability to cause disease on specific hosts. Over 70 formae speciales have been described (Kistler, 1997). Although formae speciales are not recognized by the International Code of Botanical Nomenclature, they are useful to plant pathologists because they provide informal groupings of morphologically indistinguishable isolates. Isolates within a forma specialis may be pathogenic on a specific host, such as F. oxysporum f sp. pisi, which causes a wilt disease on peas, Pisum sativum L., or may have a broader host range, such as 1. oxysporum f. sp. vasinfectum (Atk.) Snyd. & Hans., which is pathogenic on plants in the families Malvaceae, Solanaceae and Fabaceae (Windels, 1991). Two assumptions that have been made about isolates within a forma specialis are that 1) they are genetically more similar to each other than to isolates of different formae speciales, and 2) that formae speciales are monophyletic groupings (Kistler, 1997).
Fusarium wilt of banana, commonly referred to as Panama disease, is caused by F. oxysporuni Schlechtend.: Fr. f sp. cubense (E.F. Sm.) W.C. Synder & H. N. Hans.
This fungus systemically colonizes the xylem tissue of susceptible banana cultivars, disrupting normal water uptake (Beckman, 1987). This disruption eventually results in wilt and the subsequent death of the plant. Studies described by Beckman (1990) indicate that conidia in aqueous solutions are readily taken up into the vascular elements and carried in the transpiration stream to sites of entrapment at vessel endings. Subsequent secondary colonization occurs in susceptible cultivars by the dispersal of conidia through the vascular system. Resistant cultivars prevent spread by the formation of occluding gels and tyloses that wall off the lumena of the infected vessels.
To date, three races of the pathogen, 1, 2 and 4, have been described which are pathogenic on banana; these races are differentiated by the banana cultivars Gros Michel, Bluggoe, and Cavendish, respectively. Historically, this pathogen has caused large scale disease epidemics when susceptible banana cultivars have been grown in the presence of the pathogen (Ploetz, 1994).
The first recording of Fusariumn wilt of banana was made in southeast Queensland in 1874, on the cultivar Silk, known locally as Sugar (Bancroft, 1976). The history of this disease generally paralleled the dissemination of susceptible genotypes to new areas. The edible bananas are asexually propagated, typically using rhizomes or sucker plants. Occasionally, infested rhizomes were brought to new plantations and this is likely how the pathogen was widely distributed. Interestingly, one region where the pathogen has been found only recently is the vast area of Oceania where susceptible cultivars have been grown for millennia (Shivas and Philemon, 1996).
Prior to the late 1950s, monoculture of the race 1 -susceptible cultivar Gros Michel resulted in severe and wide-spread epidemics in the export banana trade. The disease became economically less important in export production once the trade converted to race 1-resistant Cavendish cultivars. However, approximately 20-40 years after Cavendish cultivars were planted widely in a number of subtropical regions, they too began to succumb to the disease. Outbreaks on Cavendish cultivars were reported in the Canary Islands in 1926 (Ashby, 1926), the 1940s in South Africa, the 1950s in Australia, and the 1960s in Taiwan (Ploetz et al., 1990). In the late 1970s outbreaks on Cavendish were observed in the Mindinao province of the Philippines. However, in contrast to outbreaks in the subtropics, disease outbreaks in the Philippines occurred in localized patches that
tended not to spread (Ploetz et al., 1990). All isolates in VCG 0122 are from the Philippines and they all have been recovered from Cavendish cultivars.
In the late 1970s Su et al. (1977) conclusively demonstrated that isolates capable of causing disease on Cavendish were different from race I and 2 isolates. These isolates, which also cause disease on race 1- and race 2- susceptible cultivars, were designated race 4.
In addition to economic losses in the export trade, Panama disease is a serious
problem on locally consumed dessert bananas and cooking bananas (Ploetz, 1992; Ploetz et al., 1992; 1994; 1995). Only 14% of the bananas that are produced in a given year are exported. The remaining 86% (65 million metric tons in 1992) are consumed locally (FAO, 1993) and are often staple foods for poor, subsistence farmers in the tropics (INIBAP, 1993). In total, the impact of Fusarium wilt on these bananas is far greater than on bananas produced for export.
The F. oxysporum f. sp. cubense: banana pathosystem likely developed from a process of co-evolution between the host and pathogen, resulting in complementary evolutionary changes in both organisms. Bananas belong to the order Zingiberales, which consists of eight families of tropical and subtropical rhizomatous perennial plants that inhabit moist forested areas or invade disturbed sites (Kress, 1990). Bananas belong to the family Musaceae, which contain three genera (Musa, Musella, and Ensete) whose natural range extends from West Africa to the Pacific. Edible bananas are the most important cultivated crop of this family. Two additional minor crops of economic
importance are abacd (M. textilis), a fiber crop, and cultivated forms of Ensete ventricosum, a food and fiber plant of upland Abyssinia.
Edible bananas have been domesticated from two species, M acuminata and M balbisiana. Simmonds and Shepherd (1955) used 15 phenotype characters and scored each on the scale of I (accuminata expression) to 5 (balbisiana expression) to develop a system to classify edible bananas. From this, the genetic constitution of any edible banana can be specified using a genome formula indicating the contributions from M accurninata
(AA) and M balbisiana (BB).
Edible M. acuminata cultivars are either diploid (AA) or triploid (AAA), and their center of origin includes Southern Thailand, Sumatra, Java, Borneo, Burma, and Malaysia, which is also the primary center of diversity for the species. This species is most diverse in peninsular Malaysia and the equatorial islands where the wild species is native. Six subspecies of M. acuninata are recognized. Although the banana clone Gros Michel has phenotypic characters that suggest descent from M acunzinata subsp. malaccensis (Stover, 1990), the parents of most clones are not known.
Simmonds (1962) argues that edibility in bananas arose in Malaysia through parthenocarpy in M. acurninata diploids followed by triploidy (AAA). Through the process of domestication, man has selected for both parthenocarpy and sterility, which are genetically independent characteristics. Additionally, domestication has led to selection of triploidy because it contributes to high plant vigor and fruit growth rate, and confers gametic sterility (Simmonds, 1962).
In contrast, M balbisiana has a center of origin in a region including the countries of India, Burma, and Northern Thailand. It is less genetically diverse than M acuminata.
It is presumed that edibility did not arise in M balbisiana because no naturally occurring parthenocarpic clones have been identified. Similarly, no purely M. balbisiana triploids or tetraploids are known. M balbisiana is naturally cross fertile with a number of Musa species, including M. acuninata, but it is reproductively quite isolated (Simmonds, 1962). The hybrid groups (AB, AAB, ABB) are peripheral, being diversified in India and, to a lesser extent, in Indochina and eastern Malaysian areas from which wild M acuninata is locally absent but in which M balbisiana is native (Simmonds, 1962). The transport by humans of AA-type bananas into areas occupied by wild M. balbisiana is inferred to have provided the opportunity for interspecific hybridization (Simmonds, 1962). Characteristics such as plant hardiness, drought resistance, disease resistance to nematodes and yellow Sigatoka, and starchiness of fruit that are found in interspecific hybrid bananas are believed to be contributed from M balbisiana (Simmonds, 1962). Interspecific hybrids between M balbisiana and M. acuminata occur in nature because the species are sympatric over a large area of southeast Asia and introgression occurs often.
A secondary center of diversity for bananas is East Africa. In East Africa,
mixtures of genotypes are grown for local use and particular dietary needs. The East African baking and beer type bananas (AAA) are unique to the region and are often found in mixed plantings with cultivars of exotic origin, such as ABBs and various dessert types (Ploetz et al., 1990). Inland cultivation of banana in East Africa is based primarily on triploid clones (AAA) but the coastal areas have much more diverse populations. Coastal East Africa is the only location outside of Southeast Asia where there is an assemblage of edible diploids cultivars and extensive local diversification. The most striking
accumulations of unique clones are in upland East Africa and the Pacific Islands, where it appears that subsistence farmers actively preserved these clones (Simmonds, 1962).
Murdock's (1959) discussion of plant introduction in relation to African history suggests that Malaysian food plants entered Africa in the hands of Indonesian migrants about 2000 B.P. Bananas, as well as taro, yams, breadfruit, coconut, some rices and sugarcane, are adapted to moist climates and became staple foods of the wetter parts of Central Africa (Simmonds, 1962). Madagascar appears to be the principal point of entry of bananas into Africa. Simmonds' (1962) view is that the early African cultivars reached Madagascar from Malaysia late in the first half of the first millennium AD. From West Africa, Portuguese travelers took plants to the Canary Islands early in the fifteenth century, and a least one clone traveled to Hispaniola in 1516. Simmonds (1962) believed that this represented the first of many introductions made into the New World. However, Langdon (1993) suggests, based on linguistic, botanical, and archeological data, that banana was present in the New World before the Spaniards arrived.
Presently, the only location with a high level of genetic diversity of M acurninata is Papua New Guinea and surrounding islands. Here, the indigenous people have preserved an intact population of primitive cultigens, a relic of an early stage in the evolution of the group. Jarret (1992) used 98 phylogenetically informative restriction fragment length polymorphisms (RFLPs) to characterize eight species representing three sections of Musa from Papua New Guinea, as well as one species of Ensete. The study found only a distant relationship between M balbisiana and other species of Musa; M. balbisiana was not in the same clade as the other species within the section Musa (also
referred to as Eumusa). The diploid land races from this section clustered with M. acunfitiata. All accessions of M acurninata shared a large number of alleles.
Genetic Characterization of Fusarium oxvsvorurn f. sp. cubenise
Cladistic analyses of the phylogenetic relationships among isolates of a fungal
species are based on the concept that members of a species have evolved from a common ancestor (Scotland, 1992). This ancestor possessed primitive genetic characteristics which may have changed through the evolutionary history of the organism. Derived character changes, those changes that are different from the ancestral type, are the result of evolutionary processes which consist of both genetic (mutation, recombination, and transposon rearrangements) and spatial (migration, divergence, and convergence) forces.
Evolutionary studies of organisms aim to determine phylogeny by examining the genetic relatedness among members of a group. Studies may involve closely related organisms, such as fungal isolates of a given species, or organisms more distantly related, such as all fungal isolates within a location. However, many evolutionary studies are conducted at the genus or species level. These types of studies rely on identifying genetic markers that discern among individuals sampled in the study. Prior to the development of molecular techniques, researchers primarily relied on phenotypic traits to develop evolutionary theories. However, the development of molecular techniques, such as isozyme analysis, restriction fragment length polymorphism analysis, random amplified DNA analysis and gene sequencing, have made it possible to objectively assess traits at the genotypic level.
Isolates ofF oxysporuni have been characterized using morphological,
biochemical and genetic markers (Kistler, 1997). Additionally, vegetative compatibility
groups (VCGs) based on work on Fusarium moniliforme also have been used to differentiate similar isolates of oxysporumn (Puhalla and Spieth, 1985). Isolates ofF. oxysporumn which share identical alleles at the loci governing heterokaryon incompatibility, commonly referred to as het or vic, are vegetatively compatible. Conventionally, this is determined via the ability of nitrate-nonutilizing (nit) auxotrophic mutants to complement one another for nitrate utilization (Leslie, 1993). Since F. oxysporum has no known sexual stage, it is not possible to determine genetically the number of loci involved in vegetative incompatibility. However, genetic studies performed on F. mnoniliforme (teleomorph- Gibberellaffuijikuroi) indicate that a minimum of 10 vic loci govern vegetative compatibility in this closely related species (Puhalla and Spieth, 1985). Unfortunately, the relationship among isolates comprising different VCGs is impossible to discern based solely on this technique because a single change of alleles at any of the vic loci will generate a new VCG.
Vegetative compatibility group diversity varies among the known formae speciales of F. oxysporumn. For example, among 115 isolates ofF. oxysporum f sp. lycopersici, Elias and Schneider (1991) identified one major and two minor VCGs, as well as a number of single-member VCGs. In contrast, a worldwide collection of 52 isolates ofF. oxysporum f sp. vasinfectum isolates comprised 10 VCGs and isolates belonging to distinct races of the pathogen were never in the same VCG (Fernandez et al., 1994). Currently, at least 17 VCGs have been described for F. oxysporum f sp. cubense (Ploetz, 1994). The majority of VCGs appear to have limited geographical distribution. However, the majority of isolates belong to a few major VCGs that are geographically widespread and contain more than one race of the pathogen (Ploetz, 1994).
Until the advent of tissue culture, the rhizome of the banana plant was used to
propagate banana. It is likely that the worldwide movement of planting stock (rhizomes and associated soil particles) from region to region is responsible for much of the geographical distribution of the pathogen outside its region of origin. In terms of evolution, the diversity of the pathogen in areas outside of its center of origin will depend on the number of years the pathogen has been present in the new location and whether single or multiple clonal lineages were introduced. For example, one might expect the populations ofF. oxysporum f. sp. cubense in East Africa to be more genetically variable than populations in the USA because the crop was introduced into East Africa thousands of years prior to the introductions to Florida.
Many earlier genetic characterizations of various formae speciales ofF. oxysporum focused on a small number of isolates with limited geographical distribution. These studies gave an incomplete picture of genetic diversity when more extensive testing using a larger number of isolates with a large geographical distribution was performed. For example, when 170 isolates of the pathogen F. oxysporum f. sp. dianthi were collected from 42 cultivars at 40 different sites in Israel, all of the isolates belonged to the same VCG (Katan et al., 1989). However, when isolates of the same pathogen were sampled from a number of geographical locations, including Italy, Spain, The Netherlands, the United States and Germany, eight VCGs were identified (Aloi and Baayen, 1993). The latter researchers found not only single races within single VCGs but also multiple races within single VCGs (Aloi and Baayen, 1993).
Isolates ofF. oxysporun f sp. cubense also have been grouped based on
electrophoretic karyotype, random amplified polymorphic DNA (RAPD) analysis and the
production of volatiles in culture. Boehm et al. (1994) proposed two groupings of isolates ofF oxysporuwn f sp. cubense based on similarities in chromosome number and genome size. In general, these groupings agreed with RAPD data (Bentley et al., 1995, Sorensen et al., 1993) and groups determined when the formation of aldehydes by different isolates were examined (Brandes, 1919; Moore et al., 1991, 1993; and Stover, 1959).
Clonality and Its Consequences in the Study of Population Genetics of Asexually Reproducing Organisms
Fusarium oxysporum is considered to be an asexual fungus based on the fact that, although many researchers have searched for a sexual stage in both the laboratory and field, none has been reported. For an asexually reproducing organism like F oxysporunl, it is generally assumed that isolates within a VCG are genetically similar and represent clonal populations. Many asexually reproducing populations consist of a number of clonal lineages with wide geographical distribution. Kohli et al. (1992), using a DNA fingerprinting probe and mycelial compatibility, detected 39 clones of Sclerotinia sclerotioruni on canola from three provinces in Canada. The most widely distributed clone was found in all three provinces. Overall, they found that clones of S. sclerotiorum were distributed over large geographic areas and field populations were composed of more than one clone.
Patterns of genetic variability in natural populations of fungi may reflect strict clonality, panmixia or a mixture of clonality and mating (Anderson and Kohn, 1995). In asexually reproducing organisms, there is presumably no contribution to genotypic variability through meiotic recombination or introgression. In general, genotypic variability in asexual fungi may be generated through mutation, random genetic drift,
parasexuality, somatic recombination and transposable elements. Although mutation rates are generally low per locus, mutations are persistent and over time they generate much variability, summed over all loci (McDonald, et al., 1989). Tibrayrenc et al. (1990) established criteria that characterize asexually reproducing organisms. They include over-represented, ubiquitous multilocus genotypes in a population; the absence of recombinant genotypes; correlation between two independent sets of genetic markers; and the presence of linkage disequilibrium. Genetic analyses of a large collection of isolates using neutral, single-copy, polymorphic loci ideally can test for the absence of recombination.
Recently there has been an increased interest in genetically characterizing field populations of plant pathogenic fungi. Many field studies have been conducted which have characterized the genetic diversity of fungal populations within a defined geographical location. The concept of a fungal population has been described by McDonald and McDermott (1993) as a group of individuals sharing a common gene pool and present in a limited geographical area. Population genetic analysis involves the identification of genetic markers that are unambiguous and informative [polymorphic] (McDonald and McDermott, 1993). These types of markers have been used to characterize populations of a number of fungal species, including Cryphonectria parasitica, Fusarium oxysporum, Sclerotinia sclerotiorum, Phytophthora infestans, Mycosphaerella graminicola, Cochhiobolus carbonum, Phytophthora megaspernia f sp. glycinea, Magnaporthe grisea and Stagonospora nodorum (Appel and Gordon, 1994; Goodwin et al., 1992; Koh et al., 1994; Kohli et al., 1992; Kohli et al.,1995; Leonard and Leath, 1990; Levy et al., 1991; Levy et al., 1993; McDonald and Martinez, 1990;
McDonald et al., 1994; McDonald et al., 1995; Sujkowski, et al., 1994; Whisson et al., 1992).
