Louisiana State University LSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1992 The olecM ular Systematics of Unitunicate, Perithecial Ascomycetes. Joseph William Spatafora Louisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses

Recommended Citation Spatafora, Joseph William, "The oM lecular Systematics of Unitunicate, Perithecial Ascomycetes." (1992). LSU Historical Dissertations and Theses. 5469. https://digitalcommons.lsu.edu/gradschool_disstheses/5469

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contact [email protected]. INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information Company 300 North Z eeb Road. Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600

Order Number 9317001

The molecular systematics of unitunicate, perithecial ascomycetes

Spatafora, Joseph William, Ph.D.

The Louisiana State University and Agricultural and Mechanical Col., 1992

UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106

THE MOLECULAR SYSTEMATICS OF UNITUNICATE, PERITHECIAL ASCOMYCETES

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Doctor of Philosophy

in

The Department of Botany

by Joseph William Spatafora B.S., Louisiana Tech University, 1986 December 1992 ACKNOWLEDGEMENTS

Although the work of this dissertation is attributed to one individual, it would not have been possible without the assistance, guidance and friendship of many. Furthermore, I have been most fortunate to participate in one of the better interdisciplinary groups studying systematics and evolutionary biology in the country.

I am deeply indebted to my major professor, Dr. Meredith Blackwell, beyond the limits of words. She has provided me with unparalleled guidance in research and has always shown unwavering patience to my relentless questions.

Her generous support regarding research and travel removed much of the stress and anxiety associated with the financial reality of competitive research.

I would like to thank my labmates including Elly Van Eeckhout, Steve

Cassar and Brandye Sawyer for their support and friendship. I am especially grateful to Brandye who has been an undergraduate worker in Dr. Blackwell’s lab for the past three years and has assisted me beyond measure in day-to-day lab work.

Additionally, I would like to thank people that I have collaborated with over the years including Keith Hamby, Bernhard Kaltenboeck, Eldon Jupe,

Rick Zechman, Tom Kantz, Mark Buchheim, Deborah Waters and all of the

Botany graduate students and postdocs. Also, I want to thank the graduate students in the Museum of Natural Science, especially Shannon Hackett, John Bates and Jim Demastes for their friendship and stimulating rhetoric throughout the years at Molecular Evolution seminars and Friday afternoon beers at Library Joe’s.

Gratitude is also extended to members of my committee including Drs.

Russel Chapman, William Platt, Robert Zink and Sue Bartlett. They have offered me sound advice concerning research and professional issues.

There are many people outside of LSU who have contributed substantially to my graduate career either through direct assistance or intellectual fervor. These include Gary Samuels, Bud Uecker, Stephen Rehner,

Amy Rossman, John Taylor, Mary Berbee, Tom Bruns, Tom Harrington, David

Malloch, Jack Rogers, Gerald Benny and Clark Rogerson. Also, I would like to acknowledge the National Science Foundation for their financial support through the Doctoral Dissertation Improvement grant BSR-9101088.

None of this would have been possible without my family to whom I owe everything. My mother and father, Billie and Gene Spatafora, have provided me with wisdom and unconditional love and support throughout life.

I would not have made these gains in life without them. Also, my wife and best friend, Elizabeth Rossi Spatafora, has been there for me through the often stressful times of preparing this dissertation. But I thank her most for our daughter, Anna, whose mere smile overshadows any importance these pages may hold.

i i i TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF TABLES...... vi

LIST OF FIGURES ...... viii

ABSTRACT ...... xii

CHAPTER

1 INTRODUCTION ...... 1 Literature C ite d ...... 15

2 THE CLAVICIPITALES-HYPOCREALES CONNECTION ...... 19 Materials and Methods ...... 21 Results ...... 25 D iscussion...... 30 Literature C ite d ...... 40

3 THE POLYPHYLETIC ORIGINS OF QPHIOSTOMATALEAN FUNGI ...... 44 Materials and Methods ...... 48 Results ...... 50 D iscussion...... 51 Literature C ite d ...... 64

4 THE NONMONOPHYLY OF UNITUNICATE PERITHECIAL ASCOMYCETES ...... 68 Materials and Methods ...... 70 Results ...... 75 D iscussion...... 78 Literature C ite d ...... 92

5 THE EVOLUTION OF CENTRUM DEVELOPMENT IN DERIVED UNITUNICATE PERITHECIAL ASCOMYCETES AS INFERRED BY CLADISTIC ANALYSIS OF THE NUCLEAR-ENCODED SMALL SUBUNIT RIBOSOMAL D N A ...... 95 Materials and Methods ...... 98 Results ...... 98

iv D iscussion...... 101 Literature Cited ...... I l l

6 SUMMARY AND CONCLUSIONS ...... 115 Literature Cited ...... 122

VITA ...... 123

v LIST OF TABLES

Table 2.1 List of taxa included in study, source of taxon and regions sequenced. Primers are those of White et al (1990)......

Table 2.2 Bootstrap confidence levels for pairs of clavicipitalean taxa from 100 bootstrap replications. C. paspali is paired with all species sampled for the order. Note that C. capitata (fungal parasite) and A. take (endophyte, Japanese isolate) did not receive bootstrap confidence levels of greater than 9 0 % ......

Table 3.1 Nonmolecular characters of conidiogenesis, antibiotic sensitivity, cell wall carbohydrate composition and production of asci that have been used to separate Ceratocystis s.s. and Ophiostoma H. & P. Sydow......

Table 3.2 List of taxa included in study, source of taxon and regions sequenced. Primers are those of White et al (1990)......

Table 4.1 List of taxa included in study, source of taxon and regions sequenced. Primers are those of White et al. (1990)......

Table 4.2 Degenerate, inosine-containing primers for amplification and sequencing of the B subunit of vacuolar ATPase. Primers lBvATP and BvATP8 amplify a 750 bp fragment that corresponds to positions 1803-2552 ofN. crassa (Bowman et al, 1988). Primer BvATP7 anneals to the noncoding strand at positions 2436-2455. Degenerate nucleotide codes: N=ACGT, H=ACT, D=AGT, R=AG and Y=CT. 1. Amino acid sequence. 2. Degenerate nucleotide sequence. 3. Primer sequence ...... Summary of Cl, RI and RC for the two ssrDNA and the two BvATPase data sets. N = number of characters. RCs decrease with the addition of new taxa (ssrDNA) and new characters (BvATPase). The RC decreased most for the BvATPase data. Note that the RI did not decrease in the ssrDNA data did in the BvATPase. The results suggest that the third position of the codon is characterized by an inordinate amount of variation with respect to character states and character state distribution ......

v i i LIST OF FIGURES

Figure 1.1 (a) Schematic diagram of a cladogenic event resulting in species B and C. (b) Venn diagram of overlapping characters and character states for members of species B and C. Character 4 is in conflict with characters 1 and 2. Character 3 defined B and C as sister taxa. (c) Cladogram that represents the hierarchical correlation of characters 1-4 with respect to species B and C ...... 12

Figure 2.1 Strict consensus tree of 42 equally most parsimonious cladograms containing 40 taxa (474 steps, Cl = .530, RI = .766, RC = .406). The designated outgroup is T. deformans and the Endomycetales. Branch lengths are proportional to the number of steps...... 27

Figure 2.2 Single cladogram from one round of successive approximaton using the maximum fit option. Characters were reweighted based on their rc from the initial parsimony analysis. This cladogram is one of the 42 equally most parsimonious cladograms inferred by the unweighted data. Branch lengths are distorted by the reweighting of characters and are not indicative of the number of steps ...... 28

Figure 2.3 Strict consensus of eight equally most parsimonious cladograms (396 steps, Cl = .591, RI = .780 and RC = .461). Bootstrap confidence levels are given above those nodes that were supported in 90% or more of the trees produced from 100 bootstrap replications. Corresponding decay indices are presented below their respective nodes ...... 29

Figure 2.4 Bootstrap confidence levels for different pairs of clavicipitalean taxa derived from 100 bootstrap replications. The Clavicipitales is an unresolved monophyletic clade in the

v i i i maximum parsimony analysis. Bootstrap confidence levels for pairs of taxa suggest that homoplasy is distributed nonrandomly across the inferred monophyletic group...... 31

Figure 2.5 Enlargement of successive approximation clade containing the Hypocreales and Clavicipitales. Nutritional modes are mapped on the tree. Emphasis is placed on the trend from parasite to endophyte. *Nectria and other hypocrealean genera contain species that are parasitic on fungi ...... 38

Figure 3.1 Strict consensus of eight equally most parsimonious cladograms containing 33 taxa (481 steps, Cl = .530, RI = .736, RC = .390). The designated outgroup is T. deformans and the Endomycetales. Bootstrap confidence levels of 90% or greater from 100 bootstrap replications and decay indices are given above and below the corresponding nodes, respectively. Branch lengths are proportional to the number of steps. (*See Fig. 3.2; Fig. 3.3)...... 52

Figure 3.2 Strict consensus of eight equally most parsimonious cladograms (424 steps, Cl = .559, RI = .769 and RC = .430). S. fimicola has been excluded from the analysis. Note that the bootstrap values are elevated significantly in subclade A (Fig. 3.1), and the clade containing the Microascales and Ceratocystis (Fig. 3.3). The dampening effect of S. fimicola on decay indices appears to be more localized with respect to subclade A (Fig. 3.1) ...... 53

Figure 3.3 Bootstrap confidence levels for three different taxon samplings within the Microascales-Ceratocystis clade. (a) 5 taxa withoutS. fimicola, (b) 6 taxa with S. fimicola and (c) 5 taxa with S. fimicola. Support for the clade is contingent on the inclusion or exclusion ofS. fimicola.

ix Branch lengths are not proportional to the number of steps...... 54

Figure 4.1 Strict consensus tree of 42 equally most parsimonious cladograms containing 54 taxa (898 steps, Cl = .427, RI = .702 and RC = .299). The designated outgroup is the Basidiomycetes. Specified ordinal rankings follow Hawksworthet al. (1983)...... 76

Figure 4.2 Bootstrap confidence levels are mapped on the single most parsimonious cladogram of 37 taxa (689 steps, Cl = .489, RI = .696 and RC = .340). Bootstrap confidence levels are given for the nodes supported in 90% or more of the trees from 100 bootstrap replications. Corresponding decay indices are given below their respective nodes ...... 77

Figure 4.3 Characters of ascomata, evanescent unitunicate asci and location of asci within the ascomata are mapped on the strict consensus tree. Unitunicate, perithecial ascomycetes are nonmonophyletic. (* = evanescent asci; B = asci basal, S = asci scattered, BS = asci basal becoming scattered; circles denotes cleistothecial ascomata)...... 81

Figure 4.4 Strict consensus tree of 153 equally most parsimonious BvATPase trees inferred from the first and second nucleotide positions of the codon (63 steps, Cl = .571, RI = .691 and RC = .341). The topology of the resolved nodes is in agreement with the ssrDNA gene trees (Fig. 4.1; Fig. 4.5) ...... 84

Figure 4.5 Strict consensus tree of four equally most parsimonious ssrDNA gene trees containing 14 taxa (239 steps, Cl = .649, RI = .690 and RC = .448). The topology agrees with the 54 taxa ssrDNA tree (Fig. 4.1) except that the Xylariales and Sordariales, and the Hypocrealcs and Clavicipitales clades are unresolved between one step grades or clades (data not shown)...... 85

x Figure 4.6 Strict consensus tree of two equally most parsimonious BvATPase trees inferred from an unweighted analysis of all three nucleotide positions of the codon (487 steps, Cl = .478, RI = .371 and RC = .177). The topology does not agree with the ssrDNA gene tree (Fig. 4.1; Fig. 4.5), or the topology of the BvATPase tree inferred from only the first and second positions of the codon (Fig. 4.4)......

Figure 5.1 Strict consensus of 42 equally most parsimonious cladograms containing 54 taxa presented in Chapter 4 (898 steps, Cl = .427, RI = .702 and RC = .299)......

Figure 5.2 Centrum types as defined by Luttrell (1951) are mapped on the strict consensus tree. The Xylaria and Ophiostoma centrum types do not define monophyletic groups as inferred by cladistic analysis of partial sequences from ssrDNA......

Figure 5.3 Enlargement of the clade comprising derived unitunicate, perithecial ascomycetes. Centrum development is mapped on the tree denoting the genus and investigator of the centrum types that are congruent with the ssrDNA gene tree. * Genera were not included in centrum development ...... ABSTRACT

Maximum parsimony analysis of partial sequences from the small subunit ribosomal DNA (ssrDNA) rejects the monophyly of unitunicate, perithecial ascomycetes (pyrenomycetes). The generaKathistes, Pyxidiophora and

Subbaromyces are separated from a large clade of relatively derived pyrenomycetes. This derived pyrenomycete clade is supported by a bootstrap confidence level of 100% and a decay index of > 10 steps, and contains two subclades, A and B. Subclade A comprises the orders Clavicipitales,

Hypocreales, Microascales and the genus Ceratocystis; subclade B consists of the Diaporthales, Sordariales, Xylariales and the genus Ophiostoma. The

Clavicipitales is a sister taxon to the Hypocreales, which consists of two paraphyletic lineages. The ophiostomatalean taxa sampled do not form a monophyletic group; the data support the removal ofOphiostoma H. & P.

Sydow from Ceratocystis Ellis and Halst. as proposed by several investigators.

The Xylaria and Ophiostoma centrum types of Luttrell (1951) are inconsistent with the monophyletic groups inferred from the ssrDNA. Hypotheses concerning the evolution of centrum development are refined in light of these data. An additional data set was derived from the B subunit of vacuolar

ATPase (BvATPase) to provide an independent test of the ssrDNA gene tree and morphologically based hypotheses. Analysis of the first and second position of the codon support the separation of the generaPyxidiophora and

Subbaromyces from the larger clade of relatively derived pyrenomycetes.

x i i Inclusion of the third position of the codon does not support the separation of these genera from the larger clade of relatively derived pyrenomycetes or relationships within the derived clade that are mutually supported by the ssrDNA and morphology. Over 95% of the third codon positions sampled were variable, which constituted 76% of the synapomorphies of the BvATPase data set. Retention indices (RI) and rescaled consistency indices (RC) of the two

BvATPase data sets suggest that the third codon position is too variable with respect to frequency of characters state variation and distribution of character states for use in the systematics of the taxa sampled. The nonmonophyly of unitunicate, perithecial ascomycetes is suggested to be the better of two hypotheses. CHAPTER 1

INTRODUCTION

Members of the Kingdom Fungi are a large and diverse group of organisms estimated to number over one million species (Hawksworth, 1991).

Four major divisions are recognized in modern taxonomic treatments: the

Chytridiomycotina, Zygomycotina, Basidiomycotina and Ascomycotina (Barr,

1992). The largest and most diverse of the four divisions is the Ascomycotina, commonly referred to as the ascomycetes.

Ascomycetes are considered by most mycologists to form a natural, monophyletic group. Modern ascomycetes are found in the fossil records from the Cretaceous, 145 million years ago (MYA) (Pirozynski, 1976). Possible ascomycete ancestors have been described from the Triassic, 250 MYA (White &

Taylor, 1988) and may date back to the Silurian, 430 MYA (Sherwood-Pike and

Gray, 1985). Ascomycetes are cosmopolitan in distribution and occupy a diversity of habitats. They are represented in arctic, temperate, tropical and antarctic regions of the world, and they have evolved an array of nutritional modes. They are associated with living plants, other fungi, insects and animals as parasites. They may be saprophytic on wood, organic litter, soil and dung. Many ascomycetes, and fungi in general, participate in symbiotic relationships with insects, plants, animals, cyanobacteria and algae. It is estimated that half of the species of ascomycetous fungi interact with a cyanobacterium or alga to form a lichen (Alexopoulos and Mims, 1979).

1 2

Ascomycetes are important to human beings for a number of economic

reasons. They are plant pathogens of many of our staple grains, fruits and vegetables. Also, they are pathogens of animals and human beings; for example,

an increasing number of species are being recognized as opportunistic pathogens of persons with compromised immune systems. Nevertheless, their use in medical applications has had an immeasurable, beneficial impact upon humanity. The discovery of antibiotics produced by many ascomycetes has removed the mortal threat inflicted by an array of pathogenic bacteria. They are also important in production of food, vitamins, organic acids, etc. Yeasts are essential in the preparation of leavened breads, cheeses and other foods and alcoholic beverages around the world. Some ascomycetes, such as morels and truffles are considered by many to be delicacies of unparalleled taste.

The ascomycete life cycle is essentially haploid; a short-lived diploid phase is present only prior to ascospore cleavage. Ascomycete life cycles often comprise both a sexual, or perfect, stage (teleomorph) and an asexual, or imperfect, conidial stage (anamorph). For some taxa, however, only the sexual or the asexual stage is known. Because classification schemes are based primarily upon the sexual state, the lack of an asexual state poses no significant phylogenetic problem. However, the absence of a known sexual state does have negative consequences for our understanding of the phylogeny of many ascomycetes. Some ascomycetes possess conspicuous stromata from which sexual and asexual reproductive structures are produced. Others lack a stroma, or develop only a diminuitive stroma, and produce reproductive structures from a less specialized mycelium.

The ascus is the defining character of ascomycetes. A species of fungus that possesses asci is placed within the ascomycetes regardless of the presence or 3

absence of any other character(s). Asci are sac-shaped structures that are the site

of karyogamy, meiosis and ascospore formation. They may be one of several

shapes ranging from ovoid to clavate to cylindrical. Moreover, the ascus wall(s)

may be persistent or may deliquesce prior to ascospore maturity. They may have

one or two functional wall layers termed unitunicate and bitunicate, respectively.

Asci may occur within an ascoma or not. In the absence of an ascoma, asci are

produced on a mycelium or in non-mycelial forms by conversion of a single

vegetative cell. The absence of an ascoma is most common among yeasts.

There are four basic types of ascomata in which asci may occur. Apothecia are cup-shaped ascomata that produce asci in a distinct hymenial layer

and usually discharge their ascospores forcibly. The second type is the

cleistothecium, a completely closed fruiting structure with asci most commonly

scattered throughout the cleistothecium. The third type of ascoma is a

perithecium. These are flask-shaped structures with an opening, or ostiole, through which ascospores are discharged either forcibly or passively from a basal

hymenium, or from scattered locations within the perithecium. The final type, a

pseudothecium, may have the appearance of a cleistothecium or a perithecium;

however, the development is different from a true ascoma. True ascomata are produced from hyphae that are consigned to the sexual reproduction system, whereas pseudothecia are derived from somatic, or stromatic, hyphae.

Furthermore, unitunicate asci are produced within a true ascoma, and bitunicate asci are usually associated with a pseudothecium.

The majority of the taxa discussed in this study bear perithecial ascomata, which produce the ascogenous system from transformed hyphal cells termed initials. Ascogonial initials may commence development as a simple ascogonium, an ascogonial coil, an ascogonium and antheridium, or an ascogonium and 4

trichogyne. The ascogenous system becomes surrounded by an envelope of

hyphae that develops from the base of the ascogonial coil or ascogonium,

adjacent hyphae, or both. The envelope produces the nonascogenous portion of the perithicium including the outer wall (the peridium) and inner nonascogenous

(sterile) hyphae (cells). These sterile cells differentiate into paraphyses,

periphyses and other hyphal components; sterile cells are considered to play a

major role in formation of the centrum cavity. The entirety of the perithecium, asci and ascogenous cells, paraphyses, periphyses, and other sterile cells

excluding the peridium, is designated the centrum. Sterile cells may or may not

be present; if they exist, they may vary in production and morphology according

to species.

There is extensive diversity among ascospores with respect to morphology, formation, dispersal and germination. Among species they range from spherical

to filiform, and may be unicellular to multi-septate. Typically, they are produced

eight to an ascus but may vary in number from one to several hundred according

to the species. The difference in composition and ornamentation of ascospore walls results in some of the more conspicuous characters of species or groups of

taxa. Ascospores may be dispersed by wind, water, air, insects and other animals.

Dispersal is considered to exert a strong selection pressure on ascomycetes and,

therefore, similar morphologies in different groups are thought to be the product

of convergent or parallel evolution. Also, ascomycetes display a variety of germination strategies among their spores,i.e. germ tubes, slits or pores.

The characters described above are the principal ones historically

employed by mycologists in constructing phylogenetic hypotheses for ascomycetes. Earlier in this century a change in use of particular morphological characters was proposed. This shift was championed by Nannfeldt (1932) and 5

Luttrell (1951). They proposed de-emphasizing stromatic characters and

accentuating the development of the ascoma, its centrum anatomy and the

number of functional layers that comprise the ascus wall. Their ideas and

hypotheses, as well as others, will be discussed further throughout this

dissertation.

The phylogenetic study presented here focuses on relationships among

ascomycetes that produce unitunicate asci within perithecial ascomata. These organisms are commonly referred to as pyrenomycetes. This study tests rivaling

phylogenetic hypotheses, which are based on morphological and developmental

characters by employing independent molecular data sets. These data sets

comprise nucleotide characters derived from the nuclear-encoded small subunit ribosomal DNA (ssrDNA) and the B subunit of vacuolar ATPase (B vATPase)

genes. Both genes are distributed ubiquitously throughout eukaryotes, and

represent a multitude of characters available for comparison across a broad

taxonomic range.

The use of molecular characters in studies of systematics has led to a rejuvenation of the discipline. Molecular biology has provided a multitude of

characters for systematic analysis that are independent of morphology.

Variability in the rates of evolution throughout different regions of the genome allows for the inference of relationships at different taxonomic levels. In general, more conserved regions are more useful at higher taxonomic levels, and more variable regions are more useful at lower taxonomic levels. One of the more conserved regions (genes) of the genome, and one of the most frequently used genes in molecular systematics, is the ssrDNA. It has provided insights into the ancient divergences of prokaryotes and eukaryotes (Woese et al., 1985; Lake,

1988) and has aroused controversy and criticism concerning the monophyly of 6 metazoans (Field, 1988). Nucleotide characters have been derived from this gene for phylogenetic studies in protoctistans (Sogin, 1989; Lee and Taylor,

1992), green algae (Chapman and Buchheim, 1991), higher plants (Hamby and

Zimmer, 1988; Zimmer et a l , 1989) and vertebrates (Hillis and Dixon, 1989).

Hillis and Dixon (1991) review pertinent studies involving the use of ssrDNA in studies of molecular evolution and systematics.

The ssrDNA is part of a rDNA repeat unit that comprises the small subunit, the internally transcribed spacer 1 (ITS1), the 5.8S rDNA, the ITS2, the large subunit rDNA (IsrDNA) and the intergenic spacer (IGS). This order of coding and non-coding (spacer) regions is tandemly repeated 50 to 200 times in the haploid genome of fungi (Rubtsov et al, 1980). In other eukaryotes the copy number is considerably greater with the repeat unit present in thousands of copies (Jorgenson and Cluster, 1988). The repeat units may vary in copy number and length at inter- and intraspecific levels. The major site of length variation is within the IGS, which is also the most variable region of the repeat unit with respect to nucleotide sequence. In general, the nucleotide sequence of the non­ coding, spacer regions (ITS1, ITS2 and IGS) is more variable than that of the coding regions (SSU, 5.8S and LSU) (Jorgensen and Cluster, 1988). A hallmark of rDNA repeat units is the nucleotide homogeneity among individual repeat units within a tandem array. This phenomenon has been termed concerted evolution (Zimmer, 1980; Arnheim, 1983) or molecular drive (Dover, 1982).

Two mechanisms, gene conversion (Fogelet al, 1978; Hillis et al, 1991) and unequal crossing-over (Tartof, 1975; Smith, 1976), have been theorized to account for the homogeneity, but neither has been universally accepted.

The ssrDNA is approximately 1750 base pairs (bp) in length in most fungi; however, some isolates have been reported to possess intervening 7

sequences (IVS) of several hundred bp (Edman et al, 1988). Presumably, the

IVS is spliced out of the initial transcript and is not part of the mature ribosomal

RNA (rRNA). ssrDNA encodes for the ssrRNA transcript, which is the major

RNA component of the 40S subunit of the ribosome. The ssrRNA is proposed to have a secondary structure that consists of double-stranded regions of internal base pairing (stems) and single-stranded regions in which no base pairing occurs

(loops) (Gutell, 1985). In general, the stems are more conserved than loops with respect to nucleotide changes. This is attributed to the functional constraints of maintaining the internal base pairing . The difference in functional constraints results in some regions of the molecule evolving at slower, or faster, rates than others. Variation in rates of evolution allows the ssrDNA to provide phylogenetically informative characters across a broad taxonomic range.

Furthermore, functional constraints have produced an ssrDNA motif that corresponds to regions of the ssrDNA secondary structure that are conserved across kingdoms (Gray, 1984). Most studies in molecular systematics have worked at the level of the primary structure, the nucleotide sequence; however, the utility of characters derived from the secondary structure may be particularly suited for investigations of ancient evolutionary events.

ssrDNA has been employed in several studies of fungal systematics. It was used to place the human pathogen, Pneumocystis carinii, within the higher fungi

(Edman et al, 1988). Bowmanet al (1992) used the ssrDNA to place the chytridiomycetes within the Kingdom Fungi. Chytridiomycetes are the only fungi that possess a flagellated stage, and represent the most primitive division of fungi.

Bruns et al (1991) have elucidated relationships among ectomycorhizzal basidiomycetes, and have found that convergence and parallelism is common among gilled and pored species. Berbee and Taylor (1992a; 1992b) used the 8 ssrDNA to test relationships based upon ascomatal morphology within the ascomycetes. Their findings support two classes of unitunicate, ascomycetes

(Plectomycetes and Pyrenomycetes). Bruns et al (1992) provides an excellent review of the current state of molecular systematics in fungi with an emphasis on rDNA based phylogenies.

The second data set presented here (Ch. 4) is derived from the B subunit of vacuolar ATPase gene. Vacuolar ATPase (vATPase) is a multimeric enzyme complex that is part of the proton (H + ) ATPase family of enzymes. It is located on the endomembrane system of eukaryotic cells, and is proposed to be homologous to FjF0 ATPase of mitochondria and chloroplasts (Nelson and Taiz,

1990). However, unlike FjF0 ATPase it functions only in the direction of hydrolysis of ATP producing ADP, Pi and an H + (Johnsonet al., 1982). The H + is transported to the inside of the membrane system thereby setting up a proton gradient. The establishment of this gradient is thought to assist in the transport of proteins across the membrane and maintain the acidic pH of the vacuole

(Nelson and Tiaz, 1990).

Vacuolar ATPase (vATPase) consists of two large subunits and several smaller ones. The two large subunits are termed A and B, and are 70 and 60 kilodaltons (kD), respectively. A is the catalytic subunit and B is proposed to be regulatory in nature. In pairwise comparisons between fungi and plants, the A and B subunits proved to be 62% and 70% similar at the nucleotide level, respectively (Nelson and Taiz, 1990).

The A and B subunits of the vATPase were used to infer the phylogenetic placement of the methanobacterium, Methanococcos thermolithotrophicus, among the eubacteria, archaebacteria and eukaryotes (Gogarten et al, 1989a). Their results support the placement of methanobacteria within archaebacteria, and 9 suggest that archaebacteria and eukaryotes are sister taxa relative to eubacteria. They proposed that vATPase of eukaryotes and the membrane ATPases of archaebacteria are homologous and represent an ancient enzyme complex that predates the archaebacteria-eukaryote split. The B subunit was chosen for this study because it is more conserved than the A subunit, and therefore might provide a more accurate data set for comparison with the ssrDNA. An advantage that genes encoding for a protein have over rDNA is the data may be analyzed at either the nucleotide or amino acid level. Due to the degeneracy of the genetic code, the amino acid sequence is more conserved than that of the nucleotide.

