Quick viewing(Text Mode)

Systematics and Phylogeny of Cordyceps and the Clavicipitaceae with Emphasis on the Evolution of Host Affiliation

Systematics and Phylogeny of Cordyceps and the Clavicipitaceae with Emphasis on the Evolution of Host Affiliation

AN ABSTRACT OF THE DISSERTATION OF

Gi-Ho Sung for the degree of Doctor of Philosophy in Botany and Pathology presented on December 1. 2005.

Title: Systematics and Phylogeny of and the with Emphasis on the Evolution of Host Affiliation

Abstract approved: Redacted for Privacy

Cordyceps is a of perithecial ascomycetes. It includes over 400 that are pathogens of and parasites of the truffle genus . Based on the morphology of cylindrical asci, thickened ascusapices and fihiform , it is classified in the Clavicipitaceae (), which also includes and epiphytes of the grass (e.g., Claviceps and Epichloe). Multi-gene phylogenetic analyses rejected the monophyly of the Clavicipitaceae.

Clavicipitaceous fungi formed three strongly supported clades of which one was monophyletic with the . Furthermore, the monophyly of Cordyceps was also rejected as all three clades of Clavicipitaceae included species of the genus.

These results did not support the utility of most of the characters (e.g., morphology and host affiliation) used in current classifications and warranted significant taxonomic revisions of both Cordyceps and the Clavicipitaceae. Therefore, taxonomic revisions were made to more accurately reflect phylogenetic relationships. One new family was proposed and two families

(Clavicipitaceae and ) were emended for the three clavicipitaceous clades. Species of Cordyceps were reclassified into Cordyceps sensu stricto,

Elaphocordyceps gen. nov., gen. nov., and and a total of 147 new combinations were proposed. In teleomorph-anamorph connection, the phylogeny of the Clavicipitaceae s. 1. was also useful in characterizing the polyphyly of sect. Prostrata and in integrating the following major anamorphic genera with their teleomorphic genera in a new classification system(teleomorphic genera: anamorphic genera): Cordyceps: ,, ; :

Aschersonia; Metacordyceps: , ; Ophiocordyceps: ,

Hymenostilbe. In addition to multi-gene phylogenetic analyses, phylogenetic dating and ancestral host state reconstruction were conducted to understand the evolution of host affiliation in Clavicipitaceae s. 1. and the Hypocreales. These analyses suggested that fungal- symbioses of the Clavicipitaceae s. 1. initiated in the Jurassic by a series of inter- host jumps from plant-based nutrition to animal pathogens. A total of 9-14 inter-kingdom host jumps were hypothesized for the Clavicipitaceae s. 1., including 5-7 onto fungi, 2-4 onto , and 2-3 onto , with dominant ecology as pathogens of arthropods. The total evidence of the Cretaceous radiation of the Clavicipitaceae s. 1. indicated that the current diversity of its fungal- symbioses may be correlated with the coevolutionary interactions between arthropods and angiosperms in the Cretaceous. ©Copyright by Gi-Ho Sung December 1, 2005 All Rights Reserved Systematics and Phylogeny ofCordycepsand the Clavicipitaceae with Emphasis on the Evolution of Host Affiliation

by

Gi-Ho Sung

A DISSERTATION

submitted to

Oregon State University

in partial fulfillment of the requirement for the degree of

Doctor of Philosophy

Presented December 1, 2005

Commencement June 2006 Doctor of Philosophy dissertation of Gi-Ho Sung presented on December 1. 2005.

APPROVED:

Redacted for Privacy Major Professor, representing Botany and

Redacted for Privacy

Chair of the Depaltment ofl3otany and Plant Pathology

Redacted for Privacy Dean of th4Iiraduate School

I understand that my dissertation will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my dissertation to any reader upon request.

Redacted for privacy

Gi-Ho Sung, Author CONTRIBUTION OF AUTHORS

Dr. Joseph W. Spatafora was involved in the initiation, design, phylogenetic analyses, and the writing and editing of each manuscript. Drs. Rasoul Zare, Kathie T.

Hodge, and Walter Gams were involved in the design and editing of chapter two. Drs.

Nigel L. Hywel-Jones and Jae-Mo Sung were not only taxonomic consultants but also contributed to chapter five in providing valuable cultures and specimens. Dr. George

0. Poinar, Jr. was involved in the design and intellectual development of chapter six. TABLE OF CONTENTS

Chapter 1: INTRODUCTION ...... 1 and classification ...... 3

Dissertation scope...... 6 References ...... 8

Chapter 2: A REVISION OFVERTICILLIUMSECT.PROSTRATA. II. PHYLOGENETIC ANALYSIS OF SSU AND LSU NUCLEAR rDNA SEQUENCES FROM ANAMORPHS AND TELEOMORPHS OF THE CLAVICIPITACEAE...... 13 Abstract ...... 14

Introduction...... 16

Materials and methods...... 19

Taxon sampling...... 19

DNA extraction...... 20

Phylogenetic analyses...... 24

Results...... 26

Discussion...... 29

Familial relationships of Verticillium sect. Prostrata ...... 29

Polyphyly of Verticillium sect. Prostrata...... 32

Relationships of V. sect. Prostrata with anamorphs and teleomorphs

of the Clavicipitaceae...... 34

Evolution of host association...... 37 TABLE OF CONTENTS (Continued)

Conclusions ...... 39

Acknowlegements...... 40

References...... 40

Chapter 3:CORD YCEPS CARDINALIS SP. NOV.,A NEW SPECIES OF CORD YCEPSWITH AN EAST ASIAN-EASTERN NORTH AMERICAN DISTRIBUTION...... 45

Abstract...... 46

Introduction...... 47

Materials and methods...... 49

Field collection...... 49

Morphological observations...... 50

Phylo genetic analyses...... 51

Taxonomy...... 53

Results...... 57

DNAsequencing...... 57

Phylogenetic analyses...... 58

Discussion...... 60

Morphological comparison of C.cardinaliswith other

C. militaris-like species...... 60

Anamorphic and cultural differences among C.cardinalis,C.

militaris and C. pseudomilitaris...... 62 TABLE OF CONTENTS (Continued)

Phylogenetics of C. cardinalis ...... 64

Biogeography of C. cardinalis ...... 66 Conclusions ...... 67 Acknowledgements ...... 68 References ...... 68

Chapter 4: A MOLECULAR PHYLOGENETIC STUDY OF THE CLAVICIPITACEAE BASED ON MUTIPLE INDEPENDENT LOCI. . ..71 Abstract ...... 72

Introduction...... 74

Materialsand methods...... 77

Selection of taxa...... 78

DNA isolation, PCR amplification, Cloning and Sequencing...... 78

Phylogenetic analyses...... 85

Results...... 91

Characteristics of each gene...... 91

Separate analyses and test of incongruence...... 95

The Phylogenetic analyses of the combined data set...... 96

Combinational analyses...... 103

Contribution of each gene partition to nodal support...... 105

Discussion...... 109

ILD test and the heterogeneity among gene partitions...... 109 TABLE OF CONTENTS (Continued)

Comparison of phylogenetic methods in combined analyses.. . .111

Incongruence, sampling and bootstrap support ...... 113

Congruence vs. combinability ...... 115

Phylogenetic relationships of the Clavicipitaceae ...... 117

Conclusions...... 121

References...... 122

Chapter 5: PHYLOGENETIC CLASSIFICATION OF CORD YCEPS AND THE CLAVICIPITACEOUS FUNGI ...... 130

Abstract...... 131

Introduction...... 133

Materials and methods...... 136

Taxon and character sampling...... 136

Sequence alignment and phylogenetic analyses...... 137

Results...... 147

Sequence alignment...... 147

Test of incongruence and phylogenetic analyses...... 148

Phylogenetic relationships of the clavicipitaceous fungi...... 150

Discussion...... 155

Phylogenetic implications on the systematics

of the genus Cordyceps...... 155

Cordyceps species in Clavicipitaceae dade A...... 157 TABLE OF CONTENTS (Continued)

Cordyceps species in Clavicipitaceae dade B ...... 158

Cordyceps species in Clavicipitaceae dade C ...... 164

The taxonomic utility of anamorphic forms in

classification of Cordyceps ...... 170

Clavicipitaceae dade A ...... 170

Clavicipitaceae dade B ...... 173

Clavicipitaceae dade C...... 177

The taxonomic revisions of Cordyceps and the Clavicipitaceae. . .180

Accepted names and new combinations...... 191

References...... 207

Chapter 6: THE FUNGAL-ARTHROPOD SYMBIOSES OF THE HYPOCREALES (FUNGI) REPRESENT A CRETACEOUS RADIATION EVENT...... 216

Abstract...... 217

Introduction...... 219

Materials and methods...... 222

Fossil records for calibration points...... 222

Taxon sampling, construction of data sets

for phylogenetic dating...... 223

Taxon sampling, construction of data sets for ancestral state

reconstruction of host affiliation...... 228 TABLE OF CONTENTS (Continued)

Sequence alignment and phylogenetic analyses ...... 229

Divergence time estimation ...... 231

Ancestral state reconstruction of host affiliation ...... 233

Results and discussion ...... 233

Phylogenetic analyses and the relationships of the hypocrealean fungi ...... 234

Phylogenetic dating and evolution of host affiliation in the

hypocrealean fungi...... 237

Internal consistency of age estimates between Hypocreales and

Sordariomycetes data sets ...... 253 Conclusions ...... 259 References ...... 259

Chapter 7: CONCLUSIONS ...... 268 References ...... 272 BIBLIOGRAPHY ...... 274 LIST OF FIGURES

tnct consensus tree of most parsimonious trees based on SSU and LSU

)NA dataset...... 28

ne of the most parsimonious trees based on SSU and LSU rDNA

ataset...... 31

4orphology and ecology of C. cardinalis...... 55

me drawing of variation in anamorph of C cardinalis...... 56

3.3 One of the most parsimonious trees based on SSU and LSU rDNA

Dataset...... 59

4.1 Topological incongruence of the phylogenies among seven genes...... 94

4.2 One of the most parsimonious trees of the Clavicipitaceae based on seven genes and 91 taxa ...... 97

4.3 50% majority consensus tree of the Clavicipitaceae based on seven genes and91 taxa...... 99

4.4 50% majority consensus tree of the Clavicipitaceae based on six genes and91 taxa...... 100

4.5 The effect of character and taxon sampling in resolving ten major

relationships in the Clavicipitaceae...... 105

4.6 Schematic diagrams of the contribution of each gene to total support...... 107

5.1 One of the most parsimonious trees of the Clavicipitaceae based on

five genes and 162 taxa...... 149 LIST OF FIGURES (Continued)

Figure

5.2 50% majority consensus tree of the Clavicipitaceae based on five genes

and162 taxa...... 151

5.3 Schematic diagrams of the phylogenies to compare the resolution of

the Clavicipitaceae...... 154

5.4 Phylogeny of the Clavicipitaceae dade A and macromorphology of its

representative species...... 159

5.5 Phylogeny of the Clavicipitaceae dade B and macromorphology of its

representative species...... 161

5.6 Phylogeny of the Clavicipitaceae dade C and macromorphology of its

representative species...... 166

5.7 A new classification ofCordycepsand the clavicipitaceous fungi...... 181

6.1 Phylogeny of the Hypocreales based on five genes and 74 taxa to infer

divergence times...... 236

6.2 Chronogram of the Hypocreales...... 239

6.3 Results of ancestral host state reconstruction based on the phylogeny

with five genes and 147 taxa...... 242

6.4 Phylogeny of the based on five genes and 60 taxa to infer

divergence times...... 254

6.5 Schematic diagrams of chronograms of the Sordariomycetes...... 256 LIST OF TABLES

Figure

2.1 List of cultures and specimens...... 21

3.1 Morphological comparision of C.cardinalisand C. militaris-like species. .61

4.1 List of cultures and specimens...... 79

4.2 List of primes...... 84

4.3 Summary of data used in separate and combined analyses...... 92

4.4 Nucleotide base frequencies and results of chi-square tests...... 93

4.5 Results of incongruent length difference (ILD) tests...... 96

4.6 Nodal supports for ten major relationships in the Hypocreales...... 101

5.1 List of cultures and specimens...... 138

6.1 List of cultures and specimens...... 224

6.2 Results of estimating the divergence times in the Hypocreales...... 240

6.3 Results of estimating the divergence times in the Sordariomycetes...... 258 SYSTEMATICS AND PHYLOGENY OF CORD YCEPS AND THE CLAVICIPITACEAE WITH EMPHASIS ON THE EVOLUTION OF HOST AFFILIATION

Chapter 1

INTRODUCTION

The Clavicipitaceae (: Hypocreales) is a family of perithecial ascomycetes that is unique in comprising entirely obligate symbionts of other organisms including pathogens of arthropods, parasites of truffles, and endophytes and epiphytes of the grass family (Diehi, 1950; Rogerson, 1970). Species of the family have had considerable impact on human civilization and modern agriculture (White et al., 2003). Ergotism (causal agent: ) has caused human toxicoses and has significantly influenced human civilization for centuries (White et al., 2003).

Infected grasses with the clavicipitaceous endophytes (e.g., genera Epichloe and

Neotyphodium) cause toxic syndromes in grazing animals and grass endophytes have applications to improve commercial grass cultivars for drought tolerance, growth rate and resistance to herbivory (Clay, 1989; Clay and Holah, 1999; Clay and Schard!,

2002). Species of the genus Cordyceps produce biologically active compounds that are used in research and medicine and numerous asexual species of the genus (e.g.,

Beauveria bassiana and ) have been used as biological control agents of pests (Evans, 2003; Samson et al., 1988). 2

Cordyceps is the largest and most diverse genus in the Clavicipitaceae with respect to numbers of species, morphological variation, and the breadth of host range.

It is classified in the family by its possessing cylindrical asci, thickened apices, and fihiform ascospores that often disarticulate into partspores (Diehi, 1950; Rossman et al., 1999). More than 400 species of the genus have been reported and the majority of them produce superficial to completely immersed perithecia on stipitate stromata that emerge from arthropods (Index Fungorum; Kobayasi, 1941). Morphology of the stromata varies according to species ranging from clavate, to fihiform, to capitate, but species of the genus are essentially astromatic and produce perithecia on loosely organized hyphae or subiculum on the external cuticle of arthropods (Kobayasi, 1941).

Approximately 20 species of the genus are known as the parasites of the false-truffles genus Elaphomyces with the remaining species of Cordyceps and its close relatives

(e.g., Hypocrella and Torrubiella) parasitizing and other arthropods from eight orders of Arthropoda, including Arachnida, Coleoptera, , Hymenoptera,

Isoptera, and (Kobayasi, 1941; Mains, 1958). Host specificity varies according to species. A few species of Cordyceps are capable of infecting a broad host range of species in multiple families or orders of arthropods; however, the majority of species are restricted to a single species or a few closely related species. In addition, most species of Cordyceps only infect, or at least fruit from, a particular stage of the host life cycle (e.g., larva, pupa, etc.).

As seen in other pathogens of plants and vertebrates, pathogens of arthropods in the Clavicipitaceae infect their host through the external cuticle (Samson et al., 1988). The insect integument is penetrated by a germ tube of spores with enzymatic reactions of proteases, lipases, and chitinases (St. Leger, 1991; Samson et al., 1988).

Once the pathogens successfully reach the hemoceol of the hosts, they circulate the body cavity of the hosts in a -like phase by producing blastospores. After the completely consumes the internal components of the hosts, it forms an endosclerotium from which a stroma or stromata are produced, rupturing through the cuticle and releasing spores into the environment.

Taxonomy and classification

The nomenclatura! history of Cordyceps began in the18thcentury. Clavaria militaris () was first described by Linnaeus (1753) and later recominbined in Sphaeria by Persoon (1801). In 1823, Sphaeria was segregated by

Fries into a number of tribes for species that are classified in many current genera of ascomycetous fungi. In the Friesian classification of Sphaeria, one of the tribes, which was designated Cordyceps for perithecial ascomycetous species that produce fleshy and stipitate stromata, was later elevated to genus by Link (1933). The current concept of Cordyceps was originated by Tulasne (1853) who included pathogens of arthropods and parasites of Elaphomyces in Cordyceps, although he proposed the invalid name

Torrubia (a homonym of Torrubia Vell., 1825).

Several subgeneric classifications of Cordyceps were proposed in the19thand early20th1centuries, largely based on macroscopic characters (e.g., presentation of 4 perithecia and morphology of stromata) with considerable emphasis on host affiliation.

The genus was divided by Tulasne (1865) into two groups to separate the entomogenous species as Entomogenae and the mycogenous species in Mycogenae.

Both of these groups were again segregated into Clavatae and Capitatae based on whether fertile parts of stromata are fihiform or capitate (Tulasne, 1865). Saccardo

(1883) also emphasized host affiliation and recognized Cordylia for the mycogenous species. However, he divided the entomogenous species into Eucordyceps (peritheciis stromate subimmersis) and Racemella (peritheciis stromate subsuperficialibus) based on whether the perithecia are superficial or immersed. Later, Lindau (1897) and

Schroeter (1908) regrouped the entomogenous species with different emphasis on the macroscopic characters in their subgeneric classification.

Although Lindau (1897) recognized that Cordyceps is a member of the family

Clavicipitaceae based on its unique microscopic characters (e.g., cylindrical asci, thickened ascus apices, and fihiform ascospores), it was not until Petch (1931) that microscopic characters were used in the classification of Cordyceps. He erected the genus Ophiocordyceps for the species of Cordyceps that possess nondisarticulating ascospores and a less pronounced thickening of the ascus apex (Petch, 1931).

Subsequently, the most comprehensive taxonomic study of Cordyceps was undertaken by Kobayasi (1941), who emphasized the morphologies of ascospores and partspores and orientation of perithecia in his subgeneric classification. In this work, he classified

124 species of Cordyceps into three subgenera (i.e., C. subg. Cordyceps, C. subg.

Ophiocordyceps (Ophiocordyceps Petch) and C. subg. Neocordyceps), which were further divided into 19 subtaxa based on macroscopic characters used in previous classifications (e.g., presentation of perithecia and the shape of the fertile parts in stromata). Later, Mains (1958) added two additional subgenera to the classification of

Kobayasi and greatly advanced our knowledge of North American taxa. Currently, four subgenera are recognized including three subgenera of Kobyasi and C. subg.

Bolacordycepsthat possess bola-ascospores (Eriksson and Hawksworth, 1986).

The contribution of Kobayasi to the taxonomy ofCordycepscannot be overstated. He described over 100 species ofCordycepsand his monographic work in

1941 provided the essential treatment for the current classification ofCordyceps

(Kobayasi, 1941; Kobayasi, 1982). Although he provided the most inclusive taxonomic key for 282 species ofCordyceps(Kobayasi, 1982), its taxonomic utility is considered limited. The use of Kobayasi's key for identification has led to confusion in species identification due to its lack of the information of microscopic characters.

Kobayasi (1982) recognized that these characters are missing from descriptions of most of the species published before 1900. The taxonomy ofCordycepsis further complicated by missing or undesignated type specimens in both early European and

Kobayasi's collections. Unfortunately, most of Kobayasi's collection are considered lost by many mycologists and the type materials need to be reestablished.

In systematics of Ascomycota and , asexual states (termed anamorphs) are allowed to be named under Article 59 of the International Code of

Botanical Nomenclature (Greuter et al., 2000), thus providing a binomial nomenclature system for naming both the sexual (termed teleomorphs) and anamorphs. Anamorphs are classified based on the morphology of structures associated with spore production (e.g., conidiophore and phialide) and more than 25 anamorphic genera (e.g., Beauveria, Metarhizium, Hirsutella, ,

Paecilomyces sect. Isarioidea and Verticillium sect. Prostrata) are currently linked or possibly related to Cordyceps (Hodge, 2003). The anamorphic species of Cordyceps are that resemble asexual species of other hypocrealean fungi. The conidiogenesis is typically phialidic or sympodial and produces ameroconidia in dry chains or slimy drops on pale and brightly colored conidiogenous structures. Since

Persoon (1932) erected the anamorph genus Isaria, numerous asexual genera (e.g.,

Beauveria, , and sect. Prostrata) have been segregated from

Isaria and Spicaria by significant nomenclatural changes. The connection to

Cordyceps was first established by Tulasne (1857), although he mistakenly demonstrated that I. farinosa was the anamorph of C. militaris. The anamorph of C. militaris is now known to be a Lecanicillium anamorph, but teleomorph-anamorph connections are poorly understood for most anamorphic species that are linked or hypothesized to be closely related with Cordyceps.

Dissertation scope

With the advent of molecular systematics, Clavicipitaceae has been confidently demonstrated to be a member of the Hypocreales (Castlebury et al., 2004; Rehner and

Samuels, 1995; Spatafora and Blackwell, 1993). The monophyly of the family, 7 however, has not been convincingly addressed. Previous studies based on limited sampling of both taxa and characters resulted in both paraphyletic and monophyletic hypotheses of the relationships for the family with varying levels of support

(Artjariyasripong et al., 2001; Castlebury etal., 2004; Gams etal., 1998; Glenn etal.,

1996; Nikoh and Fukatsu, 2000; Rehner and Samuels, 1995; Rossman et al., 2001;

Spatafora and B!ackwe!l, 1993; Stensrud et al., 2005; Sung et al., 2001). A!though these studies have suggested that Cordyceps is polyphy!etic and in need of taxonomic revision, it is difficult to make taxonomic revisions without robust phylogenetic hypotheses. In this dissertation, I examine the phylogenetic relationships of Cordyceps and its allies in the C!avicipitaceae. Numerous phylogenetic analyses based on broad taxon and character sampling are performed with the result being the recognition of monophyletic taxa of Cordyceps and the Clavicipitaceae. Moreover, these analyses infer the relationships between Cordyceps and its putative anamorphic forms (e.g.,

Beauveria, Hirsutella, and Verticillium sect. Prostrata) and provide robust evolutionary hypotheses of host affiliation in Cordyceps and the Clavicipitaceae.

Chapter 2 presents a molecular phylogeny of V. sect. Prostrata based on nuclear large and small ribosomal DNA. V. sect. Prostrata is a large and diverse anamorphic group of fungi that presumably belongs to the Clavicipitaceae (Gams,

1971; Hodge, 2003). The molecular phylogeny in this chapter clarifies the phylogenetic relationships of the species in V. sect. Prostrata with other anamorphic and teleomorphic taxa (e.g., genera Cordyceps and Torrubiella) in the Clavicipitaceae. el for incorporating other anamorphic taxa into the phylogeny and ultimately classification of the Clavicipitaceae.

Chapter 3 describes a new species, C. cardinalis, with an East AsianEastern

North American distribution. This chapter provides a description of C. cardinalis and uses a molecular phylogeny based on nuclear large and small ribosomal DNA to compare it with macroscopically similar species (e.g., C. militaris and C. pseudomilitaris). Overall aspects of its morphology, anamorph, and host association were compared to address the plasticity of stromatal morphology in lower levels of the classification of Cordyceps. This chapter provides a model by which phylogenetic species concepts can be applied to species complexes of Cordyceps whereby molecular data are used to refine our understanding of morphological homologies.

Chapter 4 addresses the higher-level phylogeny of the Clavicipitaceae. Data sets were constructed based on the sequences of seven genes (nrSSU, nrLSU, /3- tubulin, EF-la, RPBJ, RPB2 and mtATP6) for a broad and inclusive sampling of taxa.

Separate analyses of each gene partition were conducted to identify localized incongruence among gene partitions. The distribution of support or conflict of each gene partition was also generated to verify the results from the separate analyses. The phylogeny from the combined analyses was compared to the current subfamilial classification of the Clavicipitaceae.

Chapter 5 presents a molecular phylogeny of Cordyceps and the clavicipitaceous fungi based on the sequences of five genes (nrSSU, nrLSU, EF-Ja,

RPBJ and RPB2). The combined analyses were performed with extensive taxon sampling for Cordyceps and the Clavicipitaceae. The phylogeny was compared to the current subgeneric classification of Cordyceps and used to evaluate the utility of diagnostic characters as currently defined. As a result, a new phylogenetic classification of Cordyceps and the clavicipitaceous fungi is proposed.

Chapter 6 discusses the evolution host affiliation and phylogenetic dating of the Clavicipitaceae and other families of the Hypocreales. Because the origin of the hypocrealean fungi can not predate the divergences of three kingdoms that serve as hosts or substrates (Animalae, Fungi, and Plantae), their evolutionary history must be explained by some level of inter-kingdom host shifts (Nikoh and Fukatsu, 2000). To develop hypotheses for the evolutionary history of the host affiliation, ancestral states of host affiliations were reconstructed for the major lineages of the hypocrealean fungi based on multi-gene phylogenies. These analyses highlight dynamic process of inter- kingdom host shifts that characterize the evolution of the Hypocreales. Phylogenetic dating was also conducted to develop the geologic age estimates for the major lineages of the hypocrealean fungi. The phylogenetic dating, coupled with the ancestral host state reconstruction, results in the evolutionary hypothesis that correlates the origin and diversification of the major lineages of the Hypocreales with diversification of angiosperms and insects.

References

Artjariyasripong, S., J. I. Mitchell, N. L. Hywel-Jones, and E. B. Garath Jones. 2001. Relationships of the genus Cordyceps and related genera, based on parsimony 10

and spectral analysis of partial I 8S and 28 S ribosomal gene sequences. Mycoscience 42:503-517.

Castlebury, L. A., A. Y. Rossman, G.-H. Sung, A. S. Hyten, and J. W. Spatafora. 2004. Multigene phylogeny reveals new lineage for Stachybotrys chartarum, the indoor air fungus. Mycological Research 108:864-872.

Clay, K. 1989. Clavicipitaceous endophytes of grasses: their potential as biocontrol agents. Mycological Research 92:1-12.

Clay, K., and J. Holah. 1999. Fungal symbiosis and plant diversity in successional fields. Science 285:1742-1744.

Clay, K., and C. L. Schardl. 2002. Evolutionary origins and ecologicalconsequences of endophyte symbiosis with grasses. The American Naturalist 160:S99-S127.

Diehl, W. W. 1950. and the in America. Agricultural monograph No.4. U. S. D. A. Washington. 82pp.

Eriksson, 0. B., and D. L. Hawksworth. 1986. Notes on ascomycete systematics. Systema Ascomycetum 5:114-174.

Evans, H. C. 2003. Use of clavicipitalean fungi for the biological control of arthropod pests. Pages 517-548 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Gams, W. 1971. Cephalosporium-artige Schmmelpilze (Hyphomycetes). G. Fischer, Stuttgart.

Gams, W., K. O'Donnell, H.-J. Schroers, and M. Christensen. 1998. Generic classification of some more hyphomycetes with solitary conidia borneon phialides. Canadian Journal of Botany 76:1570-1583.

Glenn, A. E., C. W. Bacon, R. Price, and R. T. Hanlin. 1996. Molecular phylogenyof and its taxonomic implications. Mycologia 88:369-383.

Greuter, W., J. McNeill, F. R. Barrie, H.-M. Burdt, V. Demoulin, T. S. Filgueiras,D. H. Nicolson, P. C. Silva, J. E. Skog, P. Trehane, N. J. Turland, and D.L. Hawksworth. 2000. Internationsi Code of Botanical Nomenclature (St Louis Code). Konigstein: Koeltz Scientific Books.

Hodge, K. T. 2003. Clavicipitaceous anamorphs. Pages 75-123 inClavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts(J. F. 11

White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku 84:53-260.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of 23:329-364.

Lindau, G. 1897. Pyrenomycetineae. Engler & Prantl Pflanzenfamilien. 1 Abt. 1:321 - 505.

Linnaeus, C. 1753. Species Plantarum. p. 1182.

Link, K. 1833. HandbUch zur Erkennung der Nutzbarsten und am Haufigsten Vorkommenden Gewächse 3: 347.

Mains, E. R. 1958. North american entomogenous species of Cordyceps. Mycologia 50:169-222.

Nikoh, N., and T. Fukatsu. 2000. Interkingdom host jumping underground: Phylogenetic analysis of entomopathogenic fungi of the genus Cordyceps. Molecular Biology and Evolution 17:629-638.

Petch, T. 1931. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 16:55-75.

Rehner, S. A., and G. J. Samuels. 1995. Molecular systematics of the Hypocreales: a teleomorph gene phylogeny and the status of their anamorphs. Canadian Journal of Botany 73 (Suppi. 1):S816-S823.

Persoon, D. C. H. 1801. Syn. Meth. Fung. p. 1.

Persoon, D. C. H. 1832. Syst. Mycol. 3:270.

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

Rossman, A. Y., J. M. MaKemy, R. A. Pardo-Schultheiss, and H. J. Schroers. 2001. Molecular studies of the using large subunit rDNA sequences. Mycologia 93:100-110.

Rossman, A. Y., G. J. Samuels, J. S. Rogers, and R. Lowen. 1999. Genera of Bionectriaceae, and Hypocreaceae (Hypocreales, Ascomycetes). Studies in 42:1-238. 12

Samson, R. A., H. C. Evans, and J.-P. Latge. 1988. Atlas of Entomopathogenic Fungi. Springer-Verlag, Berlin Heidelberg New York.

Saccardo, P. A. 1883. Cordyceps. Syll. Fung. 2:566-578.

Schroeter, J. 1908. Die Pilze Schiesiens. Cohn's Kryptogamen-Flora von Schlesien 3: 1-597.

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes: the Clavicipitales-Hypocreales connection. Mycologia 85:912-922.

Stensrud, 0, N. L. Hywel-Jones, and T. Schumacher. 2005. Towards a phylogenetic classification of Cordyceps: ITS nrDNA sequence data confirm divergent lineages and paraphyly. Mycological Research 109:41-56.

St. Leger, R. J., D. W. Roberts and R. C. Staples. 1991. A model to explain differentiation of appressoria by germlings of Metarhizium anisopliae. Journal of Invertebrate Pathology 57:299-310.

Sung, G.-H., J. W. Spatafora, R. Zare, K. T. Hodge, and W. Gams. 2001. A revision of Verticillium sect. Prostrata. II. Phylogenetic analyses of SSU and LSU nuclear rDNA sequences from anamorphs and teleomorphs of the Clavicipitaceae. Nova Hedwigia 72:311-328.

Tulasne, L. R. 1853. Aim. Sci. Nat. III. 20:43.

Tulasne, L. R. 1857. Les Isaria et Sphaeria entomogènes. Ann. Sci. Nat. Bot. Sér 4. 8:35-43.

Tulasne, L. R. 1865. Selecta Fungorum Carpologia 3.p. 221.

White, J. F., Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora. 2003. Historical perspectives: Human interactions with Clavicipitalean fungi. Pages 1-16 in Clavicipitalean Fungi (J. F. White, Jr., M. Backlund, N. L. Hywel- Jones, and J. W. Spatafora, eds.). Marcel Decker, New York. 13

Chapter 2

A REVISION OFVERTICILLIUMSECT.PROSTRATA. II.PHYLOGENETIC ANALYSIS OF SSU AND LSU NUCLEAR rDNA SEQUENCES FROM ANAMORPHS AND TELEOMORPHS OF THE CLAVICIPITACEAE

Gi-Ho Sung', Joseph W. Spatafora',

Rasoul Zare2, Kathie T. Hodge3, and WalterGams4

'Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR

9733 1-2902, USA

2CABI Bioscience, Bakeham Lane, Egham, Surrey TW2O 9TY, UK. Present address:

Department of Botany, Plant Pests and Diseases Research histitute, P. 0. Box 1454,

Tehran 19395, Iran

Departments of Plant Biology and Plant Pathology, Cornell University, Ithaca, N.Y.

14853-290 1, USA

Centraalbureau voor Schimmelcultures, P.O. Box 85167, 3508 AD Utrecht, The

Netherlands

Nova Hedwigia72: 311-328 14

Abstract

Parsimony analyses were conducted on partial nucleotide sequences from the

small and large subunits of the nuclear nbosomal DNA from representatives of

Verticillium sect. Prostrata and related ascomycetes. The majority of species from V.

sect. Prostrata were supported as members of the Clavicipitaceae, but they did not

form a monophyletic group within the family. Three to five groups of fungi in V. sect.

Prostrata were inferred in these analyses and were designated groups B-D following

the convention of Zare et al. (2000). These groups integrated with other anamorph and

teleomorph genera including Cordyceps, which was also not supported as being

monophyletic. Group B 1 included the anamorph of C. militaris, V. lecanii, V. psalliotae, V. fusisporum, V. araneareum, and V. antillanum. It was part ofa larger

dade designated Cordyceps s.s., which included entomopathogenic species of

Cordyceps and anamorphic species of Beauveria, Engyodontium, Microhilum, and

Paecilomyces. Group B2 included V. lamellicola, Cephalosporium lanosoniveum, and

Acremonium obclavatum and was the most closely related dade to Cordycepss.s.

Group C represented a monophyletic dade of nematophagous species that included V. balanoides, V. campanulatum, and V. sinense. It was part of a weakly supported dade designated the C. ophioglossoides dade, which included fungicolous and entomopathogenic species of Cordyceps and anamorphic species of Hirsutella,

Harposporium, and Paecilomyces. Within the C. ophioglossoides dade, V. sect.

Prostrata group C was well supported as closely related to C. gunnii,a parasite of 15 lepidopteran larvae. Group D was not monophyletic and consisted of three lineages including the rust parasite V. epiphytum (Dl), the mainly nematophagous species V. chiamydosporium, V. suchiasporium, V. cf. bactrosporum, and V. gonioides (D2), and the homopteran pathogen V. pseudohemipterigenum (D3). These data did not confidently address the relationships of the three lineages of group D to one another or to other groups within the Clavicipitaceae. 16

Introduction

Verticillium Nees is a large and heterogeneous anamorph genus that has been linked to several families of the Ascomycota including Clavicipitaceae, Hypocreaceae,

Nectriaceae, and Phyllachoraceae (Gams and van Zaayen, 1982; Gams, 1988;

Samuels, 1988; Messner et al., 1996; Zare et al., 2000). The genus is characterized by phialidic conidiophores produced directly from the mycelium as either erect or prostrate structures. The verticillate phialides are generally aculeate and are typically produced in whorls of three to five. Conidia are hyaline and range in morphology from cylindrical to ellipsoidal to falcate according to species. In addition, some species produce varying forms of chiamydospores that are thought to serve as resistant propagules. Classification of Verticillium currently consists of four different sections and a residual group based on morphological and cultural characters (Gams andvan

Zaayen, 1982). This classification is recognized as polyphyletic and amore natural classification is necessary. Current attempts to produce such a classification, however, are hampered by the fact that several of the important morphological traits can vary with culture conditions and age of isolate (Gams, 1971; Gams andvan Zaayen, 1982;

Gams, 1988).

Verticillium sect. Prostrata W. Gams was introduced by Gams (1971) for species producing prostrate conidiophores that are often poorly differentiated froma fine, white or yellowish mycelium. Phialides eitherpossess or lack inflated bases according to species. A number of species produce dictyochiamydospores, but 17 production varies amongst and within species and across a range of culture conditions.

Conidia often aggregate on the tips of phialides in heads, but in a few species they adhere in chains. In addition to variation in morphological traits, delineation of the section is complicated by the fact that some species are known to produce differentiated erect conidiophores (e.g., Verticillium suchiasporium). Recognition of monophyletic lineages within V. sect. Prostrata is also complicated by the fact that several other anamorph genera, including , Engyodontium,

Aphanocladium, and Acremonium, are morphologically similar to V. sect. Prostrata and are hypothesized to be closely related, but the exact nature of their relationships remains disputed (reviewed in Zare et al., 2000).

Several species of V. sect. Prostrata have been linked to entomopathogenic

(e.g., Torrubiella confragosa Mains) and fungicolous (e.g., Cordyceps ophioglossoides (Fr.) Link) teleomorphs, of the Clavicipitaceae, including the anamorph of C. militaris (L.Fr.) Link. which is the type of V. sect. Prostrata. The majority of species within the section, however, are only known from anamorphs and their relationship to teleomorphs of the Clavicipitaceae is speculative. Molecular phylogenetics holds great promise in developing more robust phylogenetic hypotheses through the integration of anamorphic and teleomorphic fungi in systematic studies

(Taylor, 1993; 1995). Broad taxon sampling of representatives with diverse life histories will not only result in more accurate hypotheses of relationships (Blackwell,

1993), but they will also result in a better understanding of the evolution of morphologies and ecologies, such as host affiliation and nutritional mode. Verticillium has a broad host range with mainly plant-pathogenic, fungicolous,

entomogenous, and nematophagous groups (van Zaayen and Gams, 1982). V. sect.

Prostrata includes the majority of entomogenous and nematophagous species and

some fungicolous taxa within the genus and overlaps in part of its host range with

teleomorphs of the Clavicipitaceae. Hosts of the entomogenous (s.l.) species include

species of Arachnida, Coleoptera, Homoptera, and Lepidoptera (Gams 1971). Hosts of

the fungicolous include a variety of Ascomycota and Basidiomycota, including fleshy

fungi of and rusts of the Urediniomycetes (van Zaayen and Gams,

1982; Lim and Wan, 1983; Gams et al., 2001). Nematophagous species of V. sect.

Prostrata have a fairly broad host range, but species specialize on either adultsor

cysts (Barron 1977; Gams 1988). Many species, however, are isolated from soil samples and plant litter and the exact nature of their nutritional mode and potential host affiliation are unknown. The host range of V. sect. Prostrata is therefore quite diverse and the patterns and processes by which this host rangearose are largely unknown. Like other fungi of the Clavicipitaceae, species of V. sect. Prostrataare the subject of numerous biological control studies, including those of

(Mankau, 1980; Kerry and Crump, 1977), arthropods (Schuler et al., 1991;

Vestergaard et al., 1995), and fungal pathogens (Saksirirat and Hoppe, 1991; Verhaar et al., 1996). A more accurate understanding of their systematics and phylogenetic affinities is desirable for the design of biological control experiments and the application and introduction of such agents in nature. 19

In a phylogenetic study of ITS rDNA, Zare et al. (2000) revealed at least three

distinct groups of V. sect. Prostrata that were consistent with differences in

and phialide morphology, host affiliation, and to a lesser extent production of

dictyochlamydospores. Their data also supported a phylogenetic affinity with the

Clavicipitaceae, but few teleomorphs were included in their analyses. Zare et al.

(2000) provide the most thorough insight into the polyphyly of V. sect. Prostrata and

laid the foundation for an eventual revision of its nomenclature. As part of a

continuing series of systematic investigations into Verticillium and the

Clavicipitaceae, we initiated a molecular phylogenetic study of nuclear ribosomal

DNA (rDNA) sequences from V. sect. Prostrata. The major goals of the study were 1)

to integrate data from V. sect. Prostrata with data from the Clavicipitaceae and other

related members of the Ascomycota, 2) to test the monophyly of V. sect. Prostrata, 3) to develop hypotheses for its relationship to the major teleomorphs and anamorphs of the Clavicipitaceae, and 4) to develop hypotheses for character evolution and host association.

Materials and methods

Taxon sampling

Taxa were sampled to include a broad representation of V. sect. Prostrata, representative teleomorphs of the Clavicipitaceae (e.g., Cordyceps, Torrubiella,

Claviceps,Epichloe),and other possible anamorphs of the Clavicipitaceae (e.g., 20

Metarhizium, Beauveria, Paecilomyces). Both host affiliation and morphology were

used in taxon selection. Several sequences were included to represent a diversity of

outgroup taxa from other families and orders of perithecial ascomycetes. A total of 93

isolates, representing 80 species and varieties, were included in this study. Sources of

isolates and host affiliation of specimens used in this study are listed in Table 2.1.

DNA extraction

DNA was extracted from both dried herbarium specimens and live cultures.

Crude DNA extracts were prepared by a modified CTAB method (Gardes and Bruns

1993, Spatafora et al., 1998). Samples were ground in microcentrifuge tubes with disposable pestles attached to an electric hand drill. CTAB bufferwas then added to the ground tissue and incubated at 65 °C for 30 mm. Two chloroform: isoamylalcohol

(24:1) extractions were conducted after the incubation. DNAwas purified from the aqueous phase of the second chloroform extraction using the GeneClean III kit (Bio

101 Inc., Vista, CA).

To integrate V. sect. Prostrata into a database of clavicipitalean fungi, 1150 bp of the small subunit (SSU) and 950 bp of the large subunit (LSU) nuclear ribosomal

DNA (nrDNA) were amplified in the polymerase chain reactions (PCR) (Mullis and

Faloona, 1987). The SSU rDNA was amplified with primers NS1 and NS4 (Whiteet al., 1990). The LSU rDNA was amplified with primers LROR and LR5 (Vilgalysand

Sun, 1994). PCRs were performed in 50-pi reactionsas follows: [94 °C (1 mm), 50-52

°C (30 sec), 72 °C (1 mm)] 35-40 cycles. Success of PCRswas confirmed 21 Table 2.1. The listANAMORPHS of cultures and specimens used in this study. Taxon Specimen voucher flostlsubtratum GeneBank accession number SSU LSU AphanocladiumBeauveria brongniartiibassiana.bassiana album (Preuss)(Bals.) (Sacc.) Vuill. W. Petch. Gams. IFONRRLCBS 52994848 401.70 28020 myxomycete AB027335AF049144AB027336AF339568 AB027381AF049164AB027382AF339518 HirsutellaHarposporiumEngydontiumBeauveria thompsonii caledonica helicoides aranearumaranearum. Fisher Bissett Drechsier. (Cavara) & Widden. W. Gams. ATCCArsefCBS 309.85 256753542929 24874 .citrusspider.saprobe. rust mite. U32406AF339577AF339575AF339576AF339570 AF339528AF339527AF339525AF339526AF339520 MicrohilumMetarhiziumfiavovirideMetarhizium oncoperae anisopliae Yip (Metch.)Gamsvar.frigidum.var. & majus. Rath. & Sorok.Rozsypal. var. minus. ArsefIFO 5940 4358203731454606 NilaparvataOncoperaOryctes rhinoceros lugens intricala (Homoptera: (Lepidoptera:(Coleoptera: Delphacidae). Scarabaeidae). I-lepialidae). AF339581AF339580AF339578AF339579AB027337 AF339532AF339531AF339529AF339530AB027383 PaecilomycesPaecilomycesjavanicusNeotyphodium tenuipeslilacinus coenophialum (Peck)(Thom)(Friderichs (Morgan-JonesSamson. Samson. & BalIly) &Brown W. Gams) & Smith. Glenn et al. OSCArsefATCC 76403 2181322 52274 Nematoda:LepidopteraLitodactylusTall fescue, Tylenchida, Festucaleucogaster eggarundinacea (Coleoptera:mass of Meloidogyne Curculionidae). sp. AB027334U46880AF339583AF339582U45942 AB0273801347838AF339534AF339533U57681 Verticillium sect.CephalosporiumAcremoniumRot (ferophthora obclavatum lanosoniveum.lanosoniveumangustispora W. Gams.(Barron) (van Beyma) Banon. W. Gams. Prostrata. IMICBS 317442 101437704.86311.74 Hemileiaairrotifer, above decaying sugarpinevastatrix straw. field AF339603AF339602AF339567AF339584 AF339554AF339553AF339517AF339535 Verticillium balanoides.balanoidesantillanumaranearum (Drechsler)Castaneda(Petch) W. & Garns.Dowsett, G. Arnold. Reid & Hopkins. CBS 522.80335.80726.73a350.85 Ditylenchustriformisnematodespider.agaric in soil AF339590AF339589AF339586AF339585 AF339541AF339540AF339537AF339536 VerticilliumVerticillium chiamydosporium bulbillosumcfcampanulatum bactrosporum W. var.Glockling.Gams (Drechsler) calenulatum & Malla. Subram. (Barron) W. Gams. CBSIMI 356051 101433504.66145.70250.82 Soildungrootnematode of of Picea nematode in soil abies AF339593AF339587AF339592AF339591AF339588 AF3AF339538AF339543AF339542AF339539 39544 Table 2.1. (Continued) 22 VerticilliumVerticilliumfusisporum epiphytum.epiphytumgonoides Drechsler. Hansf.W. Gams. CBS 384.81154.61891.72164.70 nematodesCollriciaperennisHemileiauredium of vastatrix rust AF339599AF339598AF339596AF339597 AF339550AF339549AF339547AF339548 VerticilliumVertjcjl/jum lecanii. lecaniiincurvumlamellicola (Zimm.)(Torrubiella W. FEW. Heifer. Viegas. Smith.confragosa). CBSNRRLIMICBS. 304807 126.27 460.88116.25 28023 AgaricusCoccusGanoderma bisporus viridis. lipsiense AF049156AF339604AF339601AF339600 AF049176AF339555AF339552AF339551 Verticillium psalliozae.psa!liotae.psalliotae.psalliota Treschow. [MICBS 163640 532.81639.85363.86 rhizosphereIceryapurchasi()soilforestAgaricus soil ofbisporus Pseudotsuga menziesii AF339607AF339610AF339609AF339608AF339605 AF339558AF339561AF339560AF339559AF339556 VerticilliuniVerticilliumVerticilliu,n sinense.pseudohemipterigenum.sinensepseudohemipterigenumpsalliotae. K. G. Zhang, L. H.C.EvansCao & Z.Q.Liang. & Y. Jun. CBS196-1013Arsef 131.95567.95100172 5687 barknematodeCoccusnymphoflxodes(tick)Homoptera. near viridis. rootnear of root Act of inidia Actinida deliciosa deliciosa. AF339595AF339594AF339612AF33961AF339606 I AF339546AF339545AF339563AF339562AF339557 TELEOMORPHSVerticillum sect. Nigricentw.VerticjlljumVerticillium sp.suchlasporiumdahijac Klebahn. W. Gams & Dackm. var. catenatum. CBS 101284402.78464.88 eggsleaf litter of Heterodera of Acer saccharum. avenae. AF104926AF33961AF339613AF339614 5 AF104926AF339566AF339564AF339565 Perithiciat AscomycotaAphysiostromaChaetomiumCercophora septentrionalis stercorariumglobosum Kunze: Barassa, N. Lundq.Fr. Martinez and Moreno. U20379U32400U32398 U47825U478231.147820 GlomerellaDiatrypeDiaporthephaseolarumDaldiniaColletotrichum concentricadisc(formis cingulata gloeosporioides (Hoffm.:Fr)(Bull.:Fr.)(Stoneman) (Cooke &Ces. Fr.Spauld. Ellis) (Penz.) & de Sacc. & Not.Penz. H. Sehrenk. & Sacc. U48427U32403L36985M55640 U48428U47829U47828Z18999U47830 Nirenberg MicroHypomycesHypocreaHaematonectria ascus schweinitzii polyporinustrigonosporus haematococca (Fr.) Peck. C. Sacc. W. (Berk. Emmons & Br.) & B.Samuels 0. Dodge. & L36987U32410L36986N32413 U47835AF04913U47833AF04914 23 TableNeocosmosporaNectria 2.1. cinnabarina (Continued) vasinfecta (Tode: E.F. Fr) Fr.Smith. U32414U32412 U00748 SphaerostilbellaNeurosporaPetOphiostomapiliferum ne/la setiferacrassa aureonitens Shear(J. C. Syd. Schmidt) (Tul.) SeifertCurzi. et al. & & B. 0. Dodge.P. Syd. N32415N32421N20377X04971 N00755N48421N47837M38154N47836 ClavicipitaceaeXylaria hypoxyloncuria Fr. (L.:Fr.) Grey. Diehl. N20378U32417 N47841N47840 BalansiaAiricordycepsAtkinsonella henningsianaaristidae hypoxylon harpospor(fera (Moell.) (Peck) Samuels. Diehl. Atkinson. Stevens & Arsef 5472 .Poaceac.millipede. 1J44036U44035AF339569U44034 U57678U57677AF3395U57087 19 CordycepsCordycepioideusClavicepspurpureaClavicepspaspali bifusispora bisporus F. (Fr.:Fr.) 0. E.Stifler. Tu1.. Eriksson. J. G. Hall. Lepidoptera.termitePoaceae. (Isoptera). AF339571AF009651U44040U32401 AF339521AF0096541357085U47826 Cordyceps cochlidiicolacoccidiicolacapitata (Fr.) Y, Y,Link. Kobayasi Kobayasi & & 0. Shimizu.D. Shimizu. Arsef 5690 Lepidoptera.mothscaleElaphomyces. insect (Homoptera). (Lepidoptera). AB027331ABO3AB0273 1195 18 AB027377ABO3AB027364 1196 CordycepsCordyceps ophioglossoidesmilitarisroseostromata gunnii (L. (Berk.) Y, Sacc.Kobayasi : Fr.) (Ehrh.) Link. Link. & 0. Shimizu. ArsefOSC 76404 4871 Elaphomyces.Lepidoptera. (Lepidoptera). AF339573AB027321AB027333AF339572 AF339523AB027367AB027379AF339522 MyriogenosporaEpichloeCordyceps typhinaamanillans scarabaeicola atramentosa(Pers.) J. F. Tul. Y,White. Kobayasi (Berk. & & Curt.) Diehl. 0. Shimizu. Arsef 5689 Poaceae.beetlePoaceae. (Coleoptera: Scarabaeidae). 134415513324051335034AF339574 135708413173961357680AF339524 24

by agarose gel electrophoresis of 5 il of the reaction mix. PCR products of SSU and

LSU rDNA were purified using QlAquick PCR purification kits (Qiagen Inc.,

Valencia, CA). Purified PCR products were visualized on a 1% agarose gel stained

with ethidium bromide and quantified using Gibco-BRL low DNA Mass Ladder. The

purified product was sequenced using ABI Prism BigDye Terminator Cycle

Sequencing chemistry with AmpliTaq DNA polymerase, PS on an ABI Prism Model

377 (version 2.1.1) automated DNA sequencer (Perkin-Elmer) at the Central Services

Laboratory of the Center for Gene Research and Biotechnology at Oregon State

University. The template strands of purified PCR-products were directly sequenced utilizing the primers NS1, SR7, N53, NS4 for SSU rDNA and LROR, LR5 for LSU rDNA (White et al., 1990; Vilgalys and Sun, 1994).

Phylogenetic analyses

DNA sequences were edited in SeqEd Ver. 1.0.3, manually aligned using a color font, and appended to a preexisting data set of the Clavicipitaceae and other perithecialfungi(Spatafora et al., 1998). The phylogenetic analyses were performed with PAUP* 4.0 (Swofford, 1998). Parsimony analyses were performed on the combined data set of SSU and LSU nrDNA using the following heuristic search options: 100 replicates of random sequence addition, TBR (Tree bisection- reconnection) branch swapping, and MulTrees ON. Insertions and deletions (indels) were minimized in alignments and gaps were treated as missing data in the analyses.

Ambiguously aligned sequence regions were excluded from the data matrix before 25

analysis. Weighted parsimony analyses were performed using a step-matrix to weight

nucleotide transformations based on the reciprocal of the observed transition:

transversion (TN: TV) ratio. TN: TV ratios were estimated a-priori from an average of

pairwise comparisons calculated in PAUP using the "Pairwise base differences"

command. Relative support for the resulting trees was determined by 2500 bootstrap

replications on informative characters only with the previously mentioned search

options except that only one tree was retained during each replication (Moncalvo et

al., 2000). The phylogenetic trees generated from the combined data set with SSU and

LSU nrDNA data sets were rooted with Xylaria curta and Xhypoxylon.