Population genetic studies of Fusarium oxysporuni f sp. melonis have addressed genetic structure in both agricultural and native soils and also have examined the genetic relationships among pathogenic and nonpathogenic isolates by characterizing their VCGs and mitochondrial haplotypes (Appel and Gordon, 1994, Gordon and Okamoto, 1992; Gordon et al., 1992,). Based on parsimony analysis of the mtDNA data, nonpathogenic isolates from Maryland were more closely related to F. oxysporum f sp. melonis than to the California nonpathogens (Appel and Gordon, 1994). Based on mtDNA, both pathogen VCGs were more closely related to a nonpathogen VCG than to each other (Appel and Gordon, 1994). Essentially all diversity in mtDNA was found among the nonpathogenic strains of the species, whereas pathogenic strains were represented by only a single mtDNA haplotype at any one location (Gordon and Okamoto, 1992; Jacobson and Gordon, 1990b) indicating that pathogenic strains are less genetically diverse than non pathogenic strains. In contrast, a study of the genetic relatedness among 120 strains ofF. oxysporum in native and cultivated soils consisted of 23 different mtDNA haplotypes, of which the five most common haplotypes accounted for 78% of the isolates. Isolates representing these haplotypes were found in both native and agricultural soils (Gordon et al., 1992). Seventy two percent of the isolates found in cultivated soil were associated with the same mtDNA haplotype as one or more of the isolates in the native soil (Gordon et al., 1992).
Host populations can act as powerful selective forces on pathogen populations and vice versa (McDonald et al., 1989). For example, the rice blast fungus, Magnaporthe
grisea, can be divided into a large number of pathotypes. Pathotypes are analogous to races in that specific pathotypes of a pathogen are virulent on certain genotypes of the host. Using a repetitive fingerprinting probe that serves as a genealogical index among rice blast isolates, Levy et al. (1993) examined the population structure and associated virulence properties among isolates in a resistance-breeding nursery in Colombia. Sampling the nursery population enabled the recovery of 39 pathotypes among 151 isolates. Using the fingerprinting probe called MGR 586, they could distinguish six distinct clonal lineages, each with a generally non-overlapping cultivar range and an associated set of several related pathotypes. Eighty-five percent of the pathotypes were lineage specific. Their results support the view that cultivar specificity and pathotype evolution have developed within constraints imposed by the genetic background of each lineage (Levy et al., 1993).
The overall objectives of this dissertation research were to determine the
phylogenetic relationships among isolates ofF. oxysporum f. sp. cubense and to determine the level of genetic variability and its distribution in field populations of the pathogen on banana. An RFLP study of a world collection of isolates was undertaken to assess the phylogenetic relationships among isolates. A study which characterized isolates by their mitochondrial haplotype and distinguished between different clones by utilizing a DNA fingerprinting probe was undertaken to assess the level and distribution of genetic variability in two Honduran field populations. Both of these studies should contribute to the overall understanding of the genetic structure and relationships among isolates ofF. oxysporum f, sp. cubense.
FUSARIUMOXYPSORUMF. SP. CUBENSE CONSISTS OF A SMALL NUMBER OF
DIVERGENT AND GLOBALLY DISTRIBUTED CLONAL LINEAGES Introduction
Fusarium wilt of banana, commonly referred to as Panama disease, is caused by Fusarium oxysporumn Schlechtend.: Fr. f. sp. cubense (E.F. Sm.) W.C. Synder & H. N.
Hans. Isolates ofF. oxysporum f. sp. cubense previously have been characterized by morphology or biochemical and genetic markers. On morphological bases, F. oxysporumn was included in section Elegans by Wollenweber and Reinking (1935). No teleomorph has been described for the species.
Vegetative compatibility groups (VCGs) have been used to categorize isolates of F. oxysporum, including Fusarium oxysporum f. sp. cubense. Based on data from other fungi, isolates that share identical alleles at the loci governing heterokaryon incompatibility, commonly referred to as het or vie, are vegetatively compatible (Leslie, 1993). Conventionally, this is determined by the ability of nitrate-nonutilizing (nit) auxotrophic mutants to complement one another for nitrate utilization (Correll et al., 1987). Currently, at least 17 VCGs have been described for this forma specialis (Ploetz and Correll, 1988; Ploetz, 1994; Ploetz et al., 1997). The majority of isolates belong to two major VCGs that have pantropical distribution and contain more than one race of the pathogen. Minor VCGs were found to have a more limited geographical distribution (Ploetz and Correll, 1988; Ploetz, 1990). For an asexually reproducing organism like F.
oxysporurn, it is generally assumed that isolates within a VCG are genetically similar and represent clonal populations (Appel and Gordon, 1994; Jacobson and Gordon, 1991; Tantaoui et al., 1996).
In addition to characterizing isolates by VCG, isolates ofF. oxysporum f. sp. cubense have been grouped based on electrophoretic karyotype, randomly amplified polymorphic DNA (RAPD) analysis and the ability to produce volatile organic compounds in culture. Boehm et al. (1994) proposed two groupings of isolates ofF. oxysporum f sp. cubense based on similarities in chromosome number and genome size. Group I comprised isolates in VCGs 0124, 0125, 0124/0125, 01210 and 01214. Group II comprised isolates in VCGs 0120, 0121, 0122, 0123, 0129 and 01213. In general, these groupings agreed with RAPD data (Bentley et al.,1995; Sorensen et al., 1993). When the presence or absence of RAPD bands were treated as binary data and subjected to phenetic analysis based on the unweighted pair group method with arithmetic mean (UPGMA), isolates in VCGs 0120, 0121, 0122, 0126, 01210, 01211 and 01212 formed one group and isolates in VCGs 0123, 0124, 0124/0125 and 0125 formed a second group.
Similar major groups of isolates were evident when differentiation was based on the formation of aldehydes in culture. Brandes (1919) noted that certain isolates of the pathogen produced these odorous compounds when grown on steamed rice, whereas others did not; the latter isolates were classified as variety inodoratum. Stover (1959) examined a larger set of isolates from tropical America and the Caribbean and noted that production of odorous compounds was a consistent and repeatable trait. Those isolates which produced the aldehydes were referred to as cultivar Odoratum whereas, those which did not were referred to as cultivar Inodoratum. More recently, Moore et al. (1991;
1993) analyzed the production of these compounds with high pressure liquid chromatography. Isolates in VCGs 0120, 0129 and 01211 produced characteristic volatile profiles, whereas isolates in VCGs 0123, 0124, 0124/0125, 0125 and 0128 did not.
Restriction fragment length polymorphisms (RFLPs) also have been employed to determine the genetic relationships among isolates ofF. oxysporum. These markers are ideally suited to genetic diversity studies because of the following characteristics: i) most are selectively neutral; ii) polymorphisms tend to be more numerous compared to other types of markers, such as isozymes; iii) they are reproducible; and iv) those identifying random single-copy loci avoid problems that are associated with linkage of markers.
In this study, probes from F. oxysporum f. sp. lycopersici that correspond to
single-copy, anonymous loci (Elias et al., 1993) were used to identify polymorphic alleles in F. oxysporum f sp. cubense. The objectives of the study were to determine whether isolates within VCGs ofF. oxysporum f. sp. cubense were clonally derived and if clonal lineages correlated with previously determined VCGs. Phylogenetic relationships among the various VCGs ofF. oxysporum f sp. cubense and F. oxysporum from other hosts were also characterized to determine if host specificity is monophyletic.
Materials And Methods
One hundred and sixty-five isolates representing worldwide distribution and 17
VCGs ofF. oxysporum f. sp. cubense from the collection located at the Tropical Research and Education Center in Homestead, Florida, were selected for analysis (Table 2-1). Additionally, three isolates ofF. oxysporum f. sp. lycopersici and a single isolate of
Table 2-1: List of isolates, their vegetative compatibility groups (VCG), clonal lineages, the cultivar, and geographical regions from which they were collected
VCG' Isolate CultivarV Origin and collectors Lineage
0120 IC2 Cavendish Icod de los Vinos, Canary Islands,d II
22425 Cavendish Wamuran, Queensland, Australia,g II
ORT2 Cavendish La Orotava, Canary Islands,d II
0-1220 Mons Queensland, Australia,c II
GAL2 Cavendish Las Galletas, Canary Islands,d II
C2 Cavendish Canary Islands,f II
ADJ2 Cavendish Adeje, Canary Islands,d II
Cl Cavendish Canary Islands, f II
22424 Lady Finger Moorina, Queensland, Australia,g II
0-1222 Mons Queensland, Australia,c II
0-1219 Mons Queensland, Australia,c II
A2 Mons Mari Australia,f II
ADJ1 Cavendish Adeje, Canary Islands,d II
STGM1 Gros Michel Costa Rica,i II.
3 S1 Highgate Honduras,i IP
PAJ1 Cavendish Pajalillos, Canary Islands,d II
ORT1 Cavendish La Orotava, Canary Islands,d II
GALl Cavendish Las Galletas, Canary Islands,d II
BUEl Cavendish Buenavista, Canary Islands,d II
NW Williams Natal, South Africa,f II
NH Williams Natal, South Africa,f II
NB Cavendish Natal, South Africa,f II
F9127 Grand Naine South Africa,g II
15638 x Malaysia,a II
FCJ7 Lacatan Jamaica,q II
Pacovan Pacovan Bahia, Brazil,n II
MGSA1 SH3142 South Africa,s II
SA6 Cavendish Transvaal, South Africa,b II
SA4 SH3362 Natal, South Africa,b Ir
SA3 Williams Tansvaal, South Africa,b II
0121 GM Gros Michel Taiwan,h
9130 Cavendish Taiwan,g III
0-1124 x Taiwan,c IIIx
H1 Cavendish Taiwan,e III
ML Cavendish Taiwan,h III
TBR Cavendish Taiwan,h III
0122 Ph3 Cavendish Philippines,l VI
Ph6 Cavendish Philippines,l VI
P79 Cavendish Philippines,h VI
LAP Cavendish Philippines,h VI
SABA Saba Philippines,h VI
PW3 Cavendish Philippines, m VI
Table 2- 1 --continued 19
VCG" Isolate Cultivar' Origin and collector Lineage
PW6 Cavendish Philippinesm VI
PW7 Cavendish Philippinesm VI
0123 DAVAO Silk Philippinesh VII
TI Gros Michel Taiwanf VII
PhL2 Latundan Philippinesi VII
Ph12 Latundan Philippinesl VII
9129 Latundan Taiwang VII
JLTH4 Klue namwa Smoeng Hwy 1269, Thailandv x
JLTH5 Klue namwa Smoeng Hwy 1269, Thailandv
0124 A36 x Brazilk
GMB Gros Michel Braziln
Maca Maca Braziln
STPAI Pisang Awak Burundii
STD2 Highgate Hondurasi
BLUG Bluggoe Hondurash
Sx Tetraploid 1242 Bodles, Jamaica, i
FCJ2 Bluggoe Jamaicaq
FCJ3 x Jamaicaq
FCJ8 x Jamaicaq
FCJ9 Tetraploid 1242 Jamaicaq
STJ2 Grande Naine Jamaicaj
MW43 Harare Chitipa, Karonga, Malawib
MW45 Harare Chitipa, Karonga, Malawib
NM47 Harare Chesenga, Malawib Ix
MW50 Harare ChitipaKaronga, Malawib I
MW52 Sukali Karonga South, Malawib I
MW58 Harare Karonga, Malawib I
NM64 Harare Kaporo North, Malawib I
MW67 Kholobowa Thyolo, Blantyre, Malawib I
MW69 Kholobowa Thyolo, Blantyre, Malawib I
MW71 Kholobowa Mulanje, Blantyre, Malawib I
MW78 Harare Vinthukutu, Karonga, Malawib I
STN2 Bluggoe Corinto, Nicaraguai I
STN5 Bluggoe Corinto, Nicaraguai I
STN6 Bluggoe Corinto, Nicaraguaj I
STN7 Bluggoe Corinto, Nicaraguaj I
STPA2 Pisang Awak Tanzaniai I
B 1 Burro(Bluggoe) FloridaUSAb 1
0124 B2-1 Burro FloridaUSAb I
JCB I Burro FloridaUSAb I
YLTH2 Klue namwa Smoeng Hwy 1269, Thailandv I
JLTH7 Klue namwa Smoeng Hwy 1269, Thailandv I
JLTH 15 Klue namwa Chai Yo Hwy, Thailandv I
Table 2-1 --continued 20
VCGu Isolate Cultivarv Origin and collector' Lineagze
MW9 Zambia Kaporo, Malawi,r I
MWI 1 Harare Kaporo, Malawi,r I
M1W39 Harare Chitipa, Karonga, Malawi,b I
MW53 Sukali Karonga, Malawi,b i
MW56 Zambia Karonga, Malawi,b I
MW 60 Zambia Karonga, Malawi,b I
M W6 I Harare Vinthukutu, Karonga, Malawi,b I
NMW63 Harare Karonga, South, Malawi,bI
MIW66 Kholobowa Thyolo, Blantyre, Malawi,bI
MW70 Kholobowa Thyolo, Blantyre, Malawi,bI
MW 86 Mbufu Chitipa, Karonga, Malawi,bI
JLTIi Klue namwa Ban Nok, Thailand,vI
JLTH16 Klue namwa Ban Nok, Thailand,v I
JLTH 17 Klue namwa Ban Nok, Thailand,v JX
JLTH 18 Klue namwa Ban Nok, Thailand,v JX
JLTH1 9 Klue namwa Ban Nok, Thailand,v I
0125 A4 Lady Finger Australia,f 1
8606 Lady Finger Currumbin, Queensland,g I
8611 Lady Finger Currumbin, Queensland,g 1
22468 Lady Finger Tomewin, Queensland,g I
22479 Ducasse Bowen, Queensland, Australia,g Py
22600 Lady Finger Murwillumbah, New S. Wales, Australia,g 1
22417 Lady Finger Rocksberg, Queensland, Australia,g 1
22541 Lady Finger Murwillumbah, New S. Wales, Australia,g I
0-1223 Mons Queensland, Australia,c I
1 5X Williams Bodles, Jamaicaji I
STPA3 Pisang Awak Ugandaji I
JLTH20 Klue namwa Ban Nok, Thailand,v 1
0126 51 Highgate Honduras,i 1I
STA2 Highgate H-ondurasji 1I
STM3 Maqueno Hondurasji I1
STB2 Highgate Hondurasji II
0128 22993 Bluggoe South Johnstone, Queensland,g 1
22994 Bluggoe South Johnstone, Queensland, Australia,g I
A47 Bluggoe Comores Islands~j Ix
0129 N5331 Cavendish Yandina, Queensland, Australia,g 11
0-122 1 Mons Queensland, Australia,c 11
N5443 Cavendish Doonan, Queensland, Australia,g II
8627 Cavendish North Arm, Queensland, Australia,g 11
22402 Cavendish Wamuran, Queensland, Australia,g 1I
8604 Cavendish North Arm, Queensland, Australia,g I1
Table 2-1--continued 21
VCGu Isolate Cultivarv Origin and collector"' Lineage
01210 A2-1 Apple Florida, USA,b IV
A4-1 Apple Florida, USA,b IV
CSB Apple Florida, USA,b IV
JC14 Apple Florida, USA,b IV
A15 Apple Florida, USA,b IV,
A3-I Apple Florida, USA,b IV
JC8 Apple Florida, USA,b IV
F2 Apple Florida, USA,b IV
F3 Apple Florida, USA,b IV
JC1 Apple Florida, USA,b IV
GG1 Apple Florida, USA,b IV
01211 13721 x xa IXx
SH3142 SH3142 Queensland, Australia,g IX
01212 STNP1 Ney Poovan Pemba Island, Zanzibar, Tanzania,i VIII
STNP2 Ney Poovan Tenguero Station, Tanzania,i VIII
STNP4 Ney Poovan Bukava Station, Tanzania,i VIII
01213 1-1 Cavendish Taiwan,u III
2-1 Cavendish Taiwan,u III
6-2 Cavendish Taiwan,u III
5-1-1 Cavendish Taiwan,u III
4-2-1 Cavendish Taiwan,u III
4-1-1 Cavendish Taiwan,u III
2-2 Cavendish Taiwan,u III
ES2-1 Cavendish Taiwan,u
01214 MW2 Harare Misuku, Karonga, Malawi,r V
MW40 Harare Misuku Hills, Karonga, Malawi,b V
MW41 Mbufu Misuku Hills, Karonga, Malawi,b V
MW42 Harare Misuku Hills, Karonga, Malawi,b V
MW44 Harare Misuku Hills, Karonga, Malawi,b V
MW46 Harare Misuku Hills, Karonga, Malawi,b V
MW48 Harare Misuku Hills, Karonga, Malawi,b V
MW51 Harare Misuku Hills, Karonga, Malawi,b V
MW89 Harare Misuku Hills, Karonga, Malawi,b V
01215 CR1 Gros Michel Isolona,Costa Rica,b II
CR2 Gros Michel Hamburgo,Rio Reventazon,Costa Rica,b II
CR4 Gros Michel Hamburgo,Rio Reventazon,Costa Rica,b II
CR5 Gros Michel Hamburgo,Rio Reventazon, Costa Rica,b II
Table 2-1--continued 22
VCGU Isolate Cultivarv Origin and collectors Lineage
INDO20 Musa spp. Jatesari, East Java,Indonesia,g III
INDO15z Musa spp. Jatesari, East Java,Indonesia,g III
INDO 18Z Musa spp. Jatesari, East Java,Indonesia,g III
Fusarium oxysporumn f sp. lycopersici
SC548 Homestead Bradenton, Florida
SC626 Oristano Italy
SC761 Sunny Bradenton, Florida
Fusarium oxysporum f sp. niveum
0082 CS85-4 Florida,w
uVegetative compatibility groups (VCGs) were assigned using nitrogen metabolism (nit) mutants according to the protocols of Cove (1976) as modified by Puhalla (1985).
vCultivars are inter- and intraspecific diploid or triploid hybrids ofMusa acuminata (AA) and Musa balbisiana (BB). The ploidy levels and constitutions of the cultivars are as follows: AAA = Gros Michel, Highgate, Mons (Mons mari), Cavendish, Dwarf Cavendish, Grande Naine, Williams, Lacatan; AA = SH3142 and SH3362 (synthetic clones); AAB = Lady Finger, Pacovan, Prata, Silk, Latundan, Maquefio; ABB = Saba, Bluggoe, Harare, Kholobowa, Pisang awak, Klue namwa, Ducasse, Mbufu, Burro, Zambia; AB = Ney poovan, Sukali; AAAA = Tetraploid 1242
missing data made it impossible to determine coefficient of similarity but based on all other data it is presumed to be in the lineage indicated.