The advent of the polymerase chain reaction (PCR) has facilitated the incorporation of molecular data into systematics. PCR is anin vitro enzymatic reaction involving Taq polymerase, genomic DNA, and oligonucleotide primers that flank a DNA region of interest (Mullis and Fallona, 1987; Saikiet el, 1988).

The result is an exponential increase in the DNA region of interest by up to 10^- fold. This DNA fragment may then serve as a template in other enzymatic manipulations, e.g. sequencing and restriction digest analysis. Prior to PCR SSU rRNA was the molecule of choice in systematics because it could be purified in large quantities and sequenced using reverse transcriptase (Hamby et al., 1988).

PCR allows the use of primers to amplify other regions of DNA, thereby increasing the number and diversity of genes that may be exploited in systematics and other fields of biology.

Ascomycetes are good candidates for studies involving molecular characters. Morphological characters in ascomycetes are few in number, and are often phenotypically plastic. The characters that do exist assist in placing a taxon within a family or order but do not aid in understanding the relationships among orders. Many studies have emphasized a single character such as ascospore 10

morphology, but seldom have a number of morphological characters been

integrated into a single data matrix and analyzed phylogenetically.

Historically, three schools of thought exist concerning the philosophy of

systematics, and therefore data analysis; these are evolutionary taxonomy,

phenetics and phylogenetics (cladistics) (Wiley, 1981). Evolutionary taxonomy is

a diverse field comprising various positions rather than a strict theory. It is

opposed to both phenetics and cladistics, and most importantly it accepts

paraphyletic taxa as natural groupings. The acceptance of paraphyletic taxa reflects the primary difference between evolutionary taxonomy and cladistics.

Evolutionary taxonomists validate paraphyletic taxa by invoking morphological

gaps, species richness, adaptive zones and monophyly. However, this definition

of monophyly includes paraphyly and, therefore, is less restrictive than that of

cladistics (Wiley, 1981).

Phenetics comprises various techniques that attempt to group taxa into larger, more encompassing groups based on an estimate of overall similarity

(Sneath and Sokal, 1973). The resulting relationships do not necessarily depict phylogenetic relationships, although they are often interpreted as such. The

debate between pheneticists and cladists emanates from the pheneticists1

contention that relationships based on overall similarity produce stable, natural

classifications. However, the derived phenograms have not been stable, and thereby call to question the representation of natural classifications (Mickevich,

1978). In estimating overall similarity, characters are not qualitatively distinguished as primitive (pleisiomorphic), derived (apomorphic) or nonhomologous (homoplastic). In such treatments, pleisiomorphic and homoplastic characters contribute to the overall similarity and result in groupings 11 that are, at least in part, based on primitive and nonhomologous characters rather than strictly derived, homologous ones. Phylogenetics, or cladistics, was pioneered by Willi Hennig (1950) in his classic treatise Grundzuge einer Theorie der Phylogenetischen Systematik.

However, it was not until Phylogenetic Systematics was printed in English (Hennig,

1966) that his theory drew widespread attention and, still later, acceptance. The major ideas proposed in his work are as follows: 1) relationships among organisms are genealogical relationships; 2) these relationships exist within and among species; 3) all other biological phenomena are correlated with genealogical descent and are best understood in the context of evolution (descent with modification); 4) genealogical relationships among groups may be recovered by employing characters that have recorded these relationships; 5) and, classifications of organisms should be based on genealogical relationships of the organisms in question (Hennig, 1966; Wiley, 1981)

Cladograms are graphical representations of genealogical relationships among biological entities. In using cladograms one acknowledges the hierarchical nature of these relationships,i. e. the product of descent with modification. To recover this hierarchy one searches for characters that provide the best chance of having accurately recorded the speciation events in question.

Suppose an ancestral species A that diverged into species B and C (Fig. 1.1).

One would examine phenotypic and/or genotypic characters that exhibit variable states, and are shared among individuals within species B and C. This scenario is represented in a Venn diagram (Fig. 1.1) where taxa of species B are united by character states la and 2a, and taxa of species C are united by character states lb and 2b. Character 3 unites them as sister taxa relative to their common ancestor.

Individuals are united in groups based upon the observance that they have unique 12

B

(a)

(b)

BB

(c)

Figure 1.1 (a) Schematic diagram of a cladogenic event resulting in species B and C. (b) Venn diagram of overlapping characters and character states for members of species B and C. Character 4 is in conflict with characters 1 and 2. Character 3 defines B and C as sister taxa. (c) Cladogram that represents the hierarchical correlation of characters 1 -4 with respect to species B and C. 13

character states in common. It is with these shared, derived characters -

synapomorphies - that cladistic theory reconstructs genealogical relationships

among taxa (Fig 1.1).

The practice of grouping organisms based on synapomorphies seems

simple in theory; however, when one considers the large numbers of characters

that molecular biology can provide, the task becomes more involved. Moreover,

invariably not all characters are in agreement and some similarities in character

states may be homoplastic; that is, they are similar due to some reason other than common descent, e.g convergence, parallelism, reticulation. Character four

is an example of a homoplastic character. Phylogenetics approaches this

dilemma by invoking parsimony (Farris, 1970; Fitch, 1971). Parsimony chooses

the hypothesis that requires the fewest assumptions, i.e. the most prudent

explanation of the data at hand. In systematics this translates as follows: the

preferred phylogenetic hypothesis (cladogram) is the one, which requires the

fewest numbers of evolutionary events (changes) to explain the distribution of

characters and character states across a given taxonomic statement. This

approach attempts to minimize the detractive effects that homoplastic characters

have on reconstructing genealogies.

An inherent aspect of this analysis is that an assumption must be made

regarding the derived nature of a character or character state. How does one

polarize the respective states of a character? The method that is implemented

almost universally with molecular data is outgroup analysis (Stevens, 1980). In an

outgroup analysis one uses a sister group of the ingroup, the taxon in question, to polarize the character states. A phylogenetically informative character, found in both the outgroup and the ingroup, by definition is variable and assumed to be homologous. The state that is exhibited by the outgroup is assumed to be 14 ancestral and the one possessed by some or all members of the ingroup is assumed to be the derived state. This does not equate to the outgroup being an ancestral taxon to the ingroup; rather, the character states in common between the ingroup and the outgroup are primitive (symplesiomorphic), and the character states unique to the ingroup are treated as derived (apomorphic) relative to the outgroup.

Data analysis in molecular systematics is often computationally intensive.

The large number of characters and the need to search an even greater number of tree topologies require computer assistance. The recent advances in software packages have played an immeasurable role in molecular systematics and systematics in general. The data presented here were analyzed using phylogenetic software packages, which will discussed where appropriate.

This dissertation is an exercise in molecular phylogenetics. Ordinal and supraordinal level relationships within unitunicate, perithecial ascomycetes were inferred using cladistically analyzed molecular data from the ssrDNA and B vATPase. The specific objectives of the research were:

1) to test the monophyly of unitunicate, perithecial ascomycetes

(pyrenomycetes);

2) to construct phylogenetic hypotheses for supraordinal level

relationships within monophyletic clades of pyrenomycetes;

3) to test Luttrell's centrum concept for pyrenomycetes (1951);

4) to construct a molecular data set from the B subunit of vATPase

that is independent of ssrDNA; and,

5) to infer non-molecular characters, both morphological and

ecological, that are congruent with molecular data. 15

Literature Cited

Alexopoulos, C. J. andW. C. Mims. 1979. Introductory Mycology, 3rd ed. John Wiley & Sons, N. Y. 632 p.

Arnheim, N. 1983. Concerted evolution of multigene families, pp. 38-61. In Nei, M. and R. K. Koehn (eds.) Evolution o f Genes and Proteins. Sinaur Assoc., Inc., Sunderland, MA.

Barr, D. J. S. 1992. Evolution and kingdoms of organisms from the perspective of a mycologist.Mycologia 84: 1-11.

Berbee, M. L. and J. W. Taylor. 1992a. Two ascomycete classes based upon fruiting-body characters and ribosomal DNA sequence. Molec. Biol. Evol. 9: 278-284.

and . 1992b. Convergence in ascospore discharge mechanism among pyrenomycete fungi based on 18S ribosomal RNA gene sequence. Molec. Phylog. Evol. 1:59-71.

Bruns, T. D., T. J. White, and J.W. Taylor. 1991.Fungal molecular systematics. Ann. Rev. Ecol. Syst. 22: 525-564.

Bowman, B. H., J. W. Taylor, A. G. Brownlee, J. Lee, S. Lu, and T. J. White. 1992. Molecular evolution of the Fungi: Relationship of the Basidiomycetes, Ascomycetes, and Chytridiomycetes. Molec. Biol. Evol. 9: 285-308.

Chapman, R. L. and M. A. Buchheim. 1991. Ribosomal RNA gene sequences: analysis and significance in the phylogeny and taxonomy of green algae. Crit. Rev. PI. Sci. 10: 343-368.

Dover,G. 1982. Molecular drive: a cohesive mode of species evolution. Nature 299: 111-117.

Edman, J. C., J. A. Kovacs, H. Masur, D. V. Santi, H. J. Elwood and M. L. Sogin. 1988. Ribosomal RNA sequence showsPneumocystis carinii to be a member of the fungi. Nature 334: 519-522.

Farris, J. S. 1970. Methods for computing wagner trees. Syst. Zool. 19: 83-92.

Field, K. G., G. J. Olsen, D. J. Lane, S. J. Giovannoni, M. T. Ghiselin, E. C. Raff, N. R. Pace, and R. A. Raff. 1988. Molecular phylogenyof the animal kingdom. Science 239: 748-753.

Fitch, W. M. 1971. Toward finding the course of evolution: minimum change for a specific tree topology.Syst. Zool. 20: 406-416. 16

Fogel, S., R. Mortimer, K. Lusnak and F. Tavares. 1978. Meiotic gene conversion: a signal of the basic recombination event in yeast. Cold Spring Harbor Symp. Quant. Biol 43: 1325-1341. Gogarten, J. P., T. Rausch, P. Bernasconi, H. Kibak, and L. Tiaz. 1989a. Molecular evolution of H +ATPases. I.Methanococcus and Sulfolobus are monophyletic with respect to eukaryotes and eubacteria. Z. Naturforsch. 44c: 641-650.

, H. Kibak, P. Dittrich, L. Tiaz, E. J. Bowman, B. J. Bowman, M. F. Manolson, R. J. Poole, T. Date, T. Ashima, J. Konishi, K. Denda, and M. Yoshida. 1989b. Evolution of the vacuolar H +-ATPase: Implications for the origin of eukaryotes. Proc. Natl Acad. Sci. USA 86: 6661-6665.

Gray, M., W. D. Sankoff and R. J. Cedergren. 1984. On the evolutionary descent of organisms and organelles: a global phylogeny based on a highly conserved structural core in small ribosomal RNA. Nuc. Acids. Res. 12: 5837-5852.

Gutell, R ., B. Weiser, C. R. Woese, and H. F. Noller. 1985. Comparative Anatomy of 16-S-like Ribosomal RNA. Prog. Nucleic Acids Res. 32: 155- 216.

Hamby, R. K. and E. A. Zimmer. 1988. Ribosomal RNA sequences for inferring phylogeny within the grass family (Poaceae). PI. Syst. Evol. 160: 29-37.

, L. E. Sims, L. E. Issel and E. A. Zimmer. 1988. Direct ribosomal RNA sequencing: optimization of extraction and sequencing methods for work with higher plants. Plant Mol. Biol. Rep. 6: 175-192.

Hawksworth, D. L. 1991. The fungal dimension of biodiversity: magnitude, significance, and conservation. Mycol. Res. 95: 641-655.

Hennig, W. 1950. Grundziige einer Theorie der phylogenetischen Systematik. Deutscher Zentralverlag, Berlin.

1966. Phylogenetics Systematics. University of Illinois Press, Urbana, IL. p.263.

Hillis, D. M., C. Moritz, C. Porter and R. J. Baker. 1991. Evidence for biased gene conversion in concerted evolution of ribosomal DNA.Science 251: 308-310.

Hillis, D. M. and M. T. Dixon. 1989. Vertebrate phylogeny: evidence from 28S ribosomal DNA sequences, pp. 355-367. In Fernholm, B., K. Bremer and H. Jornvall (eds.) The Hierarchy of Life. Elsevier Science Publishers B. V., Amsterdam.

. 1991. Ribosomal DNA: molecular evolution and phylogenetic inference. Quart. Rev. Biol. 66: 411-453. 17

Johnson, R. G., M. F. Beers and A. Scarpa. 1982. H +ATPase of chromaffin granules: kinetics, regulation and stoichiometry. J. Biol. Chem. 257: 10701-10707.

Jorgensen, R. A. and P. D. Cluster. 1988. Modes and tempos in the evolution of nuclear ribosomal DNA: new characters for evolutionary studies and new markers for genetic and population studies. Ann. Missouri Bot. Gard. 75: 1238-1247. Lake, J. A. 1988. Origin of the eukaryotic nucleus determined by rate-invariant analysis of rRNA sequences. Nature 331: 184-186.

Lee, S. B. and J. W. Taylor. 1992. Phylogeny of five fungus-like Protoctistan Phytophthora species, inferred from the internal transcribed spacers of Ribosomal DNA. Molec. Biol. Evol. 9: 636-653.

Luttrell, E. S. 1951. Taxonomy of the pyrenomycetes.University o f Missouri Studies 3:1-120.

Mickevich, M. 1978. Taxonomic congruence.Syst. Zool. 27: 112-118.

Mullis, K. B. and F. A. Fallona. 1987. Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reaction. Meth. Enzym. 155: 335-350.

Nannfeldt, J. A. 1932. Studien iiber die morphologie und systematik der nichlichenisieten inoperculaten Discomyceten. Nova Acta Regiae Sco. Scien. Upsal, Ser. IV. 8:1-368.

Nelson, N. and L. Tiaz. 1990. The evolution of H + ATPases. Trends Biochem. Sci. 14: 113-116.

Pirozynski, K. A. 1976. Fossil Fungi. Ann. Rev. Phytopath. 14: 237-246.

Rubtsov, P. M., M. M. Musakhanov, V. M. Zakhaiyev, A. S. Krayev, K. G. Skryabin and A. A. Bayev. 1980. The structure of the yeast ribosomal RNA genes. I. The complete nucleotide sequence of the 18S ribosomal RNA gene from Saccharomyces cerevisiae. Nuc. Acids. Res. 8: 5779-5794.

Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Huiguchi, G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with thermostable DNA polymerase. Science 239: 487-491.

Sherwood-Pike, M. A. and J. Gray. 1985. Silurian fungal remains: probable records of the class Ascomycetes. Lethaia 18: 1-20.

Smith, G. P. 1976. Evolution of repeated DNA sequences by unequal crossover. Science 191: 528-535.

Sneath, P. H. A. and R. R. Sokal. 1973. Numerical Taxonomy. Freeman & Co. San Francisco. 18

Sogin, M. L., U. Edman, and H. Elwood. 1989. A single kingdom of eukaryotes, pp. 133-143. In Fernholm, B., K. Bremer and H. Jornvall (eds.) The Hierarchy o f Life. Elsevier. Amsterdam.

Stevens, P. F. 1980. Evolutionary polarity of character states.Ann. Rev. Ecol. Syst. 11: 333-358.

Tartof, T. N. 1975. Redundant genes. Ann. Rev. Genet. 9: 355-385.

Troitsky, A. V., Y. F. Melekhovets, G. M. Rakhimova, V. K. Bobrova, K. M. Valiejo-Roman and A. S. Antonov. 1991. Angiosperm origin and early stages of seed plant evolution deduced from rRNA sequence comparisons. J. Molec. Evol. 32: 253-261.

White, J. F. and T. N. Taylor. 1998. Triassic fungus from Antarctica with possible ascomycetous affinities. Amer. J. Bot. 75: 1495-1500.

Wiley, E. O. 1981. Phylogenetics: the Theory and Practice o f Phylogenetics. Systematics. John Wiley & Sons, N. Y. p.439.

Woese, C. R., E. Stackebrandt, T. J. Macke, and G. E. Fox. 1985. A phylogenetic definition of the major eubacterial taxa.Syst. Appl. Microbiol. 6: 143-151. Zimmer, E. A. and R. K. Hamby, M. L. Arnold, D. A. Leblanc and E. C. Theriot. 1989. Ribosomal RNA phylogenies and flowering plant evolution, p.205- 214. In Fernholm, B., K. Bremer, and H. Jornvall (eds.) The Hierarchy of Life. Elsevier. Amsterdam. CHAPTER 2

THE CLAVICIPITALES-HYPOCREALES CONNECTION

The Clavicipitales are unitunicate, perithecial ascomycetes that include

endophytes of grasses and parasites of insects and other fungi (Rogerson, 1970).

The geographic distribution of these organisms is panglobal with representatives in the tropics and temperate regions of the world. As a whole, the order is thought to be centered in the tropics (Diehl, 1950). Species of the order have had a dramatic influence on human societies and continue to have an economic impact. The notorious effects of ergotism (primarily caused by the ingestion of rye infected with Claviceps purpurea) on human civilization have been documented in numerous societies for centuries (Matossian, 1989). The genera within the tribe Balanseae are endophytes of grasses (Poaceae) and sedges

(Cyperaceae) (Diehl, 1950), and have a large impact upon the host plant and its community (White, 1992). They have been linked with staggers in cattle and other livestock (Fletcher and Harvey, 1981; Bacon et al., 1977), increased resistance to insect herbivory (Prestridge et al, 1982; Clay et al., 1985), and increased resistance to fungal infections (Stovall and Clay, 1991). Species of the genus Cordyceps are of particular interest because they are parasites of both fungi and insects (Mains, 1957; 1958; Rogerson, 1970), and represent a model system for studying host switching and specificity.

The taxa of this order possess unique morphological characters that have been interpreted as shared, derived characters (synapomorphies). The stromata are fleshy and brightly colored to dark. The asci are long and cylindrical, and

19 20

possess a pronounced apical cap; they are produced in a strict basal cluster.

Lateral paraphyses are present but deliquesce prior to ascospore maturity. The

ascospores are long and filiform; at maturity they are septate and disarticulate

into part spores (Rogerson, 1970). In general, researchers have agreed on the

monophyly of this group; however, two major alternative classifications exist that

differ over taxonomic ranking and placement. One treats the order as a being a sister group to the Hypocreales (Gaumann, 1952; Rogerson, 1970;

Wehmeyer,1975). The second considers the Clavicipitales as being a near

relative to members of the Xylariales (Sphaeriales) (Miller, 1949; Luttrell, 1951;

Barr, 1990).

The two schools of thought disagree on emphasis of different

morphological and developmental characters. One places the Clavicipitales as a

sister group to the Hypocreales (Gaumann, 1952; Wehmeyer, 1975; Rogerson,

1970), or as a member of the order (Gaumann and Dodge, 1928). This model

emphasizes characters of the stroma and the asexual states (anamorphs). Fungi of both orders produce brightly colored to dark, fleshy stromata with true perithecia immersed becoming erumpant to superficial on the stroma. Also, both orders are linked to anamorphs that exhibit phialidic production of conidia. The second hypothesis proposes that the Clavicipitales are more closely related to members of the Xylariales (Miller, 1949; Luttrell, 1951; Barr, 1990). In doing so, it emphasizes characters of centrum anatomy and development. The aparaphysate basal cluster of asci with lateral, evanescent paraphyses of the

Clavicipitales has been interpreted as a variation of the Xylaria centrum type

(Luttrell, 1951; Barr, 1990). The pronounced apical cap of the ascus in clavicipitalean fungi has been employed by proponents of both schools, because 21 there are taxa in the Hypocreales and Xylariales that possess a less pronounced apical cap.

The study presented here is a cladistic analysis of nucleotide characters sampled from the nuclear-encoded small subunit ribosomal DNA (ssrDNA).

Use of molecular characters and cladistic analysis provides an independent test of morphologically based phylogenies. This approach assists in overcoming the previous inability to polarize non-molecular characters and identify synapomorphies necessary to determine ordinal level ranking and placement.

The objective of this work is to test the monophyly and infer the ordinal placement of the taxa sampled for the Clavicipitales.

Materials and Methods

The taxa sampled for this study were chosen based upon their position in past and contemporary classification systems and their possession of particular morphological and ecological characters. Taxa were sampled to survey the centrum types of unitunicate, perithecial ascomycetes as defined by Luttrell

(1951). Also, certain aspects of the life history such as host affiliation and nutritional modes (parasitic, saprophytic and biotrophic) were considered. The taxa and source of isolates are listed in Table 2.1. The ingroup consisted of 33 unitunicate, perithecial ascomycetes, Aureobasidium pullulans and Stictis radiata; the outgroup comprises five species of yeasts from the Endomycetales and

Taphrina deformans.

Organisms were cultured in modified Leonian's liquid medium (Cain and

Weresub, 1957) for approximately one week. Cultures were harvested on filter paper (Whatman no.l), frozen in liquid nitrogen and stored at -20°C. Total 22

Table 2.1 List of taxa included in study, source of taxon and regions sequenced. Primers are those of Whiteet al. (1990).

Aciculosporium take Miyake ATCC 36483 NS2, NS4 Ambrosiozyma platypodis (Baker & Kreger ATCC 36174 NS, NS2, NS3, NS4 - van Rij) van der Walt Aphysiostroma stercorarium Barassa & Moreno A. Rossman NS2, NS4 Aureobasidium pullulans (de Bary) Amaud Illingsworthet al., 1991 Balansia clserotica (Pat.) Hoehn. ATCC 16582 NS1.NS2, NS3.NS4 Candida albicans (Robin) Berkout Genbank Ysall6s Candida tropicalis (Castellani) Berkhout Genbank Ysarmab Cephaloascus fragrans Hanawa ATCC 36174 NS1, NS2, NS3, NS4 Ceratocystis fimbriata Ellis & Halst. Harrington C89NS1.NS2, NS3, NS4 Ceratocystis virescens Davids. Harrington C69 NS1.NS3 Cercophora septentrionalis Lundq. D. Malloch NS1.NS2, NS3 Chaetomium globosum Kunze ex Fr. ATCC 44699 NS1.NS2, NS3 Claviceps paspali Stevens & Hall ATCC 13892 NS2, NS4 Cordyceps capitata (Fr.) Link J. Spatafora NS1, NS2, NS3, NS4 Daldinia concentrica (Bulliard ex Fr.) ATCC 36659 NS1, NS3, NS4 Cesati et de Notaris Diaporthe phaseolarum (Cookeet Ellis) Sacc. FAU-458 NS2, NS4 Diatrype discoformis (Hoffm.rFr.) Fr. CBS 197.49 NS2, NS4 Epichloe typhina (Pers.) Tulasne ATCC 56429 NS2, NS4 Hirsutella thompsonii Fischer ATCC 24874 NS2, NS4 Hypocrea lutea (Tode:Fr.) Petch GJS 84-474 NS1.NS2, NS3 Hypocrea pallida Ellis & Everh. GJS 89-83 NS1.NS2, NS3.NS4 Hypocrea schweinitzii (Fr.)Sacc. CTR 79-225 NS1, NS2, NS3.NS4 Hypocrella sp. Sacc. GJS 89-104 NS1.NS2, NS3.NS4 Hypomyces polyporinus Peck ATCC 46844 NS2, NS4 Hypoxylon atroroseum J. Rogers NS1.NS3 Leucostoma persoonii (Nits.) Hoehn. Berbee & Taylor, 1992 Microascus trigonosporus Emmons & Dodge RSA 1942 NS1.NS2, NS3.NS4 Nectria cinnabarina (Tode ex Fr.) Fr. GJS 89-107 NS1, NS2, NS3, NS4 Nectria haematococca Berk.& Br. GJS 89-97 NS1.NS2, NS3.NS4 Neocosmospora vasinfecta Smith RSA 1898 NS2, NS4 Neurospora crassa Shear & Dodge Genbank Neurmas Ophiostoma piliferum H. & P. Syd. Harrington C300 NS2, NS4 Ophiostoma ulmi (Buisman) Nannf. Berbee & Taylor, 1992 Saccharomyces cerevisieae Hansen Genbank Yscrgea Sphaerostilbella aureonitens (Tul.) Seifert et al. GJS 83-286 NS1.NS2, NS3.NS4 Sphaerostilbella NZ GJS 82-40 NS2, NS4 Stictis radiata Per. ex S. F. Gray D. Malloch NS2, NS4 Taphrina deformans (Fuckel) Tul. ATCC 11124 NS1.NS3 Xylaria curta Fr. J. Rogers NS2, NS3, NS4 Xylaria hypoxylon (L.:Fr.) Grev. ATCC 42768 NS2, NS3, NS4 23

nucleic acids were extracted following the procedure of Lee and Taylor (1990).

Five microliters of the nucleic acid preparation was electrophoresed on 1%

agarose gel (Amresco, biotech grade) and stained with ethidium bromide to visualize high molecular weight DNA. The nucleic acid sample was diluted 100- fold for use in polymerase chain reactions (PCR).

Twenty microliters of the diluted nucleic acid sample provided the template (genomic DNA) in symmetric PCRs (Mullis and Fallona, 1987; Saiki et al, 1988). The nuclear-encoded small subunit ribosomal DNA (rDNA) was amplified symmetrically using primers NS1 and NS4 (White et al, 1990).

Symmetric PCRs were optimized following the procedure of Kaltenboecket al

(1992). Single-stranded PCR (ssPCR) products were amplified following the technique of Kaltenboeck et al (1992). ssPCR products were purified from unincorporated nucleotides and primers using centrifugal filtration cartridges (Millipore, UFC3 TTK 00) following manufacturer's protocol. Samples were concentrated to a final volume of 25 microliters.

The regions of the ssrDNA sequenced, or source of sequence, for the taxa included in this study are noted in Table 2.1. Seven microliters of the purified ssPCR products served as template in sequencing reactions using a Sequenase 2.0 sequencing kit (USB). Reactions were performed according to manufacturer's protocol except dGTP labeling mix was diluted 25-fold, extension time was for three minutes, and termination reactions were performed at 42°C for three minutes. Samples were subject to electrophoresis on 5% polyacrylamide gels

(Amresco). Samples were double loaded; four microliters were loaded and run for 2.5 hours followed by an additional load in the adjacent lanes and subsequent

2.5 hours of electrophoresis. This procedure maximized the number of nucleotides obtained from a single gel. Gels were fixed in a 10% acetic acid and 24

12% methanol solution for 30 minutes. Gels were then blotted onto filter paper

(Whatman no. 3) and dried under vacuum at 60°C for 1.5 hours. Dried gels were exposed to x-ray film (Kodak XRP-5) for 48 hours and developed in Kodak GBX developer.

Sequences were read a total of three times to minimize any error that may have been introduced by the investigator and entered by hand into a sequence file using the University of Wisconsin Genetics Computer Group (GCG) software package. Initial alignments were facilitated using the programs GAP,

LINEUP, and PRETTY; and, final alignments were optimized by hand. The data were analyzed with PAUP 3.0q (Swofford, 1990) on a Macintosh Ilex computer. Due to the number of taxa, only heuristic searches were employed; these include general, stepwise addition (random sequence addition) and branch swapping (tree bisection-reconnection) algorithms. Ten replications of each of the three types of heuristic searches were performed.