To test alternative phylogenetic hypotheses for V. sect. Prostrata, i.e., monophyly of V. sect. Prostrata, constraint topologies were constructed in MacClade

3.0 (Maddison and Maddison, 1992). Constraint topologies forced the monophyly of

V. sect. Prostrata, but left all other nodes of the tree as unresolved. These topologies were used as starting trees in maximum parsimony analyses; search options were as described above except that MaxTrees was set to 1000. The most parsimonious trees recovered from the constraint analyses were statistically compared to the trees recovered from the maximum and weighted parsimony analyses using the Templeton

WSR test implemented in PAUP* 4.0b3 (Swofford, 1998). Host affiliation was mapped onto one of the most parsimonious trees using MacClade 3.0 (Maddison and

Maddison, 1992) with equal weights for all character state transformations.

Results 26

The combined SSU and LSU nrDNA dataset included 2,040 aligned nucleotide

positions, with SSU rDNA comprising 1100 and the LSU rDNA comprising 940

positions. One-hundred and ten SSU rDNA positions and 128 LSU rDNA positions

were excluded due to either ambiguously aligned regions or an excess of missing data

near the 5' and 3' ends. The final data set included 1,802 nucleotide positions of which

347 positions (146 from the SSU rDNA and 201 from the LSU rDNA) were identified

as parsimony-informative. Maximum parsimony analysis of the 93 taxon dataset yielded 329 equally most parsimonious trees of 1,710 steps. For each of these trees, the consistency index (CI) was 0.322 and the retention index (RI) was 0.690. Although a large number of trees were recovered in these analyses, many of the nodes amongst the major genera and groups of the Clavicipitaceae were resolved in the strict consensus tree (Figure 2.1).

The averaged TN: TV ratios estimated for SSU and LSU nrDNA sequences from a-priori pairwise comparisons were 3.0 and 2.4, respectively. These TN:TV ratios violated the triangular inequality assumption, and therefore two sets of weighted parsimony analyses were performed. In the first set of analyses, the SSU and LSU partitions of the data set were weighted 2.0 and 1.6; a weighting scheme that simply reflects the initial averaged TN:TV, but scaled to 2.0. In the second set of weighted analyses, transversions were weighted 2 times greater than transitions in both the SSU and LSU partitions of the data. The heuristic search options were as described previously for the maximum parsimony analyses except only 50 replications were 27 performed. The first set of weighted parsimony analyses resulted in 720 trees of

2113.2 steps (CI = 0.338, RI = 0.699); the second set of weighted parsimony analyses resulted in 2832 trees of 2250 steps (CI = 0.339, RI = 0.70 1). The major difference between the maximum parsimony and the weighted parsimony analyses was an increase in the number of trees recovered in the weighted analyses and a decrease in resolution of the strict consensus trees (data not shown). This observation is consistent with recent studies, which suggest that weighting character state transformations often does not result in substantially better hypotheses with respect to well supported nodes and may actually decrease the resolution power of some data sets (Broughton et al.,

2000, Tehler et al., 2000). The decrease in resolution was centered around the poorly resolved region of the tree that included the grass symbionts and V. sect. Prostrata group D. One of the most parsimonious trees from the maximum parsimony analyses was chosen at random and is shown in Figure 2.2 for the purpose of displaying branch lengths and mapping host affiliation.

Both the maximum and weighted parsimony analyses of the combined SSU and LSU rDNA data support the inclusion of all isolates sampled from V. sect.

Prostrata in the Clavicipitaceae except for V. incurvum. However, the monophyly of the Clavicipitaceae is weakly supported by bootstrap values and is characterized by a relatively short branch. The clavicipitalean isolates of V. sect. Prostrata did not form a monophyletic group and were placed in at least three separate parts of the

Clavicipitaceae dade. To test the monophyly of V. sect. Prostrata, two topological constraint analyses were performed. One topological constraint that forced the 28

Xyiana wv Xylwia hyJKxylC.' Perneila mftm A/4aa80rna ae,rci,. Hypr.a ,kø64nlli Hyespdyp*jem.gi1beIIa awe,,i!em V. ,ncwvwntBS Necl,w Hcrmalo.'wcf no /a,ndcccrca Mra,rn,u vayi,fecta Arciwn alb,,n CBS 40170 OSC 76403 'Pxcikenycesteaa NRRL 220231 V ki V. sect. Prostrata 17v11 304207I 029B groupBl B. birzigniaflii Beea.wna ca!eehca C.,wrJ&.:cc/a P jovr.cIo C,uasm,n0/0 Codycep ,,,iIilo,u Ve#ki/Ii.,, CBS 402 Ve24ie,/ho 7eea, CBS 12627I CBS 1(01721 8641163640 " I CBS 53281 V. sect. Prostrata CBS 363.86I I VUU.0 CBS 401224 IgroupBl Mc,thIwn azoeeor Arf 4358I V. pae1Iio CBS 639.85 I V xo,wn CBS 726. 73a I VJ6owwnCBS 164.70 I V. ooi1hm.,, CBS 35024 I CBS 3(24.851 E,wktzw,, owzea,um Arf2929 I IMI 317442 CBS 704.86C io.ewn V. sect. Prostrata Ae,ao,,iwn Iawhl, CBS 31 I% V. !neI1&daCBS I16, group18h

5461V. eJlp'ytM,l V. sect. Prostrata CBS 38481 I Af 2037 group 01 Are4856 I Ai,ef 3145 I !F05940 I V. aWathe CBS 464.88 I V geds CBS 89472 I V. sect. Prosirala V. ci bocImpntn CBS 1014331 V. b.o CBS 14570 I group 02 V. chk.,th9,aiwn CBS 504.861 AthinaeIIo hypory/a7 Bak.aa an.toke B. gow Myeiogeno.0,c0,en1ao Ck64cepjzupoo C. jxath aemIJ1e,igemm V. SCCt. ProstraJa Ro//èpvp6t!r,a group D3 Ce,rlycezxiekoe icuc 85radefla !hcmJm0 C. cerWidiccla C ccidiccb Ho,Tex,,,um helicc,des

CBS 335801 V.lcthoaeks I v. I V. sect. Prostrata CBSIiI95 v00 CBS 567951 I groupC V. &*,tkk. CBS 580 I C C. capsab C 46ogloidca P. l,leri,00 Cd/elc64ch/.n glceo,w.ks GIr,,,e,rd/a ci80dasa Ve,iicillu,, ckèJia Cercq*em ejaemh Qijetceoio,, giobaiwn Newv4wa craw, Op'uasta si!/ènm Diaj1he do/ow,,, Dailwjek.a.sfomn DoId,,ia cathepanw

Figure 2.1. Strict consensus tree of 329 most parsimonious trees from the maximum parsimony analyses of the combined SSU and LSU rDNA dataset. Species represented by multiple isolates are differentiated by accession number; acceession numbers for species represented by a single isolate are listed in Table 1. Bootstrap values of >5Q%are noted above their respective nodes. The three major groups of Verticillium sect. Prostrataare designated to the right of the species names following the convention of Zare et al. (2000); subgroupsare designated numerically, i.e., BI, Dl, D3. The nodes representing the Hypocreales and Clavicipitaceae are noted by H and C, respectively. 29

monophyly of V. sect. Prostrata including V. incurvum (constraint 1), while the second constraint forced the monophyly of V. sect. Prostrata without V. incurvum

(constraint 2). Maximum parsimony analyses using constraint 1 and constraint 2 topologies as starting trees resulted in trees that were 75 and 55 steps longer, respectively. The most parsimonious trees from both constraint 1 and constraint 2 were both rejected as being significantly worse explanations of the data than the trees from both the maximum and weighted parsimony analyses (P = <0.0001). This finding is consistent with the observation that the ITS rDNA sequences of V. incurvum could not be aligned with the remaining taxa of Verticillium (Zare and Gams, unpublished). The most parsimonious trees recovered from the maximum and weighted parsimony analyses were not significantly different.

Discussion

Familial relationships ofVerticillium sect. Prostrata

Verticillium is known to be phylogenetically related to many families of perithecial ascomycetes (Gams, 1971; Gams and van Zaayen, 1982; Samuels, 1988;

Zare et al., 2000). In these analyses, the specimens sampled from V. sect. Prostrata grouped within the Clavicipitaceae with the exception of V incurvum, which grouped more closely with members of the Hypocreaceae. The clavicipitaceous affinity of V. sect. Prostrata is consistent with hypotheses from morphological studies (Gams, 1971; 30

Gams and van Zaayen, 1982) and phylogenetic analyses using the ITS rDNA (Zare et

al., 2000). Although the Clavicipitaceae were inferred to be monophyletic, their

monophyly was not strongly supported by the data. The clavicipitaceous region of the

rDNA tree is characterized by relatively short basal branches that received low bootstrap support (Figure 2.2). The monophyly of the family is, however, consistent with morphology. The Clavicipitaceae are united by the synapomorphies of long

cylindrical asci with a conspicuously thickened apex and filiform ascospores that typically disarticulate into part-spores (Diehi, 1950; Mains, 1949; 1958; Rogerson,

1970; Kobayashi, 1982; Spatafora and Blackwell, 1993).

These data are also consistent with a distant relationship between fungi of V. sect. Prostrata and those of V. sect. Nigrescentia (Zare et al., 2000). In a previous study of the SSU rDNA, the plant pathogen V. dahliae Kieb. was shown to be closely related to the Phyllachoraceae (Messner et al., 1996). This relationship was not strongly supported by the data and the node in question was characterized by long branch lengths. Here we included sequence data from both the SSU and LSU rDNA and a larger sampling of perithecial ascomycetes, which may serve to disrupt long branch attraction (Felsenstein, 1985; Graybeal, 1998). These analyses increased the support for a close relationship between V. dahliae and Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. and its teleomorph Glomerella cingulata

(Stonem.) Spauld. & Schrenk and the inclusion of V. dahliae in the Phyllachoraceae

(Figure 2.1). These data also strengthened the hypothesis of the disparate relationship between V. sect. Prostrata and V. sect. Nigrescentia (Zare et al., 2000). 31

o AphysiosiromaSIercorarium Hyvocrea schweinitzii aureonitens

Vincurvu°CS46(c'8 Nectria cinnab anna -f::I:;.t.r-- Haematonectria huematococca vasinjecta phanocladium albumCBS40I .70 g Pacci/omyces tenuipesOSC 764Q3 I 00 P.tenuipesAB02734 I Lepidopteia Cbtfusi.spora V.aranearumCBS 726.73a - Arachnila Vlecan,,NRRL 28023 V.lecanuIMI 304807 - Homoptera '00 B. baoc,anaNRRL 28020 63 B. bass,anaIFO 4848 56 B. brongniartu 63 Beauveria caledon,ca-Saprthe C.scarabaeicola -Coleoptera VerI,c,lliumsp. CBS 101284 88 * Microh,lum oncoperae- Lepidoptera 00 C.mseo,ctromata Cm,l,tar,.c Lepidoptera P.Javanicus - oleoptera V.psalhotaeIMI 163640 - Soil VpsalliotaeCBS 100172 - Arachnila V. fiisisoorumCBS 164 70 - Aphyllophorales Vant,1lanumCBS 350.85 - Agaricabs 98 Engyodontium aranearurnCBS 309.85 Arar Engyodontium aranearumArsef 2929 VpsalliotaeCBS 639.85 - Rhizosphere VpsalliolaeCBS 532.81 Soil Vpsall,otaeCBS 363.86 - Agaricabs Vertic,lliumsp. CBS 402.78 - Leaf litt Veri,c,ll,um Ic canh,BS 126.27 - Homoptera 66 C lanosoniveumIMI 317442 I 86 C lanoson,ve urnBS 704.86 Rust 00 Acremoniurn obclavatum -Air Vlarnell,colaCBS 116.25 - Agaricabs V.chlamydosporium - Soil amaril Ions Ep,chloe typhina I Poese Weotyphod,um coo nophialum I

)9 V epiphytum I CBS 154.61 Rust VepipIiytumCBS 384.81 I 99 M.flavov,r,devar.minus- Homoptera ss Al.anisopliaevar.frigidum ,00 Al.anbsopl,aevar.majus -Coleoptera EMelarhi:ium a nisopl:ae V.suclporium- Nematoth V ono,de.e -Aphyllophorales 98 ulbillosum -Pinese V. cf.bac!ro.sporum g AIkinsonella hypoxylon

Balansia aristhdae I .J98MyniogenosPoraB. hennings,ana I Poae a/name ntosaI P purpurea C I too Vpseudohem,pten,g,enumArsef 5687 I Vose udohem,pter,ginurn196-1013 - Homoptera i?oufèrop hthora an gust i.spo ra - Rotifera 90 (ordvcepio,deus 6 isporus - Isoptem thompsonh, -Arachniia 99 C cochlhducola- Lepidoptera C Ccocc,diicola- Homoptera so helicoid s- Nematoth AIr,cordyceps harposporhJra- Diplopoda

Vbalanoides I CBS 250.82 Nematoth V.balanoidesCBS 335.80 I V.campanulatum- dung of nematode V..cinenseCBS 131.95 - bark V.s,nenseCBS 567 95 - Nematoth Vhalano,de.cCBS 522.80 - Nematoth C. gunnhi- Lepidoptera capita/a - Coph,oglo.s.coides Elaphomycci - P.lilacinu.,- Nematoch - 5 changes

Figure 2.2. A phylogram of one of the 329 most parsimonious trees. Only the Hypocreales section of the tree is shown to emphasize the Clavicipitaceae and Verticillium sect. Prostrata. Tree descriptors are as in Figure 2.1 with the exception that nodes which collapse in the strict consensus are designated by asterisks. Host or substratum is provided to the right of species names. 32

Polyphyly ofVerticillium sect. Prostrata.

Although V. sect. Prostrata was confirmed as being included in the

Clavicipitaceae, the monophyly of V. sect. Prostrata was rejected. The most parsimonious trees from the analyses of constraint topologies, which forced the monophyly of V. sect. Prostrata with and without V. incurvum, were rejected as significantly worse explanations of the data (Templeton WSR test P = < 0.0001).

These results suggest that V. sect. Prostrata is polyphyletic and does not represent a natural group of fungi within the Clavicipitaceae. At least three, and possibly six, separate groups of V. sect. Prostrata were resolved within the Clavicipitaceae. We emphasize groups rather than clades out of convenience as several of thegroupswere not supported as monophyletic in the maximum parsimony analyses (Figure 2.1 and

2.2), but represent closely related sets of species. Also, certain morphological and ecological traits are discussed in the context of thesegroups.We follow the convention of Zare et al. (2000) and designate thegroupsof V. sect. Prostrata as B-D with group A representing V. dahliae of V. sect. Nigrescentia. Groups B and D are not supported as monophyletic. Group B comprises two closely relatedgroupsthat are designated B 1 and B2. Group D includes three separate lineages of V. sect. Prostrata

(D1-D3), a finding consistent with Zare et al. (2000), that were part of a poorly resolved region of the tree (Figure 2.1 and 2.2). Group Bi includes the type of V. sect.

Prostrata, the anamorph of C. militaris, and morphologically similar species of V. lecanii, V. psalliotae, V. fusisporum, V. araneareum, and V. antillanum;groupB2 33 includes V. lamellicola, Cephalosporium lanosoniveum, and Acremonium obclavatum.

Species in group B are characterized by phialides produced on prostrate conidiophores, nonadhesive conidia that vary in morphology from oval to cylindrical to falcate, and the lack of dictyochiamydospore production. Group C includes the nematophagous species V. balanoides, V. campanulatum, and V. sinense. It is distinguished from the other groups of V. sect. Prostrata by the basal inflation of phialides, which are similar to Tolypocladium, and adhesive conidia that are balanoid and rarely elongated.

Group D is grossly paraphyletic in these analyses and includes three lineages that are part of an unresolved region of the tree (Figure 2.1). Group Dl consists of the rust parasite V. epiphytum, which produces falcate conidia that are morphologically similar to V. psalliotae, but differs from the latter species in its sparse production of thick-walled, cyanophilic . Group D2 comprises the species V. chiamydosporium, V. suchiasporium, V. cf. bactrosporum, and V. gonioides. These species occur primarily on nematode cysts and eggs and are characterized by the often prominent production of dictyochiamydospores and nonadhesive conidia that are oval to subglobose to truncate. Verticillium suchiasporium is unusual among V. sect.

Prostrata in that it forms erect conidiophores. Group D3 consists of the homopteran pathogen V. pseudohemipterigenum, which produces a compact whorl of phialides on erect conidiophores that distinguishes it from other species of the group (Hywel-Jones et al., 1997). The considerable morphological differences that exist among the fungi of 34 group D may serve to divide them into more natural groups in future analyses (Zare et al., 2000), however, the data presented here do not robustly address this issue.

Discrepancies exist whether some species of V. sect. Prostrata (e.g., V. balanoides) are capable of dictyochlamydospore production and considerable debate exists over the taxonomic utility of dictyochlamydospore production among species of

Verticillium. Because of its variable formation in culture, Gams (1971, 1988) regarded the dictyochlamydospore as a poor taxonomic character, unsuitable for a subdivision of V. sect. Prostrata. In contrast, Kamyschko (1962) erected Diheterospora and

Batista and Fonseca (1965) introduced Pochonia, largely based on the basis of dictyochlamydospores, both genera with type species that turned out to be identical with V. chiamydosporium. This taxonomy was followed by Barron and Onions (1966) and Barron (1985) who considered the dictyochiamydospore as a valuable character in the taxonomy of nematophagous fungi and parasites of (e.g. Rotferophthora).

These analyses suggest that while production of dictyochlamydospores may distinguish some taxa, i.e., group D2, their production may be polymorphic for some verticilliate fungi of the Clavicipitaceae and the ability to produce such structures may have been lost and gained several times. A more accurate character state reconstruction of this trait is not possible until its distribution is confirmed for more anamorphs of the Clavicipitaceae.

RelationshipsofV. sect. Prostrata with anamorphs and teleomorphsofthe

Clavicipitaceae. 35

Isolates of V. sect. Prostrata integrate with anamorphs and teleomorphs of the

Clavicipitaceae at several points in the rDNA phylogeny. Importantly, not only was

Verticillium revealed to be polyphyletic, but Cordyceps was also inferred not to be

monophyletic. It consisted of two separate clades; one that included C. militaris and

relatives and will be referred to as Cordyceps sensu stricto, and one that included the

parasites of Elaphomyces (e.g., C. ophioglossoides) and pathogens of a diversity of

arthropods and will be referred to as the C. ophioglossoides dade. Furthermore, the

grass endophytes of the Clavicipitaceae did not form a monophyletic group, although

their monophyly could not be rejected by parsimony analyses of constraint topologies

that forced their monophyly (Templeton WSR, P >>0.05). V. sect. Prostrata group B

is included with teleomorphs of Cordyceps s.s. and the closely related anamorphic

species of Beauveria, Engyodontium, Microhilum, and Paecilomyces. The

teleomorphs of the dade include C. militaris, the type species of Cordyceps, and

morphologically similar and closely related Cordyceps species that are pathogens of

Lepidoptera (e.g., C. bfusispora O.E. Erikss.) and Coleoptera (e.g. C. scarabaeicola

Y. Kobay.). Of particular interest are the isolates of Verticillium that are known to be

linked with teleomorphs of Torrubiella, a genus hypothesized as closely related to

Cordyceps because of its ecology as an entomopathogen and its ascus and ascospore morphology (Mains 1949, Kobayasi 1982). Verticillium lecanii (isolate IMI 304807) was established from an isolate of Torrubiella confragosa pathogenic on scale insects

(Jun et al., 1991). Verticillium aranearum has also been linked to Torrubiella (T. alba

Petch), although the culture sampled here was not isolated from a teleomorphic 36 specimen. Another link to Torrubiella may exist in V. sect. Prostrata group B2, the most closely related dade to Cordyceps s.s. Cephalosporium lanosoniveum CBS

740.86 was isolated from a teleomorph tentatively identified as Torrubiella sp. (H. C.

Evans, pers. comm.). These results suggest that Torrubiella may display its closest phylogenetic affinity among the brightly pigmented, fleshy stromatic species of

Cordyceps s.s.

In the strict consensus tree (Figure 2.1), V. sect. Prostrata group C is resolved as being related to teleomorphs of the C. ophioglossoides dade (e.g., C. cap itata, C. gunnii) and the closely related anamorphs of Hirsutella, Harp osporium, and

Paecilomyces (Figure 2.1 and 2.2). Teleomorphs of this dade are represented by darkly pigmented, stromatic species that show considerable diversity in morphology and host affiliation to the point that two isolates, bisporus and

Atricordyceps harposporioides, are not currently classified in Cordyceps. The C. ophioglossoides dade is not strongly supported by the data (Figure 2.2) and the current sampling certainly underrepresents the phylogenetic diversity of teleomorphs of the Clavicipitaceae. However, C. gunnii, an Australian parasite of lepidopteran larvae (Kobayasi 1941, 1982), is well supported as a teleomorph closely related to V. sect. Prostrata group C (Figure 2.2). An anamorph has not been established by culture methodology for C. gunnii and these data do not establish that C. gunnhi is the teleomorph of any member of this group of V. sect. Prostrata. Rather, these results provide predictive value for future sampling of teleomorphs that may further illuminate teleomorph-anamorph relationships and connections within the 37

Clavicipitaceae. Production of Verticillium anamorphs is known for C. ophioglossoides, a parasite of Elaphomyces, but the phialides lack the basal inflation as in V. sect. Prostrata group C and these data do not confidently establish a close relationship between C. ophioglossoides and V. sect. Prostrata group C.

Although V. sect. Prostrata group D is confinned as a member of the

Clavicipitaceae, its relationships to teleomorphs and other anamorphs of the family are poorly resolved (Figure 2.1 and 2.2). Group Dl (V. epiphytum) is included in the dade containing the teleomorphs and anamorphs of Epichloë and group D3 (V. pseudohemipterigenum), along with the nematophagous species and Rotferophthora angustispora, grouped most closely with the teleomorphs of , Balansia and Claviceps. None of these relationships were strongly supported by the data (Figure

2.2). All of these teleomorphs are symbionts of the Poaceae (Rogerson 1970, Clay

1988) and some are linked to anamorphs that are morphologically similar to V. sect.

Prostrata (e.g., Acremonium, ) (Diehi 1950, Morgan-Jones and Gams

1982, Glenn et al., 1996, Schardi et al., 1997). The similarity in anamorphs, however, exists for many teleomorphs of the Clavicipitaceae and is by itself not proof of a close relationship. Group D2 of V. sect. Prostrata and Metarhizium, an entomopathogenic anamorph, are also placed in this poorly resolved region of the tree, but again their relationship is not strongly supported by the data (Figure 2.2) and remains speculative.

Evolutionofhost association In the classification of V. sect. Prostrata and Clavicipitaceae, host affiliation has been regarded as an important taxonomic character (Die! 1950, Gams and van

Zaayen 1982, Zare et al., 2000). In an attempt to better understand the evolution of host-jumping, host association was mapped onto one of the most parsimonious trees

(Figure 2.2). Because of the lack of support for much of the basal nodes of the

Clavicipitaceae, and the presence of multiple most parsimonious trees, we present these results as working hypotheses and emphasize only the more strongly supported resolutions (Figure 2.2).

V. sect. Prostrata group B 1 includes pathogens of insects, mites, and fungi. Its placement within Cordyceps s.s. expands the host range of this dade beyond that of

Lepidoptera and Coleoptera and bolsters arguments that the inclusion of anamorphic taxa in phylogenetic analyses improves our understanding of the evolution of host affiliation and life histories. Pathogens and parasites of fungi are located in four regions of the rDNA tree of the Clavicipitaceae (Figure 2.2), a pattern consistent with multiple origins of fungal pathogens on distantly related groups of fungi, i.e., homobasidiomycetes, rusts, and ascomycetous truffles. However, these data do support independent origins of fungal pathogens on the same host (Hemileia vastatrix), as two clades (B2 and Dl) of the rust pathogens are resolved in these analyses (Figure 2.2). Nikoh and Fukatsu (2000) proposed that the parasites of

Elaphomyces originated from a host jump from arthropods onto truffles. While these data are consistent with their finding, they do not provide unequivocal support for the polarity of a single arthropod to Elaphomyces host-jump (Figure 2.2). At least two separate groups of nematophagous fungi exist among the species of Verticillium, suggesting that two independent origins of of nematodes may have occurred within the Clavicipitaceae (Figure 2.2). However, the poorly resolved basal nodes of the Clavicipitaceae do not allow us to distinguish between the hypotheses of two independent origins of parasitism of nematodes and a single origin followed by multiple losses or host-jumps. Furthermore, many isolates of nematophagous Verticillium species are isolated from plant or fungal material and the extent to which they may exist saprotropically in nature is unknown.

Conclusions

The data presented here illustrate the need for the inclusion of anamorphs and teleomorphs in common phylogenetic analyses. Gene phylogenies provide testable phylogenetic hypotheses of anamorph-teleomorph relationships and their evolutionary history, despite disparate morphologies and ecologies. These data supported the inclusion of most isolates of V. sect. Prostrata in the Clavicipitaceae, but rejected the monophyly of these fungi within the family. The Verticillium morphology is distributed throughout much of the Clavicipitaceae, but differences in phialide and conidium morphology, and to a lesser extent dictyochiamydospore production are consistent with certain groups of V. sect. Prostrata. Finally, many of the basal nodes in the rDNA phylogeny of the Clavicipitaceae were characterized by short branch lengths and poor statistical support. Molecular analyses that include increased 40 sampling of additional loci and taxa of the Clavicipitaceae and their anamorphic fungi are needed to more confidently resolve these nodes and better define monophyletic, infrafamilial clades.

Acknowlegements

This research was supported by a grant from the National Science Foundation

(DEB-9806936). Most cultures were provided by the Centraalbureau voor

Schimmelcultures, some by USDA-ARS Entomopathogenic Fungal Collection, and

CABI Bioscience.

References

Barron, G. L., and A. H. S. Onions. 1966. Verticillium chiamydosporium and its relationships to Diheterospora, Stemphyliopsis, and Paecilomyces. Canadian Journal of Botany 44:861-869.

Barron, G. L. 1977. Predators and parasites of microscopic animals. Pages 167-200 In Biology of conidial fungi, Vol. 2 (Cole, G. T. & B. Kendrick, eds). Academic Press, New York.

Barron, G. L. 1985. Fungal parasites of bdelloid rotifers: Diheterospora. Canadian Journal of Botany 63:211-222.

Barron, G. L. 1991. A new genus, Rotferophthora, to accommodate the Diheterospore-like endoparasites of rotifers. Canadian Journal of Botany 69: 494-502.

Broughton, R. E., S. E. Stanley, and R. T. Durett. 2000. Quantification of homoplasy for nucleotide transitions and transversions and a reexamination of 41

assumptions in weighted phylogenetic analysis. Systematic Biology 49: 617- 627.

Clay, K. 1988. Clavicipitaceous endophytes of grasses. Coevolution and the change from parasitism to mutualism. Pages 79-105 In Coevolution of Fungi with Plants and Animals (Hawksworth, D. L. & K. Pirozynski, eds). Academic Press, London. Diehi, W. W. 1950. Balansia and the Balansiae in America. Agricultural monograph No.4. U. S. D. A. Washington.

Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39:783-79 1.

Gams, W. 1971. Cephalosporium-artige Schmmelpilze (Hyphomycetes). G. Fischer, Stuttgart.

Gams, W. 1988. A contribution to the knowledge of nematophagous species of Verticillium. Neth. J. P1. Pathol. 94:123-148. Gams, W and A. V. Zaayen. 1982. Contribution to the taxonomy and pathogenicity of fungicolous Verticillium species. I. Taxonomy. Neth. J. P1. Path. 88:57-78.

Gardes, M and T. D. Bruns. 1993. ITS primers with enhanced specificity for basidiomycetes-application to the identification of mycorrhizae and rusts. Molecular Ecology Ecology 2:113-118.

Glenn, A. E., C. W. Bacon, R. Price, and R. T. Hanlin. 1996. Molecular phylogeny of Acremonium and its taxonomic implications. Mycologia 88:369-383.

Graybeal, A. 1998. Is it better to add taxa or characters to a difficult phylogenetic problem? Systematic Biology 47:9-17.

Jun, Y., P. D. Bridge, and H. C. Evans. 1991. An integrated approach to the taxonomy of the genus Verticillium. Journal of General Microbiology 137:1437-1444.

Kamyschko, 0. P. 1962. De monilialibus terrestribe novis notula. Bot. Mater. Notul. Syst. Sect. Cryptogram. Inst. Bot. Acad. Sci. SSSR, 15:137-141.

Kerry, B. R. and D. H. Crump. 1977. Observations on fungal parasites of females and eggs of the cereal cyst-nematode, Heterodera avenae, and other cyst nematodes. Nematologica 23:193-201.

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku (Section B, no. 84) 5:53-260. 42

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364.

Lim, T. K and Z. W. N, Wan. 1983. Mycoparasitism of the coffee rust pathogen, Hemileia vastatrix, by Verticillium psalliotae in Malaysia. Pertanika. 6:23-25.

Maddison, D. R., and W. P. Maddison. 2000. MacClade 4: Analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts.

Mains, E. B. 1949. New species of Torrubiella, Hirsutella and Gibellula. Mycologia 41 :303-3 10.

Mains, E. B. 1958. North American entomogenous species of Cordyceps. Mycologia 50:169-222.

Mankau, R. 1980. Biocontrol: fungi as nematode control agents. Journal of Nematology. 12:213-232.

Messner, R., W. Schweigkofler, M. Ibi, G. Berg, and H. Prillinger. 1996. Molecular characterizattion of the plant pathogen Verticillium dahliae Kleb. using RAPD- PCR and sequencing of the 18SrRNA-gene. Journal of Phytopathology 144:347-354.

Moncalvo, J.M., F.M. Lutzoni, S.A. Rehner, J. Johnson & R. Vilgalys. 2000. Phylogenetic relationships of agaric fungi based on nuclear large subunit ribosomal DNA sequences. Systematic Biology 49:278-306.

Morgans-Jones, G. and W. Gams. 1982. Notes on Hyphomycetes XLI. An endophyte of Festuca arundinacea and the anamorph of Epichloe typhina, new taxa in one of two new sections of Acremonium. Mycotaxon. 15:311-318.

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

Nikoh, N., and T. Fukatsu. 2000. Interkingdom host jumping underground: Phylogenetic analysis of entomopathogenic fungi of the genus Cordyceps. Molecular Biology and Evolution 17:629-638.

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

Saksirirat, W., and Hoppe, H.-J. 1991. Secretion of extracellular enzymes by Verticillium psalliotae Treschow and Verticillium lecanii (Zimml.) Viégas 43

during growth on uredospores of the soybean rust fungus (Phakopsora pachyrhizi Syd.) in liquid cultures. Journal of Phytopathology 131:161-173.

Samuels, G. J. 1988. Fungicolous, lichenicolous, and myxomyceticolous species of , Nectriopsis, Nectria, Peristomialis, and Trichonectria. Mem. N. Y. Bot. Gdn. 48:1-78.

Leuchtmann, D. Chung, D. Penny, and M. R. Siegel. 1997. Coevolution by common descent of fungal symbionts (Epichloe spp.) and grass hosts. Molecular Biology and Evolution 14:133-143.

Schuler, T. M., Hommes, H.-P. Plate and G. Zimmermann. 1991. Verticillium lecanii (Zimmermann) Viégas (Hyphomycetales: Moniliaceae): Geschichte, Systematik, Verbreitung, Biologie und Anwendung im Pflanzenschutz. Biologische Bundesanstalt, Braunschweig.

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes I. The Clavicipitales-Hypocreales connection. Mycologia 85:912-922.

Spatafora, J. W., B. Volkmann-Kohlmeyer, and J. Kohlmeyer. 1998. Independent terrestrial origins of the Halosphaeriales (marine Ascomycota). American Journal of Botany 85:1569-1580.

Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.OblO. Sinauer, Sunderland, Massachusetts.

Taylor, J. W. 1993. A comtemporary view of the holomorph: nucleic acid sequence and computer databases are changing fungal classification. Pages 3-13 In The Fungal Holomorph: Mitotic, Meiotic and Pleomorphic Speciation in Fungal Systematics (Reynolds, D.R. & J.W. Taylor, eds): 3-13. CAB International, Wallingford, UK.

Taylor, J. W. 1995. Making the Deuteromycota redundant: a practical integration of mitosporic and meiosponc fungi. Canadian Journal of Botany 73:S754-S759.

Tehler, A., J. S. Fan-is, D. L. Lipscomb, and M. Kallersjo. 2000. Phylogenetic analyses of the fungi based on large rDNA data sets. Mycologia 92 :459-474.

Verhaar, M. A., T. Hijwegen, and J. C. Zadoks. 1996. Glasshouse experiments on biocontrol of cucumber powdery mildew (Sphaerothecafuliginea) by the mycoparasites Verticillium lecanii and Sporothrix rugulosa. Biol. Control 6:353-360. 44

Vestergaard, S. A., T. Gillespie, T. M. Butt, G. Schreiter, and J. Eilenberg. 1995. Pathogenicity of the hyphomycete fungi Verticillium lecanii and Metarhizium anisopliae to the western flower , Frankliniella occidentalis. Biocontrol Sci. Technol. 5:185-192.

Vilgalys, R., and B. L. Sun. 1994. Ancient and recent patterns of geographic speciation in the oyster mushroom Pleurotus revealed by phylogenetic analysis of ribosomal DNA sequences. PNAS 91:4599-4603.

White, T. J., T. D. Bruns, S. Lee, and J. Taylor. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pages 315-322 in PCR protocols, a guide to methods and applications. (M. A. hmis, D. H. Gelfand, J. J. Sninsky, and T. J. White, eds.). Academic Press, San Diego.

Zaayen, A. V., and W. Gams. 1982. Contribution to the taxonomy and pathogenicity of fungicolous Verticillium species. II. Pathogenicity. Neth. J. P1. Pathol. 88: 143-154.

Zare, R., W. Gams, and A. Cuiham. 2000. A revision of Verticillium sect. Prostrata. I. Phylogenetic studies using ITS sequences. Nova Hedwigia 71:465-480. 45

Chapter 3

RDINALIS SP. NOV,A NEW SPECIES OF CORDYCEPS WITH kSIAN-EASTERN NORTH AMERICAN DISTRIBUTION

Gi-Ho Sung and Joseph W. Spatafora

Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR

9733 1-2902, USA

Mycologia 96:658-666 46

Abstract

Cordyceps cardinalis, a pathogen of lepidopteran larva, is described as a new

species from the southern Appalachians Mountains of eastern and

southeast Japan. It is macroscopically similar to both C. militaris and C. pseudomilitaris, however, microscopic, cultural, and molecular evidence support it

being a phylogenetically distinct species. Cordyceps cardinalis is most similar to C. pseudomilitaris in the microscopic characters of nondisarticulating ascospore

morphology and its host affinity for lepidopteran larvae, which contrast with the

characters of C. militaris. Aspects of morphology, host association, phylogeny, and

biogeography are discussed. 47

Introduction

Cordyceps (L.: Fr.) Link includes pathogens of species from nine orders of arthropods and parasites of one genus of fungi, Elaphomyces (Kobayashi, 1941; 1982;

Mains, 1957; 1958). The type of the genus, and possibly best known species, is C. militaris (L.: Fr.) Link. It is characterized by the production of orange to orangish-red colored stromata that erupt from the carcass of the dead host (Mains, 1958; Kobayasi

1982). C. militaris is one of the most frequently collected species with collections from all major continents except Antarctica; however, it is also one of the most variable species in the genus with respect to morphology and host affiliation. It is most commonly reported from numerous families of lepidopteran pupae with less frequent reports from lepidopteran larvae and coleopteran pupae (Kobayashi, 1941;

Mains, 1958). Therefore, the species is generally recognized as a cosmopolitan taxon with a relatively broad host range that is characterized by a propensity for attacking lepidopteran pupae, but with the potential to attack multiple stages and hosts from more than one of arthropods (Kobayashi, 1941; Mains, 1958).

A number of macroscopically similar species, which had either been considered conspecific with C. militaris, close relatives, or had simply gone unnoticed among large collections from a variety of hosts and life cycle stages, have been described (Kobayashi, 1982; Mains, 1947; Hywel-Jones, 1994). Some examples include Cordyceps washingtonensis Mains from lepidopteran pupae (Mains, 1947), C. rosea Kobayasi and Shimizu from lepidopteran larvae (Kobayasi and Shimizu, 1982), C. roseostromata Kobayasi and Shimizu from coleopteran larvae (Kobayasi and

Shimizu, 1983), and C. pseudomilitaris Hywel-Jones & Sivichai from lepidopteran

larvae (Hywel-Jones, 1994). In the example of C. pseudomilitaris, Hywel-Jones

(1994) emphasized the macroscopic similarity to C. militaris, but also recognized

diagnostic differences between the two species. Cordyceps pseudomilitaris possesses

nondisarticulating ascospores in contrast to C. militaris, which produces ascospores

that disarticulate into partspores at the septations. The anamorph of C. militaris is a

Lecanicillium Gams & Zare (= Verticillium Nees), whereas C. pseudomilitaris mainly

produces simple, unbranched phialides that give rise to a single conidiogenous cell and

conidium (Hywel-Jones, 1994), a morphology that is more appropriately

accommodated in the recently described taxon Simplicillium Gams & Zare (Zare and

Gams, 2001). Finally, C. pseudomilitaris is only known from lepidopteran larvae of

undetermined familial affinity, whereas C. militaris is most frequently collected from pupae of a number of different families of . Recent molecular data for C. pseudomilitaris supported its distinction from C. militaris (Artjariyasripong et al.,

2001) and that these characters may be informative in recognizing unique and possibly cryptic species of the C. militaris complex.

For the past several years we have been conducting a phylogenetic investigation of the genus Cordyceps and related entomopathogenic fungi. This research has involved several collecting trips to the southern Appalachians Mountains of North America, the richest region of United States with respect to Cordyceps species diversity (Mains, 1958). As part of this survey, we made numerous collections 49

of C. militaris sensu lato from both lepidopteran pupae and larvae. Here we report

that morphological, cultural, and molecular data are all consistent with there being two

phylogenetic species present among these collections. One is specific to lepidopteran

pupae and is consistent with C. militaris sensu stricto. The second is specific to

lepidopteran larvae and is more closely related to C. pseudomilitaris, but represents a

previously undescribed phylogenetic species of the genus Cordyceps. During this

investigation, specimens of C. militaris-like, lepidopteran pathogen from Japan were

made available to us. Upon morphological and molecular investigations, these

specimens were found to be arguably conspecific with the undescribed pathogen of

lepidopteran from the southern Appalachians, supporting an east Asian-eastern

Northern American disjunct distribution.

Materials and methods

Field collection

Collecting expeditions were conducted during July of 1999 and 2000.

Specimens were collected with host when possible. Notes of stromatal color and size

were taken on fresh specimens. Individual perithecia were dissected from the stromata

and ascospores were shot onto Potato Dextrose Agar (PDA) amended with Ampicillin to a final concentration of 100 jig/ml. All cultures were deposited in the ARS collection (ARSEF) and the Centraal Bureauvoor

Schimmelcultures (CBS). Specimens were dried on a food dehydratoron low heat 50 with air movement. Dried specimens were placed in sealed plastic bags with silica gel and stored in specimen boxes for transport and shipment. Four specimens from Zaita- cho, Mitoyo-gun, Kagawa Prefecture, Shikoku Island, Japan of a C. militaris-like fungi were also sampled. These latter specimens are also pathogens of lepidopteran larvae and were selected due to their morphological similarity to the C. militaris-like species from the southern Appalachians and existing preliminary molecular data from one specimen. All specimens were deposited in the Mycological Collection of the

Oregon State University Herbarium (OSC).

Morphological observations

For examination of ascus and ascospore characteristics, dried specimens were rehydrated and mounted in sterilized water. Microscopic examinations of anamorphic characters were made from cultures maintained on Potato-Carrot Agar (PCA) at 24 C in the dark for 10 d. Conidiophore branching and conidial arrangementwere observed as described by Zare and Gams (2001) and slide preparations were mounted in lactic acid-cotton blue. Twenty-five measurements were made for asci, ascospores, phialides and conidia. To examine the structure of perithecia and their attachment to stromata, dried specimens were fixed in FAA (five parts stock formalin: five parts glacial acetic acid: 90% parts 50% ethanol) and dehydrated ina graded ethanol series.

Samples were transferred from 95% ethanol into 1:1, plastic infiltration solution: 95% ethanol under vacuum. After 12 hours, samples were vacuum-infiltrated with straight solution. Samples were embedded in Historesin (Leica) GMA (glycol methacrylate) 51

plastic and sectioned (5-6 tm) on a rotary microtome with steel knives. Sections were

stained in 0.5% TBO (Toluidine Blue 0). All microscopic examinations were made

using either bright field or Nomarski DIC on a Leica DMRB compound microscope.

DNA extraction, PCR, andsequencing.DNAwas extracted from cultures by a

modified CTAB method as previously described in Sung et al. (2001). Approximately

1150 base pairs (bp) of the nuclear small subunit (SSU) ribosomal DNA (rDNA) and

ca 950 bp of nuclear large subunit (LSU) rDNA were amplified using PCR.

Amplifications and sequencing reactions were performed as described in Sung et al.

(2001).

Phylogenetic analyses

Sequences of nuclear SSU rDNA and LSU rDNA from 35 taxa were selected

from the major phy!ogenetic groups of the Clavicipitaceae based on the result of

Nikoh and Fukatsu (2000) and Sung et al. (2001). Three representative taxa of the

Hypocreaceae were used as outgroup. Species included in this study and their

GenBank accession numbers of rDNA sequences (SSU rDNA; LSU rDNA) were as

follows: Aphysiostroma stercorarium Barassa et al. (U32398; U47820), schweinitzii (Fr.) Sacc. (L36986; U47833), Sphaerostilbella aureonitens (Tu!.) Seifert

et al. (U32415; U00755), Atkinsonella hypoxylon (Peck) Diehi (U44034; U57087),

Balansia aristidae Atkinson (U44035; U57677), Balansia henningsiana (Moe!!.)

Diehi (U44036; U57678), (Bals.) Vuill. (AB027336; AB027382),

Claviceps paspali F. Stevens & J. G. Hall (U3240 1; U47826), Clavicepspurpurea 52

(Fr. :Fr.) Tul. (U44040; U57085), Cordycepioideus bisporus Stifler (AF00965 1;

AF009654), Cordyceps bfusispora 0. E. Eriksson (AF339571; AF339521), C. capitata (Fr.) Link (AB0273 18; AB027364), C. coccidiicola Kobayasi & Shimizu

(ABO31 195; ABO31 196), C. cochlidiicola Kobayasi & Shimizu (AB027331;

AB027377), C. heteropoda Kobayasi (AB027327; AB027373), C. militaris

(AF049146; AF049166), C. militaris (AB027333; AB027379), C. ophioglossoides

(Ehrh.) Link (AB027321; AB027367), C. pseudomilitaris (AF327394; AF327376), C. roseostromata Kobayasi & Shimizu (AF339573; AF339523), C. scarabaeicola

Kobayasi & Shimizu (AF339574; AF339524), C. sobolfera Berk (AB027328;

AB027374), C. takaomontana Kobayasi (AB04463 1; AB044637), Cordyceps sp.

(AB027332; Ab027378), Epichloe amarillans J. F. White (U35034; U57680),

Epichloe typhina (Pers.) Tul. (U32405; Ui 7396), Lecanicillium psalliotae (Treschow)

Zare & W. Gams (AF339607; AF339558), atramentosa (Berk. &

Curt.), Diehl (U441 15; U57084), Paecilomyces lilacinus (Thom) Samson (AF339583;

AF339534), Paecilomycesjavanicus (Friderichs & Balily) Brown & Smith

(AF339582; AF339533), Paecilomyces tenuipes (Peck) Samson (AB027334;

AB027380), Simplicillium lamellicola (F. E. W. Smith) Zare & W. Gams (AF339601;

AF339552), S. lanosoniveum (van Beyma) Zare & W. Gams (AF339603; AF339554),

S. obclavatum (W. Gams) Zare & W. Gams (AF339567; AF339517), Torrubiella confragosa Mains (AF339604, AF339555). GenBank accession numbers of sequences determined in this study are listed in the "Results" section. 53

Sequences were edited in SeqEd Ver. 1.0.3 and manually aligned in PAUP*

4.0 (Swofford, 2002) using a color font. Intron sequences for nSSU rDNA of four species (AB02733 1, ABO3 1195, AB027328 and AB027327) were deleted and the final alignment was deposited in TreeBase SN1280. PAUP* 4.0 was used to perform the maximum parsimony analyses on the combined data set of nSSU and nLSU rDNA. Heuristic searches were conducted using the following options: 100 replicates of random sequence addition, TBR branch swapping, and MulTrees option in effect.

Insertions and deletions are minimized and treated as missing data. Ambiguously aligned sequence regions were excluded from the data matrix before analysis.

Relative support of the resulting trees was determined by 1000 bootstrap replications on informative characters only with the previously mentioned search options

(Felsenstein, 1985). The phylogenetic trees generated from the combined data set with nSSU and nLSU rDNA data sets were rooted with A. stercorarium, H. schweinitzii and S. aureonitens.

Taxonomy

Examination of perithecium, ascus, ascospore, and anamorph characters combined with phylogenetic analyses of molecular data revealed that the C. militaris- like fungus which parastizes lepidopteran larva in the southern Appalachians of eastern United States and Japan does not match either C. militaris, C. pseudomilitaris, or other closely related species. Therefore, a new species, C. cardinalis, is proposed. 54

Cordyceps cardinalis G.-H. Sung & J. W. Spatafora, sp. nov.

Stromata solitaria vel plures, simplicia vel raro ramosa, in larvis

Lepidopterorum. Stipes carnosus, ochraceo-aurantius ye! cardinalis, cylindricus ye!

versus apicem amplificatus, 4-50X0.5-1.5 mm. Area fertilis terminalis, armeniaca

vel cardinalula, cylindrica, ellipsoidea ye! fusiformis, 2-9X1-4 mm. Perithecia

congesta, laxe inclusa, in orientatione ordinalia, ellipsoidea, fusiformia ye! obclavata,

230-540X110-240 Jim. Asci 8-spori, hyalini, cylindrici, 175-330X3-5 Jim, apice

prominenti. Ascosporae laeves, fihiformes, hyalinae, irregulariter multiseptatae, into

partisporas non rumpentes, 160-320X1 tm.

Stromata solitary or occasionally several, simple or rarely branched, on larvae

of Lepidoptera. Stipe fleshy, ochraceous orange to red, cylindrical to enlarging

apically, 4-50X0.5-1.5 mm. Fertile area terminal, reddish orange to reddish,

cylindrical, elliptical to fusiform, 2-9X1-4 mm. Perithecia crowded, loosely

embedded, ordinal in orientation, elliptical to fusifonn to obclavate, 23 0-540X110-

240 tm (Figure 3.1 E, F). Asci 8-spored, hyaline, cylindrical 175-330X3-5 j.tm,

possessing a prominent apex (Figure 3.1 B). Ascospores smooth, filiform, hyaline,

irregularly multiseptate, not fragmenting into part-spores, 160-320X1 jim (Figure

3.1 A).

Knowndistribution.Southern Appalachian Mountains of eastern United States,

southeast Japan

Specimens examined.USA.TENNSSEE: Great Smokey Mountains National Park,

Hen Wallow Trail, 19 Jul 2000, OSC 93609 (HOLOTYPE), OSC 93610, OSC 93611, 55 //

Figure 3.1. Morphology and ecology of C. cardinalis. A. Irregular septation of ascospores. B. Distal end of ascus showing pronounced ascus cap. C. Proximal end of ascus showing ascus foot characteristic of the basal fascicle. D. Host of C. cardinalis, larval stage of Archolophinae lying in its web. E. Cross section of stroma, showing presentation of perithecia on a stroma. F. Enlargement of single perithecium with an ascus projecting through the ostiole. G. Typical stromata of C. cardinalis fruiting from a larva amongst moss and forest litter. (Figure 3.1 A-C, Figure 3.1 E, F) in Nomarski interference contrast, Figure 3.1 D with a digital Nikon 990 COOLPLX, Figure 3.1 G with a 35mm Cannon EOS-ID with a 100mm macro lens. Scale bars in Figure 3.1 A-C = 4.5 jim; Figure 3.1 D =1.5 mm; Figure 3.1 E = 87.5 jim; Figure 3.1 F = 30.5 jim; Figure 3.1 G = 2.25 mm.

USC 93612, OSC 93613; Roan Mountain State Park, 25 Jul 2000, USC 93614, USC

93615, OSC 93616, OSC 93617, OSC 93618, OSC 93619. JAPAN. KAGAWA:

Shikoku, 23 Jun 1997, OSC 93619, USC 93620, USC 93621, USC 93622.

Etymology .-Red, in reference to both the color of the stroma and the pigment produced in culture.

Commentary.-Cordyceps cardinalis is most frequently collected as a single, small, reddish stroma emerging from lepidopteran larvae buried in the upper 1-2 cm of soil or from under well-developed moss mats (Figure 3.1 G). The host is frequently enclosed in a web or silken matrix (Figure 3.1 D). In ascus maturation, young asci are 56

Figure 3.2. Line drawing of variation in anamorph of C.cardinalis.Variation in phialide morphology ranges from simple Si,nplicillium-like phialides to branched verticillate Lecanicillium-likephialides. Conidium ontogeny is basipetal with conidia arranged in partially imbricate chains. Scale bar in Figure 3.2 = 10.0 Jim.

characterized by a pronounced apical cap or apex that possesses a distinct canal. In

the mature ascus, the apical cap becomes less pronounced, flattened and the canal is

usually not as apparent (Figure 3.1 B). The arrangement of ascospores within asci was

more or less parallel for the entire length of the asci, indicating that the ascospores

were of approximately the same length as the asci (Figure 3.1 C). The non-

disarticulating ascospores of C. cardinalis are irregularly septate with the size of each

segment ranging from 4-13 urn long (Figure 3.1 A). In the conidial stage of C.

cardinalis, cultures are moderately fast growing in PDA, attaining approximately 25 ±

8 mm at 18 C for 10 d. Aerial mycelium is cottony in texture and whitish to whitish yellow and the reverse side of cultures is reddish. Phialides are solitary or in whorls of

2 or 3 and swollen at the base or slightly flask-shaped, measuring 7-25X1.5-3 J.tm wide near the base and tapering to 0.5-1.3 tm at the apex (Figure 3.2). Conidia are 57 ellipsoidal to elliptical, measuring 4-6X1.5-2.5 tm, and are produced in long divergent sympodially imbricate chains (Figure 3.2). Chiamydospores are absent.

