Evaluation of the parasitic relationship of Melanospora and other allied genera with Fusarium oxysporum


Material Information

Evaluation of the parasitic relationship of Melanospora and other allied genera with Fusarium oxysporum
Physical Description:
vii, 129 leaves : ill. ; 29 cm.
Harveson, Robert Martin, 1960-
Publication Date:


Subjects / Keywords:
Watermelons -- Diseases and pests   ( lcsh )
Fusarium oxysporum   ( lcsh )
Wilt diseases   ( lcsh )
Fusarium diseases of plants   ( lcsh )
Plant Pathology thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Plant Pathology -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1999.
Includes bibliographical references (leaves 120-128).
Statement of Responsibility:
by Robert Martin Harveson.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030474449
oclc - 43343326
System ID:

This item is only available as the following downloads:

Full Text







I would first like to extend my thanks to Dr. G. N. Agrios and the committee in

charge of awarding plant pathology departmental assistantships. Without this possibility, I

would not have been able to come to Florida to complete my education. Most

importantly, I want to thank my major professor, Dr. Jim Kimbrough for listening to my

unorthodox research ideas and being willing to support them at a very difficult time in my

career. His constant sense of humor, enthusiasm, and guidance will always be appreciated

and never forgotten. I would also like to thank my other committee members-Dr. Don

Hopkins, Dr. Gary Simone, Dr. Andy Ogram, and Dr. Bill Zettler-for agreeing to serve on

my committee at the last minute and their input throughout the course of this project. Dr.

Hopkins also agreed to support peripheral areas related to this work by furnishing his

research plots and laboratory in Leesburg. Unfortunately, none of this research will

appear in this dissertation. Dr. Simone was instrumental in also sharing his lab facilities

and being there to listen to me vent and advise me when my course of study changed. Dr.

Ogram has been helpful in finding flaws in this work, which subsequently led to an

improved final product. Dr. Zettler's contribution is particularly important. He served as

proxy for both the oral exam and the final defense, hence he has become a formal

committee member in literally the twelfth hour. Without the encouragement and

assistance from Dr. Gerry Benny and Ulla Benny over the last three years, this project and

my sanity would probably not be where they are now. I want to also thank Drs. Dave

Mitchell, Dan Purcifill, and Tom Kucharek for advice and academic discussions over the

years. The influence and mentoring ability of Dr. Charlie Rush on my career is

incalculable. I owe him a debt of gratitude that can never fully be expressed or repaid. I

am also grateful to Dr. Corby Kistler for so succinctly illustrating to me that my ultimate

future led down a different path than I first began when starting my doctoral work. I want

to also express my appreciation to Galen Jones and Jack Bishko for their consultation and

recommendations for the various statistical analyses used in this study. I need to also

acknowledge my parents, Wayne and Maralyn Harveson, and my grandmother, Marietta

Martin, for always believing in me and giving unconditional support and encouragement

for any endeavor that I have chosen to undertake. Finally, I wish to thank my wife,

Tammy, for understanding my sometimes unreasonable drive to complete this work and

for reluctantly tolerating the 12-14 hour work days spent over the last year and a half.

Her willingness to consider moving to a new and totally foreign location has been a

tremendous comfort to me while finishing my work. The completion of this project would

not have been possible without her.


ACK NOW LEDGM ENTS........................................................................... ii

A B ST R A C T ..................................................................................... vi


1 IN TR O D U C TIO N ............................................... ........................... 1


Introduction........................... ....................... 17
Materials and Methods............................ ............... 22
Results................ ........... ...... ............. ........ 25
D iscussion................................... ............................ ..... 38

Introduction..................................................................... 40
M materials and M ethods................................................ ......... 43
R esults.............................................. ................. ......... 51
D iscussion.................................. ....................... ......... 63

SPHAERODES AND PERSICIOSPORA................... .......... 70

Introduction........................... ...................................... 70
M materials and M ethods....................................................... 72
Results..................................................... .... ................. ....... 79
D iscu ssio n.............................................................. ............ 90

OF WATERMELON ............................................................ 93

Introduction.............................. ..... 93
Materials and Methods............................. .......... ...... 95

R esults.................................................................. ............. 100
D iscussion.................................................................. ............ 110

6 SUMMARY AND CONCLUSIONS..................................... 116

REFERENCES ................................................ 120

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

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



Robert Martin Harveson
August 1999

Chairman: J. W. Kimbrough
Major Department: Plant Pathology

Fusarium wilt, caused by Fusarium oxysporum, has been a problematic disease for

many different crops for over 100 years. The discovery of a group ofpyrenomycetous

ascomycetes found naturally in association with severalformae specials ofF. oxysporum

from field samples, has generated the interest to investigate the use of these fungi as

biocontrol agents for watermelon wilt caused by F. oxysporum f sp. niveum (FON). The

first of these fungi was isolated from the sugar beet pathogen, F. oxysporum f. sp. betae,

and was identified asMelanospora zamiae. Later findings from other wilt-infected crops

yielded another five isolates that very closely resembled the original M zamiae. However,

based upon current taxonomic criteria, these five isolates were determined to consist of

two distinctly different genera, Sphaerodes retispora var. retispora, and Persiciospora


All six isolates were found to be strongly dependent upon F. oxysporum for

growth and sporulation. Linear growth and production ofperithecia were significantly

improved for all isolates in dual culture with F. oxysporum or growing in filtrate from a

liquid culture ofFusarium. Significant differences in growth were observed among and

between the different pyrenomycetes. Without the presence ofF. oxysporum or its

metabolic growth products, growth and reproduction were generally very poor for all

isolates. The better media without Fusarium tended to be those composed of simple

sugars like dextrose or maltose, while those producing the worst results contained starch

or cellulose. All isolates were determined to be parasitic on F. oxysporum through the use

of several cultural assays, but not all were pathogenic to FON. Persiciospora caused no

detectable reduction in hyphal growth, whereas the Sphaerodes isolates were the most


Isolate LE (S. retispora) was selected for disease reduction evaluations in the

greenhouse. It was incorporated into sodium alginate pellets and delivered to the infection

court by either root treatment or amendment to infested soils before transplantation. Yield

data were collected after eight weeks, and consisted of mortality counts and total plant dry

weights. This technique did significantly improve plant dry weights of treated plants

compared to the FON-inoculated controls. These are promising results for an alternative

method for managing Fusarium wilt of watermelon. However, it is not known whether

this procedure is efficacious or even feasible under field conditions. If proven to be

successful in the field, this procedure could also presumably be utilized for disease control

with other vegetable crops in Florida that are similarly grown as transplants. However,

care must be taken for the selection of the potential biocontrol isolates, because not all

included in this study were found to equally colonize or induce deleterious effects upon all

formae specials ofF. oxysporum.


The genus Fusarium was first erected in 1809 by Link when he described F.

roseum as the first species (Booth 1984). Fusarium species are ubiquitous soil fungi

found in many climatic extremes such as deserts and the arctic, in addition to temperate

and tropical regions of the world. They can be saprophytes and pathogens infecting both

animals and plants (Nelson 1981). Because of variation in culture, it has been difficult for

workers to correctly classify this genus into species. The basis for Fusarium classification

was established by Appel and Wollenwebber (1910). They placed species studied into

sections based on various morphological characteristics (Coons 1928). These sections

were later named and expanded to include a type species for each section (Wollenwebber


Of the 43 species of Fusarium listed by Booth (1971), 27 are pathogenic to plants

and cause some of the most important diseases in agriculture. The pathogens may be

divided into three main categories: the wilts, the root rots caused primarily by F. solani,

and those fungi attacking graminaceous plants, such as F. moniliforme, F graminearum,

F. culmorum, and F. avenaceum (Price 1984). The vascular wilts, caused byformae

specials ofF oxysporum, are considered to be the most economically important due to

the diversity of plants attacked and the destructiveness of the diseases (Nelson 1981).

Fusarium oxysporum was the first pathogenic species described, and was included

in the section Elegans by Wollenwebber and Reinking (1935). They divided this section

into three subsections and further subdivided these subsections into 10 species, 18

varieties, and 12 forms (Nelson 1981). Their taxonomic system was difficult to use by

others because they placed such great emphasis on spore measurements and septations,

which are highly variable.

Snyder and Hansen (1940) simplified Wollenwebber and Reinking's system by

combining all members of the section Elegans into one species, F. oxysporum, by the

utilization of standard culture media and incubation temperatures. The different parasitic

forms were recognized by their selective pathogenicity to specific crops and were

designated asformae specials. Formae specials are also further subdivided into

biological races based on pathogenicity to differential cultivars containing different

resistance genes.

Recognition of the limiting effect of the Fusarium wilts on crop production goes

back to the 1890's. Atkinson (1892) described the first Fusarium wilt disease, and

referred to it as "Frenching." It had been causing serious losses to Alabama cotton

growers. He demonstrated the presence of gummy substances blocking the vascular

elements and illustrated the phialids producing the microconidia, which has proven to be

the main diagnostic criterion for identification ofF. oxysporum. He also named the causal

agent as F. vasinfectum (Atkinson 1892), which is now known to be the form name for

Fusarium wilt of cotton. There are many additional early examples of crops threatened by

Fusarium wilts including potatoes, celery, tomatoes, flax, bananas, cabbage, cowpea, and


The cultivated watermelon, Citrullus lanatus (Thunb.) Matsum and Nakai, is the

most economically important member of the genus Citrullus, and is grown for fresh

consumption as a desert vegetable. The citron, C. lanatus var. citroides (L. H. Bailey)

Mansf, is not eaten fresh, but is candied. It has also been important in breeding for

disease resistance because it readily crosses with the watermelon, and is resistant to

Fusarium wilt. Watermelons are an important truck crop in the South. Commercial

production is approximately 184,000 acres nationally, concentrated in Florida, Texas,

Georgia, and California (Peirce 1987). It has an estimated gross market value of $150

million (Martyn 1986).

The watermelon is considered to be native to the southern deserts of Africa.

David Livingstone reported that wild populations covered vast areas of the Kalahari

Desert after unusually heavy rains. Both natives and animals were observed consuming

the fruit (Livingstone, 1859). Carrier, however, concluded wrongly that it was indigenous

to North America based on descriptions by French explorers of Indians growing melons in

the Mississippi Valley (Carrier 1924).

The actual time and place of domestication of the watermelon is unknown, but by

2000 B.C. it was being cultivated along the Nile by Egyptians and widely grown by

prehistoric agricultural peoples of sub-Sahara Africa (Peirce 1987, Sauer 1994).

Watermelons were introduced to India and China by A.D. 1100, and brought to Europe by

the Moors. Popularity in Europe spread slowly, presumably because of the long, hot

growing season required for good yields (Sauer 1994). In North America, Spanish

colonists were growing them in Florida during the 1500's, and by 1650, Dutch and British

colonies in the Northeast were cultivating them abundantly (Peirce 1987). Watermelons

were enthusiastically adopted by the North American Indians. By 1700, they were being

grown from the Great Lakes region to the American southwest and lower Rio Grande

Valley of Texas (Sauer 1994). This probably explains Carrier's mistaken conclusion that

they originated in North America. Hawaiians also began growing watermelons after being

given seeds by Captain Cook on one of his voyages (Sauer 1994).

Fusarium wilt of watermelon was first described in 1894 from Georgia and South

Carolina by E. F. Smith (Smith 1894). By the time of his second report (Smith 1899), it

had been identified throughout the melon-growing areas of the Southeast ranging from

Texas to Maryland. Taubenhaus began investigations on this disease in Texas in the late

teens because it was considered to be the most important and destructive disease affecting

watermelons (Taubenhaus 1920). He showed that the organism was widely distributed at

different soil depths, that disease became more severe and spread more rapidly after using

manure than with commercial fertilizers, and that activity of the pathogen was greater

under warm conditions than cool. He further concluded that after a three year absence of

melons in fields, pathogen levels were reduced, but not completely eradicated

(Taubenhaus 1920). Elsewhere throughout the South, large tracts of land, previously

cropped profitably to watermelons, were rendered unfit and abandoned because of this

disease. It is now known to be widespread throughout the U.S. and wherever else

watermelons are grown, and is considered to be the primary limiting factor to production

in many of these areas.

Smith's (1899) research on the watermelon wilt was combined with a wilt of

cowpea and extended work on the wilt disease of cotton originally described by Atkinson.

He gave the fungi causing these diseases different names: F. niveum, F. tracheiphila, and

F. vasinfectum for watermelon, cowpea, and cotton, respectively. This was done because

they could not be distinguished in pure culture, but he determined that they differed in

pathogenicity after cross inoculation studies on each crop host. He further observed that

land infested with the melon fungus should not be planted with watermelons, but could be

grown with cotton or cowpea. He also demonstrated that the pathogen could be spread to

new land on seeds, in run-off water, barnyard manure, and in soil clinging to farm

equipment (Smith 1899).

These outstanding researches were marred however, by Smith's erroneous

conclusion that Neocosmospora, found on dead stems, was the teleomorph of the

Fusarium wilt pathogen (Smith 1899). This conclusion caused some confusion until it was

first challenged by Higgins in 1909 (Higgins). He proved that Neocosmospora was a

saprophyte and distinct from the organism causing the wilt disease of watermelon. This

was subsequently confirmed by others (Taubenhaus 1920, Wollenwebber, 1913).

The pathogen survives in soil primarily as chlamydospores and may remain

dormant for many years. This stage serves as the predominant source of inoculum

although infection may also occur via macro- and microconidia (Martyn 1996).

Chlamydospores may germinate in response to host or nonhost roots, or by contact with

fresh noncolonized plant debris (Nelson 1981). After chlamydospore germination, conidia

and new chlamydopsores may be formed, in addition to hyphae. Penetration then occurs

on the host plant either directly or through wounds.

Once the pathogen enters a susceptible host, it invades the vascular system where

it becomes systemic and moves throughout the plant by means of hyphal growth or

movement of microconidia in the vascular stream (Martyn 1996, Nelson 1981). It is

generally confined to the xylem vessels and tracheids initially, but may also enter the

parenchyma later. As disease development progresses, the fungus may also move into

adjacent tissues such as the pith, cambium, phloem, or cortex (Nelson 1981). At this point

wilt symptoms are severe and often result in plant mortality. After the plant dies,

chlamydospores are reformed from mycelium or macroconidia and serve as the

overwintering propagule. In resistant plants, colonization is generally limited to the cortex

and several vessel elements (Martyn 1996).

Fusarium oxysporum f. sp. niveum has been shown to be seedborne in low levels,

however the major mechanisms of spread are via movement of infested soil or plant debris.

The disease is usually more severe on light, sandy soils with a low pH, and has an

optimum soil temperature requirement between 20-27 C (Martyn 1996).