Y= unique isolate that had no lineage affinity based on defined criteria; the lineage assigned is based on coefficients of similarity to the most closely related isolates. = isolates analyzed with this designation do not correspond to any isolates currently held in the Homestead collection.
WCollector or original source: a, American Type Culture Collection; b, R.C. Ploetz, Homestead, Florida; c, Paul E. Nelson, Fusarium Research Center, University Park, Pennsylvania; d, J.H. Hemrnandez, Tenerife, Canary Islands; e, S.C. Hwang, Taiwan Banana Research Institute, Pingtung; f, B. Manicom, Nelspruit, South Africa; g, K. Pegg, Brisbane, Australia; h, S. Nash Smith, Alameda, California; i, R.H. Stover, La Lima, Honduras; j, IFRA, Montpellier, France (via R.C. Stover); 1, A.M. Pedrosa, Philippines; m, N.I. Roperos, Philippines; n, Z. J. M. Cordeiro, EMBRAPA, Cruz das Almas, Brazil (via E.D. Loudres); q, J. Ferguson-Conie, Banana Board, Kingston, Jamaica; r, B. Braunworth, Oregon State University, Corvallis; u, Tsai-young Chuang, National Taiwan University, Taipei; v, J. Leslie, Kansas State University, Manhattan; w, F. Martin, USDA, ARS, Salinas, CA.
F. oxysporurn f sp. niveum were analyzed to compare results with genotypes representing different formae speciales.
All cultures were derived from single microconidia and were stored at 4C on
strips of Whatman filter paper (Correll et al., 1987). Paper strips were plated onto potato dextrose agar (PDA) (Difco, Detroit, MI). After approximately 7 days of growth at 270C, a 5-mm3 block was excised from the margin of colonies on PDA and transferred to 4-L Erlenmeyer flasks containing at least 100 ml of potato dextrose broth (Difco 24 g/L). After 7 to 10 days of growth in still culture, the contents of the flask were filtered through sterilized cheese cloth to collect mycelium. The mycelium was placed into 13-ml plastic tubes, frozen at -80'C and lyophilized for at least 12 hours.
Buffer for DNA extraction was comprised of a 1:1.0.4 volume of the following solutions: Buffer A (0.3 M Sorbitol, 0.1 M Tris, 20 mM EDTA at pH 7.5), Buffer B (0.2 M Tris at pH 7.5, 50 mM EDTA and 0.2 mM CTAB) and 5% Sarkosyl. Ten ml of the extraction buffer were mixed with approximately 0.5 g of ground mycelial powder, and the tubes were placed in a 65C water bath for 30 minutes. The contents of the tubes were then shaken and 1 ml of solution was transferred to each of 10, sterile, 1 .5-ml microcentrifuge tubes. Five hundred microliters of chloroform:octanol (24:1) solution were added to each tube. The solution was mixed using a vortex shaker for approximately
2 minutes before centrifugation for 10 minutes at 12,000 g in a microcentrifuge at room temperature. Unless noted, all further centrifuge steps were done in a microcentrifuge at room temperature. The supernatant was transferred to sterile 1.5-ml tubes and treated with 5 al of 20 mg RNase A (Sigma Chemical Co., St. Louis, MO)/ml of solution for 30
minutes at 37C. Following RNase treatment, 5 pl of 20 mg proteinase K (Sigma)/ml of solution were added to the tubes, and the tubes were incubated for 20 minutes at 37C. Approximately one volume isopropanol was added, and the tubes were centrifuged for 15 minutes at 12,000 g. The isopropanol then was discarded, and 100 1 of ice-cold 70% ethanol were added before centrifugation for five minutes. The ethanol was discarded, and the DNA sample was air-dried for at least 30 minutes in a laminar flow hood. At least 100 pla of TE buffer (10 mM Tris, 1 mM EDTA at pH 7.4) were added to the tubes, and they were placed in a water bath at 65C until the DNA pellet was dissolved.
Samples which were difficult to bring into solution were subjected to a LiCI
treatment. Three hundred microliters of ice-cold 4M LiCI solution were added to each tube, and the tubes were placed on ice for 30 minutes before centrifugation at 12,000 g for 10 minutes at 4C. The supernatant was transferred to a sterile 1.5-ml tube containing 600 ll of isopropanol. This solution was mixed, and the tubes were kept at room temperature for 30 minutes. After centrifugation at 12,000 g for 10 minutes at 4C, the supernatant was discarded and 100 pl of ice-cold 70% ethanol were added to the tubes. After centrifugation at 12,000 g for 5 minutes, the ethanol was discarded and the DNA was airdried. TE buffer (100 l) was added to the tubes and they were placed in a water bath at 650C until the pellet was dissolved.
The concentration of DNA in the samples was estimated by running 3 p1 of each sample on an agarose gel along with DNA fragments (bacteriophage lambda DNA digested with PstI) of a known concentration and making visual comparisons of their relative fluorescence in the presence of ultraviolet (UV) light and 0.5 !ag of ethidium bromide/ml of solution.
Southern Blotting and Hybridization
Approximately 10 ig of DNA were digested with at least 10 units of either Dral, EcoRV or HaeIII restriction enzymes (Bethesda Research Laboratories, Gaithersburg, MD) and incubated for at least 3 hours. These restriction enzymes were chosen based on their insensitivity to DNA methylation and their ability to digest DNA consistently. Restriction fragments were separated by electrophoresis in 0.7% agarose for EcoRV- and DraI-digested DNA or 1.5% agarose for HaeIII-digested DNA in TBE buffer at pH 7.0. Gels were run at either 30 volts for approximately 16 hours or 40 volts for approximately 12 hours. Bacteriophage lambda DNA digested with either PstI or HindIII was included on each gel and used to calculate the molecular mass of restriction fragments obtained from F. oxysporuni f sp. cubense DNA. Ethidium bromide (10 mg/ml) was dissolved in the agarose gel at a concentration of 1 p11-100 ml, and the digested DNA was illuminated by UV irradiation and photographed. The DNA was transferred to Nytran membranes (Schleicher & Schuell, Inc., Keene, NH) using the capillary transfer method (Sambrook et al., 1989). The DNA transfer proceeded for at least 12 hours, and the DNA was immobilized by UV cross linking (UV 254-nm cross linker, model FB UVXL 1000, Fischer Scientific, Pittsburgh, PA).
To reduce the incidence of repeatedly scoring similar regions of the genome or hypervariable regions, clones containing single-copy DNA sequences, obtained from Talma Katan (The Volcani Center, Bet Dagan, Israel), were utilized (Elias, et al., 1993). Clones were considered to be single-copy based on the criteria defined by Elias et al. (1993) as hybridizing to only a single DNA fragment in any of the isolates tested using at least one of three restriction enzymes (Figure 2-1).
1 4 5 6 71 0 11 12 13 14 15 16 17 18 19
W eight ::.:.::::::::::::::.::.:::
N .. ::.:.:. :... K ii..:.:....:, -,::.F.f ++:...:::::: :>
1 XA 201 001 0
2 ADJ 0.2 .010.00 ...
........... . .
........... ...A .. .. .. .. ..C
Isolate VCG Polymorphic State
IGAL2 0120 01000
2 ADJ2 0120 01000
3 01222 0120 01000
4 F9127 0120 01000
5 Pacovan 0120 01000
6 GM 0121 00100
7 SABA 0122 01000
8 FCJ9 0124 01000
9 MW60 0124/25 01000
10 CVA 0124/25 01000
11 S? 0125 01000
12 JLTH1 0124/25 01000
13 JLTH21 0125 01000
14 22994 0128 01000
15 A47 0128 01000
16 MW2 01214 10010
17 1-2 01213 10000
18 22507 0129 01000
19 SI 0126 01000
Figure 2-1: Single-copy sequences of selected Fusarium oxysporum f. sp. cubense isolates. DNA of isolates was digested with Eco RV restriction enzyme and probed with clone 187. Polymorphic states indicate the presence (1) or absence (0), respectively of the
6.5, 5.1, 4.6, 4.2, or 3.4 Kb bands
DNA for each clone was labeled using random hexamer primers to incorporate fluorescein- 12-dUTP following the procedures provided by the manufacturer (Dupont NEN "Renaissance," E. I. du Pont de Nemours & Co. Inc., Boston, MA). DNA labeling, hybridization and detection followed the procedures provided by the manufacturer (Dupont, Boston, MA). Prehybridization, hybridization and washing were performed at 65C using a Hybaid hybridization oven (Dot Scientific Inc., Flint, MI). Membranes were placed between acetate sheets and exposed to X-ray film for at least 5 hours.
Mitochondrial DNA ofF. oxysporum f sp. cubense isolate 3S1 (VCG 0120) was isolated following the procedures of Kistler and Leong (1986) and labeled as described above. The mitochondrial DNA profiles of a subset of isolates were obtained by digesting approximately 10 gg of total DNA with at least 10 units of the restriction enzyme HaeIII and probing with the mitochondrial DNA of isolate 3S 1. Restriction fragments were separated by electrophoresis in 1.5% agarose in TBE buffer at pH 7.0. Gels were run at 30 volts for approximately 22 hours. Southern blotting and hybridization followed the procedures described above.
Initially, a subset of 38 geographically widespread isolates ofF. oxysporum f. sp. cubense that represented all 17 VCGs was used to determine if a particular probe-enzyme combination was polymorphic. Only polymorphic loci were considered informative for phylogeny determinations. If polymorphisms were detected in the subset, then all 165 isolates were analyzed for that probe-enzyme combination. If all isolates within the subset were monomorphic, it was assumed that the entire collection was monomorphic for that
probe-enzyme combination. The different restriction size fragments generated for each combination of probe and enzyme were considered to be alleles at a single RFLP locus, and their presence or absence was scored for each isolate. RFLP patterns for each combination of probe and enzyme were combined to assign an RFLP haplotype to each isolate.
The data were analyzed by a cladistic approach based on parsimony analysis using the computer program PAUP Version 3.1.1 (Swofford, 1993), and by a phenetic approach using distance matrix methods (UPGMA clustering, Sneath and Sokal, 1973) and the neighbor-joining algorithm of Phylip Version 3.5c (J. Felsenstein, University of Washington). For parsimony analysis, phylogenies were derived by using the heuristic search option, and the degree of support was evaluated using 500 bootstrap replicates. In addition, coefficients of similarity based on simple matching were calculated for those isolates in which data were available for every RFLP loci scored, based on the formula described by Sneath and Sokal (1973). Isolates were arbitrarily considered to be within the same clonal lineage based on coefficients of similarity ranging from 0.94 to 1.00. Since many of the isolates had identical multilocus haplotypes, only a single isolate was used to represent each haplotype in data analyses.
To determine whether this collection of isolates provided evidence for clonal
reproduction, the gametic disequilibrium coefficient (D) was calculated among pairs of alleles at different loci by methods described by Weir (1990). Clone-corrected allele frequencies, using only a single representative for each haplotype, were employed for the calculations. Also, to avoid the potential problem of repeatedly scoring similar regions,
only data from a single restriction enzyme digestion were used for each probe in the analyses. Nine hundred and eighty-eight pairwise comparisons were performed to test for disequilibrium between multiple alleles at nine loci. A test for the significance of the disequilibrium coefficient between each pair of alleles at two loci was formulated with the chi-square statistic
where n was the number of individuals in the sample and D!,Vwas the maximum likelihood estimator for the coefficient of disequilibrium between alleles u and v.
The observed allele frequencies for the loci were Pt and P, respectively
(McDonald et al., 1994; Weir, 1990). The chi-square statistic had one degree of freedom and the pairs of loci that showed significant departure from random expectations (P <0.05) were considered to be in disequilibrium. A test for significance of the disequilibrium coefficient across all alleles for each pair of loci was formulated with the chi-square test statistic as described by McDonald et al. (1994).
The isolates tested in the study were assigned previously to a VCG, and one to several representative isolates within each of the examined VCGs were pathogenic on at least one and as many as eight different banana cultivars (Ploetz, personal communication).
Although one sample of DNA was isolated from each isolate, many isolates were tested at least once and as many as three times for each probe enzyme combination examined. Additionally, a subset of 24 F. oxysporumr f sp. cubense and 2 F. oxysporum f
sp. lycopersici isolates tested here were separately analyzed using DNA sequencing data (O'Donnell, personal communication). Similar phylogenetic relationships were observed using this separate analysis, further validating the groups resolved in this analysis.
In this study, 38 isolates ofF. oxysporum f. sp. cubense were screened for
polymorphisms using 19 probe-enzyme combinations. Only six of the 19 probe-enzyme combinations were monomorphic among the 38 selected isolates, indicating a high degree of genetic diversity among the isolates. The entire collection of 165 isolates was then scored for polymorphisms using the 13 probe-enzyme combinations that were found to be informative during the initial screening of isolates. A multilocus RFLP haplotype was assigned to each isolate based on the allelic data for all probe-enzyme combinations. Only 72 distinct multilocus haplotypes were detected among the 165 isolates, 50 of which were represented by a single isolate (Figure 2-2). The five most common haplotypes represented 45% of the isolates. The median number of alleles per locus was three, and if three alleles were present at each locus, theoretically 1.16 x 109 (3 possible haplotypes could exist for this collection ofF. oxysporum f. sp. cubense. However, the majority of single-isolate haplotypes found were the result of one to a few allelic differences from a more common haplotype.
To determine relatedness among isolates, the 72 RFLP haplotypes were subjected to phenetic and cladistic analyses. Both types of analyses produced trees with similar branching patterns. The 50% majority rule bootstrap consensus tree generated by PAUP 3.1.1. is presented in Figure 2-3. A dichotomy with strong bootstrap support (99%) was
0 E z
LID LO N LO r I
Figure 2-2: Frequency distribution of RFLP haplotypes among the 165 isolates representing a world collection of Fusarium oxysporum f sp. cubense
SFOL SC626 F oxysporum
4) 1 L5(90) 14fOL SC548 l sp lycopersici
L FOL SC761 and
1 VCG 0122 Ph3 F oaysporun
VCG 0122 LAP f sp. nivean
8(96)] VCG 0122 PW7 FOCVI
VG 0122 P79
VCG 0122 SABA01214 MW VCG 01214 MW40 VCG 01214 MW41
4(69) VOCG 01214 MW42 f 01VCG 01214 MW48 FOCV
VCG 01214 MW44
GVCG 01214 MW46 V 0VCG2 01214 MW51
5 VCGVCG 01214 MW89 VCG 0120 C2.. S VCG 0120 F9127
VCG 0120 PAJ1 FOC II
VCG 0120 IC2
2 VCG 0120 Pacovan
04 VCG 0121 GM IFOCX
1-- VCG 0121 F9130 I
SVCG 0121 H1 5(58) VCG 01213 6-2
5(8 VCG 01213 1-2 FOC II
-I- VCG 01213 4-1-1 VCG 0121 0-1124 VCG 0120/01215 INDO20 2
3 VCG 0123 DAVAO FOC VI
14(74) VCG 0123 Phl2 F
1 VCG 0123 T1 VCG 0123 PhL2
15 VCG 0123 JLTH4 FOC X
1 VCG 0123 JLTH5
VCG 0120 MGSAI VCG 0L20 SA4 .. ... VCG 0120 SA3 FOC II
7(83) VCG 0126 STA2
_VCG 0126 STB2
5 VCG 0126 S1
VCG 0129 N5443
VCG 01210 A2-1 VCG 01210 GG 1 FOC IV
,, VCG 01210 A15
VCG 01211 13721 4 FOC IV
VCG 01210 JC1 FOC IV A VCG 01211 SH3142 r4FOCa VOG 01213 ES2-1 A VCG 0124 A36 ......... VCG 0124 GMB
-... VCG 0124 Maca ....... VCG 0124 STJ2 VCG 0124 MW43 VpG 0124 MW47
VCG 0124 JLTH15 VCG 0124 MW64
VCG 0124 STN5 FOC I
VCG 0124 STN7 VCG 0124/25 MW9 13(99) VCG 0124/25 MW11
VCG 0124/25 MW53 VCG 0124/25 MW63 VCG 0124/25 MW86 nVCG 0125 8606 VCG 0125 22479 VCG 0125 22541
VCG 10128 A47
4 83 VCG 01212 STNP1 FOC VIII
VCG 01212 STN?4
2 VCG 0124/25 JLTH16 VCG 0124/25 JLTH17 FOCI
VCG 0124/25 JLTH18 VCG 0124/25 JLTH19
Figure 2-3: Midpoint rooted 50% majority rule consensus tree representing 500 bootstrap
replicates; One isolate represents each of the 72 RFLP haplotypes of Fusarium oxysporum
f. sp. cubense, three isolates ofF oxysporum f sp. lycopersici and one isolate E
oxysporum f. sp. niveum. Branch lengths are indicated on each branch and bootstrap
values are in parenthesis. Tree length= 351; Consistency Index= 0.214; Homoplasy
Index = 0.786; Retention Index = 0.705.
observed on the midpoint-rooted tree among the 72 haplotypes representing the 165 F. oxysporuni f sp. cubense isolates and four isolates from other formae speciales. Isolates ofF. oxysporum f. sp. cubense belonging to VCGs 0124, 0124/0125, 0125, 0128, and 01212 formed one main phylogenetic branch, while isolates belonging to VCGs 0120, 0121, 0122, 0123, 0126, 0129, 01210, 01211, 01213, 01214, 0120/01215 and 01215, as well as the isolates ofF. oxysporum f sp. lycopersici and f. sp. niveum, were found on the second main branch.