The mitotic stage is best described as being Clonostachys-like or Mariannaea-like. In comparison with other species, C. cardinalis differs from C. militaris in color and size of stromata, disarticulation of ascospores, type of anamorph produced in culture, and developmental stage of host in which stromata are produced.Cordycepscardinalis is most similar to C. pseudomilitaris in both morphology and ecology but differs in size of asci, ascospores, type of anamorph, and in their color characteristics in culture.

Results

DNA sequencing

PCR amplification yielded ca 1150 bp of nSSU rDNA and ca 950 bp of nLSU rDNA with no introns detected in the specimens sampled as part of this research. Ten new sequences of C. cardinalis and closely relatedCordycepsspecies obtained from

ARSEF and CBS culture collections were determined as part of this research and deposited in GenBank. Their GenBank accession numbers as well as their voucher numbers (nSSU rDNA; LSU rDNA; Voucher number) are as follows: C. cardinalis

(AY184973; AY184962; OSC 93609), C. cardinalis (AY184974; AY184963; OSC

93610), C. cardinalis (AY184975; AY184964; OSC 93619), C. cardinalis

(AY184976; AY184965; OSC 93620), C. militaris (AY184977; AY184966; OSC 58

93623), Mariannaeapruinosa (Liang) (AY184979; AY184968; ARSEF 5413),

Torrubiella wallacei H. C. Evans (AY184978; AY184967; CBS 101237).

Host identfication.Entact or nearly entact hosts were collected with approximately

one-quarter of the specimens. Hosts that retained enough characters for identification

were identified as members of the subfamily Archolophinae (Lepidoptera: Tineidae)

and were often enclosed in a loosely woven web or silken matrix. In contrast, the

hosts of C. militaris s.s., which were collected throughout the same region, were all

pupae of the lepidopteran families Arctiidae and Hesperiidae.

Phylogenetic analyses

The combined alignment of the nuclear SSU and LSU rDNA dataset is 2,017 bases in length, 1,099 from the NS1/NS4 region of the SSU rDNA and 918 from the

LRORILR5 region of LSU rDNA. Of these, 111 SSU rDNA positions and 86 LSU rDNA positions were ambiguously aligned or contain an excess of missing data near the 5' and 3' ends. These positions were excluded and the final data set included 1,820 nucleotide positions. One hundred and eighty-seven characters were parsimony informative, 75 from the SSU rDNA and 112 from the LSU rDNA. Maximum parsimony analysis of the 42 taxa dataset yielded 6 equally most parsimonious trees of

613 steps with consistency indices (CI) of 0.4078, and retention indices (RI) of

0.7105. One of 6 equally parsimonious trees is shown in Figure 3.3 with nodes that collapse in the strict consensus denoted with asterisks. The topology of the most parsimonious trees was globally similar to the results of a larger phylogenetic analysis 59

Aphysiostroma stercorarium Hypocrea schweinitzii Sphaerostilbella aureonitens 71 Cordyceps ophioglossoides

Paediomyces lilacinus Cordycepioideus bisporus 83 Cordyceps cochlldiicola 70 Cordyceps coccidllcola Cordyceps sobolifera Cordycep hetempoc LAtkinsonella hypoxylon Myriogenospora atramento

I 921 Balansia aristidae rrycepscapiateBalansia henningsiana 100 Epichloe amarillans Epichloe typhina 90 CIa viceps purpurea Claviceps pasp a/i Cordyceps militaris 100Cordyceps miitaris (USA) (Lecenici/hum) Cordyceps roseostromata (Korea) * Cordyceps militaris(Japan) 78Beauveria bassiana 100 * Cord ycops scarabaeicola (Beauveria) Cordyceps bifusispora 100 Paediomycestenuies 8 Cordyceps takaomontana (Paediomyces) * Mariannaea pruino * Paeci/omycesjavanicus Torrubiella confragosa (Lecanicililum) Paediomyces farinosus Lecanicillium psalliotae * 9Cordyceps cardinals (USA) 100Cordyceps cardinals (USA) Cordyceps cardinalls (Japan) (Clonostachys or Marianna) * 99 Cordyceps cardinalls (Japan) 89 Cordyceps cardinalis (Japan) 88 Cordyceps pseudomilitaris (, Simpilcililum) Torrubiella wallacei (Simpilcililum) Simpilcililum obclavatum Simpilcillium lanosoniveum 5 changes Simplidiilum lam ellicola

Figure 3.3. One of 6 most parsimonious trees from the maximum parsimony analysis of combined SSU and LSU rDNA data. Asterisks denote branches that collapse in a strict consensus. Numbers above nodes are nonparametric bootstrap values from 1000 replications. of the Clavicipitaceae with more intensive taxon sampling (Sung et al., 2001).

Cordyceps was inferred to be polyphyletic and consisted of two separate clades. One

of the clades includes C. militaris and other brightly colored Cordyceps species and

will be referred to as Cordyceps s.s. dade. The second dade includes C.

ophioglossoides and other darkly pigmented Cordyceps species and will be referred to

as the C. ophioglossoides dade. All C. militaris-like species (C. militaris, C.

roseostromata, C. pseudomilitaris and C. cardinalis) were placed within strongly

supported Cordyceps s.s. dade. Within this dade, C. cardinalis is distantly related with C. militaris and formed a statistically well-supported group with C. pseudomilitaris.

Discussion

Morphological comparisonofC. cardinalis with other C. militaris-like species

Cordyceps cardinalis is macroscopically similar to numerous species of

Cordyceps with the two best examples for purpose of comparison being C. militaris, and C. pseudomilitaris. These three species differ, however, in several microscopic, cultural, and ecological characteristics, which in this case, are indicative of phylogenetic species. The characteristics of C. militaris and C. pseudomilitaris are summarized according to the previous studies of Mains (1958) and Hywel-Jones

(1994) with the characteristics for differentiating these three species listed in Table

3.1. In overall size of stromata, C. militaris is relatively large compared to C. 61 Table 3.1. A morphologicalSpecies comparison of C. cardinalis with C. militaris-like species. Host Rhizomorph Stromata Perithecium Asci size C. cardinalis Lepidoptera larva Present X6-59size (mm) 0.5-4 230-540X size (jim) 110-240 X175-330 (jim) 3-5 AscosporeX160-320size (jim) I multiseptateIrregularlyAscosporeseptation NAsizePartspore (jim) reddish(reverse)Colony C. pseudomilitaris Lepidoptera larva Present 15-30x 0.9-4 290-570x 120-245 290-410x 5-6 280-390x 1 * Multiseptate NA Cream C. militaris LepidopteraMainly pupa Absent * X8-70 1.5-6 500-720X 300-480 300-5x 35_5 10 Multiseptate 2-4.5X 1-1.5 yellowWhitish * C. rosea Lepidoptera larva * X11 1 330-380x 160-230 X100 2-5 X120 1-1.5 8-10 septate NA * C. washingtonensisroseostromata ColeopteraLepidoptera larva pupa Present 1.2-5x15-30 2-6 280-300480-644x 252-386 300-418x 3-3.5 * 80-1x 1-1.510 * Multiseptate NA * Information not provided in original description; NA = not applicable, ascospores do not disarticulate.X1.5-2.2 x 140-160 4-5x 1 62

cardinalis and C. pseudomilitaris and there is slight, but observable, difference in the

color of stromata, at least in the southern Appalachians. C. militaris is more of an

orange to orangish-red, whereas C. cardinalis and C. pseudomilitaris are orangish-red

to red. We had previously interpreted the characters of stromatal size and color to be

variable traits within C. militaris that may have been linked to ecology (e.g., host) or

simply represented intraspecific variation. This interpretation of C. militaris in the

southern Appalachians was not supported by the data presented here (Figure 3.3).

However, the integration of C. roseostromata amongst C. militaris isolates in the

molecular analyses (Figure 3.3) suggests that there may well be greater color variation

within C. militaris in other parts of the world (e.g., east Asia).

Characteristics of perithecia, asci, ascospores and anamorphs were consistent

with stromatal traits in separating C. cardinalis and C. pseudomilitaris from C.

militaris (Table 3.1). The most easily distinguishable trait being size and morphology

of ascospores. The ascospores of C. cardinalis and C. pseudomilitaris do not

disarticulate into partspores as do those of C. militarislike many species of

Cordycepsand they are also smaller than those of C. militaris. Although the

ascospores of C. cardinalis do not disarticulate, they are septate, but the septations are

irregular in their spacing and unlike the condition found in C. militaris.

Anamorphic and cultural differences among C. cardinalis, C. militaris and C.

pseudomilitaris. 63

The anamorph of C. militaris in culture is a Lecanicillium, which exhibits

fairly standard verticillate awl-shaped phialides and conidium production. The

anamorph of C. pseudomilitaris consists of simple phialides resembling a solitary

phialide of Lecanicillium or Simplicillium-like species. In contrast, the anamorph of

C. cardinalis exhibits a solitary or verticillate phialide that is swollen at the base and

its conidia are produced in partially imbricate chains (Figure 3.2). This latter

morphology is somewhat intermediate between the anamorph genera Clonostachys

Corda and Mariannaea Arnaud & Samson (Samson, 1974), which are described as possessing similarly shaped phialides with swollen basal portions and producing

conidia in imbricate chains. While the anamorph of C. cardinalis also produces

imbricate chains of conidia it differs from Mariannaea species due to its partially imbricate arrangement of conidia (Figure 3.2). The phylogenetic analyses are consistent with these differences and indicate that C. cardinalis does not form a monophyletic group with other clavicipitaceous species of Mariannaea (e.g., M pruinosa). The anamorph of C. cardinalis differs from typical species of

Clonostachys in that the phialides of C. cardinalis are produced more sparsely and do not form dense heads of phialides and chains typical of Clonostachys (Schroers,

2001).

While C. cardinalis and C. pseudomilitaris do differ in anamorph morphology, the most distinguishing nonmolecular trait is the color of the reverse side of the culture

(Table 3.1). Cordyceps cardinalis imparts a red pigment to the medium, whereas the reverse side of C. militaris and C. pseudomilitaris cultures are cream or yellowish 64 white. Among other clavicipitaceous species, L. psalliotae (Treschow) Zare & W.

Gams exhibits the same color in culture (Zare and Gams, 2001) as that of C. cardinalis. However, L. psalliotae possesses awl-shaped phialides and molecular data do not support its conspecificity with C. cardinalis or any of the other known teleomorphs of Cordyceps (Figure 3.3).

PhylogeneticsofC. cardinalis

The main objectives of the phylogenetic analyses presented here were to determine the phylogenetic affinity of C. cardinalis to C. militaris and C. pseudomilitaris, to test the taxonomic utility of the aforementioned morphological and ecological traits in distinguishing closely related species of Cordyceps, and to determine if the morphologically and ecologically similar specimens from eastern

North America and Japan were conspecific. Although the Clavicipitaceae is a morphologically and ecologically unique family of fungi, molecular phylogenetic analyses have never strongly supported or rejected its monophyly (Spatafora and

Blackwell, 1993; Suh et al., 1998; Gams et al., 1998; Sung et al., 2001). These analyses presented here are not designed to test the monophyly of the family and should not be interpreted as such. The three main clades of clavicipitaceous fungi inferred as part of this study are consistent with phylogenetic hypotheses based on more extensive taxon sampling, but the interrelationships of the three clades remains equivocal (Sung et al., 2001). These data do reject the monophyly of Cordyceps, which includes species that are members of at least two clades of the Clavicipitaceae 65

(Figure 3.3): the Cordyceps5.5.dade and the C. ophioglossoides dade (Sung et al.,

2001). Cordyceps s.s. dade is defined by the inclusion of C. militaris, the type of the

genus, and numerous closely related and morphologically similar species. The C.

ophioglossoides dade includes many well-known fungal and arthropod pathogens that

cannot be accommodated in a monophyletic generic concept of Cordyceps.

Ultimately a complete generic revision of Cordyceps is needed, however it is a

problem beyond the scope of this study and will require a considerable amount of data

from independent genes to confidently address the issue.

Cordyceps cardinalis is well supported as a member of the monophyletic

Cordyceps s.s. dade and thus it is appropriate for its nomenclature to reflect this

phylogenetic affinity. The molecular data and analyses also support a clear separation

of C. cardinalis from C. militaris and a close relationship of it with C. pseudomilitaris.

These findings are consistent with stromatal color, disarticulation and morphology of

ascospores, cultural characteristics, and stage of host attacked by the pathogen as all being informative traits in distinguishing the three phylogenetic species. While C. cardinalis and C. pseudomilitaris are morphologically and ecologically very similar, and while they are each other's closest relatives in these analyses, these data are consistent with two separate species. Although sequence from only one specimen of

C. pseudomilitaris is available for comparison, the monophyly of the isolates sampled from C. cardinalis, the amount of sequence divergence between C. cardinalis and C. pseudomilitaris as compared to other species of Cordyceps, and the distinguishing morphological traits are sufficient evidence for recognition of distinct species. Cordyceps pseudomilitaris is only known from Thailand and C. cardinalis is only known from the southern Appalachians of eastern United States and southeast Japan.

The molecular divergence between these two taxa is in stark contrast to C. militaris, which also includes specimens from Asia and eastern North America as well as other continents (Figure 3.3).

Biogeography ofC. cardinalis

The east Asian-eastern North American disjunct species distribution is well documented in plants (reviewed in: Qian and Ricklefs, 2000; Tiffiiey, 1985; Wen,

1999), but relatively understudied in fungi (Vilgalys and Sun, 1994, Wu and Mueller,

1997, Wu et al., 2000). The current known distribution of C. cardinalis (Japan and the southern Appalachians of United States) is consistent with that of an east Asian- eastern North American disjunct species distribution; however, for two main reasons, we offer this as a hypothesis that is in need of further testing. First, additional data are needed from existing "C. militaris" collections as numerous cryptic or previously unrecognized species are likely to be housed in herbana under the name C. militaris.

Cordyceps militaris is one of the more frequently collected species of Cordyceps and many collections were placed in herbaria without having been subjected to microscopic examination. Second, much of what we know about fungal geographic distributions is more reflective of where mycologists have collected and not a systematic survey of geographic regions of the planet. There are numerous regions where Cordyceps has not been collected in a systematic manner and our current 67

understanding of the biogeography of the genus must take this lack of knowledge into

account.

Conclusions

Numerous existing names and species delimitations of Cordyceps were

considered for possible application to C. cardinalis. Descriptions and examination of

specimens of C. washingtonensis did not match morphologically (Table 3.1) and

sequence data are not available. Descriptions of C. rosea are similar to both C. pseudomilitaris and C. cardinalis, but differed in size of perithecia, asci, ascospores

and its 8-10 septated ascospores (Kobayasi and Shimizu, 1982, Table 3.1). Cordyceps

rosea is only known from the type collection, whose whereabouts are unknown and

thus its utility in Cordyceps taxonomy is extremely limited. Finally, C. roseostromata

is a morphologically similar species that grouped strongly among C. militaris

sequences. This finding is consistent with its original morphological description,

especiallyinthe size of partspores, but differ in the size of the perithecium (Kobayasi

and Shimizu, 1983; Table 3.1). This suggests that either it is potentially conspecific with C. militaris and indicates its variability in the size of penthecium or the cultures that we used are not indicative of C. roseostromata sensu Kobayasi. Additional research is required before we are able to address the taxonomic status of C. roseostromata. In discovering and describing C. cardinalis we attempted to apply all known names to this species, however, none of them proved to be suitable in comparison of morphological and/or molecular data.

Acknowledgements

We would like to acknowledge K. Cook for the assistance in preparing and sectioning of stromata, Drs. R. A. Humber and W. Gams for providing ARSEF and

CBS cultures, Dr. J. Trappe for providing a Latin diagnosis, Dr. D. Murray for assistance in identifying the lepidopteran host of C. cardinalis, H. Manabe for providing collections of Japanese specimens and photographs of C. cardinalis, and the

National Science Foundation for its financial support (DEB-9806936 and DEB-

0129212) to J. W. Spatafora.

References

Artjariyasripong, S., J. I. Mitchell, N. L. Hywel-Jones, and E. B. Garath Jones. 2001. Relationships of the genus Cordyceps and related genera, based on parsimony and spectral analysis of partial 18S and 28S ribosomal gene sequences. Mycoscience 42:503-517.

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

Gams, W., K. O'Donnell, H.-J. Schroers, and M. Christensen. 1998. Generic classification of some more hyphomycetes with solitary conidia borne on phialides. Canadian Journal of Botany 76:1570-1583. Hywel-Jones, N. L. 1994. Cordyceps khaoyaiensis and C. pseudomilitaris, two new pathogens of lepidopteran larvae from Thailand. Mycological Research 98:939-942.

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku (Section B, no. 84) 5:53-260.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364.

Kobayasi, Y and D. Shimizu 1982. Cordyceps species from Japan 5. Bull. Natn. Sci. Mus., Tokyo, 5cr. B. 8:111-123.

Kobayasi, Y and D. Shimizu 1983. Cordyceps species from Japan 6. Bull. Natn. Sci. Mus., Tokyo, 5cr. B. 9:1-21.

Mains, E. B. 1947. New and interesting species of Cordyceps. Mycologia 39:535-545.

Mains, E. B. 1957. Species of Cordyceps parasitic on Elaphomyces. Bulletin of The Torrey Botanical Club. 84:243-251.

Mains, E. B. 1958. North American entomogenous species of Cordyceps. Mycologia 50:169-222.

Nikoh, N., and T. Fukatsu. 2000. Interkingdom host jumping underground: Phylogenetic analysis of entomopathogenic fungi of the genus Cordyceps. Molecular Biology and Evolution 17:629-638.

Qian, H. and R. E. Ricklefs. 2000. Large-scale processes and the Asian bias in species diversity of temperate plants. Nature 407:180-182.

Samson, R. A. 1974. Paecilomyces and some allied hyphomycetes. Studies in Mycology No 6.

Schroers, H.-J. 2001. A monograph of Bionectria (Ascomycota, Hypocreales, Bionectriaceae) and its Clonostachys anamorphs. Studies in Mycology No 46.

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes. I. The Clavicipitales-Hypocreales coimection. Mycologia 85:912-922.

Suh, S.-O., J. W. Spatafora, G. R. S. Ochiel, H. C. Evans, and M. Blackwell. 1998. Molecular phylogenetic study of a termite pathogen Cordycepioideus bisporus. Mycologia 90:611-617. 70

Sung, G.-H., J. W. Spatafora, R. Zare, K. T. Hodge, and W. Gams. 2001. A revision of Verticillium sect. Prostrata. II. Phylogenetic analyses of SSU and LSU nuclear rDNA sequences from anamorphs and teleomorphs of the Clavicipitaceae. Nova Hedwigia 72:311-328.

Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.OblO. Sinauer, Sunderland, Massachusetts.

Tifffiey, B. 1985. Perspectives on the origin of the floristic similarity between eastern Asia and eastern North America. J. Arnold Arboretum 66: 73-94.

Vilgalys, R., and B. L. Sun. 1994. Ancient and recent patterns of geographic speciation in the oyster mushroom Pleurotus revealed by phylogenetic analysis of ribosomal DNA sequences. PNAS 91:4599-4603.

Wen, J. 1999. Evolution of eastern Asian and eastern North American disjunct distributions in flowering plants. Annu. Rev. Ecol. Syst. 30:421-455.

Wu, Q.-X. and G. M. Mueller. 1997. Biogeographic relationships between the macrofungi of temperate eastern Asia and eastern North America. Canadian Journal of Botany 75: 2108-2116.

Wu, Q.-X., G. M. Mueller, F. M. Lutzoni, Y.-Q. Huang, and S.-Y. Guo. 2000. Phylogenetic and biogeographic relationships of eastern Asian and eastern North American disjunct Suillus species (Fungi) as inferred from nuclear ribosomal RNA ITS sequences. Molecular Phylogenetics and Evolution 17:37-47.

Zare, R., and W. Gams. A revision of Verticillium sect. Prostrata IV. The genera Lecanicillium and Simplicillium gen. nov. Nova Hedwigia 73:51-86. 71

Chapter 4

A MOLECULAR PHYLOGENETIC STUDY OF THE CLAVICIPITACEAE (ASCOMYCOTA, FUNGI) BASED ON MULTIPLE INDEPENDENT LOCI

Gi-Ho Sung and Joseph W. Spatafora

Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR

9733 1-2902, USA

For submission to Molecular Phylogenetics and Evolution or Systematic Biology 72

Abstract

Multi-gene phylogenetic analyses were conducted to address the evolution of the Clavicipitaceae. Data are presented here for approximately 5,900 base pairs from portions of seven loci: the nuclear ribosomal small and large subunit DNA(nrSSUand nrSSU), fi-tubulin,elongation factorI a (EF-la),the largest and second largest subunits of RNA polymeraseII (RPBJandRPB2)and mitochondrial ATP synthase subunit 6 (mtATP6). Separate phylogenetics analyses, with data partitioned according to genes, produced conflicting results. The results of separate analyses fromRPBJand

RPB2are in agreement with the combined analyses that support three well-supported monophyletic clades in the Clavicipitaceae (i.e., Clavicipitaceae dade A, B and C), whereas the tree obtained from mtA TP6 is strongly conflicted in the monophyly of

Clavicipitaceae dade B and dose not the sister-group relationship of the Hypocreaceae and Clavicipitaceae dade C. Results from saturation analyses and ILD tests suggest the gene partitions are heterogeneous. The distribution of relative contribution of nodal support for each gene partition was made based on partitioned Bremer support

(PBS) values and bootstrap proportions from combinational analyses of all possible combinations of the seven gene partitions. These results suggest that combining heterogeneous gene partitions, which individually support a limited number of nodes, results in increased support for overall tree topology. Furthermore, identifying the localized incongruence of mtA TP6 in this relationship improves its phylogenetic confidence and provides evidence rejecting the monophyly of the Clavicipitaceae. The 73 data presented here also reject the current subfamilial classification of the

Clavicipitaceae. Although the monophyly of the grass-associated subfamily

Clavicipitoideae (e.g., Claviceps, Balansia and Epichloë) is strongly supported, species of the subfamily Cordycipitoideae (e.g., Cordyceps and Torrubiella) are members of all three recognized clavicipitaceous clades. In particular, species of the genus Cordyceps, which are pathogens of arthropods and truffles, are found in all three clavicipitaceous clades. These results imply that most characters (e.g., host affiliation and the morphology of ascomata) used in the current familial classification of the Clavicipitaceae are not diagnostic of monophyly. 74

Introduction

The Clavicipitaceae (Ascomycota, Hypocreales) is a large group of fungi with more than 35 genera (Diehi, 1950; Rogerson, 1970; White et al., 2003). Its distribution is cosmopolitan in virtually all terrestrial ecosystems, including the boreal, tropical and temperate regions of the world. The family is easily recognized by its unique morphology of asci and ascospores (Diehl, 1950; Luttrel, 1955; Spatafora and

Blackwell, 1993). The asci are long and cylindrical with a pronounced apical apex and contain long, fihiform and multiseptate ascospores that often disarticulate into partspores at maturity (Diehi, 1950; Hywel-Jones, 2002; Rogerson, 1970). Genera within the Clavicipitaceae represent ecologically diverse groups of fungi that include pathogens of arthropods (e.g., Cordyceps and Torrubiella), parasites of truffles (e.g.,

Cordyceps), and endophytes and epiphytes of the grass family (e.g., Claviceps,

Balansia and Epichloe) (Diehi, 1950; Kobayasi, 1982; Rogerson, 1970).

The first attempt to classify the clavicipitaceous fungi in a subfamilial classification was by Gaumann (1926), who divided the family into three groups, including the -Ascopolyporus group [Oomyces Berk. & Broome is not currently classified in the Clavicipitaceae (Erikkson, 1981)], the Epichloë-Claviceps group, and the Cordyceps group. The classification, however, was simple and artificial with the three recognized groups based primarily on the macromorphology of ascomata (e.g., developmental morphology of stromata). Significant progress and more advanced biological insight into the family were incorporated by Diehl (1950) 75 resulting in the current subfamilial classification. Diehi divided the family into three subfamilies (Clavicipitoideae, Cordycipitoideae, and Oomycetoideae) based on morphology (i.e., developmental morphology of ascomata and asexual characters) with considerable emphasis on host affiliation. As a result, the current subfamilial classification implicitly assumes that host affinity is a diagnostic character; the

Clavicipitoideae includes all species of grass symbionts (e.g., Claviceps, Balansia and

Epichloë) and the Cordycipitoideae and Oomycetoideae contain all of the pathogens of arthropods and fungi (e.g., Cordyceps, Hypocrella and Torrubiella).

In addition to the ascomatal characters, asexual states in the life cycles of the

Clavicipitaceae have been considered taxonomically informative (Diehl, 1950; Hodge,

2003). For example, Diehi (1950) recognized three tribes in the subfamily

Clavicipitoideae (i.e., Clavicipiteae, Balansiae and Ustilaginoideae) primarily based on the characters in asexual states (e.g., the morphology of conidiophores and phialides). Currently, more than 40 asexual genera are hypothesized to be members of the Clavicipitaceae, although many of the species of these asexual genera (e.g.,

Beauveria, Lecanicillium and Paecilomyces) have not been definitively linked to sexual species and the full utility of asexual forms in the classification of the

Clavicipitaceae is not fully understood (Hodge, 2003).

Numerous phylogenetic analyses of the Clavicipitaceae have been conducted using ribosomal DNA sequences (Artjariyasripong et al., 2001; Gams et al., 1998;

Gernandt and Stone, 1999; Gleim et al., 1996; Rehrier and Samuels, 1995; Spatafora and Blackwell, 1993; Sung et al., 2001). Although these studies support the 76 monophyly of subfamily Clavicipitoideae, the monophyly of the Clavicipitaceae has been both supported and rejected. Furthermore, these studies included a limited taxon sampling that did not cover the morphological and ecological diversity of the family

(Gams et al., 1998; Glenn et al., 1996; Rehner and Samuels, 1995; Spatafora and

Blackwell, 1993). On the other hand, Sung et al. (2001) included a broad taxon sampling to represent the diversity of the family but could not address the monophyly of the Clavicipitaceae due to limited character sampling. Therefore, none of the phylogenetic studies based on ribosomal DNA offers a fully resolved hypothesis of the family, although they provide many points of phylogenetic resolution within the family.

After more than a decade of the reliance of fungal systematics on ribosomal

DNA, the growth of multi-gene phylogenies in fungal systematics has resulted in more robust phylogenetic hypotheses for numerous groups (reviewed in Lutzoni et al.,

2004). Castlebury et al. (2004) constructed a well-resolved phylogeny for the

Hypocreales based on five independent loci and demonstrated that the Clavicipitaceae is monophyletic. The taxon sampling in that study, however, included only eight clavicipitaceous species, and thus lacked the broad taxon sampling necessary to provide a robust phylogenetic hypotheis of the Clavicipitaceae. Therefore, we conducted a series of phylogenetic analyses with seventy representative taxa in the

Clavicipitaceae and DNA sequence data from seven independent loci: the nuclear ribosomal small and large subunit DNA,Ji-tubulin,elongation factor la, the largest 77 and second largest subunits of RNA polymerase II and mitochondrial ATP Synthase subunit 6.

With the advent of multi-gene phylogenies, particular emphasis has been placed on congruence or combinability of independent and possibility heterogeneous datasets, which may possess different underlying histories (e.g., gene duplication, lineage sorting and horizontal transfer) (Barker and Lutzoni, 2002; Bull et al., 1993;

Chippindale and Wiens, 1994; Levasseur and Lapointe, 2001). Statistical tests (e.g.,

Incongruence length difference test and Likelihood-ratio test) have been developed in parsimony and model-based methods to detect global level of incongruence among the heterogeneous data sets (Bull et al., 1993; Farris et al., 1994; Huelsenbeck et al., 1996;

Rodrigo et al., 1993). There has also been research directed at identifying the localized incongruence in particular regions of a tree (Baker et al., 1998; Gatesy et al.,

1999; Thornton and DeSalle, 2000; Wiens, 1998).

In this study, we improve the phylogenetic hypotheses of the relationships in the Clavicipitaceae by identifying the localized incongruence among heterogeneous data sets. By doing so we are able 1) to infer a well-supported phylogeny of the

Clavicipitaceae with increased taxon sampling using molecular sequences fromseven independent loci, and 2) to identify the distribution of nodal support and localized incongruence among data partitions.

Materials and Methods i11

SelectionofTaxa

Taxa of this study were selected to represent the morphological and ecological diversity of Clavicipitaceae. Representative taxa from other families in Hypocreales were also selected to test the monophyly of Clavicipitaceae. A total of 91 taxa were sampled from the representative families in Hypocreales: six taxa in Bionectriaceae, seven in Nectriaceae, five in Hypocreaceae and seventy in Clavicipitaceae. Taxa in the

Clavicipitaceae consist of 42 sexual taxa and 28 asexual taxa (Table 4.1). This study includes eight representative sexual genera and eleven asexual genera out of over 40 possibly linked asexual genera to the Clavicipitaceae (Table 4.1). Although the current taxon sampling does not cover all sexual and asexual genera in Clavicipitaceae, the most representative and ecologically diverse taxa were included and provide the means to evaluate and refine the higher level of relationships of the family.

Glomerella cingulata and Verticillium dahliae from Phyllachorales were chosen as outgroups based on previous phylogenetic analyses (Castlebury et al., 2004; Spatafora et al., 1998).

DNA isolation, PCR amplification, Cloning and Sequencing.

Total genomic DNA was isolated from both dried herbarium specimens and live cultures as previously described (Sung et al., 2001). To generate a robust phylogenetic hypothesis of Clavicipitaceae and the Hypocreales, one mitochondrial and six nuclear gene regions were amplified and sequenced. The nuclear regions sequenced were from nuclear ribosomal small and large subunit DNA (nrSSU and Table 4.1. The list of cultures and specimens used in this sequences available in GenBank. An asterisk (*) denotes the Taxon Specimen voucher a study. GenBank accession numbers are provided for the GeneBank accession numberdetermined by the first author of this study. AschersoniaAphysiostroma badia stercorarium ATCC 62321 AF543769 nrSSU * AF543792 nrLSU * fi-tubulin * AF543782 EF-la * AY489633 nrRPBJ * nrRPB2 * AY489566 mtATP6 * AschersoniaAschersonia placenta placenta * * * * * * * BalansiapilulaeformisBeauveriaBalansiahenningsiana caledonica ARSEFAEGGAM 97-2 16112 2567 AF3AF543764AY545723 39570 * AF3AF543788AY545727 39520 * AY489610 * * AY489643 * * AY489576 * BionectriaBionectria ochroleuca cf. aureofulva ATCCCBSGJS 71-328114056 26019 AY489684 * AY489176 * * AY48961 I * * * * * ClavicepspurpureaClavicepspaspaliClavicepsfusiformis GAMATCC 12885 13892 AF543765 * AF543789 * * AF543778 * AY489648 * * CordycepsCordycepsagriotaClaviceps aphodii purpurea ARSEF 54985692 * * * * * * * * CordycepsCordycepscapitata brunneapunctatacardinalis USCNHJ 71233 12565 AY489689 AY489721 * * AY489615 * AY489649 * * AY489581 * Cordycepscf.acicularisCordycepscardinalis N1-IJ0SC936100SC93609 12592 AY184974AY184973 * AY184963AY184962 * * * * * * * * * CordycepsgunniiCordycepsfractaCordyceps chiamycosporia USCCBS 10124476404110989 AF339572 * AF339522 * * AY489616 * AY489650 * * AY489582 * Cordyceps irangiensisheteropoda NHJNHJUSC 12575 12572106404 AY489690 * AY489722 * * AY489617 * * AY489651 * * CordycepsCordycepsmelonlontheCordycepsjaponica ,nilitaris OSC 93623110991110993 AY184977 * AY184966 * * * * AY545731 * * Table 4.1. (Continued) Cordyceps nutans USC 106405110994 AY489691 * AY489723 * * AY48961 8 * AY489652 * * AY489583 Cordyceps ophioglossoidesraveneliiscarabaeicola EFCCUSC 110995 5689 AF339574 * AF339524 * * * * * * CordycepsstylophoraCordycepssphecocephala USC 110999110998 * * * * * * * CordycepsCordycepstaii subsessilistuberculata USCARSEF 111002111001 5714 AF543763 * AF543787 * * AF543775 * * * * Cordyceps variabilisunilateralis ARSEFNHJ 12523 5365 114050 AY489702 * AY489734 * * * AY489629 * AY489667 * * * AY489596 * Epich/oëlyphinaEngyodontiumCosmospora aranearum coccinea ATCCCBS 56429 309.85 AF339576U32405 AF339526U17396 * AF543777 * AY489653 * * * AY489584 * GlomerellaGlemerella cingulata FAUCBSI 513 250.8214054 AF339588AF543762 U48427 AF543786AF339539U48428 * AF543772AF543773 * AY489659 * * AY489590 * Haptocillium zeosporumsinensebalanoides CBSCBS 335.80567.95 AF339589AF339594 * AF339540AF339545 * * * * * * * Hirsute/laHirsute/la sp. ATCCNHJ 1252512527 36093 AY545722 * AY545726 * * * * AY545732 * * HymenostilbeHydropisphaeraHydropisphaerapeziza aurantiaca erubescens NHJCBS 12574 102038 AY489698 * AY489730 * * AY489625 * AY489661 * * AY489591 HypocrellaHypocrea lutea schizostachyi GJSNHJATCC 8910412605 208838 AF543768 * AF543791 * * AF543781 * AY489662 * * AY489592 * LecanicilliumHypomycespolyporinusHypocrellasp. antillanum CBSATCC 532.81350.85 76479 AF339585AF543771 AF339536AF543793 * - AF543784 * AY489663 * * AY489593 * LeuconectriaLecanicilliumpsalliotae clusiae ATCCCBS 22228 101270 AY489700AF339609 * AY489732AF339560 * * AY489627 * AY489664 * * AY489595 * Table 4.1. (Continued) MelarhiziumMariannaeapruinosa album ARSEF 20825413 AY184979 * AYI 84968 * * * * * * MetarhiziumfiavovirideMetarhizium anisopliae var. minusmajus AEGARSEF 96-32 20373145 AY489701AF339580AF339579 AY489733AF339531AF339530 * AY489628AF543774 * AY489665 * * * NectriaMyrigenosporacf. Nectria cinnabarina sp. atramentosa CBS 478.75114055 U47842U32412 U17404U00748 * AF543785 * AY489666 * * * PaecilomycesfarinosusOphionectria trichospora OSCCBS 111006 109876 AF543766 * AF543790 * * AF543779 * AY489669 * * PaecilomycesPaecilomycesfarinosus tenuipes OSC 111007111005 891.72 AF339599 * AF339550 * * * * * * * PseudonectriaPochoniagonioides rousseliana CBS 101437114049 AF339584AF543767 * AF339535 U17416 * * * AF543776AF543780 * AY489670 * * AY489598 * RoumeguieriellaRotferophthora angustispora rufula EFCCGJSCBS 91-164 6279 346.85 * * * * * * * * SimplicilliumShimizuomycesShimizuomycesparadoxa lamellicola paradoxa CBSEFCC 6564 116.25 AF339601 * AF339552 * * * * * * * Simplicillium lanosoniveum CBSCBS 102308704.86101267 AF543770AF339602AF339603 AF339553AF339554U00756 * AF543783 * * * * TorrubiellaratticaudataTorrubiellaconfragosaSphaerostilbella berkeleyana ARSEFCBS 1915 101247 AF339604 * AF339555 * * * * * * * VerticilliumTorrubiella wallaceichiamydosporiumdahliae ATCCCBS 504.66101237 16535 AY489705AF339593AY184978 AY489737AYIAF339544 84967 * AY489632 * AY489673 * * AY489600 * ViridisporaVertjcjlliumVerticillium dipariezisporaincurvumepiphytum CBS 460.88384.81102797 AY489703AF339600AF339596 AY489735AF339547AF339551 ** AY489630 * AY489668 * * AY489597 * Corvallis,collection,FungalTypeaAEG, Culture A.Culture OR. E.GJS, GlennCollections, Collection, G. J.personal Samuels Manassas, Chuncheon, collection; personal VA; collection;Korea;ARSEF, CBS, GAM, Centraalbureau USDA-ARS NHJ, Julian Nigel H. Collection Hywel-Jones Millervoor Schimmelcultures, Mycological personal Herbarium collection; Utrecht, Athens, OSC,the ; Oregon GA; FAU, State EFCC, F. University A. Uecker Entomopathogenic personal of Entompathogenic Fungal cultures, Ithaca, NY; ATCC, American Herbarium, 83 nrLSU), /3-tubulin, elongation factor 1 a (EF-la), and the largest and second largest subunits of RNA polymerase II (RPBJ and RPB2). The mitochondrial region was from the protein-coding ATP Synthase subunit 6 (mtATP6).

The PCR amplification and sequencing of nrSSU and nrLSU were conducted according to the procedure previously described in Sung et al. (2001). The nrSSU was amplified with the primers NS1 and NS4 and sequenced with the primers NS1, NS3,

SR7 and NS4 (Table 4.2). The nrLSU was amplified with the primers LROR and LR5 and sequenced with the same primers (Table 4.2). In the amplification of EF-la, RPBJ and mtATP6, we followed the procedure of Castlebury et al. (2004). The EF-la was amplified with the primers 983F and 221 8R and was sequenced with the primers

983F, 2218F, CEFF2, CEFR1 and CEFR2 (Table 4.2). For the amplification of RPB1, the primers CRPB 1 and RPB 1 C were used for most taxa in the present study. The primer CRPB 1 A was used with the primer RPB 1 C instead of the primer CRPB 1 for the remaining samples. Sequencing of RPB1 was performed with the same primers used in the amplification. For mtATP6, the primers ATP6C1A and ATPC2A were used in PCR amplification and sequencing (Table 4.2).

The PCR amplification of RPB2 and /3-tubulin was performed in a PTC-100 cycler (MJ research) programmed as follows: 35 cycles of 94 °C for 1 mm, 50-53 °C for 30 sec, and 72 °C for 2 mm. Segments of RPB2 were amplified and sequenced with the primers fRPB2-5F and fRPB2-7cR (Table 4.2). Two forward primers, RPB2-

5F1 and RPB2-5F2, were designed on the sequence alignment of RPB2 and used for

PCR amplification and sequencing with fRPB2-7cR for the samples that did not Table 4.2. List of primers used in this study.

Genes Primers Sequence, 5'- 3' References Nuclear NrSSU NS1 5'-GTAGTCATATGCTTGTCTC- 3' White et al. (1990) SR7 5'-GTFCAACTACGAGCTFTTFAA- 3' White etal. (1990) NS3 5'-GCAAGTCTGGTGCCAGCAGCC- 3' White etal. (1990) NS4 5'-CTTCCGTCAATTCCTTTAAG- 3' White Ct al. (1990) NrLSU LROR 5'-GTACCCGCTGAACTFAAGC- 3' Vilgalys and Sun (1994) LR5 5'- ATCCTGAGGGAAACTTC- 3' Vilgalys and Sun (1994) /3-tubulin T12 5'-TAACAACTGCTGGGCCAAGGGTCAC-3' O'Donne!l and Cigelnik (1997) T22 5'-TCTGGATGTTGTTGGGAATCC-3' O'Donnel! and Cige!nik (!997) EF-la 983F 5'-GCYCCYGGHCAYCGTGAYTTYAT-3 Cast!ebury et al. (2004) 2218R 5'-ATGACACCRACRGCRACRGTYTG-3' Castlebury et al. (2004) CEFF2 5'-GGCTTCAACGTGAAGAACG-3' Castlebury et al. (2004) CEFR1 5'-CCGTKCAARCCRGAGATGG-3' Cast!ebury et al. (2004) CEFR2 5'-GRGGGTCGTI'CYTGGWGTC-3' Cast!ebury et al. (2004) RPBI CRPB! 5-CC WGGYTTYATCAAGAARGT-3' Castlebury etal. (2004) CRPB1A 5'-CAYCCWGGYTTYATCAAGAA-3' Castlebury et al. (2004) RPB1Cr 5'-CCNGCDATNTCRTTRTCCATRTA- 3' Cast!eburyet al. (2004) RPB2 fRPB2-5F 5'-GAYGAYMGWGATCAYTTYGG- 3' Liu et al. (!999) RPB2-5F1 5'-TKGC!GTJGGYATCAARCC- 3' In this study RPB2-5F2 5'-GGGGWGAYCAGAAGAAGGC- 3' In this study fRPB2-7cR 5'-CCCATRGCTTGTYYRCCCAT- 3' Liu et al. (1999) Mitochondrial mtATP6 ATPC1A 5'-AGAWCAATTYGAARTRAGAG-3' Cast!ebury etal. (2004) 5'-ACAAAYACTTGWGCTFGKATWAAJGC- ATPC2A Cast!ebury et al. (2004)

amplify with fRPB2-5F (Table 4.2).

The amplification and sequencing of /3-tubulin was performed with the primers

T12 and T22. In the course of sequencing of /3-tubulin, paralogous copieswere detected for three species of genus Simplicillium, and all five taxa in Hypocreaceae.

Using a TOPO TA cloning kit (Invitrogen Inc., Carlsbad, CA), we cloned and sequenced the paralogous copies of /3-tubulin. The orthologous copies of/3-tubulin were characterized and determined through phylogenetic analyses of the predicted protein sequences (Landvik et al., 2001).

All PCR products were purified using QlAquick PCR purification kits (Qiagen

Inc., Valencia, CA) according to the manufacturer's instructions. Sequencing was conducted on an Applied Biosystems (ABI) 3100 automated DNA sequencer using

AIBI Big-dye Ready-reaction kit (Applied BiosystemsInc.,Foster City, CA) following manufacturer's instruction.

P/iylogenelic Analyses

Sequences of nrSSU and nrLSU genes were edited using SeqEd 1.0.3 (Applied

Biosystems Inc., Foster City, CA) and aligned with the multiple sequence alignment option implemented in Clustal W (Thompson et al., 1994). The program was run with default settings and the alignment was further optimized by direct examination in

MacClade 4.0 (Maddison and Maddison, 2000). Nuclear and mitochondrialsequences of protein-coding genes were also edited using SeqEd 1.0.3 and manually aligned by direct examination. The assignment of codon positions was confirmed by translating nucleotide sequences into predicted amino acid sequences using MacClade 4.0.

Highly variable regions in nrSSU and nrLSU, which were ambiguously aligned,were excluded from phylogenetic analyses. Hypervariable protein-coding regions of RPBJ were also ambiguously aligned and excluded. The alignments from each gene were combined in MacClade 4.0 to generate a concatenated alignment. In separate analyses of each gene partition, maximum parsimony (MP) analyses were performed using PAUP* 4.0 (Swofford, 2002). All characters were equally weighted and unordered. Gaps were coded as missing data. Due to the low number of parsimony informative characters, two-step maximum parsimony analyses were performed to find the most parsimonious trees for thenrSSUandnrLSU

(Olmstead REF). Initially, heuristic searches were performed with 100 replicates of random sequence addition of tree bisection-reconnection (TBR) branch swapping and

Multrees OFF with a maximum of two trees saved per replicate. The resulting trees were used as starting trees in subsequent analyses with Multrees ON and saving no more than 5000 equally parsimonious trees. Nodal support was assessed with 100 replicates of nonparametric bootstrapping (Felsenstein, 1985), each with five replicates of random sequence addition and other settings identical to step-one of the maximum parsimony analyses. For protein-coding gene partitions, we performed a heuristic search with 100 replicates of random sequence addition of TBR branch swapping and Multrees ON. Nonparametric bootstrap values were obtained with 100 replicates, each with five replicates of random taxon addition, and other settings identical to the heuristic search.

To evaluate and identify the possible global incongruence among gene partitions, incongruence length difference (ILD) tests (Farris et al., 1994)were conducted using PAUP* 4.0. We excluded 25 taxa due to the missing data insome gene partitions and performed the tests with a core data set that comprises 66 taxa

(Table 4.1). A total of twenty-one gene-by-gene comparisons in ILD testswere conducted based on 1000 ILD replicates with 50 random sequence additions per replicate. In interpreting the results of ILD tests, recent studies have shown that the utility of the ILD test is limited as a measure of the incongruence among data partitions (Barker and Lutzoni, 2002; Dowton and Austin, 2002; Yoder et al., 2001).

Therefore, we used the ILD tests as a measure of heterogeneity between gene partitions and the results in ILD tests were not interpreted as a measure for a combinability test (Barker and Lutzoni, 2002).

The separate analyses of each gene partition were used to detect localized incongruence among gene partitions for higher-level relationships (Figures 4.1 and

4.2). Because the emphasis of this study pertains to higher-level relationships of the

Clavicipitaceae, the possibility of localized incongruence restricted to more terminal clades was not assessed. Congruence between gene partitions was assessed using the methodology proposed by Wiens (1998), whereby the level of bootstrap support was used to detect significant levels of localized incongruence among gene partitions. We interpreted a 70% bootstrap support (BP) as a strong support for a particular node and identified the conflicted nodes by comparing each gene partition with a threshold between conflicting (BP? 70%) and non-conflicting (BP70%) nodes (Hillis and

Bull, 1993; Wiens, 1998).

One significant localized conflict was detected between mitochondrial mtATP6 and two nuclear partitions(RPB1andRPB2)for a major node in combined analyses

(Figure 4.1). Two combined analyses with 7-gene and 6-gene data sets (w/ mtATP6vs. w/o mtATP6) were performed to address the influence or impact of mtATP6 on bootstrap support for the conflicting major node. In addition, we conducted the

combined analyses with two different taxon samplings (91 taxa vs. 66 taxa) to

investigate the potential problems associated with missing data (Wiens, 2003a; Wiens,

2003b). These four combined maximum parsimony analyses were performed with

1000 heuristic replicates of random sequence addition, TBR branch swapping, and

Multrees ON. Nodal support was evaluated by bootstrapping (Felsenstein, 1985) and

Bremer support (Bremer, 1988; Bremer, 1994). We used 1000 bootstrap replicates, each with five replicates of random sequence addition, with other settings identical to the heuristic search as previously described. Bremer support (BS) was calculated using

TreeRot 2 (Sorenson, 1998). To assess the effect of nucleotide compositional bias (see results) on phylogenetic inference, the phylogenetic analyses were conducted using the neighbor-joining method by employing log determinant distances (LogDet) in PAUP*

4.0 (Lockhart et al., 1994).

Bayesian Metropolis coupled Markov chain Monte Carlo (B-MCMCMC) phylogenetic analyses as implemented in MrBayes 3.0 (Huelsenbeck et al., 2000;

Huelsenbeck and Ronquist, 2001) were performed on the four combined data sets described above. We used ModelTest 3.06 (Posada and Crandall, 1998) to select the model and the parameters of DNA substitution that best fit each of the partitions.

Model selection was based on either the hierarchical likelihood ratio test (hLRTs)

(Huelsenbeck and Rannala, 1997) or Akaike Information Criteria (AIC) (Akaike,

1974). ThenrSSU, nrLSUand each codon of each gene were assigned its own model of evolution in Bayesian analyses. A recent simulation study (Lemmon and Moriarty, 2004) demonstrated that the parameter-rich models perform better in estimating likelihoods and posterior probabilities. As such, we selected the more parameter-rich model when hLRTs and AIC disagree in unequivocally selecting a model for a given partition. Bayesian analyses were performed for 10,000,000 generations, with four independent chains and sampled every one-hundredth tree. Two additional runs of

10,000,000 generations were performed to insure the consistent convergence on stable log-likelihoods. The first 10,000 trees were identified as the burnin and deleted from constructing consensus trees and calculation of posterior probabilities. A total of

90,000 trees were imported into PAUP* to construct a 50% majority-rule consensus tree (Figures 6.3 and 6.4).

To evaluate the influence of combining genes on nodal support of the higher- level relationships of the Hypocreales, ten major nodes were selected (Figures 4.2, 4.3 and 4.4). Nodal support was assessed by a series of bootstrap analyses, which we refer to here as combinational bootstrap analyses. For combinational bootstrap (CB) analyses maximum parsimony analyses were performed on 126 data sets, which represent all possible combinations of the seven gene partitions in datasets that contain from one to seven partitions. Nonparametric bootstrap values were calculated for each data set with the same options described above. To demonstrate the influences of character (gene) and taxon sampling, mean bootstrap supports were calculated by averaging the bootstrap support values from analyses that combined the same number of genes but in different combinations (e.g., 1 dataset of 7 gene partitions, 7 datasets for 6 gene partitions, etc.) with two different taxon samplings (the 91-taxon data set vs. the 66-taxon data set). For example, we can generate a total of 21 data sets in different combinations when we combine five genes from seven gene partitions. The mean bootstrap support for a given node was calculated by averaging the bootstrap values from MP analyses of these 21 data sets for that node.

Partitioned Bremer Support (PBS) analyses were conducted to localize the relative contribution of each gene partition and to verify the impact of mtA TP6 partition on the total support in combined analyses. PBS values of each gene partition were calculated for ten major nodes using the command file of constraint trees as implemented in TreeRot 2 (Sorenson, 1998). A total of 100 heuristic searches were performed and tree lengths were averaged if more than one parsimonious tree was found (Baker and DeSalle, 1997). We calculated PBS values for ten major nodes under two taxon samplings as described above, i.e., the 91-taxon vs. the 66-taxon datasets. The utility of PBS values has been questioned in recent studies (Danforth et al., 2004; Koepfli and Wayne, 2003). Therefore, we compared the results of CB and

PBS analyses to further test their applicability and to demonstrate the contribution of each gene partition to total nodal support. Out of 126 combinational bootstrap (CB) analyses, any one gene partition is present in only half or 63 of the CB analyses. For each gene partition, the percentage of 63 CB analyses, which have more than 70% bootstrap supports, was calculated for ten major relationships and used as indicators to measure the relative contribution of each gene partition to the total supports in combined analyses. For example, a gene partition that receives a CB of 50%means 91

that half of the datasets in which that gene partition is present results in bootstrap

values of >70% for the node in question.

Results

Characteristicsofeach gene partition

In PCR amplification of nrSSU, the NS 1 and NS4 amplicons are approximately 1kb. For 22 taxa, however, longer PCR products were amplified due to the presence of group 1 introns. Amplification of mtA TP6 using ATPC 1 A and

ATPC2A also produced longer than average PCR products for 11 taxa and group 1 introns were also inserted in mtATP6. The group 1 introns were deleted in the final alignment and were not analyzed further. After the exclusion of group 1 introns, the entire dataset consisted of 6,261 aligned nucleotide sites for the seven gene partitions

(Table 4.3). After the exclusion of ambiguously aligned sites, the alignment included

5,914 sites, 2,281 of which were parsimony informative (Table 4.3). The number of completed sequences differs among gene partitions (Table 4.1). A total of 66 taxaare complete for all seven gene partitions and 91 taxa are complete for at least threegene partitions. The 66-taxon data set (no missing data) was used for performing the ILD tests and comparing the phylogenetic results from the 91 -taxon data set (missing data).