Control of this disease is very difficult to achieve. It was noted long ago that

spraying plants with Bordeaux mixture (CaSO4 and lime) or a similar fungicide had little

or no effect because of the location of the fungus inside the vascular system (Taubenhaus

1920). Control with broad spectrum soil fumigants is erratic and expensive (Martyn

1986). However, it has been shown to be effective under certain conditions. Hopkins and

Elmstrom (1979) successfully obtained commercial yields on severely infested lands in

Florida utilizing sodium azide and chlorinated hydrocarbons combined with methyl

isothiocyanate (DD-MENCS). This was only possible with moderate or highly resistant

cultivars, but not with susceptible ones. They further demonstrated that control was more

consistent and effective when applications were made prior to bed formation due to

deeper penetration of the chemicals (Hopkins and Elmstrom 1979). However, some

studies since this time have provided evidence that recolonization ofFusarium after

fumigation still can occur in some cropping systems (Marois and Mitchell 1981).

Biological control is another form of disease management that is becoming

increasingly popular and important for Fusarium wilt control. Cook and Baker (1983)

define biological control as "the reduction of the amount of inoculum or disease-producing

activity of a pathogen accomplished by or through one or more organisms other than

man". Their very broad-based definition includes cultural practices, breeding for

resistance, or the use of antagonists, nonpathogenic, or other beneficial microorganisms

either introduced or resident (Cook and Baker 1983).

Numerous cultural methods that attempt to modify the environment to the

advantage of the host have proven to be of some value for Fusarium wilt management.

Avoiding monoculture of watermelons is one of the simplest and most important of these

methods. In most cases, continuous culture tends to increase inoculum potential of soils.

An exception to this was noted with the use of the moderately resistant cultivar Crimson

Sweet. After four-five years, it induced a wilt suppressive soil (Hopkins and Elmstrom

1984, 1987). Because of the capacity of Fusarium to survive as chlamydospores or on

roots of plants as symtomless carriers, control through crop rotation alone appears to be

unlikely. Rotation also should be combined with long periods (five-seven years) without

growing watermelons. Planting melons on new land is another excellent strategy to avoid

wilt, however, this is also very impractical for many producers. Soil solarization is

another cultural method that has shown promise by delaying onset of wilt and reducing

overall disease incidence. It reduced initial inoculum and overall population levels in soil,

but was only effective in fields with moderate to low inoculum levels (Martyn 1986,


The management of crop nutrition has also resulted in varying levels of success in

controlling Fusarium wilt diseases. These diseases have been demonstrated to be less

severe in soils with a high pH. Since the addition of many elements profoundly influences

soil pH, it is not known whether the pH itself or the increased plant nutrition is responsible

for wilt control. One possible explanation for this phenomenon has been hypothesized that

at high pH, many micronutrients tend to become unavailable. Fusarium oxysporum has a

high requirement for these elements, and its inability to grow and sporulate under these

conditions may explain reduced disease at high pH (Jones et al 1989). Evidence for this

idea was obtained when tomato wilt still occurred in limed soils in Florida after certain

combinations of trace elements (iron, manganese, and zinc) were applied (Jones and Woltz

1981). Regardless of the mechanism responsible for reducing disease severity, Jones et al.

(1989) demonstrated that the use of lime, N03-N instead of NH4-N, and low phosphorus

was highly effective for Fusarium wilt of tomato. However, Hopkins and Elmstrom

(1976) found no significant differences in watermelon wilt control comparing the nitrogen

fertilizer treatments of NO3 and NH4.

The most effective control measure for any Fusarium wilt disease, other than long

rotations and planting on new land, is generally considered to be the use of resistant

cultivars. The first known instance of developing a plant variety with resistance to a

Fusarium disease by hybridization was for watermelon wilt. Orton (1907, 1909) was able

to create the wilt resistant cultivar Conquerer by crossing the inedible but resistant citron

melon with the susceptible cultivar Eden. Conquerer was not widely used commercially

due to certain horticultural limitations, but it was employed as a parent for early resistance

breeding (Porter and Melhus 1932). Since that time, extensive breeding programs have

successfully produced a number of wilt resistant cultivars, which had their beginning with

this original cultivar (Walker 1957). The major disadvantage of using genetic resistance is

that performance of individual cultivars may vary among different locales because of

different disease reactions of pathogen races. Currently there are 3 known races ofF.

oxysporum f. sp. niveum (races 0,1,and 2). Race 0 is pathogenic only to cultivars with no

resistance genes. Since most modern cultivars have some Fusarium wilt resistance, this

race of the pathogen is no longer important. High levels of race I resistance are available

for many good commercial cultivars. Race 2 isolates have only been identified, to date,

from Israel, Texas, Oklahoma, and Florida (Martyn, 1996), but will undoubtedly be

problematic for watermelon growers from other locations within the near future.

Induced resistance and suppressive soils are two of the best known and successful

biological control methods that utilize microorganisms. Induced resistance, also known as

cross protection and immunization, has been extensively documented against all classes of

pathogens in many different plant pathosystems, however, it has been best exemplified for

disease control of foliar pathogens of cucurbits, beans, and tobacco (Kuc 1982, 1990).

Induced resistance has also been shown to be effective for a number of vascular

wilt diseases. Reduced Fusarium wilt has been reported by inoculation of susceptible

hosts with nonpathogenic races or different wilt pathogens prior to challenge with the

target pathogen. The selection and subsequent inoculation of nonpathogenic strains ofF.

oxysporum have been shown to successfully reduce wilt by pathogenic strains in celery

(Schneider 1984), cucumber (Paulitz et al. 1987), sweet potato (Ogawa and Komada

1985), and watermelon (Biles and Martyn 1989, Larkin et al. 1996).

The inoculation of non-host pathogens has additionally been documented in

reducing disease severity. Phillips et al. (1967) observed reduced wilt of tomato by

inoculations of Cephalosporium spp. Another interesting example by Gessler and Kuc

(1982) demonstrated that root protection from F oxysporum f. sp. cucumerina on

cucumber could be induced sytemically with otherformae specials (melonis and

conglutinans), or by leaf inoculation with Colletotrichum lagenarium or tobacco necrosis

virus. Specifically, watermelon wilt has similarly been proven to be effectively reduced by

preinoculation with the non-watermelon pathogens F oxysporum f.sp. cucumerina (Biles

and Martyn 1989) and F oxysporum f. sp. lycopersici, Verticillium alboatrum, and

Helminthosporium (Cochliobolus) carbonum (Shimotsuma et al. 1972).

Conversely, Wymore and Baker (1982), concluded that using this approach to wilt

control of tomatoes under greenhouse conditions was not practical. The induced

resistance initially obtained required inoculum densities of the cross protection agents to

be of equal or greater than that of the challenge pathogen. Additionally, the protection

that did occur was only temporary by delaying symptoms and eventually produced disease

equal to that of the inoculated controls (Wymore and Baker 1982).

Over the last 25-30 years, the greatest progress achieved in biological control

employing soil microorganisms has been investigating and understanding the principles

responsible for pathogen suppressive soils (Cook 1990). Most soils are suppressive some

degree, but those that are pathogen suppressive have a unique ability to reduce the impact

of soilborne pathogens following certain agronomic management systems. Suppressive

soils have been defined by Cook and Baker (1983) as those soils where disease never

becomes established or persists, becomes established but causes little damage, or is

established and causes disease for a short time but afterward becomes less problematic

even though the pathogen is still detectable in these soils.

The occurrence of pathogen suppressive soils over the years has been recognized

primarily because of observed disease severity in crops being lower than was anticipated

(Baker and Cook 1974). In general, suppressiveness may act directly on the pathogen by

affecting propagule growth and survival, or it conceivably can work indirectly through the

host in the form of cross protection or induced resistance (Cook 1990, Cook and Baker,

1983). This example of induced resistance is different from the one previously described

by being resident or naturally-occurring in the soil rather than being introduced.

Pathogen suppressive soils are now known for 10-15 plant diseases (Cook 1990,

Cook and Baker 1983). With the possible exception of take-all suppressive wheat soils

Gaeumannomyces graminis var tritici), the most extensively studied suppressive soils

worldwide are those of Fusarium wilt. The existence of wilt suppressive soils has been

recognized for over 100 years, and research on them has been conducted for greater than

70 years. Atkinson (1892) first noticed that cotton wilt was more severe on sandy soils in

Alabama. Later work on wilt suppressive soils from Central America involving banana

wilt (F. oxysporum f sp. cubense) and the pea wilt caused by F.oxysporum f. sp. pisi in

Wisconsin similarly noted that disease severity was less devastating on heavier clay soils

than on lighter, sandy ones (Toussoun 1975, Reinking and Mann 1933, Walker and

Snyder 1933)

A number of other agricultural soils throughout the world exhibit a similar

phenomenon of disease reduction, including several areas of the San Joaquin and Salinas

valleys of California (Scher and Baker 1980, Smith and Snyder 1971), the Chateaurenard

region of the Rhone Valley in France (Alabouvette 1986, 1990), and on an experimental

farm in Central Florida (Hopkins and Elmstrom 1984, 1987). Most of these soils have a

number of characteristics in common. They tend be alkaline and heavy-textured.

Suppressiveness is destroyed by heat treatments, but can be restored by the addition of

small quantities of suppressive soil. This suggests a biological origin for induction and

maintenance of suppression. Suppression to a particularforma specialist is also

suppressive to other pathotypes ofF. oxysporum, but not to nonpathogenic F. oxysporum

or other soil pathogens. This also holds true for other species ofFusarium. Pathogen

suppressive soils also tend to have higher concentrations of overall soil microflora than do

wilt conducive soils (Cook and Baker 1983 ) Competition for iron or carbon between the

pathogens and resident bacteria and/or nonpathogenic fusaria is believed to be responsible

for suppression in many of these soils. The mechanism of suppression in other soils seems

to be a general resistance in plants induced by nonpathogenic F oxysporum or growth-

promoting rhizobacteria (GPR). They have the ability to infect plants before the

pathogens due to a higher degree of aggressiveness.

The watermelon wilt suppressive soil from Florida is the curious exception to

several of these traits characteristic of most wilt suppressive soils. It has similarly been

shown to harbor higher populations of soil microorganisms than wilt conducive soils, but

is different by its sandy, acidic nature and low organic matter content. This soil is also

unique by suppression being induced by monoculture of the moderately resistant cultivar

Crimson Sweet (Hopkins and Elmstrom 1984, 1987). The suppressive effect of the other

soils is believed to be constituitive. The primary mechanism of suppression in the Florida

soil is not competition for micronutrients among native soil organisms, but an induced

resistance caused by the high populations of nonpathogenic F. oxysporum isolates.

Mycoparasitism is another method that has been utilized in controlling soilborne

pathogens, and this term is defined as the parasitism of one fungus by another

(Hawksworth et al. 1995). Barnett and Binder (1973) divided mycoparasites into two

major categories based upon their mode of obtaining nutrition. The highly aggressive

necrotrophic parasites make contact with hosts, secrete hyphal wall-degrading enzymes

causing death of the host cell, and finally utilize nutrients released through this process

(Barnett and Binder 1973, Jeffries and Young 1994). Necrotrophic parasites tend to have

large host ranges and are largely unspecialized as to their mechanism of parasitism. Most

members of this group are opportunisitc and capable of living saprophytically or

competing with other organisms for space and nutrients (Jeffries and Young 1994).

The biotrophic parasites, conversely, obtain their nutrition from living host cells.

The living host supports the parasite and may appear unaffected at first. Throughout the

years of close association, some biotrophic parasites have lost the ability the synthethize

one or more required nutrients, and must rely upon the host to supply them. Not only

must the host supply these food stuffs, but the parasite must also be capable of

successfully retrieving them from the host. Thus a physiological balance is established in

this relationship and the parasite appears to be highly adapted (Jeffries and Young 1994).

This is probably why biotrophic parasites tend to be more restrictive in host range than do

necrotrophs. They also form specialized infection structures, or what Jeffries and Young

(1994) call "host-parasite interfaces." Additionally no biotrophic parasite has ever been

demonstrated to produce cell-degrading enzymes or exotoxins (Jeffries and Young 1994).

There are 3 separate types of biotrophic relationships that can be distinguished

based upon the morphology of this interaction. There are those in which the whole thallus

of the parasite enters the hyphae of the host, and are called intracellular biotrophs. The

haustorial biotrophs send a lobed haustorium into host cells via invagination of the

plasmalemma. Nutrients are absorbed from host cytoplasm back to the parasite through

this haustorium. The third type is apparently very unusual and the most specialized of the

three. It involves physical contact between host and parasite, but no penetration occurs.

Until the ultrastructural studies of Hoch (1977), it was not understood specifically how

this relationship functioned. His work showed that nutrients are obtained by the parasite

through the formation of channels of contact between closely appressed host-parasite

hyphae. Specialized contact or buffer cells form on the parasite and contact the host

hyphum without penetration. These cells serve as a stable interface for the transfer of the

nutrients through channels similar to plasmodesmata (Hoch 1977). The plasma membrane

of both organisms come into contact and fuse, therefore cytoplasmic continuity is achieved

(Hoch 1977, Jeffries and Young 1994). Since this is such an important feature of this type

of parasitism, it has been renamed fusion biotrophism, to replace the older term, contact

biotrophism (Barnett and Binder 1973, Jeffries and Young 1994).

Fusion biotrophs consist primarily ofanamorphic hyphomycetes, although several

teleomorphic states of the Ascomata have also been reported (Jeffries and Young 1994).

These include several closely related genera like Gonatobotrys, Gonatobotryum, and

Nematogonum (Whaley and Barnett 1963, and Barnett and Binder 1973). The

pyrenomycete Melanospora zamiae has also been reported to parasitize a large number of

fungi, (Hanlin et al 1993, Jordan and Barnett 1978), which is considered to be unusual for

a fusion biotrophic mycoparasite.

Interestingly, one of the first biotrophic parasites discovered was that of

Gonatobotrys simplex (Whaley and Barnett 1963). It also happened to serve as the model

for Hoch's ultrastructural work with fusion biotrophs (Hoch 1977). The teleomorph of G.

simplex has further been identified asMelanospora damnosa, which is itself a fusion

biotrophic parasite ofFusarium sacchari, causal agent of corn stalk rot (Vakili, 1989).

Because of the unique ability ofF. oxyporum to inhabit and colonize the

specialized niche inside a plant's vascular system, no studies have been published

attempting to control wilt diseases with mycoparasites that act directly on the pathogen.