Isolates comprising these two branches could be further divided into eight major clades which have moderate to strong bootstrap support (values greater than 70%). Within one of the branches, isolates in VCGs 0124, 0124/0125, 0125, and 0128 formed one clade. Isolates in VCG 01212 were genetically similar but distinct (83% bootstrap support) (see below) from these isolates. Within the other branch, six clades were identified. Isolates in VCGs 0122 (96% bootstrap support), 0123 (74% bootstrap support), 0126 (83% bootstrap support), 01214 (97% bootstrap support), and isolates of F. oxysporum f sp. lycopersici (100% bootstrap support) each formed their own clade. Isolates in VCGs 0121, 01213, and three isolates in 0120/01215 formed a clade of weak support (58%), and therefore could not be confidently differentiated from isolates in VCGs 0120, 0123 (two isolates) 0129, 01210, 01211, 0120/01215 and the single isolate ofF. oxysporum f sp. niveum.
All clades with strong bootstrap support are comprised of isolates that have identical or nearly identical multilocus haplotypes and are referred to here as clonal lineages. Additionally, many of the isolates which could not be resolved using bootstrap
analysis shared nearly identical multilocus haplotypes with other isolates. To further understand the genetic relationships among unresolved isolates, a simple matching coefficient of similarity for comparison was used (Table 2-2). Isolates with coefficients of similarity ranging from 0.94 to 1.00 were considered to be within a clonal lineage. This range reflects apparent natural groups (Figure 2-3); isolates within a lineage possess either small or no genetic differences. In cases where isolates did not fall within this range for all pairwise comparisons, isolates were included in the lineage if they shared values near or within the specified range with the majority of isolates comprising the lineage. These isolates are marked with an asterisk in Table 2- 1. Two isolates, ES2-1 and GM, could not be assigned to lineage based on these criteria.
A similarity matrix, which includes coefficients of similarity for selected isolates representing the major RFLP haplotypes and VCGs, is presented (Table 2-2). In general, VCGs aligned with single clonal lineages; exceptions to this were isolates in VCGs 0 123. Table 2-3 lists each lineage (with the prefix FCC), the number of isolates represented, the VCG of each lineage, and its geographical distribution. Seventy-four percent of the isolates studied were represented by Lineages FCC I, FCC 11 and FCC 111. Each of these three lineages contain more than one VCG, with lineages FCC I and FCC 11 having a pantropical distribution. Isolates in FCC IV through X each belong to a single VCG and represent one to a few geographical regions.
Coefficients of similarity between isolates of FCC I and FCC II ranged from
0.66 to 0.74. By comparison, two isolates of F. oxysporlim f. sp. lycopersici had coefficients of similarity ranging from 0.55 to 0.71 compared to isolates in FCC I and 0.64
Table 2-2: Similarity matrix of simple matching coefficients based on restriction fragment length polymorphisms for selected isolates of Fusarium oxysporum f sp. cubense, F. oxysporum f sp. lycopersici and F. oxysporum f sp. niveum.
C0 () t C9 C C) t C) C C) C) to t, C, C 0 0 C) 0 C) C)
VCG 0120 9127 I
VCG 0120 IC2 0.98 1
VCG 0121 F9130 0. 82 0.85 1
VCG 0121 HI 0.86 0.88 096
VCG 0122 SABA 0.91 0 88 0 82 0 86 1
VCG 0123 DAVAO 0.78 0 75 0.74 0.75 0.75 1
VCG 0123 TI 08 0M78 074 078 0.78 093 I
VCG 0123 Phl2 0.76 0.76 0.73 0.76 0.74 0.94 0.92 1
VCG 0123 PhL2 0.81 0.79 0(75 0.79 0.79 0.94 0(99 0.93 1
VCG 0123 JLTH4 0.78 0075 0.72 0075 0.8 088 0.93 0 87 0.94 1
VCG 0124 MW43 0.73 071 0.62 0.66 0.75 074 079 073 0.8 0.81 1
VCGO0124MW64 0.72 0.69. 0,61 065 0.74 0.75 0.78 0.74 0.79 0.8 0.99 I
VCG0124/0125 MV11 0.74 0.72 0. 64 0.67 076 073 078 0.72 0.79 0.85 096 0.95 1
VCG 0125 8606 0.71 0.68 06 064 073 0.74 0.76 073 0.78 079 0.98 099 0.94 1
VCG 0126STA2 096 0.96 0.84 085 0.87 076 0.79 0.75 0.8 0.76 072 0.71 073 0.69 1
VCG 0126 SI 0.94 094 0.81 0.82 085 0.74 0.79 0.75 0.8 076 0.72 0.71 073 069 0.98 1
VCG 0129 N5443 096 0.99 0.86 087 0.87 0.74 076 0.75 0.78 0.74 069 0.68 0.71 0.67 0.95 0.93 1
VCG 01210 A2-1 0.89 092 091 0.92 87 07 .81 0.84 0.82 085 0.81 072 0.71 0.73 0.69 0.93 0.91 0.91 1
VCG01211 SH3142 0.91 0.93 0.85 0.86 0.86 0.75 0.75 0.74 0.76 0.78 0.71 0.69 0.69 0.68 0.89 0.87 0.92 0.89
VCG01212 STNP4 0.73 0.71 0.6 0.64 0.75 072 0.76 071 0.78 0.79 0.95 0.94 0.92 0.93 0.72 0.72 0.69 0.69 0.71 1
VCG01213 1-2 0.86 0.88 0.94 0.98 0.84 075 0.8 0.79 0.81 0.78 0.68 0.67 0.69 0.66 0.85 0.85 0.87 0,92 086 0.66 1
VCG01214MW2 0.76 0.74 0.75 074 0.79 0.75 08 0,74 0.79 0.8 0.68 0.67 0.69 0.66 078 0.75 0.75 0.8 0.76 068 0_74
VCG 01214 MW41 0.76 0.74 0.78 0.74 0.79 0.78 08 0.74 0.79 0.8 0.68 067 0.69 0.66 0.78 0.75 0.75 0.8 0.79 0.71 0.74 0.95
VCG 01214 MW51 079 0.76 075 074 0.79 0.75 0.8 074 0.79 0.8 0.68 0.67 0.69 0.66 08 0.78 0.78 0.8 0.79 0.68 0.74 0.98 0.95 1
VCG01214MW89 0,8 0.78 074 0.78 0.82 0.79 084 0.78 0.82 0.84 0.72 0.71 0.73 069 0.79 0.76 0.76 0.81 0.8 0.72 0.78 0.94 0.94 0.94 1
VCG012001215Indo20 0.87 0.89 0.95 0.99 085 0.76 0.79 078 0.8 0.76 067 0.66 0.68 0.65 0.86 0.84 0.88 0.93 087 065 0.99 0.75 0.75 0.75 079 I
SC626 lycopersici 071 073 0.76 075 075 069 0.72 06 0. 0.73 0 74 069 0.68 0.71 0.67 0.72 0.72 0.74 0.76 075 0.67 075 0.66 0.68 0.66 067 0.74 1
SC761 lycopersici 0.64 066 072 071 0 .71 06 062 6 0641 0.64 065 06 0.59 0.61 0.58 065 065 0.9 0. 0. 0. 0 67 0.69 068 062 071 0.64 0.66 0.61 0.62 0.69 088
CS85-4 niveum 074 0.74 075 076 0.76 08 085 79 0.8 .86 0.85 0.8 079 079 0.78 075 0.75 075 0.8 076 078 0.79 0.81 079 0.81 0 8 0 78 075 0.71 1
Table 2-3: Clonal lineages of Fusarium oxysporuni f sp. cubense isolates, their geographical distributions and corresponding vegetative compatibility groups (VCG).
No. of VCG
Lineage Isolatesz Geographic Distribution Represented
FOC I 65 Australia, Brazil, Burundi, 0124,0124/25,
Comores Islands Honduras 0125, and
Jamaica, Malawi, Nicaragua, 0128
Tanzania, Thailand, Uganda,
and the United States (Florida)
FOC II 43 Australia, Brazil, Canary Islands, 0120, 0126,
Costa Rica, Honduras, 0129, and
Jamaica, Malaysia, South Africa, 01215
FOC III 15 Indonesia and Taiwan 0121, 01213,
FOC IV 11 United States (Florida) 01210
FOC V 9 Malawi 01214
FOC VI 8 Philippines 0122
FOC VII 5 Philippines, Taiwan 0123
FOC VIII 3 Tanzania 01212
FOC IX 2 Australia 01211
FOC X 2 Thailand 0123
ZIsolates ES2-1 in VCG 01213 and GM in VCG 0121 did not align with any lineage based on coefficient of similarity data
to 0.74 compared to isolates in FOC II. The single isolate ofF. oxysporum f sp. niveurn had coefficients of similarity ranging from 0.75 to 0.81 compared to isolates in FOC I and from 0.73 to 0.76 compared to isolates in FOC II. Thus, the two largest lineages ofF. oxysporuni f sp. cubense each are more genetically similar to the F. oxysporun f sp. niveun isolate than to each other. Similarly, they are roughly as genetically distinct from each other as either is to the F. oxysporun f. sp. lycopersici isolates.
All isolates in VCG 01214 comprise FOC V. Isolates in this lineage formed a clade exhibiting the longest branch length compared to all other clades representing F. oxysporuni f sp. cubense isolates. In fact, its length was comparable to the branch lengths of the clades containing the isolates from the other formae speciales (Figure 2-3). Additionally, isolates comprising this lineage had coefficients of similarity which did not align closely to any other lineage. For example, isolates in FOC V had coefficients of similarity ranging from 0.62 to 0.73 compared to isolates in FOC I, from 0.72 to 0.81 compared to isolates in FOC II and from 0.61 to 0.69 compared to the F. oxysporun f. sp. lycopersici isolates. This group had a large number of lineage-specific alleles at several RFLP loci which accounted for its relative lack of similarity to other F oxysporum f sp. cubense isolates. All of the isolates in VCG 01214 have a very limited geographical distribution. Isolates in this group also do not form chlamydospores (R.C. Ploetz, unpublished) as other F oxysporum isolates do, which is usually a defining trait for the species.
FOC VII consisted of the majority of isolates in VCG 0123. FOC X consisted of two additional isolates in VCG 0123. Similar to FOC V, isolates comprising these lineages had coefficients of similarity which did not align closely to any other lineage.
Surprisingly, the RFLP multilocus haplotypes of isolates within these lineages consisted of alleles similar to those in both FOC I and II, in addition to some lineage-specific alleles (Table 2-4). A representative FOC VII isolate (PhL2) shared 43% (6/14) of polymorphic alleles with isolates in FOC II and 36% (5/14) of these alleles with isolates in FOC I. By contrast, the isolates in FOC I displayed only 29% (4/14) allelic similarity with isolates in FOC II. With the exception of FOC V, isolates in the other lineages more closely aligned with isolates in either FOC I or II. Unlike isolates in other lineages, the multilocus haplotype of isolates in FOC VII and X appear to represent a combination of alleles from FOC I and II.
The mitochondrial DNA haplotypes of a subset of 55 isolates representing the
major VCGs were examined to determine their genetic relationships based on an additional independent genetic marker. Isolates within the same VCG shared identical mitochondrial DNA haplotypes. Isolates could be divided into three major groups based on visual assessment of similar though not identical, mitochondrial RFLP patterns (Figure 2-4). Isolates in VCGs 0123, 0124, 0124/0125, 0125, 0128, 01212, and 01214 formed one group. Isolates in VCGs 0120, 0121, 0122, 0129, 01213, 0120/01215 and 01215 formed a second group. Isolates in VCGs 0126 and 01210 formed a third group. In general, these groupings aligned with those based on RFLP analysis of single-copy loci, although more groups could be resolved using the latter method.
Measures of gametic disequilibrium were performed for alleles at nine loci (Table 2-5). Even though many of the individual comparisons were not significant, 34 of the 36 pairwise comparisons among alleles at different RFLP loci showed significant
Table 2-4: A comparison of allelic data for isolates in FOC I, FOC II and FOC VII.
Isolate (VCG)y Allelesz
A36 (VCG 0124) 11111111111111
IC2 (VCG 0120) 22122122221222
PhL2 (VCG 0123) 12132123331113
YIsolates A36 and IC2 represent the two largest RFLP haplotypes in FOC I and FOC II, respectively, and PhL2 represents an isolate in FOC VII.
zAlleles found in the most common haplotypes were given the number 1. Alleles in the second most common haplotype, if different from haplotype 1, were given the number 2. Alleles in isolate PhL2, if different than those in the two most common haplotypes were given the number 3.
Molecular Weight 40
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 23.1
9.9 g ,v
... .. ...... ..... .
1 ORTZ 0120 2
2 NW 0120 2
3 Ph3 0122 2I
4 DAVAO 0123 1
5 S? 0124 1
6 MW60 0124/25 1
7 8606 0125 1
8 STNP4 01212 1
9 22993 0128 1
10 DAVO 0123 1
11 STM3 0126 3
12 N5443 0129-- 2
13 TBR 0121 2
14 ES2-1 01213 2
15 JLTIH7 0124 1
16 SC548 F oxysporum other
Figure 2-4: Mitochondrial DNA haplotypes of selected Fusarium oxysporum f sp. cubense isolates.
Table 2-5: Clone-corrected measurements of gametic disequilibrium among pairs of alleles in a world-wide collection of Fusarium oxysporum f sp. cubense.
Probe / Enzyme Combinations
120 162 177 187 204 228 260 261
Hae III Dra I EcoRV EcoRV Hae IlI EcoRV EcoRV EcoRV
7 13/35 4/10 4/18 5/30 5/15 7/30 5/20 22/40
Dra 1 77.7 (24) 25.1 (5) 56.8 (10) 174.4 87.4 (8) 50.0 (24) 183.1 86.6
(20) (12) (28)
120 8/14 17/21 9/42 11/21 9/42 13/28 14/56
Hae 11I 31.33(6) 68.3(12) 122.6 142.4 121.1 87.5(18) 184.3
(30) (12) (30) (42)
162 4/6 8/12 4/6 5/12 5/8 6/16
Dra I 4.91 (2)* 37.7 5.49 (2)* 128.7 (5) 29.3 23.3
(5) (3) (7)
177 5/18 4/9 7/18 2/12 9/24
EcoRV 33.0(10) 73.09 78.1 (10) 37.7 (6) 54.6(14)
187 6/18 17/36 6/24 6/48
EcoRV 40.5 126.2 63.54 65.2 (35)
(10) (25) (15)
204 7/18 6/12 11/24
Hae I11 51.3(10) 48.2(6) 85.8(14)
228 8/24 7/48
EcoRV 53.7 (15) 57.2 (35)
All tests had at least 68 multilocus haplotypes in the comparisons and a maximum of n=72. The first row in each box represents the number of significant chi-square tests (P<0.05) between individual alleles at different RFLP loci over the total number of tests made for each pairwise comparison; the second row in each box represents the results of a chisquare test for the significance of association between all alleles at the two loci. The number in the parenthesis is the degrees of freedom for the test. All are significant at P<
0.05 and P<0.01 except those noted with *
nonrandom associations at the 1% level. Many individual allele combinations were not present in the population, but the chi-square test for these combinations was not significant; based on the overall allelic frequency, the expected number of these combinations was small. In contrast, for the most common alleles the large number of nonrandom associations was indicated by significantly larger or smaller numbers of observations when compared to the expected frequency of these combinations.
Several lines of evidence support the concept that F. oxysporum f. sp. cubense has a clonal population structure in line with criteria established by Tibayrenc et al. (1990). A unifying feature of clonally reproducing organisms is widespread geographic distribution of a few successful clones. Even though this study identified 72 multilocus haplotypes in a worldwide collection ofF. oxysporum f sp. cubense, the five most common haplotypes accounted for nearly half of the isolates. Additionally, the two most common haplotypes were found on all five continents sampled in this study, indicating the pantropical distribution of a small number of genotypes.
Further evidence of clonal reproduction is the absence of recombinant genotypes. Significant gametic disequilibrium for alleles at 34 of 36 loci tested supported nonrandom association between alleles of different loci. In addition, the strong correlation between independent genetic markers (VCG, mitochondrial and multilocus RFLP haplotype) also are indicative of a clonally reproducing organism (Tibayrenc et al., 1990).