In analyzing base composition of each gene and codon partition, the nucleotide frequencies of mtATP6 gene partition were A+T rich (Table 4.4), especially for the Table 4.3. Summary of data used in separate and combined MP analyses. No. of sites No. of observed No. of parsimony NrLSUNrSSU No. of taxa 91 in alignments 1,088768 variable sites. 247299 informative sites 177182 Tree length 903699 Tree number 3,7785,000 Consistency index (CI) 0.28130.3577 Retention index (RI) 0.71870.6473 RPBIEF-la/3-tubulin 889187 1,020616 514270 411237 3,6802,342 464 17 0.18970.1892 0.81030.5414 MrATP6RPB2 7583 1,048700674 424610409 335569370 4,2007,366 294 1 0.17520.155 0.50460.845 7-gene6-gene 91 5,9145,214 2,7732,349 2,2811,946 22,18619,7072,330 480 5 1 0.17940.25410.186 0.53950.61760.826 93

Table 4.4. Nucleotide base frequencies fornrSSU, nrLSUand different codon positions of each protein-coding gene and chi-square testsofnucleotide bias among taxa.

A C G T A+T Chi-square P value nrSSU 0.26972 0.21050 0.25821 0.26157 0.53129 48.97 1.0 nrLSU 0.26846 0.21040 0.31461 0.20653 0.47499 175.55 0.999 fi-tubulin (1) 0.22452 0.25478 0.3 1900 0.20171 0.42623 24.45 1.0 f3-tubulin(2) 0.24993 0.25553 0.16803 0.32652 0.57645 7.11 1.0 /3-tubulin (3) 0.02920 0.53442 0.22276 0.21362 0.24282 603.1368 <0.001 EF-Ia(1) 0.30056 0.19675 0.37169 0.13100 0.43156 31.12 1.0 EF-la(2) 0.30800 0.26720 0.15237 0.27243 0.58043 17.6489 1.0 EF-la(3) 0.02262 0.5 1992 0.22643 0.23 103 0.25365 579.29 <0.001 RPBJ (1) 0.28804 0.24834 0.35365 0.10997 0.39801 51.6 1.0 RPBJ (2) 0.34576 0.20854 0.20270 0.24301 0.58877 39.86 1.0 RPBI (3) 0.13 197 0.34600 0.26303 0.25901 0.39098 686.43 <0.001 RPB2(1) 0.26981 0.24691 0.33244 0.15084 0.42065 77.398 1.0 RPB2(2) 0.33438 0.21457 0.18152 0.26953 0.60391 33.06 1.0 RPB2 (3) 0.14983 0.32743 0.29845 0.22428 0.37411 1160.92793 <0.001 mtATP6(1) 0.33500 0.08974 0.22858 0.34668 0.68168 93.31 1.0 mIATP6(2) 0.18387 0.20019 0.14084 0.47510 0.65897 41.68 1.0 mtATP6(3) 0.52166 0.10371 0.03732 0.33731 0.85897 501.83 <0.001 All nuclear 0.23956 0.26840 0.26630 0.22541 0.46497 691.11 <0.001 All 0.24950 0.25580 0.25438 0.24045 0.48995 1052.57 <0.001

third codon (85.8 % A+T). In contrast, the third codon of/3-tubulinandEF-lawas

biased towards G+C content(/3-tubulin;24.2% A+T,EF-la;25.3% A+T). Although

the compositional biases were found in third codons of protein-coding genes, the base

frequencies (48.9 % A+T) of the combined dataset with all seven genes were not

biased (Table 4.4). A significant level of heterogeneity in the base composition was

detected among taxa in the combined dataset (P < 0.00 1) for the third codon partition

of five protein-coding genes using a chi-square test, but not for the first and second

codon positions (P = 0.999-1.00). This indicates that the base composition difference

among taxa in the combined dataset (P < 0.00 1) results from that of the third codon partitions of protein-coding genes. (A) nrSSU (B) nrLSU

Blon,cthac,a, (6) Blonectaiac.ae (6)

N,ct,iacea, (7) N.ct,laca,e (7)

EHypocleaceae (5) Hypocreac.a. (5) Clavicipilaceae (5) Clad. B Claviclpithc,a. (19) Clade C Claviclpllace.. (20) Claviclpitacea, 1) Clad. A Clad, B Clavicipilac,aa (24) ClavlcIpltacea. (24) Clad. A Clad. B

ClavlCipitac.a. (21) Clavlclpllaceae (26] Clad. B Clad, C

(C) P-tubulin (0) EF-la (E) RPBI Clavicipitac,ae (5) BlonaclIlac... (4) Clad. B 100 BiocecUiacaa, (6) Clavlcipitaceao (17) ______Clad. C Necfrlac.ae (7) Nectalacea. (1) N,cfrlac,a, (7) Claviclpltac.ae (13) Clad. B Blon.cUiacaa. (6) Clavicipllac.a. (3) Hypoc,oac.ae (4) 100 Clad, C Hypocceac.ae (5) Hypocaac,ae (3) Clavicipitac,a. (20) Clad. C Clavicipilacea, (24) I ClavlclpItacaa. (20) Clad, A Clad. C Clavicipilac.aa (24) Clad. A Hypocreac.a, (2) ClavIcIpItacea. (24) Clavicipitacaae (23) ClavicIpitacea. (9) ______Clad. A dade B Clad, B 1-1EClavlpllac.a. (25) t Clad. B (F) RPB2 (G) mtATP6 Bion.ctcac,a, (6) Bion.cfrlac.a. (4) 96 Clavicipltacaae (17) Clad. C Necl,iacaae (7) BlOn.ct,lac.a, (2)

Neclrlac.a, (7)

Clavicipitac.a. (23) Hypoc,eac.ae (2) dade B Clavlcipltac.ae (3) Clad, B Clavicipilac.a. (17) Clavlclpitac.a. (3) Clad, C Clad. B llypoaeaceae (3)

Hypocreac.ae (4) Clavlclpltac,a. (3) 92 Clad. C Clavlcipitac.ae (19) Clad. A Clavlclpllacea. (23) 98 Clad, A -i Claviclpllacaaa (9) Clad. B

Figure 4.1. Topological incongruence of the phylogenies from individual analyses ofseven gene partitions in maximum parsimony (MP) framework. Schematic diagrams of the phylogenies from seven genes (Figure 4.1 .A-G) are presented to show the topological incongruence in higher level of relationships based on the phylogeny from combined analyses in Figure 4.2. In each diagram, the number of taxa included in six major clades (i.e. Bionectriaceae, Nectriaceae, Hypocreaceae, Clavicipitaceae dade A, B and C) is provided in the parentheses. More than 70 % bootstrap values in MP analysesare presented below the corresponding nodes. The bootstrap values (BP? 70%) in the rectangular box correspondto the nodes that are in significant incongruence among gene partitions. 95

Separate analyses and testofincongruence

None of the seven separate analyses supported a monophyletic relationship of

Clavicipitaceae (Figure 4.1). Gross tree topologies of each gene partition are significantly different from one another and none is identical to that of combined data

(Figure 4.2). The separate analyses of nrSSU, nrLSU and /3-tubulin do not resolve any often major relationships in combined analyses and EF-la only resolves the

Nectriaceae (Figure 4.1). However, independent analyses of RPBJ, RPB2 and mtATP6 resolve several of the higher-level relationships (Figure 4.1). In particular, RPB2 provides the "best" phylogenetic signal among seven gene partitions and confidently resolves five major relationships with strong statistical supports (BP? 70%). The tree statistics of the separate analyses are presented in Table 4.3.

In tests to detect the localized incongruence for the ten major relationships identified in the simultaneous analysis (Figure 4.1), significant examples of localized incongruence were detected between mtA TP6 and two nuclear gene partitions (RPB1 and RPB2). The separate analyses of RPB1 and RPB2 strongly support the

Clavicipitaceae dade B (BP70%) in Figure 4.2, 4.3 and 4.4. These separate analyses indicate thatRPB]andRPB2gene partitions possess the phylogenetic signal to strongly support the Clavicipitaceae dade B in combined analyses (Figure 4.2). On the other hand, the separate analysis of mtATP6 strongly supports the grouping of six taxa in Clavicipitaceae dade B and three taxa in Hypocreaceae (BP = 87 % in Figure

4.1). This relationship is in strong conflict with the results from the combined analyses

(Figure 4.2). Except for this one strong localized incongruence among ten major Table 4.5. Incongruent length differences (ILD) for the pairwise comparisons of gene partitions. Numbers in the top of the matrix are the number of extra steps introduced by combining gene partitions. Numbers in the bottom of the matrix are the ILD measurements normalized by the length of the most-parsimonious trees for the combined data set of each pairwise comparison. Asterisks indicate the pairwise comparisons that are statistically significant (P < 0.05).

nrSSU nrLSU ,8-tubulin EF-la RPBJ RPB2 mtATP6 nrSSU 59* 70* 99* 72* 63* 71* nrLSU 0.001* 77* 102* 89* 100* 92*

/3-tubulin 0.001* 0.001* 84* 51 82* 71* EF-la 0.001* 0.001* 0.001* 86* 128* 135* RPBI 0.001* 0.001* 0.143 0.001* 50 87* RPB2 0.001* 0.001* 0.001* 0.001* 0.624 84* mtATP6 0.001* 0.001* 0.001* 0.001* 0.001* 0.001*

nodes, the nodes that are strongly supported in one separate analysis are also supported in others (BS ? 70%) and nodes that are discordant among the separate analyses are always weakly supported. In ILD tests, our results (Table 4.5) indicate that the degree of heterogeneity is significant among gene partitions. Only two (Ji- tubulin vs. RPBJ and RPBJ vs. RPB2), out of2l pairwise comparisons, show homogeneity between gene partitions.

The phylogenetic analysesofthe combined data set

MP analyses of the combined data set with 91 taxa and seven gene partitions

(the 91-taxon 7-gene data set) recovered five equally most parsimonious trees (Figure

4.2; Table 4.3). Although the separate analyses of each gene partition do not resolve ten major relationships, combining seven gene partitions provide well-supported phylogenetic estimates for most of the higher relationships except for the sister-group 97

GIom.r..IIa clngulata Glomer.Ila clnguleta VerticIIllum daltllao 100 BIo.motrIe of.,reofutra too lotBionectna ochroleuca Hydroplsphpaclra Bionectriaceae

oo floumeguMrIeIIa rufula 100 Rowoteguierfella tufula Psoudonecfrle roussellana 00 2 Ophl0000frla ttichospore 00 01 Leucon.ctrla cluslae Vfrldlspora diparletlspora Nectriaceae Nec ftc clnnabaflna cOCtInea of. Necfrfa op. Hyyce:polyporlrnos Spheerostllbella beo*.Ieyaoa Hypocreaceae I r- Aphyslosttoma sleruo,a,lum / ypocree lutea I Hypocreales too SImpUcIllium lamellicola SbmplIdilIlum Ianosonl ye urn SImplbcilbi urn Ianosonlveum Toroubl.11a tvallacel to Engyodontlurn aranearurn

Looanlcilhiurn antllbaooum Coodyc.ps cardlnalls Co,oiyceps cardlnalls Lecanlcllbun, psallbotae Clavicipitaceae dade C Lecanicllllun, psaIlIotae Cordyceps tub.rculata * Tom,bI.Ua confregosa I, lot Peedllomyces farinosus * I Peecllomyces fart nosus * I Beauverla eal.donlca lotCoodyc.ps scarabaelcola .,ii Paadflornyceo tenulpes 00 Cordyceps mISterm so Marbanneoa pnJInosa - Ascha,sonla badia - Hypocrolla schlaostachyi , Hypocrebla op. too Aschsosonla placenta boo Auchersonla placenta Shltniauomyc.o p.oradoaa Shirnlwomyces paradoora - Pochonle gonboides too Co,dyceps CS! ao,oydosporla too Pochonla oh! am ydosporlum Rotlferophthora angustlspora Malarhl,.lum album toot Meterhizlt.,m flavovlrid. Clavicipitaceae dade A tooCordyc.ps tall tooMemarhlzlum anlsopliao Veyticlblburn eplphyf urn lot Myriogenospora atramentosa lot Balansla henningslana too Balaotsla pita! aoforn, Is ar Eplchboe typhina toor CIa clomps panpali too too Claviceps fuslfarmis 00 Clavbcepo purpuroa too Clavlceos ourourea - Cordycops gunnhl Haptocibhbum sbnense LHaptodblllutl, balanoldes Haptoclhllut,t z005porum Cortiycops capbtata - Cordyc.ps fracta - CordycepaJaponica Cord ycaps ophiogbosoobdes - Cortiyoeps subsasshlls - Cordyceps rarlabblls Cordyceps ravenelti Cord yceps heferopoda Co,dycops t,tololonthao agPbota Clavicipitaceae dade B .. Cordycops otylopho.'e so Cotdyceps of. eoIcuIa,ta too Hlrsutelta op. too Hlroutilla op. Cord yceps unilateralbs 100 Cordycep, aphodll so Cordytteps brunneapuc! eta 100 stipi Cordycops sphaoocephala Cord yceps nutans too Hyrnenostllbe aurantlaca too Co,dyoreps iranglenols Coadyceps Irangloo,ols Figure 4.2. Phylogenetic relationships among 91 taxa in Clavicipitaceae and other families in Hypocreales based on combined evidence of seven gene partitions (i.e. nrSSU, nrLSU, fi- tubulin, EF-la,RPBJ, RPB2and mtATP6). The phylogeny is from one of five equally parsimonious trees inferred from maximum parsimony analyses and the resulting bootstrap values that are more than or equal to 70 % are provided above internal branches. The bootstrap values from combined dataset of six gene partitions (i.e. nrSSU, nrLSU, /3-tubulin, EF-la, RPBJandRPB2)are also shown below internal branches to show the influence of mtATP6 gene partition on the bootstrap values. The internal branches that are supported with more than or equal to 70% bootstrap values in both 7-gene and 6-gene analyses are shown in a thicker line. The internal branches that are collapsed in strict consensus tree are marked with asterisks (*). The numbers in circles correspondto ten major selected internodes in Figure 4.5 and 4.6. relationship of Clavicipitaceae dade C and Hypocreaceae (node 9 in Figure 4.2; Table

4.6). The neighbor-joining tree analyses employing a LogDet model produced a topology identical to that of MP analyses for ten major relationships (results not shown). These results indicate that the compositional bias in third codons of protein coding genes (Table 4.3) does not adversely affect the phylogenetic reconstruction for ten major relationships.

MP analyses with the 91 -taxon 7-gene data set resolved the non-monophyly of the Clavicipitaceae and three well-supported monophyletic groups labeled as

Clavicipitaceae dade A, B and C (Figure 4.2). However, the non-monophyly of the

Clavicipitaceae is not strongly supported because of the poor statistical support in the sister-group relationship of Hypocreaceae and Clavicipitaceae dade C (node 9 in

Figure 4.2; BP = 47%, BS = 3 in Table 4.6). Based on the separate analyses, the strong conflict in mtATP6 gene corresponds to the sister-group relationships of the

Hypocreaceae and Clavicipitaceae dade C (Figure 4.1). Thus, we excluded the mtA TP6 gene partition and generated the 91 -taxon 6-gene data set to address the sister-group relationships of the Hypocreaceae and Clavicipitaceae dade C. The topology from MP analyses based on the 91-taxon 6-gene data set is in agreement with that of the analyses with the 91-taxon 7-gene data set for ten major relationships

(Figure 4.2 and 4.3). However, the sister-group relationship of the Hypocreaceae and

Clavicipitaceae dade C (node 9 in Figure 4.2 and 4.4) received strong statistical support (BP = 74%, BS = 9 in Table 4.6). These results indicate that the conflict in Bionectria ct aurecfuwa Bionectriaceae

000'Roumegulenella ,ututa - cf Nectrla op. Nectriacinnatoorina I - Cosmospora coccinea I Leuconeclrla cluslae I Nectrjaceae - Vtndispooa dlparlelispera I Oploloneclrla t,lcflospo,a I Pseudonectrla rousseliana I Verticlillum lncojr,um Nypomoes polyporinus SphaerosWbella bes*eleyana Hypocreaceae Apfly00001rOma smrco.ranum Hypocrea lutea 00 Slmpllcillium Iamelllcola

l 000r Simpllcillium IanosonWeum 100 Slmplicillium lanosonlveum - Torrublla wallacel Engyodrsltlum araneanjm 000 I Torruboella ratlicaudata

O Lecanlcilllum anlillanum Cord yceps cardlnalis 00 Coniyceps card lot ails Lecanlclllium psalliotae Clavicipitaceae dade C Lecanlcllllum psalliofae Paediomyces faolnoSus 00 r O Paediomyces lannosus O Cord yceps tuberculata 000 Torrubiella con fragosa 100 000 Beauverla caledonoca Coodyveps scarabaeicola O a Paecilomyces tenulpes 000 Cordyceps militaris 000 Mariannaea pruinosa POehmoIa golnokles RoIiferopIo'o0ra angustispora Cordyceps chlamydosporia Pochon2 chlamydosporlum Metarhlzium album - Metan'olrium flavorirlde 100 Cord yceps tail too Metarhizium anisopffae 000 Shlmizuomyces paoad000us 000 Shimlzuomyces paradOxus Aschersonla badia Hypocrellaschlzosfachyl Clavuciputaceae dade A 00 00 Hypocretla op. 1001 Aschersonha placenta 100 Aschersonla placenta Verticillium epiphytum 0 Myologenospora atsamentosa 000 Balansia Ioenningslana too Balansia pllulaefonnis Eplclttoetyphlna 100 Claoolcepspaopali 00 Claviceps fusifonnis 000 Claviceps purpurea 000 Claviceps purpu,ea Cootiyceps guwoil o Haptocillium sineaose Haptociilium balanoides Haptoclllium aeospooum Coodyceps capItals Cordyceps fracla Cordycepsjaponlca CooTl)veps ophloglossolcies Co,dycepssubsessilis Clavicipitaceaedade B Cordyceps varlabilis Cordyceps ravenelll Cordyceps heteropoda Cordyceps melolonlhae Cord yceps unliateralis Cordyceps agnola o slylophora Cf. acleularls Hlrsu0010a op. Hirsute ha op. 00 Cordyceps aphodil Cordycepsbrurooeapuclata tool Cordyc.ps sphecocephala 000Lj Coniyceps nutans 0.05 subtaitutlonslslte 100 HymenosUlb. auranUaca O004Cordyceps iran glensla tooCordyceps Iran glensls Figure 4.3. Phylogenetic relationships among 91 taxa in Clavicipitaceae and other families in Hypocreales based on combined evidence of seven gene partitions (i.e. nrSSU, nrLSU, fi- tubulin, EF-la, RPB1, RPB2 and mtATP6). The phylogeny is constructed from a 50% majority consensus tree of 90,000 trees sampled from Bayesian analyses and the resulting posterior probabilities (PP) that are more than or equal to 95% are provided above internal branches. The posterior probabilities (PP) from combined dataset of six gene partitions (i.e. nrSSU, nrLSU, /3-tubulin, EF-la, RPB] and RPB2) are also shown below internal branches to show the influence of mtA TP6 gene partition on the posterior probabilities. The internal branches that are supported with more than or equal to 95% posterior probabilities in both 7- gene and 6-gene analyses are shown in a thicker line and the internal branches that are disparately supported between 7-gene and 6-gene analyses are marked with asterisks (*). The numbers in circles correspond to ten major selected internodes in Figure 4.5 and 4.6. Blonecbla ci. aureofuka BlonechIa ochroleuca ______Hyoroplspfla.apozlza Hydroplsphaera erub.scensIBionectriaceae 100 Roumeguleiella rufula I 100 Roumeguieriella rufula I Cf. Neclrla sp. . Necbacinnabarina I Cosmospora cocdnea I - leuconecpla cluslae - Nectriaceae 000 Viridlspotd Ulparmuspooa OpflloneQrla trlchospofa Pseudonec Via rounelana Ver?Jcllllum incurium I StercOrurium I Hypocreaceae paIuIHypomyspponus 000 Sphaerosttbeila beo*eleyana I Hypocre ,00 SImpliclIlium lamellicola 000 : SImpIlcilIlum lanosonlveum 000 Simpllcillium lanosoniveum TobleNa wahac& Engyodonlum waneanjm lX\J Tom,blela ratlicaudal 000 Leca,lcNllum antiJlanum

l Co,ciyceps cadinalls 1001 Coniyceps cardh,alls LI Lecanicllllum psaWotae Clavicipitaceae dade C , ------Lecanicllllum psalilota. I Paecilomyces fa,lno$us 1-H Paecllomyces farinosus Cordyceps .0 Tomiblela confragosa 0000 00,B.auvmia caledonka I4U_Co.dycepsscarabaeicola 6 Paecilomyces tenu,es ,oiJ Mariannaeapn,inosaCordyceps mWlarls 0000 Pochonia goniokies Rolifemplothota angusispora 100 Cordyceps chlamydospo,la ICC Pochonla chlamydospoflum Metarhlrlum album flavoviride izlum anlsopllae Shlmlzuomyces parado.cus Shimizuomyces paradoxuo Aschersonla badla Hypocreila schlzoslachyl

100 Hypoc,eila op. 000 Asch.rsonia placenta Clavicipitaceae dade A '00 Asche,sonla placenta Ve,liCillium epiphy&om Myrlogenospam atramen loss Balansla hennfrigslana - Balansia plluiaefonnis Eplchloe typhina Clavlceps paspal CIa vlceps fusiformis '°°rClaviceps pu,puoea 00 Claviceps pu,purea ioor Coniycups gunni Ioor------Haploclilium sinense

I Li Haotocllllum balanoides Haptocilliumzeospooum - Cordycepssubsessills L...... ( 6 r Cordyceps ophloglossoides °°''rl Llj_____Coniyceps capilata CIavicipitaceae dade B I I U::: Cordyceps fracta psjaponlca I I ._- Coniyceps heleropoda 00 100 Cord yceps ravenelll Cordyceps varlabili,

C Cordyceps unhlateralls Coidyceps agtlola Coniyceps ,tylophora 00006 Cordyceps ci. acicularl. 100 Hlrsutehia op. ICC Hireuteilasp. Coidyceps aphodi Coodyceps bnmn.suctata 100 000 CoVdyceps sph.cocw,hala ICC Corolyceps nutCCos CC iou Hymenosiuib.auranllaca O.O5substltutionslsite 100 Cordyc.pslrangtensls 000 Cordyc.ps lranglensis

Figure4.4.Phylogenetic relationships among 91 taxa in Clavicipitaceae and other families in Hypocreales based on combined evidence of six gene partitions (i.e. nrSSU, nrLSU, /3-tubulin, EF-la, RPBJandRPB2).The phylogenyisconstructed from a50%majority consensus tree of 90,000trees sampled from Bayesian analyses. Resulting posterior probabilities (PP) that are more than or equal to95%are provided below internal branches. Bootstrap values that are more than or equal to 70 % are obtained from combined data set of six gene partitions (i.e. nrSSU, nrLSU, /3-tubulin, EF-la, RPBJ and RPB2) in maximum parsimony framework and shown above internal branches. The internodes that are supported by both bootstrap values and posterior probabilities are considered strongly supported and in a thicker line. The numbers in circles correspond to ten major selected internodes in Figure4.5and 4.6. 101 to Tablethose of4.6. the Nodal 66-taxon supports analyses. for ten The major nodal relationships supports from in bothFigures 7-gene 4.2, (w/4.3 mtAand 4.4.TP6) The and nodal 6-gene (w/o mtA TPO) data sets supports from the 91 -taxon analyses are compared address the Nodeimpact of the localized conflict in mtATP6 91-taxonanaiss gene partition (Figure 4.1) to the total supports. 66-taxon anal ses are also provided to - 1 1 00% 99% 100% 100% 26 22 100 i'o 10 / 423 100%100%96% 100%95% 100% 100% 332919 281918 100%99%93% 100%93% 100% 100%100% 203714 2933 11 765 100%76% 100%76%100% 100% 100% 3357 4 30145 100%99% 100%95% 100% 100% 315619 343013 10 98 84%47%87% 75%74%78% 100% 100% 12 103 9119 54%53%85% 57%72%79% 100% 100% 12 2 759 102

mtA TP6 lowers the phylogenetic support for node 9 in the analyses of the 91 -taxon 7-

gene dataset. On the other hand, the nodal supports for other nodes (e.g., node 8 and

node 10) slightly decreased by excluding mtATP6 gene partition (Table 4.6). This pattern of increasing and decreasing bootstrap support for different nodes based on the

inclusion or exclusion of mtATP6 is consistent with the mtATP6 gene partition possessing congruent phylogenetic signal for other major relationships (node 8 in

Figure 4.2) and the conflict of mtATP6 being localized over a particular region of the tree (node 9 in Figure 4.2).

Two additional MP analyses were performed with 66-taxon data sets (the 66- taxon 7-gene data set and the 66-taxon 6-gene data set). These analyses were used to address the potential effect of missing data on the phylogenetic reconstruction and address the impact of taxon sampling in resolving the higher relationships. As seen in the analyses of 91 -taxon data sets, support for node 9 was greatly improved after mtATP6 gene partition was excluded (Table 4.6). We found, however, different levels of nodal support for the sister-group relationship of Clavicipitaceae dade A and B

(node 10) between the 91 -taxon and the 66-taxon analyses (Table 4.6). Node 10 is not well supported in either the 66-taxon 7-gene analyses (BP = 54%, BS = 2 in Table 4.6) or the 66-taxon 6-gene analyses (BP = 57%, BS = 2 in Table 4.6). These results are in dramatic contrast to the 91 -taxon analyses, which showed the strong support for node

10 in either the 91-taxon 7 gene (BP = 84%, BS = 10 in Table 4.6) or the 91-taxon 6 gene analyses (BP = 75%, BS9 in Table 4.6). As a result, taxon sampling is 103 considered important in resolving the relationship in node 10 despite the missing data associated with the 91-taxon data set.

Bayesian analyses were conducted on the four previously described data sets using a GTR + I + F model of evolution as determined by Modeltest analyses. For all analyses, three independent runs converged on the same stable log-likelihood scores and thus we sampled 90,000 trees from one of the three runs. A 50% majority-rule consensus tree (Figure 4.3) from the 91-taxon 7-gene data set is well resolved and most nodes (94%) were supported by posterior probabilities of greater than or equal to

95%. The topology of Bayesian tree is identical to the parsimony tree with respect to the relationships often major deep nodes (Figure 4.3 and 4.4). The posterior probabilities in Bayesian analyses, however, were not dramatically altered by excluding the mtATP6 gene partition. The ten major deep relationships were supported

(PP? 95%) in either 9l-taxon 7-gene or 91-taxon 6-gene analyses (Table 4.6). These results indicated that the detected localized conflict in mtATP6 at node 9 is not reflected in the posterior probabilities. In addition, the 10 deeper nodes were supported

(100% PP) in both the 66-taxon 7-gene and 66-taxon 6-gene analyses, indicating that unlike the bootstrap values in MP analyses, the posterior probabilities are less sensitive to changes in taxon sampling (e.g., node 10; Table 4.6).

Combinational analyses

To investigate the effect of character and taxon sampling on the resolution of ten major relationships in more detail, we performed a series of combinational 104 bootstrap (CB) analyses. These analyses were performed on 126 data sets, from all combinations of seven gene partitions, using maximum parsimony. Mean bootstrap supports were calculated by averaging the bootstrap support values from analyses that combined the same number of genes but in different combinations. These mean bootstrap supports were plotted against number of genes in Figure 4.5 and used as an indicator to address how many genes are required to confidently resolve (BP ? 70%) these ten relationships (Figure 4.5). Two nodes (nodes 7 and 10) differ greatly in number of genes required to receive strong support between 91 -taxon and 66-taxon

CB analyses. For node 7, the mean bootstrap supports of the 66-taxon CB analyses are generally much greater than those of the 91 -taxon CB analyses. Node 7 requires six genes to be strongly supported in 91-taxon CB analyses. In contrast, only two genes are necessary to receive strong support in 66-taxon CB analyses (Figure 4.5). On the other hand, the mean bootstrap supports for node 10 are much higher in 91 -taxon CB analyses than 66-taxon CB analyses. In receiving strong mean bootstrap support, node

10 is only strongly supported in 91 -taxon CB analyses when six genes are combined.

For the remaining eight deeper relationships, the mean bootstrap supports from 91- taxon CB analyses tend to be slightly higher than those of 66-taxon CB analyses

(Figure 4.5), indicating that these eight nodes are not sensitive to taxon samplings (91- taxon vs. 66-taxon) (Figure 4.5). Although the remaining eight deeper nodes received consistently higher bootstrap support as the number of genes increased (Figure 4.5), these nodes required different numbers of genes in the combined data sets to receive 105

TfT11T iTtt: ITI'tT : to

F !: iNfl: !::ii!iil:

T4+ff Iii1ii I 2 3 4 5 6 7

Figure 4.5. The effect of character (gene) and taxon samplings on mean bootstrap supports of ten selected major relationships in Figures 4.2, 4.3 and 4.4. The mean bootstrap supports were based on the results of the combinational (CB) analyses with 126 data sets in different combinations (see materials and methods). The mean bootstrap supports were calculated by averaging the bootstrap supports of the analyses from the combined same number of genes in different combinations and plotted against the number of genes in combining data sets. To address the impact of taxon sampling, the mean bootstrap supports were calculated for the 91-taxon and 66-taxon data sets. strong support. For example, node 1 requires only two genes to be strongly supported, whereas node 8 is only strongly supported when six genes are combined. Node 9 is not strongly supported by the mean bootstrap supports due to the inclusion of mtATP6 gene partition in CB analyses (Figure 4.5). Due to the localized conflict of mtATP6 at node 9, it is only supported by 6-gene combined analyses that do not include mtATP6

(Table 4.6).

Contributionofeach gene partition to nodal support 106

Partitioned Bremer Support (PBS) values were calculated to generate the distribution of relative contribution of each gene partition to the total support present in the 7-gene analyses. The PBS values often major deeper nodes were calculated for both the 91-taxon and 66-taxon 7-gene data sets in order to examine the influence of missing data in 91-taxon data set (Figure 4.6). While five major relationships (nodes 1,

6, 7, 9 and 10) are supported by the same gene partitions from both the 91-taxon and

61-taxon analyses, other relationships (nodes 2, 3, 4, 5 and 8) show dissimilar patterns of PBS values between the 91-taxon and 66-taxon analyses (Figure 4.6). In demonstrating the lowest contribution of each gene partition to the total support, six relationships (nodes 1, 2, 5, 7, 8 and 9) are not identical in gene partitions between 91- taxon and 61-taxon analyses (Figure 4.6). In particular, the dissimilar patterns in PBS values were detected for node 9. Although the 66-taxon data set is complete for all gene partitions, the strong conflict of mtATP6 at node 9 was not reflected in the PBS results based on the 66- taxon data set. While mtA TP6 demonstrated the strongest conflict at node 9 in 91-taxon analyses (PBS -33 in Figure 4.6), the gene partition that is in strongest conflict in 66-taxon is nrLSU gene partition (PBS -4 in Figure 4.6).

The dissimilar patterns of PBS values (Figure 4.6) between 91-taxon and 66- taxon analyses prevented us from identifying particular gene partitions that contain the highest or lowest levels of phylogenetic signals for the relationships. Instead, we used the results of CB analyses to demonstrate the contribution of each gene partition to the total support for the ten major relationships. All ten major nodes are supported by the same pattern of gene partition from both 91-taxon and 66-taxon data sets (Figure 4.6). 107

,*SSU -5 8 76 1OO ,VLSU 0.5 6 90 100 4ubulin 0.5 2.5 77 98 EF-h 13.5 13 79 100 ______RP8I 3,5 31.5 77 100 ('\ ssL RPB2 15* 365*100 lOOrn \. 91tx66tx91tee 66tox* nSATPS -2 6,5 79 100 ,vSSu -6.5 0 49 49 ToI 26 104 100%lO0/ ,*,tSU -14.5 -10 49 49 /HuSuI,, 19.2 14 49 49 EF-1 7.25 20* 100* 100* PBS Cor,. And. RPBI 28.7 -4 49 49 -i;;; -r--- RPB2 11.7 -15 i i' -jj; jj %ATP6 -27 9 _! _..- ,vLSU 4.5 -5 65 88

iQ9_ i* ,buS, 1 105 90 95 EF-1 11.5* 35 95 93 PBS Con, Anol. RPBI 6 5.5 90 93 PBS -2' - 91tXo 66tex091t*xO 0*oo RPB2 2.5 0 93 96 66e 6.5 95 95 os 1 16 22 &5 ,o$SU -4.85 2 20 61 n,LSL -15.5 -7 14 26 .I. ..i...... L. i2 i22. VLSU -16.5 -7 20 66 /4,buIin 13.5 1.5 IS 19 %buSn 17.31 2.5 22 61 8 23 :; p *265 95* 25* 36* RPB2 5.14 -4 49 92 ,VSSU -2.75 3.5 77 87 AlPS 33 4 5 14 n(ATPS -31.42 . -0.5 87 87 To 4 19 100% 100% -- pubu!m 21.5 6 81 87 SF-b -3 jj5* 79 93 RPBI *3563 96 '- PBS Con, Anol 5.5 65 8_)_-i RPB2 16.87 -3.5 100* 100* -g r OMTPS -25.12 .45 77 85 ,,,tsu :2' -2' 38 42 !s!i. _.. _9...i i2. jibubu/in 2.75 4.5 36 44 -13 1.5 41 42 ,'___\ And. RPBI 13.5 5 41 44 79* -!L_ Bitaxo 66toxo RPB2 *1975 45 90* ,n'ssu T 11 'L? _-._±t- .__ ,ntSU -13.5 -2 61 88 i99* !22x !e!5!_ _JL. _!L P4USOSO 19.5 7 64 90 SF-to .6.5 7 84 90 100* 100* '"'\ PRO Con, Anal. RPBI 48.5 19 RP62 10 4 95 100*

OrSSU 1 30 20 92 100k nrLSU -8.5 7 31 20 ...L 56 100% 100% /toSoSn 5.5 1.5 25 15 SF-To

,nATPS 2 -4 30 19 c.

Tsal 10 2 100% 100% '

SF-To 12.5 0.5 68 76 RPBI 34.5 14.5W 100* 100* RP82 20.5 10 85 100* ,nTATP6 -8 -1 69 74

Figure 4.6. Schematic diagram of ten major relationships in Figures 4.2, 4.3 and 4.4 to demonstrate the contribution of each gene partition to the total supports. Partitioned bremer support (PBS) values were calculated for ten major relationships from 91 -taxon and 66-taxon data sets. The results from combinational (CB) analyses in both 91 -taxon and 66-taxon data sets were also used to demonstrate the relative contribution of each partition to the total supports. Out of 126 CB analyses (see materials and methods), 63 CB analyses include one particular gene partition in combined data sets. For each gene partition, the percentage of 63 CB analyses, which have more than 70% bootstrap supports, is used as supplementary information for generating the distribution of the supports or conflicts in each gene partition to each of ten maj or relationships. For each calculation, the gene partitions, which show the highest and lowest contributions toa particular relationship, were denoted with the PBS values and the percentages in bold letters. Asterisks (*) indicate the gene partition that shows the highest contribution. Although only six deeper nodes contain the lowest contribution from the same gene partition, the difference in the percentage of CB analyses is within 2% between the

lowest and the second lowest contributed gene partitions (Figure 4.6). Therefore, the results of CB analyses (Figure 4.6) are more consistent than PBS values in

demonstrating the relative contribution of each gene partition to the total supports. The

inconsistencies of PBS values suggest that they are not good indicators for addressing the nodal support from a particular gene partition and may be particularly sensitive to

stochastic properties associated with homoplasy of heterogeneous data.

Nodes 8, 9 and 10 are the most crucial nodes in inferring a well-resolved phylogeny of Clavicipitaceae, as the resolution of these nodes imply that the family is either monophyletic or paraphyletic. The results from the CB analyses demonstrate that the phylogenetic signal necessary to resolve these three nodes is not present in any one gene partition. The CB analyses showed that node 8, which unites the

Hypocreaceae and the three clades of Clavicipitaceae, is strongly supported by the

RPB2gene partition (79% CB analyses in 91-taxon) (Figure 4.6). Node 9, which supports the monophyly of Hypocreaceae and Clavicipitaceae dade C, receives it highest level of support fromRPBJ (25%CB analyses from the 91-taxon data set)

(Figure 4.6). In addition,RPBJis the most important gene partition (49 % CB analyses from the 91 -taxon data set) for resolving node 10, which supports the monophyly of Clavicipitaceae clades A and B (Figure 4.6). This pattern of localized support is also present in other nodes of the tree. For example, node 2 (monophyly of the Nectriaceae) is supported solely by theEF-lagene partition (100% CB analyses 109

from both the 91-taxon and 66-taxaon data sets) (Figure 4.6). Despite the strongEF-

la support at node 2 in CB analyses, however, PBS values indicate thatRPBJ(PBS

28.75 in Figure 4.6) is considered to be the greatest contributor of signal for node 2 in

91-taxon dataset.

Discussion

ILD test and the heterogeneity among gene partitions

One of the most common methods used in detecting the overall incongruence among data partitions, is the incongruence length difference (ILD) test (Farris et al.,

1994). The ILD test is based on the degree of homoplasy that results from combining data sets from multiple independent sources. In this study, significant heterogeneities were detected in most of pairwise comparisons (19 out of2l) in the ILD tests (Table

4.5). The utility of ILD test as a combinability test has been questioned in recent studies (Barker and Lutzoni, 2002; Dowton and Austin, 2002; Yoder et al., 2001), which demonstrated that the difference of among site-rate variation and evolutionary rates between data partitions leads to significant ILD results by generating excessive homoplasies in combined data sets. The relative size of the data partitions also influences the significance of the test (Dowton and Austin, 2002). Therefore, the ILD tests should be used with caution as a combinability test (Barker and Lutzoni, 2002;

Dolphin et aL, 2000; Dowton and Austin, 2002). 110

The significant higher level of discordance in the relationships between separate and combined analyses (Figures 4.1 and 4.2) may be due to the saturation of each gene partition that results in the poor resolution for the higher relationships in separate analyses (Baker and DeSalle, 1997; Engstrom et al., 2004). To investigate the substitutional saturation of nrSSU, nrLSU and each codon position of protein-coding genes, we calculated the observed pairwise distances for transversions and transitions for each pair of taxa and plotted against the corrected distance with the F8 1 model

(data not shown). The nrSSU and nrLSU showed linear relationships in both transversions and transitions, exhibiting no saturation. Therefore, the poor resolution in the separate analyses of nrSSU and nrLSU (Figure 4.1) may be from the lower number of parsimony informative characters (Table 4.3). In contrast, saturation was detected in the third codons of all five protein-coding genes, in both transversions and transitions, at higher level of divergences. These results indicate that the relative substitutional rates of the third codons in protein-coding genes are substantially different from those of the nrSSU and nrLSU. In addition, the first codons of protein- coding genes also appear to be saturated for transitions at higher level of divergences except for EF-la gene partition. The saturation patterns in protein-coding genes may result in the significant noise of first and third codon positions, and the significance of the ILD tests in this study may be due to the high level of homoplasies in gene partitions (Baker and DeSalle, 1997; Dolphin et al., 2000). Therefore, the results of

ILD tests in this study (Table 4.5) are not interpreted as a measure of incongruence and combinability between data partitions. Rather, they are considered as preliminary 111 indicators of overall level of homoplasies from combining the heterogeneous data partitions.

Comparison ofphylogenetic methods in combined analyses.

Although strong localized conflict was detected at node 9 between mtATP6 and other nuclear partitions (RPBJ and RPB2) (Figure 4.1), our results based on MP and

B-MCMCMC analyses are topologically congruent for the ten major deeper nodes.

Exclusion or inclusion of the mtATP6 partition does not alter the topologies for ten major relationships in the 91-taxon or the 66-taxon data sets analyses in MP or B-

MCMCMC analyses (Figures 4.2, 4.3 and 4.4). Despite the topological congruence between the MP and B-MCMCMC analyses, the posterior probabilities support the sister-group relationships (PP = 100%) of the Hypocreaceae and Clavicipitaceae dade

C (node 9 in Figure 4.2) and do not reflect the strong conflict in mtATP6 gene partition

(Figure 4.1; Table 4.6). These results are consistent with the tendency of the Bayesian posterior probabilities to overestimate the phylogenetic confidence of relationships.

The use of Bayesian posterior probabilities in assessing phylogenetic confidence has been intensely debated (Alfaro et al., 2003; Cummings et al., 2003; Doaudy et al.,

2003a; Huelsenbeck et at., 2001). They have been considered as an equivalent measure to bootstrap support in some studies (Cummings et al., 2003; Huelsenbeck et al., 2001; Larget and Simon, 1999), while other empirical and simulation studies showed the significant difference between bootstrapping and posterior probabilities

(Doaudy et al., 2003a; Doaudy et al., 2003b; Erixon et al., 2003; Gontcharov et at., 112

2004; Reeb et al., 2004; Suzuki et al., 2002; Whittingham et al., 2002). Our results are in agreement with these latter empirical and simulation studies in that posterior probabilities were insensitive in detecting the localized incongruence or conflict associated with the mtATP6 data. Although the conditions are not all known, the posterior probabilities appear to overestimate the phylogenetic confidence for short internodes such as node 9 and 10 (Figure 4.3) in this study (Alfaro et al., 2003; Kauff and Lutzoni, 2002).

Patterns similar to overestimation of nodal support by Bayesian posterior probabilities were also evident among more closely related taxa within the major clades (Figures 4.3 and 4.4). The Bayesian analyses of the 91-taxon 7-gene and 6-gene data sets strongly support most of relationships (94% of internal nodes) by posterior probabilities (PP95%) (Figures 4.3 and 4.4). However, different topological resolutions between the 7-gene and 6-gene analyses of more closely related species

(marked with asterisks in Figure 4.3) received posterior probabilities of ? 95%. In contrast, MP analyses showed that none of these topological incongruences is strongly conflicted (BP ? 70%) between 7-gene and 6-gene analyses (Figure 4.2); the strongly supported relationships in 7-gene analyses are also strongly supported in 6-gene analyses. Furthermore, the overall number of strongly supported nodes (BP70%) in

MP analyses is much lower (73%) than that (PP? 95%) of Bayesian analyses (Figures

4.2 and 4.3) and many of the well-supported nodes in Bayesian analyses were not supported in MP analyses of either the 7-gene or 6-gene data sets. 113

Incongruence, sampling and bootstrap support

We detected strong localized conflict in mtATP6 gene partition at node 9 in the comparison of separate analyses (Figure 4.1). Identifying the conflicted gene partition greatly improved the support for phylogenetic relationships in the combined analyses

(Figure 4.2; Table 4.6). To better understand the incongruences among gene partitions and demonstrate the trends of combining the heterogeneous data in more detail, combinational bootstrap (CB) analyses were performed for the ten major deep nodes by MP analyses of 126 data sets (see materials and methods) (Figure 4.5). These analyses suggest that different numbers of combined genes were required for any given node to provide robust phylogenetic estimates. The short internodes (node 9 and

10 in Figure 4.3 and 4.4) required more character and taxon sampling to resolve the relationships than longer branches (e.g., nodes 2, 3, 5 and 6 in Figure 4.3 and 4.4).

Node 9, for example, is not supported by mean bootstrap supports due to the localized conflict in mtATP6 and it is only supported in 6-gene combined analyses without mtATP6 in both 91-taxon and 66-taxon data sets (Table 4.6). On the other hand, longer branches (e.g., nodes 2, 3, 4, 5 and 6) receive the strong mean bootstrap supports (BP

? 70%) by combining three or four genes (Figure 4.5).

Based on CB analyses, one to few gene partitions contribute most of the phylogenetic signal of the total support for a given node with no one gene partition being the major support contributor for all nodes (Figure 4.6). For example, the monophyly of the Nectriaceae (node 2 in Figure 4.6) is only supported whenEF-la gene partition is included in the combined analyses, whereas the support fromRPB2 114 gene partition is essential to resolve node 8 (Figure 4.6). Similarly, the contribution of

RPBJis considered important to support the sister-group relationships (the

Hypocreaceae and Clavicipitaceae dade C and Clavicipitaceae dade A+B) in nodes 9 and 10 (Figure 4.6). These results indicate that the contributions of each gene partition are biased for certain nodes and represent the heterogeneous nature of gene partitions as evolutionary markers (Wiens, 1998). Our results indicate that combining the heterogeneities of gene partitions are beneficial to construct a well-resolved phylogeny through supporting different relationships from different gene partitions.

We analyzed data in two different taxon samplings (91-taxon vs. 66-taxon) to demonstrate the impact of missing data in 91 -taxon dataset. The sister-group relationship of Clavicipitaceae dade A and B (node 10) is only strongly supported by the 91-taxon analyses (Table 4.6). These results suggest that taxon sampling is more important than character sampling in confidently addressing the relationship in node

10. Numerous studies have focused on the attributes of sampling more characters versus more taxa. Many simulations have suggested that increasing character sampling is beneficial for phylogenetic accuracy and increases the phylogenetic confidence for certain branches (Braum et al., 1994; Mitchell et al., 2000; Rosenberg and Kumar,

2001; Sullivan, 1996). Other studies, however, have addressed the beneficial effect of taxon sampling in addressing problematic lineages characterized by long-branch attraction (Graybeal, 1998; Zwickl and Hillis, 2002). When we increased the character sampling, we observed dramatic increases in bootstrap support for most of the ten major relationships (Figure 4.6); however, our results also demonstrated that taxon 115 sampling is important in resolving the relationship of node 10 (Figure 4.5). These results support that in general increased character sampling increases phylogenetic confidence for a given node, but that increased taxon sampling may also be important, but difficult to predict priori.

Congruence vs. combinability.

The strong conflict in mtATP6 at node 9 begs the question whether or not mtATP6 should be included in the combined dataset to infer a final working hypothesis of the relationships in the Clavicipitaceae. Proponents of total evidence

(character congruence) would argue that mtATP6 gene partition should be included in the combined dataset (Barrett et al., 1991; Kiuge, 1989; Kluge, 1998; Kluge and Wolf,

1993; Miyamoto and Fitch, 1995). On the other hand, the supporters of a conditional combinability approach might claim that mtATP6 should be excluded in combined dataset because of the incongruence in mtATP6 gene partition (Bull et al., 1993;

Chippindale and Wiens, 1994; de Queiroz, 1993). Our results from nucleotide content, saturation analyses, and ILD tests are consistent as characterizing these data partitions as heterogeneous. Combining these heterogeneous gene partitions, however, resulted in increased support for different nodes by different gene partitions (e.g., node 2 by

EF-la, node 7 by RPB2, and nodes 9 and 10 by RPBJ in Figure 4.6) and an overall robust estimate of phylogeny. Nonuniform distributions of congruence and conflict complicate the interpretation of the mtATP6 partition. If the decision was made to exclude mtATP6 gene partition in the combined data set, the phylogenetic signal 116 would be significantly diminished for other relationships. For example, node 5 is supported by mtATP6 gene partition, equal to that of theRPBJandRPB2gene partitions (Figure 4.6). Several similar cases in Figure 4.6 indicate that one particular gene partition varies significantly in the level of support for different relationships

(e.g.,RPB2for nodes 8 and 10). Therefore, excluding certain partitions in combined analyses may not be desirable when the conflict is localized, as the same gene partition may provide important signal in the combined analyses that increases the overall robustness of the inference (Wiens, 1998).

With an increasing number of taxa, several studies have shown that incorporating data from multiple genes is necessary to confidently resolve higher-level relationships (e.g., Lutzoni et al., 2004; Rokas and Carroll, 2005). Due to the complexities of large datasets that include many genes and taxa, the evolutionary processes affected by constraints unique to gene partitions may not be sufficiently modeled for lineage specific parameters (Rokas and Carroll, 2005). For heterogeneous data, the violations of assumptions associated with phylogenetic algorithms may happen in one particular gene partition, or for certain lineages, and result in its strong localized conflict against the phylogenetic estimate of other gene partitions when analyzed separately or combined. In this study, we provide a working hypotheses on the relationships in the Clavicipitaceae by demonstrating the contribution of each gene partition for specific nodes. Our results suggest that it is not advantageous to simply exclude one particular gene partition in the combined dataset because of its localized conflict, rather analyses should explore the complexity of the conflict and the 117 ramifications of the potential loss of signal for other nodes as well. That is, analyses should be performed with and without the conflicting data to best assess the overall inference of phylogenetic relationships.

Phylogenetic relationshipsofthe Clavicipitaceae

The Clavicipitaceae comprises a morphologically and ecologically diverse group of fungi. Robust phylogenetic hypotheses of the family are essential to assess its biodiversity and to gain insight into the dynamic evolution of its ecology (e.g., host affiliation). Identifying and understanding the localized conflict in the mtATP6 partition enabled us to confidently construct a well-resolved hypothesis for the higher relationships of the Clavicipitaceae. One recent simulation study (Kolaczkowski and

Thorton, 2004) demonstrated that if a dataset comprises heterogeneous partitions, the phylogenetic estimate from parsimony analyses is more robust than that from

Bayesian analyses. Both the insensitivity of the B-MCMCMC posterior probabilities to the localized conflict of mtATP6 and the pattern of well-supported conflicting terminal nodes observed in this study are consistent with this finding. Therefore, the topologies with strong bootstrap support are used in constructing the working phylogenetic hypotheses for the Clavicipitaceae and posterior probabilities from

Bayesian analyses are used as supplementary measure of support. h interpreting nodal support and proposing a working hypothesis for the phylogeny of the

Clavicipitaceae, we chose nodes with the highest bootstrap support and posterior 118 probability between the 6- and 7-gene analyses in both the 91 and 66 taxon datasets as the most plausible hypothesis of phylogenetic relationships.

Our results agree with a previous multigene phylogenetic study (Castlebury et al., 2004) in supporting the monophyly of other families in Hypocreales

(Bionectriaceae, BP = 100%, PP = 100%; Nectriaceae, BP = 96%, PP = 100%;

Hypocreaceae, BP = 100%, PP = 100%) and additional taxon sampling within the

Clavicipitaceae did not affect the monophyletic relationships of these three families.

The basal placements of Nectriaceae and Bionectriaceae with respect to

Clavicipitaceae and Hypocreaceae also agree with previous phylogenetic results

(Castlebury et al., 2004). Our results, however, reject the monophyly of the

Clavicipitaceae and support the paraphyly of the Clavicipitaceae due to the sister- group relationship (BP = 74%, PP = 100%) of the Hypocreaceae and Clavicipitaceae dade C (Figures 4.2, 4.3 and 4.4). Therefore, the support for monophyly of the

Clavicipitaceae in the previous phylogenetic study (Castlebury et al., 2004) arguably results from the limited taxon sampling of representative taxa in the Clavicipitaceae.