Cook and Baker (1983) point out that in general, the more internal a pathogen is during

the host-pathogen interaction, the less vulnerable it is to antagonists because of the degree

of protection afforded by the plant. The vascular wilt fungi, and F. oxysporum in

particular, have only a brief epiphytic stage and exist almost entirely as endophytes within

the root cortex and vascular tissue. However, these same authors also point out that the

specialization required for this type of life strategy can also lead to an increased

vulnerability to the physical environment and antagonists when not residing within the host

(Cook and Baker 1983, Baker and Cook 1974).

Owing to the difficulties associated with direct observation of living fungi in the

soil, most information concerning mycoparasites and their hosts have been obtained, by

necessity, from dual cultures in vitro, or by inference from field observations (Jeffries and

Young 1994). Mycoparasites are also notoriously difficult to observe in the lab because

of slow or poor growth in axenic culture. Very little is known about the natural

occurrence of fusion biotrophic mycoparasitism, however because of these reasons, this

field is wide open for exploration.

The impetus behind this research began while collecting isolates ofF. oxysporum f

sp. betae from sugar beet in Texas. After several weeks, approximately 10-15% of the

isolates from the field became contaminated with an unknown ascomycete that seriously

affected growth and vigor ofFusarium in culture. This unknown fungus was identified as

Melanospora zamiae, and was determined to be occurring naturally with F. oxysporum f.

sp. betae. Over the last year, five additional isolates ofPyrenomycecious ascomycetes

have been found in Florida associated with field populations of three additional plant

pathogenicformae specials ofF. oxysporum. These organisms are closely related to, but

distinct from, M. zamiae.

This study will attempt to address some of these aspects involving the

mycoparasitism ofF. oxysporum byM. zamiae and other allied genera. None of these

additional genera have been conclusively shown to be parasitic. Chapter 2 will cover the

identification and taxonomy of this group of Pyrenomycetes. Chapter 3 will address the

stimulation of growth and reproduction of the parasites by the F. oxysporum hosts.

Chapter 4 will determine host range and measurable damage caused by the parasites to F.

oxysporum in culture. Chapter 5 will delve into some ecological aspects of the parasite

and its ability to colonize both plant hosts and F. oxysporum, and to survive in soil.

Chapter 5 will also attempt to combine knowledge from other aspects of this study and

attempt to utilize one or more of these parasites as a biological control agent for

watermelon wilt at the greenhouse level.



During the course of studying soilborne isolates ofFusarium causing root diseases

of a number of crops, species resembling Melanospora were found in constant association

with these organisms. These Melanospora-like strains appeared to be mycoparasitic and

causing damage to the fusaria. Because there has been confusion as to the limits of

Melanospora, a study of this genus and its closely related allies appeared warranted.

The genus Melanospora was first erected in 1837 by Corda (1837). He described

three species in the following order, M. zamiae, M. chionea, and M. leuchotricha (Shear

and Dodge, 1927). Because he did not designate a type species, most authors have

considered M zamiae to be the type (Von Arx 1981, Cannon and Hawksworth, 1982),

although Clements and Shear described M chionea as the type (1931).

Melanospora is currently classified as belonging to the Ceratostomataceae in the

Sordariales (Hawksworth et al, 1995, Barr, 1990). However, this genus has long been a

taxonomic problem for mycologists and has resided in at least five orders and six families

since its first description. Gilmer (1957), and Gwynne-Vaughn and Barnes (1927) both

included it in the Nectriaceae of the Hypocreales while Alexopoulos (1962) and Stevens

(1974) placed it in the Melanosporaceae of the same order.

Numerous authors have placed Melanospora in the now invalid order Sphaeriales.

Within this order, Melanospora has been included in the Melanosporaceae (Dennis 1978,

Bessey 1935,1961, Munk 1957), Hypocreaceae (Clements and Shear 1931),

Ceratostomataceae (Fitzpatrick 1934), and the Xylariaceae (Nannfeldt 1932).

Based on centrum development, Luttrell (1951) included Melanospora in the

Melanosporaceae of the Diaporthales. This placement was established to include genera

from the Hypocreales that had a Diaporthe-like centrum, but resembled members of the

Diaporthaceae in centrum-type only.

Weymeyer (1975) recognized three families in the Melanosporales,

Chaetomiaceae, Melanosporaceae, and Ceratocystaceae, and placed Melanospora in the

Melanosporaceae. He regarded the order as intermediate between the Plectascales and the

Hypocreales or Sphaeriales. Alexopoulos et al. (1996), listed the genus in the

Melanosporales, but designated it as a monotypic order with no familial rank.

Over one hundred species have been named Melanospora (Goh and Hanlin 1994),

but Doguet (1955) considered most to be synonyms. Doguet introduced a broad concept

of the genus and arranged species into groups based on shape and ornamentation of the

ascospores. Recently this concept of Melanospora has been narrowed, and many genera

have now been separated from Doguet's broad treatment of the genus (Udagawa and Cain

1969, Hawksworth and Udagawa 1977, Cannon and Hawksworth 1982). Based on

Cannon and Hawksworth's (1982) study utilizing scanning electron microscopy,

Melanospora is now restricted to those fungi with smooth ascospore walls and depressed

germ pores.

Of all the many synonyms used for species ofMelanospora, the two most

commonly accepted have been Sphaeroderma Fuckel, and Microthecium Corda.

Sphaeroderma was distinguished from Melanospora by the absence of an ostiolar neck,

but this trait is considered to be highly variable (Cannon and Hawksworth 1982).

Microthecium was retained for cleistothecial species, but otherwise identical to

Melanospora (Udagawa and Cain, 1969, Hawksworth and Udagawa 1977). Scanning

electron microscopy studies have now confirmed that both cleistothecial and perithecial

species with reticulately ornamented ascospores constitute a distinct genus for which the

name Sphaerodes Clem. was available (Von Arx 1981). Because species ofMicrothecium

are characterized by smooth-spored ascospores with depressed germ pores, they are now

synonomized with Melanospora, while those with reticulated spores formally placed there

have been transferred to Sphaerodes (Cannon and Hawksworth 1982).

Sphaerodes was erected by Clements in 1909 to include S. episphaeria, and was

distinguished from Sphaeroderma by the lack of a subiculum (Clements and Shear 1931).

Besides the reticulated spores, Sphaerodes differs from Melanospora in the structure of

the germ pores. It has protruding, tuberculate pores, while Melanospora has pores either

level with spore surface or slightly sunken or depressed (Cannon and Hawksworth 1982).

Persiciospora is a newly created genus to retain Melanospora-like pyrenomycetes

with yet another type of ascospore morphology. It is distinguished from Melanospora's

smooth spores and Sphaerodes' reticulated ones by producing spores with delicately

pitted walls similar to a peach seed (Cannon and Hawksworth 1982).

Apart from ascospore structure and morphology, Melanospora and its allies,

Sphaerodes and Persiciospora, are similar in numerous respects. They are characterized

by superficial, nonstromatic perithecia composed of angular pseudoparechymatous cells.

Perithecia are spherical, amber to pale yellow in color and are solitary to gregarious.

Four-to eight-spored, clavate asci are soon deliquescent with spores filling the perithecial

cavity and exuding out of the ostiolar neck in a slimy cirrhus. Ascospores are dark, and

ellipsoid or oval-shaped. The yellowish, translucent perithecia appear black when filled

with mature ascospores. Ostiolate species have long or short beaks, often ringed with

hyaline setae of varying lengths. Centra are composed of thin-walled psuedoparenchyma

and lack paraphyses at maturity. Many species have additionally been suspected of being

parasitic on other fungi.

The primary reason for the extensive taxonomic changes for Melanospora is that

its characteristics are not consistently applicable to all the families and orders within which

it has been classified. It does not fit well within any of the ranks. It was placed in the

Hypocreales because of its fleshy, unpigmented perithecium, and phialidic anamorphs

(Rehner and Samuels 1995) Many species of this order, such asMelanospora, also have

clavate asci and a presumed ecological role as mycoparasites. However, Melanospora

differs from other members of the Hypocreales in its aseptate, dark ascospores with germ

pores at each end, and deliquescent asci. Melanospora also lacks the apical paraphyses or

pseudoparaphyses of the Nectria-type centrum development characteristic of this order

(Rehner and Samuels 1995).

There has never been a standard, universally recognized classification for the

Sphaeriales. It once included all pyrenomycetes, but the Hypocreales, Clavicepitales, and

Laboulbeniales were removed based on centrum development and ascal characteristics

(Luttrell 1951). The remainder constituted the single, heterogeneous order Sphaeriales,

which did not fit well within any other order of the Euascomycetes (Wehmeyer 1975).

Therefore, it is no longer recognized as an acceptable taxonomic rank. Members of this

old order were characterized by clavate or cylindric asci, that were borne in a hymenial

layer. Asci opened by a round, apical pore, and were forcefully ejected. They additionally

contained haemathecia consisting of varying numbers of paraphyses (Alexopoulos 1962).

Melanospora and its allies were considered members of this catch-all group probably

because of their varied, hybrid-like characteristics. Since they had to be placed

somewhere, this was considered to be the best location. They were similar in the

possession of clavate asci borne in a true perithecium, but were different by lacking the

paraphyses or poroid asci.

Luttrell (1951) placed Melanospora in the Diaporthales based on similar early

perithecial development. However, this group is different from Melanospora in almost

every other respect, including persistent, poroid asci with a refractive ring at the apex, and

ascal development at several levels within the perithecium (Alexopoulos et al. 1996).

The current placement ofMelanospora in the Sordariales is not without

inconsistencies either. This order is characterized by dark perithecia or cleistothecia with

tough, leathery peridia. Asci are produced in persistent, basal fascicles and usually have a

thin refractive apical ring, although ascus walls may be evanescent in some species. The

Sordariales also have both deliquescent paraphyses and psuedoparenchyma. Spores are

variable in color, shape, and ornamentation, and many species have germ pores.

Melanospora and its allies have a number of like characteristics, such as evanescent asci

and variation in spore morphology and ornamentation, but still defy placement

comfortably within this order.

Very little information exists concerning the biology and natural occurrences of

this group of fungi. In addition, extensive confusion has prevailed over the years

regarding the taxonomic placement of various genera within this ascomycete complex.

Therefore, this study was conducted in order to correctly identify and characterize a

number ofMelanospora-like pyrenomycetes isolated from production agriculture soils in

Texas and Florida. A secondary goal was to identify and reclassify isolates of similar fungi

that have been deposited in the University of Florida fungal herbarium.

Materials and Methods


During the fall of 1994, while surveying Texas sugar beet-producing counties for

isolates of the fusarial root rot pathogen, Fusarium oxysporum f. sp. betae,, approximately

10-15% of cultures became contaminated after two-three weeks with an unknown

pyrenomycetous ascomycete. Growth and vigor of the Fusarium isolates was severely

affected in vitro with the dual cultures. The consistent association with and observed

damage to the pathogen generated the concept of mycoparasitism and the utilization of

this ascomycete as a potential biological control agent for fusarial wilt. Several cultures

containing both organisms were dried and saved on sterile filter paper as previously

described, and stored at 40C (Harveson and Rush 1997).

Over the last 14 months, an additional five isolates closely resembling the Texas

sugar beetMelanospora were obtained from agricultural fields in Florida. All were

observed growing in dual culture with variousformae specials of Fusarium oxysporum

recovered from diseased plant roots. These isolates reacted like the one from Texas by

not becoming evident until after a period of several week's incubation. The recovered

isolates were given a two letter acronym and are illustrated in Table 2-1 along with their

provenance. The Texas sugar beet isolate was designated as TX, while the Florida

isolates were named FL, FP, LE, GI, and GI-2. FL was found in association with the

causal agent of Fusarium wilt of Canary Island date palm, F. oxysporum f. sp. canariensis.

Isolate FP was recovered from the tomato wilt pathogen, F. oxysporum f. sp. lycopersici,

while the remaining three were isolated from the watermelon wilt pathogen, F. oxysporum

f. sp. niveum (Table 2-1).

All isolates were separated from their respective hosts by one of two methods.

Individual ascospore cirrhi were picked off culture plates and streaked on potato dextrose

agar (PDA) or Mycological agar (MA). However, because of the difficulty of avoiding

the numerous Fusarium microconidia, this was not always successfully accomplished.

The other method entailed removing mature perithecia and shaking in two ml water blanks

amended with 50% lactic acid. After several minutes, the water was spread on PDA and

left at room temperature overnight. The following day, individual ascospores were

removed with the aid of a dissecting microscope and transferred to PDA or MA amended

with 100 m/1 of a filtrate from a three to four day old liquid culture of Fusarium. This

greatly enhanced the germination and growth of the ascomycetes (Chapter 3).

Morphological Characterization and Identification

Cultures produced from germinated ascospores were transferred to and observed

on 12 different growth media. Results of growth characteristics on these various media

and subsequent inferred nutritional requirements are included in Chapter 3. The

identification and correct taxonomic placement of these fungi was based upon current

protocol for the Ceratostomataceae by Cannon and Hawksworth (1982), and Horie et al.

(1986). This was accomplished by viewing and measuring the fungal morphological

Table 2-1. Isolates collected and used for taxonomic identification

Isolate Host* Geographic Taxonomic
Designation Origin Identity

FL FOCA Brevard Co., FL Sphaerodes retispora var. retispora

FP FOL Indian River Co., FL Sphaerodes retispora var retispora

GI FON Gilchrist Co., FL Persiciospora moreaui

GI-2 FON Gilchrist Co., FL Sphaerodes retispora var. retispora

LE FON Lake Co., FL Sphaerodes retispora var. retispora

TX FOB Randall Co., TX Melanospora zamiae

SFOCA = F. oxysporum f. sp. canariensis; FOL = F. oxysporum f. sp. lycopersici; FON =
F. oxysporum f. sp. niveum; and FOB = F. oxysporum f sp. betae

structures with both light and scanning electron microscopy. The shape, ornamentation,

and germ pore structure were central criteria for establishing the identity of these fungi.

In addition to the six isolates found in association with F. oxysporum, another

collection of isolates identified as Melanospora and deposited in the University of Florida

fungal herbarium were examined and reclassified according to modem taxonomic criteria.

These fungi included three isolates originally named as Melanospora sphaerophila, M.

ornata, and M. cameronse. TheM. sphaerophila isolate was collected originally in

Kittery Point, ME in 1906 in association with Plowrightia morbosa on branches of Prunus

virginiana (chokecherry). TheM. ornala isolate was found in deer dung from Cheboygan

Co., Michigan in 1969, while that ofM. cameronse was found in Gainesville, Florida

colonizing dead asparagus fern stems (Asparagus plumosus) in 1927.