This study confirms that in phylogenetic analysis ofF. oxysporum f sp. cubense using parsimony, VCG is a strong predictor of cladistic groupings. Further differentiation into lineages may be done based on coefficients of similarity of RFLP haplotypes. At least 17 VCGs have been described for F. oxysporum f sp. cubense (Ploetz and Correll,
1988; Ploetz, 1990; Ploetz et al., 1992; Ploetz, 1994; Ploetz et al., 1997), and representatives from each of these VCGs were included in this study. With the exception of two isolates in VCG 0123, all isolates within a VCG were in the same clade and clonal lineage. The correlation between VCG and RFLP patterns has been observed previously in F. oxysporum formae speciales, including albedinis, conglutinans, dianthi, gladioli lycopersici, melonis, pisi, raphani, and vasinfectum (Elias and Schneider, 1991; Elias et al., 1993; Kim et al., 1992; Kistler et al., 1987; Manicom et al., 1990, Manicom and Baayen, 1993; Mes et al., 1994; Namiki et al., 1994; Tantaoui et al., 1996; Whitehead et al., 1992).
Although VCGs are good indicators of genetic similarity among the individuals comprising them, they do not provide information regarding the genetic similarity to individuals in different VCGs. In fact, this study shows that isolates belonging to different VCGs could have identical or nearly identical RFLP haplotypes. With two exceptions, the entire collection of isolates consisted of only ten distinct clonal lineages (Table 2-2). Clonal lineages provide a conservative system for grouping similar isolates, and the coefficient of similarity provides a numerical value to assess genetic relationships among isolates representing different lineages. This is in contrast to VCG groupings where, for an asexually reproducing organism such as F. oxysporum, it is impossible to determine the quantitative differences among individuals in different VCGs. Thus, the use of clonal lineages rather than VCG is proposed to genetically characterize similar isolates ofF. oxysporum f sp. cubense. In most instances, VCGs can be used to predict lineage.
Many of the isolates used in this study also have been classified based on their
electrophoretic karyotype (Boehm et al., 1994) and RAPD profile (Bentley et al., 1995). Based on their electrophoretic karyotype, Boehm et al. (1994) divided 118 isolates into
two major groups. Group I contained isolates from VCGs 0 124, 0124/0125, 0125, 012 10 and 0 1214 and was characterized by high chromosome number and large relative genome size [39.9-58.9 megabase pairs (Mbp)]. Group II contained isolates from VCGs 0120, 0121, 0122, 0123, 0129 and 01213, which had correspondingly fewer chromosomes and smaller genome sizes (32.1-44.9 Mbp). Using RAPD analysis, Bentley et al. (1995) similarly found that 54 isolates, representing 11I VCGs, could be divided into two major groups. Group I contained isolates in VCGs 0120, 0121, 0122, 0126, 01210, 01211 and
0 1212 while group II contained isolates in VCGs 0 123, 0124, 0124/0125 and 0 125. Cluster analysis indicated that VCG 0 1212 was distinct from the other VCGs in group I and 0 123 was distinct from group II.
Although the results presented here corroborate most of the broad conclusions made previously, this study provides additional and sometimes disparate conclusions regarding the affinities of some of these isolates. The bootstrap 50% majority rule consensus tree showed strong support for more than two clades among isolates ofFE oxysporurn f sp. cubense. The midpoint rooted tree divides isolates into two major groups. One group is comprised of isolates in five VCGs, which represent two significant clades. Isolates in this major group are remarkably homogenous, and the branch lengths that separate isolates are minimal. In contrast, the second branch encompasses isolates representing eight lineages, I1I VCGs, a large number of significant clades, as well as isolates belonging to other formae speciales of F oxysporum. Isolates representing the second group had more variable branch lengths compared to isolates in the first major branch.
With the exception of VCGs 0 122, 0123, 0126, 01210, 01212, and 0 1214, the relationship of isolates in 10 of the 17 VCGs correspond to those defined by previous
studies (Bentley et al., 1995; Boehm et al., 1994). Unlike previous investigations, isolates in VCGs 0122, 0126, 01212 and 01214 each formed individual clades with bootstrap values greater than 70%. Additionally, because this study used numerous, independent clones, had a large sample size, and provides bootstrap support for the clades, it gives a greater resolution of the genetic relationships among isolates ofF. oxysporum f. sp. cubense than do previous studies.
A number of the clonal lineages described here are phylogenetically distinct. Some isolates ofF. oxysporum f. sp. cubense are as genetically dissimilar to one another as they are to other formae speciales ofF. oxysporum (niveum and lycopersici). One interpretation of these results is that isolates belonging to the dissimilar groups acquired their ability to be pathogenic on bananas independently.
FOC V contains isolates from the Misuku Hills in Malawi, a relatively small area (approximately 400 square kilometers) on the country's northern border with Tanzania (Ploetz et al., 1992). All isolates in FOC V are in VCG 01214, and this is one of the few VCGs which has not been found in Southeast Asia, the center of origin of banana. Due to numerous lineage-specific alleles, FOC V is distant from all other lineages. One hypothesis is that this lineage ofF oxysporum f sp. cubense may have evolved independently of other members of the taxon in East Africa.
Alternatively, isolates within FOC V could have arisen by a founder effect. Bananas probably first arrived at the island Madagascar in the later half of the first millennium A.D. and from there moved to the coastal and then interior regions of the African continent (Ploetz et al., 1992). Diverse genotypes of banana are now found in East Africa, many of which are found nowhere else. It is possible that the pathogen was moved from Southeast Asia on the bananas introduced to Africa, and, as a result of
mutation and selection or through adaptation to endemic bananas, isolates in VCG 01214 may have diverged from their Asian progenitors.
FOC VII and X contain isolates in VCG 0123. The RFLP haplotype of isolates
belonging to these groups carry an assortment of alleles from the two major lineages (FOC I and II) as well as a number of lineage-specific alleles. Additionally, isolates comprising this clade are quite heterogenous; the five isolates comprising FOC VII belong to two significant sister groups. Also, all of the isolates in VCG 0123 fall neither into the same lineage nor in a single clade. Based on this information, this group may provide evidence of an ancient genetic exchange between individuals in FOC I and II. Alternatively, it may represent an ancestral group possessing primitive character states found in FOC I and FOC II.
In conclusion, isolates ofF. oxysporum f. sp. cubense represent a genetically diverse group of organisms, many of which are distantly related. Previous studies on other formae speciales indicate that many are genetically diverse. However, this study is the first based pm RFLP data of nuclear DNA to present evidence that a forma specialis of F oxysporum may be polyphyletic. The implied independent origin of pathogenicity to banana in some of the lineages has practical implications for work on this disease. Much effort is devoted towards developing cultivars of banana which are resistant to Fusarium wilt. Clearly, new hybrids should be screened against isolates representing the two most common lineages of the pathogen (FOC I and II). Ideally, breeding programs could screen new hybrids against isolates from each clonal lineage to increase the probability of developing cultivars that resist genetically distinct populations of the pathogen.
GENETIC VARIATION IN TWO HONDURAN FIELD POPULATIONS OF FUSARIUM OXYSPORUM F. SP. CUBENSE Introduction
Fusariumu oxysporum Schlechtend, Fr. forma specialis cubense (E.F. Sm.) W.C.
Synder & H. N. Hans., is the causal agent of Fusarium wilt of banana, a disease which has been responsible for significant yield losses in banana. This disease had a devastating effect on the export banana trade in Central America during the early part of this century (Stover, 1962).
Bananas were introduced to Honduras well over 200 years ago and commercial trade began about 1876. Until the late 1950s, the race 1 susceptible banana cultivar Gros Michel was the most widely planted commercial cultivar in Honduras and elsewhere (Stover, 1962). However, disease epidemics in many commercial banana plantations were exacerbated by an increased prevalence of race I isolates due to the perennial monoculture production of this susceptible, clonally derived cultivar. Eventually, Cavendish cultivars, which are resistant to race I and 2, were used to replace Gros Michel for export production.
Three races of the pathogen that are pathogenic on banana have been reported in prior work. Only two races of the pathogen have been identified in Honduras (Stover, personal communication), and these isolates belong to only three vegetative compatibility groups (VCGs) (0120, 0124 and 0126) (Ploetz, 1990). Race I isolates occur in all three
of these VCGs, while race 2 isolates appear to be limited to VCG 0124. In a number of subtropical countries (Australia, South Africa, Canary Islands and Taiwan), many commercial plantations of Cavendish cultivars succumbed to Fusarium wilt shortly after they were planted (Ploetz, 1990). More recently, isolates in VCGs 01213 and 01216 have caused disease on Cavendish cultivars in the tropics. Isolates capable of causing disease on Cavendish cultivars were identified by Su et al. (1977) as race 4, a new race of the pathogen. Race 4 isolates also are pathogenic on all cultivars susceptible to race 1 and race 2 but presently have a more limited geographical distribution than races 1 and 2. There have been no reports of race 4 in Central America.
Fusarium wilt has been a problem in most major banana-producing regions of the world and the occurrence of the pathogen in areas where bananas are not indigenous is most likely a consequence of human dissemination of infested rhizomes. Bananas are rhizomatous, perennial plants that are grown in subtropical and tropical climates. Cultivated bananas are commonly sterile diploid, triploid, or tetraploid clones derived from Musa acuminata or interspecific hybridization between M. acuminata and M balbisiana (Simmonds, 1962). Since most edible bananas do not produce viable seed, they must be asexually propagated, typically using rhizomes or sucker plants. As such, they are usually also accompanied by soil and associated microorganisms; it is through the movement of infested plant material that the pathogen has been introduced to nonendemic regions. For example, isolates that have been introduced to the Americas on infested plant material comprise only a limited number of VCGs and clonal lineages compared to those found in southeast Asia, the center of origin of M. acurninata (Ploetz and Correll, 1988; Chapter 2).
A study of a world collection of F oxysporum f. sp. cubense isolates representing 17 vegetative compatibilty groups (VCGs) revealed that this pathogen is comprised of at least 10 clonal lineages (Chapter 2). The greatest diversity of VCGs and lineages are present in Malesia (Ploetz, 1994; Chapter 2), which is also the region of origin of the host (Simmonds, 1962).
Although the level of genotypic diversity that was represented in a worldwide
collection of F. oxyspornrn f sp. cubense isolates has been described previously (Bentley et al, 1995; Chapter 2), the amount of diversity found in individual field populations of the pathogen is unknown. Similarly, factors that may affect the levels of diversity within a field have not been explored for F. oxysporuni f. sp. cubense. McDonald et al. (1995) discussed several factors which may cause pathogen populations to undergo significant changes in genetic structure over time. For a perennial, tropical crop such as banana, cropping practices such as rotation, field burning, and fungicide application may affect the pathogen population. However, the use of host disease resistant genotypes also likely plays a major role. Similarly, the reproductive strategy and mutational rate of the pathogen, as well as the frequency of immigration of new genotypes, may also have a profound effect on the genetic composition of a field population.
The objectives of this study were to: i) determine the amount of genotypic
diversity in field populations of the pathogen in Honduras; and ii) determine whether specific genotypes of the pathogen preferentially infect specific host genotypes.
Materials and Methods
Description of Sampling Procedure
Sampling ofF. oxysporum f sp. cubense isolates was conducted in two adjacent disease screening fields (Fields 1 and 2) at the banana breeding station operated by the Fundaci6n Hondurefia de Investigaci6n Agricola (FHIA) in La Lima, Honduras. Historically, Field 1 had been planted to cultivar Highgate, which is susceptible to race 1, and had a high incidence of plants with Fusarium wilt based on visual assessment of symptoms. Therefore, it was an optimal location to evaluate the resistance to race 1 in seven of FHIA's tetraploid hybrids (Rivera, pers. comm.). Prior to planting the trial, inoculum levels were artificially increased in the field using a number of procedures. First, freshly infected pseudostem and rhizome tissue from Highgate plants infected with F. oxysporum f sp. cubense was uniformly spread throughout the field. Subsequently, the entire field was planted to the race 1-susceptible cultivar Maquefio. Prior to planting the Maquehio rhizomes, approximately 250-gram (g) chunks of symptomatic pseudostem and rhizome tissue from Highgate were placed in the bottom of each hole. After a year, 94% of the resulting plants exhibited internal symptoms of Fusarium wilt (vascular discoloration of rhizomes and pseudostems). Prior to planting the disease screening trial, infected Maquefio plants were incorporated into the soil. Additionally, prior to planting each experimental hybrid, 250 g of symptomatic Maquefio tissues were placed in each planting hole.
The experiment was arranged in a randomized complete block design with nine treatments (cultivars) and seven replications (Table 3-1). The banana plants used in the experiments were derived from tissue culture. The treatments were planted in the same
Table 3-1. Genotypes of banana planted in Field 1
FHIA 1: AAB x AA = AAAB
Dwarf Prata x 3142 (Pisang Jari Buaya derivative)
FHIA 2: AAA x AA = AAAA
Williams*x 3397 (3142 (AA) x 3217 (AA); both are Pisang Jari Buaya derivatives)
FHIA 3: ABB x BB = ABBB; ABBB x AA = AABB; AAB or ABB x AA = AAA,
AAB, or ABB
Cardaba x BB = ABBB x 2741 (AA) = 3386 [selected triploid x 3320 (AA)]
FHIA 6: AAB X AA = AAAB
Maquefio x 3437 (Black sigatoka resistant Mura acuminata subsp. burimanica)
FHIA 15: AAB x AA = AAAB
Same parentage as FH1A 6, but this selection is faster to ratoon
FHIA 17: AAA x AA = AAAA
Highgate x 3362 (3142 x 3217)
FHIA 23: AAA x AA = AAAA
Same parentage as FH1A 17 but different selection
Highgate: AAA (race 1 susceptible control (race 2 resistant)) Williams: AAA ( race 1 and race 2 resistant control) Bluggoe: ABB (race 2 susceptible control (race 1 resistant))
Note: With the exception of FHIA 3, all FHIA hybrids are either interspecific or intraspecific tetraploids obtained by performing crosses on Musa acumninata (AA) and M. balbisiana (BB) parental lines. Highgate, Williams and Bluggoe are banana cultivars used as host differentials to differentiate among the three races of Fusarium oxysporum f sp. cubense.
* It is questionable as to whether or not Williams was actually the parent of this cross, due to the cultivar Williams extreme sterility.
rows in which the cultivar Maquei'io had been planted. Each plot contained three plants spaced at 2.5 m, and the entire experimental area encompassed approximately 6500 in2. The field consisted of seven rows; three of these contained the seven replications of the experiment planted in a serpentine fashion throughout the plot. Experimental rows were separated from one another by a single row of the race 2-susceptible cultivar Bluggoe (Table 3-2).
The first samples were collected in February 1995 using a nondestructive method, which consisted of removing the first few outer layers of pseudostemn tissue from symptomatic and asymptomatic plants (Ploetz, 1992). A second set of samples were collected from the same planting in September 1995 using a destructive sampling method, which consisted of excising inner tissue of both the rhizome and pseudostem. When collecting symptomatic tissue, attempts were made to obtain samples displaying necrotic vascular strands. Since sampling occurred during two periods, some plants were sampled twice. Samples were obtained from every plant showing visible disease symptoms in all treatments and also from the symptomatic race 2 susceptible Bluggoe plants in the adjacent rows. Approximately one of every 12 asymptomatic plants in the entire field also were sampled to determine incidence of infection on these plants.
Field 2 was planted to the cultivar Bluggoe in an effort to increase inoculum levels for a future field trial to evaluate hybrids for resistance to race 2. The field was approximately 3500 m' and consisted of eight rows (Table 3-3). Seventy-six percent of the plants showed visible symptoms of Fusariumn wilt during an initial assessment of the field. Symptomatic plants were fairly evenly distributed throughout the field, and, based on this distribution, a systematic sampling strategy was followed in which one of every three plants was sampled following a serpentine pattern through the plot. Samples were
Table 3-2. Experimental design of banana plants in rows of Field 1.
2 3 4 5 6 7
B F15 B F15 B F2 B
B F15 B F15 B F2 B
B F15 B F15 B F2 B
B F6 B F3 B F I B
B F6 B F3 B F I B
B F6 B F3 B F I B
B F23 B W B F6 B
B F23 B W B F6 B
B F23 B W B F6 B
B F17 B H B F6 B
B F17 B H B F6 B
B F17 B H B F6 B
B F I B F17 B F17 B
B F I B F17 B F17 B
B F1 B F17 B F17 B
B H B F23 B F2 B
B H B F23 B F2 B
B Hz B F23 B F2 B
B W B F6 B F15 B
B W B F6 B F15 B
B W B F6 B F15 B
B F2 B Hz B H B
B F2 B Hz B H B
B F2 B H B H B
B F3 B F3 B F23 B
B F3 B F3 B F23 B
B F3 B F3 B F23 B
B W B F17 B F3 B
B W B F17 B F3 B
B W B F17 B F3 B
B F17 B F23 B W B
B F17 B F23 B W B
B F17 B F23 B W B
B F15 B F15 B F1 B
B F15 B F15 B F I B
B F15 B F15 B F1 B
B F6 B W B F2 B
B F6 B W B F2 B
B F6 B W B F2 B
B F2 B F2 B W B
B F2 B F2 B W B
1 2 3 4 5 6 7
B F2 B F2 B W B
B Fl B Fl B F6 B
B Fl B Fl B F6 B
B F1 B Fl B F6 B
B F3 B F2 B H B
B F3 B F2 B H B
B F3 B F2 B H B
B Hz B H B F17 B
B Hz B H B F17 B
B H B H B F17 B
B F23 B F1 B F1 B
B F23 Bz Fl B F1 B
B F23 B Fl B Fl B
B F23 B F6 B F15 B
B F23 B F6z B F15 B
B F23 B F6 B F15z B
B F17 B F3 B F3 B
B F17 B F3 B F3 B
B F17 B F3 B F3 B
B W B F15 B F23 B
B W B FI5 B F23 B
B W B F15 B F23 B
Y' Each letter or letter/numeral combination represents a single plant; bold type indicates plants that were sampled for the presence of Fusarium oxysporum f. sp. cubense; B= cultivar Bluggoe, H= cultivar Highgate, W= cultivar Williams, F l =FHIA 1, F2=FHIA 2 F3=FHIA 3, F6=FHIA 6, F 15=FHIA 15, F 17=FHIA 17, F23=FHIA 23
indicates multiple samples were obtained from these plants
Table 3-3. Field design of banana plants in rows of the race 2 disease screening plot.