The Clavicipitaceae has long been recognized based on the diagnostic characters of long cylindrical asci, a thickened ascus apex, and fihiform ascospores that often break into partspores at maturity (Diehi, 1950; Rogerson, 1970; Rossman et al.,

1999; Spatafora and Blackwell, 1993). Species of the Hypocreaceae also possess long cylindrical asci and ascospores that break into partspores, but they differ from the species of the Clavicipitaceae in a less pronounced ascus apex and ovoid to ellipsoidal ascospores (Rossman et al., 1999). These results suggest that the diagnostic characters 119 historically used to recognize the Clavicipitaceae (e.g., long cylindrical asci with pronounced apices and filiform ascospores) are symplesiomorphic and were inherited from the common ancestor of Hypocreaceae and all three clades of Clavicipitaceae, i.e., node 7. Furthermore, the ovoid to ellipsoidal ascospores of Hypocreaceae

(Rossman et al., 1999) are derived, inherited synapomorphies from the common ancestor of the Hypocreaceae, i.e., node 3.

Our results reject the monophyly of the Clavicipitaceae and indicate that it comprises three strongly supported clades (Clavicipitaceae dade A, BP100%, PP =

100%; Clavicipitaceae dade B, BP = 100 %, PP100%; Clavicipitaceae dade C, BP

= 100%, PP = 100%). Clavicipitaceae dade A consists of four strongly supported subclades, but their interrelationships are not strongly supported. Subclade Al (BP =

99%, PP 100%) includes one asexual species, Verticillium epihytum, and the grass symbionts (e.g., Claviceps, Balansia and Epichloe) of subfamily Clavicipitoideae.

Subclade A2 (BP = 95%, PP = 100%) comprises members of subfamily

Oomycetoideae that includes genus Hypocrella, which are pathogens of scale insects and white flies, and its asexual genus Ashersonia. Subclade A3 (BP = 98%, PP =

100%) includes two species of Cordyceps and members of the asexual genera

Pochonia, Metarhizium and Rotzferophthora and the subclade A4 contains only one species, Shimizuomyces paradoxa, which fruits on seeds of Smilax.

Clavicipitaceae dade B contains three major subclades that all include members of Cordyceps and several closely related asexual genera (e.g., Hirsutella,

Haptocillium and Hymenostilbe). Subclade Bl (BP = 100%, PP = 100%) contains 120 species of Cordyceps (e.g., C. ophioglossoides) that parasitize the truffle genus

Elaphomyces and scarab larvae (e.g., C. subsessilis). Subclade B2 (BP = 100%, PP =

100%) contains one species of Cordyceps (e.g., C. gunnii) and three species of the asexual genus Haptocillium, which are pathogens of free-living nematodes. Subclade

B3 (BP = 99%, PP = 100%) consists of members of Cordyceps and the asexual genera of Hirsutella and Hymenostilbe. Clavicipitaceae dade C consists of two major subclades. Subclade Cl is a strongly supported asexual lineage (BP = 100%, PP =

100%), which includes three species of asexual genus Simplicillium. These asexual taxa are primarily isolated as parasites of fungi and are not linked to any sexually reproducing species of the Clavicipitaceae. Subclade C2 (BP = 75%, PP100%) includes the members of genera Cordyceps and Torrubiella as well as several members of asexual genera (e.g., Beauveria, Paecilomyces, and Lecanicillium) with knownlinksto Cordyceps and Torrubiella.

The three clavicipitaceous clades not only reject the monophyly of the

Clavicipitaceae, but they are also inconsistent with the recognition of the subfamilies

Clavicipitoideae, Cordycipitoideae and Oomycetoideae (Figure 4.2). Species of the subfamily Clavicipitoideae (e.g., Claviceps, Balansia and Epichloe) form subclade Al within Clavicipitaceae dade A, which also includes species of the subfamily

Oomycetoideae (e.g., Hypocrella), subclade A2. On the other hand, species of the subfamily Cordycipitoideae (e.g., Cordyceps and Torrubiella) belong to all three clavicipitaceous clades, indicating that both Cordyceps and the Cordycipitoideae are not monophyletic (Figure 4.2). These results point to the need for significant 121 taxonomic revisions for the clavicipitaceous fungi at numerous taxonomic ranks ranging from familial to generic.

Conclusions

The phylogenetic analyses presented here provide a robust hypothesis for the higher-level relationships within the Clavicipitaceae and the Hypocreales. We demonstrated that identifying the localized conflict in mtA TP6 gene partition improves the phylogenetic estimates of the relationships, but that complete exclusion of this gene partition is not without cost for other nodes. These analyses revealed that the

Clavicipitaceae is paraphyletic and comprises three statistically well-supported groups

(Figure 4.2). The paraphyly of the Clavicipitaceae here is defined by the monophyly of the Hypocreaceae and Clavicipitaceae dade C (node 9 in Figure 4.2). The mtATP6 partition is in conflict with this relationship (Figure 4.1), which is strongly supported byRPBJgene partition (Figure 4.6). Although most of the higher-level relationships are not sensitive to the taxon samplings (91-taxon vs. 66-taxon), the sister-group relationship of Clavicipitaceae dade A and B (node 10 in Figures 4.2 and 4.5; Table

4.6) is only supported in the 91-taxon analyses. These results indicate that while increased character sampling generally increases nodal support for most nodes, taxon sampling can be more important for particular nodes (e.g., node 10). Our results also showed that multi-gene phylogenies produced by heterogeneous data sets are advantageous in resolving the higher relationships by combining the complementary 122 support for different relationships from different gene partitions. Therefore, unequivocal exclusion of a gene partition due to the localized incongruence is not advantageous in inferring a robust phylogenetic hypothesis for the Clavicipitaceae and the Hypocreales.

References

Akaike, H. 1974. A new look at the statistical model identification. IEEE Trans. Automatic Control 19:716-723.

Alfaro, M. E., S. Zoller, and F. Lutzoni. 2003. Bayes or bootstrap? A simulation study comparing the performance of Bayesian markov chain monte carlo sampling and bootstrapping in assessing phylogenetic confidence. Molecular Biology and Evolution 20:255-266.

Artjariyasripong, S., J. I. Mitchell, N. L. Hywel-Jones, and E. B. Garath Jones. 2001. Relationships of the genusCordycepsand related genera, based on parsimony and spectral analysis of partial 1 8S and 28S ribosomal gene sequences. Mycoscience 42:503-5 17.

Baker, R. H., and R. DeSalle. 1997. Multiple sources of character information and the phylogeny of Hawaiian drosophilids. Systematic Biology 46:654-673.

Baker, R. H., X. Yu, and R. Desalle. 1998. Assessing the relative contribution of molecular and morphological characters in simultaneous analysis trees. Molecular Phylogenetics and Evolution 9:427-436.

Barker, F. K., and F. Lutzoni. 2002. The utility of the incongruence length difference test. Systematic Biology 51:625-637.

Barrett, M., M. Donoghue, and E. Sober. 1991. Against consensus. Systematic Zoology 40:486-493.

Braun, E. L., and R. T. Kimball. 2002. Examining basal avian divergences with mitochondrial sequences: model complexity, taxon sampling, and sequence length. Systematic Biology 5 1:614-625. 123

Braum, E. L., R. K. Jansen, B. Oxelman, M. Backlund, H. Lantz, and K.-J. Kim. 1994. Examining basal avian divergences with mitochondrial sequences: model complexity, taxon sampling, and sequence length. Systematic Biology 51:614- 625.

Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42:795-803.

Bremer, K. 1994. Branch support and tree stability. Cladistics 10:295-304.

Buckley, T. R. 2002. Model misspecification and probabilistic tests of topology: Evidence from empirical data sets. Systematic Biology 51:509-523.

Bull, J. J., J. P. Huelsenbeck, C. W. Cunningham, D. L. Swofford, and P. J. Waddell. 1993. Partitioning and combining data in phylogenetic analysis. Systematic Biology 42:384-397.

Cameron, K. M., M. W. Chase, W. M. Whitten, P. J. Kores, D. C. Jarrell, V. A. Albert, T. Yukawa, H. G. Hillis, and D. H. Goldman. 1999. A phylogenetic analysis of the Orchidaceae: evidence from rbcL nucleotide sequences. Am. J. Bot. 86:208-224.

Castlebury, L. A., A. Y. Rossman, G.-H. Sung, A. S. Hyten, and J. W. Spatafora. 2004. Multigene phylogeny reveals new lineage for Stachybot,ys chartarum, the indoor air fungus. Mycological Research 108:864-872.

Chippindale, P. T., and J. J. Wiens. 1994. Weighting, partitioning, and combining characters in phylogenetic analysis. Systematic Biology 43:278-287.

Cummings, M. P., S. A. Handley, D. S. Myers, D. L. Reed, A. Rokas, and K. Winka. 2003. Comparing bootstrap and posterior probability values in the four-taxon case. Systematic Biology 52:477-487.

Danforth, B. N., S. G. Brady, S. D. Sipes, and A. Pearson. 2004. Single-copy nuclear genes recover Cretaceous-age divergences in bees. Systematic Biology 53:309- 326.

Darlu, P., and G. Lecointre. 2002. When does the incongruence length difference test fail? Molecular Biolology Evolution 19:432-437. de Queiroz, A. 1993. For consensus (sometimes). Systematic Biology 42:368-372.

Diehl, W. W. 1950. Balansia and the Balansiae in America. Agricultural monograph No.4. U. S. D. A. Washington. 124

Doaudy, C. J., F. Delsuc, Y. Boucher, W. F. Doolittle, and B. J. P. Douzery. 2003a. Comparison of the Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Molecular Biology and Evolution 20:248-254.

Doaudy, C. J., M. Dosay, M. S. Shivji, and M. J. Stanhope. 2003b. Molecular phylogenetic evidence refuting the hypothesis of Batoidea (rays and skates) as derived sharks. Molecular Phylogenetics and Evolution 26:215-22 1.

Dolphin, K., R. Belshaw, D. L. C. Orme, and D. L. J. Quicke. 2000. Noise and incongruence: Interpreting results of the incongruence length difference test. Molecular Phylogenetics and Evolution 17:40 1-406.

Dowton, M., and A. D. Austin. 2002. Increased congruence dose not necessarily indicate increased phylogenetic accuracy-The behavior of the incongruence length difference test in mixed-model analyses. Systematic Biology 51:19-31.

Engstrom, T. N., H. B. Shaffer, and W. P. McCord. 2004. Multiple data sets, high homoplasy, and the phylogeny of softshell turtles. Systematic Biology 53:693- 710.

Erixon, P., B. Svennblad, T. Britton, and B. Oxelman. 2003. Reliability of bayesian probabilities and bootstrap frequencies in phylogenetics. Systematic Biology 52:665-673.

Farris, J. S., A. G. Kallersjo, A. G. Kluge, and C. Butt. 1994. Testing significance of incongruence. Ctadistics 10:315-319.

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

Gams, W., K. O'Donnell, H.-J. Schroers, and M. Christensen. 1998. Generic classification of some more hyphomycetes with solitary conidia borne on phiatides. Canadian Journal of Botany 76:1570-1583.

Gatesy, J., G. Amato, M. Norell, R. DeSalle, and C. Hayashi. 2003. Combined support for wholesale taxic atavism in gavialine crocodytians. Systematic Biology 52:403-422.

Gatesy, J., P. O'Grady, and R. H. Baker. 1999. Corroboration among data sets in simultaneous analysis: Hidden support for phylogenetic relationships among higher-level artiodactyt taxa. Cladistics 15:271-3 13. 125

Gernandt, D. S., and J. K. Stone. 1999. Phylogenetic analysis of nuclear ribosomal DNA places the nematode parasite, Drechmeria coniospora, in Clavicipitaceae. Mycologia 91:993-1000.

Glenn, A. E., C. W. Bacon, R. Price, and R. T. Hanlin. 1996. Molecular phylogeny of Acremonium and its taxonomic implications. Mycologia 88:369-383.

Gontcharov, A. A., B. Mann, and M. Melkonian. 2004. Are combined analyses better than single gene phylogenies? A case study using SSU rDNA and rbcL sequence comparisons in the Zygnematophyceae (Streptophyta). Molecular Biology and Evolution 21:612-624.

Graybeal, A. 1998. Is it better to add taxa or characters to a difficult phylogenetic problem? Systematic Biology 47:9-17.

Hillis, D. M., and J. J. Bull. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Systematic Biology 42:182-192.

Hodge, K. T. 2003. Clacicipitaceous anamorphs. Pages 75-123 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Huelsenbeck, J. P., J. J. Bull, and 119. 1996. A likelihood ratio test to detect conflicting phylogenetic signal. Systematic Biology 45:92-98.

Huelsenbeck, J. P., B. Larget, R. E. Miller, and F. Ronquist. 2002. Potential applications and pitfalls of bayesian inference of phylogeny. Systematic Biology 51:673-688.

Huelsenbeck, J. P., B. Larget, and D. Swofford. 2000. A compound poisson process for relaxing the molecular clock. Genetics 154:1879-1892.

Huelsenbeck, J. P., and B. Rannala. 1997. Phylogenetic methods corn of age: Testing hypotheses in an evolutionary context. Science 276:227-232.

Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.

Huelsenbeck, J. P., F. Ronquist, R. Nielsen, and J. P. Boliback. 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294:2310-2314. 126

Hywel-Jones, N. L. 2002. Multiples of eight inCordycepsascospores. Mycological Research 106:2-3.

Kauff, F., and F. Lutzoni. 2002. Phylogeny of the Gyalectales and Ostropales (Ascomycota, Fungi): among and within order relationships based on nuclear ribosomal RNA small and large subunits. Molecular Phylogenetics and Evolution 25:138-156.

Kluge, A. G. 1989. A concern for evidence and a phylognetic hypothesis of relationships among Epicrates (Boidea, Serpentes). Systematic Zoology 3 8:7- 25.

Kluge, A. G. 1998. Total evidence or taxonomic congruence: cladistics or consensus classification. Cladistics 14:151-158.

Kiuge, A. G., and A. J. Wolf. 1993. Cladistics: What's in a word? Cladistics 9:183- 199.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364.

Koepfli, K.-P., and R. K. Wayne. 2003. Type I STS markers are more informative than cytochrome b in phylogenetic reconstruction of the Mustelidae (Mammalia: Carnivora). Systematic Biology 52:571-593.

Kolaczkowski, B., and J. W. Thorton. 2004. Performance of maximum parsimony and likelihood phylogenetics when evolution is heterogeneous. Nature 431:980- 984.

Landvik, S., 0. E. Briksson, and M. L. Berbee. 2001. -a fungal dinosaur? Evidence from beta-tubulin amino acid sequences. Mycologia 93:1151-1163.

Larget, B., and D. Simon. 1999. Markov chain Monte Carlo algorithms for the Bayesian analysis of phylogenetic trees. Mol Biol Evol 16:750-759.

Lemmon, A. R., and E. C. Moriarty. 2004. The importance of proper model assumption in Bayesian pylogenetics. Systematic Biology 53:265-277.

Levasseur, C., and F.-J. Lapointe. 2001. War and peace in phylogenetics: A rejoinder on total evidence and consensus. Systematic Biology 50:881-891.

Liu, Y., S. Whelen, and B. Hall. 1999. Phylogenetic relationships among ascomycetes: evidence from an RNA polymerse II subunit. Molecular Biology Evolution 16:1799- 1808. 127

Lockhart, P. J., M. A. Stell, M. D. Hendy, and D. Penny. 1994. Recovering evolutionary trees under a more realistic model of sequence evolution. Molecular Biology and Evolution 11:605-612.

Luttrel, B. S. 1955. The ascostromatic ascomycetes. Mycologia 47:511-532.

Lutzoni, F., F. Kauff, C. J. Cox, D. McLaughlin, G. Celio, B. Dentinger, M. Padamsee, D. Hibbett, T. Y. James, B. Baloch, M. Grube, V. Reeb, V. Hofstetter, C. Schoch, A. E. Arnold, J. Miadlikowska, J. Spatafora, D. Johnson, S. Hambleton, M. Crockett, R. Shoemaker, G.-H. Sung, R. Lucking, T. Lumbsch, K. O'Donnell, M. Binder, P. Diederich, D. Ertz, C. Gueidan, K. Hansen, R. C. Harris, K. Hosaka, Y.-W. Lim, B. Matheny, H. Nishida, D. Pfister, J. Rogers, A. Rossman, I. Schmitt, H. Sipman, J. Stone, J. Sugiyama, R. Yahr, and R. Vilgalys. 2004. Assembling the fungal tree of life: progress, classification, and evolution of subcellular traits. Am. J. Bot. 91:1446-1480.

Maddison, D. R., and W. P. Maddison. 2000. MacClade 4: Analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts.

Mitchell, A., C. Mitter, and J. C. Regier. 2000. More taxa or more characters revisited: combining data from nuclear-protein encoding genes for phylogenetic analyses of Noctuoidea (Insecta: Lepidoptera). Systematic Biology 49:202-224.

Miyamoto, M. M., and W. M. Fitch. 1995. Testing species phylogenies and phylogenetic methods with congruence. Systematic Biology 44:64-76.

Nylander, J. A. A., F. Ronquist, J. P. Huelsenbeck, and N.-A. J. Luis. 2004. Bayesian phylogenetic analysis of combined data. Systematic Biology 53:47-67.

O'Donnell, K., and Cigelnik, E. 1997. Two divergent intragenomic rDNA ITS2 types within a monophyletic lineage of the fungus Fusarium are nonorthologous. Molecular Phylogenetics and Evolution 7:103-116.

Posada, D., and K. A. Crandall. 1998. ModelTest: Testing the model of DNA substitution. Bioinformatics 14:817-818.

Reeb, V., F. Lutzoni, and C. Roux. 2004. Contribution of RPB2 to multilocus phylogenetic studies of the euascomycetes (Pezizomycota, Fungi) with special emphasis on the -forming Acarosporaceae and evolution of polyspory. Molecular Phylogenetics and Evolution 32:1036-1060. 128

Rehner, S. A., and G. J. Samuels. 1995. Molecular systematics of the Hypocreales: a teleomorph gene phylogeny and the status of their anamorphs. Canadian Journal of Botany 73 (Suppi. l):S816-S823.

Rodrigo, A. G., M. Kelly-Borges, P. R. Bergquist, and P. L. Bergquist. 1993. A randomization test of the null hypothesis that two cladograms are sample estimates of a parametric . N. Z. J. Bot. 31:257-268.

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

Rokas, A., and S. B. Carroll. 2005. More Genes or More Taxa? The Relative Contribution of Gene Number and Taxon Number to Phylogenetic Accuracy. Molecular Biology and Evolution 22:1337-1344.

Rosenberg, M. S., and S. Kumar. 2001. Incomplete taxon sampling is not a problem for phylogenetic inference. PNAS 98:10751-10756.

Rossman, A. Y., G. J. Samuels, J. S. Rogers, and R. Lowen. 1999. Genera of Bionectriaceae, Nectriaceae and Hypocreaceae (Hypocreales, Ascomycetes). Studies in Mycology No. 42.

Simmons, M. P., K. M. Pickett, and M. Miya. 2004. How meaningful are Bayesian support values? Molecular Biology and Evolution 21:188-199.

Sorenson, M. D. 1998. TreeRot, version 2. Univ. Michigan, Ann Arbor.

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes I. The Clavicipitales-Hypocreales connection. Mycologia 85:912-922.

Spatafora, J. W., B. Volkmann-Kohlmeyer, and J. Kohlmeyer. 1998. Independent terrestrial origins of the Halosphaeriales (marine Ascomycota). American Journal of Botany 85:1569-1580.

Sullivan, J. 1996. Combining data with different distributions of among-site variation. Systematic Biology 45:375-380.

Sung, G.-H., J. W. Spatafora, R. Zare, K. T. Hodge, and W. Gams. 2001. A revision of Verticillium sect. Prostrata. II. Phylogenetic analyses of SSU and LSU nuclear rDNA sequences from anamorphs and teleomorphs of the Clavicipitaceae. Nova Hedwigia 72:311-328. Suzuki, Y., G. V. Glazko, and M. Nei. 2002. Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. PNAS 99:16138-16143.

Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.OblO. Sinauer, Sunderland, Massachusetts.

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673-4680.

Thornton, J. W., and R. DeSalle. 2000. A new method to localize and test the significance of incongruence: Detecting shuffling in the nuclear receptor superfamily. Systematic Biology 49:183-201.

Vilgalys, R., and B. L. Sun. 1994. Ancient and recent patterns of geographic speciation in the oyster mushroom Pleurotus revealed by phylogenetic analysis of ribosomal DNA sequences. PNAS 91:4599-4603.

White, J. F., Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora. 2003. Historical perspectives: Human interactions with Clavicipitalean fungi. Pages 1-16 in Clavicipitalean Fungi (J. F. White, Jr., M. Backlund, N. L. Hywel- Jones, and J. W. Spatafora, eds.). Marcel Decker, New York.

Whittingham, L. A., B. Slikas, D. W. Winkler, and F. H. Sheldon. 2002. Phylogeny of the tree swallow genus, Tachycineta (Ayes: Hirundinidae), by Bayesian analysis of mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 22:430-441.

Wiens, J. J. 1998. Combining data sets with different phylogenetic histories. Systematic Biology 47:568-58 1.

Wiens, J. J. 2003 a. Incomplete taxa, incomplete characters, and phylogenetic accuracy: Is there a missing data problem? Journal of Vertebrate Paleontology 23: 297-3 10.

Wiens, J. J. 2003b. Missing data, Incomplete taxa, and phylogeneticaccuracy. Systematic Biology 52:528-538.

Yoder, A. D., J. A. frwin, and B. A. Payseur. 2001. Failure of the ILDto determine data combinability for slow loris phylogeny. Systematic Biology 50:408-424.

Zwickl, D. J., and D. M. Hillis. 2002. Increased taxon sampling greatly reduces phylogenetic error. Systematic Biology 51:588-598. 130

Chapter 5

[YLOGENETIC CLASSIFICATION OFCORD YCEPS AND THE CLAVICIPITACEOUS FUNGI

Gi-Ho Sung', Nigel L. Hywel-Jones2, Jae-Mo Sung3, and Joseph W. Spatafora'

1Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR

9733 1-2902, USA

2 Laboratory, National Center for Genetic Engineering and Biotechnology,

Science Park, Pathum Thani, Thailand

'Division of Environmental Biology, Kangwon National University, Chuncheon 200-

701, Republic of Korea

For submission to Mycologia 131

Abstract

Cordyceps is a large genus of ascomycetous fungi with over 400 species. It is a member of Clavicipitaceae based on its cylindrical asci, thickened ascus apices, and fihiform ascospores, which often disarticulate into partspores. Cordyceps is distinguished from other genera (e.g., Claviceps, Epichloë, and Torrubiella) of the family by the production of well-developed often stipitate stromata and its ecology as a pathogen of arthropods and truffles. Current classifications emphasize orientation of perithecia, ascospore morphology and host affiliation. To refine the current classification of Cordyceps and clavicipitaceous fungi, the phylogenetic relationships of 162 taxa were estimated based on five loci: the nuclear ribosomal small and large subunits (nrSSU and nrLSUJ, elongation factor lix (EF-la), and the largest and second largest subunits of RNA polymerase II (RPB] and RPB2). Our results support the existence of three clavicipitaceous clades and reject the monophyly of both

Clavicipitaceae and Cordyceps. Most diagnostic characters used in current classifications were not supported as being phylogenetically informative; the characters that were most consistent with the phylogeny are texture, pigmentation and morphology of stromata. Therefore, we revise the taxonomy of Cordyceps and the

Clavicipitaceae to be consistent with the multi-gene phylogeny. The family

Cordycipitaceae is emended based on placement of the type of Cordyceps, C. militaris, and includes Cordyceps species that possess brightly colored fleshy stromata. The new family Ophiocordycipitaceae is proposed based on Ophiocordyceps 132

(Petch), which we emend. The majority of species in this family produce darkly pigmented, tough to pliant stromata that often possess sterile apices. A new genus

Elaphocordyceps is proposed for a subclade of Ophiocordycipitaceae, which includes all the species of Cordyceps that parasitize the truffle genus Elaphomyces. The

Clavicipitaceae s.s. is emended and restricted to the core dade containing Claviceps.

In addition, a new genus Metacordyceps is proposed for Cordyceps species that are closely related to the grass symbionts in Clavicipitaceae s.s. 133

Introduction

Cordyceps (L: Fr) Link is the most diverse genus in the family Clavicipitaceae in terms of number of species and host range (Kobayasi, 1941; Kobayasi, 1982;

Spatafora and Blackwell, 1993). There are estimated to be more than 400 species

(Index Fungorum; Kobayasi, 1982; Mains, 1958; Stensrud et al., 2005) although its diversity is expected to be underestimated in nature (Hawksworth and Rossman,

1997). Its host range is broad, ranging from eight orders of arthropods to the truffle genus Elaphomyces, although most species are restricted to a single host species or a set of closely related host species (Kobayasi, 1941; Kobayasi, 1982; Mains, 1957;

Mains, 1958). Its distribution is cosmopolitan, including all terrestrial regions except

Antarctica, with the height of species diversity occurring in subtropical and tropical regions, especially East Asia (Kobayasi, 1941; Samson et al., 1988). The genus belongs to the family Clavicipitaceae, based on its cylindrical asci, thickened ascus apices, and filiform ascospores that often disarticulate into partspores (Hywel-Jones,

2002; Kobayasi, 1982; Mains, 1958; Rossman et al., 1999). It is characterized and distinguished from other genera of the family by its production of superficial to completely immersed perithecia on stipitate and often clavate to capitate stromata and its ecology as a pathogen of arthropods and truffles (Kobayasi, 1941; Mains, 1957;

Mains, 1958; Rogerson, 1970).

The subgeneric classification of Cordyceps is primarily based on the taxonomic study of Kobayasi (1941). Kobayasi proposed three subgenera (C. subg. 134

Cordyceps, C. subg. Ophiocordyceps, and C. subg. Neocordyceps) mainly based on

the orientation of perithecia and the morphology of ascospores and partspores. Later,

Mains (1958) expanded the subgeneric classification with a different emphasis on the

diagnostic characters and erected two additional subgenera; C. subg. Racemella and C.

subg. Crytpocordyceps. The difference of these two classifications is exemplified by their contrasting treatments of the genus Ophiocordyceps Petch. Ophiocordyceps was originally erected by Petch (1931) for Cordyceps species that possess nondisarticulating ascospores and a less pronounced apical ascus apex. It was not accepted by subsequent workers and later assigned to C. subg. Ophiocordyceps by

Kobayasi (1941) to accommodate species that possess nondisarticulatingascospores and asci with a less pronounced ascus apex. In contrast, Mains (1958) placed only the type, C. blattae Petch, in C. subg. Ophiocordyceps based on its lack of a thickened ascus apex, thus deemphasizing the importance of asco spore disarticulation at the subgeneric level. Currently, the subgenera C. subg. Cordyceps, C. subg.

Ophiocordyceps, and C. subg. Neocordyceps have been accepted by most taxonomists

(Artj ariyasripong et aL, 2001; Hywel-Jones, 2002; Stensrud et al., 2005; Sung and

Spatafora, 2004; Zang and Kinjo, 1998) following the concept of Kobayasi with the addition of C. subg. Bolacordyceps that is characterized by its possession of bola- ascospores (Eriksson and Hawksworth, 1986).

In addition to the morphological characters, host affiliation has played important roles in the classification of Cordyceps (Kobayasi, 1982; Massee, 1895).

Cordyceps species that parasitize the truffle genus Elaphomyces have often been 135 recognized as a taxonomic unit in the classification. The genus Cordylia Fr. was once

assigned for the mycogenous Cordyceps species (Massee, 1895) although it is a homonym of Cordylia Lour. (Leguminosae). Kobayasi (1941; 1982) also recognized the mycogenous Cordyceps species as taxonomic units (e.g., C. subg. Cordyceps sect.

Cystocarpon subsect. Eucystocarpon ser. Mycogenae) and emphasized the utility of host affiliations in delimiting closely related species of arthropod pathogens.

Several phylogenetic studies employing ribosomal DNA (Artjariyasripong et al., 2001; Stensrud et al., 2005; Sung et al., 2001) have been conducted to test and refine the classification of Cordyceps. Ribosomal DNA sequences have limited resolution power, however, and limited conclusions could be drawn regarding systematics of the genus. Recent phylogenetic studies (Chapter 4) based on multiple independent loci provided a greater level of resolution and support, and revealed that neither Cordyceps nor the Clavicipitaceae is monophyletic. Three monophyletic groups of the clavicipitaceous fungi have been recognized, all of which include species of Cordyceps. These results reject the current subfamilial classification (Diehi,

1950) and indicate that the phylogenetic diversity of Cordyceps is representative of the entire family Clavicipitaceae (Chapter 4). Therefore, a new classification of

Cordyceps and the clavicipitaceous fungi is necessary to reflect the current hypotheses of phylogenetic relationships and to be predictive in nature.

Here, we conducted the most extensive multi-gene phylogenetic analyses to provide a basis of phylogenetic classification of Cordyceps and clavicipitaceous fungi.

The main objectives of this study are 1) to reassess the morphological traits used in the 136 current classifications of Cordyceps in the context of the phylogeny, 2) to investigate the taxonomic utility of the anamorphic forms in classification of Cordyceps and better understand the teleomorph-anamorph connections, and 3) to revise the genus

Cordyceps and the clavicipitaceous fungi to be consistent with phylogenetic relationships.

Materials and Methods

Taxon and character sampling

A total of 162 taxa was selected in the Clavicipitaceae and other families of the

Hypocreales with Glomerella cingulata and Verticillium dahliae included as outgroups (Table 5.1). In constructing the full data set (1 62-taxon data set), we used five genes: the nuclear ribosomal small and large subunits (nrSSU and nrLSU), elongation factor Ia (EF-la), the largest and second largest subunits of RNA polymerase H (RPBJ and RPB2). Sequences from 91 taxa were obtained from the previous work (Chapter 4), however we excluded /3-tubulin and ATP Synthase subunit

6 (mtATP6) in further sampling due to the paralogy of /3-tubulin, as well as the large size of the group I intron in mtA TP6. DNA extractions were conducted usinga

FastDNA kit (Qbiogene) following the manufacturers instructions, with minor modifications. Polymerase chain and sequencing reactions were performedas previously described (Chapter 4). Information pertaining to voucher numbers concerning the sequences is provided in Table 5.1. 137

Sequence alignment and phylogenetic analyses

Sequences were edited using SeqEd 1.0.3 (Applied Biosystems Inc., Foster

City, CA) and contigs were assembled using CodonCode Aligner 1.4 (CodonCode

Inc., Dedham, MA). Sequences of each gene partition were initially aligned with

Clustal W 1.64 (Thompson et al., 1994) and appended to an existing alignment

(Chapter 4). This initial alignment was manually edited as necessary in MacClade 4.0

(Maddison and Maddison, 2000). All five gene regions were concatenated into a single, combined data set with ambiguously aligned regions excluded from phylogenetic analyses.

In order to detect incongruence among the five individual data sets, bootstrap proportions were used for each individual data set with the 108 core taxa (108-taxon data set) that was complete for all five genes (Table 5.1). Bootstrap proportionswere determined in a maximum parsimony framework using the program PAUP* 4.OblO

(Swofford, 2002). Only parsimony informative characters were used with the following search options: 100 replicates of random sequence addition, TBR branch swapping, and MulTrees OFF. The incongruence was assumed to be significant iftwo different relationships for the same set of taxa were both supported with greater than

70 % bootstrap proportions (Mason-Gamer and Kellogg, 1996; Wiens, 1998). withGenBank.Table C4. 5.1. AnThe asterisk list of cultures (*) denotes and specimensthe used in this sequences determined by the first author of this study. Sequences determined in Chapter 4 are denoted study. GenBank accession numbers are provided for the sequences available in 138 Aphysiostroma stercorarium Taxon ATCC 62321 Specimen voucher a on cow dung Hostlsubstratum AF543769 nrU AF543792 nrLU GeneBank accession number AF543782 Ji-1a AY489633 It'iJ1 Kt'152 C4 BalansiaAschersonia epichloe placentabadia AEG 96-15a PoaceaeScale insect (l-Jemiptera)(Hemiptera) C4 * C4 C4 * C4 * C4 * BjonectriaBalansiapilulaeformisBalansia henningsiana cf. aureofulva GJSAEGGAM 71-328 97-2 16112 PoaceaePanicumPoaceae sp. (Poaceae) AF543764AY545723 C4 AF543788AY545727 C4 AY489610 C4 AY489643 C4 C4 ClavicepspaspaliClavicepsfusiformisBionectria ochroleuca ATCCCBS 114056 1389226019 Poaceaeon bark AY489684 C4C4 AY489176 C4 AY48961 I C4 C4C4 C4 Claviceps purpurea GAM 12885 PoaceaeDactylis glomerata (Poaceae) AF543765 C4 * AF543789 C4 AF543778 C4 AY489648 C4 C4 Cordyceps agriotaacicularis ARSEFOSC 5692110988110987 Larva (Coleoptera) C4 * C4 * * C4 * * C4 * C4 Cordyceps bifusisporaaphodii EFCCARSEFEFCC 8260 5498 5690 PupaAphodius (Lepidoptera) hewitli (Coleoptera) C4 * C4 * C4 * * C4 * Cordyceps cardinaliscapitatabrunneapunctata OSCNHJ 125659360971233 LarvaElaphomycesColeoptera (Lepidoptera) sp. (Euascomycetes) AY1AY489689 84973 C4 AYIAY489721 84962 C4 AY489615 C4 AY489649 C4 C4 CordycepsCordyceps cf. cardinalisochraceosiromataacicularis ARSEFNHJOSC 12592 569093610 ColeopteraLepidopteraLarva (Lepidoptera) AY184974 C4 * AY184963 C4 * C4 * C4C4 * C4C4 * Table 5.1. (Continued) 139 CordycepsCordycepscf.takaomontana coccidiicolachiamydosporia CBSNHJ12623 101244 ScaleEggLepidoptera of Insect mollusc (flemiptera) (Diplopoda) ABO3 1195 C4 * ABO3 1196 C4 * C4 * C4 * C4 * Cordyceps entomorrhizaelongatacochlidiicola KEWOSC 110989 53484 LarvaPupa (Lepidoptera) (Coleoptera)(Lepidoptera) AB027331 * AB027377 * * * * CordycepsCordycepsgracilisCordycepsfracta gracilis EFCCOSC 110990 31018572 LarvaElaphomyces (Lepidoptera) sp. (Euascomycetes) C4 * C4 * C4 * C4 ** C4 * CordycepsCordycepsCordycepsgunnii heteropoda heteropoda OSCOSC 106404 76404 NymphLarva of (Lepidoptera) Cicada (Hemiptera) AY489690AF339572 * AY489722AF339522 * AY48961AY489616 7 AY489651AY489650 * C4 * CordycepsCordyceps irangiensis inegoensis NHJ 1257512572 AntNymph (Hymenoptera) of Cicada (Hemiptera) AB027322 C4C4 AB027368 C4 C4 C4 C4 CordycepsjezoensisCordycepsjaponica OSC 110991 Elaphomyces sp. (Euascomycetes) AB027320 C4 * AB027365 C4 C4 C4 * C4 * Cordyceps konnoana EFCC 73157295 Larva (Coleoptera) * * * * * Cordyceps kyushuensisliangshanensis EFCC 152314525886 LarvaPupa (Lepidoptera) (Lepidoptera) * * * * Cordyceps longissimalongisegmentismelolonthae OSCEFCC 110993110992 6814 PhyllophagaNymphsElaphomyces of Cicada sp. sp. (Coleoptera) (Euascomycetes) (Hemiptera) C4 C4 * C4 * C4 * * Cordyceps nutansmilitarisnigrella OSCEFCC 93623110994 9247 StinkLarvaMoth bug(Lepidoptera).(Lepidoptera) (Hemiptera) AYI 84977 C4 * AY1 84966 C4 * C4 * C4 * AY54573 1 * Table 5.1. (Continued) 140 Cordyceps ophioglossoides OSC 106405 - Elaplzomyces sp. (Euascomycetes) AY489691 AY489723 AY48961 8 AY489652 C4 CordycepsCordycepsparadoxa raveneliirobertsii KEWOSC 110995 27083 LepidopteraPhyllophagaNymphs of Cicada sp.(Coleoptera) (Hemiptera) AB027323 C4 AB027369 C4 * C4 * C4 C4 Cordyceps sinensisscarabaeicola EFCC 72875689 LarvaScarab (Lepidoptera) (Coleoptera) AF339574 * * AF339524 * C4 * C4 * C4 * Cordyceps sp.sobolifera EFCC 2131 EFCCKEW 788422131 PupaNymph (Lepidoptera) of Cicada (Herniptera) * * * * * Cordyceps sp. EFCC 25352135 EFCC 25352135 ColeopteraPupa (Lepidoptera) * * * * * Cordyceps sp. EFCC 56935197 EFCCEFCC 5693 5197 Lepidoptera * * * * CordycepsCordyceps sp. sp. NHJ NI-V 10684 10627 NHJNW 1062710684 LepidopteraLepidoptera * * * * * CordycepsCordycepssp.NHJ sp. NHJ 12522 12118 NHJ 1252212118 ColeopteraLepidoptera * * * * Cordycepssp.NHJCordyceps sp. NHJ 1258112529 NHJN1-IJ 12581 12529 Coleoptera * * * * * * * Cordyceps sp. OSCNHJ 12582110996 OSCNW 110996 12582 LepidopteraColeoptera * * * * * * * CordycepsCordyceps sphecocephala sp. OSC 110997 OSC 110998110997 BeeAnt (1-lymenoptera) (Hynlenoptera) C4 * C4 * C4 C4 C4 Cordyceps stylophorastaphylinidicola OSCEFCC 111000110999 5718 LarvaColeoptera (Coleoptera) C4 * C4 * C4 * * C4 * C4 * CordycepsCordyceps taii subsessilissupeificialis ARSEFMICHOSC 5714111001 36253 Larva (Coleoptera) AF543763 C4 * AF543787 C4 AF543775 C4 C4 * C4 Table 5.1. (Continued) 141 Cordyceps takaomontanatricentri Lepidoptera AB027330AB044631 AB027376AB044637 Cordyceps variabilisunilateralisluberculata ARSEFNHJOSC 12523111002 5365 DipteraAntMothSpittlebug (Hymenoptera) (Lepidoptera) (Hemiptera) C4 C4 C4 C4C4 C4 CosmosporaCordyceps yakushimensisvariabilis coccinea CBSOSC 114050309.85111003 InonotusNymphDiptera nodulosus of cicada (Hymenomycetes) (Hemiptera) AY489702AB044632 * AY489734AB044633 * AY489629 * AY489667 * C4 * GlomerellaEpich/oeEngyodontiumGlomerella typhina cingulata cingulataaranearum FAUCBSIATCC 513 14054 56429 FragariaFestucaSpider rubra sp. (Arachnida) (Poaceae) AF543762AF339576 U32405 AF543786AF339526UI 7396 AF543773AF543777 C4 AY489659AY489653 C4 C4 Haptocillium zeosporumsinensebalanoides CBSCBS 250.82335.80 567.95 NematodaFragaria sp. AF339589AF339594AF339588U48427 AF339540AF339545AF339539U48428 AF543772 C4 C4C4 C4 HydropisphaeraHirsute/la sp. NHJ erubescens 1252712525 CBSATCCNHJ 1252712525102038 36093 onCordylineColeoptera bark banksii (Laxmanniaceae) AY545722 C4 AY545726 C4 C4 C4 AY545732 C4 HypocrellaHypocreaHyinenostilbeHydropisphaerapeziza schizostachyi lutea aurantiaca NHJATCC 1260512574 208838 ScaleonAnt decorticated (l-iymenoptera) insect (Hemiptera) conifer wood AY489698AF543768 C4 AY489730AF543791 C4 AF543781AY489625 C4 AY489662AY489661 C4 C4 IsariaHypomycespolyporinusHypocrella cf farinosa sp. GJS 89104 OSCATCCGJS 111004 76479 89104 TrametesScale insect versicolor (Flemiptera) (Hymenomycetes) AF543771 C4 * AF543793 C4 * AF543784 C4 * AY489663 C4 * C4 IsariaIsariafarinosa tenulpes OSC 111007111006111005 PupaPupa (Lepidoptera) (Lepidoptera) C4 C4 C4C4 C4 C4 Table 5.1. (Continued) 142 Lecanicillim attenuatumaranearum CBS 402.78726.73a leafSpider litter of (Arachnida) Acer saccha ruin $ * * * * LecanicillimpsalliotaeLecanicillimLecanicillimfusisporum psalliotae CBS 101270363.86164.70 AgaricusSoilColtricia bisporus perennis (Basidiomycetes) (Basidiomycetes) C4 * C4 * C4 * C4 * C4 LeuconectrjaLecanicilliumLecanicillimpsalliotae clusiaeantillanum ATCCCBS 532.81350.85 22228 SoilAgaicforest soilmushroom (Hymenomycetes) AY489700AF339585AF339609 AY489732AF339536AF339560 AY489627 C4 C4 C4 MetarhiziumMariannaea pruinosaalbumanisopliae var. majus ARSEF 314520825413 Oryctes1-lemipteraPupa (Lepidoptera) rhinoceros (Coleoptera) AYIAF339579 84979 C4 AYIAF339530 84968 C4 AF543774 C4 AY489664 C4 C4 MyriogenosporaMicrohilumMetarhiziumfiavoviride oncoperae atramentosa var. minus AEGAFSEFARSEF 96-32 43582037 AndropogonNilaparvataOncopera lugens virginicusintricata (Homoptera) (Lepidoptera) (Poaceae) AY489701AF339580 * AY489733AF339531 * AY489628 C4 * AY489665 C4 * C4 * NomuraeaNectria cinnabarina alypicola CBSCBS 114055 744.73 BetulaSpider sp. (Arachnida)(Betulacaceae) U32412 * U00748 * AF543785 * AY489666 * C4 * OphionectriaNomuraea rileyi trichospora CBSCBS 109876 806,71 lianaLepidoptera AF543766AY624205 * AF543790AY624250 * AF543779 * AY489669 * C4 * Paecilomyces cicadaecarneus KEWCBS 399.59239.32 78839 NymphSoilDune of sand Cicada (Hemiptera) * * * * * Paecilomyces cinnamomeus CBS 398.86828.88 livingScale insectleaf of (Hemiptera) Syzygiumjambos (Myrtaceae) AY624174 * AY624213AY624214 * * Paecilomyces lilacinus CBSARSEF 431.87 2181 Meloidogyne sp. (Nematoda) AY6241AF339583 88 AF339534 * * * PhytocordycepsPaecilomyces marquandiililacinus ninchukispora CBS 182.27284.36 Soil AY624189 * AY624227 * * * * Table 5.1. (Continued) 143 PochoniaPhylocordyceps chiamydosporiabulbillosa ninchukispora CBS 145.70504.66 SoilRoot of Picea abies (Pinaceae) AF339593 * AF339544 * * * * PseudonectriaPochonjaPochonia rubescensgonioides rousseijana CBS 114049464.88891.72 BuxusHeleroderaNematoda sempervirens avenae (Nematoda) (Buxaceae) AF543767AF339599 * AF339550LJ17416 * AF543780 C4 * C4 * C4 * RoumegueriellaRotferophthora angustisporarufula GJSCBS 91-164101437346.85 GloboderaRotifer (Rotifera) rostochiensis (Nematoda) AF339584 C4 AF339535 C4 AF543776 C4 AY489670 C4 C4 SimpliciliumShimizuomycesShimizuomycesparadoxus lanosoniveum paradoxus CBSEFCC 704.86 65646279 HemileiaSmilax sieboldii vastatrix (Smilacaceae) (Urediniomycetes) AF339602 C4 AF339553 C4 C4 C4 C4 SiinplicilliumSimplicilium lanosoniveum latnellicola CBS 101267116.25 AgaricusHemileia bisporusvastatrix (Hymenomycetes)(Urediniomycetes) AF339601AF339603 * AF339552AF339554 * C4 * C4 C4 SphaerostilbellaSimplicilliumTolypocladium obclavatunz berkeleyana parasiticum ARSEFCBSCBS102308 311.74 3436 BdelloidPolyporeAir above (Hymenomycetes) sugarcane (Rotifera) field AF543770 * U00756 * AF543783 * * * Torrubiella wallaceiconfragosaratticaudata CBSARSEF 101247101237 1915 LarvaSpiderCoccus (Lepidoptera) (Arachnida) viridis (Hemiptera) AYAF339604 184978 C4 AY1AF339555 84967 C4 C4 C4 C4 Verticillium epiphytumdahliae CBSATCC 154.61384.81 16535 HemHemileiaCrataegus ileia vastatrix vastatrixcrus-galli (Urediniomycetes)(Urediniomycetes) (Rosaceae) AF339596AY489705 C4 AF339547AY489737 C4 AY489632 C4 * AY489673 C4 * C4 * ViridisporaVerticillium diparietisporaincurvumsp.CBS 102184 CBS 102797102184460.88 CrataegusSpiderGanoderma (Arachnida) crus-galli lipsiense (Rosaceae) (Hymenomycetes) AY489703AF339600 * AY489735AF339551 * AY489630 C4 * AY489668 C4 * C4 * 144 Herbarium,KEW,H.Manassas,a Miller mycology Mycological VA; Corvallis,AEG, CBS, collection A. CentraalbureauOR. HerbariumE. Glenn of Royal personal Athens, Botanicalvoor collection; GA;Schimmelcultures, Garden,FAU, ARSEF, F. KEW, A. Uecker USDA-ARSUtrecht, Surrey, personal UK;the Netherlands;Collection MICH, collection, University of GJS,EFCC, Entompathogenic G. of Entomopathogenic J. Michigan Samuels FungalpersonalHerbarium, cultures,Fungal collection; Culture NHJ, Collection, Nigel Ithaca, NY; ATCC, American Type Culture Collections, Ann Arbor, MI; OSC, Oregon State University Hywel-Jones personal collection;Chuncheon, Korea; GAM, Julian 145

Maximum and weighted parsimony (MP and WP) analyses were conducted on the combination of five individual data sets for the 162 taxa (Table 5.1). All characters were equally weighted and unordered in the MP analyses. For the WP analyses, the unambiguously aligned regions were subjected to symmetric step matrices for eleven partitions to incorporate the differences in substitution rates and patterns. The eleven partitions consisted ofnrSSU, nrLSUand nine codon positions of three protein-coding genes(EF-la, RPBJandRPB2). Tocalculate the eleven symmetric step matrices, changes between all character states at each partition were summarized using the "full detail" character status option in PATJP* 4.OblO (Swofford, 2002). The proportional frequency of changes was calculated using the program STMatrix 2.1 (available at http://www.lutzonilab.net/pages/download.shtml) and converted into cost of changes using a negative natural logarithm (Felsenstein, 1981; Wheeler, 1990). MP and WP analyses were performed using only parsimony informative characters with the following settings: 100 replicates of random sequence addition, TBR branch swapping, and MulTrees ON. The phylogenetic confidence was assessed by nonparametric bootstrapping (Felsenstein, 1985). A total of 1000 bootstrap replicates were used to calculate bootstrap proportions; bootstrapping used the same search options as in the MP and WP analyses with five replicates of random sequence addition per bootstrap replicate.

Bayesian Metropolis coupled Markov chain Monte Carlo (B-MCMCMC) analyses were performed on combined dataset using MrBayes 3.0b4 (Huelsenbeck and

Ronquist, 2001). The appropriate model and parameters of substitution in each 146 partition were selected using the program ModelTest 3.06 to accommodate site- specific models of rate variation (Posada and Crandall, 1998). In estimating the likelihood of each tree, we used the general time-reversible model, with invariant sites and gamma distribution (GTR+I+F) and employed the model separately for each partition. In an initial analysis, a B-MCMCMC analysis with 10,000,000 generations and four chains was conducted in order to test the convergence of log-likelihood.

Trees were sampled every 100 generations, for a total of 100,000 trees. For a second analysis, ten independent Bayesian runs with 1,000,000 generations and different starting trees were conducted to reconfirm the adequate log-likelihood convergence and mixing of chains. Convergence of log-likelihoods was assessed by inspecting the

MrBayes log file. The first 1,000 trees in the 10,000,000 generation analysis were identified as the burnin and deleted to exclude trees prior to the convergence of log- likelihood. A total of 99,000 trees were imported in PAUP* and used to generate a

50% majority-rule consensus tree and calculate posterior probabilities.

In addition to the analyses with 1 62-taxon data set, a series of analyses were conducted in MP, WP and Bayesian frameworks with different taxon samplings (108- and 153-taxon data sets) to address the potential adverse effects of missing data on phylogenetic inferrence. Previous phylogenetic and simulation studies demonstrated that the phylogenetic hypothesis is not affected if less than 50% characters are missing for each taxon in the phylogenetic analyses (Philippe et al., 2004; Wiens, 2003b). In this study, we assumed that the phylogenetic analysis is not confounded if the taxa were complete for at least three gene partitions out of five. Therefore, a total of nine 147 taxa (Table 5.1) in the 1 62-taxon data set were complete for only two gene partitions and thus excluded to generate the 153-taxon data set. A 108-taxon data set that does not contain any missing data in gene partitions was also prepared to compare the phylogenetic relationships between 108-taxon and 153-taxon analyses. MP, WP and

Bayesian analyses based on 1 62-taxon data set (Figure 5.1 and 5.2) showed that the C. sphecocephala dade is characterized by long branch lengths relative to the rest of the clavicipitaceous fungi. To address the impact of the C. sphecocephala dade on the phylogenetic resolution, we excluded all members of the C. sphecocephala dade

(Table 5.1) from the 153-taxon data set and instead constructed a 148-taxon data set.

Results

Sequence Alignment

The combined data set of five genes consists of 4926 base pairs of sequence data for the 162-taxon data set (nrSSU: 1102 bp, nrLSU: 953 bp, EF-la: 1020 bp,

RPBJ: 803 bp, RPB2: 1048 bp). As a result of excluding ambiguously aligned regions, the alignment comprised 4548 base pairs (nrSSU: 1172 bp, nrLSU: 750 bp, EF-la:

1014 bp, RPBJ: 667 bp, RPB2: 1045 bp), 1870 of which were parsimony informative

(nrSSU: 223 bp, nrLSU: 224 bp, EF-la: 462 bp, RPBJ: 382 bp, RPB2: 579 bp). A total of 108 taxa are complete for all five genes and the number of complete taxa for each gene is as follows; nrSSU: 158 taxa, nrLSU: 156 taxa, EF-la: 152 bp, RPBJ: 142 taxa, RPB2: 130 taxa (Table 5.1). 148

Testofincongruence and phylogen etic analyses

The reciprocal comparisons of 70% bootstrap trees from individual data sets did not reveal any significantly supported contradictory nodes (data not shown). These results were interpreted as indicating that no strong incongruence existed among the individual data sets that would be indicative of different phylogenetic gene histories

(e.g., lineage sorting or horizontal gene transfer). As a result, all five individual data sets were combined in simultaneous analyses.