On PDA, all isolates except GI produced hyaline to white mycelium. Colony

diameters ofFL, FP, LE, and GI-2 reached 80 mm after six-seven days. GI grew at the

same rate, but its mycelium was a creamy tan to straw color. Isolate TX was much

slower-growing, attaining growth approximately one half that of the others in the same

period of time. Mycelium on the agar surface was thick (4-8 grm) and verrucolose,

containing numerous bulbils (Fig.2-1). These were particularly evident with increasing

age of cultures. Aerial mycelium was much finer in structure measuring 0.5-2.0 pm in

diameter with phialids scattered profusely along hyphal branches (Figs. 2-2 and 2-3).

Most of the perithecia appeared to form within the branches of the aerial mycelium and

not as readily on the agar surface.

Hyphal growth on hemp agar was quite different than that on PDA. Aerial

mycelium was sparse, and closely appressed to substrate surface. Mycelial growth

occurred in an irregular, non-radial pattern, and perithecia were not as abundantly found

as on PDA due to the lack of extensive aerial mycelium.

Although all isolates were approximately the same size and possessed the same

morphological structures, several differences were also noted. FP was much more prolific

in the production of this phialidic anamorph than the others. TX had the least amount of

phialids, while the remaining four fungi were intermediate. In general, the production of

perithecia likewise, differed greatly among isolates. Isolates LE and FL tended to form

greater numbers of perithecia, while TX and GI produced the fewest. FP and GI-2 were

intermediate between these two extremes. GI additionally formed its perithecia more

often gregariously while the remaining taxa produced perithecia scattered more or less

evenly over the substrate.

Perithecia were superficial, globose to ampullaceus, and amber to yellowish-brown

in color. They formed fleshy, membranous walls composed of angular

pseudoparenchymatous cells. All taxa had clavate, eight-spored asci that soon became

deliquescent (Fig.2-4). Ascospores filled the perithecial cavity before being exuded out in

a slimy cirrhus. Ascospores produced by all taxa were dark brown to olivaceous in color

and ellipsoid to citriform in shape with two apical germ pores (Fig. 2-5). The light,

translucent perithecia appeared dark to the naked eye due to the spore mass inside. All

isolates were ostiolate with perithecial necks varying from 80-150 pm (Figs.2-6-2-11).

The presence and lengths of coronal setae varied depending on growth substrate.

Morphological structures ofMelanospora and allies
2-1. Hyphae at agar surface showing numerous bulbils (x 350)
2-2. Aerial hyphae showing anamorphic conidia or spermatia (x 350)
2-3. Close-up view of anamorphic state (x 500)
2-4. Squash mount ofperithecium depicting asci in various states of
maturity (x 500)
2-5. Germinating ascospore showing two polar germ pores (x 350)

Figures 2-1-2-5.

0 D% .


4 5a
-* : 1

Figures 2-6-2-11. Perithecia ofMelanospora and allies. Note that all are ostiolate.
2-6. Sphaerodes retispora var. retispora (isolate FL) (x 138)
2-7. Sphaerodes retispora var. retispora (isolate FP) (x 123)
2-8. Sphaerodes retispora var. retispora (isolate GI-2) (x 127)
2-9. Sphaerodes retispora var. retispora (isolate LE) (x 126).
2-10. Melanospra zamiae (isolate TX) (x 150)
2-11. Persiciospora moreaui (isolate GI) (x 141)




1- ..-

However, despite numerous similarities among isolates, there were some major

differences that are commonly used now to separate members of this group of fungi.

These include the structure of the germ pore and the morphology and ornamentation of

the ascospore. Spores of isolates LE, FL, FP, and GI-2 were all coarsely reticulate with

strongly apiculate germ pores. They contained 8-10 deep lumina on each face of the

spore (Figs. 2-12-2-15). The spores of TX were smooth with sunken or depressed germ

pores (Fig. 2-16). GI produced spores that were also smooth in outline, but differed

from the others by having finely pitted spore walls reminiscent of a peach seed, with

slightly apiculate germ pores (Fig.2-17). A comparison of morphological structures and

sizes among isolates is presented in Table 2-2.

Based upon all discussed taxonomic criteria, isolates LE, FL, FP, and GI-2 were

concluded to be Sphaerodes retispora var. retispora. TX was classified as the type

species forMelanospora, M. zamiae. GI was determined to be the third reported finding

for the newly created taxon, Persiciospora moreaui (Cannon and Hawksworth 1982)

(Table 2-1).

One of the three herbarium specimens was not able to be identified to species.

Melanospora cameronse was correctly named at the generic level due to its dark, smooth

ascospores with depressed germ pores (Fig. 2-18). However, the sample was in poor

shape and perithecia were seldom able to be observed. They disintegrated upon contact

with the cover slip and all that remained was a mass of spores in the place previously

occupied by the perithecium (Fig. 2-19). Since no immature perithecia could be located,

no asci could be examined. An extensive search of the literature did not reveal other

Table 2-2. A comparison of biological and morphological characteristics among Melanospora and its allies.

Specimen Genus Geographic Perithecium (pm) Ascus (pm) Ascospore (pm) Ostiole Substrate Spore
Origin Morphology

FL Sphaerodes Florida 240-320 35-50 x 18-22 18-22 x 8-12 Yes F oxysporum Reticulated

FP Sphaerodes Florida 280-320 35-55 x 18-22 20-22 x 8-12 Yes F oxysporum Reticulated

GI Persiciospora Florida 250-300 40-45 x 20-22 20-22 x 8-10 Yes F oxysporum Pitted

GI-2 Sphaerodes Florida 240-300 35-50 x 18-21 18-20 x 8-10 Yes F oxysporum Reticulated

LE Sphaerodes Florida 240-350 35-50 x 18-22 18-22 x 10-12 Yes F. oxysporum Reticulated

TX Melanospora Texas 230-300 40-65 x 20-25 16-18 x 12-14 Yes F. oxysporum Smooth

sphaeophila Scopinella Maine 250-375 Yes Prunus virginiania Smooth

ornata Sphaerodes Michigan 220-400 Yes Deer dung Reticulated

mer Mel a loid A s plu u Soo
camerense Melanospora Florida ? ? Asparagus plumosus Smooth

Figures 2-12-2-17. Mature ascospores for Melanospora and allies.
2-12. Sphaerodes retispora var. retispora (FL), note the reticulations
with 8-10 lumina on the face of each spore. (x 1500)
2-13. Sphaerodes retispora var. retispora (FP), note same structures.
(x 1476)
2-14. Sphaerodes retispora var retispora (GI-2), note same structures
(x 1578)
2-15. Sphaerodes retispora var. retispora (LE), note same structures.
(x 1550)
2-16. Melanospora zamiae (TX), note smooth spores with no
additional ornamentation. Lighter spore is immature. (x 1470)
2-17. Persiciospora moreaui (GI), note pores or pitted surface of
spore characteristic of this genus. (x 1309)


reports of similar fungi containing the specific epithet, cameronse. Therefore, the name

used for this specimen, M. cameronse, was retained. Conversely, the remaining two

specimens were easily identified and redesignated with their currently accepted generic

names. The specimen labeled as Melanospora ornata needed to be renamed. This taxon

had been synonomized with Sphaerodesfimicola (Cannon and Hawksworth 1982). The

sample easily fit into this category because of its eight-spored asci (Fig. 2-20), globose

perithecium (220-400 pm), very short neck and poorly developed coronal setae, and dark,

citriform spores with strongly apiculate germ pores and coarse reticulations (Fig. 2-21).

The other specimen from the herbarium that needed renaming was Melaospora

sphaerophila. It had been synonomized with and described as Scopinella sphaerophila by

Malloch in 1976. The herbarium sample proved to be identical with Malloch's description

(1976) by possessing black, carbonaceous perithecia measuring 250-375 pm in diameter

with very long necks (450-700 pm) (Fig. 2-22). Its smooth, barrel-shaped ascospores

contained a thick walled, dark brown band in the middle of the spore with hyaline,

protruding walls at each end that collapsed bilaterally after drying (Malloch 1976) (Fig. 2-

23). This feature was thought to be a vertical germ slit by Cannon and Hawksworth

(1982), but Scopinella was later proven to germinate from both lateral, hyaline

protuberances (Tsuneda et al. 1985). Its identification was further confirmed by the

presence of the unique, 2-spored asci (Malloch 1976, Tsuneda et al. 1985) (Figs. 2-24 and


Figures 2-18-1-25. Morphological characteristics from specimens deposited in University
of Florida fungal herbarium.
2-18. Mature ascospore ofMelanospora cameronse. Note the
smooth spore with no ornamentation (x 1500)
2-19. Ascospores ofM. cameronse after mounting. Clump of spores
are the only structures remaining after degradation of
perithecium (x 500)
2-20. Eight-spored, clavate ascus ofSphaerodesfimicola (x 404)
2-21. Mature ascospores of S.fimicola. Note the fine reticulations on
spore surface (x 1300)
2-22. Perithcium of Scopinella sphaerophila, note the long neck
(x 160)
2-23. Mature ascospores of S. sphaerophila (x 1375)
2-24. Immature asci ofS. sphaerophila arising from an ascogenous
hyphal system (x 1100)
2-25. Nearly mature ascus of S. sphaerophila. Note characteristic
two-spored nature (x 1535)



The difficulty in classifying allied members of the Ceratostomataceae has been well

documented. Horie et al. (1986) state that the reason for this fact is that superficially

similar ascomycetes that are placed in this group have all been derived from other

unrelated sources through convergent evolution. For example, the presence of the

deliquescent ascus, long necks, and ostiolar setae are adaptations for passive dispersal of

spores by animal movement or rain splashing, rather than active dispersal by air currents

via forcibly discharged ascospores and poroid asci. Horie et al. (1986) further point out

that the characters now utilized for taxonomic separation such as ascospore morphology

and ornamentation, and germ tube structure, are useful as indicators of true evolutionary

relationships since they would not be subjected to strong selection pressures.

Recently, the tremendous insurgence of analyses using molecular techniques and

characters and the resulting inferred phylogenies among pyrenomycetes has not

extensively addressed this group of fungi. Very few members of this group consisting of

Melanospora and its closely related allied genera have been included in these types of

studies. However, two recent ones have been attempted utilizing M zmniae, and M

fallax (Rehner and Samuels 1995, Spatafora and Blackwell 1994). In both cases,

Melanospora was derived within or near the Hypocreales, specifically in clades containing

the genus Nectria. Although Melanospora and its allies were once included in this order

by some workers, they have many distinctly non-Hypocrealean features. They are similar

to the Hypocreales in the enteroblasitc, phialidic anamorphs, and the fleshy, light colored,

true perithecia. However they differ substantially from hypocrealean fungi by containing

deliquescent asci, and dark, aseptate ascospores with polar germ pores.

In a separate study, inference with another group of fungi has been found. These

same two Melanospora species, M. zamiae and M fallax have additionally been aligned

phylogenetically with the anamorphic, ambrosial species Raffaela hennebertii (Jones and

Blackwell 1998). Members of this genus have been difficult to derive because of the lack

of a known teleomorphic state. However, when R. hennebertii was excluded from the

analysis, seven different species from this genus formed a lineage that was resolved as a

sister group with the sexual genus Ophiostoma (Jones and Blackwell 1998). These

modern taxonomic studies have therefore proven to be equally as difficult to discern and

understand as have the morphological ones.

As pointed out earlier, the primary reasons for the problems of taxonomic

placement for this group ofpyrenomycetes is shared characteristics, both morphological

and molecular, with other fungal taxa. This can be further illustrated with the example of

Sordaria destruens. This taxon was first referred to Anthostomella before being

transferred toMelanospora by Shear (Shear and Dodge 1927). It was used for almost

thirty years as the model for studying fungal physiology and nutrition before being

transferred to Sordaria in 1951 (Asthana and Hawker 1936, Hawker 1936, 1951). The

inclusion of this taxon in the literature for many years as Melanospora has likely

contributed to the confusion between this genus and its allied genera in subsequent

studies, as well as contributing to the incorrect placement of these species.



The phenomenon of enhanced sporulation of fungi by other organisms has been

observed for many years. The first account of increased reproduction of one organism by

another was recorded by Molliard in 1903. He found that species ofAscobolus growing

on carrots produced numerous apothecia in the presence of a contaminating bacterium,

but not when growing in pure culture (Molliard 1903). The first report of the

enhancement of sporulation by one fungus on another was made by Heald and Poole

(1909) in Nebraska. They found that Melanospora pampeana (syn. M zamiae) produced

perithecia abundantly in a mixed culture with Fusarium moniliforme, but only gave sparse

mycelial growth after being in pure culture (Heald and Poole 1909). A recent paper by

Watanabe (1997) has also demonstrated a similar process for Sordariafimicola and two

species ofArmillaria. In this case, A. mellea and A. tabescens were antagonistic to S.

fimicola by inhibiting its growth. However, at the same time, these Armillaria species

increased perithecium and ascospore production of Sordaria.

Many different fungi have also been demonstrated over the years to need certain

nutrients for adequate growth and sporulation in pure culture. A growth substance found

to be important for maximum culture of certain yeasts was identified in the late 1920's.

Eastcott (1928) was able to separate this substance into two parts, one of which was

identified as inositol. This compound is now known to be a hexahydroxycyclohexane that

occurs widely in the phospholipids of many microorganisms (Singleton and Sainsbury

1987). Robbins and Kavanagh (1938) showed that the addition of vitamin Bi (thiamin)

promoted the growth ofPhycomyces blakesleeans in synthetic culture. Using an isolate of

Melanospora destruens (syn. Sordaria destruens), Hawker (1938,1939) found that its

growth on a medium containing glucose and various inorganic salts was negligible. The

addition of thiamin had no beneficial growth effect on the fungus. Biotin (vitamin K)

added to the same medium, gave good vegetative growth, but no perithecial production.

However, when both biotin and thiamin were included, both reproduction and mycelial

growth were vastly improved.

Jordan and Barnett (1978) were similarly able to induce sporulation ofM. zamiae

better in dual culture with other fungi than by itself. They additionally attained greater

sporulation and hyphal growth with biotin and thiamin added to a glucose-based medium,

but discovered that the addition of zinc enhanced these properties to an even greater

extent. No growth of the fungus occurred without the inclusion of these nutrient

supplements. (Jordan and Barnett 1978). Conversely, Vakili (1989) found that the

majority ofMelanospora damnosa isolates produced from single ascospores would

germinate on PDA. His results determined that only 6% and 2% of the isolates required

thiamin and biotin respectively for germination and axenic growth.

A number of fungi have also been observed occurring and growing in close

association with other fungi. These types of relationships have also been noticed to be

necessary for the growth and sporulation of some members. Those members of these

complexes presumed to be parasitic were often not successfully grown in pure culture

(Jordan and Barnett 1978, Whaley and Barnett 1963). Gonatobotrys simplex was found

in association with Alternaria tennis and attempts to grow it axenically failed unless an

extract of the Alternaria was added to the medium (Whaley and Barnett 1963). A growth

factor was isolated and partially purified from this extract and termed mycotrophein. It

was assumed to be vitamin-like because it was only needed in minute amounts (Whaley

and Barnett 1963), but its nature is still not fully understood.