Plant Number Row Number'
1 2 3 4 5 6 7 8
D D D D D D D D
2 D D D D D D D H
3 H D D D D D D D
4 D D D D D D D D
5 D D D D D D D D
6 H D D D D D D D
7 H D D D D D D D
8 H D D D D D D D
9 D D D D D D H
10 H D D D D D D H
I I H D H D D D H H
12 H D D D H H D D
13 D H D H D H D D
14 H H D D D H D H
15 H H D D D D D D
16 D H D D D D H H
17 D H D D D D D D
18 D H D D H D D D
19 H H D D D D D D
20 H D D H D D D D
21 H D D D D D D D
22 H D D D D D D D
23 H D D H D D D D
24 H D D D D D D D
25 H D D D D H D D
26 H D H D D D D D
27 H D D D H D D H
28 H H D D D D D D
29 D D D D D D D D
30 D H D D D D D D
31 H H D D D D D D
32 D H D H D H D H
33 D H D D D D D D
34 H D D D D H H H
35 H H D D D D H D
36 H D D D D D D H
37 H D D H H H D D
38 D D D D D D D H
39 H H D D H H D D
'The entire field planted to the race 2 susceptible cultivar Bluggoe. Each letter represents an individual plant. H = asymptomatic plants and D= symptomatic plants based on visual assessment-, samples were obtained from plants in bold type. Total plants = 312; Total D = 238 (76%); Total H = 74 (24%).
collected in March 1995 and consisted of destructive sampling of inner rhizome and pseudostem tissues. Sampling included both symptomatic and nonsymptomatic tissue. Fungal Isolations
Tissue samples were allowed to air dry for 3 to7 days prior to transport to Florida. In Florida, each sample was cut into approximately eight, 1 -cm pieces, which each contained at least one discolored and presumably infected vascular element. The tissue pieces were surface disinfested by placing them in 70% ethanol for 10 seconds, followed by a 2-min soak in 10% bleach, and a final rinse for at least 1 minute in sterile water. Pieces were air dried in a laminar flow hood and then placed into two petri dishes to provide four pieces of tissue in each dish. One of two media were poured over the tissue pieces. Medium 1 consisted of amended water agar, streptomycin sulfate (10 mg/L; Sigma Chemical Company, St. Louis, Mo.), rifampicin (12 mg/L; Sigma) and 10 A1 of danitol/L of medium (Chevron Corporation, San Francisco, CA.). Medium 2 consisted of amended potato dextrose agar (PDA) containing streptomycin (10 mg/L), rifampicin (12 mg/L), danitol 10 gl/L/L, and terigitol NP-10 (100 /A/L; Sigma). Medium 1 allowed for quick growth of Fusarium spp. while suppressing the growth of other microorganisms. However, identification ofF. oxysporum isolates was difficult on this medium because of poor mycelial growth and spore formation. Medium 2 allowed for better identification of F. oxysporum colonies but the additional nutrients were also more conducive to the growth of other microorganisms. Plates were incubated at room temperature, and within 3 days mycelial growth was observed on tissue pieces. Putative colonies of Fusariunm spp. were selected based on visual growth characteristics and the formation of microconidia or macroconidia. Agar sections (1 mm2) were excised from these colonies and placed on the
reciprocal medium (from medium 1 to 2 and vice versa). Colonies that were confirmed to be F. oxysporum based on mycelium pigmentation and growth characteristics, microconidia or macroconidia formation, phialide anatomy, or chlamydospore formation were streaked onto plates containing PDA. Within 14 hours, four single, germinated spores were picked off using a compound microscope and placed with a sterile needle on plates containing PDA; at least one single spore isolate was collected for each sample. DNA Extraction, Southern Blotting and Hybridization
Procedures for DNA extraction have been described previously (Chapter 2). Approximately 10 gg of DNA were digested with at least 10 units of either HaeIII or HindIII restriction enzymes (Bethesda Research Laboratories, Gaithersburg, MD) and incubated for at least 3 hours at 37C. Restriction fragments were separated by electrophoresis in 0.7% agarose for HindIII-digested DNA or 1.5% agarose for HaeIIIdigested DNA in TBE buffer at pH 7.0. Gels were run at 30 volts for approximately 24 hours. Bacteriophage lambda DNA digested with either PstI or HindIII was included on each gel and used to calculate the molecular mass of restriction fragments obtained from F. oxysporum f sp. cubense DNA. The DNA of each isolate was visualized by incorporating 100 ng of ethidium bromide/ml of solution into the agarose gels and then illuminating the gels using UV irradiation. The DNA was transferred to Nytran membranes (Schleicher & Schuell, Inc., Keene, NH) using the capillary transfer method (Sambrook et al., 1989). The DNA transfer proceeded for at least 12 hours, and the DNA was immobilized by UV crosslinking (UV 254-nm crosslinker, model FB UVXL 1000, Fisher Scientific, Pittsburgh, PA).
Mitochondrial DNA (mtDNA) was isolated from Honduran isolate 3 S1 (VCG 0120) following a procedure described previously (Kistler and Leong, 1986). Total mtDNA was labeled using random hexamer primers to incorporate fluorescein- I 2-dUTP. DNA labeling; hybridization and detection followed the procedures provided by the manufacturer (Dupont NEN "Renaissance," E. I. du Pont de Nemours & Co. Inc., Boston, MA). Prehybridization, hybridization and washing were performed at 65oC using a Hybaid hybridization oven (Dot Scientific Inc., Flint, MI). Membranes were placed between acetate sheets and exposed to x-ray film for 15 to 30 minutes.
Preliminary work on a world-wide collection ofF. oxysporum f sp. cubense
isolates revealed that mitochondrial haplotypes were correlated with VCG and isolates tested within a VCG had identical mitochondrial haplotypes (Kistler and Momol, 1990, Chapter 2). Isolates belonging to the three VCGs found in Honduras (0120, 0124 and 0126) could be differentiated based on their unique mitochondrial profiles, obtained by probing blots of total HaellI-digested DNA with mtDNA isolated from Honduran isolate 3S1. Using this method, a mitochondrial haplotype was assigned to each isolate. Honduran isolates which had mitochondrial profiles identical to isolates in VCG 0120 were designated mitochondrial haplotype (mthap) 0120, those which had mitochondrial profiles identical to isolates in VCG 0124 or VCG 0126 were designated mthap 0124 and mthap 0126, respectively.
Similarly, F. oxysporum f sp. cubense isolates were differentiated based on their DNA fingerprint by digesting total DNA with the restriction enzyme HindIII and probing with clone pEYI0 (Kistler et al, 1991). The clone pEY10 hybridizes to a nuclear dispersed, middle repetitive DNA sequence (Benny and Kistler, unpublished) and contains
an open reading frame encoding a polypeptide with 51% amino acid identity to the transposase of the Magnaporthe grisea transposon Pot3 (Farman et. al., 1996; Pettway and Kistler, unpublished). All previously tested isolates from a world-wide collection of F. oxysporunm f sp. cubense had unique DNA fingerprints using pEY10 (Kistler et al., 1991). The DNA labeling, hybridization and detection procedures followed those described above.
DNA fingerprints were scored by visual assessment. Only bands that were of strong intensity and consistently present were scored. Similar patterns were assigned a number based on their frequency of occurrence in the field; the most common pattern was assigned the number 1. Two chi-square analyses were performed. In the first test, observed percent infected and noninfected of each FHIA hybrid and the two controls, Highgate and Williams, were obtained and compared against the null expectation of no differences among cultivars. The expected number of infected plants for this test was 6. 1, which was obtained by summing the number of isolates obtained from the FHIA hybrids and the cultivars Highgate and Williams and dividing by 189 the total number of plants in the experiment, excluding the cultivar Bluggoe. This number was then divided by 9, which represents the total number of cultivars in the experiment. The expected number of noninfected plants was 14.8, and this number was obtained as described above but based on the total number on noninfected plants. The chi-square analysis had 8 degrees of freedom. The second test determined whether the pathogen genotypes obtained from infected plants were randomly distributed among the different host genotypes. In this test, the observed frequency of each pathogen genotype in the field and the host genotype from
which they were obtained were compared against the null expectation of no differences among the host genotypes. The expected frequency for each genotype was estimated by dividing the total number of isolates obtained from each cultivar by three. For example, a total of 16 F. oxysporurn f sp. cubense isolates were obtained from the cultivar Highgate, and the expected number of isolates obtained for each mtDNA haplotype was 5.3. The chi-square analysis had 16 degrees of freedom.
Field I was sampled twice and isolates were collected from both symptomatic and asymptomatic tissue during each period. Duplicate isolates were collected from symptomatic tissue of treatments 104-2, 106-3, 208-2, 306-2, 308-2, 408-1, and 408.3. Additionally, duplicate samples were obtained from one Bluggoe plant. Field 2 was sampled once and duplicate isolates were analyzed for four of the Bluggoe plants. Additionally, many of the isolates from both fields were duplicated on different gels to validate the results.
Ninety-seven isolates were collected from Field 1 (the race 1 -screening field). Of these 97 isolates, 12 isolates were obtained from plants that were symptomatic during both sampling periods and represent multiple samples taken from the same plant. A mtDNA haplotype was assigned to each isolate by probing blots of HaeJII-digested DNA from each isolate with labeled mtDNA isolated from Honduran isolate 3S1I (Figure 3 -1). Isolates collected from the field were assigned a mtDNA haplotype as described in the materials and methods. The incidence and distribution of pathogen mtDNA haplotypes varied with host genotype (Table 3-4). The Highgate and Cavendish cultivars; used as susceptible and resistant experimental controls had the highest and lowest incidence of
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Molecular
4 .5 ...... ..... .
Isolate/Host Cultivar mtDNA Haplotype
2 1-5-4Bluggoe 0124
3 1-3-63 Bluggoe 0124
4 1-3-62 Bluggoe 0124
5 1-1-52 Bluggoe 0124
6 27-5-11 Bluggoe 0124
7 2-8-26 Bluggoe 0124 ... .. ..........
8 1-3-51 Bluggoe Unknown type
9 1-3-3 Bluggoe 0124
10 106-2 Highgate 0120
11 601-2 FHIA 6 0126
12 308-1 Highgate 0126
............::: :: :::::::::::::
....... ..... i" : iii~ i i i i :: !i ........ .:. .. .....ii:!i!!i:!ii:i:::
13 707-3 FHIA 15 0126
14 502-1 FHIA 17 0126
..........~~~~~~ ~~~~....... .::: .:: .... ::i :ii::::i:::~ii!::!,i~i
Isolate/15 304-3 FHIA 15 0126ost Cultivar tDNA Haplotype
1 y- 1-20 Bluggoe Unknown type
16 1-508-1 FHIA 2Bluggoe 0126
17 707-3-63 BluggoeFHIA 15 0124
18 408-3-62 Bluggote 0126
5 1-1-52 Bluggoe 0124
6 2Y-5-11 Bluggoe 0124
Figure 3-1: mtDNA haplotype of selected Honduran field isolates.
7 2-8-26 Bluggoe 0124
8Isolates with 1-designation were1-3-51collected from the cultivar Bluggoe Unknowin Field 1.type
9Isolates with 2-designation were collected from the cultivar1-3-3 Bluggoe 0124in Field 1.
10FHIA hybrids designated according to their lot number in Field 1.20
11 601-2 FHIA 6 0126
12 308-1 Highgate 0126
13 707-3 FHIA 15 0126
14 502-1 FHIA 17 0126
15 304-3 FHIA 15 0126
"i6 108-1 FHIA 2 0126
17 707-3 FHIA 15 0124
18 408-3 Highgate 0126
19 409-3 FHIA 6 0126
Figure 3 1: mtDNA haplotype of selected Honduran field isolates. Isolates with 1 -designation were collected from the cultivar Bluggoe in Field 1. Y Isolates with 2-designation were collected from the cultivar Bluggoe in Field 1. SFHIA hybrids designated according to'their plot number in Field 1.
infection, respectively. No isolates were obtained from the Cavendish cultivar Williams, which is resistant to both race I and race 2 of the pathogen. In contrast, 17 isolates were collected from the Highgate plants. Fourteen of the 17 isolates had 0126 mtDNA haplotypes, whereas one each had 0120 and 0124 mtDNA haplotypes. One isolate collected from the cultivar Highgate had a previously uncharacterized mtDNA haplotype for F. oxysporum f sp. cubense.
Percent infection for the seven hybrid bananas ranged from less than 1% for FHIA
1 to 57% for FHIA 15 (Table 3-4). The observed percent infection of each cultivar was obtained and compared against the null expectation of no differences among cultivars with a series of chi-square tests. The differences in infection on the cultivars were highly significant (p < .001) indicating that cultivars are not equally likely to be infected. This suggests differences in host resistance genes. The distribution of pathogen genotypes varied on the different host genotypes. The overwhelming majority of isolates collected from the hybrids had 0126 mtDNA haplotypes. Isolates with 0120 and 0124 mtDNA haplotypes were recovered from some of the hybrids but at a lower frequency compared to isolates with 0126 mtDNA haplotypes. In contrast, 96% of the isolates collected in the same field but obtained from the race 2 susceptible cultivar Bluggoe, had the 0124 mtDNA haplotype (Table 3-4).
The clone pEY1 0 was used to generate DNA fingerprints of the isolates (Figure 3-2). Isolates which had identical DNA fingerprints were presumed to represent the same clone. Clone frequency and distribution on host genotypes were examined in the two field populations. The nuclear DNA fingerprints of the isolates with 0126 mtDNA haplotypes showed little variation. Only two patterns, which were 99% similar, were
Table 3-4. Isolates of Fusarium oxysporum f sp. cubense collected in Field 1 and their incidence on different host genotypes
Number of Isolates with mtDNA Haplotypes
Cultivar or Hybrid 0120 0124 0126 Total Infections
Highgate (race 1
susceptible) 1 1 14 16 76
Williams (race 1 &
2 resistant) 0 0 0 0 0
Bluggoe (race 2
susceptible) 0 26 1 27 11
FHIA 1 0 1 0 1 <1
FHIA 2 1 1 7 9 43
FHIA 3 0 0 2 2 14
FHIA 6 0 0 8 8 38
FHIA 15 0 1 11 12 57
FHIA 17 0 0 3 3 14
FHIA 23 0 1 3 4 19
XFor plants where more than one isolate was collected, only one is represented if the mtDNA haplotypes were identical.
Y Total percentage of plants in Field 1 from which isolates were recovered Chi-square analysis: p < .001; df= 8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 64
W eight .... ......
4 .7 .
*: i-.', :' i
4.5 K .1
Isolate/Host Cultivar mtDNA Haplotype
1 2x-3-18 Bluggoe 0124
2 102-3 FHIA 6 0126
3 2-7-37 Bluggoe 0124
4 707-2 FHIA 15 0126
5 503-2 Highgate 0126
6 1-5-66 Bluggoe 0124
7 503-3 Highgate 0124
8 707-3 FHIA 15 0124
9 309-1 FHIA 2 0124
10 208-1 Highgate 0126
11 1-1-59 Bluggoe 0124
12 304-1 FHIA 15 0126
13 208-2 Highgate 0126
14 2-8-16 Bluggoe 0124
15 1-1-29 Bluggoe 0124
16 1-5-4 Bluggoe 0124
Figure 3-2: HindlII digested DNA of selected Honduran field isolates that were probed with clone pEY10 to generate DNA fingerprints. x Isolates with 2-designation were collected from the cultivar Bluggoe in Field 2. Y FHIA hybrids designated according to their plot number in Field 1. Isolates with 1-designation were collected from the cultivar Bluggoe in Field 1.
observed, and they represented the same clonal lineage. The two isolates with 0120 mtDNA haplotypes, which infected the cultivars Highgate and FHIA 2, had identical nuclear DNA fingerprints that were distinct from isolates representing other mtDNA haplotypes. With the exception of a single isolate, all of the isolates with 0124 mtDNA haplotypes, which infected the cultivar Bluggoe and FHIA hybrids 1,2,15 and 23, had identical nuclear DNA fingerprints that were distinct from isolates representing other mtDNA haplotypes. An isolate collected from the cultivar Highgate with a 0124 mtDNA haplotype had a unique nuclear DNA fingerprint.
In contrast to Field 1, which contained ten different genotypes of banana, Field 2 was entirely planted to the cultivar Bluggoe. Fifty-nine isolates ofF. oxysporum f sp. cubense were collected in this field, and all of them shared the same 0124 mtDNA haplotype (Table 3-3). Similarly, all tested isolates had a nuclear DNA fingerprint identical to those isolates in Field 1 which had a 0124 mtDNA haplotype and were recovered from Bluggoe and FHIA hybrids 1, 2, 15 and 23.