Decay indices (Mishler, pers. comm.) and bootstrapping (Felsentein,

1985) were used to estimate putative stability and support of the inferred clades.

Bootstrap confidence levels were calculated on PAUP 3.0q. It is a computationally intensive exercise with the limiting components being the number of taxa and characters. Therefore, the number of taxa was pruned to 30 so as to represent only the major clades of interest. 100 bootstrap replications were performed. Admittedly, 100 is an undesirably low number of replications, however, it is the upper limit for the computers at our disposal with this number of taxa. Decay indices were also calculated using PAUP 3.0q. Decay indices up to > 10 steps were calculated for the 30 taxon statement; I was unable to find all trees ten steps longer. Decay indices were calculated for 8700 trees and are presented with the bootstrap values in Figure 2.2. The 8700 trees were filtered, 25

decreasing one step each round of filtering; a strict consensus tree was

constructed after each round. Each of the ten consensus trees was compared to

the strict consensus of the eight most parsimonious trees. This procedure

allowed estimation of the number of steps away from the most parsimonious

cladogram (or the consensus of the most parsimonious cladograms) that a

particular clade is no longer supported (Mishler, pers. comm.).

Results

The ssrDNA data set consists of 40 taxa and approximately 900 bp/taxon,

450 bp from the two regions designated by the primer pairs NS1/NS2 and

NS3/NS4 (White et al., 1990). Of the 900 bp sampled, 168 are phylogenetically

informative. Cladistic analysis of the 168 phylogenetically informative positions

produced 42 equally most parsimonious cladograms (trees) of 463 steps with

consistency (Cl), retention (RI) and rescaled consistency (RC) indices of .542,

.779 and .422, respectively. A strict consensus of these trees is 474 steps in length

(Fig. 2.1) and has a Cl of .530, a RI of .766 and a RC of .406.

The taxa sampled for the Clavicipitales form a natural, monophyletic group (clade), although the unweighted data were unable to resolve relationships within the clade. A successive approximation (Farris, 1969), whereby the characters were weighted based upon their rescaled consistency indices, chose one of the 42 most parsimonious trees produced by the unweighted data (Fig.

2.2). These data suggest that the Clavicipitales is a derived group of filamentous ascomycetes that is a sister group to the Hypocreales, and not a relatively close neighbor to the Xylariales. The taxon sampled from the Ostropales,Stictus 26

radiata, is not closely related to the Clavicipitales as proposed (Gaumann and

Dodge, 1928; Gaumann, 1952; Rogerson, 1970; Sherwood, 1977).

Cladistic analysis of the pruned taxonomic statement resulted in eight most parsimonious trees. Unlike the 40 taxa cladogram (Fig. 2.1) the

Clavicipitales-Hypocreales node is unresolved in a strict consensus of the eight

trees; nodes that received confidence levels of 90% or greater are noted on the

strict consensus tree (Fig. 2.2). The taxa sampled for the Hypocreales represent two paraphyletic lineages. One consists of the generaHypocrea, Hypomyces, Sphaerostilbella and Aphysiostroma, and is supported by bootstrap values and decay indices of 92% and 7 steps, respectively (Fig. 2.3); this clade is a sister group to the Clavicipitales. The other clade comprises taxa from the genera

Nectria and Neocosmospora and is not strongly supported.

The Clavicipitales and Hypocreales form a clade that is a sister group to the genera Microascus and Ceratocystis (Fig. 2.1; Fig. 2.3). It is supported by a bootstrap value of 91% and a decay index of 6. These taxa comprise one of the two subclades of pyrenomycetous taxa, subclade A. The other subclade of pyrenomycetes, subclade B, includes taxa sampled from the orders Diaporthales,

Sordariales, Xylariales and the genus Ophiostoma. The clade containing the pyrenomycetes is supported by a bootstrap value of 100% and a decay index > 10.

Aureobasidium pullulans and Stictis radiata form the most basal clade of the ingroup. The yeasts sampled from the Endomycetales form a monophyletic group that is supported by a bootstrap value of 100% and a decay index of > 10 steps. The monophyly of the ingroup was maintained with respect to the outgroup. However, this taxonomic statement (sampling) does not address the monophyly of Pyrenomycetes(sensu Luttrell, 1951) as a natural grouping due to incomplete taxon sampling. This question is considered in Chapter 4. 27

■ Aciculosporium take

— Balansia sclerotica • Claviceps paspali — Cordyceps capilata CLAVICIPITALES Epichloe typhina

— Hirsutella thompsonii — — Hypocrella sp.

j - Aphysioslroma slercorarium P— Hypocrea schweinitzii L - Hypocrea lutea Sphaerostilbella NZ Sphaerostilbella aureonitens ri HYPOCREALES Hypocrea pallida Hypomyces polyporinus Nectria haemalococca Neocosmospora vasinfecta Nectria cirmabarina Ceratocystis fimbriata OPHIOSTOMAT ALES • Ceratocystis virescens ■ Microascus trigonosporus MICROASCALES I Diaporthe phaseolarum DIAPORTHALES L- Leucostoma persoonii p Ophiostoma piliferum OPHIOSTOMATALES L Ophiostoma ulmi • Cercophora septentrionalis ' Chaetomium globosum SORDARIALES • Neurospora crassa j — Daldinia concenlrica B __P Diatrype discoformis I— Hypoxylon atroroseum XYLARIALES r -X• Xylaria y h curia i—■ Xylaria ; hypoxylon • Aureobasidium pullulans DOTHIDEALES • Stictis radiata OSTRAPALES • Ambrosiozyma platypodis Saccharomyces cerevisiae i— Candida albicans ENDOMYCETALES L■ CandidaC.aih tropicalis • Cephaloascusfragrans ■ Taphrina deformans I TAPHRINALES

Figure 2.1 Strict consensus of 42 equally most parsimonious cladograms containing 40 taxa (474 steps, Cl = .530, RI = .766 and RC = .406). The designated outgroup is T. deformans and the Endomycetales. Branch lengths are proportional to the numbers of steps. 28

Aciculosporium take — Balansia sclerotica Claviceps paspali Epichloe typhina Hypocrella sp. Hirsutella thompsonii Cordyceps capitata r— Aphysiostroma stercorarium j >— Hypocrea schweinitzii I L . Hypocrea lutea ]r— Sphaerostilbella NZ * Sphaerostilbella aureonitens Hypocrea pallida Hypomyces polyporinus j - Nectria haematococca P Neocosmospora vasinfecta 1— Nectria cinnabarina Ceratocystis fimbriata ■ c Ceratocystis virescens Microascus trigonosporus j Diaporthe phaseolarum I— Leucostoma persoonii [ Ophiostoma piliferum Ophiostoma ulmi Cercophora septentrionalis Chaetomium globosum - Neurospora crassa Daldinia concentrica B Diatrype discoformis Hypoxylon atroroseum -XyXylaria curta Xylaria hypoxylon Aureobasidium pullulans Stictis radiata ■ Ambrosiozyma platypodis Saccharomyces cerevisiae rCECephaloascus fragrans Candida albicans " t CandidaCat tropicalis Taphrina deformans

Figure 2.2 Single cladogram from one round of successive approximation using the maximum fit option. Characters were reweighted based on their rc from the initial parsimony analysis. This cladogram is one of the 42 equally most parsimonious cladograms inferred by the unweighted data. Branch lengths are distorted by the reweighting of characters and are not indicative of the number of steps. 29

Ilypomyces polyporinus 92 — Hypocrea schweinitzii Sphaerostilbella aureonitens i- Nectria haematococca 11 Neocosmospora vasinfecta L - Nectria cinnabarina

91 98 i Balansia sclerotica c Claviceps paspali 100 | Ceratocystis fimbriata 94 >10 L— Ceratocystis virescens

—— Microascus trigonosporus Diaporthe phaseolarum 100 99 >10 Leucostoma persoonii 100 r Ophiostoma piliferum >10 {L. Ophiostoma ulmi

98 1 - Cercophora septentrionalis 99 [ T t _Chaetomium globosum 95 - Neurospora crassa Daldinia concentrica rC Hypoxylon atroroseum Xylaria curta Xylaria hypoxylon Aureobasidium pullulans Stictis radiata

- Ambrosiozyma platypodis — Saccharomyces cerevisiae 100 97 ■ 1 Candida albicans >10 5 LL CandidaCam tropicalis — Cephaloascus fragrans Taphrina deformans

Figure 2.3 Strict consensus of eight equally most parismonious cladograms containing 30 taxa (396 steps, Cl = .591, RI = .780 and RC = .461). Bootstrap confidence levels are given above those nodes that were supported in a 90% or more of the trees produced from 100 bootstrap replications. Corresponding decay indices are presented below their respective nodes. 30

Discussion

Cladistic analysis of nucleotide characters from the ssrDNA supports the hypothesis that the taxa sampled from the Clavicipitales are more closely related to the taxa sampled from the Hypocreales than either group is to the Xylariales

(Fig. 2.1). All three orders are part of a larger derived pyrenomycete clade that contains two subclades, A and B. Subclade A consists of the Clavicipitales,

Hypocreales, and the genera Microascus and Ceratocystis. This clade is supported at a 91% bootstrap confidence level with a decay index of 6 steps (Fig. 2.3). The clavicipitalean clade, represented by Balansia sclerotica and Claviceps paspali, is supported by a confidence level of 98%. Further support for the clavicipitalean clade was pursued by bootstrapping several pairs of clavicipitalean taxa. The taxon pairs of C.paspali and Hypocrella sp., C. paspali and Hirsutella thompsonii,

C. paspali and Epichloe typhina received bootstrap confidence levels of 95%, 92% and 93%, respectively. However, neither the C. paspali and Cordyceps capitata pair and nor the C. paspali and Aciculosporium take pair received bootstrap confidence levels above 90% (Fig. 2.4; Tbl. 2.2). These findings suggest that homoplasy is not uniformly distributed within an inferred monophyletic group.

One could criticize the subsampling of taxa in a bootstrap analysis as not being indicative of the larger taxonomic statement. However, any taxonomic statement would be inappropriate unless it comprise all taxa of a monophyletic group.

None of the supraordinal relationships in subclade B received bootstrap confidence levels above 90%, although support is received for intraordinal relationships within the orders Diaporthales and Sordariales, and within the genus Ophiostoma at the levels of 97%, 100% and 100%, respectively. These 31

Claviceps paspali 72%

Aciculosporium take

Claviceps paspali 98%

Balansia sclerotica

Claviceps paspali 93%

Epichloe typhina

Claviceps paspali 95%

Hypocrella sp.

Claviceps paspali 92%

Hirsutella thompsonii

Claviceps paspali 73%

Cordyceps capitata

Figure 2.4 Bootstrap confidence levels for different pairs of clavicipitalean taxa derived from 100 bootstrap replications. The Clavicipitales is an unresolved monophyletic clade in the maximum parsimony analysis. Bootstrap confidence levels for pairs of taxa suggest that homoplasy is distributed nonrandomly across the inferred monophyletic group. 32

Table 2.2 Bootstrap confidence levels for pairs of clavicipitalean taxa from 100 bootstrap relications.C. paspali is paired with all species sampled for the order. Note thatC. capitata (fungal parasite) and A. take (endophyte, Japanese isolate) did not receive bootstrap confidence levels of greater than 90%.

Clavicipitalean taxon paired Bootstrap Nutritional modes with C. paspali confidence levels

Aciculosporium take 72% endophyte Balansia sclerotica 98% it tt Claviceps paspali — M Epichloe typhina 93% Hypocrella sp. 95% insect parasite Hirsutella thompsonii 92% tt Cordyceps capitata 73% fungal parasite 33

same nodes received higher decay indices than nodes not supported by bootstrap

analysis (Fig. 2.3).

In general, nodes supported by high bootstrap values received higher

decay indices than nodes that are not supported by significant bootstrap values. This apparent correlation between bootstrap confidence levels and decay indices

suggests that they may be measuring the same putative phylogenetic signal.

However, there exist discrepancies to this general trend; there are nodes that

received bootstrap confidence levels of 95% or greater, but received decay

indices lower than nodes that are not supported by bootstrapping, and vice versa

(Fig. 2.3). These discrepancies may be attributed to sampling error, i.e. a low number of bootstrap replications combined with the inability to find all trees 10 steps longer in a decay analysis, or an inherent difference in what component of the putative phylogenetic signal they are measuring.

The phylogeny inferred from the ssrDNA provides an independent test of the two rival, morphologically based phylogenetic hypotheses. This allows the identification morphological and ecological characters that are congruent with the ssrDNA gene tree, and the polarization of these non-molecular characters

(Brooks and McLennon, 1992). Species of the Clavicipitales and Hypocreales produce soft, or fleshy, stromata that are often brightly pigmented. In contrast, the stromata of the Xylariales are typically dark to black in color and carbonaceous in texture. However, stromatic characters have been de­ emphasized in recent systematic treatments. They have been deemed more phenotypically plastic than other characters, and therefore may not be good characters for ordinal level relationships (Miller, 1949; Luttrell, 1951). However, this approach assumes that all ordinal level rankings are equal with respect to evolution. The phylogeny presented here proposes that the Clavicipitales and 34

Hypocreales are sister taxa and may be members of the same order. If one

accepts this close relationship, the stromatic characters may be interpreted as

good synapomorphies uniting these two groups of fungi. This is not to say that

stromatic characters will serve as good characters for inferring all ordinal

relationships.

The conidia of the Clavicipitales may develop directly on the stroma, or from sporodochia or cupulate structures. Conidiogenous cells are enteroblastic phialidic, and produce hyaline, one-celled, dry or gultinous conidia. They were categorized by Diehl (1950) in three groups: ephelidial that produce waxy conidia from cupulate structures; sphacelial that produce glutinous conidia from sporodochia; and typhodial that produce dry conidia on the stroma. Examples are the form generaAschersonia, Hirsutella, Paecilomyces, Sphacelia and Ephelis.

Anamorphs of the Hypocreales exhibit enteroblastic phialidic conidiogenous cells that are similar to those of the Clavicipitales. They may be hyphomycetous or grouped in sporodochia; conidiophores may be unbranched or branched penicillately or verticillately (Samuels and Rossman, 1979; Samuels and

Seifert, 1987). Examples of form genera includeAcremonium, Gliocladium,

Penicillifer, Trichoderma and Tubercularia. Moreover,Acremonium, an admittedly nonnatural grouping, has been described as an anamorph of the genus

Epichloe (Morgan-Jones and Gams, 1982).

The condiogenous cell of xylarialean anamorphs is most frequently holoblastic. All of the taxa sampled in this study for the orders Xylariales(sensu

Barr, 1990) are linked with anamorphs that exhibit holoblastic conidiogenous cells. However, the sister clade to the Xylariales contains taxa sampled from the

Sordariales, Diaporthales and Ophiostoma that are linked to anamorphs, which possess entero- and holoblastic conidiogenous cells. Although, enteroblastic 35

phialides appear to be informative with respect to the Clavicipitales and Hypocreales, the usefulness of conidiogenesis among others orders is questioned.

Centrum anatomy is defined as the whole of the ascoma excluding the

peridium (Luttrell, 1951; Reynolds, 1981). Centrum development in the

Clavicipitales is described as a variation of the Xylaria centrum type (Luttrell,

1951). The asci are produced in a strict basal cluster with lateral evanescent

paraphyses. However, the typical Xylaria centrum type produces asci from a

hymenial layer that occupies the inner basal and lateral walls of the ascoma.

Furthermore, the asci are interspersed among persistent paraphyses from their

points of origin. The asci are never interspersed among the paraphyses in the

Clavicipitales.

The Hypocreales is described as having a Nectria centrum type

development (Luttrell, 1951). This centrum type is typified as producing asci

from the base of the ascoma with evanescent paraphyses produced from the

apical region of the inner ascoma. These apical paraphyses extend downward

and may intercalate between the asci; however, the asci and paraphyses are not interspersed from a common point of origin. The phylogenetic hypothesis presented here challenges the interpretation of the centrum development in the

Clavicipitales as being a variation of the Xylaria centrum type, and proposes that it is a derived state of the Nectria centrum type. A more detailed discussion of the evolution of centrum anatomy and development is provided in Chapter 5 .

Clavicipitalean fungi sampled in this study are obligately biotrophic on other fungi, insects and plants. Cordyceps capitata is parasitic on hypogeous fungi in the genus Elaphomyces (Mains, 1957). Hirsutella thompsonii, a member of an form genus linked to Cordyceps (Evans and Samson, 1982; 1984; Samson and

Evans, 1985), and Hypocrella sp. (Mains, 1959) is a parasite of scale insects. The 36

remaining taxa are associated with grasses (Poaceae) and sedges (Cyperaceae),

and are endophytes of their respective hosts (Diehl, 1950; Morgan-Joneset al.,

1992). Some species within the endophytic tribe Balansieae are proposed

commensalists of their hosts (Clay, 1988b; White, 1992). Often, the fungal

infection results in the prevention of seed set, which results in inbreeding or the

plants foregoing sexual reproduction (Clay, 1982; 1988b). However, the

endophyte-plant association has been proposed to decrease mammalian and

insect herbivory (Bacon et al., 1977; Prestridge, 1982; Clay, 1985), increase fungal

disease resistance (Stovall and Clay, 1991), and drought tolerance (West et al.,

1990) of the host plant.

The taxa sampled for the Hypocreales are also biotrophic on plants and

other fungi; however, this study is not a thorough sampling of hypocrealean fungi.

This is especially true for the genus Nectria, which comprises saprobes, plant

pathogens and fungal parasites (Samuels, 1988). Typically, the taxa sampled for

the Xylariales are saprophytic or only weakly parasitic on hardwoods and are not

representative of any major biotrophic line of evolution. There are pathogens or

parasites within, or closely related to, the Xylariales, but the overwhelming

majority of Xylarialean fungi are not associated with living material,i.e. they are

not biotrophic.

Of particular interest are the sister groups of the Clavicipitales and

Hypocrea clade. The placement of these two clades as sister groups is very unstable. The resolution of theHypocrea or the Nectria clade as the sister group to the Clavicipitales is taxon dependent (Fig. 2.1; Fig. 2.3). The relatively long branches of the Clavicipitales and the Hypocrea clades may represent attraction of relatively long branches (Felsenstein, 1978), or accelerated rates of evolution relative to the Nectria clade. By and large, the Hypocrea clade depicts a line of 37

evolution dominated by fungal parasites. If the former two clades are sister

groups, one is justified in polarizing the fungicolous tendency within the

Clavicipitales as primitive. This polarization would still be possible if the Nectria

clade is a sister group to the Clavicipitales because the genus Nectria contains

many fungicolous species (Samuels, 1988). The unweighted data did not resolve

relationships among the clavicipitalean taxa (Fig. 2.1). However, after a single

round of successive approximation (Farris, 1969) one tree was found, which is

one of the 42 most parsimonious trees constructed by the unweighted data (Fig.

2.3). The result placed C. capitata, a fungal parasite, at the base of the

clavicipitalean clade. Furthermore, the nutritional modes were polarized in the order of fungal parasite to insect parasite to plant endophyte (Fig. 2.5). The inference of endophytes as the most derived taxa fails to reject the hypothesis that commensalists are derived from parasites (Clay, 1988b). He proposed that some plant-endophyte interactions are mutually beneficial to both organisms, and that they represent derivations of host-parasite relationships. The treatment of commensalism as derived parasitism has also been offered for other systems, as well as fungi (Price, 1988).

The result of successive approximation also questions the monophyly of

Cordyceps. Hirsutella thompsonii is an anamorph ofCordyceps that is parasitic on insects. The maximum parsimony analysis puts it within the unresolved group of the Clavicipitales, and the successive approximation places it as a sister taxon to

Hypocrella sp., another insect parasite. Successive approximation is a controversial analysis that is not widely practiced within the systematics community, and the use of it concurrently with bootstrapping is unprecedented.

Therefore, the approximated relationships within the Clavicipitales are not 38

— Aciculosporium take

Balansia sclerotica PLANT — Claviceps paspali ENDOPHYTES Epichloe typhina — Hypocrella sp. INSECT PARASITES — Hirsutella thompsonii Cordyceps capitata I FUNGAL PARASITE j - Aphysiostroma stercorarium COPROPHILE \ J L - Hypocrea schweinitzii *- Hypocrea lutea Sphaerostilbella NZ FUNGAL PARASITES tSphaerostilbella , aureonitens Hypocrea pallida ‘ rHypomyces * polyporinus Nectria haematococca Neocosmospora vasinfecta SAPROBES AND PLANT PATHOGENS* Nectria cinnabarina

Figure 2.5 Enlargement of successive approximation clade containing the Hypocreales and Clavicipitales. Nutritional modes are mapped on the tree. Emphasis is placed on the trend from parasite to endophyte. *Nectria and other hypocrealean genera contain species that are parasitic on fungi. 39 offered as a proposed genealogy, but rather as a working hypothesis for future studies within the order.

Bootstrap confidence levels were unstable for the clavicipitalean clade, when either Cordyceps capitata orAciculosporium take was used as a sister taxon to Claviceps paspali. However, B. sclerotica and Hypocrella sp. were supported significantly as sister taxa toC. paspali, and Hirsutella thompsonii and Epiclhoe typhina were supported marginally. Because bootstrap confidence levels are sensitive to taxon sampling, extrapolations across an inferred monophyletic group of organisms may not be appropriate.

Cladistic analysis of nucleotide characters sampled from the ssrDNA supports a close phylogenetic affinity between the Clavicipitales and the

Hypocreales. A strict consensus suggests that the Hypocreales consists of at least two paraphyletic lineages; however, the branches connecting the Clavicipitales and the two Hypocrealean lineages are relatively short and unstable.

Furthermore, the Hypocreales is a large and diverse group of organisms that have not been thoroughly sampled in this study. Therefore, the

Clavicipitales/Hypocreales clade may be viewed as a polytomy comprising at least three lines of evolution that may be the products of a radiation event. The congruence of the molecular data with morphological characters of the stromata, anamorphs and centrum anatomy provides insight into synapomorphies that are useful at this taxonomic level, and assists in polarizing their respective character states. Also, it allows ecological characters, such as host affiliation and nutritional mode, to be viewed in a phylogenetic context. 40

Literature Cited

Bacon, C. W., J. K. Porter, J. D. Robbins, and E. S. Luttrell. 1977. Epichloe typhina from toxic tall fescue grasses.Appl. Environ. Microbiol. 34: 576- 581. Barr, M. E. 1990. Prodromus to nonlichenized, pyrenomycetous members of class Hymenoascomycetes. Mycotaxon 39: 43-184.

Berbee, M. L. and J. W. Taylor. 1992. Convergence in ascospore discharge mechanism among pyrenomycete fungi based on 18S ribosomal RNA gene sequence. Molec. Phylog. Evol. 1:59-71.

Brooks, D. R. and D. A. McLennon (eds.). 1992. Phytogeny, Ecology and Behavior: A Research Program in Comparative Biology. Univ. Chicago Press, Chicago, p.434.

Bruns, T. D., R. Fogel, T. J. White, and J. D. Palmer. 1989. Accelerated evolution of a false-truffle from a mushroom ancestor.Nature 339: 140- 144.

Cain, R. F. and L. K. Weresub. 1957. Studies of coprophilous ascomycetes. V. Sphaeronaemella fimicola. Can.J.Bot. 35: 119-131.

Clay, K. 1988a. Fungal endophytes of grasses: a defense mutualism between plants and fungi. Ecology 69:10-16.

, 1988b. Clavicipitaceous fungal endophytes of grasses: coevolution and the change from parasitism to mutualism, p.79-105. In Hawksworth, D. L., and K. Pirozynski (eds.). Academic Press, N Y.

Clay, K., T. N. Hardy and A. M. Hammond, Jr. 1985. Fungal endophytes of grasses and their effects on an insect herbivore. Oecologia 66: 1-6.

Diehl, W. W. 1950. Balansia and the Balansiae in America. Agriculture Monograph no. 4. USD A, Washington, D.C. p.78.

Eckblad, F. 1967. The genus Cordyceps in Norway. Nytt Magasin for Botanikk 14:68-76.

Evans, H. C. and R. A. Samson. 1982. Cordyceps species and their anamorphs pathogenic on ants (Formicidae) in tropical forest ecosystems. I. The Cephalotes (Myrmicinae) complex.Trans. Br. Mycol. Soc. 79:431-453.

a n d . 1984. Cordyceps species and their anamorphs pathogenic on ants (Formicidae) in tropical forest ecosystems. II. TheCamponotus (Formicinae) complex.Trans. Br. Mycol. Soc. 82:127-150.

Farris, J. S. 1969. A successive approximation approach to character weighting. Syst. Zool. 18: 374-385. 41

. 1970. Methods for computing Wagner trees. Syst. Zool. 19: 83-92.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using bootstrap. Evol. 39: 783-791.

Fletcher, L. R. and I. C. Harvey. 1981. An association of aLolium endophyte with ryegrass staggers. New Zealand Veterin. J. 29: 185-186.

Gaumann, E. A. and C. W. Dodge. 1928. Comparative Morphology of Fungi. New York:McGraw-Hill Book Co., Inc. p.701.

Gaumann, E. A. 1952. The Fungi: A Description o f Their Morphological Features and Evolutionary Development. New York: Hafner publishing Co. p.420.

Hawksworth, D. L., B. C. Sutton, and G. C. Ainsworth. 1983. Ainsworth and Bisby's Dictionary o f the Fungi. 7th edition. CMI, Kew. p.445.

Illingsworth, C. A., J. H. Andrews, C. Bibeau and M. L. Sogin. 1991. Phylogenetic placement oiAthelia bombacina, Aureobasidium pullalans and Colletotrichum gloeosporioides inferred from sequence comparisons of small-subunit ribosomals RNAs. Exp. Mycol. 15:65-75.

Kaltenboeck, B., J. W. Spatafora, X. Zhang, K. Kousoulas, M. Blackwell, and J. Storz. 1992. Efficient production of single-stranded DNA as long as 2 kb for sequencing of PCR-amplified DNA. Biotechniques 12: 164-171.

Lee, S. B. and J. W. Taylor. 1990. Isolation of total DNA from fungi for amplification by the polymerase chain reaction, p. 282-287. In Innis, M. A., D. H. Gelfand, J. J. Sninsky and T. J. White (eds.) PCR Protocols: A guide to Methods and Applications. Academic Press, N. Y.

Luttrell, E. S. 1951. Taxonomy of the pyrenomycetes.University o f Missouri Studies. 3:1-120.

Mains, E. B. 1957. Species of Cordyceps parasitic onElaphomyces. Bull. Torrey Bot. Club 84:243-251.

. 1958. North American entomogenous species ofCordyceps. Mycologia 50:169-222.

. 1959. Species of Hypocrella. Mycopathologia et Mycologia Applicata. 11:311-326.

Matossian, M. K. 1989. Poisons o f the Past: Molds, Epidemics and History. New Haven and London: Yale University Press, p. 190.

Miller, J. H. 1949. A revision of the classification of the ascomycetes with special emphasis on the pyrenomycetes. Mycologia 33: 74-81.

Morgan-Jones, G. and W. Gams. 1982. Notes on Hyphomycetes. XLI. An endophyte ofFestuca arundinaceae and the anamorph ofEpichloe typhina, 42

new taxa in one of two new sectionsAcremonium. of Mycotaxon 15: 311- 318.

Morgan-Jones, G., R. A. Phelps, and J. F. White. 1992. Systematic and biological studies in the Balansieae and related anamorphs. I. Prologue. Mycotaxon 43: 401-415.