Gonatobotrys and several species ofMelanospora have also been demonstrated to

be fusion biotrophs of other fungi (Jordan and Barnett 1978, Jeffries and Young 1994,

Vakili 1989). The biological behavior of these types of parasites appears to often be

associated with a nutritional dependency (Jeffries and Young 1994). It is assumed that

whatever growth factor or vitamin-like molecule that is responsible for enhanced growth

and sporulation of these organisms in culture can be obtained directly from the host

hyphae in nature.

The collection of pyrenomycetes found naturally occurring in association with F.

oxysporum-infected plant roots (Chapter 2) may also be reacting in a similar manner.

Species of Sphaerodes and Persiciospora, like Melanospora, have been suspected as

being parasitic on other fungi (Cannon and Hawksworth 1982). It is not known, however

if any of them exhibited a similar dependency upon their hosts or host products for

improved sporulation and growth. Therefore, this study was initiated to explore this

possibility and to better characterize the association and interrelationships between

members of this fungal group and the variousformae specials ofF. oxysporum from

which they were originally found.

Materials and Methods

Since all pyrenomycete isolates were found occurring in dual culture with F.

oxysporum, it was first necessary to separate the two organisms and isolate them in pure

culture. Once separation had been accomplished, it was quickly observed how slow the

pyrenomycetes germinated and grew on a selected group of standard synthetic media.

Preliminary attempts to grow the isolates using different concentrations of a

Fusarium culture filtrate provided promise for increased sporulation and growth. These

experiments consisted of filtrates from four-five day old F. oxysporum cultures grown in a

liquid medium composed of sucrose, yeast extract, and inorganic salts (Esposito and

Fletcher 1961). Cultures were vacuum-filtered through 150 ml bottle top flasks (0.45 um

capacity, Coming, Coming, NY) directly into media bottles using a hand-held Nalgene

(Cucamonga, CA) vacuum pump. Concentrations of the filtrate consisted of 25, 50, and

100 ml per one half liter of media. These experiments also included the addition of filtrate

before and after autoclaving. Filtrates produced by the still culture from the fourformae

specials ofF. oxysporum, niveum, lycopersici, canariensis, and betae, were all evaluated

on PDA and mycological agar with the corresponding pyrenomycetes in all possible

combinations. Once methods and techniques for optimum efficiency were established, all

subsequent tests routinely included filtrate from F oxysporum f. sp. niveum (FON) at a

rate of 50 ml per one half liter of media added prior to autoclaving.

Effects of Media with and without the Addition of Filtrate

Isolates used for this study included the four Sphaerodes strains and the one isolate

of Melanospora (Chapter 2). Persiciospora moreaui was include in the preliminary

studies, but was eventually lost in culture and was never able to be reconstituted. It has,

however, been deposited in the University of Florida fungal herbarium and is available for

microscopic examination.

There were 12 different media employed for this portion of the study. They and

their ingredients are illustrated in Table 3-1. These media were used in to determine the

most efficient source for growth and reproduction of these fungi for eventual mass

production of inoculum. It was also hoped that inferences could be made on the

nutritional requirements or preferences for this group of pyrenomycetes.

An eight mm diameter cork borer agar plug was placed in the center of each plate

utilizing six replications per medium. Each medium was evaluated both with the added

filtrate (50 ml/1/2 liter) and without. Plates were incubated in the dark at 250C. Colony

growth measurements were made after six days. After measurements were recorded,

plates were further incubated for another two weeks.

After the two week incubation period, growth morphology and characteristics

were recorded. Four cork borer plugs (eight mm) were also removed approximately one

inch away from the point of plate inoculation at the 12, 3, 6, and 9 o'clock positions Each

plug was viewed under a dissecting microscope and perithecia were counted. Four of the

24 plugs utilized for counting perithecia were randomly selected, prepared as squash

mounts, and observed microscopically. With the use of 10X magnification, 12 perithecia

from each plug were randomly measured for diameter at the widest point.

Table 3-1. Media used for evaluation of growth, perithecial development, and perithecial
size for isolates ofMelanospora zamiae, and Sphaerodes retispora var. retispora.


AK sporulating agar (AK)

Brain infusion agar (BIA)

Cornmeal agar (CMA)

Emerson's YPS (EYPS)

Hemp agar (HA)

J&B agar

Komada's agar (KA)

Malt extract agar (ME)

Mycological agar (MA)

Potato dextrose agar (PDA)

Starch agar (SA)

V82 (double strength)


Pancreatic digest of gelatin, pancreatic digest of
casein, yeast extract, beef extract, dexstrose,
and MgSO4

Infusion of calf brains and beef hearts, proteose
peptone, dextrose, NaCI, Na2PO4

Infusion of cornmeal

Yeast extract, soluble starch, K2P04, MgSO4

Sterilized hemp seeds

glucose (dextrose), casein hydrolysate, MgSO4,
thiamin, biotin, microelement soln (Mn, Zn, Fe)

galactose, asparagine, K2PO4, KCL, MgSO4
Fe-Na EDTA, NaB407, PCNB, oxgall powder,
stretomycin SO4

maltose, dextrin, glycerol, peptone

soytone, dextrose

Infusion of potatoes, dextrose

Soluble starch, beef extract

V8 juice, CaCO

Factorial Study

This study was initiated to more fully explore and understand possible

interrelationships between each of the four Fusariumformae specials and their

corresponding fungal partner with which they were recovered. There were two primary

questions which were being addressed. One was whether there was an added benefit with

the inclusion of a filtrate above that achieved with the host Fusarium alone. The second

was whether evidence could be gathered that suggested if a host specificity or preference

was involved with eachforma specialist and the specific fungus found with it. The major

assumption being made with this approach was that the ability of the each fungus to

reproduce in conjunction with the Fusarium was an important factor in host-parasite

evaluations. Therefore, perithecial production was the central criterion that was used for

evaluating this process. Peripheral goals that could be achieved at the same time included

evaluation of medium effect, relative efficiency of parasites to sporulate, and efficiency of

formae specials to induce formation of perithecia.

This study was designed as a four-level factorial using four media, four

formae specials, four suspected parasites, and the inclusion or exclusion of filtrate in

media (Table 3-2). Each host and parasite were paired on each medium one half inch

apart in all possible combinations using five replications for each combination. Each

medium was also evaluated with filtrate added or excluded. After two week's incubation

in the dark, plugs were removed along two transect lines on each plate, four plugs spaced

equidistantly from each transect (Fig. 3-1). Perithecia were counted on the eight plugs as

before with the aid of a dissecting microscope.


Table 3-2. Media, pyrenomycete isolates and Fusarium oxysporum isolates utilized in the
factorial study.

Medium F. oxysporum forma specialist Parasite

Potato dextrose agar betae TX

Double strength V8 agar canariensis FL

Emerson's YPS agar lycopersici FP

Starch agar niveum LE

TX = M. zamiae
FL, FP, and LE = S. retispora var. retispora
Each host-parasite combination was evaluated on each medium in all possible
combinations, and each medium was tested both with and without the addition of a filtrate.

Figure 3-1. Plate inoculated with both F. oxysporum and pyrenomycete for evaluation of
perithecial development in the factorial study. Note the plugs removed along
two transect lines. Perithecial counts were made from each plug. Tiny black
structures are mature perithecia (x 0.8)


Both studies were repeated once, and all data were analyzed by the general linear

model procedure. The means involving growth, sporulation, and perithecial size in the

presence or absence of filtrate on different media were separated by Duncan's mean

separation test. In the factorial study, all possible interactions and main effects were

tested before separating means by the least squares procedure (LS Means). The pairwise

probability values obtained were additionally adjusted by the Bonferroni correction.

Filtrate Evaluation

Finally, two separate assays were attempted to establish some metabolic or

biological properties of the filtrate produced by the Fusarium isolates. Both involved the

use of dialysis tubing with a cut-off size of 14-16,00 MW (Spectra/Por, Spectrum Medical

Industries, Los Angeles, CA). Strips of tubing were taped to pieces of wire mesh and

autoclaved. These obstructions were placed in the center of Petri plates, dividing them in

half. Plugs of both parasite and Fusarium were inoculated on separate sides with the

parasite alone on separate plates as controls. Colony diameters of the parasite from both

treatments were measured after one week.

The other assay consisted of filling a six inch piece of tubing with 15 mls of PDA

broth and two plugs of the Fusarium. The ends were sealed and placed into sterile media

bottles containing 500 mis of PDA broth and two plugs of the parasite. This was

compared to bottles of liquid media containing only the parasite. After 10-14 days,

dialysis tubing was removed, contents were strained through cheesecloth and weighed.


After separation of isolates of the Melanospora and Sphaerodes from the F.

oxysporum, all isolates were evaluated after growing from a single ascospore. Because

they all grew and sporulated to some extent on the different media, they were concluded

to be homothallic and also not obligate parasites of the variousformae specials. From

the preliminary experiments, it was further determined that no differences existed in terms

of growth or perithecial production in response to the filtrates produced by the Fusarium

isolates. Because watermelon wilt was the pathosystem to be eventually evaluated for

disease control, filtrate from the Fusarium niveumforma specialist was routinely used for

all subsequent studies. It was also observed that few differences existed with the volume

of filtrate added to each medium. Therefore, a 1:10 ratio of filtrate to medium volume

was selected for all subsequent evaluations and general maintenance of isolates (50 ml/1/2


Linear growth after six days was significantly enhanced for all isolates, as was

perithecial production (Tables 3-3 and 3-4). However, the addition of filtrate had no

effect upon perithecial diameter (Tables 3-3 and 3-4). Significant differences were

observed among the three variables, and were influenced by choice of medium and

pyrenomycete isolate. These differences among growth, perithecial production and size

are illustrated in Tables 3-5 and 3-6 for combined media and isolates respectively.

With the exception of J&B, the better media for growth tended to be better for

sporulation as well. This medium consisted of glucose and inorganic salts amended with

biotin, thiamin, and a micronutrient solution (Mn, Zn, Fe), and was developed exclusively

for aM zamiae from West Virginia (Jordan and Barnett 1978). However, theM. zamiae

isolate used in this study from Texas did not respond well to this medium. In terms of

sporulation, none of the other Sphaerodes isolates in this study responded well either

(Table 3-5).

Although among the leaders at producing linear growth, J&B medium was extremely poor

in inducing perithecial production (Table 3-5). The media that overall allowed the best

combination of growth and sporulation were PDA, MA, AK, and ME (Table 3-5). These

media also tended to produce the greatest amount of fluffy, aerial mycelium, which is

where the majority of the perithecia were formed. Those media that did not induce overall

good, vigorous growth and perithecial formation included Komada, starch, Emerson,

hemp, and cornmeal agars. Brain infusion agar was similar to J&B by producing extensive

mycelial growth, but the numerous perithecia produced after two weeks were primarily

immature. Interestingly, the Komada medium allowed absolutely no growth at all for any

of the pyrenomycete isolates. This is a F. oxysporum-selective medium that truly worked

correctly by inhibition of the pyrenomycetes completely. It was determined that a plug of

any of these isolates could be placed on this medium for at least a month without

producing any growth. However, the isolates were not killed, because the plugs could be

removed from the Komada, placed on a suitable growth medium and afterward function


Those media having high soluble starch (starch, Emerson) or cellulose

content (V82, hemp, cornmeal) produced very little aerial mycelium. Growth was often

extensive linearly but very thin, sparse and flat to the agar surface. Perithecial

development was additionally poor (Table 3-5). In general, perithecial diameter was


Table 3-3. General linear model data for growth, perithecial production, and perithecial
diameter of pyrenomycete isolates with and without an added filtrate from F. oxysporum f.
sp. niveum.

Source df growth (mm) perithecia perithecial size (mn)

Form 1 0.0001 0.0001 0.8743*

Medium 11 0.0001 0.0001 0.0001

Isolate 4 0.0001 0.0001 0.0001

Replication 5 0.9254* 0.9481* 0.3830*

* = not significant at P = 0.01


Table 3-4. The effect of an added fusarial filtrate to growth, perithecial production and
perithecial diameter of pyrenomycete isolates.

Growth (mm) Perithecia Perithecial size (unm)

Filtrate 68.3 a 26.4 a 245.6 a

No Filtrate 38.9 b 10.6 b 245.7 a

Means followed by the same letter are not significantly different according to Duncan's
multiple range test
(P =0.01)

Table 3-5. Growth, perithecial production and perithecial size of all pyrenomycete isolates
combined on different growth media.

Medium Growth (mm)" Perithecia' Perithecial size (unm)

AK 49.0d 29.9c 224.8ef

brain infusion 56.7c 4.9f 220.8f

cornmeal 30.5f 10.9e 256.8b

Emerson's YPS 50.2d 14.1e 277.6a

hemp 47.4d 12.2e 251.2bc

J&B 57.2c 3.7f 244.0cd

Komada's 0.Og 0.0h 0.0h

malt extract 58.9c 20.2d 199.2g

mycological 72.3b 38.4b 234.4de

potato dextrose 79.7a 42.2a 231.2ef

starch 47.5d 11.6e 277.6a

V82 38.9e 12.6e 273.6a

Means followed by the same letter are not significantly different according to Duncan's
multiple range test.
(P = 0.01)


Table 3-6. Growth, perithecial production and perithecial diameter for each pyrenomycete
isolate on all combined growth media.

Isolate' Growth (mm)b Peritheciab Perithecial size (nm)b

FL 54.3b 28.3a 261.6a

FP 59.7a 20.6a 237.6b

GI-2 54.2b 20.5a 232.0b

LE 59.6a 22.8a 259.2a

TX 39.9c 5.7b 233.6b

"FL, FP, GI-2, and LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
bMeans followed by the same letter are not significantly different according to Duncan's
multiple range test.
(P =0.01)

likewise influenced by choice of growth media and isolate. Those that produced the

poorest growth and sporulation also tended to induce significantly larger perithecia (Table

3-5). Isolates FL and LE were observed to form larger, but not more numerous perithecia

than the other pyrenomycetes with the exception of the M. zamiae isolate from Texas

(Table 3-6).

Because of the significant differences detected among media and isolates, means of

the measured variables were sorted by fusarial filtrate effect and evaluated

individually. Tables 3-7 and 3-8 depict the effects of filtrate and medium upon the growth,

and sporulation, respectively, of the various isolates being evaluated in this study. Isolate

TX (M. zamiae) was consistently the slowest-growing and produced the fewest perithecia

among the other Sphaerodes isolates evaluated. This is particularly evident in Table 3-8

showing its production of perithecia without filtrate. In general, most media with the

added filtrate greatly improved both sporulation and growth, but ME was an exception.