The observed frequency of each pathogen genotype in the entire field and the host genotype from which they were obtained were compared against the null expectation of no differences (no host selection) with a series of chi-square tests (Figure 3-3). The differences in clone frequencies on the various host genotypes were highly significant. Clones of the pathogen did not randomly infect banana plants representing the various host genotypes. Instead clones were preferentially distributed on specific host genotypes.
Clonal diversity was low in the two field populations of F. oxysporum f sp.
cubense; only five clones were recovered from the 10 genotypes of banana. Although the
S Chi-square analysis: p < 001; df=16
15 a mthaplotype 0124
gmthaplotype 0126 E 10
)Observed frequency of m thaplotype 0r-20 = .02 in Field 1.
Observed frequency of mt haplotype 0124 = .38 in Field 1.
Observed frequency of mt haplotype 0126 = .60 in Field 1.
Figure 3-3: Incidence and distribution of mtDNA haplotypes ofFusarium oxysporum f.
s. cubense on host notess in Field 1.
-~ T 44
-= L LL LL LL 7 T I LL LL LL F
H ost c ultiva r
Observed frequency of m thaplotype 0120 = .02 in Field 1.
Observed frequency of mt haplotype 0 124 = .3 8 in Field 1.
Observed frequency of mt haplotype 0 126 = .60 in Field 1.
Figure 3 -3:i Incidence and distribution of mtDNA haplotypes of Fusariuni oxysporum f. sp. cubense on host genotypes in Field 1.
number of clonal genotypes in these two populations was low, there appeared to be a relationship between the level of genotypic diversity in the pathogen and host populations. By increasing the number of genotypes of banana in a field, more genotypes of the pathogen were detected. In fact, the number of different clones of the pathogen which were recovered in each field was proportional to the number of host genotypes planted in the field. For example, all of the isolates recovered from host tissues in Field 2, which was planted to a single cultivar, belonged to a single mtDNA haplotype and represented a single clone. In contrast, isolates recovered from host tissues in Field 1, which was planted to 10 different genotypes of banana, comprised three mtDNA haplotypes and five clones. Each mtDNA haplotype was represented by only one or two clones.
These findings were consistent with field studies which compared the genetic diversity of nonpathogenic F. oxysporum in native soils to F. oxysporum f sp. melonis isolates. In contrast to populations of native, nonpathogenic Fusarium oxysporum, pathogen populations ofF. oxysporum f. sp. melonis were less diverse based on their VCGs and mtDNA haplotypes (Gordon and Jacobson, 1990a and b; Gordon and Okamoto, 1992). In contrast, field populations of other asexually reproducing fungi, including Stagonospora nodorum and Magnaporthe grisea as well as sexually reproducing fungi including Sclerotinia sclerotiorum, Mycosphaerella graninicola and some populations of Phytophthora infestans, contained a larger number of genotypes, indicating higher levels of clonal diversity (Goodwin et al., 1992; Kohli et al.,1995; Kohn, 1995; Levy et al., 1993; McDonald et al., 1994; McDonald et al., 1995).
The recovery of different clones ofF. oxysporum f sp. cubense in these two fields was dependent on the genotype of the host. Chi-square analyses were performed to test
whether isolates of the pathogen were randomly recovered from the different banana genotypes; these showed a significant association between a mtDNA haplotype and its incidence on specific host genotypes. In fact, isolates with certain mtDNA haplotypes were more frequently and sometimes almost exclusively recovered on certain host genotypes (Figure 3-3). For example, 99% of the isolates recovered from the race 2-susceptible cultivar Bluggoe in Fields 1 and 2 had a 0124 mtDNA haplotype and represented a single clone based on their identical DNA fingerprints. In contrast, 87% of the isolates recovered from Highgate had a 0126 mtDNA haplotype, representing two nearly identical clones.
The occurrence of a specific genotype of the pathogen associated with virulence on a specific host genotype is likely the consequence of linkage disequilibrium found in clonally reproducing organisms (Tibayrenc et al., 1991). Since sexual reproduction does not occur in these organisms, the entire genome is effectively linked (Anderson and Kohn, 1995). Therefore, alleles at different loci, such as those for mt haplotypes, VCG, virulence, are repeatably associated. Additionally, in these field populations it appears that the host genotype drives selection for particular clones of F. oxysporum f sp. cubense. In fields planted to a susceptible cultivar, genotypes of the pathogen which are virulent on the host genotype may have greater fitness. They may, for example, produce more propagules. The pattern of introducing strains ofF. oxysporum f sp. nielonis to agricultural soils and the correspondingly intense selection pressure on the susceptible crop, has been described previously (Gordon and Okamoto, 1992). Such selection pressure undoubtably has an even more significant effect on the pathogen populations in perennial, monocultural production systems, such as banana plantations, since susceptible
host plants are exposed to the pathogen for even longer times. In these systems, one would expect a large increase in the number of propagules of virulent pathogen genotypes when susceptible cultivars are grown.
In contrast to certain plant pathogens, such as rusts, where new pathotypes appear annually and breeders are continually challenging the pathogens by introducing cultivars with new combinations of resistant genes, resistance genes to Fusarium wilt in Cavendish cultivars have provided stable, relatively long-term protection. In fact, Cavendish cultivars have been planted in Central America for at least 40 years and no substantial reports of Fusarium wilt on these cultivars have been made. However, in a limited number of subtropical and tropical countries, race 4 isolates of the pathogen, which are capable of causing disease on Cavendish cultivars, do occur. The asexual nature of this fungus may preclude rapid change in the pathogen population. However, lack of genetic change due to recombination in populations of the pathogen actually may be detrimental in commercial banana plantations where monoculture production of clonally propagated banana cultivars are typical. In these cropping systems, when a virulent pathotype evolves, lack of recombination in the pathogen can serve to stably perpetuate the rapid selection and growth of the most virulent pathotype.
This study, in combination with information from Chapter 2, indicates that race 1 isolates belong to more than one distinct lineage. The study of a world collection ofF. oxysporum f. sp. cubense found that isolates in VCG 0124 belonged to one clonal lineage, while isolates in VCG 0120 and 0126 belonged to a second lineage. Isolates in these two lineages were as genetically dissimilar to one another as they were to isolates ofF. oxysporun f. sp. niveuni and f. sp. lycopersici isolates. In the present study, the DNA
fingerprint of isolates representing these two lineages were distinct, further substantiating their genetic disimilarity. Race I isolates are known to occur in all three of the VCGs found in Honduras (Ploetz and Correll, 1988) and the present results showing mitochondrial haplotypes corresponding to these three VCGs, recovered from diseased Highgate plants (Table 3-4), support this observation. The definition of race assumes that all race I isolates share identical virulence characteristics corresponding to susceptibility in the cultivar Gros Michel (and Highgate) and resistance in the cultivar Cavendish. However, in this study, the frequency with which genotypes representing these two lineages were isolated differed in the race 1 -susceptible cultivar Highgate and the FHIA hybrids. Isolates in the lineage represented by mthap 0126 preferentially infected the cultivar Highgate and most of the FHIA hybrids. In contrast, isolates with mthap 0124, which represent a second phylogenetic lineage, were recovered only at a low frequency from these same host genotypes. Because isolates with mthap 0124 were abundantly present in the field and were frequently recovered from Bluggoe plants, the basis for the race 1 phenotype may be different in these two lineages. This may imply genetic differences in the virulence determinants of these two nominally race I lineages. This of course assumes that isolates with identical mthap and DNA fingerprints in field I have the identical pathogenic phenotype.
Similar situations have been observed in other fungi. In a survey of Magiiaporthe grisea pathotypes from the United States, Levy et al. (1991) found that isolates of a single pathotype had significantly different DNA fingerprints, and this suggested that the pathotypes were derived from independent clonal lineages. In F. oxysporurm f sp. cubense, isolates in VCG 0126 that cause disease in Highgate may possess genetic factors
that make them better adapted than those in VCG 0 120 and 0 124 to the ecological and environment factors in Honduras.
In this study, the majority of isolates recovered from Highgate had mthap 0126, whereas the majority of isolates recovered from Bluggoe had mthap 0124 (Table 3-2). Isolates representing other haplotypes were recovered on these host genotypes, but at low frequencies. Additionally, similar results were observed for some of the FHIA hybrids. The presence of a low rate of mixed infections may provide one explanation for this observation. During the initial selection of isolates from the infected tissue, occasionally multiple samples were taken from a single plant when fungal colonies displayed different growth characteristics. Of these samples, only one plant was infected with isolates having both 0 124 and 0 126 mtDNA haplotypes. Additionally, five plants from Field 1, each representing different genotypes of banana, were sampled multiple times by excising different xylem elements within a single plant. A total of 45 samples were obtained and all of the sampled plants were infected by isolates representing a single mtDNA haplotype. Thus, it appears that mixed infections occur, albeit at a low frequency.
Banana breeders in Honduras as well as other regions are involved in developing cultivars of banana with resistance to all races of this pathogen. Because of the polyploid nature of bananas and the sterility of many of the important edible cultivars, it is impossible to do traditional plant breeding and disease screening. Additionally, in Situi disease screening in regions where the pathogen is nonendemic is limited to isolates found within the country. Since isolates belonging to only three VCGs and two clonal lineages have been described in Honduras, yet over 17 VCGs and 10 clonal lineages have been described worldwide, breeding for resistance using endemic populations in Honduras may
not result in resistance to populations of the pathogen in other locations (Ploetz, 1995; Stover and Buddenhagen, 1986). Ideally, new genotypes should be screened against isolates that encompass the range of genetic diversity present in the entire formae speciales. Therefore, there is a need to assess the range of variation in the pathogen populations for virulence on particular breeding lines (Milgroom and Fry, in press). Using both the mtDNA haplotype and a nuclear DNA fingerprinting probe to estimate genetic diversity within a field site can improve in situ screening for Fusarium wilt. In this pathosystem there is a relatively long-term interaction between host and the pathogen that may not be assessed in artificial inoculation tests (Buddenhagen, 1990). Using molecular techniques for assessing the pathogen population at a given site and determining the lineage of isolates infecting particular host genotypes can increase the effectiveness of disease screening and assist in the evaluation of new breeding lines.
SUMMARY AND CONCLUSIONS
This research has provided a genetic characterization and phylogenetic analysis of isolates of F. oxysporurn f sp. cubense representing both a world collection and population from two fields. Prior investigations, which genetically characterized isolates based on VCGs, RAPIs and electrophoretic karyotyping, provided only limited information on the genetic relationships among isolates (Bentley et aL, 1995; Boehm et al., 1994; Ploetz and Correll, 1988; Sorensen et al., 1993). However, the present research, which utilized numerous, independent DNA clones to identify polymorphic loci among a worldwide collection of isolates, has provided a comprehensive assessment of the genetic relationships among isolates. Parsimony analysis of the data using bootstrapping also provided a statistical evaluation of the RFLP data as well as an unbiased assessment of significant groupings among the isolates.
In addition, this research represents the first attempt to determine the amount of genotypic diversity present in field populations of F. oxysporumf. sp. cubense based on DNA fingerprinting. The study of two Honduran field populations provides information regarding the amount of genetic variability present in a defined location. By identifying the mtDNA haplotypes of individual isolates and determining the extent of clonality in the population using a DNA fingerprinting probe, this study assessed the genetic diversity of
the collected isolates and demonstrated how the genotype of the host can affect the pathogen populations.
Both studies demonstrated the importance of choosing an appropriate DNA
marker in phylogenetic and population genetic investigations. An important benefit to using RFLPs in phylogenetic analysis is that the number of RFLP markers available for use is large because any cloned, low-copy-number piece of DNA can be used as a probe and several restriction enzymes can be assayed to identify RFLPs with each probe (Michelmore and Hulbert, 1987). One important characteristic of the RFLP markers used in this phylogenetic analysis is that they were likely selectively neutral. This is in contrast to several markers that have been employed in traditional studies, such as pathogen virulence genes or resistance to fungicides, which are under strong selection pressure in agricultural systems and may provide a biased estimate of genetic diversity (McDonald and McDermott, 1993).
Other molecular markers that are commonly used in phylogenetic analyses of fungi include isozymes, RAPDs, the nuclear and mitochondrial rRNA genes and internal transcribed spacers (ITS), and mtDNA maps. Isozymes have been used to characterize fungal isolates but their disadvantage is that the possible number of enzymes that are informative for an organism is limited and many loci are monomorphic and, thus, noninformative. (Michelmore and Hulbert, 1987).
Analysis ofF. oxysporum f sp. cubense isolates using RAPDs was attempted
initially in this research. However, this technique proved to be unreliable for phylogenetic studies. The largest obstacle to using them was that polymorphisms could not be scored unambiguously without cloning the putative polymorphic bands and then performing an
RFLP analysis. Even when RAPDs were used that could be excised with an appropriate restriction enzyme, results were difficult to interpret (see Appendix I). Two problems associated with methods based on polymerase chain reaction (PCR), such as RAPDs, are that 1) nonamplification can result from base pair substitutions at any of the different nucleotides complementary to the primers at the ends of the amplified fragment, and 2) results may be affected by changes in environmental conditions and at times are not reproducible (Rosewich & McDonald, 1994).
Mitochondrial DNAs and, in certain organisms, mitochondrial plasmids have been used to characterize isolates ofF. oxysporum f sp. melonis, f sp. conglutinans, and f sp. cubense (Jacobson and Gordon, 1990; Kistler et al., 1987; Kistler and Momol, 1990). Two disadvantages of using mtDNA for phylogenetic analyses is that the genome is small, so consequently fewer polymorphisms may be found, and the rate of evolution of mtDNA may be different than that of nuclear DNA. Mitochondrial DNA was used in both the phylogenetic and population studies. In the phylogenetic study, groupings of isolates based on mtDNA haplotypes were congruent with groups based on the RFLP analysis; however, fewer groups could be distinguished based on mtDNA analysis. In the population studies, a mtDNA haplotype was used to group similar isolates, which were then subjected to DNA fingerprint analysis to determine whether isolates were clonally derived.
Sequencing of both the nuclear and mitochondrial rRNA genes as well as ITS regions have been described for fungi and used in phylogenetic studies (Bruns et al., 1991). An underlying assumption of sequence analysis is that the phylogeny of the region is a good indicator of the phylogeny of organisms (Bruns et al., 1991). Another approach
to identifying sequence differences in the ITS region is through an RFLP analysis using the PCR-amplified region. In this study, a number of primers were used to amplify ITS regions ofF oxysporum f sp. cubense isolates. Several four- and six-base restriction enzymes were used to digest the amplified products (Appendix II). No polymorphisms were observed using the restriction enzymes studied, and this methodology was discontinued because it did not prove useful for a phylogenetic analysis.
A central point of discussion in both studies was the overwhelming evidence of clonality in this organism. With the exception of isolates in VCG 0123, the genetic evidence which resulted from this research is consistent with the hypothesis that F. oxysporuni f sp. cubense is a clonally reproducing organism. Based on criteria established by Tibayrenc et al. (1990), a unifying feature of clonally reproducing organisms is widespread geographic distribution of a few successful clones. The RFLP study identified 72 multilocus haplotypes in a worldwide collection ofF. oxysporurn f, sp. cubense. The five most common haplotypes accounted for nearly half of the isolates, while the two most common haplotypes were found on all five continents sampled in this study, indicating the pantropical distribution of a small number of genotypes. Additional evidence of clonal reproduction is the absence of recombinant genotypes. Significant gametic disequilibrium for alleles at 34 of 36 loci tested supported nonrandom association between alleles of different loci. Finally, the strong correlation between independent genetic markers (VCG, mitochondrial and multilocus RFLP haplotype) also are indicative of a clonally reproducing organism (Milgroom, 1996; Tibayrenc et al., 1990).
The Honduran field population study also indicated that clonal reproduction accounted for the occurrence of a specific genotype of the pathogen associated with
virulence on a specific host genotype. This observation is likely the consequence of linkage disequilibrium found in clonally reproducing organisms where alleles at different loci, such as mtDNA haplotypes, VCG, virulence, are observed to be repeatedly associated. The clonal structure in agricultural populations of plant pathogenic fungi also has been found in some populations of Scierotinia sclerotiorum, Fusarium oxysporumn, Phytophthora infestans and Magnaporthe grisea (Appel and Gordon, 1994; Goodwin et al., 1992; Gordon et al., 1992; Kohli et al., 1995; Kohn, 1995; Levy et al., 1991).
A significant finding of the phylogenetic analysis was that a number of the clonal lineages are phylogenetically distinct. Host specialization, such as the ability to cause disease on a particular host, appears to be polyphyletic in F. oxysporumn f sp. cubense. Based on coefficients of similarity, isolates in the two largest lineages ofF. oxysporum f sp. cubense are genetically more similar to the F oxysporum f sp. niveum isolate than to each other. Similarly, they are roughly as genetically distinct from each other as either is to the isolates of F. oxysporum f. sp. lycopersici. An interpretation of these results is that isolates belonging to the dissimilar groups acquired their ability to be pathogenic on bananas independently. This is the first reporting of a potential polyphyletic acquisition of pathogenicity on a particular host for F. oxysporum f. sp. cubense. Appel and Gordon (1994) observed a close relationship between nonpathogenic strains ofF. oxysporum and F. oxysporum f. sp. melonis isolates in VCG 0131, suggesting an independent origin as a melon pathogen separate from FE oxysporum f. sp. melonis isolates in VCG 0134. However, this study was based on mtDNA haplotype data and interpretation may be complicated by the possibility of horizontal transfer of mtDNA.