Maximum parsimony (MP) analyses of the combined 5-gene, 162-taxon data set resulted in 195 equally parsimonious trees. These trees were 19,461 steps with consistency indices (CI) of 0.1620 and retention indices (RI) of 0.6175. One of 195 equally parsimonious trees is shown in Figure 5.1. Collapsed nodes in the strict consensus tree are denoted with asterisks. Weighted parsimony analyses of the 162- taxon data set resulted in one tree of 34518.15 steps. In the Bayesian analyses, the

10,000,000-generation analysis converged on the plateau of the log-likelihood at approximately around 60,000 generations. The results of ten 1,000,000-generation analyses also showed a convergence on the log-likelihood at approximately 60,000 generations. As a result, the 1,000 trees from the first 100,000 generations were deleted from all eleven analyses to generate a 50% majority rule consensus tree. The topologies from the eleven consensus trees were identical and all analyses converged on the same optimal log-likelihoods. 149

Bionectriaceae Nectriaceae Hypocreaceae 100

EE Clavicipitaceae dade C

Mad.t. Md.oa Ca,.p. d aa,.e.oanna MC.CdO.

C.Synpfl.*.o.ao..BB .MIo oenp*. - V..*AJkaa, .p. CBS 102004

Ø CaSya.p/. NHJ 10047

'14Phylaa.,dya.p. .4,4ukIopo.i

I Shimizuomyces dade I] Claviceps dade

Clavicipitaceae dade A

C. tail dade

JO. C. sphecocephala dade

C. gunnhl

000

I P. hilaclnus dade

C. ophioglossoldes dade

Clavicipitaceae dade B

* C. unilateralls dade

*

100 steps 150

Figure5.1.Phylogenetic relationships among162taxa in the Clavicipitaceae and other families in Hypocreales based on combined evidence of five gene partitions (i.e.nrSSU, nrLSU, EF-la, RPBJandRPB2).The phylogeny is from one of195equally parsimonious trees of19,461steps inferred from maximum parsimony (MP) analyses. Bootstrap proportions of ? 70% are provided above their corresponding nodes and in a thicker line. The internal nodes that were collapsed in the strict consensus tree are marked with asterisks (*).

A 50% majority consensus tree (Figure 5.2) was generated from the

10,000,000-generation analysis. Posterior probabilities of the tree were then

determined to provide a measure of nodal support. Since the topology of WP analyses

is nearly identical to that of the Bayesian consensus tree, the bootstrap proportions of

WP analyses are provided above the corresponding nodes in Figure 5.2. Previous

studies have shown, that in interpreting the supports of the phylogenetic estimates of relationships, the posterior probability tends to overestimate the phylogenetic confidence (Doaudy et al., 2003a; Lutzoni et al., 2004; Reeb et al., 2004). As a result, the probability can be used as a supplementary indicator to bootstrap proportions. In this study, nodes were considered strongly supported when supported by both bootstrap proportions (BP ? 70) and posterior probabilities (PP? 95).

Phylogenetic relationshipsofthe clavicipitaceous fungi

All MP, WP and Bayesian analyses of thel62-taxon data set of five genes recognized three monophyletic well-supported clades of clavicipitaceous fungi (Figure

5.1 and 5.2), designated here as Clavicipitaceae clades A, B and C (Figures 5.1 and

5.2) following the convention of the previous phylogenetic studies (Chapter 4). These clades are statistically well supported by the bootstrap proportions of the MP and WP analyses (MP-BP and WP-BP) and posterior probabilities (PP) of the Bayesian 151

Nectriaceae Hypocreaceae I

Clavicipitaceae dade C

dade

Claviceps dade

Clavicipitaceae dade A

C ta,,clade

100 C. gunnil dade ) 100 P. lilacinus dade

J

C. ophioglossoides dade

Clavicipitaceae dade B I

100

LI iiiL: C. sphecocephala'.- dade

- C. unhlaferalls dade

0.05 btJtdI./$It. 152

Figure 5.2. Phylogenetic relationships among 162 taxa in Clavicipitaceae and other families in Hypocreales based on combined evidence of five genes (i.e. nrSSU, nrLSU, EF-la,RPBJand RPB2). The phylogeny is from a 50% majority consensus tree of 999,000 trees sampled from Bayesian analyses and outgroups (Glomerella cingulata and Verticillium dahliae) are not shown. Posterior probabilities of?: 95% are provided below their corresponding nodes. Bootstrap proportions are obtained in weighted parsimony (WP) analyses and shown above their corresponding nodes for?: 70%. The nodes that were supported with both bootstrap proportions (WP-BP?: 70%) and posterior probabilities (PP?: 95%) are considered strongly supported and in a thicker line.

analyses (dade A: MP-BP = 99%, WP-BP = 100%, PP = 100%, dade B: MP-BP

90%, WP-BP = 96%, PP100%, dade C: MP-BP = 100%, WP-BP = 97%, PP =

100). A sister-group relationship between clades A and B is also strongly supported

(MP-BP = 72%, WP-BP = 73%, PP = 100%). The monophyletic group of dade C and

Hypocreaceae is moderately supported (MP-BP = 66%, WP-BP72%, PP = 100%).

Clavicipitaceae dade A comprises four statistically well-supported subclades.

These are labeled in Figure 5.4 as the C. taii dade (MP-BP85%, WP-BP83%, PP

= 100%), the Claviceps dade (MP-BP = 90%, WP-BP = 100%, PP = 100%), the

Hypocrella dade (MP-BP = 92%, WP-BP93%, PP = 100%) and the Shimizuomyces

dade (MP-BP = 100%, WP-BP = 100%, PP = 100%). As indicated previously

(Chapter 4), internal relationships among these four subclades are not strongly

supported in any analyses (Figure 5.1, 5.2 and 5.4).

Based on relationships in MP analyses the Clavicipitaceae dade B consists of five

major subclades designated as the C. gunnii, C. ophioglossoides, C. sphecocephala, C.

unilateralis and P. lilacinus clades (Figures 5.1, 5.2 and 5.5). Nearly all of the

subclades in dade B are strongly supported by bootstrap proportions and posterior probabilities (C. gunnii dade: MP-BP98%, WP-BP = 98%, PP = 100%, C. ophioglossoides dade: MP-BP = 70%, WP-BP82%, PP = 100%, C. sphecocephala 153

dade: MP-BP = 100%, WP-BP = 100%, PP = 100%, P. lilacinus: MP-BP = 74%,

WP-BP = 100%, PP = 100%). It should be noted, however, the C. unilateralis dade

was not resolved in the MP analyses (Figure 5.1). This lack of resolution was due to

the placement of the C. sphecocephala dade, which is characterized by long branch

lengths relative to the rest of the clavicipitaceous fungi. In the MP analyses, the

internal relationships among major subclades in dade B collapsed (Figure 5.1). This

lack of resolution was due to the multiple placements of the C. sphecocephala dade

either as the basal lineages of the Clavicipitaceae dade B or as a terminal dade nested

within the members of the C. unilateralis dade (data not shown). In previous analyses

(Chapter 4) based on a combined data set with seven genes (five genes included in this

study, with the addition of/3-tubulin and mtA TP6), the C. sphecocephala dade was

placed as a terminal group of the C. unilateralis dade, a finding consistent with the

results from WP and Bayesian analyses (Figure 5.2) in this study. That placementwas

strongly supported in the previous study (Chapter 4) by both bootstrap proportion and

posterior probability (BP = 100, PP = 100%). Therefore, the C. sphecocephala dade is

best classified as a member of C. unilateralis dade, a treatment supported by the

morphological characteristics discussed below.

In comparing the relationships of dade B (Figure 5.3), our Bayesian results

indicate the C. sphecocephala dade is either a sister-group of the C. unilateralis dade

(108-taxon data set) or in the terminal group of the C. unilateralis dade (153-taxon

data set). In contrast, MP analyses places the C. sphecocephala dadeas a basal group of dade B in both analyses, suggesting that MP analyses aremore sensitive to the 154

MP analyses WP analyses Bayesian analyses

97 100 100 72 Hpocreaceae Hypocreaceae 100 100 Claeicipitaceae C Clavicipitaceae C ClaeicipitaceaeC 100 100 100 Claoicipitaceae A Claoicipitaceae A Claaicipitaceae A 100 100 I 08-taxon C. sphecocephala dade C sphecocephala dade C.gunnhi dade oo lilacinus dade 9 C. gunniidade P.ilacinus dade 100 100 C. gunnhi dade P. lilacinus dade C.ophioglossoide dade 100 100 cophioglossoide dade C. ophioglossoide dade cunilateralis dade 80 100 cH1.pocreaceaeC, unilateralls dade C. unhlateralis dade C: sphecocephala dade

99 100 100 Hypocreaceae Hypocreaceae 100 Hgocreaceae o° '° 100 ciaacipitaceaec Claacipitaceae C lI C 100 Cla6idipdaceae 1 48-taxon Claoidipitaceae A 100Clavicipitaceae A Cla6ldipitaceaeA 100 100 I i 100 C. gunnii dade 81 C. gunnhi dade C.gunnhi dade 74 87 P. lilacinus dade 0 p, lilacinus dade 9ot P.lilacinus 100 100 108 C. ophioglossoide dade C. ophioglossoide dade I _ e C.ophioglossoide dade 88 92 .100 C. unilateralis dade C. unilateralis dade cunhlaferalis dade

99 100 100 Hypocreaceae Hipocreaceae lao. Hypocreaceae rI 100 100 100 Cla6ldipitaceae C Cla6lcipitaceae C ClaeaoipitaceaeC 100 100 I 53-taxon Cladcipitaceae A Cla'acipitaceae A Cladcipdaceae A 72 100 70 100 C. spheCocephala dade C. gunnii dade C.gunnhi dade 72 C. gunnii dade plilacinus dade P.lilacinus dade 100 P. lilacinus dade C. ophioglossoide dade C.ophioglossoide dade L_.L;::::: c.ophioglossoide C. unilateral is dade cunilaferalis dade 100 100 C. unilateralis dade sphecocophala dade C.sphecocephala dade C. unhlaferalis dade C.unilaferalis dade

100 100 100 Hypocreaceae Hypocreaceae iao. Hpocreaceae 100 - 100 Cla6lcipitaceae C CIa'Acipitaceae C CladcipitaceaeC °° 100 1 62-taxon CIaacpitaceae A Claoicipitaceae A Cladcipitaceae A 100 72 100C. sphecocephala dade C. gunnii dade oat cgunnhi dade 98 84 100 cgunnii dade 96 P lilacinus dade P.liladinus dade 70 82 P. lilacinus dade C. ophioglossoide dade C.ophioglossoide dade 74 C. ophioglossoide dade C. unilateralis dade C.unhlaferalis dade 100 C. unhlateralis dade csphecocephala dade Csphecocephala dade C. unit ate ralis dade C.unilaferalis dade

Figure 5.3. Schematic diagrams of the phylogenetic hypotheses from a series of analyses. Maximum parsimony (MP), weighted parsimony (WP) and Bayesian analyses were performed with the data sets that differ in taxon sampling based on the degree of missing data. The 108- taxon data set are complete for five genes (i.e. nrSSU, nrLSU, EF-la, RPBJ and RPB2) and the 153-taxon data set are complete for at least three genes. To address the impact of the C. sphecocephala dade to the nodal support of C. unilateralis group, the 148-taxon data setwas constructed after the members of C. sphecocephala dade were excluded. The bootstrap proportions (BP? 70%) or posterior probabilities (PP? 95%) are shown above the corresponding nodes and in a thicker line. longer branches that characterize the C. sphecocephala dade. In light of long-branch attraction problems associated with the MP analyses (Figure 5.1), we used the

Bayesian tree (Figure 5.2) to further discuss the relationships in dade B andwe conclude the C. unilateralis dade includes the members of the C. sphecocephala dade

(Figures 5.2 and 5.5). In interpreting the C. unilateralis dade in terms of the statistical 155

support, we used the bootstrap proportions and posterior probabilities (MP-BP = 88%,

WP-BP = 92%, PP = 100) based on the results of 148-taxon data set (Figure 5.3).

Discussion

Phylogenetic implications on the system aticsofthe genus Cordyceps

Previous phylogenetic analyses (Chapter 4) have confidently revealed that species in the Clavicipitaceae form three strongly supported monophyletic groups based on combined data sets of six or seven genes (the five gene analyzed herein plus

Ji-tubulinand mtATP6). Although more taxa and fewer characters are used in this study, our results are consistent with the previous studies, recognizing three monophyletic groups designated as dade A, B and C (Figures 5.1 and 5.2). In addition, our results also support the paraphyly of Clavicipitaceaeas defined by the monophyly of Clavicipitaceae dade C and Hypocreaceae (Figures 5.1 and 5.2).

Although the paraphyly of Clavicipitaceae (dade C + Hypocreaceae) is moderately supported (MP-BP = 63%, WP-BP = 68%, PP = 100%) in the 162-taxon analyses, it was supported more robustly in the previous analyses which investigated localized conflicts among gene partitions and compared bootstrap proportionsamong alternative sampling strategies (Chapter 4).

The phylogenetic hypothesis presented here contradicts the current subfamilial classification of the Clavicipitaceae. Diehl (1950) proposed three subfamilies, 156

Oomycetoideae, Clavicipitoideae and Cordycipitoideae, based on the development of

ascostromata, anamorphic characters and host affiliations. However, these three

subfamilies do not represent the three clades of the Clavicipitaceae inferred in these

analyses (Chapter 4; Figures 5.1 and 5.2). Clade A includes members of all three

subfamilies (e.g., Claviceps of the Clavicipitoideae, Cordyceps of the

Cordycipitoideae, and Hypocrella of the Oomycetoideae), whereas the remaining clades only comprise members of Cordycipitoideae (e.g., Cordyceps and Torrubiella).

Importantly, all three maj or clades include the members of Cordyceps, indicating that neither Cordyceps nor the Clavicipitaceae are monophyletic (Figures 5.1 and 5.2). As a result, the three recognized well-supported clades (dade A, B and C) of the clavicipitaceous fungi represent a robust phylogenetic framework for the taxonomic revision of Cordyceps and the Clavicipitaceae.

In the current subgeneric classification of the genus, Cordyceps comprises four subgenera (C. subg. Bolacordyceps, C. subg. Cordyceps, C. subg. Ophiocordyceps and

C. subg. Neocordyceps) based on ascospore morphology and orientation of perithecia in stromata (Eriksson and Hawksworth, 1986; Kobayasi, 1941; Kobayasi, 1982).

However, most of these characters are not consistent with the current phylogenetic hypothesis and are not diagnostic of monophyletic taxa (e.g., subgenera and genera)

(Figure 5.1 and 5.2). For example, Kobayasi (1941; 1982) emphasizedascospore morphology and used the disarticulation of ascospores to delimit C. subg.

Ophiocordyceps from other subgenera. Species with intact ascospores, however,are included in all three major clades (C. acicularis Ray., C. cardinalis Sung & Spatafora 157

and Cordyceps sp. EFCC 2131) (Figure 5.1 and 5.2), indicating that nondisarticulating

ascospores are not phylogenetically informative at this level (Figure 5.1 and 5.2).

Therefore, a reassessment of diagnostic characters, in the previous and current

classification of Cordyceps, is necessary for the three major clades to provide a basis

for taxonomic revisions of Cordyceps and Clavicipitaceae.

Cordyceps species in Clavicipitaceae dade A- Clade A is comprised of four well-supported subclades (Figure 5.4). All known species of Cordyceps in the dade are included in the C. taii dade. Species of Cordyceps in the dade possess partially or completely immersed perithecia on clavate to cylindrical fertile parts of stromata

(Liang et al., 1991; Zang et al., 1982; Zare et al., 2001). They produce ascospores that are either disarticulating or nondisarticulating and include species that possess ordinal and obliquely embedded perithecia. Two species, C. chiamydosporia Evans and C. liangshanensis Zang, Liu & Hu, form ordinal perithecia and possess disarticulating ascospores, consistent with C. subg. Cordyceps (Kobayasi, 1982; Zang et al., 1982;

Zare et al., 2001). The undescribed Cordyceps species (Cordyceps sp. EFCC 2131 and

Cordyceps sp. EFCC 2135) possess ordinal perithecia and nondisarticulating ascospores, characters consistent with C. subg Ophiocordyceps. Finally, C. taii Liang and Liu, a known teleomorphic species of the genus Metarhizium Sorokin, produces disarticulating ascospores and obliquely embedded perithecia in the stromata,a trait presumably unique to C. subg. Neocordyceps (Liang et al., 1991). Therefore, dade A 158 contains species that in the current classification could be classified in three of the subgenera of Cordyceps.

These results suggest that ascospore morphology and orientation of perithecia are not phylogenetically informative in recognizing either the C. taii dade, or higher clades of clavicipitaceous fungi. Rather, they are more useful at species level classification. For example, our phylogenetic analyses revealed that C. taii is very closely related to C. brittlebankisoides Liu et al., the teleomorph of M anisopliae

(Metschn.) Sorokin var. majus (Johnst.) Tulloch (Liu et al., 2001). Although these species are similar to each other in macromorphology (e.g., greenish clavate stromata), they differ in the orientation of perithecia. C. brittlebankisoides possesses the perithecia that are ordinally placed in the stromata, whereas C. taii is characterized by the obliquely embedded perithecia. These results suggest that orientation of perithecia in the stromata is useful in delimiting these closely related Cordyceps species in the C. taii dade (Figure 5.4).

Cordyceps species in Clavicipitaceae dade B- Species of Cordyceps in this dade possess disarticulating or nondisarticulating ascospores and produce superficial to completely immersed perithecia that are ordinally or obliquely placed in the stromata. As with the Cordyceps species of dade A, this dade also includes members of C. subg. Cordyceps (e.g., C. ophioglossoides (Fr.) Link and C. variabilis Petch), C. subg. Ophiocordyceps (e.g., C. acicularis and C. unilateralis (Tul.) Sacc.) and C. subg. Neocordyceps (e.g., C. nutans Pat. and C. sphecocephala (Klotzsch ex Berk.) 159

°° Shsmszuomyces paradoxus Shimzuomyces paradoxus Shimizuomyces dade badia (A) llypocrella schizostachyi Aschersonia placenta Hypocrella dade ;;;1__f Hypocrella sp. GJS 89104 Epic hloe lyphina Clavices paspali S fusitonnis Claviceps pulpuiea Claviceps pwpuzea epi phylum CIa viceps dade Verticillium epip yftlm 100 Mynogenospoia atramentosa p,lulaefoemis ns,a epichloe lansia henningSiana bulbillosa , Pochonia gonioldes parasiticum Pochonia rubescensa1um Paedilomyces marquarothi LP0c1Pochonia100 Paecilomyces carneus Paecilomyces carneus Roiiferophthora angustispora 0 '" Cordyceps chIamydospora Pochon,a)I C. subg. Cordyceps , Pochon,a chiamydosporia - Cordyceps liangsloanensiS C. taildade Cordyceps liangshanensis I C subg. Cordyceps 005 Cordycepssp.EFCC2131 I _ Cordyceps . EFCC 2135 C. subg. Ophiocordyceps Nomuraea riley! so Metarhizium album ' Cordyceps sp, NH.) 12118 (Metarhizium) . Metarhizium flavovinde var. minus , Metarhizium anisopliae var. majus Cordyceps tag! )Metarhizium)I C. subg. Neocordyceps Cordyceps sp OSC 110996

(B)

Figure 5.4. (A) Enlargement of the Bayesian consensus tree m Figure 5.2. Only the Clavicipitaceae dade A is shown to emphasize the relationships within the dade. The respective suhgenera of Cordyceps species are provided to the right of species. The known anamorphs of Cordyceps species are in parentheses. The tree description is in Figure 5.2. (B) Examples of the nprsiilative species in Clalricipitaceae dade A.!. Shiirnzuomyces paradoxuson the seed of the plant (Smilax sieboldii; Smilacaceae). 2. Claviceps purpurea on Spartina afternifioiu (Poaceae). 3. Ilyjxxrellaschizostachyion scale insects (Hemiptera). 4. Cordvceps sp. EFCC 2131 on lepidoptera pupa. 5.Cordyceps liangshanensison lepidopteran larva. 160

Berk. & M.A. Curtis). The majority of Cordyceps species in this dade produce pliant

to tough to fibrous stromata that are typically darkly pigmented and parasitize

subterranean or wood-inhabiting hosts, which are deeply buried in soil or imbedded in

decaying wood. Exceptions to this ecology do exist; for example, members of the C.

sphecocephala dade which parasitize adult insects.

Clade B consists of five subclades. All subclades include either species of

Cordyceps or anamorphs that have been linked to Cordyceps (e.g., C. cylindrica

Petch) (Figure 5.5). The well-resolved tree in the present study provides the basis to

charactere three of the five subclades of Clavicipitaceae dade B. Due to insufficient

taxon sampling, it is not possible to characterize the members of the Cordyceps

species in the C. gunnii and P. lilacinus clades. In light of this, we focused on the

remaining subclades that include a sufficient number of Cordyceps species.

The C. ophioglossoides dade consists of Cordyceps species, that primarily parasitize

truffles in the genus Elaphomyces (e.g., C. japonica Lloyd and C. capitata (Holmsk.)

Link) and the nymphs of cicada (e.g., C. inegoensis Kobayasi and C. paradoxa

Kobayasi) deeply buried in soil (Kobayasi, 1939; Kobayasi and Shimizu, 1960;

Kobayasi and Shimizu, 1963; Mains, 1957). Species in this dade produce partiallyor

completely immersed perithecia, in clavate to capitate fertile parts of stromata thatare darkly pigmented with olivaceous tints (Kobayasi and Shimizu, 1960; Kobayasi and

Shimizu, 1963). Because they produce disarticulating ascospores and ordinal penthecia, all known species of this dade are classified in C. subg. Cordyceps. CO3dyo.p. 0, 100,0, C. gunnhi dade H.ploc/ilium ' F C. subg. Co,dyeps Nomu Tn. .lypicol. P..duomyon O.cinu. °°° P. lilacinus dade P..000myen oinn.mo moos

I.!ongis.gn,enUs CO,dy0000 trOd. j.ponlo. C. ophioglossoides dade C. subg. Cord'ceps ophiogl0550i055 yc.0100bn.sIHl rolypool.diom( Tn7.bIO5 (Syogdocf.du,m( w&dSb?!S (Syn gil OCMdIUm( 21meIoI000,00 h.1.mp009 hel,rop000 YCSPS gmcllls (P.mls.,i.( loop. gr.ollls (P.rOln,i.( C. ravenellli dade 071721g. bonbon. C. suby. Coniyceps ConS ye.ps 000no.n. - Cordyc.psSup.rficl.11s C. unilateralis dade - Cordyd.p.nlgr00 - Cord yc.ps,10005IlI

C. subg. Cordyceps No- Codtpnps ).koshom.n.sn Cordyo.ps 210.021 Cor0700ps sp000000ph.l. (Hym0000SIIbe) ordyc.pssp500005700l. (Hym.n050Ib.) 0,b TO C_sphecocephala dade C006yrep..long.t. (0710100.1 C. subq. CordyCeps C. subg. Neocordyceps

C000y0000 op. NHJ 12522 (0(700 1.11.) Hirsutella Conbyng. SF, 601 12529(Hlr.0t.11.( C. subg. Opbiocordyceps I(Hlrsol&I.( C. subg. Cordyceps La1 C. subg. Ophiocorriyceps C. unhlateralis dade 07101.IlS op. FlU12525 fllt301eII. op.SF. 601 12627 Conbyo.poolo.n.l. (00.0100.) I C. subg. OphiocordyCeps P000IIomyl:. 0(2171. I C. subg. Cordyceps 005 C. subg. Oph,oCordyceps ::: 7211i( C. Isubg. (B)I

Figure 5.5. (A) Enlargement of the Bayesian consensus tree in Figure 52. Only the Clavicipitaceae dade B is shown to emphasize the relationships within the dade. The respective subgenus of Cordyceps species are provided to the right or below of species. The known anamorphs ofCordycepsspecies are in parentheses. Numbers that above the corresponding nodes are bootstrap proportions of WP analyses (before the backslash) and posterior probabilities (after the backslash) from I 48-taxon data set. Values that below the corresponding nodes are bootstrap proportions of WP analyses (before the backslash) and posterior probabilities (after the backslash) from 162-taxon data set in Figure 5.2. The bootstrap proportions of? 70% or posterior probabilities of? 95% are shown in the 162 corresponding nodes. The nodes in a thicker line are supported by the bootstrap proportions and posterior probabilities from both 148-taxon and 162-taxon data sets. Numbers in a circle correspond to the node that is informative for placing the C. sphecocephala dade. (B) Examples of the representative species in Clavicipitaceae dade B. 1. Cordyceps ophioglossoides on truffle (Elaphomyces sp.; Euascomycetes). 2. Cordyceps subsessilis on coleopteran larva in decaying wood. 3. Cordyceps ravenelii on coleopteran larva (Phyllophaga sp.). 4. Cordyceps melolonthae on coleopteran larva. 5. Cordyceps nutans on stink bug (Hemiptera). 6. Cordyceps sphecocephala on wasp (Hymenoptera). 7. Cordyceps longissima on nymph of cicada (Hemiptera). 8. Cordyceps unilateralis on (Hymenoptera). 9. Cordyceps sinensis on lepidopteran larva.

C. subsessilis Petch is a unique species within the C. ophioglossoides dade

(Figure 5.5). It produces perithecia on white or pallid stromata, arising from a rhizomorph-like structure from scarabaeid beetle larvae (Hodge et al., 1996). It is the only member of the dade that parasitizes beetles imbedded in decaying wood (Hodge et al., 1996). Therefore, C. subsessilis differs greatly in ecology and in the morphology of its stromata from most other taxa in the C. ophioglossoides dade. C. subsessilis, however, possesses several characters shared by its close relative, C. ophioglossoides

(Hodge et al., 1996; Kobayasi and Shimizu, 1960) (Figure 5.5). Both species are the only culturable species in the dade, produce verticilliate anamorphs, possess nearly identical partspore morphologies, and produce stromata that are connected to their hosts via rhizomorph-like structure.

The C. ophioglossoides dade (Figure 5.5) also includes parasites of cicada

(e.g., C. paradoxa and C. inegoensis), which are grouped with their close relatives

(e.g., C. ophioglossoides and C. jezoensis S. Imai) that parasitize subterranean truffles

(e.g., Elaphomyces). Despite low support of interspecies relationships within the dade,

C. paradoxa and C. inegoensis are morphologically more similar to C. ophioglossoides and C. jezoensis than any other members of the dade. These taxa 163

produce clavate fertile parts of the stromata that are not present in other members of

the dade (e.g., C. capitata and C. fracta Mains). Many of these species (e.g. C.

jezoensis and C. paradoxa) are also known to be connected to their hosts via

rhizomorph-like structures (Hodge et al., 1996; Kobayasi and Shimizu, 1960;

Kobayasi and Shimizu, 1963), supporting a close phylogenetic relationship.

The C. unilateralis dade includes the most morphologically diverse

assemblages of species in the genus Cordyceps (Figure 5.5). Most of the species in the

dade parasitize larval, pupal or nymph stages of arthropods (Kobayasi, 1941; Mains,

1958). Species of this dade produce superficial to completely immersed peritheciaon

the stromata with morphologies ranging from capitate, clavate to fihiform (Kobayasi,

1941; Mains, 1958). They typically possess tough or pliant stromata thatare entirely

or partially darkly pigmented, although some exceptions (e.g., C. melolonthae (Tul. &

C. Tul.) Sacc., and C. variabilis) do exist which produce fleshy and brightly

pigmented stromata (Mains, 1958). Many species in the dade (e.g., C.

brunneapunctata Hywel-Jones, C. stylophora Berk. & Broome and C. unilateralis)are

also differentiated by sterile apices of the stromata and produce perithecia in

subterminal regions.

Similar to Cordyceps species in dade A, species of Cordyceps in the C. unilateralis dade include species producing disarticulatingor nondisarticulating ascospores. For example, some species in the C. unilateralis dade (e.g., C. sinensis

(Berk.) Sacc. and C. unilateralis) are classified as members of C. subg.

Ophiocordyceps. These species are interspersed among other species (e.g., C.agriota 164

A. Kawam. and C. stylophora) that are classified in C. subg. Cordyceps. This indicates

that while ascospore morphology is useful in delimiting closely related Cordyceps

species within the C. unilateralis dade it is not diagnostic of the dade itself (Figure

5.5).

All the members of C. subg. Neocordyceps, as classically treated by Kobayasi

and Mains, form a monophyletic group labeled as the C. sphecocephala dade within

the C. unilateralis group (Figure 5.5). The majority of species in the C. sphecocephala

dade produce long, wiry stipes with darkly or brightly colored stromata, which

terminate in clavate fertile parts and possess disarticulating ascospores (Hywel-Jones,

2002; Kobayasi, 1941; Kobayasi, 1982). Species in this dade produce perithecia,

which are partially or completely immersed in stromata at strongly oblique angles

(Kobayasi, 1941; Kobayasi, 1982; Mains, 1958). This dade is one of the best

characterized by its morphology (obliquely embedded perithecia in a well-defined

clava) and its ecology of parasitizing the adult stages of insects.

Cordyceps species in Clavicipitaceae dade C- Clade C includes C. militaris,

the type species of the genus Cordyceps (Figure 5.6). Most Cordyceps species in this

dade are currently classified in C. subg. Cordyceps (Kobayasi, 1941; Kobayasi,

1982). This dade also contains the members of C. subg. Ophiocordyceps and C. subg.

Bolacordyceps, resulting in members of C. subg. Cordyceps not forminga monophyletic group in the dade C (Eriksson, 1982; Hywel-Jones, 1994; Sung and

Spatafora, 2004). Species of Cordyceps in this dade produce threeascospore types 165

including disarticulating ascospores (e.g., C. militaris), nondisarticulating ascospores

(e.g., C. cardinalis and C. pseudomilitaris Hywel-Jones) and bola-ascospores (e.g., C.

bfusispora O.E. Erikss.). Of particular note, this dade includes Phytocordyceps

ninchukispora Su and Wang, the monotypic member of the genus Phytocordyceps Su

and Wang. The genus Phytocordyceps was originally described based on bola-

ascospores and its host affiliation as a pathogen of Beilschmiedia erythrophloia

Hayata (Lauraceae) plant seeds (Su and Wang, 1986). Morphologically, this species is

most similar to C. bfusispora in that it produces the typical bola-ascospores of C.

subg. Bolacordyceps. However, the phylogenetic analyses in this study reveal that

species producing bola-ascospores (e.g., C. bfusispora and P. ninchukispora) do not

form a monophyletic group (Figure 5.6). Rather, they are interspersed among other

Cordyceps species possessing disarticulating ascospores, most notably C. militaris.

Species of Cordyceps in dade C produce superficial to partially immersed

perithecia on fleshy stromata that are pallid to brightly pigmented. This is in contrast

to Cordyceps species in dade B, which produce darkly pigmented, tough to pliant

stromata. This suggests pigmentation and texture of stromata may be phylogenetically informative at a higher level of classification. It should be noted, however, that some species in dade C are similar to distantly related Cordyceps species in stromatal pigmentation. Although these charactersare useful in recognizing

Cordyceps species in dade C, the utility of these characters for general species level of classification is limited (Figure 5.6). For example, C. militaris is macroscopically similar to C. cardinalis and C. pseudomilitaris. All three species produce orangish 166

00 SimpiciIIio..m In,eflftol , Sin0,JidIIi..no obdo01tu,,1 100 000 SimpIidlI,um 'A' 000 01 Sa,woiklslio,m !noonbo,00u001 00 To,b.lI 00aIIac.l(Sino00iI0on4 To,,obolIa rtficdf a LecankiIIiom antillaaom Eflgyoxkzntium araneaium Lecaa,ciflium annearum '°Cordyneps rardonaiis (Manannaea a, CIonoSaloy Co,dyceps ca,th,,ahs M,oannaoa a, CIonosnd1y . SUOaSJ. p ?OCOyceps 00 Lecanicillium psalliotae Lo.canicilium fuaispooum Lecanociilium paalbotae 100 Lecan,cahum peailiotae 100 sasra fannosa 000 Isaira farinosa CaoUyceps tub ,au!ata Aha,offiomyoes or Veth1iIliur4 " Tomthoelia confragoa(Leeaoo1rouium 00 100 LecalpicIIilomo atteouatu,a 000 000 Cot dyceps scarabaeicola Beauveha) ne Co.dyreps siapo1yIrniocoIa (Beauneria) 01 Coro'aepa Cf. takaoofana C. subg.Cord yceps lSata Cf. tar,nosa 00 Cor&yceps ochraceostromata Isa,,a tennpes 100 Cordyaeps op. EFCC 2535 ' Co,o.cepo takao,,bna Penellomyres)

001 C. subg.Bolacordyceps : I 00 100Mic,06ik,,t, CO*yoo0S ooCoperae 100 C. subg.Cordyceps VerfirJIiurnsp CBS 1021U 00 100 Co,otyceps op. EFCC 5197 100 000 COrdyCepS op. EFCC 5693 000 "Co,dyaopoop.NHJ 10627 C. subg.Bolacordyceps 0.05 Phytoanchuk00poraConepo op. NHJ 106U 000 (B) 0 0

Figure 5.6. (A) Enlargement of the Bayesian consensus tree in Figure 5.2. Only the Clavicipitaceae dade C is shown to emphasize the relationships within the dade. The respective subgenus of Cordyceps species is provided to the right of species. The known anamorphs of Cordyceps species are in the parentheses. The tree description is in Figure 5.2. (B) Examples of the representative species in Clavicipitaceae dade C. 1. Cordyceps militaris on lepidopteran pupa. 2. Cordyceps tubercu/ata on moth. 3. Cordyceps sp. EFCC 5197 on lepidopteran pupa (Limacodidae). 4. Cordyceps scarabaeicola on scarabaeid beetle. 5. Torrubiella sp. on spider. 6. Cordyceps cardinalis on lepidopteran larva (Tineidae). 167 colored and fleshy stromata, however, these species differ in ascospore and anamorph morphology (Sung and Spatafora, 2004). Furthermore, C. militaris is known as exhibiting great variability in stromatal morphology (Hywel-Jones, 1994; Sung and

Spatafora, 2004). Its putative conspecific species, such as C. roseostromata Kobayasi

& Shimizu and C. kyushuensis Kobayasi differ in stromatal morphology, but are closely related to C. militaris and possess identical ascospore and ascus morphologies

(Figure 5.6; Hywel-Jones, 1994; Sung and Spatafora, 2004).

The plasticity of stromatal morphology in dade C often results in misidentification of species (Liang, 1991; Sung and Spatafora, 2004). For example, this study reveals a close relationship between the anamorphic species, Mariannaea pruinosa Z.Q. Liang, and Phytocordyceps ninchukispora (Figure 5.6). The teleomorph of M pruinosa is C. pruinosa Petch, which produces disarticulating ascospores and reddish orange stromata and parasitizes lepidopteran cocoons (Limacodidae)

(Kobayasi, 1941; Liang, 1991; Petch, 1924). Although the isolate of M pruinosa, used in Liang's study, was obtained from ascospores (Liang, 1991), the morphology of the ascospores was not well characterized. As a result, the species was identified primarily based on its host affiliation and macroscopic characters. In this study, Cordycepssp.

EFCC 5197 and NHJ 10627 were collected from the same host family in Korea and

Thailand. They are also closely related and produce reddish orangish stromata (Figure

5.6). However, these species possess bola-ascospores, suggesting the identity of C. pruinosa specimens used in Liang's study (1991) may be questionable. Furthermore, these Cordyceps species are closely related to and morphologicallyvery similar to P. ninchukispora with the exception of host affiliation, suggesting the possibility of host misidentification in the original description of P. ninchukispora.

The Clade C includes not only the members of Cordyceps but also members of the genus Torrubiella that generally parasitize spiders and scale insects (Kobayasi and

Shimizu, 1982). The genus Torrubiella is morphologically characterized by superficial perithecia on the mycelial subiculum, surrounding the host (Humber and Rombach,

1987; Kobayasi and Shimizu, 1982). Species of Torrubiella also produce disarticulating (e.g., T wallacei H.C. Evans) and nondisarticulating (e.g., T ratticaudata Humber and Rombach) ascospores. Among species of Cordyceps, C. tuberculata (Lebert) Maire, a pathogen of Lepidoptera, has been considered an intermediate species between Torrubiella and Cordyceps (Humber and Rombach,

1987; Mains, 1958; Zare et al., 2001). Phylogenetic analyses in this study indicate the members of Torrubiella do not form a monophyletic group within dade C and are interspersed among species of Cordyceps. This suggests that the stipitate stromata of

Cordyceps have been gained or lost several times during the evolution of these fungi.

In addition to dade C, dade B also includes an anamorphic species, Paecilomyces cinnamomeus (Petch) Samson and W. Gams, which is linked to the teleomorph T luteorostrata Zimm. (Hywel-Jones, 1993). Currently, more than 50 species of

Torrubiella have been described and the members of genus Torrubiella are clearly undersampled in this study (Kobayasi and Shimizu, 1982). To achieve a better understanding of the relationships between Torrubiella and Cordyceps additional sampling of Torrubiella is necessary. 169

In summary, the dominant characters (e.g., ascospore morphology and the orientation of perithecia), used in the current classification of genus Cordyceps, are not congruent with the three higher clades inferred in these analyses. These characters are useful, however, in lower level classifications, such as the delimitation of closely related species. The characters most congruent with the three higher clades of clavicipitaceous fungi are texture, pigmentation and morphology of the stromata.

Although we have divided Cordyceps species into three major clades, it is difficult to characterize Cordyceps species within dade A due to the insufficient taxon sampling.

They tend to produce green to white stromata, often with lilac tints, but additional sampling is needed to more definitively characterize the teleomorphs of these species.

However, Cordyceps species within clades B and C are morphologically or ecologically distinct (Figure 5.1 and 5.2).

The majority of Cordyceps species in dade B are characterized by darkly pigmented, tough or pliant stromata. The dominant form of parasitism exhibited by these species is on subterranean or wood-inhabiting hosts, deeply buried in soil or imbedded in decaying wood, such as larval and pupal stages of arthropods. In contrast,

Cordyceps species in the dade C have brightly pigmented and fleshy stromata and parasitize their hosts in relatively exposed environments, such as leaf litter, in moss, or in the upper most soil layer. Exceptions to these morphological and ecological traits are found in some Cordyceps species in dade B (e.g., C. melolonthae, C. sphecocephala, C. subsessilis and C. variabilis). Cordyceps melolonthae produces, for example, brightly colored stromata, although it bruises dark upon handling and its 170 hosts are the larva of June beetles, deeply buried in soil (Mains, 1958). This suggests that the traits described above are not individually informative, but collectively useful in characterizing Cordyceps species within dade B.

The taxonomic utility ofanainorphicforins iii classification of Cordyceps.

The genus Cordyceps is characterized by a diverse assemblage of more than 25 anamorphic forms (e.g., genera Beauveria, Metarhizium, Hirsutella and Hymenostilbe)

(Gams and Zare, 2003; Kobayasi, 1982; Samson et al., 1988). The anamorphic genera of Cordyceps are hyphomycetes with conidiogenous cells that are white to brightly colored and produce conidia in dry chains or slimy drops (Samson et al., 1988). Some anamorphic genera (e.g., Hymenostilbe) are known as a useful diagnostic character in recognizing monophyletic groups of Cordyceps species (Artjariyasripong et al., 2001;

Kobayasi, 1941; Kobayasi, 1982). Therefore, the distribution of anamorphic forms is discussed to evaluate their phylogenetic utility in characterizing the three clades of

Cordyceps and to better understand teleornorph-anamorph connections.

Clavicipitaceae dade A- Clade A not only comprises of teleomorphic genera

(e.g., Cordyceps, Claviceps and Balansia) but it also includes members of several anamorphic genera (Figure 5.4). This study shows that some anamorphic genera (e.g.,

Nomuraea Maublanc, Paecilomyces sect. Isarioidea Samson and Tolypocladium W.

Gams) are broadly distributed and found in at least two major clades in the 171

Clavicipitaceae. For example, the members of Paecilomyces sect. Isarioidea are found in all three major clades and the taxonomic utility of this genus is limited (Figure 5.1

and 5.2). In contrast, other anamorphic genera (e.g., Aschersonia Mont. and

Metarhizium) have more restricted distribution and are phylogenetically informative in recognizing monophyletic groups within dade A.

The C. taii dade contains numerous anamorphic genera including Nomuraea,

Pochonia Batista & Fonseca, Tolypocladium, and Metarhizium (Figure 5.4). The genera Nomuraea, Pochonia and Tolypocladium are not monophyletic and are represented in two of the three clavicipitaceous clades. Only Metarhizium is monophyletic and thus phylogenetically informative (Figure 5.4). The conidiogenous cells in the genus Metarhizium are cylindrical to clavate without a neck and produced in candelabrum-like fashion (Driver et al., 2000; Evans, 2003; Rombach et al., 1986).

The genus is most similar to the genus Nomuraea and simply differs in the compact conidiophores that form a hymenial layer (Evans, 2003). In phylogenetic relationships,

N. rileyi (Fan.) Samson groups with species of Metarhizium; however, N atypicola

(Yasuda) Samson belongs to the P. lilacinus dade in dade B. Interestingly, N rileyi produces greenish colored conidia as do species of Metarhizium in the C. taii dade and N atypicola possesses lavender colored conidia similar to those of P. lilacinus

(Coyle et al., 1990; Evans, 2003; Hywel-Jones and Sivichai, 1995). Currently, two teleomorphic species of Metarhizium have been reported (Liang et al., 1991; Liu et al.,

2001). The species M taii Liang and Liu was described with its teleomorph species,

C. taii, and is morphologically very similar to M anisopliae var. majus, which is 172 linked to the teleomorph C. brittlebankisoides (Liang et al., 1991). Although M tail and M anisopliae var. majus are very closely related (Figure 5.4), Metarhizium

species show extensive variation in the size and color of conidia (Driver et al., 2000;

Evans, 2003) and more intensive sampling of anamorphs and teleomorphs is needed for this group.

The genus Tolypocladium is characterized by producing single or whorled

(verticillate) conidiogenous cells (phialides), which are flask-shaped with enlarged bases that taper into a needle-like neck usually bent from the axis of phialides (Bissett,

1983; Gams, 1971). The type of the genus Tolypocladium is T inflatum W. Gams, which is linked to the teleomorph C. subsessilis (Gams and Zare, 2003; Hodge et al.,

1996). T inflatum is placed in dade B and is distantly related to T parasiticum G.L.

Barron in the C. taii dade. Morphologically. T parasiticum differs from other species of Tolypocladium, as it is the only member of the genus that produces chiamydospores in culture (Bissett, 1983; Gams and Zare, 2003; Zare et al., 2001). In a recent treatment of Verticillium sect. Prostrata W. Gams, the genus Pochonia was also reclassified based on production of dictyochlamydospores or at least swollen hyphal cells (Gams and Zare, 2001; Zare et al., 2001), supporting the close phylogenetic relationships ofT parasiticum and Pochonia species in this study (Figure 5.4). In addition, (Massee) Hughes also possesses chlamydospores in culture as well as the anamorphic state of undescribed Cordyceps species (i.e.

Cordyceps sp. EFCC 2131 and Cordyceps sp. EFCC 2135). As suggested by Barron 173

and Onions (Barron and Onions, 1966), the presence of chiamydospores is an

informative character in anamorph taxonomy.

The present study shows that the genus Ashersonia is a monophyletic lineage labeled as H. schizostachyi dade (Figure 5.4). The genus Ashersonia is best characterized by its pycnidial or acervular conidiomata and its ecology of parasitizing scale insects and (Hywel-Jones and Evans, 1993; Hywel-Jones and

Samuels, 1998; Petch, 1921). The teleomorph of Ashersonia has long been linked to the species of Hypocrella and more than 25 species have been reported (Mains, 1959;

Petch, 1921). This study also corroborates that the unique morphology of Ashersonia is phylogenetically informative and diagnostic of a monophyletic group of clavicipitaceous fungi (Figure 5.4).

Ctavicipitaceae dade B- The dade includes several anamorphic genera including genera Haptocillium W. Gams & Zare, Hirsutella, Hymenostilbe and

Tolypocladium (Figure 5.5). The anamorphic forms in the dade are phylogenetically informative due to the dominant occurrence of Hirsut-ella and Hymenostilbe anamorphs in C. unilateralis dade. Especially, the taxonomic utility of Hirsutella anamorph needs to be evaluated given the morphologically diverse nature of

Cordyceps species in the C. unilateralis dade.

The genus Hirsutella is characterized by its typical basally-subulate phialides, narrowing into one or more very slender needle-like necks,on synnemata or mononematous mycelium (Gams and Zare, 2003; Hodge, 1998). Most Hirsutella 174

species normally produce conidia in mucus and the phialides are not usually bent in

their needle-like necks such as the genus Tolypocladium. All Cordyceps species in the

C. unilateralis dade are not connected to Hirsute/la anamorphs (e.g., Paecilomyces,

Syngliocladium Petch and Paraisaria Samson & Brady) and the anamorphic forms are

not known for many of the Cordyceps species, especially in the C. ravenelii dade

(e.g., C. heteropoda Kobayasi and C. superfIcialis (Peck) Sacc.). However, most of

the Cordyceps species in the rest of the C. unilateralis dade have been linked to

Hirsutella anamorphs (Figure 5.5). These results suggest that Hirsutella anamorph is

phylogenetically informative for at least one of the subclades of the C. unilateralis

dade or possibly symplesiomorphic for the C. unilateralis dade as a whole.

The taxonomic utility of Hirsute/la anamorph is exemplified by the

teleomorph-anamorph connection of the genus Cordycepioideus Stifler, a termite

pathogen which does not have typical ascospore and ascus morphology of

clavicipitaceous fungi (Blackwell and Gilbertson, 1984). The genus Cordycepioideus

possesses thick-walled multiseptate ellipsoid ascospores and its asci lack the thickened

ascus tip characteristic of most clavicipitaceous fungi (Blackwell and Gilbertson,

1984; Ochiel et al., 1997). The anamorph of Cordycepioideus bisporus Stifler is a

synnematous Hirsute//a anamorph that is either conspecific with or closely related to

H. thompsonni F.E. Fisher (Ochiel et al., 1997; Suh et al., 1998; Sung et al., 2001).

Although Cordycepioideus bisporus greatly differs from other members of C. unilateralis dade in its teleomorphic characters, molecular data strongly support it as a member of C. unilateralis dade, a finding consistent with its Hirsutella anamorph. It 175

should be noted that species of Cordyceps outside of dade B have been described with atypical Hirsutella anamorphs (e.g., C. pseudomilitaris), but upon further investigation were more accurately characterized in other anamorph genera (e.g. Simplicillium

W.Gams & Zare).

The C. unilateralis dade includes members of the C. sphecocephala dade that possess a Hymenostilbe anamorph. The genus Hymenostilbe usually produces cylindrical to clavate conidiogenous cells, which are produced in more or less conspicuously dense layer in synnemata (Samson et al., 1988). It is differentiated from its closely related genera (e.g., genera Akanthomyces W. Gams and Hirsutella) by its polyblastic conidiogenous cells, which holoblastically produce a single conidium on a short denticle or scar (Hywel-Jones, 1996; Samson et al., 1988). The results from the present study (Figure 5.5) indicate that Hymenostilbe anamorphs are derived from within Hirsutella, which produces conidiogenous cells on a discontinuous layer in its synnemata (Figure 5.5). The phylogenetic relationship between Hirsutella and

Hymenostilbe anamorphs has precedent as demonstrated with the morphologically intermediate synnematous Hirsutella species. For example, Hirsutella lecaniicola

(Jaap) Petch, the anamorph of C. clavulata (Schwein.) Ellis & Everh., was previously classified in Hymenostilbe based on its extensively polyphialidic conidiogenous cells although it possesses conidia in mucus and its conidia are enteroblastically produced in a discontinuous layer (Hodge, 1998; Mains, 1950; Mains, 1958; Samson and Evans,

1975). In contrast, some Hirsute/la species (e.g., H. rubripunctata Samson, Evans &

Hoekstra) produce only a single conidium without its mucous sheath in extensively 176 polyphialidic conidiogenous cells. Therefore, the morphological boundary between

Hirsutella and Hymenostilbe modes of asexual reproduction may overlap to some extent and additional work is necessary to address the relationships between Hirsutella and Hymenostilbe (Gams and Zare, 2003).

In addition to the C. unilateralis dade, the remaining three subclades contain

Hapticillium, Tolypocladium and Vertidilhium anamorphs. The genus Haptocillium was reclassified from Verticihhium sect. Prostrata primarily based on its adhesive conidia and its ability to parasitize free-living nematodes (Zare and Gams, 2001b).

This study shows that the genus is a monophyletic group in C. gunnii dade (Figure

5.5). However, the teleomorph-anamorph connection was not established for any of the species in the dade or its close relative, (. gunnii, and thus its taxonomic utility remains unclear. The C. ophioglossoides and P. hilacinus clades include the anamorphic forms of Nomuraea, Tolypocladium and Verticillium, all of which are polyphyletic as previously discussed (Figure 5.1 and 5.2; (Luangsa-Ard et al., 2005;

Luangsa-ard et al., 2004; Obornik et al., 2001). Several teleomorph-anamorph connections have been reported for Cordyceps species in the C. ophioglossoides and

P. lilacinus clades although their taxonomic utility is limited. C. subsessilis is known to be a teleomorph ofT inflatum (Hodge et al., 1996) and C. ophioglossoides produces a Verticihlium anamorph (Gams, 1971). In the P. lilacinus dade, P. cinnamomeous is known as an anamorph of Torrubiella luteorostrata (Hywel-Jones,

1993) and the anamorph of C. cylindrica is N. atypicola (Hywel-Jones and Sivichai,

1995). 177

Clavicipitaceae dade C- The major anamorphic genera sampled here that are members of dade C include Beauveria, Isaria Pers, Lecanicillium W. Gams & Zare

and Simplicillium. Species of Lecanicillium and Simplicillium were previously placed in Verticillium sect. Prostrata and recently reclassified based on the phylogenetic studies of Sung et a! (2001), Gams and Zare (2001) and Zare and Gams (2001a). The genus Lecanicillium is characterized by producing slender aculeate phialides that are produced singly or in whorls and usually arise from prostrate aerial hyphae (Zare and

Gams, 2001a). Conidia are mostly produced in the tip of phialides and attached in heads or fascicles (Zare and Gams, 2001a). The morphological delimitation of

Simplicillium from Lecanicillium is difficult although the species of Simplicillium tend to produce phialides that more or less arise singly from prostrate aerial mycelium

(Zare and Gams, 2001a). This study shows that the species of Lecanicillium do not form a monophyletic group, as species of other anamorphic genera (e.g., Beauveria,

Engyodontium G. S. de Hoog and Isaria) are interspersed among species of

Lecanicillium (Figure 5.6). The monophyly of Simplicillium is also not supported because Simplicillium wallacei H.C. Evans, an anamorphic state of Torrubiella wallacei, and Simplicillium-like anamorph of C. pseudomilitarisare not monophyletic

(Hywel-Jones, 1994; Sung and Spatafora, 2004; Zare and Gams, 2001a).