With sporulation of isolates FL and LE, the difference between treatments was minimal.

Conversely, isolate FP actually produced more perithecia with the non-filtrate treatment

(Table 3-8). The linear growth produced by these isolates on PDA and mycological agars

could also be misleading. Although very little difference is noticed from looking at the

two treatments in Table 3-7, the filtrate was still better in inducing luxuriant, vigorous

aerial growth than was the same media without the filtrate. A more accurate picture of the

beneficial effect of filtrate is observed with the production ofperithecia (Table 3-8)

The factorial study provided further evidence of the importance or dependence of

these isolates upon F. oxysporum. After analyzing the perithecial production from the

different pyrenomycetes paired with the fourformae specials, it was determined that the

addition of filtrate to media is not a significant factor, like it had been with the

pyrenomycetes growing by themselves. The general linear model data obtained for main

effects and all possible interactions is presented in Table 3-9. Because all main effects

except for filtrate were significant, means from different treatments were separated and are

included in Table 3-10. After combining results for all fungal isolates, PDA was shown to

produce the greatest number ofperithecia followed by V82, starch and Emerson. Isolates

LE and FL formed greater numbers ofperithecia on all media andformae specials

followed by FP and TX. Finally, the sugar beet pathogen (FOB) was the poorest host for

all pyrenomycetes and media, whereas no differences were observed among the remaining

formae specials (Table 3-10).

The major questions being asked about the relationships between the

pyrenomycetes and the Fusariumformae specials were effectively answered by this

complicated factorial study. There were no significant differences observed with the

filtrate treatment, or with any interaction that included filtrate (Table 3-9). This essentially

answered the first question. Whatever advantage the pyrenomycetes received from the

filtrate was likewise obtained from the Fusarium isolates themselves, and there was no

added benefit to the filtrate in the media. However, there was a significant interaction

involving the pyrenomycete-F. oxysporum combination (Table 3-9), which alluded to the

second question being asked from these organismal relationships. Therefore, it was

important to investigate these relationships more completely with an additional analysis.

Table 3-11 shows the results of this additional analysis after subjecting data to the

least square means and pairwise comparison procedures. Isolate LE was the only one of

Table 3-7. Linear growth (mm) among pyrenomycete isolates after six day's incubation as
influenced by media and addition or exclusion offusarial filtrate.

Filtrate No Filtrate


AK 49.3c 71.3a 60.3b 62.3b 47.3c 32.3c 45.0a 39.7b 46.7a 35.3c

Brain 63.7b 73.7a 54.0c 69.0ab 47.0d 50.3b 66.0a 43.7c 66.7a 33.0d

CM 48.3a 46.3a 46.3a 47.3a 47.0a 14.0a 13.7a 15.Oa 13.7a 13.0a

Emerson 79.3b 85.0a 84.3a 85.0a 53.7c 21.0b 32.7a 32.7a 17.3c 10.7d

Hemp 70.0a 72.0a 71.7a 69.0ab 64.3b 29.0a 24.7b 24.7b 28.3a 22.7b

J&B 85.0a 85.0a 80.0a 85.0a 64.7b 30.7d 43.3b 35.3c 49.0a 14.0e

Komada O.Oa O.Oa O.Oa O.Oa O.Oa O.Oa O.Oa O.Oa O.Oa O.Oa

ME 76.7a 75.7a 72.0a 76.0a 55.7b 37.3c 54.0a 54.0a 48.0b 40.0c

Mycol. 85.0a 85.0a 85.0a 85.0a 69.0b 73.7b 74.3b 53.3c 79.3a 33.3d

PDA 85.0a 85.0a 85.0a 85.0a 85.0a 83.3a 83.0a 80.0a 84.0a 41.7b

Starch 68.3b 80.7a 81.0a 81.0a 49.7b 25.3b 22.7b 28.7a 24.0b 14.0c

V82 58.3b 51.7c 34.0d 68.0a 26.0e 29.3d 45.0a 32.2c 42.3b 12.0e

'FL, FP, GI-2, and LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
bMeans followed by the same letter are not significantly different according to Duncan's
multiple range test.

Table 3-8. Perithecial production among pyrenomycete isolates after two week's
incubation as influenced by media and addition or exclusion of Fusarium filtrate.

Filtrate No Filtrate


AK 44.7b 42.7b 37. lb

Brain 20.2a 2.7bc 5.0b

CM 21.2b 28.1a 29.1a

Emerson 20.5bc 29.3a 24.2ab

Hemp 21.lab 24.3ab 28.2a

J&B 7.3a 9.5a 0.0b

Komada 0.Oa 0.Oa 0.Oa

ME 40.2a 7.4c 24.7b

MA 66.6a 48.7b 28.2c

PDA 77.4a 67.la 84.la

Starch 22.4ab 23.5ab 27.9a

VS2 40.0a 27.7a 5.lb

61.0a 33.8b

18.5a 0.7c

19.7b ll.lc

19.0bc 17.7c

17.3ab 15.5b

8.5a 0.0b

0.Oa 0.Oa

26.3b 0.Oc

63.8a 15.1c

75.6a 1.7b

18.2b 17.5b

11.4c 26.5a 22.0ab 13.0bc 7.3c

0.9ab 0.7b 0.Oc 1.3a O.Oc

0.Oa 0.Oa 0.Oa O.Oa O.Oa

6.7b 10.5a 7.06 4.6b .l1c

1.5b 0.Oc 0.Oc 2.0a 0.Oc

0.0b 3.2a 5.0a 3.2a 0.0b

0.Oa 0.Oa 0.Oa 0.Oa 0.Oa

33.6c 22.9b 0.0d 25.1b 1.2c

41.5a 24.0b 49.2c 46.7a 0.0d

42.2b 51.7a 31.8c 42.6b 0.0d

1.4a 2.3a 1.5a 1.3a 0.0b

37.3a 2.4b 3.4a 4.5a 0.7b 4.7a 0.Oc

"FL, FP, GI-2, and LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
bMeans followed by the same letter are not significantly different according to Duncan's
multiple range test. (P= 0.01)

Table 3-9. General linear model data describing sporulation of four pyrenomycetes paired
with their four corresponding F. oxysporumformae specials.

Source df Pr > F

Medium 3 0.0001

Pyrenomycete isolate 3 0.0001

Filtrate 1 0.8866*

Fusarium isolate 3 0.0001

Replication 4 0.8492*

Filtrate*Pyrenomycete isolate 3 0.2249*

Fusarium*Pyrenomycete isolate 9 0.0001

Filtrate*Fusarium isolate 3 0.2699*

* = not significant at P = 0.01

Table 3-10. Factorial study data illustrating perithecial development for pyrenomycete
isolates, Fusarium isolates, and growth media.

Mean' Medium Pyrenomycete isolated Fusarium isolate'

63.2 PDA a

V82 b

Starch c

Emerson d





'Means followed by the same letter are not significantly different according to Duncan's
Multiple Range Test (P = 0.01).
bLE, FL, FP = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
FOL = F. oxysporum f. sp. lycopersici; FOCA = F. oxysporum f sp. canariensis; FON =
F. oxysporum f. sp. niveum; and FOB = F. oxysporum f. sp. betae


the four pyrenomycetes that was capable of producing a significantly greater number of

perithecia on its original host from the field. In fact, no other isolate was able to match

LE's sporulation efficiency on any of theformae specials. This indicated that this isolate

was better than the others at reproducing on this group of Fusarium isolates. Although

numerous other examples of differences among and between isolates were detected, there

were no further incidences of significant perithecial production associated with each

pyrenomycete and theforma specialist from which it was obtained. So this effectively

addressed the second question and proved that only the combination including isolate LE

and F oxysporum f sp. niveum appeared to exhibit a significant host specificity or

preference in terms of reproductive ability.

The experiments with the dialysis tubing provided little information on the

chemical nature of the filtrate, however a great deal was learned about its biological

properties. Whatever metabolite or growth factor that was being produced in response to

the culture of Fusarium was heat and pressure stabile. It was able to retain its activity of

enhancing growth and sporulation regardless of whether it was added to media before or

after autoclaving. Secondly, it had a molecular weight smaller than 12-14,000, which was

the cut-off size for the pores in the dialysis tubing. Both assays demonstrated that the

active ingredient in the filtrate could pass through the tubing and still maintain its

beneficial effect upon growth and perithecial production for the pyrenomycetes.


The reported nutritional requirements for Melanospora growth and sporulation have

proven to be conflicting over the years. In many cases this study's results do not correlate

with those of others. For example, Hanlin et al. (1993) and Goh and Hanlin (1994),

successfully utilized V8 juice agar for both culture maintenance and developmental studies

because of superior perithecial production over that of ME agar. TheM. zamiae isolate

originally found in Texas was more effective in terms of growth and reproduction on ME

than on VS. Perithecial production on VS by isolate TX was among the poorest of all

media evaluated, and also resulted in very slow, limited growth (Tables 3-5, 3-9, and 3-

10). However, V8 was among the better media for growth and sporulation of several of

the Sphaerodes isolates (Tables 3-9, and 3-10).

Another major difference with the M zamiae isolate from Texas used in this study

concerns the increased growth and sporulation efficiency after the addition of a filtrate

from a liquid culture of the suspected host. Jordan and Barnett (1978) found that their

isolate ofM. zamiae responded positively to a mycelial extract from F. roseum and

Verticillium albo-atrum. They therefore concluded that all nutrients required for growth

and sporulation were contained within the living host and not from metabolites or

excretions from a liquid culture.

There are also discrepancies among different studies involving required growth

factors. Much of Hawker's (1938, 1939, 1950) work with fungal nutrition has found that

vitamins or vitamin-like substances are often needed for optimum growth and sporulation

of certain fungi. She determined that M destruens (syn. Sordaria destruens) needed the

combination of both thiamin and biotin. In addition, another vitamin-like growth factor

termed mycotrophein was found to be required for Gonatobotrys simplex (Whaley and

Barnett 1963). Jordan and Barnett (1978) determined that their isolates ofM. zamiae also

Table 3-11. Results of the factorial study showing least square means and pairwise
comparison probability values for perithecial production among pyrenomycetes' and F.
oxysporumformae speciales.

Pr > mT| Pr > [T Pr > 1T Pr > ITl

FOB-FL 41.5 0.0001 0.0004 0.7676* 0.0001
FOB-FP 53.1 0.9427* 0.4890* 0.0001 0.0001
FOB-LE 42.9 0.0002 0.0022 0.4267* 0.0001
FOB-TX 52.9 0.5352* 0.0001 0.0001

FOCA-FL 51.2 0.5352* 0.0001 0.0001
FOCA-FP 39.0 0.0001 0.0001 0.5111* 0.0001
FOCA-LE 64.3 0.0001 0.0001 0.0001 0.2430*
FOCA-TX 52.3 0.8311* 0.6841* 0.0001 0.0001

FOL-FL 53.1 0.9252* 0.4753* 0.0001 0.0001
FOL-FP 40.8 0.0001 0.0001 0.0001
FOL-LE 67.1 0.0001 0.0001 0.0001 0.8757*
FOL-TX 46.3 0.0147* 0.0683* 0.0417* 0.0001

FON-FL 55.2 0.3855* 0.1370* 0.0001 0.0001
FON-FP 49.3 0.1890* 0.4876* 0.0016 0.0001
FON-LE 67.5 0.0001 0.0001 0.0001 -
FON-TX 29.4 0.0001 0.0001 0.0001 0.0001

FL, FP, LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
bFOB = F. oxysporum f. sp. betae; FOCA = F. oxysporum f sp. canariensis, FOL = F.
oxysporum f. sp. lycopersici; FON = F. oxysporum f sp. niveum.
*= not significant after use ofBonferroni correction (P = 0.0031)

needed biotin and thiamin, in addition to supplemental zinc, but were capable of

synthesizing their own mycotrophein. They developed a medium specific for M zamiae

that was glucose-based and supplemented with biotin and thiamin. It proved to be

excellent for inducing maximum growth and sporulation. The current study additionally

evaluated this same medium and referred to it as J&B (Table 3-1); however, it was still

very poor in production of perithecia for all isolates. It was also among the worst in

producing vigorous linear growth, unless supplemented with the fusarial filtrate.

Conversely, Hanlin et al. (1993) and Goh and Hanlin (1994) obtained adequate results

without the addition of these growth factors. Vakili (1994) found that most of his isolates

ofM damnosa could germinate without added vitamins, but single ascospores cultures

did not produce the phialidic, anamorphic state, Gonatobotrys simplex. However,

mycelium from compatible crosses ofM. damnosa did produce both microconidiophores

and perithecia. He concluded that the absence of the phialids in single-spored cultures and

their presence in crosses showed that they were spermatial in function and related to

sexual reproduction. The isolates used in this study still were capable of germinating and

growing without the extra nutrients although very poorly. They also formed both an

anamorphic stage and perithecia from single ascospores, proving that they were


Most physiological studies have found that nearly all fungi can utilize simple sugars

such as glucose or fructose equally as well as carbon sources (Cochrane 1958, Hawker,

1950, Griffin, 1981). These simple, monosaccharides appear to be more suitable for many

fungi than the more complex polysaccharides. In some instances the more complex sugars

seemed less able to support hyphal growth but may be more conducive to reproduction

(Hawker 1950). Galactose is another carbon source that is apparently difficult for a

number of fungi to utilize (Cochrane 1958, Hawker 1950, and Griffin 1981). Hawker

(1950) determined that M. destruens and several species of Chaetomium were unable to

grow vigorously with a galactose-based medium. However, the combination of good and

bad carbon sources had a overall beneficial effect. For example, the addition ofgalactose

to a glucose medium had no detrimental effect upon these fungi (Hawker, 1950). Perhaps

a similar phenomenon explains the effectiveness of PDA as a growth and perithecial

promoter for this group of pyrenomycetes. The combination of dextrose and starch would

not be as inhibitive as starch alone, and the beneficial effect of dextrose may override that

of the more complex sugar.