Mullis, K. B. and F. A. Fallona. 1987. Specific synthesis of DNAin vitro via a polymerase-catalyzed chain reaction. Meth. Enzym. 155: 335-350.

Price, P. W. 1988. An overview of organismal interactions in ecosystems in evolutionary and ecological time. Agri. Ecosyst. and Envim. 24: 369-377.

Prestridge, R. A., R. P. Pottinger, and G. M. Barker. 1982. An association of Lolium endophyte with ryegrass resistance to Argentine stem weevil. Proc. New Zealand Weed and Pest Control Conference 35: 119-122.

Reynolds D. R. 1981. (ed.) Ascomycetes Systematics: the Luttrellian Concept. Springer-Verlag, N.Y. p.242.

Rogerson, C. T. 1970. The hypocrealean fungi (Ascomycetes, Hypocreales). Mycologia 62:865-910.

Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Huiguchi, G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with thermostable DNA polymerase. Science 239: 487-491.

Samson, R. A. and H. C. Evans. 1985. New and rare entomogenous fungi from Amazonia. Proc. Indian Acad. ScL 94:309-317. Samuels, G. J. and A. Y. Rossman. 1979. Conidia and classification of the nectrioid fungi, p.167-182. In Kendrick, B. (ed.), The whole fungus. Vol 1. Nat'l Mus. Nat. Sci., Nat'l Mus. Canada, Ottawa.

and K. A. Seifert. 1987. Taxonomic implications of variation among hypocrealean anamorphs. p.29-56. In Sugiyama, J. (ed.) Pleomorphic Fungi: the Diversity and Its Taxonomic Implications. Kodansha, Tokyo and Elsevier, Amsterdam.

Sherwood, M. A. 1977. The ostropalean fungi.Mycotaxon 5: 1-277.

Stovall, M. E. and K. Clay. 1991. Fungitoxic effects ofBalansia cyperi. Mycologia 83: 288-295.

Swofford, D. L. 1990. PAUP: Phylogenetic analysis using parsimony, Version 3.0. Illinois Natural History Survey, Campaign, IL.

Wehmeyer, L E. 1975. The pyrenomycetous fungi. Mycologia Memoir no. 6. p.250. 43

West, C. P., P. M. Oosterhuis, and S. D. Wullschleger. 1990. Osmotic adjustment in tissues of tall fescue in response to water deficit. Environ. Exp. Bpt. 30:149-156.

White, J. F. 1992. Endophyte-host associations in grasses. XVII. Ecological and physiological features characterizingEpichloe typhina and some anamorphic varieties in England. Mycologia 84:431-441.

White, T. J., T. D. Bruns, S. B. Lee and J. W. Taylor. 1990. Analysis of phylogenetic relationships by amplification and direct sequencing of ribosomal DNA genes, pp.315-322. In Innis, M. A., D. H. Gelfand, J. J. Sninsky, and T. J. White (eds.) PCR Protocols: A Guide to Methods and Applications. Academic Press, N. Y. CHAPTER 3

THE POLYPHYLETIC ORIGINS OF OPHIOSTOMATALEAN FUNGI

Fungi that are placed in the genera Ceratocystis Ellis & Halst.,

Sphaeronaemella Marchal and Ophiostoma H. & P. Sydow have been the cause of taxonomic debate since they were first recognized. They are characterized as possessing spherical ascoma with elongated perithecial necks, although their asci have been reported as being produced in a scattered manner from an ill-defined hymenium within the ascoma. This paradox has led some investigators to classify these organisms among the pyrenomycetes (Muller and von Arx, 1973; Barr,

1990), whereas others placed them among the plectomycetes (Nannfeldt, 1932; Luttrell, 1951; Benny and Kimbrough, 1980). The ascus wall deliquesces prior to ascospore maturity resulting in free ascospores that are encased in a gelatinous matrix. At maturity they are exuded through the ostiole of a long perithecial neck in a droplet. The ascospores are hyaline and variously shaped according to species. Furthermore, these genera display a diversity of asexual states

(anamorphs), all of which display blastic conidium production (Upadhyay, 1981).

The primary carbohydrate component of fungal cell walls is chitin rather than cellulose, as found in plants and many protoctistans (Bartnicki-Garcia,

1968). However, several species of Ophiostoma have been reported as having cellulose as a major component of their cell walls, whereas species ofCeratocytis sensu stricto did not (Smith et a l , 1967; Jewell, 1974). This same pattern of carbohydrate distribution was documented for rhamnose (Weijman and de Hoog,

44 45

1975). Also, species ofOphiostoma are insensitive to cylcohexamide, whereas

Ceratocystis s. s. is sensitive. Species of Ceratocystis s. s. exhibit enteroblastic production of conidia, whereas it is holoblastic inOphiostoma. All of these characters (Table 3.1) have been utilized in separating Ophiostoma from

Ceratocystis Ellis & Halst.

The genera Ceratocystis and Ophiostoma comprise some of the better known plant pathogens; examples include rot of sweet potatoes and wilt of coffee and rubber caused by C. fimbriata, and Dutch elm disease caused by O. ulmi.

There are saprophytic species, some of which result in the blue-staining of lumber. Also, there are taxa that exhibit pathogenicity towards humans.

Sporothrix schenckii, a conidial fungus, is the causal agent of sporotrichosis and has been linked with the genus Ophiostoma (Mariat, 1971; Berbee and Taylor,

1992c). For these reasons, and others, ophiostomatalean fungi are of obvious economic importance.

Species from these genera exhibit complex symbiotic relationships with insects. Members of Ceratocystis are sometimes associated with ambrosia beetles (Scolytidae and Platypodidae), but tend to form less specialized host relationships; species ofOphiostoma are identified with bark beetles (Scolytidae)

(Dowding, 1984; von Arx and van der Walt, 1988). Sphaeronaemella fimicola is a coprophilous fungus associated with mites that are phoretic on flies (Cain and

Weresub, 1957; Malloch and Blackwell, 1992b). Insect dispersal of this nature has been implicated as a strong selection force in these fungi, and is proposed to have resulted in morphological convergence of ascomata, ascospores and anamorphs among these taxa and throughout ascomycetes (Cain and Weresub,

1957; Blackwell and Malloch, 1992a; 1992b). 46

Table 3.1 Nonmolecular characters of conidiogenesis, antibiotic sensitivity, cell wall carbohydrate composition and production of asci that have been used to separate Ceratocystis s. s. and Ophiostoma H. & P. Sydow.

Ceratocystis sensu stricto Ophiostoma H. & P. Sydow

Enteroblastic anamorphs Holoblastic anamorphs (deHoog, 1974) (de Hoog, 1974)

Cyclohexamide sensitive Cyclohexamide insensitive (Harrington, 1981) (Harrington, 1981)

Cellulose absent Cellulose present (Jewell, 1974; Smith et al, 1974) (Smith etal., 1967; Jewell, 1974)

Rhamnose absent Rhamnose present (Weijman & de Hoog, 1975; (Weijman & de Hoog, 1975; Spencer & Gorin, 1974) Spenser & Gorin, 1974)

Young asci line periphery Young asci produced from base of of inner perithecium inner perithecium (van Wyk et al., 1991) (van Wyk et al., 1991) 47

The questions remain - how closely related are the species of these three genera, and where do they belong with respect to higher-level (ordinal) classification schemes? They have been treated as congeneric (Luttrell, 1951;

Upadhyay, 1981), in different families of the same order (de Hoog, 1974; von Arx and van der Walt, 1988), and as genera of separate orders (Barr, 1990).

Sphaeronaemella has been treated as congeneric with Ceratocystis (Upadhyay,

1981), as a genus in the Ophiostomataceae (Benny and Kimbrough, 1980), and as a member of the Hypocreales (Barr, 1990).

Ceratocystis sensu lato was perceived as comprising four groups defined by ascospore morphology (Olchowecki and Reid, 1974). TheMinuta-group’ is distinguished by elongate, often curved ascospores with terminally diffused sheaths; the 'Ips-group' is characterized by cylindrical to dumbbell-shaped ascospores with gelatinous sheaths; the'Pilifera-group' possesses curved, ovoidal or cylindrical ascospores without sheaths; and, the 'Fimbriata-group' is a heterogeneous assemblage consisting of the remainingCeratocystis species. The current taxonomic treatment of these fungi equates theFimbriata -group' with

Ceratocystis s.s. (de Hoog, 1974; von Arx, 1974; de Hoog and Scheffer, 1984), the

'Piliferum-group' and 'Ips-group' with Ophiostoma H. & P. Sydow (de Hoog, 1974; von Arx, 1974; de Hoog and Scheffer, 1984; Harrington, 1987), and the'Minuta- group'With Ceratocystiopsis (Siem.) Upahyay and Kendrick (Upadhyay, 1981).

Excellent reviews of the taxonomic history of this group of fungi are found in de

Hoog (1974), Upadhyay (1981), de Hoog and Scheffer (1984) and von Arx and van der Walt (1988).

The focus of this study is to infer the ordinal level taxonomic placement and propose testable phylogenetic hypotheses concerning the evolution of the taxa sampled from the three genera. These questions are approached by cladistic 48 analysis of nucleotide characters sampled from the nuclear small subunit ribosomal DNA (ssrDNA). This analysis provides an independent test of morphological data and allows for the inference of supraordinal level relationships. Furthermore, this study will provide a framework for inferring relationships of other ophiostomatalean fungi.

Materials and Methods

The taxa sampled for this study were chosen based upon their position in past and contemporary classification systems and possession of particular morphological and ecological characters. Taxa were sampled so as to include the type species of Ceratocystis s. s., Ophiostoma and Sphaeronaemella fimicola. The taxa, source of isolate and current ordinal placement (Hawksworthet. al., 1983) are listed in Table 1. The ingroup consists of 25 unitunicate, perithecial ascomycetes, Aureohasidium pullulans and Aspergillis jumigatus; the outgroup comprises several species of yeasts and Taphrina deformans.

The following procedures were performed as described previously in

Chapter 1: culturing and harvesting of organisms; extraction of total nucleic acid and assay for presence and high molecular weight DNA; symmetric and asymmetric polymerase chain reactions (PCRs); single-stranded PCR (ssPCR) product purification; sequencing reactions, polyacrylamide gel electrophoresis and autoradiography; sequence alignments and data analysis; calculation of decay indices and bootstrap confidence levels.

The rDNA data set constructed in this study consists of 34 taxa and approximately 900 bp/taxon, 450 bp of sequence from each of the two regions designated by the primer pairs NS1/NS2 and NS3/NS4 (White et al, 1990). The 49

Table 3.2 List of taxa included in study, source of taxon and regions sequenced. Primers are those of Whiteet al. (1990).

Ambrosiozyma platypodis (Baker & Kreger ATCC 36174 NS, NS2, NS3, NS4 - van Rij) van der Walt Aspergillus fumigatus Fr. Genbank Asnda Aureobasidium pullulans (de Bary) Amaud Illingsworthet al., 1991 Balansia clserotica (Pat.) Hoehn. ATCC 16582 NS1.NS2, NS3.NS4 Candida albicans (Robin) Berkout Genbank Ysall6s Candida tropicalis (Castellani) Berkhout Genbank Ysarmab Cephaloascus fragrans Hanawa ATCC 36174 NS1.NS2, NS3.NS4 Ceratocystis fimbriata Ellis & Halst. Harrington C89NS1.NS2, NS3, NS4 Ceratocystis virescens Davids. Harrington C69 NS1.NS3 Cercophora septentrionalis Lundq. D. Malloch NS1.NS2, NS3 Chaetomium globosum Kunze ex Fr. ATCC 44699 NS1.NS2, NS3 Claviceps paspali Stevens & Hall ATCC 13892 NS2, NS4 Daldinia concentrica (Bulliard ex Fr.) ATCC 36659 NS1.NS3.NS4 Cesati et de Notaris Diaporthe phaseolarum (Cookeet Ellis) FAU-458 NS2, NS4 Diatrype discoformis (Hofftn.:Fr.) Fr. CBS 197.49 NS2, NS4 Haloasphaeriopsis mediosetigera (Kohl.) Kohl. ATCC 16934 NS1.NS3.NS4 Hypocrea schweinitzii (Fr.)Sacc. CTR 79-225 NS1.NS2, NS3.NS4 Hypomyces polyporinus Peck ATCC 46844 NS2, NS4 Hypoxylon atroroseum J. Rogers NS1.NS3 Leucostoma persoonii (Nits.) Hoehn. Berbee & Taylor, 1992c Microascus trigonosporus Emmons & Dodge RSA 1942 NS1.NS2, NS3.NS4 Nectria cinnabarina (Tode ex Fr.) Fr. GJS 89-107 NS1, NS2, NS3, NS4 Nectria haematococca Berk.& Br . GJS 89-97 NS1.NS2, NS3, NS4 Neocosmospora vasinfecta Smith RSA 1898 NS2, NS4 Neurospora crassa Shear & Dodge Genbank Neurrnas Ophiostoma piliferum H. & P. Syd. Harrington C300 NS2, NS4 Ophiostoma ulmi (Buisman) Nannf. Berbee & Taylor, 1992c Petriella setifera (Schmidt) Curzi ATCC 26490 NS1.NS2, NS3 Saccharomyces cerevisieae Hansen Genbank Yscrgea Sphaeronaemella fimicola Marchal M. Blackwell NS1, NS3, NS4 Sphaerostilbella aureonitens (Tul.) Seifert et al.. GJS 83-286 NS1.NS2, NS3, NS4 Taphrina deformans (Fuckel) Tul. ATCC 11124 NS1.NS3 Xylaria curia Fr. J. Rogers NS2, NS3, NS4 Xylaria hypoxylon (L.:Fr.) Grev. ATCC 42768 NS2, NS3, NS4 50 region defined by NS1/NS4 is approximately 1200 bp in length and corresponds to the 5' half of the ssrDNA. It was chosen based on a preliminary screening that demonstrated the 3' half of the molecule was considerably more conserved and provided fewer synapomorhies than the 5' end. This pattern of nucleotide variation is consistent with that observed for secondary structure for the ssrRNA

(Gutell, 1985). Furthermore, the 3' region possessed intervening sequences (IVS) in several taxa, which complicated the acquisition of a complete data set for all taxa in question.

Results

Phylogenetic analysis of the ssrDNA data produced eight equally most parsimonious trees of 476 steps (Fig. 3.1) with consistency (Cl), retention (RI) and rescaled consistency (RC) indices of .536, .743 and .398, respectively (Kluge and Farris, 1969). A strict consensus of these eight trees is 481 steps in length with a Cl of .530, a RI of .736 and a RC of .390. Of the 900 bp sampled, 173 are phylogenetically informative.

The pyrenomycetous fungi sampled in this study form a clade that can be viewed as comprising two subelades, A and B (Fig. 3.1). In subclade A, C. fimbriata, C. virescens and S. fimicola form a monophyletic clade, which is a sister group to the clade containing taxa from the Microascales andHalosphaeriopsis mediosetigera (Fig. 3.1). This grouping is supported by bootstrap confidence levels of 99% and a decay index of > 10 steps when S. fimicola is excluded from the analysis; however, the inclusion ofS. fimicola reduces the bootstrap confidence level to 82% and the decay index to 6 steps (Fig. 3.2; Fig. 3.3). The clade comprising the Microascales and the genera Ceratocystis and 51

Sphaeronaemella is a sister group to the taxa sampled from the Hypocreales and

Clavicipitales (Fig. 3.1). Subclade B contains taxa from the orders Diaporthales,

Sordariales and Xylariales and the genus Ophiostoma. The taxa sampled for

Ophiostoma are monophyletic and a sister group to the taxa sampled for the

Diaporthales. The clade comprising Ophiostoma and the Diaporthales is a sister group to the members sampled for the Sordariales. The supraordinal level relationships within in this clade are not strongly supported by bootstrap values or decay indices (Fig. 2.3).

Aspergillis fumigatus and Aureobasidium pullulans are sister taxa positioned between the derived pyrenomycetes and the outgroup. The unitunicate, perithecial ascomycetes sampled in this study form a monophyletic group that is supported by a bootstrap value of 100% and a decay index of > 10 steps. The monophyly of the ingroup was maintained with respect to the outgroup. However, this taxonomic sampling does not address the monophyly of

Pyrenomycetes (sensu Luttrell, 1951) as a natural grouping due to incomplete taxon sampling. This question is considered in Chapter 4.

Discussion

Ophiostomatalean fungi epitomize problematic taxa in ascomycete systematics. This is exemplified by the numerous classifications for these taxa that exist in the literature. At the heart of these taxonomic debates have been the organisms comprising the generaCeratocystis and Ophiostoma. They have been considered congeneric in the genera Ceratostomella Sacc. (Elliot, 1925),

Ophiostoma H. & P. Sydow (Luttrell, 1951) and Ceratocystis Ellis & Halst. (Hunt,

1956; Upadhyay, 1981). Species have been treated as separate genera, 52

■ Hypomyces potyporinus 94 — Hypocrea schweinitzii Sphaerostilbella aureonilens HYPOCREALES r Nectria haemalococca

II Neocosmospora vasinfecta L— Nectria cinnabarina 9 01.... Balansia sclerotica CLAVICIPITALES sL Clavicepsa paspali 100100.0 j Ceratocystis fimbriata Ceratocystis virescens OPHIOSTOMATALES

82 ri———— — Sphaeronaemella fimicola Halosphaeriopsis mediosetigera Petriella selifera MICROASCALES 99 I MicroascusMici trigonosporus >10 99 I Diaporthe phaseolarum DIAPORTHALES 9 I— Leucostoma persoonii 100 Ophiostoma piliferum OPHIOSTOMATALES >10 *■ Ophiostoma ulmi 95i“ Cercophoraseptenlrionalis 9 917L Chaetomium globosm SORDARIALES ^ I— Neurospora crassa B Daldinia concentrica rC Hypoxylon atroroseum XYLARIALES Xylaria curta Xylaria hypoxylon ■ Aspergillus fumigatus EUROTIALES h: Aureobasidium pullulans DOTHIDEALES • Ambrosioiyma platypodis rCZSaccharomyces cerevisiae 100 95 j— Candida albicans ENDOMYCETALES >10 1 L Candida tropicalis — Cephaloascus fragrans ■ Taphrina deformans I TAPHRINALES

Figure 3.1 Strict consensus of eight equally most parsimonious cladograms containing 33 taxa (481 steps, Cl = .530, RI = .736 and RC = .390). The designated outgroup is T. deformans and the Endomycetales. Bootstrap confidence levels of 90% or greater from 100 bootstrap replications and decay indices are given of above and below the corresponding nodes, respectively. Branch lengths are proportional to the numbers of steps. (*See Fig. 3.2; Fig. 3.3). 53

— Hypomyces polyporinus 96 —— Hypocrea schweinitzii — Sphaerostilbella aureonitens I" Nectria haematococca j* Neocosmospora vasinfecta Nectria cinnabarina A 100 ii .....—- Balansia sclerotica 94 ■ c Claviceps paspali Microascus trigonosporus Petriella setifera 99 i -I r ^ HalosphaeriopsisHa mediosetigera 10 100 Ceratocystis fimbriata > 10 €1U E— n a Ceratocystis virescens 100 98 g | Diaportheu ic phaseolarum 10 1 1 Leucostoma persoonii 100 r Ophiostoma piliferum >10 Ophiostoma ulmi 9898]— j— Cercophora< septentrionalis loo pr L Chaetomium globosum 97 >10 I Neurospora crassa B Daldinia concentrica Hypoxylon atroroseum Xylaria curta Xylaria hypoxylon Aspergillus fumigatus Aureobasidium pullulans Ambrosiozyma platypodis — Saccharomyces cerevisiae 100 93 b- Candida albicans >10 6 *- L CanCandida tropicalis Cephaloascus fragrans Taphrina deformans

Figure 3.2 Strict consensus of eight equally most parsimonious cladograms (424 steps, Cl = .559, RI = .769, RC = .430). S. fimicola has been excluded from the analysis. Note that the bootstrap values are elevated significantly in subclade A (Fig. 3.1), and the clade containing the Microascales and Ceratocystis (Fig. 3.3). The dampening effect of S. fimicola on decay indices appears to be more localized with respect to subclade A (Fig. 3.1). 54

Ceratocystis fimbriata

Ceratocystis virescens

Halosphaeriopsis mediosetigera a. 99%

Petriella setifera

Microascus trigonosporus

Ceratocystis fimbriata

Ceratocystis virescens

Sphaeronaemellafimicola b. 82% Halosphaeriopsis mediosetigera

Petriella setifera

Microascus trigonosporus

Ceratocystis fimbriata

Sphaeronaemella fimicola

c. 82% Halosphaeriopsis mediosetigera

Petriella setifera

Microascus trigonosporus

Figure 3.3 Bootstrap confidence levels for three different taxon samplings within the MicTOSLScales-Ceratocystis clade. (a) 5 taxa without S. fim icola, (b) 6 taxa with S. fim icola, and (c) 5 taxa with S. fim icola. Support for the clade is contingent on the inclusion or exclusion ofS. fim icola. Branch lengths are not proportional to the number of steps. 55

Ceratocystis s. s. and Ophiostoma H. & P. Sydow, in different families of the same order (de Hoog, 1974; von Arx, 1974 ; de Hoog and Scheffer, 1984; von Arx and van der Walt, 1987), and as genera of separate orders (Barr, 1990). Also, they have been placed in the Endomycetaceae (Redhead and Malloch, 1977), based primarily upon the presence of galeate ascospores, or as close relatives to the yeasts (von Arx and van der Walt, 1987). Sphaeronaemella Marchal has been treated as congeneric with Ceratocystis sensu lato (Upadhyay, 1981), as a genus in the Ophiostomataceae (Benny and Kimbrough, 1980), and as a member of the

Hypocreales (Samson, 1974; Barr, 1990).

Cladistic analysis of partial sequences of the ssrDNA rejects the monophyly of the ophiostomatalean taxa sampled in this analysis and therefore the congeneric treatments. It supports the monophyly of the generaCeratocystis s. s. and Ophiostoma H. & P. Sydow (de Hoog, 1974; von Arx, 1974) and places them in separate clades among pyrenomycetes (Fig. 3.1). In doing so, it does not agree in entirety with any one of the classifications mentioned, but rather with different components of several. Of interest are the particulars that the ssrDNA phylogeny has in common and in conflict with the classifications of Luttrell

(1951), Upadhyay (1981) and Barr (1990).

These data are inconsistent with Luttrell (1951) on two basic points. First, that Ophiostomatalean taxa are not congeneric, and second, that the

Microascales, including Ophiostomataceae, is not within the Plectomycetes. The two hypotheses are consistent in that some ophiostomatalean species,le.

Ceratocystis s. s., share a recent common ancestor with the Microascales; also, it does not reject the familial ranking within the Microascales. Luttrell placed the

Microascales, and the Ophiostomataceae, in the Plectomycetes based upon the presence of scattered asci. These data do not support the homology of scattered 56 asci, and suggest that the gross morphology of possessing scattered asci in a mature ascoma has evolved several times (Berbee and Taylor, 1992c). Thus scattered asci is not a valid character for assessing taxonomic relationships.

Upadhyay (1981) placed Sphaeronaemella into synonymy with Ceratocystis; a relationship that is generally not accepted (de Hoog and Scheffer, 1984). These data position S. fimicola as a putative sister taxon to the taxa sampled for

Ceratocystis', however, it is on a relatively long branch length. The long branch length may reflect either an erroneous grouping of taxa resulting from the attraction of long branches within a cladogram (Felsenstein, 1978), or unequal rates of evolution within a monophyletic group. The relationship has been questioned by several investigators (Samson, 1974; von Arx, 1974; 1981; Barr,

1990) all of whom placedSphaeronaemella in the Hypocreales. Their arguments lay in that Sphaeronaemella differs from Ceratocystis in morphology of ascomata, ascospores and conidium ontogeny.

Ascospores in S. fimicola are described as being allantoid in lateral view and oblong to ellipsoidal in front view (Upadhyay, 1981). A study of ascosporogenesis in C.moniliformis showed that it is the ascospore sheath that is hat-shaped and the spore itself is allantoid (van Wyk et al., 1991), although there are differences in the number of electron transparent layers described for the ascospores sheaths for different species ofCeratocystis s. s. The ascomata in

Sphaeronaemella are pale in color compared to that of some species of

Ceratocystis, and the neck, or beak of the perithecium, may be shorter (Samson,

1974; Upadhay, 1981). However, both genera exude their ascospores in a droplet at the tip of the perithecium. An additional anomaly is the difference in anamorphs between the two teleomorph genera.Sphaeronaemella spp. possess

Gabbarnaudia anamorphs while species ofCeratocystis sensu stricto have Chalara 57 anamorphs. Although both are enteroblastic phialidic, the presence of a

Gabbarnaudia anamorph is considered to be a link with the Hypocreales

(Samson, 1974; von Arx, 1974 & 1981), which also display anamorphs with enteroblastic phialidic condiogenous cells. The question of support for the long branch, which terminates in the taxon

S. fimicola, was approached in two ways. The first is inherent in the taxonomic sampling and parsimony analysis. This study includes a number of taxa that represent a broad range of morphological variation across ascomata, ascospores and anamorphs. None of the taxa added after the initial inclusion ofS. fimicola attracted it away from Ceratocystis. The second involved bootstrapping

(Felsenstein, 1985) and jackknifing (Lanyon, 1985). The placement of S. fimicola and the two species ofCeratocystis as sister taxa to the Microascales is not well supported in a bootstrap analysis (Fig. 3.1; Fig. 3.3). WhenS. fimicola is excluded from the analysis the two species ofCeratocystis are maintained as sister taxa to the Microascales, and their relationship to the order is well supported

(Fig. 3.2; Fig. 3.3). These taxonomic statements differ in the number of taxa, which may affect interpretation of the results; however, if C.virescens is omitted from the analysis, S. fimicola and C. fimbriata are not well supported as sister taxa to the Microascales (Fig. 3.3). Furthermore, maximum parsimony placesS. fimicola as a sister taxon to the Microascales when both species Ceratocystisof are excluded from the analysis (data not shown). Jackknifing, the exclusion of a taxon followed by reanalysis, suggests that the placement ofS. fimicola is relatively stable. Assuming that the placement and long branch length are valid, this may represent a lineage that has undergone host mediated isolation that resulted in an accelerated rate of evolution. That is, different dispersal agents resulted in effective isolation of ancestral populations (Cain and Weresub, 1957; 58

Malloch, 1979; Malloch and Blackwell, 1992). As noted previously,

Sphaeronaemella and Ceratocystis are associated with different insect dispersal agents and substrates. It will now be important to include other species of Ceratocystis, Sphaeronaemella, Gabbarnaudia and other conidial taxa to test the stability of this placement and branch length.

Barr (1990) proposed thatCeratocystis s. s. and Ophiostoma are pyrenomycetes of separate orders. She assigned the taxa that bear enteroblastic phialidic conidiogenous cells and lack rhamnose and cellulose in their cell walls to Ceratocystis (Lasiosphaeriaceae, Sordariales). The placement ofCeratocystis in the Sordariales stems from putative affinities with Melanochaeta (Samuels and

Muller, 1979) and Chaetosphaeria (von Arx, 1981).Ophiostoma was reserved for those taxa that possess holoblastic conidiogenous cells and contain cellulose and rhamnose in their cells walls. Ophiostoma was assigned to the Microascales following Luttrell (1951) and Benny and Kimbrough (1980); however, Barr differed over the order's plectomycete affinity. The separation of the two genera based upon these characters, and others, is reviewed in Table 3.1. The phylogeny proposed here is almost the reverse of that offered by Barr (1990). Cladistic analysis of the ssrDNA places the species sampled for Ceratocystis, and

Sphaeronaemella, as sister taxa to the Microascales. The taxa sampled for

Ophiostoma are a sister group to the Diaporthales, which is in turn a sister group to the Sordariales.