Interestingly, the results of the R-FLP study indicate that isolates in VCG 0 1214, which comprise lineage FOC V, are genetically distant from other F oxysporuni f. sp. cubense isolates. One explanation is that this group could have originated in East Africa, a region outside the host's center of origin. Another explanation is that the genetic differences in these isolates could have arisen by a founder effect. Founder effects result when small groups of individuals are isolated from larger populations. It is possible that over a period of time, isolates in VCG 0 1214 were subjected to different selection forces which altered gene frequency in this group. Additionally, mutations which occur in the founding population could be fixed in the successive generations due to the clonal nature of this organism.
The data indicate that isolates in VCG 0123, which comprise both lineages FC VII and X, are phylogenetically distinct. This group may provide evidence of an ancient genetic exchange between individuals in lineages FCC I and II, or they may represent an ancestral group possessing primitive character states.
There are numerous practical applications of this research; however, only the most important aspects are highlighted. The phylogenetic study helps clarify the genetic relationships among the isolates which represent different races of the pathogen. Isolates which represent race 1 of the pathogen occur in the two largest and divergent lineages (FCC I and II) in the collection. However, data from the Honduras field study revealed that isolates which have been considered race I based on reaction to the host differential may not share the same host infectivity. In field tests, isolates representing these two lineages were not recovered at the same frequency from susceptible host tissue, even though isolates representing FCC I and FCC II were assumed to be evenly distributed
throughout the field. Isolates which represent race 2 appear to be limited to lineages FOC I and VIII, which are associated with one main branch on the midpoint rooted tree, and FOC V, which is only distantly related to other F oxy.sporuni f. sp. cubense isolates (Figure 2-2). In contrast, isolates that represent race 4 appear to be limited to lineages FOC 11, 111, and VI, which form a second branch on the midpoint rooted tree. Therefore, these two races appear to have evolved only in certain lineages of the pathogen. Race 4 isolates occur in many of the VCGs that are present in tropical areas, such as Central America, the Caribbean, and South America, but where race 4 has not yet been reported. If the ability to cause disease on Cavendish cultivars is controlled by a single gene it is possible that race 4 isolates could evolve in regions where it has not yet been found. However, it would most likely evolve in those isolates representing VCGs that presently contain race 4 isolates. Additionally, because race 4 isolates have a wide host range, it is important to identify and develop new sources of disease resistance. These new sources of disease resistance must then be incorporated into commercially acceptable banana cultivars.
Banana breeders worldwide are involved in developing cultivars of banana with resistance to all races ofF oxysporzrni f sp. cubense but are faced with a number of obstacles. Because of the polyploid nature of bananas and the low level of fertility of important edible cultivars, it is impossible to use many of the desirable commercial clones as parental lines for breeding work. Additionally, the results of container disease screening experiments do not necessarily correlate in field studies. Stover and Buddenhagen (1986) pointed out that results from disease screening tests, which relied on
seedlings, small rhizomes or in i'dro meristemn plants grown in containers and subjected to artificial inoculation, did not always correlate with field tests.
Additionally, in sitit disease screening in regions where the pathogen is not endemic is limited to isolates found within the country. The Honduras field study demonstrated that the mtDNA haplotype and a nuclear DNA fingerprinting probe can aid in situ screening for resistance to Fusarium wilt. Using molecular techniques for assessing the pathogen population at a given site and determining the lineage of isolates infecting particular host genotypes can increase the effectiveness of screening programs and assist in the evaluation of new breeding lines. To date, disease screening programs primarily have relied on visual assessment to determine disease incidence on new breeding material. Using molecular techniques to characterize isolates of F. oxysporurn f. sp. cubense infecting banana breeding lines would enable breeders to determine host resistance and susceptibility based on the lineage of the pathogen. This could aid in directing breeding programs toward developing resistance to all lineages of the pathogen rather than the three races that are known to infect bananas. This would provide a more comprehensive approach to~breeding for resistance to this pathogen. Finally, the optimal disease screening methodology requires that the isolates that are being used encompass the range of genetic diversity present in the entire forma specialis. This could be accomplished by screening breeding materials with representatives of each of the 10 lineages of F. oxysporum f. sp. ciubense described in this research. Ideally, this would occur in a field in which all lineages of the pathogen could be introduced.
EXPLORING THE POTENTIAL OF USING POLYMERASE CHAIN REACTION METHODOLOGIES IN PHYLOGENETIC STUDIES OF FUSARIUMOXYSPORUM F. SP. CUBENSE Introduction
Randomly amplified polymorphic DNAs (RAPDs) have been used as molecular markers in a number of fungal species to distinguish among genetically different isolates, to aid in genetic mapping and for DNA fingerprinting. Randomly amplified polymorphic DNAs can be generated quickly using small quantities of DNA which do not need to be highly purified. However, RAPD techniques are prone to artifactual variation, which can result in erroneous data (Ellsworth et al., 1993). The possibility of artifactual variation prevents the direct use of the technique in genetic diversity studies aimed at describing the genetic relatedness among isolates. Similarly, the inability to determine directly if bands of similar molecular size represent similar amplification products presents a problem in scoring individuals in genetic diversity studies.
Sequencing of both the nuclear and mitochondrial rRNA genes, as well as the
internal transcribed spacer region (ITS) region, have been described for fungi and used in phylogenetic studies (Bruns et al., 1991). The ITS region and intergenic spacer of the nuclear rRNA repeat units evolve the fastest compared to other regions of the rRNA genes and may vary among species and populations (White et al., 1990). The ITS primers make use of the conserved regions of the 18S, 5.8S and 28S rRNA genes to amplify the
noncoding regions between them (White et al., 1990). The primers are based on sequences from Saccharomyces cerevisiae, Dictyosteliurn discoideun and Stylonicha pustulata (Dams et al., 1988). A diagram of these primers and their location along the nuclear rDNA region of Saccharomyces cerevisiae is provided by White et al., (1990).
Restriction fragment length polymorphisms and RAPD markers are advantageous in phylogenetic studies because they are presumed to be random traits, not subject to natural selection. In this study, the ITS regions were amplified and subjected to restriction enzyme digestion to determine if polymorphisms could be detected. Additionally, RAPD fragments were generated to provide quick identification of putative polymorphic loci. Polymorphic amplification products were labeled and used as probes for Southern hybridization, enabling the confirmation and description of each polymorphism. The objective of this study was to identify polymorphic regions of the genome that could be used in phylogenetic analysis of a large subset of a world collection of Fusarium oxysporun f sp. cubense.
Results and Discussion
ITS Region Analysis
A number of ITS primers described by White et al. (1990), were used in this study to amplify the total genomic DNAs from a subset of isolates representing a large number of VCGs and a pantropical distribution (Table A-i). The amplification protocol for this study, as well as the RAPD analysis, are provided in Table A-2. Internal transcribed spacer primers 2 and 5 were used to amplify the ITS I region of the nuclear rRNA genes and primers ITS4 and 5 were used to amplify the ITS 1, ITS2 and the 5.8S rDNA regions in F. oxysporum f. sp. cubense isolates (for primer sequences see Table A-3). The
resultant amplified products were similar for all isolates. These products were then subjected to restriction enzyme digestion using enzymes HaeIII, Sau3AI, and EcoRI. All of the tested isolates had exactly the same restriction pattern for all of the enzymes tested. RAPD Analysis
The identification of potential polymorphic regions of the genome was
accomplished using RAPD analysis following protocols described previously (Welsh and McClelland, 1990; Williams et al., 1990). A number of oligonucleotide primers were screened for their ability to amplify bands in F. oxysporum f sp. cubense (Table A-4). Many of the primers generated heterogenous products which could be used to differentiate isolates. The majority of primers generated numerous bands of varying intensity. A limited number of polymorphic bands were further tested for their potential in identifying polymorphisms using Southern analysis. Southern analysis provided a method for distinguishing whether a missing band observed in a RAPD profile was the result of the presence of multiple alleles, deletions or a problem with the amplification procedure. Randomly amplified polymorphic DNA fragments were generated using 10-base oligonucleotide primers which contained internal hexameric palindromes that may be cut by single restriction endonucleases specific to the sequence (Kit F, Operon Technologies, Inc.). Fragments were subjected to agarose gel electrophoresis, and putative polymorphic regions were identified. Polymorphic amplified products were excised from low melting point agarose gels and labeled with a nonradioactive digoxigenin (dig)-1 1-dUTP using the random primed method (Boehringer Mannheim, Genius System).
Genomic DNA isolated from a subset of 30 isolates representing 15 VCGs and a pantropical distribution (Table A-1) was digested with restriction enzymes with
recognition sequences corresponding to those found in the internal hexameric sequences of the oligonucleotide primer. Standard protocols were used for restriction enzyme digestions, blotting, and hybridizations as described in chapters 1 and 2.
When genomic DNAs were amplified with an oligonucleotide containing a BamnHI recognition sequence (GAGGATCCCT-primer F2), a 0.68-kilobase (kb) amplification product was present in isolates 0-1224, STDI, BLUG, S?, B2-1, MW65, STN3, MW1 1, CVA, STNP5, 22994 and STNP 1, but absent in the remaining isolates listed in Table A-1. This fragment, designated BLUG-2-1, was excised from a gel containing the amplification product from isolate Blug in VCG 0124 and labeled as described above. Genomic DNAs from the same isolates were digested with BamHI and probed with BLUG-2-1. A 0.68-kb fragment that hybridizes to BLUG-2-1 was present only in those isolates that amplified a .68-kb fragment by PCR. The same probe has been used on a total of 80 isolates ofF. oxysporum f sp. cubense. The probe appears to be specific for DNAs from isolates in VCGs 0124, 0124/25, 0125, 0128 and 01212.
When genomic DNAs were amplified with an oligonucleotide containing a EcoRV recognition sequence (GGGATATCGG-primer F8), a 0.95-kb fragment was present in isolates 0-1224, STD1, BLUG, S?, B2-1, MW65, STPN3, MW11, CVA, 22994 and STNP1, but absent in isolate STNP5 in VCG0125 and the remaining isolates described in Table A-1. This fragment, designated BLUG-8-1, was excised from a gel containing the amplification product from isolate Blug in VCG 0124 and labeled as described above. Genomic DNAs from the isolates were digested with EcoRV and probed with BLUG-8-1. The probe hybridized to a monomorphic 0.95-kb DNA fragment in all isolates. Additionally, the probe hybridized to a 2.1-kb DNA fragment in isolates 0-1224, STD 1,
BLUG, S?, B2-1, MW65, STPN3, MWI 1, CVA, 22994 and STNPL. The PCR amplification using the 10-mer revealed a polymorphism not detected in Southern hybridizations with restricted DNA, which suggests that the polymorphism is due to a point mutation at the F8 primer sequence, but not in the EcoRV recognition site. With the exception of isolate STNP5, the 2.1-kb fragment appears to be specific for DNAs from many isolates in VCGs 0124, 0124/25, 0125, 0128 and 01212.
When genomic DNAs were amplified with an oligonucleotide containing a EcoRI recognition sequence (CCGAATTCCC-primer F5), a 1.6-kb amplification product was present in some isolates, and either absent or less intensively amplified in other isolates. This fragment, designated SH3142-5-1, was excised from a gel containing the amplification product from isolate SH3142 and labeled as described above. Genomic DNAs from the isolates were digested with EcoRI and probed with SH3142-5-1. A 1.6kb fragment was clearly present in isolates Ph6, MW40, ORT2, 1-1, 0-1219, STGM1, F9127, 15638, FCJ7, Pacovan, F9130, DAVAO, A2-1, STB-2, and SH3142. A 1.6-kb PCR amplified fragment was also clearly visible in these isolates. Isolates 0-1224, STD1, S?, STN3, B2-1, MW1 1, CVA, STNP5, 22994 and STNP1, which contained a very faint
1.6-kb amplified product, weakly hybridized to the 1.6-kb labeled probe. In the same isolates, the probe hybridized more strongly to a 4.6-kb fragment. No hybridizations were observed in isolates BLUG and MW65, which did not contain a 1.6 amplification product. A monomorphic 3.0-kb fragment was present in isolates that contained either a 1.6-kb or
Informative polymorphic single-copy and multi-copy loci can be identified quickly by PCR using random 10-base oligonucleotide primers. However, a correlation between a
polymorphism observed through PCR amplification and a similar polymorphism observed when the DNA of the same isolates is subjected to Southern analysis, does not always occur. This poses a problem for directly interpreting the polymorphisms observed from RAPD analysis in a phylogenetic study.
Polymorphic products generated from single primer amplification may be used as probes for Southern hybridization. Restriction fragment length polymorphisms identified through Southern hybridization can be used to discern genetic relationships among isolates of the world collection. However, some RAPD bands may represent multiple-copy sequences and therefore are not ideal in phylogenetic studies. However, probes that appear to be lineage specific, such as BLUG-2- 1, may provide a quick method for determining the VCG of new isolates and for genetic studies of field populations.
In conclusion, using RAPD analysis for the determination of phylogenetic
relationships among isolates is difficult because in many cases interpretation of the data is ambiguous. Clark and Lanigan (1993) have set forth criteria that should be met when using RAPD analysis to estimate nucleotide divergence of closely related taxa. One criterion identified is that polymorphic bands must be shown to behave as Mendelian factors. Since F. oxysporum f. sp. cubense does not have a known sexual state, Mendelian inheritance can not be proven. Another criterion is that allelism of bands must be ascertained by Southern blotting or segregation analysis. For F. oxysporum f. sp. ciibense, segregation analysis is not an option and Southern blotting in two of the three examples described above did not provide definitive answers to the allelic nature of the polymorphisms present in the isolates.
Table A-1. Isolates ofFusarium oxysporumn f sp. cubense used in RAPD analysis and presence or absence of major RAPD bands.
Primer F-2 Primer F-5 Primer F-8
Isolate VCG 0.68-kb band 1.6-kb band 0.95-kb band
0-1219 0120 +
STGM1 0120 +
F9127 0120 +
15638 0120 +
FCJ7 0120 +
Pacovan 0120 +
ORT2 0120 +
F9130 0121 +
Ph6 0122 +
DAVAO 0123 +0-1224 0124 + +* +
STD1 0124 + +* +
BLUG 0124 + +
S? 0124 + +* +
B2-1 0124 + +* +
MW65 0124 + +
STN3 0124 + +* +
MW11 0124/25 + +* +
CVA 0124/25 + +* +
STNP5 0125 + +*
STB-2 0126 +
22994 0128 + +* +
N5331 0129 +
A2-1 01210 +
Al 01210 +
SH3142 01211 +
STNP-1 01212 + +* +
1-1 01213 +
MW40 01214 +
* Weak band present
Table A-2. Amplification protocol for RAPD and ITS analysis
25 ng template DNA 15 ng primer DNA (Operon Technologies, San Jos6, CA)
0.3 p1 10mM dATP 0.3 /l 10mM dCTP 0.3 4 10mM dTTP 0.3 l 10mM dGTP
0.2 U Taq polymerase (Boehringer Mannheim, Indianapolis, IN)
1 l 10 mM MgCI,
2.5 ul 10Ox Taq polymerase buffer sterile water to volume of 25 4l
Amplifications were performed in a Gene Machine II thermal cycler. The machine was programmed for one cycle for 30 seconds at 940C followed by 35 cycles of 1 minute at 94C, 1 minute at 350C, and 2 minutes at 72C, followed by a final cycle of 10 minutes at 720C.
Table A-3. Primer sequence for the ITS regions
ITS Primerx Sequence
"Primers obtained from Dr. Kerry O'Donnell with the sequences based on White et al. (1990).
Table A-4. List of oligonucleotide primers tested on Fusarium oxysporum f sp. cubense isolates
Operon Primer Recognition Sequence (Yes/No)
BI GTTTCGCTCC Y
B2 TGATCCCTGG Y
B3 CATCCCCCTG Y
B4 GGACTGGAGT Y
B5 TGCGCCCTTC Y
B6 TGCTCTGCCC Y
B7 GGTGACGCAG Y
El CCCAAGGTCC Y
E2 GGTGCGGGAA Y
E3 CCAGATGCAC Y
E5 GTGACATGCC Y
E6 AAGACCCCTC Y
E7 AGATGCAGCC Y
Ell GAGTCTCAGG Y
E 12 TTATCTCCCC Y
E13 CCCGATTCGG Y
E 14 TGCGGCTGAG Y
E15 ACGCACAACC Y
E16 GGTGACTGTG Y
E17 CTACTGCCGT Y
E18 GGACTGCAGA Y
E19 ACGGCGTATG Y
E20 AACGGTGACC Y
Fl ACGGATCCTG Y
F2 GAGGATCCCT Y
F4 GGTGATCAGG Y
F5 CCGAATTCCC Y
F6 GGGAATTCGG Y
F7 CCGATATCCC N
F8 GGGATATCGG Y
FIO GGAAGCTTGG Y
Fll TTGGTACCCC Y
F12 ACGGTACCAG Y
F13 GGCTGCAGAA Y
F14 TGCTGCAGGT N
F15 CCAGTACTCC N
F16 GGAGTACTGG Y
F17 AACCCGGGAA Y
F18 TTCCCGGGTT Y
F19 CCTCTAGACC N
F20 GGTCTAGAGG Y