Some Lecanicillium species are known to be anamorphic forms of Cordyceps and Torrubiella (Evans and Samson, 1982; Petch, 1932; Zare and Gams, 2001a). For example, C. militaris produces a Lecanicillium anamorph in the culture (Zare and 178

Gams, 2001a) and the anamorph of Torrubiella alba Petch isL.aranearum (Petch)

Zare & W. Gams (Petch, 1932). The type species of Lecanicillium isL.lecanii

(Zimm.) Zare & W. Gams, which is connected to the teleomorph T confragosa Mains, a pathogen of scale insects (Evans and Samson, 1982). In addition to Lecanicillium anamorphs, other genera (e.g., genera Akanthomyces, Gibellula Cavara, Hirsutella,

Paecilomyces and Simplicillium) have also been linked to Torrubiella (Kobayasi and

Shimizu, 1982; Samson et al., 1988; Samson et al., 1989; Zare and Gams, 2001a).

Clade C also includes the species of Isaria, which was recently lectotypified based on the phylogenetic placement of I farinosa (Holmsk.) Fr. for some of the clavicipitaceous Paecilomyces species (Gams et al., 2005; Luangsa-Ard et al., 2005).

The genus Paecilomyces is a diverse genus, with molecular studies indicating its polyphyletic status (Luangsa-Ard et al., 2005; Luangsa-ard et al., 2004; Obornik et al.,

2001). The type species, P. variotii Bainier, belongs to the order

(Ascomycota) and is distantly related to the Clavicipitaceous Paecilomyces species that were previously classified in Paecilomyces sect. Isarioidea (Luangsa-ard et al.,

2004; Samson, 1974). The previous taxonomy of Paecilomyces is primarily based on the monographic study of Samson (1974), which included approximately 30 species in

Paecilomyces sect. Isarioidea. In a recent molecular study, Luangsa-Ard et al. (2005) demonstrated that species in Paecilomyces sect. Isarioidea are subdivided into four monophyletic groups, three of which are statistically supported. As a result, eleven species of Paecilomyces sect. Isarioidea are reclassified to the genus Isaria (e.g., I. fumosoroseus Wize, I. javanicus (Frieder. & Bally) Samson & Hywel-Jones and I. 179

tenuipes Peck.) (Luangsa-Ard et al., 2005). The present study indicates that the species of Isaria do not form a monophyletic group in the dade C, as they are interspersed among other anamorphic forms in the dade. Thus, the taxonomic utility of Isaria anamorph is limited to the dade C, as seen with Lecanicillium and Simplicillium anamorphs. Furthermore, the teleomorph connections of Isaria species have not been well studied. The anamorph of C. takaomontana Yakush. & Kumaz. was reported as

Isaria tenuipes (Kobayasi, 1941); I. farinosa, is an anamorphic form of C. memorabilis (Ces.) Ces. (Pacioni and Frizzi, 1978), but was once mistakenly linked to

C. militaris (Figure 5.6; Petch, 1936).

The closely related species, C. scarabaeicola Kobayasi and C. staphylinidaecola Kobayasi & Shimizu produce Beauveria anamorphs (Figure 5.6;

Sung, 1996). In addition, a few additional species of Cordyceps have been linked to

Beauveria anamorphs. C. bassiana Li, Li, Huang & Fan and C. brongniartii Shimazu are known as teleomorphs of B. bassiana (Bals.) Vuill. and B. brongniartii (Sacc.)

Petch, respectively (Li et al., 2001; Shimazu et al., 1988). Thegenus Beauveria is morphologically well characterized by producing basally-inflated conidiogenous cells that sympodially produce conidia on a rachis structure (De Hoog, 1972; Macleod,

1954). Species of Beauveria possess a cosmopolitan distribution witha considerably broad host range (Mugnai, 1989; Rehner and Buckley, 2005; Evans, 2003). Recent molecular study (Rehner and Buckley, 2005) included 87 isolates of five Beauveria species (e.g., B. amorpha (HOhn.) Samson & H.C. Evans, B. bassiana, B. caledonica

Bissett & Widden and B. vermiconia de Hoog & V. Rao) and demonstrated that the genus is monophyletic and one of the more phylogenetically informative anamorphs of

dade C.

In fungal systematics, the naming of anamorphic forms is allowed for Phylum

Ascomycota and Basidiomycota by Article 59 of the International Code of Botanical

Nomenclature (Greuter et al., 2000) and multiple names exist for the same organisms of teleomorphic and anamorphic taxa. Recently, molecular phylogenetics has played an important role in integrating teleomorphic and anamorphic forms in a unified classification system in the clavicipitaceous fungi (Luangsa-Ard et al., 2005; Reynolds and Taylor, 1993; Sung et al., 2001). In such efforts, Verticillium sect. Prostrata and

Paecilomyces sect. Isarioidea were recently reclassified into several anamorphic genera (e.g., Lecanicillium, Sirnplicilliium, Isaria, Pochonia and Hapticillium) to be consistent with the current hypotheses of relationships (Luangsa-Ard et al., 2005; Zare and Gams, 2001a; Zare and Gams, 2001b; Zare et al., 2001). The phylogeny presented here also improves our understanding of the teleomorph-anamorph connections in

Cordyceps and implies that several anamorphic genera (e.g., Aschersonia, Beauveria,

Hirsutella, Hymenostilbe and Metarhizium) are more restricted in their phylogenetic distribution and are phylogenetically informative in characterizing Cordyceps species

(Figure 5.4, 5.5 and 5.6).

Taxonomic revisionsofCordyceps and the Clavicipitaceae 181

Y Hypocreaceae

Co,dyO.psC,th,,II r-LEILC,dyCpS (dftlI0,

Cordycipitaceae Clavicipitaceae dade C Op. EFCC 2535 komon0,,, b,los,spora 511lSlSpOl

Cordyceps s.s. IS 102164

!S. Co,oycsps op. NHJ 10627 7]iCordycops opNI4J 60684 Phy000ponloohoklspors Phy8000rld 1150115 nlnchulospors

balSa Hypocrellascfllzoslachyl

Hypocstilasp.0JS89104 __CO3olfom0o 70l70l05 CIavcipitaceae s.s Myg Clavicipitaceae dade A

Metacordyceps C.tail dade

-M l0,000, Ids 085.6,0,08 00 Melarhialuw atJisopIiae oar. majos -Co,oyoopa Sail Cordycepssp.OSC1IOSSG -CordyoOpogoo,,o H 00 Ophiocordycipitaceae -6, Clavicipitaceae dade B C0rS'SSC50,SaSa 0/96 - 74I8 100 100 Elaph ocordyceps 1, c ophioglosso,desdade c C012ycops50550ssdjs== CO,OyoOps m510lOOrSrao

CSrc/Cep5 9100,05 Co,dycepS 1120,/Is C0Idyleps ko,,,,000.o 88 y,psh II 100 CSldySspoe,gon,o6,1860 C. unilateral/s dade

Cd1 pyd Sm TTo'SYPSaPS.oOoal. _y,,, Ophiocordyceps C.sphecocepha!a dade C002yoeps 11009012 '(0 Cordyceps sp. 1/HJ12581 Cor2)?OOps Sp. NOd 0552 codyoopsop. 518212522 IS ColdyoOpl 511.518212529 Cordyoeps558,/S 5( Is

C. unilateralis dade

,I*nwnm..,Tw1ln 182

Figure 5.7. A new classification of Cordyceps and the clavicipitaceous fungi based on the consensus tree from Bayesian analyses in Figure 5.2. The portions of the Bionectriaceae and Nectriaceae are not shown to emphasize the nomenclatural changes in Cordyceps and the clavicipitaceous fungi. The tree description is the same as in Figure 5.2. For the nodes that are related with the nomenclatural changes, the bootstrap proportions of MP analyses (MP-BP) in Figure 5.1 are shown above the corresponding nodes before the backslash. Bootstrap proportions of WP analyses (WP-BP) and posterior probabilities in Figure 5.2 are shown above the internal nodes after the backslash and below the internal nodes, respectively. For the corresponding node of the genus Ophiocordyceps, the bootstrap proportions (MP-BP and WP- BP) and posterior probabilities (PP) are obtained from analyses based on 1 48-taxon data set in Figure 5.3. The portions of the tree in the gray rectangular boxes indicate the nomenclatural changes of the genus Cordyceps.

The present phylogenetic analyses of the 1 62-taxa combined revealed three strongly supported monophyletic groups (i.e. Clavicipitaceae dade A, B and C in

Figure 5.1 and 5.2) of the clavicipitaceous fungi (Figure 5.1 and Figure 5.2). In reviewing the diagnostic characters used in current classification schemes, most characters are not consistent with the phylogeny presented here. Therefore, the phylogenetic relationships of Cordyceps and the clavicipitaceous fungi provide the evidence for the rejection of most of the classification of Cordyceps and the

Clavicipitaceae (Diehl, 1950; Kobayasi, 1941; Kobayasi, 1982; Mains, 1958). Here, we propose a new phylogenetic classification for Cordyceps and the Clavicipitaceae

(Figure 5.2 and Figure 5.7).

Clavicipitaceae (Lindau) Erikss. emend. prov. Sung & Spatafora

Type genus- Claviceps Tul.

Teleomorphic genera- Ascopolyporus A. Möller, Atkinsonella Diehl, Balansia,

Claviceps, Epichloë, Heteroepichloe Tanaka, Tanaka, Gafur & Tsuda, Hypocrella, 183

Metacordyceps, Myriogenospora G.F. Atk., Neoclaviceps Kobayasi, Parepichloe

White & Reddy, Regiocrella Chaverii, Hodge & Samuels, Shimizuomyces Kobayasi.

Anamorphic genera- Aschersonia, Ephellis E. M. Fries, Metarhizium,

Neotyphodium A. B. Glenn, C. W. Bacon & Hanlin, Nomuraea, Paecilomyces sect.

Isarioidea, Pochonia, Sphacelia Levéillé, Tolypocladium, Verticillium sect Prostrata.

Clavicipitaceae s. s. is recognized as the well-supported, monophyletic

Clavicipitaceae dade A (MP-BP =99 %, WP-BP = 100%, PP = 100% in Figure 5.1,

5.2 and 5.7). The family name Clavicipitaceae has been used for more than 100 years based on its unique diagnostic characters (e.g., cylindrical asci and fihiform ascospores) since it was first used in 1901 by Earle for Hypocreaceae subf.

Clavicipiteae Lindau (Earle, 1901). However, Earle (1901) first used it without description and reference to its basionym as Hypocreaceae subf. Clavicipiteae and thus the name had been invalidly used by subsequent workers suchas Nannfeldt and

Diehl until it was validated in 1982 by Eriksson (Diehi, 1950; Eriksson, 1982;

Nannfeldt, 1932). Although Clavicipitaceae is well characterized by cylindrical asci, thickened ascus apices, and fihiform ascospores that tend to disarticulate at maturityas in the original description, Clavicipitaceae s.s (Figure 5.7) is only applied to the members of Clavicipitaceae dade A due to the non-monophyly of Clavicipitaceae s.l.

(Chapter 4). These findings suggest that the character states of cylindrical asci and filiform ascospores that disarticulate at maturity are symplesiomorphic for the

Clavicipitaceae s.l.IHypocreaceae dade. Importantly, the Hypocreaceae alsopossess 184

cylindrical asci and while its ascospores are subglobose to fusiform, they show a high

frequency of disarticulation (Rogerson, 1970; Rossman et al., 1999).

Clavicipitaceae s.s includes grass-associated genera (e.g., Balansia, Claviceps,

Epichloe and Myriogenospora) that were classified in Clavicipitaceae subf.

Clavicipitoideae sensu Diehi 1950 (Figure 5.7). In addition, the recent molecular studies show that many genera (e.g., Atkinsonella, Heteroepichloe, Neoclaviceps and

Parepichloe) placed in this subfamily are closely related with the grass-associated genera, indicating that these genera are placed in Clavicipitaceae s. s (Sullivan et al.,

2001; Tanaka et al., 2002; White and Reddy, 1998). Clavicipitaceae s. s also includes the recently described genus, Regiocrella, which includes species that are pathogens of scale insects (Chaverri et al., 2006).

Metacordyceps Sung & Spatafora gen. prov.

Type species- Cordyceps taii Liang and Liu

Description- Stromata solitary or several, simple or branched. Stipe fleshyor tough, whitish, greenish yellow to greenish, cylindrical to enlarging in fertile part.

Fertile part cylindrical to clavate. Perithecia partially or completely immersed in stromata, ordinal or oblique in orientation. Asci cylindrical with thickenedascus apex.

Ascospores cylindrical, multiseptate, disarticulating into partsporesor remaining intact at maturity. Anamorphic genera- Metarhizium, Nomuraea, Paecilomyces sect. Isarioidea,

Pochonia, Tolypocladium.

The genus Metacordyceps is proposed for species of Cordyceps s. 1. in

Clavicipitaceae s. s. based on the phylogenetic placement of C. taii (Figure 5.1, 5.2 and 5.7). The genus is applied to C. taii dade that is strongly supported (MP-BP = 85

%, WP-BP = 83 %, PP = 100 % in Figure 5.1, 5.2 and 5.7). Among the members of the dade, the most well-known taxon is the anamorphic genus, Metarhizium, due to its importance in biological control (Evans, 2003; Samson et al., 1988). Currently, two species of Cordyceps (i.e., C. brittlebankisoides and C. tall) are known as teleomorphs of Metarhizium (Liang et al., 1991; Liu et al., 2001). The genus name Metacordyceps is here used to emphasize that the dade includes the species of Cordyceps s. 1. that produce Metarhizium anamorphs although other species of Cordyceps (e.g., C. chiamydosporia) in the dade are not connected to Metarhizium anamorphs.

Cordycipitaceae Wehmeyer emend. prov. Sung & Spatafora

Type genus- Cordyceps (L: Fr) Link.

Teleomorphic genera- Ascopolyporus, Cordyceps, Hyperdermium White,

Sullivan, Bills & Hywel-Jones, Torrubiella.

Anamorphic genera- Beauveria, Engyodontium, Isaria, Lecanicillium,

Mariannaea, Simplicillium. 186

The family name Cordycipitaceae was originally used in 1976 by Wehmeyer based on the type genus Cordyceps for Clavicipitaceae subf. Cordycipitoideae

(Wehmeyer, 1976). Due to the non-monophyly of Cordyceps, Cordycipitaceae is here applied to a strongly supported monophyletic Clavicipitaceae dade C that includes the type species, C. militaris (MP-BP =100 %, WP-BP = 97%, PP100% in Figure 5.1 and 5.2). Most of the species in the family parasitize hosts in leaf litter, moss, or upper soil layers and produce superficial to completely immersed perithecia on fleshy stromata or subiculum that are pallid or brightly colored.

The family contains species of Cordyceps and Torrubiella (Figure 5.5 and 5.7). the unispecific genus Phytocordyceps is also recognized as a member of this family and transferred to Cordyceps (Figure 5.6). In addition, the recent molecular study shows that species of the genera Ascopolyporus and Hyperdermium, both pathogens of scale insects, are also inferred to be the members of the family (Bichoff et al., 2005;

Sullivan et al., 2000).

Cordyceps (L.: Fr.) Link emend. prov. Sung & Spatafora

Type species- Cordyceps militaris (L.: Fr.) Fr.

Anamorphic generaBeauveria, Lecanicillium, Isaria, Mariannaea,

Microhilum Yip & Rath. 187

Species of Cordyceps s. s. are characterized by possessing fleshy stromata that

are pallid or brightly colored. However, its application in Cordycipitaceae is not clear,

because species of Torrubiella are interspersed among Cordyceps species in the basal

part of the Cordycipitaceae dade (Figure 5.7). The genus Torrubiella was erected in

1885 by Boudier with the type species T aranicida Boud. (Kobayasi and Shimizu,

1982). Cordyceps s. s. is narrowly applied to the strongly supported dade (MP-BP =

88%, WP-BP = 93%, PP = 100% in Figure 5.1, 5.2 and 5.7) that only includes

Cordyceps species closely related to C. militaris. The full extent to which the names

Cordyceps and Torrubiella will ultimately be applied awaits additional sampling of

Torrubiella, especially that of T aranicida.

Ophiocordycipitaceae Sung & S patafora Jam. prov.

Type genus- Ophiocordyceps Petch

Description- Stromata or subiculum darkly pigmented or rarely brightly

colored, tough, fibrous to pliant, rarely fleshy, often with sterile apices. Perithecia

superficial to completely immersed, ordinal or oblique in orientation. Asci cylindrical with thickened ascus apex. Ascospores cylindrical, multiseptate, articulatingor nondisarticulating into partspores.

Teleomorphic genera- Elaphocordyceps, Ophiocordyceps, Torrubiella Anamorphic genera- Haptocillium, Harposporium Lohde, Hirsutella,

Hymenostilbe, Nomuraea, Paraisaria, Paecilomyces sect. Isarioidea, Syngliocladium,

Tolypocladium, Verticillium

The new family Ophiocordycipitaceae is proposed with the type genus

Ophiocordyceps Petch and applied to the strongly supported Clavicipitaceae dade B

(MP-BP = 90%, WP-BP = 96%, PP100% in Figure 5.1, 5.2 and 5.7). Most of species in Ophiocordycipitaceae produce darkly pigmented stromata that are pliant, fibrous, or tough in texture. Ecologically, many species of the family are known as pathogens of subterranean or wood-inhabiting hosts, deeply buried in soil or imbedded in decaying wood.

Elaphocordyceps Sung & Spatafora gen. prov.

Type species- Cordyceps ophioglossoides (Fr.) Link

Description- Stromata solitary to several, simple or branched. Stipe fibrous to tough, rarely fleshy, dark brownish to greenish with olivaceous tint, rarely whitish, cylindrical to enlarging in fertile part, rarely connected to fertile part directly through rhizomorph-like structure. Fertile part clavate to capitate, rarely undifferentiated.

Perithecia partially or completely immersed in stromata, ordinal in orientation. Asci cylindrical with thickened ascus apex. Ascospores cylindrical, multiseptate, disarticulating into partspores. Anamorphic genera- Tolypocladium, Verticillium sect. Prostrata.

C. ophioglossoides dade is strongly supported (MP-BP = 74%, WP-BP = 82%,

PP = 100% in Figure 5.1, 5.2 and 5.7) and include species of Cordyceps s. 1. that parasitize the truffle genus Elaphomyces and cicadas (e.g., C. inegoensis and C. paradoxa) and beetles (e.g., C. subsessilis) (Figure 5.5, Figure 5.7). Currently, 25 species are anticipated to be included in C. ophioglossoides dade, of which more than

18 species are known as parasites of Elaphomyces (Kobayasi and Shimizu, 1960;

Kobayasi and Shimizu, 1963; Mains, 1957). The host affiliation of Elaphomyces parasites has long been recognized as a diagnostic character in Cordyceps classification (Kobayasi, 1941; Kobayasi, 1982; Mains, 1957; Mains, 1958; Massee,

1895). The oldest applicable genus name is Cordylia Fr. 1823 (Massee, 1895).

However, its application to C. ophioglossoides dade is problematic because it is a homonym of Cordylia Pers. 1807 (Mains, 1958), which is also homonym of Cordyla

Lour. 1790 (Leguminosae). Therefore, the genus Elaphocordyceps is proposed based on the phylogenetic placement of C. ophioglossoides and applied to the well- supported C. ophioglossoides dade. Although C. subsessilis is morphologically and ecologically distinct, the genus is well recognized by its dominant ecology as being pathogens of Elaphomyces and cicadas. The darkly pigmented, fibrous stromata with more or less olivaceous tint are also a good diagnostic character for recognizing the species of Elaphocordyceps. 190

Ophiocordyceps Petch emend. prov. Sung & Spatafora

Type species- Cordyceps unilateralis (Tul.) Petch

Anamorphic genera- Hirsutella, Hymenostilbe, Paraisaria and Syngliocladium

The C. unilateralis dade is strongly supported (MP-BP = 88%, WP-BP = 92%,

PP100% in Figure 5.3 and 5.7) and includes the species of Ophiocordyceps (e.g., C. acicularis and C. unilateralis) (Petch, 1931; Petch, 1933). The genus Ophiocordyceps was proposed by Petch (1931; 1933) for species of Cordyceps that produce nondisarticulating ascospores and the type of the genus is 0. blattae Petch. The genus, however, was not accepted by subsequent workers (Kobayasi, 1941; Mains, 1958) who reclassified the species of Ophiocordyceps as Cordyceps subg. Ophiocordyceps, which is not recognized as a monophyletic group in these analyses (Figure 5.4, 5.5 and

5.7). Therefore, the genus Ophiocordyceps is applied to C. unilateralis dade based on the lectotypification of C. unilateralis. The type of Ophiocordyceps Petch is 0. blattae, but it was only collected once (Petch, 1931) and is no longer available for the taxonomic treatment. Ophiocordyceps unilateralis was included in the original publication of Ophiocordyceps (Petch 1931) and thus is selected as the lectotype for the genus. The genus Ophiocordyceps includes the most morphologically diverse group of the species of Cordyceps s. 1. including the members of C. subg.

Neocordyceps (Figure 5.5 and 5.7). For most of the species in Ophiocordyceps, the stromata are fibrous to tough to pliant in texture and darkly pigmented in at least its 191

some part of the stroma. In morphology of stromata, the genus includes many species of Cordyceps s. 1. that produce perithecia in its subterminal regions resulting in sterile apices. Of particular note, Ophiocordyceps is characterized by the dominant occurrence of Hirsutella and Hymenostilbe anamorphs (Figure 5.5). The genus

Cordycepioideus is here transferred to the genus Ophiocordyceps by its placement in these analyses and possession of a Hirsutella anamorph.

Clavicipitaceae

Teleomorphic genera- Aciculosporium I. Miyake, Berkelella (Sacc.) Sacc.,

Cavimalum Yoshim. Doi, Dargan & K.S. Thind, Dussiella Pat., Epicrea (Fr.) Tul. &

C. Tul., Konradia Racib., Microstelium Pat., Bres., Mycomulus A.

Möller, Lowen, Podocrella Seaver, Romanoa Thirum., Sphaerocordyceps

Kobayasi, Stereocrea Syd. & P. Syd.

Accepted names and new corn binatio,,s

The accepted species and new combinations for the species thatwere previously classified in Cordyceps are listed following the concept of the classification in this study. We only listed the accepted names andnew combinations that were confidently assigned based on the original descriptions. The remaining species of

Cordyceps are provisionally retained within Cordycepssensu lato until further 192 phylogenetic analyses are conducted to classify them in this phylogenetic classification.

Cordyceps (L.: Fr.) Link, Handbuch 3:347. 1833.

Cordyceps belizensis Mains, Mycologia 32:21. 1940.

Cordyceps bfusispora 0. Erikss. Mycotaxon 15:185. 1982.

Cordyceps erotyli Petch, Trans. Brit. Myc. Soc. 21:40. 1937.

Cordyceps exasperata Vital, Ann. Soc. Biol. Pernambuco 14:65. 1956.

Cordycepsfasciculata Pat., Bull. Soc. Myc. France 15:206. 1899.

Cordycepsflavobrunnescens P. Henn., in Warburg, Monsunia 1:164. 1900.

Cordyceps form osana Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot)7:113. 1981.

Cordyceps grenadensis Mains, Bull. Torrey Bot. Club 81:499. 1954.

Cordycepsgrylli Teng, Sinensia 7:811. 1936.

Cordyceps hokkaidoensis Kobayasi, Series Reports of the Tokyo Bunrika Daigaku section B. no. 84. p. 91-92. 1941.

Cordyceps isarioides Curt. & Massee, Ann. Bot. 9:36. 1895.

Cordyceps kyusyuensis Kawamura, Icon. Jap. Fungi 8:841. 1955.

Cordyceps locustiphila Henn., Hedwigia 43:247. 1904.

Cordyceps macularis (Mains) Kobayasi, Science Reports of the Tokyo Bunrika Daigaku section B. 84:70. 1941.

Cordyceps militaris (L.: Fr.) Link, HandbUch 3:347. 1833. EClavaria militaris L., Sp. Plantarump. 1182. 1753. Hypoxylon militaris (L.) Meret, Nouv. Flore des Environs de Paris.p. 137. 1821. Xylaria militaris (L.) Gray, Nat. Am Brit. P1. (London)p. 510. 1821. 193

-Spharia militaris (L.) Per. Syst. Myc. 2:325. 1823. -=Torrubia militaris Tul., Sel. Fung. Carp. 3:6. 1865.

Cordyceps miryensis Henm Hedwigia 18:247. 1904.

Cordyceps mitrata Pat., Bull. Soc. Myc. France 14:196. 1898.

Cordyceps nikkoensis Kobayasi, Series Reports of the Tokyo Bunrika Daigaku section B. no. 84. p. 134. 1941.

Cordyceps ninchukispora (Su & Wang) Sung & Spatafora comb. nov. -=Phytocordyceps ninchukispora Su & Wang, Mycotaxon 26:338. 1986.

Cordyceps ochraceostromata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo 5cr. B (Bot) 6:132. 1980.

Cordyceps ogurasanensis Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:80. 1982.

Cordycepspolyarthra Möller, Phyco. u. Asco. 213. 1901.

Cordycepspruinosa Petch, Trans. Brit. Myc. Soc. 10:38. 1924.

Cordyceps rosea Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot)8:112. 1982.

Cordyceps roseostromata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:10. 1983.

Cordyceps scarabaeicola Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 2:137. 1976.

Cordyceps staphylinidaecola Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:88-91. 1982.

Cordyceps takaomontana Yakusiji & Kumazawa, Series Reports of the Tokyo Bunrika Daigaku section B. no. 84.p. 108. 1941.

Cordyceps tarapotensis Henn., Hedwigia 43:246. 1904.

Cordyceps truncata Moureau, Mem. Inst. Royal Colonial Belge 7:19. 1949.

Cordyceps typhulaeformis Berk. & Cooke, Grevillea 12:78. 1884. 194

Cordyceps washinglonensis Mains, Mycologia 39:535. 1947.

Elaphocordyceps Sung & Spatafora geli. prov.

Elaphocordyceps canadensis (Eli. & Everh.) Sung & Spatafora Comb. prov. =Cordyceps canadensis Eli. & Everh., Bull. Torrey Bot. Club 25:501. 1898.

Elaphocordyceps capitata (Holmsk.) Sung & Spatafora Comb. prov. =Cordyceps capitata (Hblmsk.) Link, Handbuck zur Erkeimung der Nutzbarsten und am Haufigsten Vorkommenden Gewächse 3: 347. 1833. =Sphaeria capitata Holmsk., Beata Runs Otia Fungis Danicis 1:38. 1790.

Elaphocordyceps delicatostipitata (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps delicatostipitata Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 5:79-80. 1960.

Elaphocordycepsfracta (Mains.) Sung & Spatafora Comb. prov. Cordycepsfracta Mains, Bull. Torrey Bot. Club 84:250-25 1. 1957.

Elaphocordyceps inegoensis (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps inegoensis Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:292-293. 1963.

Elaphocordyceps interm ed/a (Imai) Sung & Spatafora Comb.prov. ECordyceps intermedia Imai, Proc. Imp. Acad. Tokyo 10:677. 1934

Elaphocordycepsjaponica (Lloyd) Sung & Spatafora Comb. prov. Cordycepsjaponica Lloyd, Myc.Writings 6:913. 1920.

Elaphocordycepsjezoensis (Imai) Sung & Spatafora Comb.prov. Cordycepsjezoensis Imai, Trans. Sapporo Nat. Hist. Soc. 11:33. 1929.

Elaphocordycepsjuruensis (Henn.) Sung & Spatafora Comb.prov. Cordycepsjuruensis Henn., Hedwigia 43:248. 1904.

Elaphocordyceps longisegmentis (Ginns) Sung & Spatafora Comb.prov. =Cordyceps longisegmentis Ginns, Mycologia 80:219. 1988.

Elaphocordyceps minazukiensis (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. 195

=Cordyceps minazukiensis Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:117. 1982.

Elaphocordyceps miomoteana (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps miomoteana Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:118. 1982.

Elaphocordyceps nigriceps (Peck.) Sung & Spatafora Comb. prov. Cordyceps nigriceps Peck., Bull. TorreyBot. Club 27:21. 1900.

Elaphocordyceps ophioglossoides (Fr.) Sung & Spatafora Comb. prov. =Sphaeria ophioglossoides Fr., Syst. Myc. 2:324. 1823. =Cordyceps ophioglossoides (Fr.) Link, Handbuch 3:347. 1933. Torrubia ophioglossoides (Fr.) Tul., Sel. Fung. Carp. 3:20. 1865.

Elaphocordyceps paradoxa (Kobayasi) Sung & Spatafora Comb.prov. Cordycepsparadoxa Kobayasi, Bull. Biogeogr. Soc. Jap. 4:156. 1939.

Elaphocordyceps rouxii (Candussau) Sung & Spatafora Comb.prov. Cordyceps rouxii Candoussau, Mycotaxon 4:544. 1976.

Elaphocordyceps ryogamimontana (Kobayasi) Sung & Spatafora Comb.prov. ECordyceps ryogamimontana Kobayasi, Bull. Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:303. 1963.

Elaphocordyceps subsessilis (Petch) Sung & Spatafora Comb.prov. Cordyceps subsessilis Petch, Trans. Brit. Myc. Soc. 21:39. 1937.

Elaphocordyceps ten uispora (Mains) Sung & Spatafora Comb.prov. =Cordyceps tenuispora Mains, Bulletin of the Torrey Botanical Club 84:247. 1957.

Elaphocordyceps toriharamontana (Kobayasi) Sung & Spatafora Comb.prov. =Cordyceps toriharamontana Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser.B (Bot) 6:305. 1963.

Elaphocordyceps umemurai (Imai) Sung & Spatafora Comb.pray. ECordyceps umemurai Imai, Trans. Sapporo Nat. Hist. Soc. 11:32.1929.

Elaphocordyceps vallformis (Mains) Sung & Spatafora Comb.prov. Cordyceps vallformis Mains, Bulletin of the Torrey Botanical Club 84:250. 1957. 196

Elaphocordyceps valvatostipitata (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps valvatostipitata Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 5:81. 1960.

Elaphocordyceps virens (Kobayasi) Sung & Spatafora Comb. prov. ECordyceps virens Kobayasi, J. Jap. Bot. 58:222. 1983.

Metacordyceps Sung & Spatafora gen. prov.

Metacordyceps brittlebankii (McLennan & Cookson) Sung & Spatafora Comb. prov. ECordyceps brittlebankii McLennan & Cookson, Proc. Roy. Soc. 38:74. 1926.

Metacordyceps brittlebankisoides Sung & Spatafora Comb. prov. Cordyceps brittlebankisoides Liu, Liang, Whalley, Yao & Liu, J. of Invert. Pathol. 78:179. 2001.

Metacordyceps chiamydosporia (Evans) Sung & Spatafora Comb. prov. =Cordyceps chiamydosporia Evans, Nova Hedwigia 73:59. 2001.

Metacordyceps taii (Liang & Liu) Sung & Spatafora Comb. prov. =Cordyceps taii Liang & Liu, Acta Mycologica Sinica 10:257. 1991.

Ophiocordyceps Petch Trans. Brit. Myc. Soc. 16:74. 1931.

Ophiocordyceps acicularis (Ray.) Petch, Trans. Brit. Myc. Soc. 18:60. 1933. =Cordyceps acicularis Ray., Jour. Linn. Soc. 1:158. 1857.

Ophiocordyceps aemonae (Lloyd) Sung & Spatafora Comb. prov. ECordyceps aemonae Lloyd, Myc. Writ. 6:932. 1920.

Ophiocordyceps agriota (Kawam.) Sung & Spatafora Comb. prov. =Cordyceps agriota Kawam., Icon. Jap. Fungi 8:837. 1955.

Ophiocordyceps ainictos (Moller) Sung & Spatafora Comb. prov. =Cordyceps ainictos Moller, Phyco. u. Asco.p. 226. 1901.

Ophiocordyceps amazonica (Henn.) Sung & Spatafora Comb.prov. =Cordyceps amazonica Henn., Hedwigia 43:247. 1904. 197

Ophiocordyceps aphodii (Mathieson) Sung & Spatafora Comb. prov. Cordyceps aphodii Mathieson, Trans. Brit. Myc. Soc. 32:134. 1949.

Ophiocordyceps aphrophorae (Yasuda) Sung & Spatafora Comb. prov. Cordyceps aphrophorae Yasuda, Bot. Mag. Tokyo 36:51. 1922.

Ophiocordyceps appendiculata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps appendiculata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:6. 1983.

Ophiocordyceps arbuscula (Teng) Sung & Spatafora Comb. prov. Cordyceps arbuscula Teng, Sinensia 7:812. 1936.

Ophiocordyceps asyuensis (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps asyuensis Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:138-139. 1980.

Ophiocordyceps aurantia (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps aurantia Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:125. 1980.

Ophiocordyceps australis (Speg.) Sung & Spatafora Comb. prov. Cordyceps australis Speg., Syll. fung. 2:571. 1883.

Ophiocordyceps barnesii (Thwaites ex Berk. & Br.) Sung & Spatafora Comb. prov. =Cordyceps barnesii Thwaites ex Berk. & Br., Journal. Bot. Linn. Soc. 14:110. 1875.

Ophiocordyceps bicephala (Berk.) Sung & Spatafora Comb. prov. Cordyceps bicephala Berk., Hooker's J. Bot. 8:278. 1856. Cordyceps bicephala subsp. bicephala (Berk.) Moureau, Mem. Inst. Roy. Col. Beige 7(fasc.5):50. 1949.

Ophiocordyceps blattae (Petch) Petch, Trans. Brit. Myc. Soc. 16:74. 1931. Cordyceps blattae Petch, Trans. Bnt. Myc. Soc. 10:35-36. 1924.

Ophiocordyceps bisporus (Stifler) Sung & Spatafora Comb.prov. Cordycepioideus bisporus Stifler, Mycologia 76:764-765. 1984. Ophiocordyceps brunneapunctata (Hywel-Jones) Sung & Spatafora Comb. prov. Cordyceps brunneapunctata Hywel-Jones, Mycol. Res. 99:1195. 1995.

Ophiocordyceps carabi (Qué!.) Sung & Spatafora Comb. prov. ECordyceps carabi Qué!., Comp. Rend. Assoc. Franc. Avanc. Sci. 26:452. 1898.

Ophiocordyceps carabdiicola (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps carabdiicolai Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:85. 1980.

Ophiocordyceps cicadicola (Teng) Sung & Spatafora Comb. prov. Cordyceps cicadicola Teng, Sinensia 6:191. 1935.

Ophiocordyceps cinerea (Tul.) Sung & Spatafora Comb. prov. Torrubia cinerea Tul., Sd. Fung. Carp. 1:61. 1861. ECordyceps cinerea (Tul.) Sacc., Michelia. 1:320. 1879.

Ophiocordyceps clavata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps clavata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:140-145. 1980.

Ophiocordyceps coccidiicola (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordyceps coccidiicola Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 4:57. 1978.

Ophiocordyceps coccigena (Tul.) Sung & Spatafora Comb. prov. Cordyceps coccigena (Tul.) Sacc., Michelia 1:320. 1879.

Ophiocordyceps cochlidiicola (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordyceps cochlidiicola Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:128. 1980.

Ophiocordyceps corallomyces (Möller) Sung & Spatafora Comb. prov. Cordyceps corallonzyces M011er, Phyco. u. Asco.p. 127. 1901.

Ophiocordyceps crassispora (Zang, Yang & Li) Sung & Spatafora Comb. prov. ECordyceps crassispora Zang, Yang & Li, Mycotaxon 37:58. 1990. 199

Ophiocordyceps curculionum (Tul.) Sung & Spatafora Comb. prov. Torrubia curculionum Tul. Se!. Fung. Carp. 3:20. 1865. Cordyceps curculionum (Tul.) Sacc. Michelia 1:320. 1879. =Cordyceps bicephala subsp. curculionum (Tul.) Moureau, Mem. Inst. Roy. Col. Beige 7(fasc.5):50. 1949.

Ophiocordyceps cusu (Pat.) Sung & Spatafora Comb. prov. =Cordyceps cusu Pat., Bull. Soc. Myc. France p. 229. 1895.

Ophiocordyceps depokensis (Koord & Overeem) Sung & Spatafora Comb. prov. =Cordyceps depokensis Koord & Overeem, Trop. Natuur 16. p. 174. 1925.

Ophiocordyceps dipterigena (Berk. & Br.) Sung & Spatafora Comb. prov. ECordyceps dzpterigena Berk. & Br. Jour. Linn. Soc. 14:111. 1875.

Ophiocordyceps discoideocapitata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps discoideocapitata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:85. 1982.

Ophiocordyceps ditmarii (Qué!.) Sung & Spatafora Comb. prov. Cordyceps ditmarii Quel., Buil. Soc. Bot. France 24:330. 1877.

Ophiocordyceps dovei (Rodway) Sung & Spatafora Comb. prov. =Cordyceps dovei Rodway, Paper Proc. Roy. Soc. for the year 1898-1899 p. 101. 1900.

Ophiocordyceps elateridicola (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. Cordyceps elateridicola Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:4. 1983.

Ophiocordyceps elongata (Petch) Sung & Spatafora Comb. prov. =Cordyceps elongata Petch, Trans. Brit. Myc. Soc. 2 1:47. 1937.

Ophiocordyceps elongatoperitheciata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordyceps elongatoperitheciata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:126. 1980.

Ophiocordyceps elongatostromata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. 200

ECordyceps elongatostromata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:12. 1983.

Ophiocordyceps emeiensis (A. Y. Liu & Z. Q. Liang) Sung & Spatafora Comb. prov. Cordyceps emeiensis A. Y. Lie & Z. Q. Liang, Mycosystema 16:139. 1997.

Ophiocordyceps entomorrhiza (Dicks.) Sung & Spatafora Comb. prov. =Sphaeria entomorrhiza Dicks., Plant. Crypt. Brit. 22. tab.3, s. 3. 1785. Cordyceps entomorrhiza (Dicks.) Link, Handbuch 3:347. 1833.

Ophiocordyceps evdogeorgiae (Koval) Sung & Spatafora Comb. prov. Cordyceps evdogeorgiae Koval, Botanicheskii Materiari 14:160. 1961.

Ophiocordycepsfalcata (Berk.) Sung & Spatafora Comb. prov. ECordycepsfalcata Berk., Decad. Fung. no. 479 in Hook. Journ. Bot. 6:211. 1854.

Ophiocordycepsfalcatiodes (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordycepsfalcatiodes Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6(3):91. 1980.

Ophiocordycepsfasciculatostromata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. Cordycepsfasciculatostromata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:83-84. 1982.

Ophiocordycepsferruginosa (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordycepsferruginosa Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:139. 1980.

Ophiocordycepsfilformis (Moureau) Sung & Spatafora Comb. prov. Cordycepsfilformis Moureau, Mem. Inst. Royal. Cot. Beige 7:14. 1948.

Ophiocordycepsformicarum (Kobayasi) Sung & Spatafora Comb. prov. Cordycepsformicarum Kobayasi, Bull. Biogeogr. Soc. Jap. 9:286. 1939.

Ophiocordycepsforquignonii (Qué!.) Sung & Spatafora Comb. prov. 201

=CordycepsforquignoniiQuel.,Fawcett, Ann. Mag. Nat. Hist. 18. 5:3 17. 1886.

Ophiocordyceps geniculata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps geniculata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:85. 1980.

Ophiocordyceps goniophora (Speg.) Sung & Spatafora Comb. prov. =Cordyceps goniophora Speg., Fungi Puiggarini no. 307 in Bol. Acad. Nat. Cordiva6. 1889.

Ophiocordyceps gentilis (Ces.) Sung & Spatafora Comb. prov. =Torrubia gentilis Ces., Myc. Born. p. 14. =Cordycepsgentilis (Ces.) Sacc., Syll. Fung. 2:569. 1883.

Ophiocordyceps glaziovii (Henn.) Sung & Spatafora Comb. prov. =Cordyceps glaziovii Henn., Naturw. Wochenschr. 6:3 18. 1896.

Ophiocordyceps gracilioides (Kobayasi) Sung & Spatafora Comb. prov. =Cordyceps gracilioides Kobayasi, Series Reports of the Tokyo Bunrika Daigaku section B. no. 84. p. 140-141.

Ophiocordyceps gracilis (Grey.) Sung & Spatafora Comb. prov. Xylaria gracilis Grey., Scot. Fl. t. 86. 1824. Cordyceps gracilis (Grey.) Durieu & Mont., F!. Algérie Crypt. 1:449. 1846.

Ophiocordyceps heteropoda (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps heteropoda Kobayasi, Bull, the Biogeogr. Soc. Jap. 9:158. 1939.

Ophiocordyceps huberiana (Heim.) Sung & Spatafora Comb. prov. =Cordyceps irangiensis Henn., Hedwigia 48:105. 1909.

Ophiocordyceps hiugensis (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps hiugensis Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:3. 1983.

Ophiocordyceps huegelii (Corda) Sung & Spatafora Comb. prov. Sphaeria huegelii Corda, Icon. Fung. 4:44. 1840. ECordyceps huegelii (Corda) Corda, Anleit. Stud. Mycol.p. 207. 1842. 202

Ophiocordyceps irangiensis (Moureau) Sung & Spatafora Comb. prov. Cordyceps irangiensis Moureau, Lej eunea 15:33. 1961.

Ophiocordycepsjaponensis (Hara) Sung & Spatafora Comb. prov. ECordycepsjaponensis Hara, Bot. Mag. Tokyo 28:35 1. 1914.

Ophiocordyceps kangdingensis (Zang & N. Kinjo) Sung & Spatafora Comb. prov. -Cordyceps kangdingensis Zang & N. Kinjo, Mycotaxon 66:221-222. 1998.

Ophiocordyceps konnoana (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =-Cordyceps konnoana Kobayasi & Shimizu, Bull. Nat. Sd. Mus., Tokyo 5cr. B (Bot) 6:84. 1980.

Ophiocordyceps koreana (Kobayasi) Sung & Spatafora Comb. prov. =Cordyceps koreana Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 7:8. 1981.

Ophiocordyceps lachnopoda (Penz. & Sacc.) Sung & Spatafora Comb. prov. ECordyceps lachnopoda Penz. & Sacc., Icons Fung. Javan. p. 55. 1904.

Ophiocordyceps larvarum (Westwood) Sung & Spatafora Comb. prov. -=Sphaeria larvarum Westwood, Proc. Ent. Soc. Lond. 2:6. 1836. =Cordyceps larvarurn (Westwood) 011iff, Agricult. Gazett. 6:410. 1895.

Ophiocordyceps iioydii (Fawcett) Sung & Spatafora Comb. prov. -Cordyceps lloydii Fawcett, Ann. Mag. Nat. Hist. 18. 5:317. 1886.

Ophiocordyceps ion gissima (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps longissima Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:292. 1963.

Ophiocordyceps mawieyi (Westwood) Sung & Spatafora Comb. prov. -=Cordyceps mawieyi Westwood, Gard. Chron. Ser. 3. 4:553. 1891.

Ophiocordyceps melolonthae (Tul.) Sung & Spatafora Comb. prov. ETorrubia melolonthae Tul., Se!. Fung. Carp. 3:12. 1865. ECordyceps meioionthae (Tul.) Sacc., Michelia 1:320. 1879.

Ophiocordyceps michiganensis (Mains) Sung & Spatafora Comb. prov. ECordyceps michiganensis Mains, Am. Phil. Soc. Proc. 74:266. 1934. 203

Ophiocordyceps minutissima (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps minutissima Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:77. 1980.

Ophiocordyceps monticola (Mains) Sung & Spatafora Comb. prov. ECordyceps monticola Mains, Mycologia 32:3 10. 1940.

Ophiocordyceps multiaxialis (Zang & N. Kinjo) Sung & Spatafora Comb. prov. Cordyceps multiaxialis Zang & N. Kinjo, Mycotaxon. 66:224. 1998.

Ophiocordyceps muscae (Henn.) Sung & Spatafora Comb. prov. =Cordyceps muscae Henn., Engler, Bot. Jahrb. 25:507. 1898.

Ophiocordyceps muscicola (Möller) Sung & Spatafora Comb. prov. ECordyceps muscicola Möller, Phyco. u. Asco.p. 221. 1901.

Ophiocordyceps myrmecophila (Ces.) Sung & Spatafora Comb. prov. =Cordyceps myrmecophila Ces., Bot. Zeit. 4:877. 1846. Torrubia myrmecophila Tul., Sd. Fung. Carp. 3:18. 1865.

Ophiocordyceps neo-volkiana (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps neo-volkiana Kobayasi, Series Reports of the Tokyo Bunrika Daigaku section B. no. 84. p. 169-170. 1941.

Ophiocordyceps nepalensis (Zang & N. Kinjo) Sung & Spatafora Comb. prov. Cordyceps nepalensis Zang & N. Kinjo, Mycotaxon. 66:224-225. 1998.

Ophiocordyceps nutans (Pat.) Sung & Spatafora Comb. prov. =Cordyceps nutans Pat., Bull. Soc. Mycol. Fr. 3:127. 1887. =Cordyceps bicephala subsp. nutans (Pat.) Moureau, Mem. Inst. Roy. Col. Belge 7:47. 1949.

Ophiocordyceps obtusa (Penz. & Sacc.) Sung & Spatafora Comb. prov. =Cordyceps obtusa Penz. & Sacc., Malpighia 11:523. 1897.

Ophiocordyceps octosporus (M. Blackwell & Gun) Sung & Spatafora Comb. prov. ECordycepioideus octosporus M. Blackwell & Gun, Mycologia 76:764. 1984.

Ophiocordyceps odonatae (Kobayasi) Sung & Spatafora Comb. prov. 204

=Cordyceps odonatae Kobayasi, Bull. Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 7:6. 1981.

Ophiocordyceps opposita (Syd.) Sung & Spatafora Comb. prov. =Cordyceps opposita Syd., Engler Bot. Jahrb. 5 7:325. 1922.

Ophiocordyceps osuzumontana (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps osuzumontana Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:77. 1980.

Ophiocordyceps ouwensis (Höhnel) Sung & Spatafora Comb. prov. Cordyceps ouwensis Höhnel, Sitz. Akad. Wiss. Wien 118:309. 1909.

Ophiocordyceps owariensis (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps owariensis Kobayasi, Bull. Biogeogr. Soc. Jap. 9:166. 1939.

Ophiocordyceps oxyceph ala (Penz. & Sacc.) Sung & Spatafora Comb. prov. =Cordyceps oxycephala Penz. & Sacc., Icon. Fung. Jay. 55. 1904. ECordyceps sphecocephala f. oxycephala (Penz. & Sacc.) Kobayasi, Trans. Myc. Soc. Jap. 23:361. 1982.

Ophiocordyceps paludosa (Mains) Sung & Spatafora Comb. nov. ECordycepspaludosa Pap. Mich. Acad. Sci. 25:83. 1940.

Ophiocordyceps pittieri (Bomm. & Rouss.) Sung & Spatafora Comb. prov. Cordycepspittieri Bomm. & Rouss., Bull. Soc. Bot. Beig. 35:160. 1896.

Ophiocordycepsprolferans (Henn.,) Sung & Spatafora Comb. prov. =Cordyceps prolferans Henn., Hedwigia 18:248. 1904.

Ophiocordyceps pseudolongissima (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. =Cordyceps pseudolongissima Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 8:119-123. 1982.

Ophiocordyceps purpureostromata (Kobayasi) Sung & Spatafora Comb. prov. Cordycepspurpureostromata Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:136-137. 1980.

Ophiocordyceps ravenelii (Berk & Curtis) Sung & Spatafora Comb. prov. ECordyceps ravenelii Berk & Curtis, Act. Soc. Lirm. 1:159. 1857.

Ophiocordyceps rhizoidea (HOhn.) Petch, Trans. Brit. Myc. Soc. 16:74. 1931. ECordyceps rhizoidea Höhn., Stzb. K. Ak. Wiss. Wien 118:307. 1909.

Ophiocordyceps rickii (Lloyd) Sung & Spatafora Comb. prov. -=Cordyceps rickii Lloyd, Myc. Writ. 6:9 14. 1920.

Ophiocordyceps ridleyi (Massee) Sung & Spatafora Comb. prov. Cordyceps ridleyi Massee., Bull. Miscell. Inform. Roy. Gard. Kew. p. 173. 1899.

Ophiocordyceps rubiginosoperitheciata (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordyceps rubiginosoperitheciata Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:14. 1983.

Ophiocordyceps ryogamiensis (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordyceps ryogamiensis Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 9:4. 1983.

Ophiocordyceps scottiana (011iff) Sung & Spatafora Comb. prov. ECordyceps scottiana 0111ff, Agr. Gaz. New South Wales 6:407. 1895.

Ophiocordyceps selkirkii (011iff) Sung & Spatafora Comb. prov. -Cordyceps selkirkii 011iff, Agr. Gaz. New South Wales 6:411. 1895.

Ophiocordyceps sherringii (Massee) Sung & Spatafora Comb. prov. Cordyceps sherringii Massee, Ann. Bot. 5:5 10. 1890.

Ophiocordyceps sinensis (Berk.) Sung & Spatafora Comb. prov. Sphaeria sinensis Berk., in Hooker, Lond. Joum. Bot. 2:207. 1843. Cordyceps sinensis (Berk.) Sacc., Michelia 1:320. 1878.

Ophiocordyceps sobolifera (Berk.) Sung & Spatafora Comb. prov. Sphaeria sobolfera Berk., Hook. Lond. Jour. Bot. 2:207. 1843. -Torrubia sobolfera (Berk.) Tul.Sd.Fung. Carp. 3:10. 1865. Cordyceps sobolfera (Berk.) Berk. & Br. Journ. Linn. Soc. 14:110. 1875.

Ophiocordyceps sphecocephala (Berk.) Sung & Spatafora Comb. prov. =Sphaeria sphecocephala Berk., Hook. Lond. Jour. Bot. 2:206. 1843. ETorrubia sphecocephala (Berk) Tul., Sel. Fngl. Carp. 3:18. 1865. ECordyceps sphecocephala (Berk) Sacc., Michelia 1:321. 1879.

Ophiocordyceps stylophora (Berk. & Broome) Sung & Spatafora Comb. prov. 206

ECordyceps stylophora Berk. & Br., Journ. Linn. Soc. Bot. 1:158. 1857.

Ophiocordyceps subdiscoidea (Henn.) Sung & Spatafora Comb. prov. ECordyceps subdiscoidea Henn., Hedwigia 41:168. 1902.

Ophiocordyceps superfIcialis (Peck) Sung & Spatafora Comb. prov. Torrubia superjicialis Peck, N. Y. State Botanist Rep. 28:70. 1876. ECordyceps superj'lcialis (Peck) Sacc., Syll. Fung. 2:574. 1883.

Ophiocordyceps surinarnensis (Henn.) Sung & Spatafora Comb. prov. ECordyceps surinamensis Henn., Hedwigia 15:169. 1902.

Ophiocordyceps takaoensis (Kobayasi) Sung & Spatafora Comb. prov. Cordyceps takaoensis Kobayasi, Series Reports of the Tokyo Bunrika Daigaku section B. no. 84. p. 130-132. 1941.