Of particular interest is the correlation of centrum development of

Ceratocystis and Ophiostoma and their respective sister groups inferred in this study. In C. moniliformis young asci are produced from ascogenous cells that line, or are adjacent to, the inner perithecial (centrum) wall; young asci are then freed into the centrum cavity; paraphyses are absent (van Wyk et al, 1991). 59

Development of the asci and ascospores ofC. moniliformis appear similar to

those ofC. fimbriata (Elliott, 1925; Steirs, 1976; van Wyk et al., 1991). In Microascus and Petriella asci are produced terminally from ascogenous cells that grow downward into the centrum (Corlett, 1963; 1966). The asci become irregularly dispersed in the centrum in a manner superficially similar to C. moniliformis, although, asci remain attached to the ascogenous hyphae.

Paraphyses are lacking, although, sterile cells develop from the peridium inward into the centrum then evanesce. In both, C. moniliformis and the microascaceous species, asci are produced from an ill-defined hymenium, ascogenous hyphae do not form croziers, asci fill the centrum cavity and ascus walls deliquesce prior to ascospore release. The development of the centrum is markedly different in Ophiostoma compared to that ofCeratocystis (van Wyk and Wingfield, 1991b). In O. ulmi

(Rosinski, 1961), O. davidsonii and O. cucullatum (van Wyk and Wingfield,

1991b; 1991c) the production of asci is restricted to the base of the perithecium.

The asci are then released upward into the centrum cavity. No paraphyses were described; however, the inner ascomatal wall of O. davidsonii and O. cucullatum possess large sterile cells, which deliquesce prior to ascospore maturity. On the other hand, paraphysis-like bands were described for O. multiannulata (Andrus,

1936) and O. major (Hutchinson, 1950). No such sterile cells have been described in Ceratocystis. The presence of scattered asci in the mature ascomata of the genera Ceratocystis and Ophiostoma is , therefore, the result of convergent or parallel evolution,i.e. nonhomologous. The sister group toOphiostoma in this analysis is the Diaporthales, species of which also produce asci from a basal hymenium and release their asci into the centrum cavity (Luttrell, 1951; Jensen, 60

1983). Centrum anatomy in the context of derived pyrenomycetes is discussed at greater length in Chapter 5.

This separation ofCeratocystis and Ophiostoma by the ssrDNA corroborates differences in ascospore production, delimitation and morphology.

The ascospores of both genera are small and hyaline, and in both some species produce galeate, or hat-shaped, ascospores. However, the galeate morphology is produced by two different mechanisms. In C. moniliformis (van Wyk et al., 1991) and C. fimbriata (Stiers, 1976) young ascospores form pairs that become encased in a gelatinous sheath; individual spores are delimited by cell wall development resulting in hat-shaped ascospores. However, there is variation, or perhaps discrepancy, in the mechanism of cell wall synthesis (Stiers, 1976; van Wyk et al.,

1991). In O. stenoceras (Garrison, 1979),O. ulmi (Rosinsky, 1961; Jengs and

Hubbes, 1980),and O. davidsonii (van Wyk and Wingfield, 1991b ) ascospores do not form pairs during development, and cell wall synthesis occurs between a double ascospore delimiting membrane. This double delimiting membrane was not observed in C. fimbriata (Stiers, 1976) or C. moniliformis (van Wyk et al.,

1991).

Characters derived from conidiogenesis appear to be inconsistent with the ssrDNA gene tree. The anamorphs of the generaCeratocystis and

Sphaeronaemella are Chalara and Gabbamaudia, respectively. Both exhibit enteroblastic phialidic conidiogenous cells; however, the microascalean anamorphs are holoblastic annellidic, exemplified by the genus Scopulariopsis.

The overwhelming majority of taxa in subclade A possess enteroblastic conidiogenous cells, so the microascalean taxa are exceptional in this clade. But

Malloch (1979) suggested that in some lines of evolution annellides may be derived from phialides. To complicate the matter further, Ophiostoma is 61 characterized as having holoblastic condiogenous cells, e.g. Sporothrix, whereas those in its sister group, the Diaporthales, are enteroblastic,e.g. Phomopsis.

Moreover, the majority of the taxa sampled in subclade B bear anamorphs that exhibit holoblastic condiogenous cells. The difference in conidiogenesis, defined as holo- and enteroblastic, follows Hawksworthet al. (1983). Minter et al. (1983) redefined the terms holoblastic and enteroblastic by emphasizing the different stages of conidiogenesis: conidial ontogeny, delimitation, secession and proliferation and regeneration of conidiogenous cells. They suggested that holoblastic sympodial proliferation results in larger, dryer conidia befitting wind dispersal, and enteroblastic percurrent proliferation results in smaller, mucous conidia for insect dispersal. Using the redefined terminology, all the taxa within the derived pyrenomycete clade, as far as I am aware, exhibit holoblastic conidial ontogeny except forCeratocystis s. s., which produces Chalara anamorphs. These data question the usefulness of characters derived from conidiogenesis at the supraordinal level, but further developmental studies are warranted, especially for the form genusChalara.

Cladistic analysis of partial nucleotide sequences from the ssrDNA rejects the monophyly of the Ophiostomatalean taxa sampled. It supports the exclusion of Ophiostoma from Ceratocystis Ellis & Halst. and places the two genera in different clades of unitunicate, perithecial ascomycetes. Ceratocystis s. s. and

Sphaeronaemella fimicola are sister taxa, and are proposed to share a recent common ancestor with members of the Microascales. This relationship to the

Microascales is not supported strongly by bootstrap confidence levels whenS. fimicola is included; however, when S. fimicola is excluded Ceratocystis is supported significantly as a sister taxon to the Microascales. Conversely, limited jackknifing suggests that the placement of S. fimicola is relatively stable. 62

Ophiostoma is placed in a separate part of the tree as a sister taxon to the

Diaporthales. This relationship is not supported strongly, although, species of

Ophiostoma and some species ofDiaporthe share similarities in centrum development. These data are consistent with the selection pressure of insect dispersal resulting in both the convergence of ascoma, ascospores and asci in

Ceratocystis and Ophiostoma, and in the divergence of these same characters between Ceratocystis and Sphaeronaemella. This highlights the complex nature of dispersal pressures, if not selection pressures in general; a single evolutionary force may have different mechanistic effects on separate evolutionary units

(groups of taxa).

Admittedly, this analysis does not represent a complete sampling of species of the genera in question. Nonetheless, in supporting the division of

Ceratocystis Ellis & Halst. into Ceratocystis s. s. and Ophiostoma H. & P. Sydow these data acquired from type species correlate with differences in carbohydrate composition of hyphal walls, conidiogenesis, cyclohexamide resistance and centrum development (Table 3.1). The independent test of these characters allows for the inclusion of other species into this phylogeny accordingly.

The instability of bootstrap confidence levels concerning Ceratocystis and

Sphaeronaemella and their placement as sister taxa to the Microascales can be attributed to the long branch length terminating in S. fimicola. A similar phenomenon is observed within the Clavicipitales, but grossly disproportionate branch lengths were not observed (Ch. 2). These findings suggest that bootstrap analyses are sensitive to taxonomic sampling within a monophyletic group.

Presumably, this sensitivity reflects unequal or nonrandom distribution of homoplasy within a clade, and is relative to the taxon sampling for that 63 monophyletic group. This instability questions the level of interpretation of bootstrap confidence levels beyond the gene tree. 64

Literature Cited

Andrus, C. F. 1936. Cell relations in the perithecium of Ceratostomella multiannulata. Mycologia 28: 133-153.

Arx, J. A. von. 1974.The Genera of Fungi Sporulating in Pure Culture. 2nd ed. Cramer, Vaduz, p.315.

1981. The Genera o f Fungi Sporulating in Pure Culture. 3rd. ed. Cramer, Vaduz, p.424

and J. P. van der Walt. 1987. Ophiostomatales and Endomycetales. p. 167-176. In de Hoog, G. S., M. T. Smith, and A. C. M. Weijman (eds.) The Expanding Realm of Yeast-like Fungi. Elsevier, Amsterdam.

Barr, M. E. 1990. Prodromus to nonlichenized, pyrenomycetous members of class Hymenoascomycetes. Mycotaxon 39: 43-184.

Bartnicki-Garcia, S. 1968. Cell wall chemistry, morphogenesis and taxonomy of fungi. Annu. Rev. Microbiol. 22: 87-108.

Benny, G. L. and J. W. Kimbrough. 1980. A synopsis of the orders and families of plectomycetes with keys to genera. Mycotaxon 12: 1-91.

Berbee, M. L. and J. W. Taylor. 1992a. Two ascomycete classes based upon fruiting-body characters and ribosomal DNA sequence. Molec. Biol. Evol 9: 278-284.

a n d ------. 1992b. 18S Ribosomal RNA gene sequence characters place the human pathogen Sporothrix schenckii in the genus Ophiostoma. Exp. Mycol. 16: 87-91.

a n d ------. 1992c. Convergence in ascospore discharge mechanism among pyrenomycete fungi based on 18S ribosomal RNA gene sequence. Molec. Phylog. Evol. 1:59-71.

Cain, R. F. and L. K. Weresub. 1957. Studies of coprophilous ascomycetes. V. Sphaeronaemella fimicola. Can.J.Bot. 35: 119-131.

Corlett, M. 1963. The developmental morphology of two species ofMicroascus. Can. J. Bot. 41: 253-266.

. 1966a. Developmental studies in the Microascaceae. Can. J. Bot. 44: 79-88.

Dowding, P. 1984. The evolution of insect-fungus relationships in the primary invasion of forest timber, p. 133-153. In Anderson, J. M., A. D. M. Rayner, and D. W. H. Walton (eds.) Invertebrate-Microbial Interactions. Cambridge, New York. 65

Elliot, J. A. 1925. A cytological study ofCeratostomella fimbriata (E. & H.) Elliot. Phytopathology 15: 417-422.

Felsenstein, J. 1978. Cases in which parsimony or compatibility methods will be positively misleading. Syst. Zool. 27: 401-410.

Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791.

Garrison, R. G., F. M ariat, K. S. Boyd and H. Fromentin. 1979. Perithecial ultrastructure and formation of ascospores ofCeratocystis stenoceras (Robak) C. Moreau. Annales de Microbiologie 130: 3-21.

Gutell, R ., B. Weiser, C. R. Woese, and H. F. Noller. 1985. Comparative Anatomy of 16-S-like Ribosomal RNA. Prog. Nucleic Acids Res. 32: 155- 216.

Hawksworth, D. L., B. C. Sutton, and G. C. Ainsworth. 1983. Ainsworth and Bisby's Dictionary o f the Fungi. 7th edition. CMI, Kew. p445.

Harrington, T. C. 1987. New combinations inOphiostoma ofCeratocystis species with Leptographium anamorphs. Mycotaxon 28:39-44.

Hoog, G. S. de. 1974. The generaBlastobotrys, Sporothrix, Calcarisporium and Calcarisporiella gen. nov. Stud. Mycol. 7: 1-84.

and R. J. Scheffer. 1984. Ceratocystis versus Ophiostoma: a reappraisal. Mycologia 76: 292-299.

Hunt, J. 1956. Taxonomy of the genusCeratocystis. Lloydia 19: 1-58.

Hutchinson, S. A. 1950. The perithecia of Ophiostoma majus (van Beyma) Goidanich. Ann. Bot. (N. S.) 14:115-125. llingsworth, C. A., J. H. Andrews, C. Bibeau and M. L. Sogin. 1991. Phylogenetic placement ofAthelia bombacina, Aureobasidium pullalans and Colletotrichum gloeosporioides inferred from sequence comparisons of small-subunit ribosomals RNAs. Exp. Mycol. 15:65-75.

Jengs, R. S. and M. Hubbes. 1980. Ultrastructure of Ceratocystis ulmi. II. Ascogenous system and ascosporogenesis.Euro. J. For. Path. 10: 104-116.

Jensen, J. D. 1983. The development ofDiaporthe phaseolarum variety sojae in culture. Mycologia 75: 1074-1091.

Jewell, T. R. 1974. A qualitative study of cellulose distribution in ceratocystis and Europhium. Mycologia 66: 139-146.

Kluge, A. G. and J. S. Farris. 1969. Quantitative phyletics and the evolution of anurans. Syst. Zool. 28: 400-411. 66

Lanyon, S. 1985. Detecting internal inconsistencies in distance data. Syst. Zool. 34: 397-403.

Luttrell, E. S. 1951. Taxonomy of the pyrenomycetes.University o f Missouri Studies 3:1-120.

Malloch, D. 1979. Plectomycetes and their anamorphs. p.153-166. In Kendrick, B. The Whole Fungus, Vol 1. National Museum of Natural Sciences, National museum of Canada. Ottawa. and M. Blackwell. 1992a. The adaptive nature of fungal traits facilitating fungal spore dispersal, pp. 147-172. In Carroll, G. C, and D. T. Wicklow (eds.) The Fungal Community: Its Organization and Role in the Ecosystem. 2nd ed. Dekker, N. Y.

a n d . 1992b. Dispersal biology of ophiostomatoid fungi.In Wingfield, M. J., K. A. Seifert and J. F. Wbber (eds.) Ceratocystis and Ophiostoma:Taxonomy, Ecology and Pathology. APS, St. Paul. In press.

Mariat, F. and E. Dies. 1971. Adaptation de Ceratocystis stenoceras (Robak) C. Moreau a la vie parasitaire chez l'animal. Etude de la souche sauvage et des mutants pathogenes. Comparaison avec Sporothrix schenckii Hektoen et Perkins. Rev. Mycol. 36: 3-14.

Minter, D. W., P. M. Kirk, and B. C. Sutton. 1982. Holoblastic phialides. Trans. Br. Mycol. Soc. 79: 75-93.

, B. C. Sutton, and B. L. Brady. 1983. What are phialides anyway? Trans. Br. Mycol. Soc. 81: 109-120.

Muller, E. and J. A. von Arx. 1973. Pyrenomycetes: Meliolales, coronophorales, Sphaeriales. p.87-132. In Ainsworth, G. G, F. K. Sparrow, and A. S. Sussman (eds.) The Fungi, Vol. IVA. Academic Press, New York, NY.

Nannfeldt, J. A. 1932. Studien iiber die morphologie und systematik der nichlichenisieten inoperculaten Discomyceten. Nova Acta Regiae Sco. Scien. Upsal., Ser. IV. 8:1-368.

Olchowecki, A. and J. Reid. 1974. Taxonomy of the genusCeratocystis in Manitoba. Can J. Bot. 52: 1675-1711.

Redhead, S. A. and D. Malloch. 1977. The Endomycetaceae: new concepts, new taxa. Can. Jour. Bot. 55: 1701-1711.

Rosinski, M. A. 1961. Development of the ascocarp ofCeratocystis ulmi. Amer. J. Bot. 48: 285-293.

and R. J. Campana. 1964. Chemical analysis of the cell wall of Ceratocystis ulmi. Mycologia 56: 738-744. 67

Samuels, G. J. and E. Muller. 1979b. Life-history studies of Brazilian ascomycetes 5. Two new species ofOphiostoma and their Sporothrix anamorphs. Sydowia 31: 169-179. Samson, R. A. 1974. Paecilomyces and some allied Hyphomycetes. Stud. Mycol 6: 1-119. Spencer, J. F. T. and P. A. J. Gorin. 1971. Systematics of the generaCeratocystis and Graphium. Proton magnetic resonance spectra of the mannose- containing polysaccharides as an aid in classification. Mycologia 63: 387- 402.

Smith, M. L., C. M. Patik, and M. A. Rosinski. 1967 . A comparison of cellulose production in the genus Ceratocystis. Mycologia 59: 965-969.

Stiers, D. L. 1976. The fine structure of ascospore formation inCeratocystis fimbriata. Can. J. Bot. 54: 1714-1723.

Sutton, B. C. 1980. The Coelomycetes. CMI. Kew. p.696. Upadhyay, H. P. 1981. A Monograph o f Ceratocystis and Ceratocystiopsis. Univ. Georgia Press, Athens, p. 176.

Weijman, A. C. M. and G. S. de Hoog. 1975. on a subdivision of the genus Ceratocystis. Anton v. Leeuw. 41: 353-360.

White, T. J., T. D. Bruns, S. B. Lee and J. W. Taylor. 1990. Analysis of phylogenetic relationships by amplification and direct sequencing of ribosomal DNA genes, pp.315-322. In Innis, M. A., D. H. Gelfand, J. J. Sninsky, and T. J. White (eds.). PCR Protocols: A Guide to Methods and Applications. Academic Press, N. Y.

Wyk, P. W. J. van, M. J. Wingfield and P. S. van Wyk. 1991. Ascospore development on Ceratocystis moniliformis. Mycol Res. 95:96-103.

and M. J. Wingfield. 1991a. Ultrastructure of ascosporogenesis in Ophiostoma davidsonii. Mycol. Res. 95:725-730.

and — —. 1991b. Ascospore ultrastructure and development in Ophiostoma cucullatum. Mycologia 83: 698-707. CHAPTER 4

THE NONMONOPHYLY OF UNITUNICATE, PERITHECIAL

ASCOMYCETES

Ascomycetes are a large group of fungi that rival flowering plants in numbers, morphological diversity, ecological heterogeneity and species richness

(Hawksworth, 1991). The diversity and sheer numbers of organisms have left ascomycete systematists with a formidable task. The trend over the past decade has been the de-emphasis of supraordinal relationships (Hawksworth et al.,

1983). The lack of confidence in supraordinal classifications stems from a relatively few useful characters and an inability to polarize the ones that do exist.

Recently, relationships above the level of the order have received renewed attention (Barr, 1987; 1990; Berbee and Taylor, 1992a; 1992b). The result has been both the rejection (Barr, 1990) and failure to reject (Berbee and Taylor,

1992a) the monophyly of pyrenomycetous ascomycetes.

Pyrenomycetes are filamentous ascomycetes that produce unitunicate asci from a hymenium within a true perithecial ascoma (Nannfeldt, 1932; Luttrell,

1951; Wehmeyer, 1975; Barr, 1987; 1990). Unitunicate asci are characterized by a single functional wall layer; however the ascus wall comprises several cell

(hyphal) layers. A true ascoma is one that is derived from hyphae consigned to the sexual reproductive system. There exists another arrangement where bitunicate asci are produced in a false ascoma, or pseudothecium, which is composed of somatic hyphae (Luttrell, 1951). At first glance these pairs of

68 69

characters appear to be mutually exclusive. However, there exist groups of fungi

that appear intermediate. Examples are the Coryneliales and Coronophorales,

which have been described as producing unitunicate asci in a false ascoma

(Luttrell, 1951; Alexopoulos and Mims, 1979; Barr, 1990). Additionally,

Pyxidiophora and Subbaromyces are genera that have been described as

pyrenomycetes, but have been problematic in their ordinal classifications.

A perithecial ascoma is a flask-shaped structure with a rounded base from

which a neck or beak extends. The ascomata are either produced associated with

a stroma or initiated from a mycelium. Development of a perithecium may

commence via either an ascogonium, ascogonium and antheridium, ascogonium

and trichogyne or ascogonial coil (Jensen, 1984). The neck may vary in both

length and diameter, and is penetrated by a pore, or ostiole, through which the

ascospores are discharged. Ascospores may either be discharged forcibly or passively from a perithecium; typically, ascospores that are discharged passively have early evanescent asci.

Unitunicate asci range from broadly clavate to cylindrical and may be

produced from a hymenium at the base of the perithecium, or in a scattered manner from an ill-defined hymenium. Furthermore, the ascus wall may persist

or deliquesce prior to ascospore maturity. Typically, eight ascospores are produced per ascus. They may vary from dark to hyaline and may or may not be

septate. If present, the septation(s) are latitudinal and never muriform.

Ascospores range from globose to elliptical to filiform in shape, and vary in size by orders of magnitude. Ornamentation of ascospores result in some of the more

diagnostic traits for species and genera. They may germinate via germ pores, slits or undefined wall areas. 70

This study is directed at clarifying the question of whether unituncate, perithecial ascomycetes form a natural monophyletic group. The approach involves sampling a relatively large number of traditional pyrenomycetes including several problematic taxa. Cladistic analysis of nucleotide data from the nuclear-encoded small subunit ribosomal DNA (ssrDNA) provides an independent assessment of the homology of morphological characters and the phylogenetic hypotheses inferred from them. An independent data set from the

B subunit of vacuolar ATPase (B vATPase) also is presented.

Materials and Methods

The taxa sampled for this study were chosen based upon their position in past and present classifications and possession of particular morphological and ecological characters. The ingroup comprises taxa that have been traditionally treated as pyrenomycetes (Luttrell, 1951; Wehmeyer, 1975; Barr, 1983; 1990).

Additional non-pyrenomycetous taxa also were included. The primary rationale of including these nonpyrenomycetous species is to detect more accurately the homoplasy of the characters sampled and the boundaries of unitunicate, perithecial ascomycetes. The outgroup includes three basidiomycetes: two wood rotters and a basidiomycetous yeast. The taxa, source of isolate and regions of the gene sequenced are listed in Table 4.1.

The following procedures were performed as described previously in chapter one: culturing and harvesting of organisms; extraction of total nucleic acid and assay for presence of high molecular weight DNA; symmetric and asymmetric polymerase chain reactions (PCRs); single-stranded PCR (ssPCR) product purification; sequencing reactions, polyacrylamide gel electrophoresis 71

Table 4.1 List of taxa included in study, source of taxon and regions sequenced. Primers are those of Whiteet al. (1990).

Aciculosporium take Miyake ATCC 36483 NS2, NS4 Ambrosiozyma platypodis (Baker & Kreger ATCC 36174 NS, NS2, NS3, NS4 - van Rij) van der Walt Aphysiostroma stercorarium Barassa & Moreno A. Rossman NS2, NS4 Aspergillus fumigatus Fr. Genbank Asnda Athelia bombacina Pers. Illingsworthet al., 1991 Aureobasidium pullulans (de Bary) Amaud Illingsworthet al., 1991 Balansia clserotica (Pat.) Hohn. ATCC 16582 NS1.NS2, NS3.NS4 Candida albicans (Robin) Berkout Genbank Ysall6s Candida tropicalis (Castellani) Berkhout Genbank Ysarmab Cephaloascus fragrans Hanawa ATCC 36174 NS1.NS2, NS3, NS4 Ceratocystis fimbriata Ellis & Halst. Harrington C89NS1.NS2, NS3, NS4 Ceratocystis virescens Davids. Harrington C69 NS1.NS3 Cercophora septentrionalis Lundq. D. Malloch NS1.NS2, NS3 Chaetomium globosum Kunze ex Fr. ATCC 44699 NS1.NS2, NS3 Claviceps paspali Stevens & Hall ATCC 13892 NS2, NS4 Cordyceps capitata (Fr.) Link J. Spatafora NS1.NS2, NS3, NS4 Cryptococcus neoformans (Sanfelice) Vuillemin Genbank Cpcda Daldinia concentrica (Bulliardex Fries) ATCC 36659 NS1.NS3.NS4 Cesati et de Notaris Diaporthe phaseolarum (Cookeet Ellis) Sacc. FAU-458 NS2, NS4 Diatrype discoformis (Hoffm.:Fr.) Fr. CBS 197.49 NS2, NS4 Eleutherascus peruvianus Huang ATCC 28390 NS1.NS3 Epichloe typhina (Pers.) Tulasne ATCC 56429 NS2, NS4 Halosphaeriopsis mediosetigera (Kohl.) Kohl. ATCC 16934 NS1.NS3 Hirsutella thompsonii Fischer ATCC 24874 NS2, NS4 Hypocrea lutea (Tode:Fr.) Petch GJS 84-474 NS1.NS2, NS3 Hypocrea pallida Ellis & Everh. GJS 89-83 NS1.NS2, NS3.NS4 Hypocrea schweinitzii (Fr.)Sacc. CTR 79-225 NS1.NS2, NS3.NS4 Hypocrella sp. Sacc. GJS 89-104 NS1, NS2, NS3, NS4 Hypomyces polyporinus Peck ATCC 46844 NS2, NS4 Hypoxylon atroroseum J. Rogers NS1.NS3 Kathistes analemmoides Malloch & Blackwell M. Blacked NS1.NS2, NS3, NS4 Kathistes calyculata Malloch & Blackwell M. Blackwell NS1.NS2, NS3.NS4 Leucostoma persoonii (Nits.) Hoehnel Berbee & Taylor, 1992b Microascus trigonosporus Emmons & Dodge RSA 1942 NS1.NS2.NS3.NS4 Nectria cinnabarina (Tode ex Fr.) Fr. GJS 89-107 NS1.NS2, NS3.NS4 Nectria haematococca Berk.& Br. GJS 89-97 NS1.NS2, NS3.NS4 Nematospora sp. FLAS 296 NS2, NS4 Neocosmospora vasinfecta Smith RSA 1898 NS2, NS4 Neurospora crassa Shear & Dodge Genbank Neurmas Ophiostoma piliferum H. & P. Syd. Harrington C300 NS2, NS4 Ophiostoma ulmi (Buisman) Nannf. Berbee & Taylor, 1992b Petriella setifera (Schmidt) Curzi ATCC 26490 NS1.NS2, NS3 Pyxidiophora spl A M. Blackwell NS1, NS2„ NS3.NS4 Pyxidiophora SPOl M. Blackwell NS2, NS4 72

Table 4.1 (cont.)

Saccharomyces cerevisiae Hansen Genbank Yscrgea Sphaeronaemella fimicola Marchal M. Blackwell NS1.NS3.NS4 Sphaerostilbella aureonitens (Tul.) Seifert et al. GJS NS1, NS2, NS3.NS4 Sphaerostilbella NZ GJS 82-40 NS2, NS4 Spongipellis unicolor (Schw.) Murr. Bowmanet al., 1992 Stictis radiata Per. ex S. F. Gray D. Malloch NS2, NS4 Subbaromyces splendens Hesseltine NRRL 2340 NS1.NS2, NS3.NS4 Taphrina deformans (Fuckel) Tul. ATCC 11124 NS1.NS3 Xylaria curta Fr. J. Rogers NS2, NS3, NS4 Xylaria hypoxylon (L.:Fr.) Grev ATCC 42768 NS2, NS3, NS4 73

and autoradiography; sequence alignments and data analysis; calculation of decay

indices and bootstrap confidence levels.

The rDNA data set constructed in this study consists of 54 taxa and

approximately 900 bp/taxon, 450 bp of sequence from each of the two regions

designated by the primer pairs NS1/NS2 and NS3/NS4 (White et al, 1990). The

region defined by NS1/NS4 is approximately 1200 bp in length and corresponds

to the 5' half of the ssrDNA. It was chosen based on a preliminary screening that

demonstrated the 3' half of the molecule was considerably more conserved and provided far fewer synapomorphies within pyrenomycetes. This pattern of nucleotide variability is in concordance with variation in secondary structure of the ssrRNA (Gutell, 1985). Furthermore, the 3' region possessed intervening sequences (IVS) in several taxa, which complicated the acquisition of a complete data set for all taxa in question. A subsample of 37 taxa was chosen for bootstrap and decay index analyses.