Several other previously published observations coincide well with the results from

this study. The best media overall for the current group of pyrenomycetes tended to

contain glucose or dextrose (PDA and MA), whereas the poorest contained more complex

carbohydrates such as cellulose or starch (cornmeal, starch, and Emerson). Griffin (1981),

points out that those fungi capable of utilizing complex carbohydrates such as starch or

cellulose were also able to synthesize exocellular, digestive or hydrolytic enzymes that

more readily broke the sugars down into glucose or maltose. Melanospora and the

closely related Sphaerodes and Persiciospora are suspected of being fusion, biotrophic

parasites. Organisms with this type of life strategy have additionally been determined to

obtain nutrition directly through contact with their hosts without producing any cellulitic

enzymes (Jeffries and Young 1994). Sordaria and Chaetomium, which were found to be

unable to use galactose, (Hawker 1950) are both closely related toMelanospora and

Sphaerodes. The current study concluded that Komada's agar would not allow growth of

any kind from these latter two genera. Komada's medium (Komada 1975) employs

galactose as the sole source of carbon (Table 3-1), and perhaps it is partially this

ingredient and not exclusively the antimicrobial supplements that are responsible for its

continued, successful use as a F. oxysporum selective medium.

Anamorphic fungi like Gonatobotrys, Gonatobotryum, and Phialophora have

often been linked with teleomorphic pyrenomycetes such as Melanospora. Melanospora

and related genera such as Sphaerodes and Persiciospora have also been found to be

ecologically associated with other fungi. Because of these close associations, many have

also been presumed to be contact parasites based upon the formation of hook cells or

contact branches on their hyphae. These types of fungi are also recognized to be slow-

growing or incapable of growing in axenic culture. They have also been demonstrated to

grow or sporulate better in mixed cultures with other organisms. Several other examples

have shown enhancement of these characteristics if extracts of suspected hosts are added

to media (Whaley and Barnett 1963). The current study discovered a very similar process

occurring with this group of pyrenomycetes and a filtrate produced by F. oxysporum. The

discovery of vitamin-like growth factors that improve growth and reproduction are

additionally assumed to be obtained from host fungi in nature or in synthetic media

insufficient for growth without added nutrients. This type of behavior has often been

associated with fusion biotrophs like Melanospora, but was completely unknown for allied

groups like Sphaerodes.

Although these nutritional relationships have been reported in the literature,

particularly with respect to Melanospora, they are very few in number. No other study

has attempted as extensively to address the aspect of enhancement of growth and

reproductive characteristics utilizing suspected hosts or host products. These

observations have been documented from several fungal complexes, but it is still not

known how common these types of interactions are outside the lab. Perhaps this type of

nutritional dependency between some fungi may explain why some can be easily cultured,

while others appear to be obligate.

This study has attempted to add new information that addressed Meanospora and

other related genera specifically with F. oxysporum. These particular organisms and their

interactions have been chosen for a number of reasons. First, these fungal complexes have

been consistently found occurring together in several geographically different locations

(Chapter 2). Secondly, because of the worldwide importance of Fusarium oxysporum and

the diseases it causes, it is desirable to find novel strategies for disease management.

Therefore the primary idea behind this chapter was to investigate these naturally-occurring

relationships and learn how to efficiently mass produce the pyrenomycetes in the attempt

to identify potential candidates for use as biological control agents for Fusarium wilt of

watermelon. Chapter 4 will focus on the establishment of host ranges and the evaluation

and measurement of damage to F. oxysporum at the bench level.



The concept of parasitism can be defined or subdivided into a number of

categories. Parasitism usually refers to a relationship between two organisms where one

benefits at the expense of the other. However, it is not always correlated with damage or

deleterious effects upon one member. Parasitic relationships can also include mutualism,

in which both organisms receive equal benefit, or commensalism, where one organism

benefits from the relationship, leaving the other unaffected.

Melanospora, Sphaerodes, and Persiciospora have all been found in conjunction

with other fungi or from soil in field isolations (Cannon and Hawksworth 1982). All have

additionally been suspected of being parasitic based upon these intimate, natural

associations. Of these three genera, only M zamiae has been formally investigated as a

mycoparasite. Due to the formation of hook or contact cells and the ability to sporulate

on living hyphal mats ofFusarium roseum, Tritirachium sp., and Verticillium albo-atrum,

M. zamiae was concluded as being parasitic (Jordan and Barnett 1978). A collection of

11 other fungi, including F. oxysporum f sp. lycopersici, were also concluded to be

parasitic hosts for M zamiae based solely on the hook-shaped hyphal branches formed in

culture with this group of fungi. No other methods of evaluation were attempted for this

second group of fungi.

It has been fully established that all three genera being used in this study were

capable of more efficient growth and sporulation in dual culture with F. oxysporum or

with some of its metabolic products in the form of a liquid culture filtrate (Chapter 3)

however, this was not clear proof of parasitism. Therefore this study was initiated to

thoroughly establish in a more quantitative manner ifM. zamiae is a true mycoparasite.

Sphaerodes retispora var. reispora had previously only been isolated from soil or from

several different polypores (Cannon and Hawksworth 1982). Persiciospora moreaui has

been additionally isolated from soil, and in association with severalformae specials of F.

oxysporum (Cannon and Hawksworth 1982, Horie et al. 1986, Krug, 1988). However,

neither of the latter two pyrenomycetes have ever been tested as parasites, thus they have

been included along with M. zamiae in these evaluations. In addition, no other study has

ever attempted to show pathogenicity or damage caused by mycoparasites on their hosts

at the bench level.

The ultimate goal to this project was to find an alternative management option for

Fusarium wilt of watermelon. This chapter was another step in the process of selecting

potential candidates for use as biological control agents. Through the use of the

pyrenomycete isolates found to be naturally occurring in the field with F oxysporum, this

chapter has attempted to demonstrate parasitism at the cultural level and to measure the

amount of damage that these isolates cause to F. oxysporum. In addition to the

establishment of parasitism and pathogenicity, another purpose of this chapter was to

define a host range for these fungi among a collection ofF. oxysporumformae specials,

based upon the ability to sporulate in dual culture.

Materials and Methods

Fusarium oxysporum Host Range Determination

A collection ofF. oxysporum isolates were utilized to determine host range for the

Melanospora and Sphaerodes strains. This collection consisted of both pathogenic

isolates from various hosts, and non-pathogenic strains from a watermelon wilt-

suppressive soil (Table 4-1). For this portion of the study, only isolates TX, FL, FP, and

LE were included. Evaluations were made by monitoring perithecial production as

described in Chapter HI. Briefly, this entailed pairing eight mm agar plugs from actively

growing cultures of Fusarium with cultures ofMelanospora and Sphaerodes utilizing six

replications per treatment. After two weeks' incubation in the dark at 250C, four plugs

were removed from each replication from the 12, 3, 6, and 9 o'clock positions at a

distance of approximately two and one half cm from paired isolates. Perithecia were then

enumerated from each plug with the aid of a dissecting microscope.

Two separate analyses were performed for the pathogenic and non-pathogenic

isolates. Both data sets were sorted by Fusarium host, and analyzed by the general linear

model procedure. Tests were repeated once, and means were separated by Duncan's

multiple range test. The set consisting of the non-pathogenic isolates also included one

pathogenic isolate for comparison.

Establishment of Parasitism

Four separate methods were tested in order to provide proof of parasitism on F.

oxysporum. They included slide cultures, an assay using Komada's medium (KA)

Table 4-1. Fusarium oxysporum isolates collected from Florida and used to establish host
range among isolates ofMelanospora and Sphaerodes.

Isolate Isolation Host Plant Pathogen



Komada (1975) and PDA, an assay using Fusarium hyphal mats, and an assay using

watermelon seeds previously coated with both watermelon wilt pathogen and suspected


These studies consisted of Sphaerodes isolates FL, FP, and LE, Persiciospora

isolate GI, and Melanospora isolate TX. The host being evaluated was restricted to the

Fusarium wilt pathogen of watermelon (FON). All experiments were repeated at least


Slide cultures consisted of small (one mm2) agar pieces taken from actively

growing cultures of both FON and suspected mycoparasites. They were mounted in water

on sterile slides one half inch apart, and incubated in sterile, glass humidity chambers at

250 C. They were observed daily under a microscope for ascospore germination and

formation of the hook and contact cells attaching to hyphae of Fusarium, as described in

the literature (Jordan and Barnett 1978, Whaley and Barnett 1963).

Watermelon seeds (cultivar Charleston Gray), previously coated with a suspension

of mycoparasite spores and FON were allowed to air dry in petri plates at room

temperature. After 17 months, seeds were placed on moist paper towels in humidity

chambers and incubated at 250 to observe growth of fungi and to additionally establish

viability and survival over time of the mycoparasites.

For the hyphal mat assay, cultures of FON were grown in still liquid culture for 10-

14 days. After this time, hyphal mats were removed and washed thoroughly. The mats

were then placed on moist paper towels in humidity chambers and inoculated with agar

plugs from the different suspected parasites. These plates were incubated at 250 C and

observed periodically for growth and perithecial formation of parasites.

Lastly, a novel assay was developed that included the use of KA, one of the

standard, universally-employed F. oxysporum-selective media, and PDA. Previous work

had clearly demonstrated that Melanospora, Sphaerodes, and Persiciospora were all

completely inhibited on this medium, whereas PDA was one of the better media for

growth and sporulation (Chapter 3, Harveson and Kimbrough 1998). A sterile 40 mm

diameter beaker was placed in the center of a petri plate and molten PDA was poured

around the outside. After the agar solidified, the beaker was removed, leaving a 40 mm

diameter hole in the middle ringed by PDA. The hole was gently filled with KA and

allowed to harden. Agar plugs of FON and all suspected parasites were paired on the

center portion of plates containing the KA and incubated at 250 C (Fig. 4-1).

Measurement of Deleterious Effects of Parasites on F. oxysporum

Three separate techniques were developed to not only provide evidence of

deleterious effects on FON by parasites at the cultural level, but more importantly, to

measure the damage inflicted by Melanospora, Sphaerodes, or Persiciospora. The

measurement of damage in vitro to Fusarium was accomplished with several categorical

assays, including hyphal weight reduction, growth inhibition, and reduction of aerial


The hyphal weight reduction tests were performed in large petri plates (100 x 150

mm) using liquid culture (PD broth). A standard volume of 15 mis was put in each plate

along with eight mm agar plugs of various Fusarium and parasite treatments. The isolates

utilized for this study consisted of the four primaryformae specials evaluated in Chapter

Assays to determine parasitism of Fusarium oxysporum
4-1. Assay plates showing both parasite and Fusarium co-inoculated in
center portion of plates containing the Komada medium. Outside
ring contains PDA (x 0.6)
4-2. Germinating ascospores showing attachment of parasite germ
tubes on Fusarium hyphae (x 500)
4-3. Germinating ascospores showing attachment of parasite germ
tubes on Fusarium hyphae (x 500)
4-4. Hyphal strands of parasite attaching to Fusarium hyphae and
forming contact cells (arrows) (x 1100)

Figures 4-1-4-4.


3, isolates FL, FP, LE, and TX and the corresponding pyrenomycete found in association

with eachforma specials, FOCA, FOL, FON, and FOB, respectively. Each Fusarium

forma specials was grown by itself which served as the control, and with each of the four

suspected parasites in all possible combinations. Each treatment consisted of five

replications. After five days' incubation at 250 resulting hyphal mats were filtered through

cheesecloth, squeezed free of liquid and weighed. Dry weights were additionally obtained

after air-drying hyphae for several days in petri plates. Data from hyphal weights were

analyzed by the general linear model procedure and means were separated by Duncan's

multiple range test.

Reduction of aerial hyphae was evaluated on F. oxysporum f sp. niveum only.

The various parasites were paired six cm apart on PDA and MA plates utilizing six

replications per assay. These plates were compared to FON growing singly on the same

media. Linear growth of FON for all medium treatments was measured daily for 12 days.

Collected data were analyzed as a repeated measures experiment and additionally

subjected to linear regression using the simple regression model: y = a(trt) + b(obs), where

the y equaled growth response and a and b were parameters to be estimated.

The growth inhibition assay was evaluated in a similar manner to that described

above for aerial hyphal reduction. This assay consisted of the same isolates, media, and

replications, only the mycoparasite was first put on plates to grow singly for three days.

After this time, FON was placed at the edge of the parasite colony. Linear growth of

FON was then measured as before on a daily basis. Data were similarly analyzed as a

repeated measures experiment before being subjected to linear regression. This assay

employed the quadratic model: y = a(trt) + b(obs) + c(obs2), where y = growth response

and a, b, and c were parameters to be estimated. Both of these two assays were repeated

once and results were separated by the least square means procedure. Pairwise probability

values obtained were additionally adjusted with the Bonferroni correction.


The results of the host range determinations are presented in Tables 4-2 and 4-3. Table 4-

2 shows the production of perithecia by Melanospora and Sphaerodes in conjunction with

the pathogenic isolates of F. oxysporum. There was no one parasite isolate that

statistically predominated over the others with respect to reproductive ability. Numerous

differences were observed between and among isolates with all four parasites being

represented at least once as being the best to sporulate on a fusarial isolate. Overall,

isolates LE and FL tended to produce higher numbers of perithecia than did FP or TX.

The data in Table 4-3 illustrate a similar story regarding differences among and between

parasite isolates and their abilities to efficiently produce perithecia on the non-pathogenic

isolates. However, large differences were observed in total perithecial production

between pathogenic and non-pathogenic isolates. There were no Fusarium isolates

among those tested that did not support at least some sporulation from the parasites.

With the exception of 96-14, sporulation for all parasites was much poorer on the non-

pathogenic isolates. This can be easily visualized by observing the one pathogenic isolate

analyzed with the rest in Table 4-3, and by also comparing overall values between Tables

4-2 and 4-3. The recognition that isolate 96-14 was capable of supporting good

perithecial production was an important observation and will be implemented and

discussed further in Chapter 5.

Table 4-2. Results of host range study for pathogenic isolates ofF. oxysporum using
isolates ofMelanospora and Sphaerodes

Fusarium isolate Parasite Isolate'


1443 60.lb 59.7b 72.6a 43.2c

1453 70.2a 33.7c 46.9b 0.0d

1480 49.lc 34.4d 72.2b 82.3a

1481 43.2bc 36.6c 51.9b 68.3a

1742 59.9a 53.6a 60.2a 54.4a

1778 64.0a 51.6a 56.1a 51.7a

1779 63.1a 67.0a 66.3a 41.2b

2095 33.5a 32.3a 40.2a 4.3b

2099 49.1c 59.7bc 77.9a 70.lab

CL58 57.3b 41.1c 75.2a 78.1a

GD8 65.0a 58.7a 63.0a 37.9b

MN21 83.0a 66.5b 82.la 52.8b

MN27 78.9a 66.1a 79.1a 43.lb

NFL1 70.la 63.1a 60.9a 34.8b

*FL, FP, and LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
average number of perithecia produced by parasite isolates
'Means followed by the same letter are not significantly different according to Duncan's
multiple range test (P = 0.01)

Table 4-3. Results of the host range study for non-pathogenic isolates ofF. oxysporum
using isolates ofMelmospora and Sphaerodes.