The B subunit of vacuolar ATPase was amplified using primers 1 BvATP and BvATPS. These are degenerate, inosine-containing primers that correspond to regions of conserved amino acid sequence, which are characterized by low levels of nucleotide degeneracy. Nucleotide positions that were 2- or 3-fold degenerate were coded as such in the primer design (sequence); nucleotide positions that displayed 4-fold degeneracy were coded as an inosine. Primer

BvATP7 was used for sequencing. Primer sequences, complementary amino acid sequences and degenerate nucleotide sequences are listed in Table 4.2.

The B vATPase data set comprises 14 taxa and 356 bp/taxon. The B vATPase data were analyzed in two different manners. The first was an analysis of the first and second codon positions, and the second was an unweighted analysis of all three codon positions. 74

Table 4.2 Degenerate, inosine-containing primers for amplification and sequencing of the B subunit of vacuolar ATPase. Primers lBvATP and BvATPS amplify a 750 bp fragment that corresponds to positions 1803 - 2552 in Neurospora crassa (Bowman et al., 1988). Primer BvATP7 anneals to the noncoding strand at positions 2436 - 2455. Degenerate nucleotide codes: N = ACGT, H = ACT, D = AGT, R = AG, Y = CT. 1. amino acid sequence. 2. degenerate nucleotide sequence. 3. primer sequence.

1. PHN E I A A QIC 2. CCN CAY AAY GAR ATH GCN GCN CAR ATH TGY lBvATP 3. 5’ CCI CAY AAT GAA ATH GCI GCI CAG ATH TG 3'

1. A I G E G M T 2. GCN ATH GGN GAR GGN ATG ACN BvATP7 3. 3' CGI TAD CCI CTY CCI TAC TG 5’

1. MKAVVGE EA 2. ATG AAR GCN GTN GTN GGN GAR GAR GCN BvATP8 3. 3’ TAC TTY CGI CAI CAI CCI CTY CTY CG 5' 75

Results

The ssrDNA data produced 42 equally most parsimonious trees of 887 steps (Fig. 4.1) with consistency (Cl), retention (RI) and rescaled consistency

(RC) indices of .432, .708 and .306, respectively. A strict consensus tree (Fig. 4.1) is 898 steps in length with a Cl of .427, an RI of .702 and an RC of .299. Of the

900 bp of raw data 259 were phylogenetically informative of which five were gaps.

Of the 898 inferred steps of the strict consensus 518 are transitions (Tn) and 375 are transversions (Tv) resulting in a Tn:Tv of 1.37 for the entire tree and 1.38 for the clade of derived pyrenomycetes. Cladistic analysis of the 37 taxon statement produced a single most parsimonious tree of 689 steps (Cl = .489, RI = .696 and

RC = .340) (Fig. 4.2). Decay indices were estimated from 6400 trees not greater than 10 steps longer than the set of most parsimonious trees.

The pyrenomycetes represented in the ssrDNA data set did not form a monophyletic group (Fig. 4.1). There exists a large clade of pyrenomycetes comprising the orders Clavicipitales, Diaporthales, Halosphaeriales,

Hypocreales, Microascales, Ophiostomatales, Sordariales and Xylariales. The taxa sampled from these orders form a monophyletic clade that is supported by bootstrap confidence levels of 100% and a decay index of > 10 steps (Fig. 4.2).

The remaining pyrenomycetes are species from the generaKathistes Malloch &

Blackwell, Pyxidiophora Bref. & Tavel, and Subbaromyces Hesselt., and are placed outside of the clade of derived pyrenomycetes. The most parsimonious placement of these relatively more primitive genera with their putative sister taxa is not well supported by bootstrap confidence levels.

The B vATPase data set comprises 14 taxa and 356 bp/taxon. Non­ overlapping ends were discarded decreasing the number to 319 bp/taxon. Of 76

Aciculosporium take - Balansia sclerotica Claviceps paspali ■ Cordyceps capitata CLAVICIPITALES Epichloe typhina - Hirsutella thompsonii - Hypocrella sp. Aphysiostroma stercorarium Hypocrea schweinitzii Hypocrea lutea Sphaerostilbella NZ Sphaerostilbella atireonitens Hypocrea pallida HYPOCREALES Hypomyces polyporinus t Nectria haematococca | Neocosmospora vasinfecta 1 - Nectria cinnabarina r Ceratocystis fimbriata _ _ j * - Ceratocystis virescens OPHIOSTOMATALES I — Sphaeronaemella fimicola i— Halosphaeriopsis mediosetigera _ P — Pelriella setifera MICROASCALES L— Microascus trigonosporus i Diaporthe phaseolarum DIAPORTHALES n - Leucostoma persoonii ~ | r Ophiostoma piliferum OPHIOSTOMATALES *■ Ophiostoma ulmi r Cercophora seplenirionalis _ J " L Chaelomium globosum SORDARIALES L — Neurospora crassa r Daldinia concentrica J t — Diatrype discoformis [ L - Hypoxylonatroroseum XYLARIALES j i - Xylaria curta Xylaria hypoxylon Aspergillus fumigatus Nematospora sp. > Kathistes analemmoides Kathistes calyculata r Pyxidiophora la C /Pyxidiophora spol Subbaromyces splendens ■ Aureobasidium pullulans — Stictis radiata ■ Eleutherascus peruvianus — Ambrosioryma platypodis Saccharomyces cerevisiae Candida albicans ENDOMYCETALES Candida tropicalis Cephaloascus fragrans — — Taphrina deformans TAPHRINALES ■ Alhelia bombacina • Spongipellis unicolor BASIDOMYCETES • Cryptococcus neoformans

Figure 4.1 Strict consensus of 42 equally most parsimonious cladograms containing 54 taxa (898 steps, Cl = .427, RI = .702 and RC = .299). The designated outgroup is the Basidiomycetes. Specified ordinal rankings follow Hawksworth et al. (1983). 77

98 | LDiaporthe phaseolarum _P"9 I— Leucostoma persoonii I 100 [■ Ophiostoma piliferum >10 ^ Ophiostoma ulmi 9 5 |- Cercophora septentrionalis 99J7LChaetomium globosum > 101 Neurospora crassa j - Dald.in.ia concentrica FI—— Diatrype discoformis I— Hypoxylon alroroseum Xylaria curta 100 41 * Xylaria hypoxylon >10 96| Balansia sclerotica [6*— Claviceps paspali I 98 rHypomyces polyporinus 9 I Hypocrea schweinitzii Nectria cinnabarina 100 | Ceratocystis fimbriata 98 >10 •— Ceratocystis virescens Microascus trigonosporus Aspergillus fumigatus — Nematospora sp. 100 —■ Kathistes analemmoides >10 - Kathistes calyculata 100 Pyxidiophora la >10 *— Pyxidiophora spol Subbaromyces splertdens Aureobasidium□ pullulans — Stictis radiata ■ Eleutherascus peruvianus 100 - Ambrosiozyma platypodis >10 Saccharomyces cerevisiae 100 _P------CepCephaloascus fragrans >10 I CandidaCand albicans Taphrina deformans Athelia bombacina ■ Spongipellis unicolor

Figure 4.2 Bootstrap confidence levels are mapped on the single most parsimonious cladogram of 37 taxa (689 steps, Cl = .489, RI = .696 and RC = .340). Bootstrap confidence levels are given for the nodes supported in 90% or more of the trees from 100 bootstrap replications. Corresponding decay indices are given below their respective nodes. 78

these 319 positions 173 are variable and 131 are phylogenetically informative.

96.22 % of the third, 39.25% of the first and 23.58% of the second codon positions were variable. Of the 131 informative sites 76.22% (100) were at the third codon position, 18.32% (24) at the first and 5.34% (7) at the second.

Only 8 of the 106 inferred amino acid positions were informative.

Furthermore, the sequences ofCercophora septentrionalis and Hypocrea schweinitzii are shorter than the those for the other taxa due to unalignable regions of their nucleotide sequences. These regions may represent variation in the number and positions of exons. Such variation exists betweenNeurospora crassa (Genbank Neuvma2a) and Canidida tropicalis (Genbank Ysaatpb) (data not shown). Therefore, the preliminary inferred amino acid sequence was not analyzed because of too few characters. Analysis of the first and second codon positions resulted in 153 equally most parsimonious trees of 59 steps with a Cl of

.610, an RI of .597 and an RC of .413. A strict consensus of these is 63 steps with a Cl of .571, an RI of .676 and an RC of .341 (Fig. 4.4). The separation of the pyrenomycetes is supported. Analysis of the nucleotide data set containing all three codon positions produced two equally most parsimonious trees. A strict consensus is 487 steps with a Cl of .478, an RI of .370 and an RC of .177 (Fig.

4.5). The separation of the pyrenomycetes is not maintained in the analysis containing all three codon positions.

Discussion

Monophyly of unitunicate, perithecial ascomycetes (pyrenomycetes) is rejected by the ssrDNA data. There exists a more derived clade representing a broad range of fungi that have been treated as pyrenomycetes, and some that 79 have been suggested to have an affinity with plectomycetes. Taxa sampled for the Clavicipitales, Hypocreales, Diaporthales, Sordariales and Xylariales have been treated as pyrenomycetes throughout modern mycology (Gaumann and

Dodge, 1928; Dodge, 1952; Nannfeldt, 1932; Miller, 1949; Luttrell, 1951;

Rogerson, 1970; Wehmeyer, 1975; Barr, 1987; 1990). The taxa representing the orders Microascales and Ophiostomatales have been treated as plectomycetes

(Luttrell, 1951; Benny and Kimbrough, 1980; Upadhyay, 1981) and as pyrenomycetes (Barr, 1990; Berbee and Taylor, 1992a). The clade containing these taxa is supported by a bootstrap confidence level of 100% and a decay index of > 10 steps. These analyses are interpreted as putative measures of support for this specific data set, i.e. the nucleotide characters sequenced for the taxa sampled.

The more primitive genera, Kathistes, Pyxidiophora and Subbaromyces, are excluded from the derived clade and placed among a heterogeneous group of taxa, the functional outgroup (Watrous and Wheeler, 1981). These three genera have been placed in several orders represented in the clade of derived pyrenomycetes (Barr, 1990; Blackwell and Malloch, 1988); however, they are excluded confidently from all those sampled in this study. These data suggest that the production of unitunicate asci within a true perithecium has evolved at least twice, and that the combination of these two characters do not circumscribe a natural, monophyletic group (Fig. 4.3).

The perithecia of the derived pyrenomycetes are characterized as being true ascoma, i. e. derived from hyphae designated to sexual reproductive system.

The asci may be produced from a hymenium that lines the base and/or walls of the inner perithecium, or in a scattered manner from an ill-defined hymenium

(Luttrell, 1951; Jensen, 1984; Barr, 1987; 1990). Also, the asci maybe persistent 80 or evanescent prior to ascospore maturity. The asci of the primitive pyrenomycetous genera are produced from the base of a true perithecium and are evanescent. This same combination of characters is seen in the genera

Chaetomium, Ophiostoma and Aphysiostroma. Also, several of the derived pyrenomycetes sampled produce evanescent asci in a scattered manner (Fig. 4.3; Berbee and Taylor, 1992b). Current perceptions of production and arrangement of asci do not assist in recognizing the division of these two groups.

The perithecia of the derived pyrenomycetes possess an outer wall (the peridium) and several inner wall layers. The peridium may vary in color, texture, and cell (hyphal) morphology (Jensen, 1985). Likewise, the inner walls, which are considered to be part of the centrum, may vary both quantitatively and qualitatively (Luttrell, 1951; Jensen, 1983; 1984; Barr, 1990). In general, the mature perithecial wall in the derived pyrenomycetes comprises several wall layers and appears to be thicker than that found in the primitive pyrenomycetous genera. The perithecia ofPyxidiophora lundqvistii (Blackwell and Malloch,

1990), Kathistes calyculata and K. analemmoides (Malloch and Blackwell, 1990) possess a single wall comprising a single layer of cells. However, not all of the species placed in these primitive genera have perithecia with a single wall layer

(Blackwell and Malloch, pers. comm.), and little is known of their development.

The lack of information underscores the need for continued ontogenetic studies to understand better the level of variation that exists in perithecium development among these taxa.

The three primitive genera have been classified in several orders in the clade of derived pyrenomycetes, which is reviewed in Blackwell and Malloch

(1988). Most notable are proposed affinities with the Ophiostomatales and

Hypocreales. These primitive genera have been considered ophiostomatalean 81

— Aciculosporium take — Balansia sclerotica Claviceps paspali Perithecial ■ Cordyceps capitata Epichloe typluna Ascomata — Hirsutella thompsonii — Hypocrella sp. Aphysiostroma stercorarium * B - Hypocrea schweinitzii ■o Hypocrea lutea Sphaerostilbella N Z Sphaerostilbella aureonitens Hypocrea pallida Hypomyces polyporinus r Nectria haematococca I Neocosmospora vasinfecta Nectria cinnabarina r Ceratocystis fimbriata _ T ~ L Ceratocystis virescens Sphaeronaemella fimicola _ j — Halosphaeriopsis medioseligera Perithecial J * — Petriella setifera *— Microascus trigonosporus Ascomata ■ Diaporthe phaseolarwn I L Leucostoma persoonii P r Ophiostoma piliferum BS *■ Ophiostoma ulmi BS r Cercophora septentrionalis _ J T - Chaetomium globosum B I— Neurospora crassa r Daldinia concentrica _ J L — Diatrype discoformis | L _ Hypoxylon atroroseum L p Xylaria curta T — Xylaria hypoxylon Aspergillus fumigatus * S' ■ O Nematospora sp. ' Kathistes analemmoides Kathistes calyculata Perithecial r Pyxidiophora la CPyxidiophora spol Ascomata Subbaromyces splendens • Aureobasidium pullulans — Stictis radiata Eleutherascus peruvianus Ambrosiozyma platypodis Saccharomyces cerevisiae Candida albicans • Candida tropicalis • Cephaloascus fragrans Taphrina deformans • Athelia bombacina • Spongipellis tmicolor ■ Cryptococcus rteoformans

Figure 4.3 Characters of ascomata, evanescent unitunicate asci, and location of asci in the ascomata are mapped on the strict consensus tree. Unitunicate, perithecial ascomycetes are nonmonophyletic. (* = evanescent asci; B = basal asci S = scattered asci, and BS = asci basal becoming scattered; circle denotes cleistothecial ascomata.) 82

based upon the morphologies of the perithecia and anamorphs and spore

dispersal. The primitive pyrenomycetes and species of Ceratocystis and Ophiostoma possess perithecia characterized by a globose base and a long neck,

or beak, from which specialized hyphae (setae) extend. Both groups produce

their ascospores from evanescent asci and exude them in a droplet at the tip of

the perithecial neck in the setae. Species of Pyxidiophora produce sticky, mucous

conidia from phialidic or holoblastic anamorphs (Blackwell and Malloch, 1988).

Phialidic conidiogenous cells are also represented in the Hypocreales (Samson,

1974) and Ceratocystis (de Hoog and Scheffer, 1984) and holoblastic ones are

reported for Ophiostoma (de Hoog, 1974). Subbaromyces has been reported as possessing Scopulariopsis anamorphs similar to those found in the Microascales

(Cole, 1974). Moreover, the overwhelming majority of species in the genera

Pyxidiophora, Kathistes, Subbaromyces, Ceratocystis and Ophiostoma and the

Microascales are intimately associated with insects. Malloch (1979) proposed that dispersal selection pressures may result in the parallel evolution of both the anamorph and teleomorph,i.e. convergent or parallel evolution among the life histories of disparate taxa. These data support the convergent evolution of this suite of characters, and is proposed to be directed by the selection pressure of insect dispersal (Cain and Weresub, 1957; Minter, 1983; Blackwell and Malloch,

1990; Malloch and Blackwell, 1992a; 1992b; Berbee and Taylor, 1992b).

If the characters mentioned above are non-homologous and represent convergent or parallel evolution among these fungi, is the same true for the production of unitunicate asci within a perithecial ascoma? The ssrDNA data reject the monophyly of taxa with these features and suggest that convergent or parallel evolution may exist here, also. To test this hypothesis a preliminary independent data set from the B subunit of vacuolar ATPase (B vATPase) has 83

been constructed. The data set consists of both derived and primitive

pyrenomycetes as inferred from the ssrDNA, two species of ascomycetous yeast

and Taphrina deformans.

A strict consensus of the 153 trees produced by the first and second

positions of the codon supports the separation Pyxidiophoraof and Subbaromyces

from the derived pyrenomycetes (Fig. 4.4). The derived pyrenomycetes sampled

form a monophyletic, although unresolved, clade. The two yeasts sampled,

Candida tropicalis and Cephaloascus fragrans, are the most basal members of the ingroup. The topology of this tree is identical to the ssrDNA phylogeny for these taxa with respect to the three major clades; however, the ssrDNA provides resolution within the derived pyrenomycetes (Fig. 4.5).

The phylogeny inferred from all three codon positions (Fig. 4.6) does not agree with the ssrDNA phylogeny, or the one inferred from the first two codon positions. Pyxidiophora and Subbaromyces are no longer sister taxa, and they are not separated from the derived pyrenomycetes. Furthermore, the analysis does not support relationships within the pyrenomycetes, which are mutually supported by ssrDNA and morphology. It is proposed that the relationships inferred in this analysis are based upon high levels of homoplasy associated with the third position of the codon. Of the 319 bp of data, 96.22 % of the third base positions are variable, which translates to 76.34 % of the synapomorphies. This compares to 39.25 % of the first and 23.58% of the second codon positions being variable (18.32% and 5.34% of the synapomorphies, respectively). Also, of the

100 informative third base positions, 63% are characterized by multiple hits.

Consistency indices (CIs), retention indices (RIs) and rescaled CIs (RCs) indicate that a substantial amount of homoplasy is associated with the third position. CIs are dependent on both the number of taxa and characters; as the 84

I Hypocrea schweinitzii

Hypocrea pallida

— Nectria haematococca

Balansia sclerotica

...... Ceratocystis fimbriata

— — Xylaria hypoxylon

— Daldinia concentrica

Cercophora septentrionalis

Neurospora crassa

— Pyxidiophora spol

Subbaromyces splendens

Candida tropicalis

Cephaloascus fragrans

Taphrina deformans

Figure 4.4 Strict consensus of 153 equally most parsimonious BvATPase trees inferred from first and second nucleotide position of the codon (63 steps, Cl =.571, RI = .597 and RC = .341). The topolgy of the resolved nodes is in agreement with the ssrDNA gene trees (Fig. 4.1; Fig. 4.5). 85

— Hypocrea pallida

Hypocrea schweinitzii

• Balansia sclerotica

— Nectria haematococca

— - Ceratocystis jimbriata

— Cercophora seplentrionalis

Neurospora crassa

Daldinia concentrica

—— Xylaria hypoxylon

Pyxidiophora spol

— Subbaromyces splendens

Candida tropicalis

»• Cephaloascus fragrans

——— — — Taphrina deformans

Figure 4.5 Strict consensus tree of four equally most parsimonious ssrDNA gene trees containing 14 taxa (239 steps, Cl = .649, RI = .690 and RC = .448). The topology agrees with 54 taxa ssrDNA tree (Fig. 4.1) except that the Xyiariales and Sordariales, and the Hypocreales and Clavicipitales clades are unresolved between one step grades or clades (data not shown). 86

■ Balansia sclerotica

Cercophora septentrionalis

Neurospora crassa

— I-—-.- Subbaromyces splendens

“ —i— Nectria haematococca

“ L— — Xylaria hypoxylon

—— I— Daldinia concentrica

Hypocrea schweinitzii

— Ceratocystis fimbriata

— Hypocrea pallida

— Pyxidiophora spol

Candida tropicalis

I- Cephaloascus fragrans

Taphrina deformans

Figure 4.6 Strict consensus of two equally most parsimonious BvATPase trees inferred from an unweighted analysis of all three nucleotide positions of the codon (487 steps, Cl = .478, RI = .371 and RC = .177). The topology does not agree with the ssrDNA gene tree (Fig. 4.1; Fig. 4.5), or the topology of the BvATPase tree inferred from only the first and second positions of the codon (Fig. 4.4). 87

number of taxa and/or characters are increased CIs decrease (Archie, 1989a).

Archie (1989b) showed that CIs may be artificially inflated and proposed the

homoplasy excess ratio (HER) as a better measure of homoplasy. Farris (1989)

argued that the HER is equivalent to the ensemble retention index (RI). He proposed that the RC is a more useful measure in comparisons of the fit of

characters with different levels of variations, /.e. different character state distributions and frequencies of variation. The RC is employed here as a measure of homoplasy, although, some investigators may disagree. The corresponding CIs and RIs are also reported.

The phenomenon of decreasing the Cl and RC by increasing the number of taxa and characters is displayed by the ssrDNA and BvATPase data sets, respectively. The strict consensus inferred from ssrDNA data set of 54 taxa and

259 characters (Fig. 4.1) has a Cl of .427, and RI of .702 and a RC of .298; the strict consensus inferred from the ssrDNA data set of 14 taxa and 121 characters

(Fig. 4.5) has a Cl of .649, and RI of .691 and an RC of .448. The Cl and RC behaved as expected. The strict consensus inferred from the BvATPase data set comprising the first and second codon positions (Fig. 4.4) has a Cl of .571, an RI of .597 and a RC of .341. The strict consensus inferred from the BvATPase data set containing all three positions of the codon (Fig. 4.6) has a Cl of .478, an RI of

.370 and a RC of .177. The Cl and RC decreased when the number of characters was increased, however, the number of taxa was held constant at 14. Therefore, the increase in homoplasy in the BvATPase data sets is attributed entirely to the addition of characters, i.e. the third position of the codon, and not the addition of taxa.

Of particular importance in the examples of the ssrDNA and the

BvATPase data sets is the behavior of the RIs relative to the RCs. The decrease Table 4.3 Summary of Cl, RI and RC for the two ssrDNA and the two BvATPase data sets. N = number of characters. RCs decrease with the addition of new taxa (ssrDNA) and new characters (BvATPase). The RC decreased most for the BvATPase data. Note that the RI did not decrease in the ssrDNA data but did in the BvATPase. The results suggest that third position of the codon is characterized by an inordinate amount of variation with respect to character states and character state distribution.

DATA SET N Cl RI RC ssrDNA 14 taxa 121 .649 .691 .448 ssrDNA 54 taxa 259 .427 .702 .299

BvATPase 14 taxa 31 .571 .597 .341 1st and 2nd codon positions BvATPase 14 taxa 131 .478 .370 .177 all 3 codon positions 89

in the RC observed for the ssrDNA data sets was not coupled to a corresponding

decrease in RI, whereas in the BvATPase data sets the decrease in RC was

coupled with a decrease in RI. The decrease in RI reflects the increased variation in distribution of characters state changes of the third codon position relative to the first and second codon positions. The 54 taxa ssrDNA data set did not result in a marked divergence from the distribution of character states relative to the 14 taxa statement.

The decrease in an RC that results from the addition of taxa potentially reflects a better understanding of the amount of variation in, and distribution of, character states for that group of taxa. A decrease in an RC that represents introduction of homoplasy from the inclusion of new characters increases the number of potential characters, but does not represent knowledge gained concerning the distribution and amount of variation associated with the initial set of characters. This is not to say that the addition of more sequence, or characters, will not be beneficial in some, if not many, circumstances. However, an increase in character sampling is most informative in the context of a concurrent increase in taxon sampling.

Curiously, a similar phenomenon is displayed by the transition- transversion ratios (Tn:Tv). The Tn:Tv for the 54 and 14 taxa ssrDNA data sets are 1.37 and 1.67, respectively. These values are lower than reported previously for the entire ssrDNA gene for a complementary group of fungi (Berbee and

Taylor, 1992b). Since, this study focused on the more variable 5' half, the exclusion of the 3' half of the gene may contribute to the discrepancy in Tn:Tv.

However, the number of taxa and the particular taxa sampled also appear to contribute to the different values. Future studies in Tn:Tv ratios relative to the number and relatedness of taxa are needed. 90

Analysis of the data derived from the ssrDNA rejects the monophyly of pyrenomycetous ascomycetes. Also, the first and second positions of the codon of

BvATPase supports their separation. Maximum parsimony places the species sampled from the genera Kathistes Malloch & Blackwell, Pyxidiophora Bref. &

Tavel, and Subbaromyces Hesselt. outside of a large clade representing several pyrenomycetous orders. These data support previous hypotheses concerning the convergent nature of perithecia, asci and ascospores. Specifically, the homologies of unitunicate, evanescent asci and true perithecial ascomata between these two groups are questioned. Further ontogenetic studies are warranted to understand better the developmental variations that may provide insight concerning the evolution of the primitive perithecial taxa.

The data set comprising all three positions of the codon of BvATPase did not reject the monophyly of pyrenomycetes. The amount of homoplasy associated with the informative third positions is too great relative to the number of characters from the first and second positions,i.e. the phylogenetic signal is being overwhelmed by noise. In general, the nucleotide data from BvATPase did not prove especially useful relative to this taxonomic statement. The amino acid sequence was far too conserved; the first and second positions of the codon were unable to resolve relationships within the more derived clade of pyrenomycetes; and the third position of the codon contained an inordinate amount of homoplasy.

The term nonmonophyletic was employed to describe the relationship of the relatively derived and primitive unitunicate, perithecial ascomycetes. A grouping may be nonmonophyletic either as polyphyletic or paraphyletic assemblage (Wiley, 1981). The use of nonmonophyly in this study is intentionally ambiguous with respect to polyphyly and paraphyly. The rationale for the 91 ambiguity stems from the taxonomic statement.Pyxidiophora, Kathistes and

Subbaromyces are placed in the functional outgroup (Watrous and Wheeler,

1981); however support for the suprageneric relationships of these three genera is questionable. The functional outgroup presented here is an heterogeneous group of fungi that spans a diversity of morphology and life histories, and taxon sampling for them is inadequate. Therefore, future research concerning the higher taxonomic relationships of these three primitive genera should involve a more intensive and expansive taxon sampling among this diverse group of fungi. 92

Literature Cited

Alexopoulos, C. J. and C. W. Mims. 1979.Introductory Mycology, 3rd ed. John Wiley & Sons, N. Y. 632 p.

Archie, J. W. 1989a. A randomization test for phylogenetic information in systematic data. Syst. Zool. 38: 239-252.

1989b. Homoplasy excess ratios: a new indices for measuring levels of homoplasy in phylogenetic systematics and a critique of the consistency index. Syst. Zool. 38: 253-269.

Barr, M. E. 1990. Prodromus to nonlichenized, pyrenomycetous members of c lass Hymenoascomycetes. Mycotaxon 39: 43-184. . 1983. The ascomycete connection.Mycologia 75: 1-13.

Benny, G. L. and J. W. Kimbrough. 1980. A synopsis of the orders and families of plectomycetes with keys to genera. Mycotaxon 12: 1-91.

Berbee, M. L. and J. W. Taylor. 1992a. Two ascomycete classes based upon fruiting-body characters and ribosomal DNA sequence. Molec. Biol. Evol. 9: 278-284.

and . 1992b. Convergence in ascospore discharge mechanism among pyrenomycete fungi based on 18S ribosomal RNA gene sequence. Molec. Phylog. Evol. 1: 59-71.