Fusarium Isolate Parasite Isolate"


101196 4.0bb.' 33.7a 29.6a 15.0b

3196-1 26.6a 34.9a 3.0b 8.0b

31196-1 26.2a 34.5a 25.9a 35.5a

96-10 22.0b 24.5b 47.6a 2.0c

96-14 72.5a 79.1a 88.6a 87.5a

96-17 26.9a 34.6a 36.1a 12.6b

96-2 23.7a 25.5a 15.4a 19.3a

96-5 24.9a 21.1a 18.1a 16.5a

96-6 18.7ab 5.0b 26.5a 7.0b

96-7 9.7b 29.9a 32.1a 24.2a

FO95-5* 76.0b 64.7b 96.5a 44.2c

'FL, FP, and LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
average number of perithecia produced by parasite isolates
*= pathogenic isolate ofF. oxysporum used for comparison
'Means followed by the same letter are not significantly different according to Duncan's
multiple range test (P = 0.01)

The slide cultures for all isolates produced similar results to those in the literature

describing mycoparasitism byM. zamiae. The isolates of Persiciospora and Sphaerodes

reacted in an identical manner. All three genera germinated and formed contact cells or

hooks on their hyphal strands that appeared to make contact with mycelial branches of the

Fusarium isolates. The ability to form these specialized branches was apparently not host

specific for each parasite and its corresponding field host. Eachforma specialist appeared

to be equally receptive to any of the parasites that were evaluated. These phenomena of

germination and attachment of ascospores and hyphae are demonstrated in Figures 4-2, 4-

3 and 4-4.

The watermelon seeds coated with both organisms simultaneously produced the

same results. After several weeks, perithecia were observed on the outer edges of the

paper towels in the humidity chambers, but no fusarial growth could be visualized.

Perithecia were removed from the periphery of the chambers and plated on media.

Resulting growth of both organisms indicated that the parasites were colonizing FON as it

grew saprophytically on the paper towel substrate. The hyphal mat assay produced the

same results. The isolates of Sphaerodes and Melanospora sporulated profusely on the

FON mats only, but not on the paper towels in the humidity chambers. The controls for

both assays consisted of parasite isolates placed singly on paper towels in chambers, and

did not result in perithecial formation. Therefore, it was concluded that the Sphaerodes

and Melanospora isolates were growing on and parasitizing FON due to the lack of any

additional nutritional source in the humidity chambers.

The assay consisting ofKA and PDA provided the final and perhaps most

convincing evidence for the parasitic abilities of these fungi with F. oxysporum. All three

genera were evaluated against FON for this test. Representatives of each (LE-

Sphaerodes, TX-Melanospora, and GI-Persiciospora) are illustrated in Figures 4-5-4-7.

In all cases, scattered perithecia were formed after several weeks on the central portion of

the plates containing the KA however there were differences in the ability to sporulate

with Sphaerodes forming the most perithecia followed by Melanospora and

Persiciospora. The outer ring of PDA in all examples consisted of a layer of perithecia

greater than that on the KA (Fig. 4-8). Komada agar was previously shown to inhibit

growth and sporulation for these isolates. However, perithecia formed on this normally

restrictive medium in conjunction with FON, and reproduction also was observed at points

removed from the initial inoculation site on the portion containing PDA. Therefore, it was

concluded that the most logical explanation for these events was the occurrence of

parasitism of Fusarium by the isolates of Sphaerodes, Melanospora, and Persiciospora as

it grew outward from the center.

The attempt to provide evidence for damage to F. oxysporum in culture was first

done by measuring the total hyphal weights of the variousformae specials alone and in

combination with the different parasites. Table 4-4 demonstrates the reductive effect on

Fusarium, but it includes only the fourformae specials and the corresponding parasites

originally found with them from the field. Although these evaluations were done with all

parasites and allformae specials in all possible combinations, there were not many

differences between the treatments containing both fungi. However, the important point

to consider is that both fungi together produced hyphal weights that were significantly

Figures 4-5-4-10. Results of Fusarium parasitism and pathogenicity assays.
4-5. Sphaerodes retispora var. retispora (LE) parasitizing FON.
Note heavier concentration of perithecia in outer ring of PDA.
(x 0.67)
4-6. Melanospora zamiae (TX) parasitizing FON. Note heavier
concentration of perithecia in outer ring of PDA, and the fewer
number of perithecia overall compared to that ofLE. (x 0.67)
4-7. Persiciospora moreaui (GI) parasitizing FON. Note the fewer
number of perithecia overall compared to both LE and TX.
(x 0.68)
4-8. Close-up view of parasitism of FON. Shows scattered perithecia
in center portion of plate with heavier concentration on outside
ring of PDA. Arrow depicts the line separating the two different
media. (x 0.58)
4-9. Plate showing the reduction of Fusarium aerial hyphae. Single
arrow depicts the final advance ofFON before being reduced.
Double arrows shows the extent of FON growth when picture
was taken (x 0.83)
4-10. Plate showing the inhibition ofFusarium hyphae. Note the clear
line of inhibition where the parasite stops FON growth (x 1.0)

Sri i1

/" 8

Table 4-4. Reduction of total hyphal weights of oxysporumformae specials by each
parasite originally found with them.

Treatment" Wet weight (g) Dry weight (g)

FOCA 0.83 ab 0.145 ab

FOB 0.75 b 0.135 b

FON 0.73 bc 0.13 bc

FOL 0.69 c 0.12 c

FOCA+ FL 0.42 d 0.075 d

FON + LE 0.33 ef 0.057 ef

FOB + TX 0.32 ef 0.055 ef

FOL + FP 0.31 ef 0.055 ef

'FOCA = F. oxysporum f sp. canariensis; FOB = F oxysporum f sp. betae; FON = F.
oxysporum f. sp. niveum; FOL = F. oxysporum f sp. lycopersici
FL, FP, and LE = Sphaerodes retispora var. retispora; TX = Melanospora zamiae
bMeans followed by the same letter are not significantly different according to Duncan's
multiple range test (P = 0.01).

lower than any of theformae specials alone. This demonstrates that these four parasites

were pathogenic and caused measurable damage to the particularforma specialist with

which they were isolated. Each also induced highly significant weight reductions to the

otherformae specials as well (data not shown), but very few differences were observed

between these different parasite treatments.

The hyphal reduction test did not initially look very promising. Until about the

sixth day of taking measurements, the fusarial growth on all parasite treatments was equal

to that of the FON control. However, after this time, the hyphal growth of FON began to

recede and appeared to melt away. This occurred on the aerial growth only, one could

still see the mycelium on the agar surface. As the hyphae of Fusarium receded, the

parasite growth advanced in the same direction and perithecia were readily formed in the

process (Fig. 4-9). Table 4-5 exhibits the average growth of FON after 12 days for each

treatment. Isolate LE caused the most reduction to FON, but was no different statistically

than isolates FL or FP. Because these three treatments were statistically different than

that of FON alone, it was concluded that they significantly reduced the amount of aerial

mycelium compared to the control. This was interpreted to mean that they were

pathogenic to FON. The remaining two isolates, GI and TX, were not different

statistically than FON. Therefore, they were not causing distinguishable damage to FON

by the reducing aerial hyphae.

The knowledge gained from these prior observations led to the development of

another, slightly different assay. The growth reduction test entailed growing the parasites

on media for three days prior to the addition of FON. The final results were similar to

Table 4-5. Reduction of aerial hyphae ofF. oxysporum f. sp. niveum by Persiciopsora,
Melanospora, and Sphaerodes.


Treatment" LSMean (mm) Pr >m Pr > I Pr > Pr > I Pr> IT Pr > I

FON 47.5 0.0001 0.0001 0.2868* 0.0001 0.2868*

FL 41.6 0.0001 0.7147* 0.0001 0.0063* 0.0001

FP 42.0 0.0001 0.7147* 0.0001 0.0025* 0.0001

GI 48.6 0.2868* 0.0001 0.0001 0.0001 0.0342*

LE 38.9 0.0001 0.0063* 0.0025* 0.0001 0.0001

TX 46.6 0.2868* 0.0001 0.0001 0.0342* 0.0001

'FON = F. oxysporum f sp. niveum; FL, FP, and LE = Sphaerodes retispora var.
retispora; GI = Persiciospora moreaui; TX = Melanospora zamiae
* = not significant after use of the Bonferroni correction (P = 0.0016)

Table 4-6. Growth inhibition ofF. oxysporum f. sp. niveum by isolates of Persiciospora,
Melanospora, and Sphaerodes.


Treatment' LS Mean (mm) Pr > Trr Pr>m Pr > TIPr Pr>IT P > IT| Pr>

FON 50.8 0.0001 0.0001 0.0024* 0.0001 0.0001

FL 47.3 0.0001 0.0955* 0.0001 0.0031* 0.7202*

FP 46.4 0.0001 0.0955* 0.0001 0.1898* 0.1898*

GI 49.0 0.0024* 0.0045* 0.0001 0.0001 0.0014

LE 45.6 0.0001 0.0031* 0.1898* 0.0001 0.0091*

TX 47.1 0.0001 0.7202* 0.1898* 0.0014 0.0091* -

FON = F. oxysporum f. sp. niveum; FL, FP, and LE = Sphaerodes retispora var.
retispora; GI = Persiciospora moreaui; and TX = Melanospora zamiae
*= not significant after use of the Bonferroni correction (P = 0.0016)

the aerial hypahal reduction test, but were more visually spectacular (Fig. 4-10). The data

from this assay are presented in Table 4-6. Like the previous test, isolate LE was the most

aggressive and effective in halting fusarial growth. However, it was again no different

statistically from isolates FL or FP. The major difference between these two inhibition

assays was observed with isolate TX. Unlike its performance in the aerial reduction

TX was effective in inhibiting FON, and was statistically insignificant from the pathogenic

isolates LE, FL, and FP (Table 4-6).


Jordan and Barnett (1978) tested over 20 different fungi, and found that 11 were

judged to be hosts forM. zamiae by their evaluation method of observing contact cell

formation. This group included the tomato wilt pathogen, F. oxysporum f sp. lycopersici,

but no study has ever exclusively evaluated many different form species ofF. oxysporum.

There has also never been any type of parasitic host ranges established for Sphaerodes.

The results obtained in this study for host range determination were based upon the ability

to sporulate on the variousformae specials. They revealed a large number of differences

among and between the parasites and their reproductive abilities. This type of information

is not really very surprising, because variation in biological characteristics is a common

phenomenon in mycology.

Chapter 3 found that the parasites differed significantly in the ability to sporulate

on different media in the absence ofF. oxysporum. It further demonstrated that

production of perithecia was better for all parasites in the presence ofF. oxysporum. In

general, it was shown that M.zamiae tended to sporulate less than Sphaerodes, regardless

of whether F. oxysporum was present or not (Chapter 3). This is an interesting finding in

light of the fact that the species of Sphaerodes studied had never been found with F.

oxysporum (Cannon and Hawksworth 1982). However, after a larger number offormae

specials was tested, M. zamiae proved to have some clear host preferences in terms of

producing perithecia. These included isolates 1480 and 1481 (Table 4-2). It additionally

was among the poorest at reproducing on others like 1443, 1453, and most of the non-

pathogenic isolates (Tables 4-2 and 4-3). This suggests that ability to sporulate by these

groups of fungi is not simply a matter of some being able to and some not, but that other

factors like the presence of an adequate host also plays a role. It also suggests that ability

to sporulate may be an excellent criterion to initially judge parasitism and host-parasite

relationships. The host range study has also demonstrated that no one parasite is going to

be effective as a biological control agent for every F. oxysporum. There will be a need for

testing potential biological agents in a similar manner as this in order to determine the

most likely candidate.

This study has brought into question the validity or accuracy of establishing

parasitism by relying strictly on the formation of the specialized contact hyphal branches of

the suspected parasites. Although this is obviously an important finding, there should be

other methods tried in combination to completely address this question. This chapter has

combined the traditional technique with newly developed ones in the attempt to provide

more conclusive evidence for parasitism among this group ofpyrenomycetes on F.

oxysporum. Taken together, these methods have strongly indicated that the fungi found

occurring naturally with F. oxysporum in the field have the capacity to function as


Another important feature of parasitism has further been discovered. Just because

an organism may sporulate well on, or be able to parasitize its host, this is not always a

good indication that it will be an effective agent for biological control. All of the genera

from this study were demonstrated to be parasitic on the watermelon wilt pathogen but

there were different levels of damage caused by these organisms. Persiciospora moreaui

was isolated in conjunction with FON (Chapter 2), and was shown to be parasitic in all

tests in which it was evaluated. However, the pathogenicity assays (reduction of aerial

hyphae and growth inhibition) indicated that deleterious effects upon FON by P. moreaui

were negligible (Tables 4-5 and 4-6). This suggests that it functions as a mutualistic or

commensalistic parasite, but not as a pathogen.

This chapter has been the first attempt to demonstrate damage in culture by

mycoparasites on their hosts, and has shown that F. oxysporum is an excellent host for

Sphaerodes retispora var retispora. Sphaerodes is actually more aggressive and harmful

on Fusarium than the previously-studied parasite, M. zamiae. This also represents the

first definitive proof of parasitism for both Sphaerodes and Persiciospora, as well as the

first report ofP. moreaui in the western hemisphere.

The next chapter will focus on the selection of one of these parasites as a candidate

for biological control based upon all information gathered throughout the entire study. It

will then describe the methods and techniques used for evaluating the ability to reduce

disease severity on watermelon by the Fusarium wilt pathogen at the greenhouse level.



The Fusarium wilt and root rot diseases induced byformae specials ofFusarium

oxysporum are among the most severe plant diseases in the world. They cause significant

economic losses to a wide range of species and occur on plants belonging to all families

except the Graminaceae (Alabouvette et al. 1996).

These fungal diseases are some of the oldest recognized and among the most

difficult to manage as well. Most control methods available today are either inefficient or

difficult to apply. Chemical fungicides are ineffective because of the location of the

pathogen in the plant's vascular system. Fumigation can be successful, but is erratic and

often cost-prohibitive, depending upon the value of the crop. The beneficial effects it does

produce are only temporary and must be reapplied each season due to the pathogen's

propensity to rapidly recolonize soils (Marois and Mitchell 1981). Short term crop

rotation is of little value because F. oxysporum can colonize non-host plants and survive

as chlamydospores for long periods of time. Therefore, genetic resistance is generally

considered the be the most effective strategy for managing wilt diseases. However, this

can be a difficult procedure if no known resistance genes are available, or if many different

pathogen genotypes are present, necessitating the need for germplasm containing multiple

resistance genes.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E3MVANVPI_URYD40 INGEST_TIME 2013-01-22T14:29:39Z PACKAGE AA00012994_00001