Blackwell, M. and D. Malloch. 1988. Pyxidiophora (Pyxidiophoraceae): a link between the Laboulbeniales and hyphal ascomycetes. Mem. N. Y. Bot. Gard. 49: 23-32.

a n d ----- —. 1990. Discovery of aPyxidiophora with Acariniola-type ascospores. Mycol. Res. 94: 415-417.

y T. J. Perry, J. R. Bridges and J. C. Moser. 1986. A new species of Pyxidiophora and its Thaxteriola anamorph. Mycologia 78: 605-612.

Bowman, B. J., R. Allen, M. A. Wechser, and E. J. Bowman. 1988. Isolation of genes encoding the Neurospora vacuolar ATPase: analysis of vma-2 encoding the 57-kDa polypeptide amd comparison to vma-1. J. Biol. Chem. 263: 14002-14007.

Bruns, T. D., T. J. White, and J. W. Taylor. 1991. Fungal molecular systematics. Ann. Rev. Ecol. Syst. 22: 525-564.

Cains, R. F. and L. K. Weresub. 1957. Studies of coprophilous ascomycetes. V. Sphaeronaemella fimicola. Can. J. Bot. 35: 119-131. 93

Cole, G. T., R. V. Hardcastle and P. J. Szaniszlo. 1974. Subbaromyces splendens: development and ultrastructure. Can. J. Bot. 52: 2453-2457.

Farris, J. S. 1989. The retention index and the rescaled consistency index. Cladistics 5: 417-419.

Gaumann, E. A. and C. W. Dodge. 1928. Comparative Morphology o f Fungi. New York:McGraw-Hill Book Co., Inc. p.701.

Gaumann, E. A. 1952. The Fungi: A Description o f Their Morphological Features and Evolutionary Development. New York: Hafner publishing Co. p.420.

Gutell, R., B. Weiser, C. Woese and H. Noller. 1985. Comparative anatomy of 16S-like ribosomal DNA. Prog. Nucl. Acid Res. Mol. Biol. 32: 155-216.

Hawksworth, D. L., B. C. Sutton, and G. C. Ainsworth. 1983. Ainsworth and Bisby's Dictionary o f the Fungi, 7th edition. CMI, Kew. p445.

Hawksworth, D. L. 1991. The fungal dimension of biodiversity: magnitude, significance, and conservation. Mycol. Res. 95: 641-655.

Hoog, G. S. de. 1974. The generaBlastobotrys, Sporothrix, Calcarisporium and Calcarisporiella gen. nov. Stud. Mycol. 7: 1-84.

and R. J. Scheffer. 1984. Ceratocystis versus Ophiostoma: a reappraisal. Mycologia 76: 292-299.

Jensen, J. D. 1984. Morphologic studies in the pyrenomycetes. Ph.D. dissertation. Washington State University, Pullman. 121 p. . 1985. Peridial anatomy and pyrenomycete taxonomy.Mycologia 77: 688-701. llingsworth, C. A., J. H. Andrews, C. Bibeau and M. L. Sogin. 1991. Phylogenetic placement ofAthelia bombacina, Aureobasidium pullalans and Colletotrichum gloeosporioides inferred from sequence comparisons of small-subunit ribosomals RNAs. Exp. Mycol. 15: 65-75.

Luttrell, E. S. 1951. Taxonomy of the pyrenomycetes.University o f Missouri Studies 3: 1-120.

Malloch, D. 1979. Plectomycetes and their anamorphs. pp.153-166. In Kendrick, B. The Whole Fungus, Vol 1. National Museum of Natural Sciences, National museum of Canada. Ottawa.

Malloch, D. and M. Blackwell. 1990. Kathistes, a new genus of pleomorphic ascomycetes. Can. J. Bot. 68: 1712-1721.

and — ----- . 1992a. The adaptive nature of fungal traits facilitating fungal spore dispersal, pp. 147-172. In Carroll, G. C., and D. T. Wicklow (eds.) The Fungal Community: Its Organization and Role in the Ecosystem. 2nd ed. Dekker, N. Y. 94

a n d . 1992b. Dispersal biology of ophiostomatoid fungi.In Wingfield, M. J., K. A. Seifert and J. F. Wbber (eds.) Ceratocystis and Ophiostoma:Taxonomy, Ecology and Pathology. APS, St. Paul. In press. Miller, J. H. 1949. A revision of the classification of the ascomycetes with special emphasis on the pyrenomycetes. Mycologia 33: 74-81. Minter, D. W., B. C. Sutton, and B. L. Brady. 1983. What are phialides anyway? Trans. Br. Mycol Soc. 81: 109-120.

Nannfeldt, J. A. 1932. Studien iiber die morphologie und systematik der nichlichenisieten inoperculaten Discomyceten. Nova Acta Regiae Sco. Scien. Upsal, Ser. IV. 8: 1-368.

Rogerson, C. T. 1970. The hypocrealean fungi (Ascomycetes, Hypocreales). Mycologia 62: 865-910. Samson, R. A. 1974. Paecilomyces and some allied Hyphomycetes. Stud. Mycol. 6: 1-119. Upadhyay, H. P. 1981. A Monograph o f Ceratocystis and Ceratocystiopsis. Univ. Georgia Press, Athens, p. 176.

Watrous, L. E. and Q. Wheeler. 1981. The outgroup comparison method of character analysis. Syst. Zool. 30: 1-11.

Wehmeyer, L E. 1975. The pyrenomycetous fungi. Mycologia Memoir no. 6 p.250. White, T. J., T. D. Bruns, S. B. Lee and J. W. Taylor. 1990. Analysis of phylogenetic relationships by amplification and direct sequencing of ribosomal DNA genes, pp.315-322. In Innis, M. A., D. H. Gelfand, J. J. Sninsky, and T. J. White (eds.). PCR Protocols: A Guide to Methods and Applications. Academic Press, N. Y.

Wiley, E. O. 1981. Phylogenetics: the Theory and Practice o f Phylogenetics. Systematics. John Wiley & Sons, N. Y. p.439. CHAPTER 5

THE EVOLUTION OF CENTRUM DEVELOPMENT IN DERIVED

UNITUNICATE, PERITHECIAL ASCOMYCETES AS INFERRED BY

CLADISTIC ANALYSIS OF THE NUCLEAR-ENCODED SMALL SUBUNIT RIBOSOMAL DNA

Centrum is a collective term that refers to the internal components of an ascoma or locule (asci, ascogenous hyphae and sterile cells) and their relationships to one another (Luttrell, 1951). The term was coined by Wehmeyer

(1926) to replace the term nucleus (Fries, 1823), which had become confused with the nucleus of the eukaryotic cell. Luttrell's revision of pyrenomycetous ascomycetes, Taxonomy of the Pyrenomycetes (1951), had a profound impact upon ascomycete systematics, and laid the ground work for modern classifications of the group. Reynolds (1981) introduced the term "Luttrellian concept" and defined it as "an emphasis on the relationship between the ascus and the totality of ascocarp components."

Luttrell (1951) divided Euascomycetes into two major groups. The

Unitunicatae consisted of taxa that produced unitunicate asci in a true ascoma, and the Bitunicatae consisted those taxa that possessed bitunicate asci within a locule or ascostroma. Within the Unitunicatae he constructed classes based upon morphology and development of the ascoma and asci. Those taxa that produced unitunicate asci from a hymenium within an ostiolate or nonostiolate ascoma were placed in the Pyrenomycetes. Within the Pyrenomycetes he established six

95 96 orders. The Xylariales, Hypocreales, Diaporthales and Erysiphales were characterized by Xylaria, Nectria, Diaporthe and Phyllactinia centrum types , respectively. The two additional orders, the Coronophorales and Coryneliales, were said to produce unitunicate asci in a putative ascostroma and were maintained for convenience. Most Pyrenomycetes, excluding the Erysiphales, were taxa that possess ostiolate ascomata (perithecia). The Ophiostomataceae

(Microascales) was described as having an Ophiotoma type centrum, and was placed in the Plectomycetes on the basis of scattered asci. Although the majority of microascalean species possess true peritheca, scattered asci were given priority and all taxa that possessed them were treated as Plectomycetes. These centrum types are described briefly as defined by Luttrell (1951).

The hallmark of the Xylaria type centrum is the presence of true paraphyses that originate from the inner walls and base of the perithecium; they have free apical ends. Paraphyses are sterile hyphae that are proposed to function in the role of creating the central cavity. An ascogonial coil gives rise to a system of ascogenous hyphae, which forms a hymenium over the base and inner walls of the perithecium. The hymenium produces asci, typically interspersed among paraphyses. In some forms, the asci are restricted to an aparaphysate basal cluster; these exceptions include species ofClaviceps, Cordyceps and

Chaetomium, which produce lateral paraphyses. The Diaporthe type centrum is distinct from the Xylaria type in that the centrum consists of pseudoparenchymatous cells that expand and disintegrate to form the cavity of the perithecium; the asci develop into the disintegrating pseuodoparenchyma and eventually occupy the base of the inner perithecium. The Nectria centrum development is characterized by the presence of apical paraphyses that grow downward and form the central cavity by exerting pressure on the base of the 97

perithecium; they are considered distinct from the apically free paraphyses of the

Xylaria type. The ascogenous hyphae give rise to a hymenium that extends across

the base and lower walls of the inner perithecium; asci are produced in a concave cluster at the base of the perithecium among the apical paraphyses that begin to

evanesce. The Ophiostoma type centrum, like the Diaporthe type, is composed of pseudoparenchymatous cells; asci are produced irregularly within the centrum

cavity and not from a hymenial layer; the pseudoparenchyma disintegrates

concurrently with the production of asci; the ascus walls deliquesce filling the

centrum cavity with free ascospores.

The majority of these centrum types were based upon relatively few studies of still fewer taxa. Since Luttrell's treatise numerous studies have been conducted on the development of the centrum and its utility in the systematics of pyrenomycetous ascomycetes. By and large, the studies have led to a refinement of Luttrell’s centrum types rather than an abandonment of them. Development in the Sordariales is considered to be a variation of the Xylaria type centrum

(Whiteside, 1961; Huang, 1976; Froeyen, 1980), and it is proposed to represent a midpoint on a continuum from theXylaria type to the Diaporthe type (Uecker,

1976; Jensen, 1983). Also, several disparate modes of development have been recognized within the Ophiostoma type centrum (Rosinski, 1961; Corlett, 1963;

1966a; Steirs, 1976; van Wyk et al. 1991). These and others are discussed at greater length below. Jensen (1984) and Barr (1990) provide excellent reviews of additional salient studies.

The four centrum types described above are the focus of this study. The taxa that display these types of development form a natural monophyletic group of relatively derived unitunicate, perithecial ascomycetes (Ch. 4). The

Pyyllanctinia centrum type is not considered in this analysis. The ssrDNA gene 98

tree is proposed to function as an independent test of homology and polarity of

the characters and character states of centrum development.

Materials and Methods

The cladogram presented here (Fig. 5.1) was proposed in Chapter 4 (Fig.

4.1). It is a strict consensus of 42 equally most parsimonious trees. The taxa

sampled represent a survey of the major centrum types for pyrenomycetes

including the Ophiostoma type as defined by Luttrell (1951). Analysis of the data

and interpretation of bootstrap confidence levels and decay indices are discussed

elsewhere (Ch. 2 - 4). Characters derived from the development of the perithecium and its centrum were mapped onto the ssrDNA gene tree.

Results

The derived unitunicate, perithecial ascomycetes form a large monophyletic clade that is comprised of the two subclades, which are designated

A and B (Fig. 5.1). The Xylaria and Ophiostoma centrum types do not corroborate the monophyletic groups inferred from the analysis of the ssrDNA.

The two groups representingXylaria type development, the Clavicipitales and the

Xylariales, are not sister taxa, but are placed in disparate clades of the tree

(Fig 5.2). The Clavicipitales is a sister taxon to the Hypocreales, and represents a variation of theNectria centrum type. Taxa sampled for theOphiostoma type centrum include species from the Microascales, Ceratocystis and Ophiostoma.

Species of the Microascales and Ceratocystis are sister taxa, and together with the 99

— Aciculosporium take - Balansia sclerotica Claviceps paspali ■ Cordyceps capitata CLAVICIPITALES Epichloe typhina - Hirsutelk thompsonii — Hypocrella sp. Aphysiostroma stercorarium Hypocrea schweinitzii Hypocrea lutea Sphaerostilbella H I Sphaerostilbella aureonitens HYPOCREALES Hypocrea pallida Hypomyces polyporinus t Nectria haematococca P Neocosmospora vasinfeda L Nectria cinnabarina r Ceratocystis fimbriala — I L. Ceratocystis virescens OPHIOSTOMATALES 1— —...... — ...... Sphaeronaemella fimicola r — Halosphaeriopsis mediosetigera JH — Petriellasetifera MICROASCALES •— Microascus lrigonosporus I Diaporthe phaseolarum DIAPORTHALES f * - Leucostomapersoonii I r Ophiostoma piliferum OPHIOSTOMATALES I Ophiostoma ulmi r Cercophora septentrionalis _ J T - Chaetomium globosum SORDARIALES L— Neurospora crassa r Daldinia concentrica _ p — Diatrype discoformis [1 — Hypoxylon atroroseum XYLARIALES | r - Xylaria curta Xylaria hypoxylon Aspergillus fumigatus Nematospora sp. < Kathistes analemmoides r Kathistes calyculata Pyxidiophora la C Pyxidiophora spol HI Subbaromyces splendens Aureobasidium pullulans — Stictis radiata tEEleutherascus peruvianas — Ambrosiozyma platypodis Saccharomyces cerevisiae Candida albicans ENDOMYCETALES Candida tropicalis Cephaloascus fragrans Taphrim deformans TAPHRINALES • Athelia bombacina • Spongipellis unicolor B ASIDOMY CETES • Cryptococcus neoformans

Figure 5.1 Strict consensus tree of 42 equally most parsimonious cladograms containing 54 taxa presented in Chapter 4(898 steps, Cl = .427, RI = .702 and RC = .299). 100

— Aciculosporium take - Balansia sclerotica Claviceps paspali Cordyceps capitala Xylaria — Epichloe typhina Hirsutella thompsonii - Hypocrella sp. Aphysiostroma stercorarium Hypocrea schweinitzii Hypocrea lulea Sphaerostilbella NZ Sphaerostilbella aureonitens Hypocrea pallida Nectria Hypomyces polyporinus Nectria haematococca Neocosmospora vasinfecta ■ Nectria cinnabarina i t Ceratocystis fimbriata _ _ j L . Ceratocystis virescens ' ...... Sphaeronaemellafimicola Ophiostoma r — Halosphaeriopsis mediosetigera J ” *— Petriella setifera L — Microascus trigonosporus i Diaporthe phaseolarum I »— Leucostoma persoonii | Diaporthe r Ophiostoma piliferum Ophiostoma 1 Ophiostoma ulmi r Cercophora septenlrionalis _ _ P - C haelom ium g b b o su m ‘— — Neurospora crassa r Daldinia concentrica Xylaria _ p — Diatrype discoformis P — Hypoxylon atroroseum | i “ Xylaria curta ^ — - Xylaria hypoxylon Aspergillus fumigatus Nematospora sp. ' Kathistes analemmoides Kathistes calyculata CPyxidiophora la Pyxidiophora spol Subbaromyces splendens Aureobasidium pullulans Stictis radiata Eleutherascus peruvianus Ambrosiozyma platypodis Saccharomyces cerevisiae Candida albicans Candida tropicalis Cephaloascus fragrans Taphrina deformans • Athelia bombacina • Spongipellis unicolor Cryptococcus neoformans

Figure 5.2 Centrum types as defined by Luttrell (1951) are mapped on the strict consensus tree. The Xylaria and Ophiostoma centrum types do not define monophyletic groups as inferred by cladistic analysis of partial sequences from ssrDNA. 101

Hypocreales and the Clavicipitales comprise subclade A of the derived unitunicate perithecial ascomycetes (Fig. 5.2).

The Xylaria - Sordaria - Diaporthe centrum development continuum

proposed by Uecker (1976) is not rejected in a maximum parsimony analysis.

The species sampled for Ophiostoma are sister taxa to the Diaporthales and may

represent derived states of this continuum. The Xylariales, Sordariales,

Diaporthales and Ophiostoma comprise subclade B of the derived unitunicate, perithecial ascomycetes.

Discussion

The phylogenetic hypothesis proposed here suggests that an early split

occurred among these fungi that resulted in two lines of evolution (subclades A

and B). Within each line there are examples of parallel and convergent evolution

of centrum development. The Nectria type centrum is diagnostic of the

Hypocreales, and is characterized by apical paraphyses (Luttrell, 1965). The

centrum cavity in Nectria haematococca (Hanlin, 1971) and N. cinnabarina

(Strikmann and Chadefaud, 1961; Parguey-Leduc, 1967) is formed by apical paraphyses that grow downward pushing the ascogenous system against a degenerative subhymenial layer of pseudoparenchyma. The asci are produced from a basal to lateral concave hymenium and grow up among the evanescent apical paraphyses. The same pattern of development is described for

Neocosmospora vasinfecta (Doguet, 1956). The pattern of development in the genera Hypomyces and Hypocrea varies somewhat from theNectria type. Species of these genera produce centripetal paraphyses from the innermost layers of the perithecium walls (Hanlin, 1963; 1964; 1965; Canham, 1969; Samuels, 1973; 102

Carey and Rogerson, 1981); however, the uppermost paraphyses of the

centripetal group elongate considerably relative to the more lateral ones. These

elongated paraphyses are treated as homologous to the apical paraphyses in

Nectria. The difference in paraphysis production is supported by the separation

of the these two groups in the ssrDNA cladogram (Fig. 5.1; Fig. 5.3).

The Clavicipitales has been described as displaying a variation of the

Xylaria type centrum similar to the development described for the genus

Chaetomium (Sordariales). This affinity is based on the presence of lateral

paraphyses and the production of asci from an aparaphysate, basal cluster

(deBary, 1887; Luttrell, 1951; Barr, 1990). However, the central cavity in

Chaetomium comprises pseudoparenchyma (Whiteside, 1961; 1962a; 1962b),

which has not been described for the Clavicipitales. Mhaskar and Rao (1976)

recognized the Claviceps type centrum to distinguish it from the Xylaria type. If

one accepts the proposed sister taxa relationship for the Hypocreales and

Clavicipitales, the development of the latter may be viewed as a variation of the

Nectria type centrum. In Epichloe typhina (Doguet, 1960) and Epichloe cinerea

(Mhaskar and Rao, 1976) the lateral paraphyses are proposed to function in

creating the central cavity in a manner similar to the apical paraphyses of the

Nectria type. The same is true for Cordyceps agariciformis', however, basal

evanescent paraphyses were also described (Jenkins, 1934). The similarities in function, evanescence and nonhymenial origin argue for the homology of apical

and lateral paraphyses of these two orders.

The remaining taxa in subclade A form a monophyletic group that is exemplified by the Microascales and Ceratocystis. The majority of these were described as exhibiting an Ophiostoma type centrum development (Luttrell,

1951); the central cavity is formed by the disintegration of pseudoparenchyma, 103

and asci are produced from an ill-defined hymenium. In Microascus (Corlett,

1963) and Petriella (Corlett, 1966a) the ascogenous system is displaced to the

apex of the centrum cavity by evanescent, sterile cells (hyphae) that grow inward

from the base and sides. The ascogenous hyphae grow downward and produce

terminal asci that become arranged irregularly throughout the centrum cavity;

ascus walls are evanescent. Similar development was described for Lophotrichus

(Whiteside, 1962). Halosphaeriopsis mediosetigera (Halosphaeriales), a marine

saprobe, is a sister taxon toPetriella setifera (Microascales). Like the Microascales, the centrum ofH. mediosetigera consists of pseudoparenchyma, hymenial paraphyses are absent, and the asci are evanescent (Kohlmeyer and

Kohlmeyer, 1966). Also, both possess black perithecia with long necks and passively discharge their ascospores. Unlike the Microascales, the asci ofH. mediosetigera are produced from a basal cluster, and sterile cells, termed catenophyses, are present. It is possible that the basal hymenium represents a nondisplaced microascalean ascogenous system, and that the catenophyses are homologous to the sterile cells of the microascalean centrum. However, more developmental studies are needed among the Halosphaeriales and Microascales, and more sampling within the two orders, and marine taxa in general, is required in the ssrDNA tree.

Ceratocystis sensu stricto is a sister taxon to the Microascales. In C. fimbriata the centrum cavity is formed by disintegration of pseudoparenchyma; the ascogenous system is scattered and fragmented; it closely approximates the pseudoparenchyma that lines the peridium; no paraphyses have been described

(Elliott, 1925; Stiers, 1976). In C. moniliformis young asci are produced around the periphery of the centrum and released into the cavity; the asci evanesce and ascospores fill the centrum cavity (van Wyk et al., 1991). Like the Microascales, 104

the centrum is pseudoparenchymatous and asci are evanescent. Also, their perithecia are black and consist of a globose base with a long neck from which

ascospores are discharged passively. Within subclade A there are two lines of centrum evolution (Fig. 5.2).

One is represented by the Nectria and Claviceps types of development of the

Hypocreales and Clavicipitales, respectively. The genera Hypomyces and Hypocrea represent a variation of theNectria type, the Hypomyces centrum type

(Samuels, 1973), that is consistent with the ssrDNA data. The other line is represented by the Microascales and Ceratocystis. These two groups form the centrum cavity by disintegration of pseudoparenchymatous cells; hymenial paraphyses are absent; and the asci are evanescent. However, there is considerable variation in the location of the mature ascogenous system within the centrum and the presence of sterile cells.

Subclade B comprises species from the orders Xylariales, Sordariales and

Diaporthales, which have been proposed to represent a continuum in the evolution of centrum development (Uecker, 1976). The Xylaria type is the classical example of centrum development in unitunicate, perithicial ascomycetes

(Luttrell, 1951; Alexopoulos and Mims, 1979). Luttrell (1951) included all species that possessed centra with true paraphyses. Xylaria hypoxlyon and X. polymorpha were described over a century ago as possessing paraphyses and a hymenial layer covering the pseudoparenchymatous wall layer of the inner perithecium (Fiiisting, 1867; deBary, 1887). InX. tentaculata andX. trachelina evanescent paraphyses develop from the inner cells of the peridium and an hymenium extends over the entire inner perithecium (Brown, 1913). During development in H. pruinatum, centrum pseudoparenchyma is replaced by paraphysis-like hyphae, which is replaced by more pseudoparenchyma. The 105

hymenium lines the entire inner perithecium and produces asci among persistent paraphysis-like hyphae (Rogers and Berbee, 1964). In H. fuscum the hymenium

is located at the base of the perithecium atop a layer of subhymenial

pseudoparenchyma (Rogers, 1967). Jensen (1981) reported the development of

H. serpens as being similar to H. pruinatum; however, he interpreted the centrum

pseudoparenchyma as swollen paraphyses. He defined true pseudoparenchyma

as being restricted to the subhymenial layer. Eutypella parasitica (Diatrypaceae)

was described as having paraphyses and a hymenium that lines the entire inner

perithecium (Lachance and Kuntz, 1970); however, little work has been done in

the Diatrypaceae (Glawe and Rogers, 1984).

Members of the Sordariales possess a Sordaria type centrum that was

treated as a variation of the Xylaria type (Huang, 1976). Mirza and Khatoon

(1973) reported Sordaria humana as having centrum pseudoparenchyma and true paraphyses. Uecker (1976) corroborated their findings. He showed that the asci are produced from a basal hymenium, and that the paraphyses are produced after the centrum cavity is developed, thus questioning their role in space-making.

Similar development is seen in Podospora arizonensii (Mainwaring and Wilson,

1968), Gelasinospora (Ellis, 1960; Jensen, 1982), Neurospora (Nelson and Backus,

1968) and Triangularia (Moreau and Moreau, 1950; Huang, 1976). Centrum development in the genus Chaetomium differs from the Sordaria type in that the basal cluster of asci is aparaphysate, although paraphyses are produced from the lateral walls (Whiteside, 1961; 1962a; 1962b). The asci are not interspersed among the paraphyses (Corlett, 1966b).

The Diaporthales is a sister taxon to the Sordariales in the ssrDNA tree.

The Diaporthe type comprises centrum pseudoparenchyma, and a basal hymenium that releases free, intact asci into the central cavity (Luttrell, 1951). 106

The centrum was orginally described as lacking paraphyses; however, later they

were demonstrated in Gnomonia (Morgan-Jones, 1959; Huang and Luttrell,

1982) and Diaporthe phaseolarum (Jensen, 1983; Uecker, 1988). The

diaporthalean paraphyses are inconspicuous due to their short length and

compression by the free asci filling the centrum cavity.

The genus Ophiostoma is a sister taxon to the species sampled for the

Diaporthales (Fig. 5.1). Curiously, the two taxa have much in common with

respect to centrum development. Development in Ceratocystis multiannulata

(= Ophiostoma) involves centrum pseudoparenchyma and paraphysis-like bands

(Andrus, 1936). Paraphysis-like bands were also described for C. major

( = Ophiostoma)', additionally, the asci were produced from a basal hymenium

(Hutchinson, 1950). A basal hymenium was also reported for O. ulmi (Rosinski,

1961). van Wyk and Wingfield (1991a; 1991b) demonstrated that the asci of O. multiannulata and O. davidsonii are produced from the base of the inner perithecium, then released into the centrum cavity. Once in the centrum cavity the ascus walls evanesce and release free ascospores. This is not unlike the

Diaporthe type centrum; asci are produced from a basal hymenium and asci are released into the centrum cavity. The difference is that the ascus walls in

Ophiostoma are evanescent while those of theDiaporthe type are persistent. The presence of paraphyses in Ophiostoma remains questionable.

The developmental continuum extending fromXylaria through Sordaria to

Diaporthe centrum types (Uecker, 1976) is not rejected by maximum parsimony analysis of the ssrDNA. At one end is the Xylaria type centrum with true paraphyses and a hymenium that covers the entire inner perithecium. A subhymenial layer of pseudoparenchyma is present in several taxa. Diatrype is included here; the hymenium is of the Xylaria type, and both the Xylariaceae and 107

the Diatrypaceae possess prosenchymatous peridial cells (Jensen, 1985). The

Sordaria type centrum is indicative of a hymenium that results in a more basal restriction of asci and a decrease in the production of paraphyses. The extreme example of theSordaria centrum type is the strict aparaphysate basal cluster and lateral paraphyses in the genus Chaetomium. Further restriction of the hymenium to the base of the perithecium, a continued decrease in prominent paraphyses, and the release of free asci into the centrum cavity characterizes the

Diaporthe type centrum. The Sordariaceae and the Diaporthaceae also both have pseudoparenchymatous peridial cells (Jensen, 1985).

The addition of Ophiostoma H. & P. Sydow to this continuum may seem odd. However, studies in several species of the genus reported the production of asci from a putative basal hymenium and subsequent release of asci into the centrum cavity. Ophiostoma differs from Diaporthe in the production of evanescent asci; however, evanescent asci have evolved several times in unitunicate, perithecial ascomycetes (Ch 3; Ch 4; Berbee and Taylor, 1992b).

The presence of evanescent asci in Chaetomium (Whiteside, 1961; 1962) provides precedence for evanescent asci in the continuum. Luttrell (1951) commented on particular similarities between centrum development in certain species of

Ophiostoma and the Diaporthales. However, the treatment ofCeratocystis as congeneric with Ophiostoma combined with relatively few studies apparently led him to discount these similarities. Thus, evanescent asci may represent a derived state of the Diaporthe type centrum.

The centrum as a whole is defined as the character and the variation observed in the components of the centrum constitute its character states.

Taking this approach, the centrum types as defined by various investigators are mapped onto the ssrDNA gene tree (Fig. 5.3). These data are congruent with the 108

proposed centrum development continuum extending fromXylaria through

Sordaria to Diaporthe types of development, and the mapping of character states

on the ssrDNA tree allows the centrum continuum to be polarized. TheXylaria

centrum type represents the most primitive character state of the continuum, and

the Diaporthe type, including the genus Ophiostoma, is inferred as the most

derived character state of the continuum. The interpretation of this polarity in

relation to the taxonomic statement is critical. As argued previously, the addition

of taxa can or may lead to the survey of more characters states (Ch. 4). If the

addition of taxa results in the survey of new character states,i.e. centrum types not previously included, the placement and polarity of the character states of the continuum may be disrupted (Donoghue et al., 1989). Obviously, this potential exists and will continue to exist because the inclusion of all known and unknown unitunicate, perithecial ascomycetes is at best a remote possibility. However, the inclusion of taxa that display unique centrum development relative to the whole of unitunicate perithecial acomycetes, e. g. species ofMelanospora and

Glomerella, should receive priority in future taxon sampling.

The development of spatial relationships between the paraphyses and the hymenium is an important tendency that partially defines the character states of the centrum (Luttrell, 1951). The development of paraphyses and resulting terminologies for them have long been a source of confusion (Luttrell, 1965).

Following Jensen (1984), the apical paraphyses of theNectria centrum are treated as true paraphyses, and are not homologous to pseudoparaphyses of loculoascomycetes (Strickman and Chadefaud, 1961). The placement of

Microascales within subclade A predicts that the evanescent sterile hyphae of the

Microascus centrum may be homologous to the lateral and apical paraphyses of the Clavicipitales and Hypocreales, respectively. Also, the presence or absence 109 Aciculosporium take - Balansia sclerotica Claviceps paspali ■ Cordyceps capitata Claviceps (Mhasker & Rao, 1976) Epichloe typhina - Hirsutella thompsonii - Hypocrella sp. Aphysiostroma stercorarium Hypocrea schweinitzii Hypocrea lutea Sphaerostilbella NZ Hypomyces (Samuels, 1973) Sphaerostilbella aureonitens Hypocrea pallida Hypomyces polyporinus Nectria haematococca Neocosmospora vasinfecta Nectria (Luttrell, 1951) t■ Nectria cinnabarina p Ceratocystisfimbriata Ceratocystis _ ] Ceratocystis virescens (van Wyk I------Sphaeronaemellafimicola etal., 1991) i— Halosphaeriopsis mediosetigera Microascus J""i— Petriella setifera (Corlett, 1963; 1966) L— Microascus trigonosporus I Diaporthe phaseolarum Diaporthe Jensen (1983) L- Leucostoma persoonii j Ophiostoma piliferum Ophiostoma (van Wyk & Wingfield, Ophiostoma ulmi 1991a; 1991b) Cercophora septentrionalis r Sordaria (Huang, 1976) _ M - Chaetomium globosum (Uecker, 1976) »— Neurospora crassa r Daldinia concentrica J l - — Diatrype discoformis I— Hypoxylon atroroseum Xylaria (Luttrell, 1951) j - Xylaria curta "L— Xylaria hypoxylon

Figure 5.3 Enlargement of the clade comprising derived unitunicate, perithecial ascomycetes. Centrum development is mapped on the tree denoting the genus and investigator of the centrum types that are congruent with the ssrDNA gene tree. * Genera were not included in centrum description. 110

and location of pseudoparenchyma are important. Pseudoparenchyma has been

described as a major component of the centra in two separate groups, the

Microascales and Ceratocystis clade and the Diaporthales and Ophiostoma clade. However, Jensen (1981) described the centrum pseudoparenchyma of Hypoxylon serpens as swollen paraphyses, and the presence of paraphysis-like bands was

reported for Ophiostoma (Andrus, 1936; Hutchinson, 1950). These

interpretations beg the question of whether the pseudoparenchyma of the microascalean clade is homologous to that in other groups. Cladistic analysis of partial sequences of the ssrDNA suggest that pseudoparenchyma is not homologous across the taxa sampled. Additional studies are warranted, especially in groups such as the Microascales, Ceratocystis and Ophiostoma, to understand better the presence, origin and variation in sterile hyphae and paraphyses of the centrum; the same is true for pseudoparenchyma. I l l

Literature Cited

Alexopoulos, C. J. and C. W. Mims. 1979.Introductory Mycology, 3rd ed. John Wiley & Sons, N. Y. 632 p.

Andrus, C. F. 1936. Cell relations in the perithecium of Ceratostomella multiannulata. Mycologia 28: 133-153.

Barr, M. E. 1990. Prodromus to nonlichenized, pyrenomycetous members of class Hymenoascomycetes. Mycotaxon 39: 43-184.

Berbee, M. L. and J. W. Taylor. 1992a. Two ascomycete classes based upon fruiting-body characters and ribosomal DNA sequence. Molec. Biol. Evol. 9: 278-284.

and . 1992b. Convergence in ascospore discharge mechanism among pyrenomycete fungi based on 18S ribosomal RNA gene sequence. Molec. Phylog. Evol. 1: 59-71.

Brown, H. B. 1913. Studies in the development of Xylaria. Ann. Mycol. 11: 1-11.

Canham, S. C. 1969. Taxonomy and morphology Hypocreaof citrina. Mycologia 61: 315-331.

Carey, S. T. and C. T. Rogerson. 1981. Morphology and cytology ofHypomyces polyporinus and its Sympodiophora anamorph. Bull. Torr. Bot. Club 108: 13-24.

Corlett, M. 1963. The developmental morphology of two species ofMicroascus. Can. J. Bot. 41: 253-266.

. 1966a. Developmental studies in the Microascaceae. Can. J. Bot. 44: 79-88.

— . 1966b. Perithecium development in Chaetomium trigonosporum. Can. J. Bot 44: 155-162. deBary, A. 1887. Comparative morphology and biology of the fungi, mycetozoa and bacteria. Oxford, Clarendon Press. 525 p.

Doguet, G. 1956. Morpholgie et organogenie duNeocosmospora vasinfecta E. F. Smith et du Neocosmospora africana von Arx.Ann. Sci. Nat. Bot. (Ser. 11) 17: 353-370.

. i960. Morphologie, organogenie et evolution nucleaire deCEpichloe typhina. La place des Clavicipitaceae dans la classification. Bull. Soc. Mycol. Fr. 76: 171-203. 112

Donoghue, M J., J. A. Doyle, J. Gauthier, A. G. Kluge and T. Rowe. 1989. The importance of fossils in phylogeny reconstruction.Ann. Rev. Ecol. Syst. 20: 431-460.

Elliot, J. A. 1925. A cytological study ofCeratostomella fimbriata (E. & H.) Elliot. Phytopathology 15: 417-422.

Ellis, J. J. 1960. Plasmogamy and ascocarp development of Gelasinospora calospora. Mycologia 52: 557-573.

Fries, E. M. 1823. Systema Mycologicum 2. Lundae.

Froeyen, O. J. 1980. Perithecium growth and expansion inChaetomium globosum. Can. J. Bot. 58: 375-383.

Fuisting, W. 1867. Zur Entwickelungsgeschichte der Ascomyceten, Vorlaufige Mittheilung. II. Bot. Zeit. 25: 305-311.

Glawe, D. A. and J. D. Rogers. 1984. Diatrypaceae in the pacific northwest. Mycotaxon 20: 401-460.

Hanlin, R. T. 1961. Studies in the genus Nectria II. Morphology of A. gliocladioides. Am. J. Bot. 48: 900-908.

. 1963. Morphology ofHypomyces lactifluorum. Bot. Gaz. 124:395-404.

. 1964. Morphology ofHypomyces trichothecoides. Am J. Bot. 51: 201- 208. . 1965. Morphology ofHypocrea schweinitzii. A m J.Bot. 52: 570-579.

. 1971. Morphology ofNectria haematococca. Am J.Bot. 58: 105-116.

Huang, L. H. 1976. Developmental morphology ofTriangularia backusii (Sordariaceae). Can. J. Bot. 54: 250-267.

and E. S. Luttrell. 1982. Development of the perithecium in Gnomonia comari (Diaporthaceae). Am. J. Bot. 69: 421-443.

Hutchinson, S. A. 1950. The perithecia of Ophiostoma majus (van Beyma) Goidanich. Ann. Bot. (N. S.) 14:115-125.

Jenkins, W. A. 1934. The development of Cordyceps agariciformia. Mycologia 26: 220-243.

Jensen, J. D. 1981. The developmental morphology ofHypoxylon serpens in culture. Can. J. Bot. 59: 40-49.

. 1982. The development of Gelasinospora reticulospora. Mycologia 74: 724-737. 113

. 1983. The development ofDiaporthe phaseolarum variety sojae in culture. Mycologia 75: 1074-1091.

. 1984. Morphologic studies in the pyrenomycetes. Ph.D. dissertation. Washington State University, Pullman. 121 p.

. 1985. Peridial anatomy and pyrenomycete taxonomy.Mycologia 77: 688-701.

Kohlmeyer J. and E. Kohlmeyer. 1966. On the life history of marine Ascomycetes: Halosphaeria mediosetigera and H. circumvestita. Nova Hedwigia 12: 189-202.

Lachance, D. and J. E. Kuntz. 1970. Ascocarp development of Eutypella parasitica. Can. J. Bot. 48: 1977-1979. Luttrell, E. S. 1951. Taxonomy of the pyrenomycetes.University o f Missouri Studies 3: 1-120.

Mainwaring, H. R. and I. M. Wilson. 1968. The life cycle and cytology of an apomicticPodospora. Trans. Br. Mycol. Soc. 51: 663-677.

Mhasker, D. N. and V. G. Rao. 1976. Development of the ascocarp inEpichloe cinerea (Clavicipitaceae). Mycologia 68: 994-1001.

Mirza, J. H. and A. Khatoon. 1973. Studies onSordaria Humana (Funkel) Winter: The cytology of ascus development and developmental morphology of the perithecium.Pak. J. Bot. 5: 19-28.

Moreau, F. and V. Moreau. 1950. Etude du developpement du Triangularia bambusae (van Beyma) Boedijn. Rev. Mycol. 15: 146-158.

Morgan-Jones, J. F. 1959. Morpho-cytological studies of the genusGnomonia. III. Early stages of perithecial development. SvensL Bot. tidsk. 53: 81-101.

Nelson, A. C. and M. P. Backus. 1968. Ascocarp development in two homothallic Neurospora. Mycologia 60: 16-28.

Parguey-Leduc, A. 1967. Recherches sur l'ontogenie et l'anatomie comparee des ascocarpes des Pyrenomycetes ascoloculaires. Seconde partie. Les ascocarpes des Pyrenomycetes ascoloculaires unituniques. Ann. Sci. Nat. Bot. (ser. 12) 8: 1-110.

Reynolds, D. R. (Ed.) 1981. Ascomycete Systematics: the Luttrellian concept. Springer-Verlag, N. Y. 242 p.

Rogers, J. D. and J. G. Berbee. 1964. Developmental morphology ofHypoxylon pruinatum in bark of quaking aspen. Phytopathology 54: 154-162.

Rogers, J. D. 1967. Hypoxylon fuscum II. Developmental morphology onAlnus tenuifolia. Bot. Gaz. 128: 186-197. 114

Rosinski, M. A. 1961. Development of the ascocarp ofCeratocystis ulmi. Amer. J. Bot. 48:285-293.

Samuels, G. J. 1973. Perithecial development in Hypomyces aurantius. Am. J. Bot. 60: 268-276.

Stiers, D. L. 1976. The fine structure of ascospore formation inCeratocystis fimbriata. Can. J. Bot. 54: 1714-1723.

Strickman, E. and M. Chafaud. 1961. Recherches sur les asques et les peritheces des Nectria et reflxions sur revolution des Ascomyc&tes.Rev. Gen. Bot. 68: 725-770.

Uecker, F. A. 1976. Development and cytology ofSordaria humana. Mycologia 68: 30-46.

. 1988. A timed sequence of development ofDiaporthe phaseolarum (Diaporthaceae) fromStokesia laevis. Mem. N. Y. Bot. Gar. 49: 38-50.

Wyk, P. W. J. van, M. J. Wingfield and P. S. van Wyk. 1991. Ascospore development on Ceratocystis moniliformis. Mycol. Res. 95: 96-103.

.—.— and M. J. Wingfield. 1991a. Ultrastructure of ascosporogenesis in Ophiostoma davidsonii. Mycol. Res. 95: 725-730.

a n d . 1991b. Ascospore ultrastructure and development in Ophiostoma cucullatum. Mycologia 83: 698-707

Wehemeyer, L. E. 1926. A biologic and phylogenetic study of stromatic Spaeriales. Am. J. Bot. 13: 574-645.

Whiteside, W. C. 1961. Morphological studies in the Chaetomiaceae. I. Mycologia 53: 512-523.

Whiteside, W. C. 1962a. Morphological studies in the Chaetomiaceae. II. Mycologia 54: 152-159.

Whiteside, W. C. 1962b. Morphological studies in the Chaetomiaceae. III. Mycologia 54: 611-620. CHAPTER 6

SUMMARY AND CONCLUSIONS

The dissertation presented here offers some new insights into the

systematics of unitunicate perithecial ascomycetes as inferred by cladistic analysis

of partial sequences of the ssrDNA. Those taxa that are characterized by unitunicate perithecia do not form a monophyletic group in a maximum parsimony analysis (Fig. 4.1). The analyses presented here reveal a large clade of relatively derived taxa that comprises two subclades. Subclade A consists of the orders Clavicipitales, Hypocreales, Microascales and the genus Ceratocystis, and subclade B contains the orders Diaporthales, Sordariales, Xylariales and the genus Ophiostoma. Three genera, Kathistes, Pyxidiophora and Subbaromyces, are confidently separated from the larger clade, thus defining the nonmonophyly.

The Clavicipitales is inferred to be a sister taxon to the Hypocreales and is not closely related to the Xylariales (Fig. 2.1). This placement of the

Clavicipitales agrees with the classifications of Gaumann (1952) and Rogerson

(1970) and counters those of Luttrell (1951) and Barr (1983; 1990). In doing so, the ssrDNA data correlates with characters derived from conidiogenesis, stromatal structure and nutritional modes. The hypocrealean taxa sampled form two clades that are paraphyletic in a strict consensus tree. One clade is exemplified by the genera Hypocrea and Hypomyces and is a sister group to the

Clavicipitales. The second clade contains species of the genera

115 116

Nectria and Neocosmospora, and is a sister group to the Clavicipitales- Hypocreales clade. However, the branch lengths connecting these three clades

are quite short, and the resolution of the larger clade containing these three

clades is dependent on taxon sampling (Fig. 3.1). The Hypocreales is a large and

diverse group of fungi that require further taxon sampling, especially in the

genera Nectria and Gibberella. Therefore, I prefer to view the Clavicipitales-

Hypocreales clade as a polytomy comprising at least three lines of evolution that

may be the product of a radiation event.

The Clavicipitales are monophyletic, although unresolved (Fig. 2.1). A

successive approximation of the data based upon the rescaled consistency indices

(RC) produced one of the 42 equally most parsimonious trees inferred by the

unweighted data (Fig. 2.2). The clavicipitalean clade is fully resolved producing a

continuum extending from fungal parasite through insect parasite to plant

endophyte (Fig. 2.3). Six pairs of taxa were used as "place-holders" in six

bootstrap analyses; the pairs consisted ofClaviceps paspali and the remaining

taxa sampled from the order. The two pairs involvingCordyceps capitata or

Aciculosporium take did not receive significant bootstrap confidence levels, whereas the remaining four pairs received marginally significant to significant

confidence levels (Fig. 2.4; Tbl. 2.2). It is interesting that the two pairs that did

not receive significant values represent the extremes of the continuum inferred by

successive approximation (Fig. 2.2). However, bootstrapping (Felsenstein, 1985)

and successive approximation (Farris, 1969) are analyses rooted in different

philosophies of phylogenetics. The concerted use of these two analyses is an

eclectic approach to data analysis that is potentially subject to criticism.

However, it does allow one to construct a working hypothesis for future studies

concerning the evolution of nutritional modes within the order. 117

The ophiostomatalean taxa sampled are polyphyletic. Two groups were

inferred representing Ceratocystis sensu stricto and Ophiostoma P. & H. Sydow

(de Hoog and Scheffer, 1984) (Fig. 3.1). This separation agrees with a host of

nonmolecular characters derived from the cell wall, conidiogenesis, centrum

anatomy and antibiotic resistance (Tbl. 3.1). Sphaeronaemella fimicola is a sister

taxon toCeratocystis and is on a relatively long branch; Ceratocystis and S. fimicola are placed as sister taxa to the Microascales. Significant bootstrap

confidence levels for the placement ofCeratocystis as a sister taxon to the

Microascales is contingent on inclusion or exclusion S. offimicola (Fig. 3.3). The

inclusion ofS. fimicola has a dampening effect throughout subclade A, also (Fig.

3.2). A similar response was displayed in the decay indices; however, it appeared to be more localized to the microascalean clade. Ophiostoma is currentlt viewed as a sister taxon to the species sampled from the Diaporthales, but this is not supported by bootstrapping or decay indices.

Bootstrap confidence levels should be interpreted with caution. One of the major assumptions in bootstrapping is that the characters were chosen at random from the pool of all characters for the taxa in question (Felsenstein,

1985). These data were not chosen randomly. The ssrDNA was chosen specifically for its conserved nucleotide sequence, which is postulated to be more phylogenetically informative (cladistically reliable). Therefore, the bootstrap values presented in this study are more appropriate for the ssrDNA gene trees than they are for the species tree. The sensitivity of bootstrapping to sampling subsets of a monophyletic group underscores this qualification.

Centrum development was treated as a single character and the variations in its components as the characters states. The placement of Clavicipitales,

Ceratocystis and Ophiostoma provide new information that assists in determining 118

homology and polarity of character states. TheXylaria and Ophiostoma centrum

types defined by Luttrell (1951) are not consistent with monophyletic groups

inferred by ssrDNA. The Hypomyces (Samuels, 1973) and Claviceps (Mhaskar

and Rao, 1976) centrum types are recognized and inferred as variations of the

Nectria centrum type. The apical paraphyses of the Nectria centrum type are

treated as homologous to the lateral paraphyses of theClaviceps centrum.

The centrum developmental continuum extending from theXylaria

through Sordaria toDiaporthe (Uecker, 1976) is supported by maximum

parsimony; however, the supraordinal relationships did not receive significant bootstrap confidence levels or relatively high decay indices (Fig. 4.2). The polyphyly ofCeratocystis and Ophiostoma supports the recognition of two

different centrum types (van Wyk et al., 1991). The Ceratocystis-Microascales clade is characterized by a relatively variable centrum anatomy. Ophiostoma is inferred as a derived state of the Diaporthe centrum type; both the Diaporthales and Ophiostoma produce their asci from the base of the perithecium and release them into the centrum cavity (Jensen, 1983; van Wyk and Wingfield, 1991a;

1991b). Evanescent asci are proposed to have arisen several times (Fig. 4.3).

Future ontogenetic studies should focus on the Microascales,Ceratocystis and

Ophiostoma to understand better the variation that exists among these taxa.

Cladistic analysis of the ssrDNA data separates the genera Pyxidiophora,

Kathistes and Subbaromyces from the clade of relatively derived unitunicate, perithecial ascomycetes. The separation is supported by a bootstrap confidence level of 100% and a decay index of > 10 steps. However, as discussed previously, these values are interpreted with caution with respect to the phylogeny of the organisms. However, the maximum parsimony analysis suggests that the search for near relatives of these three genera be directed away from the traditional 119

unituncate, perithecial ascomycetes. However, the relationships of the three

genera with their putative sister taxa are questionable. The differences in the

makeup of perithecial walls between these three genera and the relatively more

derived clade may prove to be informative in recognizing differences in

development of the ascoma. Clearly, more work is needed on the development

of the ascoma. What is interesting is the putative relationship ofSubbaromyces splendens and Pyxidiphora spp. They share many characters in common that

involve ascospores and ascoma development; combined with the perithecial walls, these characters can serve as guides in taxon sampling in future studies.

Systematics has experienced nothing less than a revolution over the past

25 years, and molecular characters and phylogenetics have been at the forefront. Currently, the overwhelming majority of studies in molecular systematics derive characters from one gene; therefore, the resulting phylogenies are gene trees.

The correlation of the molecular data with morphological and other non- molecular characters lends some credence to the extrapolation of the gene tree as an organismal tree. However, future research will undoubtedly include more regions of the genome (genes) and rely less on single genes. In this study, the B subunit of vacuolar ATPase (BvATPase) was sampled as an independent test of the ssrDNA. The BvATPase was something of a success and something of a failure. Analysis of the first and second codon position supported the separation of the three primitive genera (Subbaromyces, Pyxidiophora and Kathistes) away from the larger clade. The inclusion of the third codon position into the analysis disrupted the topology and did not support the separation of unitunicate ascomycetes. The resulting cladogram did not support relationships that are supported by ssrDNA, the first and second position of the codon of BvATPase, and morphology. 120

Comparative tests involving the retention index (RI) and the rescaled

consistency index (RC) (Farris, 1989) within the BvATPase data showed that an inordinate amount of homoplasy was associated with the third position of the

codon. The dramatic decrease in RI between the two BvATPase data sets suggests that not only was there an increase in the frequency of character state variation, but there was a substantial increase in variation of character state distribution, also. The RI did not decrease between the 14 taxa and 54 taxa data sets, rather it increased slightly. As expected the addition of taxa did not result in an increase in the variation of character state distribution. The decrease in the

RC observed in the ssrDNA resulted solely from an increase in the variation of character states. That is, characters that were invariant or uniformative

(autapomorphic) in the 14 taxa data set are putative synapomorphies in the 54 taxa data set. The terms vertical and horizontal homoplasy were introduced to differentiate between the homoplasy detected by the addition of taxa and the homoplasy introduced by the addition of characters, respectively. It is important to note that RIs and RCs were not compared between the ssrDNA and

BvATPase data sets; rather the behavior of the two indices relative to the addition of taxa and characters was compared between them.

The stability of the RI for the ssrDNA data is proposed here to provide support for the use of successive approximation. Successive approximation operates under the assumption that there exist cladistically reliable characters and cladistically unreliable characters. Cladistically unreliable characters are characterized by similarities in character states that reflect homoplasy.

Cladistically reliable characters are characterized by similarities in character states that reflect homology (Farris, 1969). The stable, relatively high RI values of the ssrDNA suggest that the homoplasy is characterized by relatively low levels 121

of variation in character state distribution, i. e. the characters states vary in a

hierarchical manner. The decrease in RI for the BvATPase data is indicative of

high levels of variation in character state distribution, i. e. the character states

vary in a random manner. Thus, a successive approximation of the ssrDNA data

produced one of the 42 equally most parsimonious trees inferred by the

unweighted data in Chapter 2 (Fig. 2.2), but it did not choose any of the equally

most parsimonious trees inferred by the unweighted BvATPase data (data not shown).

As new regions (genes) of the genome are surveyed for phylogenetically informative characters, some genes will be encountered that are appropriate

(cladistically reliable) and some that are inappropriate (cladistically unreliable) with respect to the relatedness of the taxa sampled. Then, a decision must be made as to which, if either, gene is cladistically reliable. In this dissertation two such genes were encountered. Annalysis of these two genes inferred two different phylogenetic hypotheses. One might argue that a decision can not be made favoring either gene tree; however, this approach completely discounts the relevance and importance of morphologically based hypotheses for the taxa in question. After all, studies in molecular systematics are usually initiated to test some morphologically based hypothesis. If an unlinked gene tree displays congruence with the morphologically based tree, a second gene tree, which does not display congruence with either the first gene tree or the morphological tree, does not override the first gene. This is especially true when it can be demonstrated that the second gene tree is based on data that are randomized with respect to evolution. Thus, both the ssrDNA tree and corroborating evidence from a new interpretation of morphological characters suggest that unitunicate, perithecial ascomycetes are nonmonophyletic. 122

Literature Cited

Barr, M. E. 1990. Prodromus to nonlichenized, pyrenomycetous members of class Hymenoascomycetes. Mycotaxon 39: 43-184. . 1983. The ascomycete connection.Mycologia 75:1-13.

Farris, J. S. 1989. The retention index and the rescaled consistency index. Cladistics 5: 417-419.

. 1969. A successive approximation approach to character weighting. Syst. Zool. 18: 374-385.

Gaumann, E. A. and C. W. Dodge. 1928. Comparative Morphology of Fungi. New York: McGraw-Hill Book Co., Inc. p.701.

Gaumann, E. A. 1952. The Fungi: A Description o f Their Morphological Features and Evolutionary Development. New York: Hafner publishing Co. p.420.

Hoog, G. S. de and R. J. Scheffer. 1984. Ceratocystis versus Ophiostoma-. a reappraisal. Mycologia 76: 292-299.

Jensen, J. D. 1983. The development ofDiaporthe phaseolarum variety sojae in culture. Mycologia 75: 1074-1091.

Luttrell, E. S. 1951. Taxonomy of the pyrenomycetes.University o f Missouri Study 3: 1-120.

Mhasker, D. N. and V. G. Rao. 1976. Development of the ascocarp inEpichloe cinerea (Clavicipitaceae). Mycologia 68: 994-1001.

Rogerson, C. T. 1970. The hypocrealean fungi (Ascomycetes, Hypocreales). Mycologia 62:865-910.

Samuels, G. J. 1973. Perithecial development in Hypomyces aurantius. Am. J. Bot. 60: 268-276.

Uecker, F. A. 1976. Development and cytology ofSordaria humana. Mycologia. 68: 30-46.

Wyk, P. W. J. van, M. J. Wingfield and P. S. van Wyk. 1991. Ascospore development on Ceratocystis moniliformis. Mycol. Res. 95:96-103.

and M. J. Wingfield. 1991a. Ultrastructure of ascosporogenesis in Ophiostoma davidsonii. Mycol. Res. 95: 725-730.

a n d . 1991b. Ascospore ultrastructure and development in Ophiostoma cucullatum. Mycologia 83: 698-707. VITA

Joseph William Spatafora was born in Monroe, Louisiana on the

nineteenth of May, nineteen-hundred and sixty four. He is the sixth of seven

children born to Julius Eugene and Billie Marie Spatafora. Joseph attended

Jesus the Good Shepherd Elementary School and St. Frederick's High School in

Monroe. Afterwards, he attended Louisiana Tech University in Ruston,

Louisiana, where he earned a Bachelor of Science in Zoology in 1986. Joseph

began his graduate career in the Biochemistry Department at Louisiana State

University in the fall of 1986. In the summer of 1989 he changed programs to the

Department of Botany in order to pursue research interests in evolutionary

biology. Joseph is married to Elizabeth Rossi Spatafora, and they have one child,

Anna Aucoin Spatafora.

123 DOCTORAL EXAMINATION AND DISSERTATION REPORT

Candidate: Spatafora, Joseph William

M ajor Field: Botany

T itle of D issertation: The Molecular Systematics of Unitunicate, Perithecial Ascomycetes

A p p r o v e d :

M ajor Professor and Chairman

Dean of the Gradh^te School

EXAMINING COMMITTEE:

/

lYVAaw

)

Date of Examination:

11/ 2/92