Ophiocordyceps termitophila (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. ECordyceps termitophila Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 4:56. 1978.

Ophiocordyceps thwaitesii (Lloyd) Sung & Spatafora Comb. prov. Cordyceps thwaitesii Lloyd, Myc. Writ. 6:1060. 1921.

Ophiocordyceps thyrsoides (A. Möller) Sung & Spatafora Comb. prov. ECordyceps thyrsoides A. M011er, Phyco. u. Asco.p. 221. 1901.

Ophiocordyceps tricentri (Yasuda) Sung & Spatafora Comb. prov. ECordyceps tricentri Yasuda, Myc. Wirt. 4.568. 1915.

Ophiocordyceps uchiyainae (Kobayasi & Shimizu) Sung & Spatafora Comb. prov. Cordyceps uchiyarnae Kobayasi & Shimizu, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:125. 1980.

Ophiocordyceps unilateralis (Tul.) Petch, Trans. Brit. Myc. Soc. 16:74. 1931. =Torrubia unilateralis Tul., Sel. Fung. Carp. 3:18. 1856. Cordyceps unilateralis (Tul.) Sacc., Syll. Fung. 2:570. 1883.

Ophiocordyceps variabilis (Petch) Sung & Spatafora Comb. prov. Cordyceps variabilis Petch, Trans. Brit. Myc. Soc. 21:44. 1937.

Ophiocordyceps voeltzkowii (Henn.) Sunh & Spatafora Comb. prov. 207

Cordyceps voeltzkowii Henn., In Voeltzkow, Reise Ostafrica 3:29. 1908.

Ophiocordyceps volkiana (A. Möller) Sunh & Spatafora Comb. prov. ECordyceps volkiana A. Möller, Phyco. u. Asco.p. 233. 1901.

Ophiocordyceps yakusiinensis (Kobayasi) Sung & Spatafora Comb. prov. ECordyceps yakusimensis Kobayasi, Bull. Nat. Sci. Mus., Tokyo Ser. B (Bot) 6:302. 1963.

References

Artjariyasripong, S., J. I. Mitchell, N. L. Hywel-Jones, and E. B. Garath Jones. 2001. Relationships of the genus Cordyceps and related genera, based on parsimony and spectral analysis of partial 1 8S and 28S ribosomal gene sequences. Mycoscience 42:503-5 17.

Barron, G. L., and A. H. S. Onions. 1966. Verticillium chiamydosporium and its relationships to Diheterospora, Stemphyliopsis, and Paecilomyces. Canadian Journal of Botany 44:861-869.

Bichoff, J. F., P. Chaverri, and J. F. White, Jr. 2005. Clarification of the host substrate of Ascopolyporus and description of Ascopolyporus philodendrus sp. nov. Mycologia in press.

Bissett, J. 1983. Notes on Tolypocladium and related genera. Canadian Journal of Botany 61:1311-1329.

Blackwell, M., and R. L. Gilbertson. 1984. New information on Cordycepioideus bisporus and Cordycepioideus octosporus. Mycologia 76:763-765.

Chaven-i, P., J. F. Bischoff, H. C. Evans, and K. T. Hodge. 2006. Regiocrella, a new entomopathogenic genus with a pycnidial anamorph and its phylogenetic placement in the Clavicipitaceae. Mycologia in press.

Coyle, F. A., P. A. Goloboff, and R. A. Samson. 1990. Actinopus trapdoor spiders (Araneae, Actinopodidae) killed by the fungus, Nomuraea atypicola (Deuteromycotina). Acta. Zool. Fenn. 190:89-93. De Hoog, G. S. 1972. The genera Beauveria, Isaria, Tritirachium and Acrodontium gen. nov. Studies in Mycology No 1.

Diehi, W. W. 1950. Balansia and the Balansiae in America. Agricultural monograph No.4. U. S. D. A. Washington.

Doaudy, C. J., F. Delsuc, Y. Boucher, W. F. Doolittle, and E. J. P. Douzery. 2003. Comparison of the Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Molecular Biology and Evolution 20:248-254.

Driver, F., R. J. Mimer, and J. W. H. Trueman. 2000. A taxonomic revision of Metarhizium based on a phylogenetic analysis of rDNA sequence data. Mycological Research 104:134-150.

Earle, F. S. 1901. Collections of Alabama fungi. Pages 150-263 in Plant life of Alabama. Contr. U. S. Natl. Herb. (C. Mohr, ed.).

Eriksson, 0. 1982. Outline of the ascomycetes-1982. Mycotaxon 15:203-248.

Eriksson, 0. B., and D. L. Hawksworth. 1986. Notes on ascomycete systematics. Nos 1-224. Systema Ascomycetum 5:114-174.

Evans, H. C. 2003. Use of clavicipitalean fungi for the biological control of arthropod pests. Pages 517-548 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Evans, H. C., and R. A. Samson. 1982. Entomogenous fungi from the Galapagos islands. Canadian Journal of Botany 60:2325-2333.

Felsenstein, J. 1981. A likelihood approach to character weighting and what it tells us about parsimony and compatibility. Bio. J. Linn. Soc. 16:183-196.

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

Gams, W. 1971. Cephalosporiuin-artige Schmmelpilze (Hyphomycetes). G. Fischer, Stuttgart.

Gams, W., K. T. Hodge, R. A. Samson, R. P. Korf, and K. A. Seifert. 2005. Proposal to conserve the name Isaria (anamorphic fungi) with a conserved type. Taxon 54:537. Gams, W., and R. Zare. 2001. A revision of Verticillium sect. Prostrata. III. Generic classification. Nova Hedwigia 72:329-337.

Gams, W., and R. Zare. 2003. A taxonomic review of the clavicipitaceous anamorphs parasitizing nematodes and other microinvertebrates. Pages 17-73 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Greuter, W., J. McNeill, F. R. Barrie, H.-M. Burdt, V. Demoulin, T. S. Filgueiras, D. H. Nicolson, P. C. Silva, J. E. Skog, P. Trehane, N. J. Turland, and D. L. Hawksworth. 2000. Internationsi Code of Botanical Nomenclature (St Louis Code). Konigstein: Koeltz Scientific Books.

Hawksworth, D. L., and A. Y. Rossman. 1997. Where are all the undescribed fungi? Phytopathology 87:888-891.

Hodge, K. T. 2003. Clacicipitaceous anamorphs. Pages 75-123 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Hodge, K. T., R. A. Humber, and C. A. Wozniak. 1998. Cordyceps variabilis and the genus Syngliocladium. Mycologia 90:743-753.

Hodge, K. T., S. B. Krasnoff, and R. A. Humber. 1996. is the anamorph of Cordyceps subsessilis. Mycologia 88:715-719.

Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.

Humber, R. A., and M. C. Rombach. 1987. Torrubiella ratticaudata sp. nov. (Pyrenomycetes: Clavicipitales) and other fungi from spiders on the solomon islands. Mycologia 79:375-382.

Hywel-Jones, N. L. 1993. Torrubiella luteorostrata: a pathogen of scale insects and its association with Paecilomyces cinnamomeus with a note on Torrubiella tenuis. Mycological Research 97:1126-1130.

Hywel-Jones, N. L. 1994. Cordyceps khaoyaiensis and C. pseudomilitaris, two new pathogens of lepidopteran larvae from Thailand. Mycological Research 98:939-942. 210

Hywel-Jones, N. L. 1996. Cordyceps myrmecophila-like fungi infecting in the leaf litter of tropical forest in Thailand. Mycological Research 100:613-619.

Hywel-Jones, N. L. 2002. Multiples of eight in Cordyceps ascospores. Mycological Research 106:2-3.

Hywel-Jones, N. L., and H. C. Evans. 1993. Taxonomy and ecology of Hypocrella discoidea and its anarnorph, Ashersonia samoensis. Mycological Research 97:871-876.

Hywel-Jones, N. L., and G. J. Samuels. 1998. Three species of Hypocrella with large stromata pathogenic on scale insects. Mycologia 90:36-46.

Hywel-Jones, N. L., and S. Sivichai. 1995. Cordyceps cylindrica and its association with Nomuraea atypicola in Thailand. Mycological Research 99:809-8 12.

Kobayasi, Y. 1939. On the genus Cordyceps and its allies on cicadae from Japan. Bulletin of the Biogeographical Society of Japan 9:145-176.

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku (Section B, no. 84) 5:53-260.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364.

Kobayasi, Y., and D. Shimizu. 1960. Monographic studies of Cordyceps 1. Group parasite on Elaphomyces. Bulletin of the National Science Museum, Tokyo, Ser. B. 5:69-85.

Kobayasi, Y., and D. Shimizu. 1963. Monographic studies of Cordyceps 2. Group parasitic on Cicadae. Bulletin of the National Science Museum, Tokyo, Ser. B. 6:286-314.

Kobayasi, Y., and D. Shimizu. 1982. Monograph of the genus Torrubiella. Bulletin of the National Science Museum, Tokyo, Ser. B. 8:43-78.

Li, Z.-Z., C.-R. Li, B. Huang, and M.-Z. Fan. 2001. Discovery and demonstration of the teleomorph of Beauveria bassiana (Bals.) Vuill., an important entomopathogenic fungus. Chinese Science Bulletin 9:751-753.

Liang, Z.-Q., A.-Y. Liu, and J.-L. Liu. 1991. A new species of the genus Cordyceps and its Metarhizium anamorph. Acta Mycologica Sinica 10:257-262. 211

Liu, Z.-Y., Z.-Q. Liang, A.-Y. Liu, Y.-J. Yao, K. D. Hyde, and Z.-N. Yu. 2002. Molecular evidence for teleomorph-anamorph connections in Cordyceps based on ITS-5.8S rDNA sequences. Mycological Research 106:1100-1 108.

Liu, Z. Y., Z. Q. Liang, A. J. S. Whalley, Y.-J. Yao, and A. Y. Liu. 2001. Cordycpes brittlebankisoides, a new pathogen of grubs and its anamorph, Metarhizium anisopliae var. majus. Journal of Invertebrate Pathology 78:178-182.

Luangsa-Ard, J. J., N. L. Hywel-Jones, L. Manoch, and R. A. Samson. 2005. On the relationships of Paecilomyces sect. Isarioidea species. Mycological Research 109:581-589.

Luangsa-ard, J. J., N. L. Hywel-Jones, and R. A. Samson. 2004. The polyphyletic nature of Paecilomyces sensu lato based on 1 8S-generated rDNA phylogeny. Mycologia 96:773-780.

Lutzoni, F., F. Kauff, C. J. Cox, D. McLaughlin, G. Celio, B. Dentinger, M. Padamsee, D. Hibbett, T. Y. James, E. Baloch, M. Grube, V. Reeb, V. Hofstetter, C. Schoch, A. E. Arnold, J. Miadlikowska, J. Spatafora, D. Johnson, S. Hambleton, M. Crockett, R. Shoemaker, G.-H. Sung, R. Lucking, T. Lumbsch, K. O'Donnell, M. Binder, P. Diederich, D. Ertz, C. Gueidan, K. Hansen, R. C. Harris, K. Hosaka, Y.-W. Lim, B. Matheny, H. Nishida, D. Pfister, J. Rogers, A. Rossman, I. Schmitt, H. Sipman, J. Stone, J. Sugiyama, R. Yahr, and R. Vilgalys. 2004. Assembling the fungal tree of life: progress, classification, and evolution of subcellular traits. American Journal of Botany 91:1446-1480.

Macleod, D. M. 1954. Investigations on the genera Beauveria Vuill. and Tritirachium Limber. Canadian Journal of Botany 32:818-890.

Maddison, D. R., and W. P. Maddison. 2000. MacClade 4: Analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts.

Mains, E. B. 1950. Entomogenous species of Akanthomyces, Hymenostilbe and Insecticola in North America. Mycologia 42:566-5 89.

Mains, B. B. 1957. Species of Cordyceps parasitic on Elaphomyces. Bulletin of The Torrey Botanical Club. 84:243-25 1.

Mains, E. B. 1959. North American species of Aschersonia parasitic on Aleyrodidae. J. Insect. Pathol 1:43-47.

Mains, E. B. 1958. North American entomogenous species of Cordyceps. Mycologia 50: 169-222. 212

Mason-Gamer, R. J., and E. A. Kellogg. 1996. Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Systematic Biology 45:524-545.

Massee, G. 1895. A revision of the genus Cordyceps. Annals of Botany 9:1-44.

Mugnai, L., P. D. Bridge, and H. C. Evans. 1989. A chemotaxonomic evaluation of the genus Beauveria. Mycological Research 92:199-209.

Nannfeldt, J. A. 1932. Studien uber die morphologie und systematik der nicht- lichenisierten, inoperculaten Discomyceten. Nova Acta R. Soc. Scient. upsal. ser 6 8:1-368.

Obornik, M., M. Jirku, and D. Dolezel. 2001. Phylogeny of mitosporic entomopathogenic fungi: Is the genus Paecilomyces polyphyletic? Canadian Journal of Microbiology 47:813-819.

Ochiel, G. S., H. C. Evans, and J. Eilenberg. 1997. Cordycepioideus, a pathogen of termites in Kenya. Mycologist 11:7-9

Pacioni, G., and G. Frizzi. 1978. Paecilornycesfarinosus, the conidial state of Cordyceps memorabilis. Canadian Journal of Botany 56:39 1-394.

Petch, T. 1921. Studies in entomogenous fungi. II. The genera of Hypocrella and Aschersonia. Ann. R. Bot. Gard (Peradeniya) 7:167-278.

Petch, T. 1924. Studies in entomogenous fungus. IV. Some Ceylon Cordyceps. Transactions of the British Mycological Soceity 10:28-45.

Petch, T. 1931. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 16:55-75.

Petch, T. 1932. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 17:209-245.

Petch, T. 1933. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 18:48-75.

Petch, T. 1934. Cordyceps militaris and Isariafarinosa. Transactions of the British Mycological Soceity:2 16-224.

Phylippe, H., E. A. Snell, E. Bapteste, P. Lopez, P. W. H. Holland, and D. Casane. 2004. Phylogenomics of : impact of missing data on large alignments. Molecular Biology and Evolution2 1:1740-1752. 213

Posada, D., and K. A. Crandall. 1998. ModelTest: Testing the model of DNA substitution. Bioinformatics 14:817-818.

Reeb, V., F. Lutzoni, and C. Roux. 2004. Contribution of RPB2 to multilocus phylogenetic studies of the euascomycetes (Pezizomycota, Fungi) with special emphasis on the lichen-forming Acarosporaceae and evolution of polyspory. Molecular Phylogenetics and Evolution 32:1036-1060.

Rehner, S. A., and E. Buckley. 2005. A Beauveria phylogeny inferred from nuclear ITS and EF1-a sequences: evidence for cryptic diversification and links to Cordyceps teleomorphs. Mycologia 97:84-98.

Reynolds, D. R., and J. W. Taylor (eds) 1993. The fungal homomorph: Mitotic, meitic and pleomorphic speciation in fungal systematics. CAB International, Wallingford, UK.

Rogerson, C. T. 1970. The hypocrealean fungi (Ascomycetes, Hypocreales). Mycologia 62:865-9 10.

Rombach, M. C., R. A. Humber, and D. W. Roberts. 1986. Metarhiziumfiavoviride var. minus, var. nov., a pathogen of plant-and leathoppers on rice in the phillippines and solomon islands. Mycotaxon 27:87-92.

Rossman, A. Y., G. J. Samuels, J. S. Rogers, and R. Lowen. 1999. Genera of Bionectriaceae, Nectriaceae and Hypocreaceae (Hypocreales, Ascomycetes). Studies in Mycology No. 42.

Samson, R. A. 1974. Paecilomyces and some allied hyphomycetes. Studies in Mycology No 6.

Samson, R. A., and H. C. Evans. 1975. Notes on entomogenous fungi from Ghana III. The genus Hymenostilbe. Proc. Kon. Ned. Akad. Wetensch. C 78:73-79.

Samson, R. A., H. C. Evans, and J.-P. Latge. 1988. Atlas of entomopathogenic fungi. Springer-Verlag, Berlin Heidelberg New York.

Samson, R. A., E. S. Van Reenen-Hoekstra, and H. C. Evans. 1989. New species of Torrubiella (Ascomycotina: Clavicipitales) on insects from Ghana. Studies in Mycology 3 1:123-132.

Shima.zu, M., W. Mitshhashi, and H. Hashimoto. 1988. Cordyceps brongniartiisp. nov., the teleomorph of Beauveria brongniartii. Trans. Br. Mycol. Soc. 29:323-330. 214

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes. I. The Clavicipitales-Hypocreales connection. Mycologia 85:912-922.

Spatafora, J. W., G.-H. Sung, J.-M. Sung, N. L. Hywel-Jones, and J. F. White, Jr. 2006. Phylogenetic evidence for an animal pathogen origin for and the grass endophytes. PNAS submitted.

Stensrud, 0., N. L. Hywel-Jones, and T. Schumacher. 2005. Towards a phylogenetic classification of Cordyceps: ITS nrDNA sequence data confirm divergent lineages and paraphyly. Mycological Research 109:41-56.

Su, C.-H., and H.-H. Wang. 1986. Phytocordyceps, a new genus of the Clavicipitaceae. Mycotaxon 26:337-344.

Suh, S.-O., J. W. Spatafora, G. R. S. Ochiel, H. C. Evans, and M. Blackwell. 1998. Molecular phylogenetic study of a termite pathogen Cordycepioideus bisporus. Mycologia 90:611-617.

Sullivan, R., M. S. Bergen, R. Patel, G. F. Bills, S. C. Alderman, J. W. Spatafora, and J. F. White, Jr. 2001. Neoclaviceps monostipa: features of an enigmatic clavicipitalean fungus and the phyletic status of the anamorphic genus Ephelis. Mycologia 93:90-99.

Sullivan, R. F., G. F. Bills, N. L. Hywel-Jones, and J. F. White, Jr. 2000. Hyperdermium: a new clavicipitalean genus for some tropical epibionts of dicotyledonous plants. Mycologia 92:908-918.

Sung, G.-H., and J. W. Spatafora. 2004. Cordyceps cardinalis sp. nov., a new species of Cordyceps with an east Asian-eastern North American distribution. Mycologia 96:658-666.

Sung, G.-H., J. W. Spatafora, R. Zare, K. T. Hodge, and W. Gams. 2001. A revision of Verticillium sect. Prostrata. II. Phylogenetic analyses of SSU and LSU nuclear rDNA sequences from anamorphs and teleomorphs of the Clavicipitaceae. Nova Hedwigia 72:311-328.

Sung, J. 1996. The insects-born fungus of Korea in color. Kyohak PubI. Co., Seoul.

Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.OblO. Sinauer, Sunderland, Massachusetts. 215

Tanaka, E., C. Tanaka, A. Gafur, and M. Tsuda. 2002. Heteroepichloe, gen. nov. (Clavicipitaceae; Ascomycotina) on bamboo plants in East Asia. Mycoscience 43 :87-93.

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673-4680.

Wehineyer, L. E. 1976. The pyrenomycetes fungi. Mycol. Mem. 6. Cremer, Lehre.

Wheeler, W. C. 1990. Combinatorial weights in phylogenetic anlaysis: a statistical parsimony procedure. Cladistics 6:269-275.

White, J. J. F., and P. V. Reddy. 1998. Examination of structure and molecular phylogenetic relationships of some graminicolous symbionts in genera Epichloë and Parepichloë. Mycologia 90:226-234.

Wiens, J. J. 1998. Combining data sets with different phylogenetic histories. Systematic Biology 47:568-58 1.

Wiens, J. J. 2003. Missing data, Incomplete taxa, and phylogenetic accuracy. Systematic Biology 52:528-538.

Zang, M., L. Daoqing, and H. Ruoying. 1982. Notes concerning the subdivisions of Cordyceps and a new species from . Acta Bot. Yunnanica 4:173-176.

Zang, M., and N. Kinjo. 1998. Notes on the alpine Cordyceps of China and nearby nations. Mycotaxon 66:215-229.

Zare, R., and W. Gams. 2001a. A revision of Verticillium sect. Prostrata. N. The genera Lecanicillium. Nova Hedwigia 73:1-50.

Zare, R., and W. Gams. 200 lb. A revision of Verticillium sect. Prostrata. VI. The genus Haptocillium. Nova Hedwigia 73:271-292.

Zare, R., W. Gams, and H. C. Evans. 2001. A revision of Verticillium sect. Prostrata. V. The genus Pochonia, with notes on Rotiferophthora. Nova Hedwigia 73:5 1- 86. 216

Chapter 6

THE FUNGAL-ARTHROPOD SYMBIOSES OF THE HYPOCREALES (FUNGI) REPRESENT A CRETACEOUS RADIATION EVENT

Gi-Ho Sung', George 0. Poinar,Jr.2and Joseph W. Spatafora'

Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR

9733 1-2902, USA

2Depment of Zoology, Oregon State University, Corvallis, OR 9733 1-2902, USA

For submission to Molecular Phylogenetics and Evolution 217

Abstract

The Hypocreales is an ecologically diverse group of fungi that includes decomposers of plant material and obligate symbionts of animals, plants and other fungi. Because the ancestor of the Hypocreales can not predate the ancestor of the animal, plant, and fungal kingdoms, the broad host range in the hypocrealean fungi, especially that of the obligate symbionts, is not a product of strict cospeciation and involves some degree of interkingdom host jumping. To better understand the evolution of symbioses of the hypocrealean fungi, we reconstructed the ancestral states of host affiliation and conducted phylogenetic dating analyses based on multi- gene phylogenies. The ancestral state reconstruction supports a total of 9-14 interkingdom host shifts that are concentrated in three families characterized by a high occurrence of entomopathogenic species (Cordycipitaceae, Clavicipitaceae and

Ophiocordycipitaceae), including 5-7 onto fungi, 2-4 onto animals, and 2-3 onto plants. Although these results indicate a dynamic process of interkingdom host jumping, the ancestral ecology of Clavicipitaceae and Ophiocordycipitaceae is characterized as a pathogen of animals, with pathogenicity of arthropods as a dominant theme in the evolutionary ecology of all three families. Phylogenetic dating, coupled with the ancestral host reconstruction, support the Jurrasic origin of hypocrealean fungal-animal symbioses and demonstrate that all three entomopathogenic families originated within a short, overlapping geological timeframe. Furthermore, our results support the Cretaceous radiations of the fungal- 218 arthropod symbioses in these three families and provide the hypothesis that the current diversity in fungal-arthropod symbioses in the hypocrealean fungi correlates with the diversification of interactions between arthropods and angiosperms in the Cretaceous. 219

Introduction

The Hypocreales is one of the largest and most taxonomically diverse monophyletic groups in Sordariomycetes, phylum Ascomycota (Eriksson, 2005;

Spatafora et al., 1998). It consists of more than seven families with at least 130 genera and 850 species (Chapter 5; Eriksson, 2005; Rossman et al., 1999). The distribution of hypocrealean fungi is cosmopolitan with representatives in virtually all terrestrial ecosystems and they occupy a multitude of ecologically diverse niches in which they function as biotrophic symbionts or saprobes (Rogerson, 1970; Rossman et al., 1999).

The diversity of symbioses that hypocrealean fungi form with other organisms is unique among fungi with interaction ranging from antagonistic to mutualistic and involving a broad range of eukaryotes, including animals, plants and other fungi

(Rossman et al., 1999). As symbionts of animals, they primarily parasitize arthropods, or more rarely, they exist as mutualistic endosymbionts of beetles (Kobayasi, 1982;

Suh et al., 2001). Hypocrealean plant symbionts cause diseases of various groups of plants (e.g., Fusarium on wheat) and occur as epiphytes (e.g., Claviceps on rye) and endophytes (e.g., Neotyphodium on tall fescue) of grasses in interactions that have been characterized as both antagonistic and beneficial (Clay and Schardl, 2002;

Rossman et al., 1999). As symbiotic partners of other fungi, they function as mycoparasites that infect various groups of fungi including mushrooms, bracket fungi, and truffles (Currie et al., 2003; Rossman etal., 1999, Mains 1957). 220

Based on these diverse symbioses, understanding the evolution of these interactions is central to understanding the evolution of the Hypocreales. Interest in the evolution of fungal symbioses has included numerous studies of the Hypocreales with recent findings resulting in a greater appreciation of the complexity of evolutionary patterns of host affiliation (Chapter 5; Castlebury et al., 2004; Currie et al., 2003). Spatafora et al. (submitted) demonstrated that the phylogenetic distribution of host affiliation is a product of a dynamic process of inter-kingdom host shifts. One of dramatic examples in a series of interkingdom host shifts is in the case of grass symbionts (e.g., Claviceps and Epichloe), which originated from an ancestor of animal pathogens.

The fungal-animal symbioses of hypocrealean fungi are the best-exemplified by the fungal-arthropod interactions of species formerly classified in the genus

Cordyceps (Kobayasi, 1982; Mains, 1958, Chapter 5). The arthropods pathogens of

Cordyceps sensu Kobayasi and Mains parasitize eight orders of arthropods with the most common and species rich associations with insects of orders Coleoptera,

Hemiptera and Lepidoptera (Kobayasi, 1941; Kobayasi, 1982). Based on a series of multi-gene phylogenetic analyses, species of the genus Cordyceps were recently reclassified into several genera (Cordyceps s.s, Elaphocordyceps, Metacordyceps and

Ophiocordyceps) in three families (Clavicipitaceae, Cordycipitaceae and

Ophiocordycipitaceae) (Chapter 5). These previous findings emphasize that fungal- arthropod symbioses are representative of three major lineages of the hypocrealean 221 fungi and are one of the dominant forms of symbioses with a potentially long history in the evolution of the Hypocreales.

h addition to the molecular phylogenetic studies, the evolutionary history of fungal-arthropod symbioses in the hypocrealean fungi is also documented in the fossil record (Poinar et al., submitted; Poinar and Thomas, 1984). Of particular importance is the recent fossil record of a male parasitized by the asexual stage of the genus Ophiocordyceps (Genus Hymenostilbe; Family Ophiocordycipitaceae) (Poinar et al., submitted). The Hymenostilbe fossil is described from Burmese amber and is estimated to be from the Early Cretaceous (-100 MYA), consistent with the hypothesis that the diversification of the pathogens of arthropods is an ancient ecology among the hypocrealean fungi. This discovery is not only the earliest fossil record of insect parasitism by fungi, but also provides an opportunity to estimate the divergence times of the major lineages in the hypocrealean fungi.

Based on the broad host range of its extant ecologies, we propose that the diversification of the hypocrealean fungi is correlated with the diversification of its hosts. The ability to estimate geological dates for the origins of the major lineages of the Hypocreales provides testable hypotheses, which can be compared to known evolutionary events from the fossil records of their symbiotic partners (Farrell, 1998;

Percy et al., 2004). To correlate estimated divergence times with their major interkingdom host shifts, we conducted a series of analyses that included phylogenetic dating and ancestral state reconstruction of host affiliation for the Hypocreales. The main objectives of this study are 1) to estimate the divergence times of the major 222 lineages of the Hypocreales, 2) to further develop hypotheses for ancestral character state reconstructions of host affiliation in the Hypocreales, and 3) to develop hypotheses for the geologic origin and the diversification of fungal-arthropod symbioses for the Hypocreales.

Materials and methods

Fossil records for calibration points

Two reference fossils (Hymen ostilbe and Paleopyrenomycites) were used to calibrate rates of nucleotide substitution for the Hypocreales and the Sordariomycetes.

Based on the estimation of the palynomorphs obtained from the amber beds where the fossil piece originated, the recently discovered Hymen ostilbe fossil from Burmese amber is -1 00 million years old (Poinar et al, submitted; Cruickshank and Ko, 2003).

Morphological characterization of the fossil resulted in its identification as a

Hymenostilbe asexual state linked to Op/iiocordyceps of the Hypocreales. The genus

Hymenostilbe is an asexual entomopathogenic genus that includes species which are primarily pathogens of wasps, ants and scale insects. It is characterized by the synapomorphies of clavate to cylindrical conidiogenous cells arranged in more or less conspicuously dense layers or palisades on relatively robust synnemata (Hywel-Jones,

1995a; Hywel-Jones, 1995b; Samson et al., 1988). Species of Hymenostilbe are restricted to a single terminal dade of the family Ophiocordycipitaceae and inferred to be closely related to and possibly derived from the asexual genus Hirsutella, which is 223 the predominant asexual genus linked to Ophiocordyceps (Chapter4; Chapter 5;

Hodge, 2003). The Hymenostilbe fossil, therefore, was used as a crown calibration point for Ophiocordyceps.

In addition to the Hyinenostilbe fossil, a 400 million-year-old perithecial ascomycetes fossil (Taylor et al., 1999) from Rhynie Chert was used as a basal calibration point for the Sordariomycetes. The genus Paleopyrenomycites was recently proposed based on this Sordariomycetes fossil with the descriptions of unitunicate asci and perithecial ascomata (Taylor et al., 2005). Because the ordinal placement of this fossil is not clearly defined by its morphological characters (Taylor et al., 2005), we use it as minimum age estimate for the origin of the Sordariomycetes, rather than for a more terminal dade of the Sordariomycetes.

Taxon sampling, constructionofdata sets for phylogeiietic dating

To estimate the divergence times of major lineages in the hypocrealean fungi, the representative taxa were chosen to include the members of all major families in

Hypocreales based on previous phylogenetic studies (Chapter 5; Castlebury et al.,

2004). Seventy-four taxa were selected to construct a Hypocreales data set, including the outgroup taxa, Glornerella cingulata and Verticillium dahliae in Phyllachorales

(Table 6.1). We used DNA sequences from five genes: nuclear ribosomal small and large subunit DNA (nrSSU and nrLSU), elongation factor 1 a (EF-la), the largest and second largest subunits of RNA polyrnerase II (RPBJ and RPB2). Most of the DNA sequences used in this study were obtained from the previous phylogenetic 224 (AFTOL)ChapteravailableFigure 6.1. 4 in areand The GenBank. identified 5 listare ofdenoted culturesAn with asterisk withAFTOL. and C4 (*)specimens and denotes C5, respectively. usedthe sequences in this study. Sequences determined GenBank obtained by accession the fromfirst authorthenumbers database of this ofstudy. Assembling Sequences Fungal determined Tree of inLife are provided for the sequences Hypocreales Taxon Specimen voucher a Host/substratum nrSSU nrLSU GeneBank accession number EF-la RPBJ RPB2 BionectriaBalansiaAschersoniaClaviceps pilulaeformis pityrodes badiapaspali ATCCAEG 97-2 13892208842 PoaceaeonScale bark insect (Hemiptera) AYAF543764 489696 C4 AY489728AF543788 C4 AY 489623 C4 AY 489658 C4C4 C4 * CordycepsClaviceps purpurea militarisgunniicardinalis OSCGAM 936237640493609 12885 mothLarvaDactyl/s (Lepidoptera) (Lepidoptera) glomerata (Poaceae) AY184977AY184973AF339572AF543765 AY184966AF339522AY184962AF543789 AY489616AF543778 C4 AY489650AY489648 C4C4 AY545731 C4 ElapliocordycepsCos,nosporaCordyceps tubescarabaeicola coccinea capita/a rculata CBSOSCEFCC 114050 71233111002 5689 Elaphomycesmono/usMothScarab (Lepidoptera) (Coleoptera) nodulosus sp. (Euascomycetes) (Hymenomycetes) AY489689AY489702AF339574 C4 AY489721AY489734AF339524 C4 AY489615AY489629 C4 AY489649AY489667 C4 C4 EngyodontiumElaphocordycepsElaphocordycepsjaponicaElaphocordycepsfracia aranearum ophioglossoides CBSOSC 110991309.85106405110990 SpiderElaphomycesElaphoinycesElapliomyces (Arachnids) sp. (Euascomycetes) AY489691AF339576 C4 AF339526AY489723 C4 AY489618 C4 AY489652 C4 C4 HydropisphaeraHaptocilliumEpichloë typhina zeosporumbalanoides erubescens ATCCCBS 250.82335.80 3609356429 PlantNematodaFestuca rubra (Poaceae) AF339589AF339588U32405 AF339540AF339539 U17396 AF543777 C4 AY489653 C4 C4 HypocreaHydropisphaera lutea peziza ATCCCBS 102038 208838 on barkdecorticated conifer wood AY489698AY545722AF543768 AF543791AY489730AY545726 AY489625AF543781 C4 AY489662AY489661 C4 AY545732 C4 * IsariaHypocrellaHypocrea tenuipes rufa sp.schizostachyi GJS 89104 OSCGJSNHJCBS 8910412605114374111007 LarvaScaleon bark insect(Lepidoptera) (Hemiptera) AY489694 C4 AY489726 C4 AY489621 C4 AY489656 C4 C4 225 Table 6.1. (Continued) Lecanicillium anti/lan urn CBS 350.85 Agaric mushroom (1-lymenomycetes) AF339585 AF339536 C4 C4 C4 MelanopsammapomiformisMariannaeaLeuconectria pruinosa clusiae ATCCARSEF 1887322228 5413 UlmusPupaSoil (Lepidoptera) sp. (Ulmaceae) AY489677AY1AY489700 84979 AY489709AYIAY489732 84968 AY489604AY489627 C4 AY489570AY489664 C4 C4 * MetarhiziumfiavovirideMetarhiziumMetacordyceps album (auchlamydosporia var. minus ARSEFCBS 101244 203720825714 HemipteramolluscLarva eggs (Coleoptera) (Diplopoda) AF543763 C4 AF543787 C4 AF543775 C4 C4 C4 MyriogenosporaMyrothecium cinctumatramentosa ATCCAEG 96-32 22270 SoilAndropogonNi/aparvata lugens virginicus (Homoptera) (Poaceae) AYAY489701AF339580 489678 AY489710AY489733AF339531 AYAY489628 489605 C4 AYAY48966S 489638 C4 C4 * NectriaMyrotheciumMyrotheciurn cjnnabarjna roridurnleucotrichumverrucaria CBSATCCAR 114055 9095162973506 BetulafromSoildecaying baled sp. grass (Betulacaceae) cotton leaf AY 489681489676489675 U32412 AY 489708489713489707 U00748 AYAF543785 489603489608489602 AYAY489666 489641489636489635 C4 * OphiocordycepsOchroleuca calami aphoduibrunneapunciagriota ala NHJARSEFCBS 12565125.87 54985692 ColeopteraAphodiusLarvaon palm (Coleoptera) hewitti (Coleoptera) AY 489717 C4 AY 489717 C4 AY 489612 C4 AY 489644 C4 C4 Ophiocordyceps entomorrhizamelolonthaeheteropodacf. acicu/aris OSCKEWNHJ 12592110993 53484 PhyllophagaNymphLarvaColeoptera (Coleoptera) of cicada sp. (Coleoptera) (Hemiptera) C4CSC4 C4CS C4CS C4CS C5CSC4 Ophiocordyceps sp.sinensisravenelii NHJ 1258112522 NHJEFCCOSC 1258112522110995 7287 ColeopteraPhyllophagaLarva (Lepidoptera) sp. (Coleoptera) CSCSC4 CSC4 CSC4 CSC4 CSC4 Ophiocordyceps variabi/isunilateralisstylophora OSCNHJOSC 111003 12523110999 DipteraAntColeopteraLarva (Flymenoptera) (Coleoptera) CSC4 CSC4 CSC4 C4CS CSC4C5 PseudonectrjaPeethambaraOphionectria spirostriata rousseliana trichospora CBS 114049110115109876 BuxusCrinipellisperniciosaliana sempervirens (Buxaceae) (Hymenomycetes) AY 489692AF543767AF543766 AYAF543790 489724 U17416 AY 489619AF543780AF543779 AYAY489670AY489669 489654 CS C4 * Table 6.1. (Continued) 226 SimpliciliumRoumegueriellaRotferophthora lanosoniveum angustisporarufula CBS 704.86346.85101437 GloboderaRotifer (Rotifera) rostochiensis (Nematoda) AF339584 C4 AF339535 C4 AF543776 C4 C4 C4 SphaerostilbellaSimplicillium lamellicola berkeleyana CBS 102308116.25 AgaricusHemileiaPolypore bisporusvastatrix (Hymenomycetes) (Hymenomycetes)(Urediniomycetes) AF543770AF339601AF339602 AF339552AF339553U00756 AF543783 C4 C4 C4 * Stachybotrys subsimplexechinatachartarum ATCCUAMH 3288866238 6594 watersoilSoil hyacinth from desert sand AYAY489679 489704489680 AY 489711489736489712 AYAY 489631489607 489606 AY 489639489672489640 * StephanonectriaTorrubiellaStilbocrea macrostoma confragosaralticaudata keithii CBSARSEFCBS 114057 101247114375 1915 Spiderplanton Coccusbark (Arachnida) viridis (Hemiptera) AY 489695AF339604AY489693 C4 AYAY489725 489727AF339555 C4 AYAY489620 489622 C4 AY 489657AY489655 C4 C4 * Taxa in other orders of SordariomycetesCamaropsViridisporaVerticillium sp. diparietisporaincurvum CBS 102797460.88 CrataegusGanoderma crus-galli lipsiense (Rosaceae) (Hymenomycetes) AY489703AF339600 AFTOL AY489735AF339551AFTOL AY489630 AFTOL C4 AY489668 AFTOL C4 AFTOL CSC4 GnomoniagnomonGlomerellaDiaportheCirrenalia eres cingulata macrocephala CBSCBSIJK 51 199.53109767 1405488B AF543762AFTOL AF543786 AFTOL AF543773AFTOL AY489659AFTOLAFTOL AFTOL C4 MicroascusLuiworthiaLindraLanspora thoJassiae coronata grandisporatrigonosporus CBSJK 4686.15090A5839A 218.31 AFTOL AFTOL AFTOL AFTOLAFTOL AFTOL ThyridiumPetriellaPapulosaOphiostomapi4ferum set amerospora sp. fera ARCBSJK 5547F3872 437.75158.74 AFTOL AFTOLAFTOL AFTOL AFTOL AFTOL Dothidiomyctes XylariaCapnodiumVergjcjlljum sp. sp. dahliae CBSOSCATCC 147.52100004 16535 AY489705 AFTOL AY489737AFTOL AY489632 AFTOL AY489673 AFTOL AFTOL C4 Table 6.1. (Continued) 227 aAEG, A. E. Glenn personal collection; AR, A. Y. RossmanDendryphiellaarenariaCochliobolus personal heterostrophus collection; CBS 181.58134.39 A.RSEF, USDA-ARS Collection of Entompathogenic Fungal cultures, Ithaca, NY; AFTOL AFTOL AFTOL AFTOL AFTOL AlbertaGarden,SamuelsFungalATCC, CultureAmerican Microfungi KEW,personal Collection,Surrey, collection;Type Collection UK; Culture Chuncheon, MICH, NHJ, and Collections, Herbarium,Nigel University Korea; Hywel-Jones Manassas, GA.M,Edmonton,of Michigan Julianpersonal VA; AB. Herbarium,CBS,H. Millercollection; Centraalbureau Mycological Ann Harbor, Herbarium Ml; JK, J. Kohlmeyer personal collection; KEW, mycology collection of Royalvoor Botanical Schimmelcultures, Utrecht, the Netherlands; EFCC, EntomopathogenicOregon State University Herbarium, Corvallis, OR; UAMH,Athens, University GA; FAU, of F. A. Uecker personal collection, GJS, G. J. 228 studies of the Hypocreales (Chapter 5; Castlebury et al., 2004). Fourteen RPB2 sequences were generated for selected taxa from the study of Castlebury et al. (2004) as previously described in Chapter 4.

We also constructed a Sordariomycetes data set to estimate the divergence times of major lineages in Class Sordariomycetes based on the Hymenostilbe and

Paleopyrenomycites fossils. We screened the available DNA sequences in AFTOL

(Assembling Fungal Tree Of Life) database to select the representatives of major orders that are complete for five genes (http://ocid.nacse.org!researchlaftol/data.php/).

We selected 15 taxa from the AFTOL database and included three outgroup taxa in

Dothideomycetes (Cochliobolus heterostrophus, sp., Dendryphiella arenaria) based on the results of previous studies (e.g., Lutzoni et al., 2004). A total of

60 taxa was used in constructing the Sordariomycetes data set including 42 taxa in

Hypocreales. In taxon sampling, we minimized the inclusion of taxa with missing data in an attempt to reduce the error associated with inference of topologies and estimation of branch lengths. In both the Hypocreales and Sordariomycetes datasets, a total of 90 unique taxa were used, of which only six taxa were not complete for all five-gene partitions (Table 6.1).

Taxon sampling, constructionofdata sets for ancestral state reconstructionofhost affiliations

We used the previously described Hypocreales data set in ancestral state reconstruction of host affiliations. In addition, the 162-taxon data set in Chapter 5 was 229 also used to maximize taxon sampling and to insure a more complete representation of the ecological diversity of the Hypocreales. Maximum likelihood host reconstruction was performed using the computer program Multistate (Page!, 1999; Page! et al.,

2004) as described be!ow. Due to the inabi!ity of the program to incorporate missing data for character state values, we excluded the taxa that did not have host information resulting in a 147-taxon data set (Table 5.1).

Sequence alignment and phylogenetic analyses

The DNA sequences were aligned with Clustal W 1 .64b (Thompson et al.,

1994) using default settings, then optimized in MacClade 4.0 (Maddison and

Maddison, 2000) by direct examination. The ambiguously aligned regions were excluded in the phylogenetic analyses. Weighted parsimony (WP) analyses were performed using PAUIP* 4.OblO (Swofford, 2002), incorporating different rates of substitutions for the eleven partitions ofnrSSU, nrLSUand each codon of the three protein-coding genes. Step matrices for each partition, based on the in of the pairwise base comparisons (Felsenstein, 1981; Wheeler, 1990), were constructed using the computer software STMatrix 2.1 (written by S. Zoller and available at http :I/www. lutzonilab.net/pages/download.shtml). Weighted parsimony analyses were conducted on parsimony informative characters only with the following search options: 100 replicates of random sequence addition, TBR branch swapping, and

MulTrees in effect. Nodal support was estimated using 1000 replicates of 230 nonparametric bootstrapping with the same heuristic search options has the WP analyses (Felsenstein, 1985).

Bayesian analyses were performed using MrBayes version 3.0b4 (Huelsenbeck and Ronquist, 2001; Huelsenbeck et al., 2001) specifying the appropriate mixed models for each of the eleven partitions. The appropriate mixed models for the heterogeneous data were selected by the likelihood ratio test (LRT) and Akaike

Information Criteria (AIC) in ModelTest 3.06 (Posada and Crandall, 1998) for both the Hypocreales and Sordariomycetes data sets. The general time-reversible model with invariant sites and gamma distribution (GTR+I+F) was determined to be the most appropriate model and was employed separately for each partition. To examine the convergence of log-likelihood, we performed five independent Bayesian MCMCMC analyses with random starting trees and 1,000,000 generations sampled every10thtree.

The variation in log-likelihoods of the sampled trees was investigated by observing the

MrBayes log file and confirmed the convergence of all five 1,000,000 generation analyses. [A 10,000,000 generation analysis was also performed and confirmed the results from the 1,000,000 generation analyses (data not shown).] The initial 10,000 trees were identified as the burnin and the remaining 90,000 trees from each of five

1,000,000 generation analyses were used to construct 50% majority-rule consensus trees and to calculate associated posterior probabilities. In interpreting the phylogenetic confidence, we consider nodes strongly supported when they receive both posterior probability (PP) of? 95% and bootstrap proportion (BP)? 70%

(Lutzoni et al., 2004; Reeb et al., 2004). 231

Divergence time estimation

In estimating the divergence times of early evolutionary events from molecular phylogenies, several methods have been recently developed to incorporate rate variation among lineages, including local clock methods (Yang and Yoder, 2003),

Bayesian methods (Thorne and Kishino, 2002), and nonparametric rate smoothing methods (Sanderson, 1997; Sanderson, 2002). Here, we estimate divergence times by using the smoothing methods, nonparametric rate smoothing (NPRS) and penalized likelihood (PL), based on the distances inferred from the consensus trees of the

Bayesian analyses.

Both nonparametric rate smoothing (NPRS) and penalized likelihood (PL) methods estimate the evolutionary rates for every branch based on its distance

(inferred number of substitutions). Rapid rate changes are penalized to prevent excess rate variation across the tree by establishing an autocorrelation of neighboring rates

(Sanderson, 1997; Sanderson, 2002). The NPRS method penalizes the rapid changes of evolutionary rates by estimating the rates and divergence times based on a least- squares smoothing criterion (Sanderson, 1997). The PL method is a semiparametric approach that combines a model-based likelihood method with a roughness penalty for which the degree is determined by a smoothing factor in a cross-validation procedure

(Sanderson, 2002).

We estimated the divergence times by using both NPRS and PL methods as implemented in the program "r8s" (Sanderson, 2002; Sanderson, 2003) basedon the 232 consensus trees (Figures 6.1 and 6.2) from Bayesian analyses of the Hypocreales and

Sordariomycetes data sets. To estimate the divergence times using PL method, cross validation procedures were performed to choose the most optimal rate smoothing factors. The smoothing factors 312 and 100 were chosen by cross validation procedures for the Bayesian consensus trees from the Hypocreales and

Sordariomycetes data set, respectively. To assess the phylogenetic uncertainty from the errors in topology and estimation of branch length from substitutional noise, we also estimated the divergence times for 100 randomly sampled Bayesian trees from

Sordariomycetes and Hypocreales data sets using the NPRS method. The fossil calibration points were applied to the corresponding nodes using "fixage" command in r8s program. Based on the Hymenostilbe fossil record, an age of 100 MYA was used for the crown node of genus Ophiocordyceps (Figure 6.2) as a conservative calibration point in the Bayesian trees from both the Hypocreales and Sordariomycetes data sets.

Due to the uncertainty in calibration point for the Paleopyrenomycites fossil, We conducted three independent dating analyses in both NPRS and PL methods to both verify the use of Paleopyrenomycites for the crown node of Sordariomycetes and to test for internal dating consistency of the Hymenostilbe and Paleopyrenomycites fossils. First, we estimated the divergence times by dating with a single calibration point (either Hymenostilbe or Paleopyrenomycites calibration point) to compare the age estimates. Second, two calibration points were simultaneously used to compare the estimates of divergence times with those from single calibration points. 233

Ancestral state reconstruction of host affiliations

To understand the evolutionary patterns of host affiliations among the hypocrealean fungi, we reconstructed the ancestral host affiliations by using both the

Hypocreales and the 147-taxon data sets. Based on host information (Table 5.1 and

6.1), we coded each taxon based on host kingdom as animal, fungi or plant. We used a continuous time Markov model of trait evolution, as implemented in the Multistate program (Pagel, 1999; Pagel et al., 2004), allowing independent host character transitions in each node of the phylogenetic tree. As previously noted, the taxa that caimot be coded in the kingdom level were excluded due to the inability to accommodate missing data in Multistate program (Page!, 1999; Page! et al., 2004).

The Bayesian consensus trees from both Hypocreales and 147-taxon data sets were transformed into a "pag" file, which contained information on the topology and branch lengths, using the program Phy!ip to Pag (available at http://www.rubic.rdg.ac.uk!meade/Mark!). The "pag" file was then used to estimate the likelihood and instantaneous rates of transitions among host character states. A one-parameter model was selected over a six-parameter mode! in a likelihood ratio test

(Huelsenbeck and Rannala, 1997), thus allowing uniform rates of transitions among host character states. In reconstructing the ancestral host states, the significance of one state over another was determined by the likelihood ratio test (Huelsenbeck and

Rannala, 1997).

Results and discussion 234

Phylogenetic analyses and the relationshipsofthe hypocrealean fungi

The alignment of the Hypocreales data set contained 5,014 unambiguously aligned positions; 4,257 sites were variable, 1,495 were parsimony informative. The alignment of 147-taxon data set contained 4,926 unambiguously aligned positions;

4,665 sites were variable, 1,832 were parsimony informative. The alignment of the

Sordariomycetes data set contained 5,110 positions unambiguously aligned positions;

3,729 were variable, 1,508 were parsimony informative.

The weighted parsimony (WP) analyses from the Hypocreales data set resulted in one tree of 23,972.54 steps. Because the Bayesian analyses inferred an almost identical topology, the Bayesian consensus tree is shown in Figure 6.1 with the bootstrap proportions (BP) of the WP analyses as well as the Bayesian posterior probabilities (PP). The overall topology for the higher relationships from the

Hypocreales data set is nearly congruent with the results from the previous phylogenetic analyses for the statistically well-supported relationships (Chapter 4;

Chapter 5; Castlebury et al., 2004). As shown in the study of Castlebury et al. (2004), theStachybotrysdade is the basal lineage of the Hypocreales and all other lineages form a monophyletic group with strong statistical support (BP = 96%, PP = 100% in

Figure 6.1). The basal placement of the Nectriaceae with respect to the Bionectriaceae differs from the previous phylogenetic analyses, which resolved the basal placement of the Bionectriaceae although the relationships are not strongly supported (Figure

6.1). The sister-group relationships of the Clavicipitaceae and Ophiocordycipitaceae 235 and the Hypocreaceae and Cordycipitaceae were resolved as in previous studies, but received lower BP values.

The WP analyses of the 147-taxon data set resulted in one tree of 32,379.12 steps, which again is nearly topologically identical to the results from the Bayesian analyses. Therefore, the consensus tree from the Bayesian analyses is presented in

Figure 6.3 with the posterior probabilities (PP) and the bootstrap proportions (BP) of

WP analyses. The results from these analyses are similar to those of the previous analyses (Chapter 4 and 5) and strongly support the basal placement of the

Bionectriaceae relative to the Nectriaceae (Figure 6.3). In contrast to the analyses of the Hypocreales data set, the sister group relationships among families that were are strongly supported in the analyses of the 147-taxon data set, i.e., Clavicipitaceae +

Ophiocordycipitaceae: BP = 76%, PP = 100%, Cordycipitaceae + Hypocreaceae: BP =

72%, PP = 98% in Figure 6.3.

The diminished support for interfamilial relationships in the Hypocreales data set, as compared to the analyses of 147-taxon dataset, is best explained by the exclusion of a significant portion of the alignment in theRPBJgene partition.

Alignment of theRPB1sequences from the taxa in theStachybotrysdade in the

Hypocreales data set was problematic and a total of 427RPBJaligned positions were excluded out of a total of 865 positions. This is in comparison to the analyses with thel47-taxon data set that excluded only 136 aligned positions. As discussed in

Chapter 4,RPBJgene partition provides considerable phylogenetic signal for resolving the sister-group relationships of the 236

Glomerella cingulata Verticillium dahllae Melanopsamma pomiformls I Sfachybolrys chartarum I Sfachybotrys echinata I Stachybobys subslmplexI Peel hambara splrostrlata Stachybotrys Clade ,wy,UI,IVUIUnIVUr,IJCd,Id I leucotrichum I Myrotheclum cinctum I Myrotheclum roridum I Nectrla clnnabarina Cosmospora cocclnea I Ophlonectria trlchospora I Pseudonectrla roussellana I Nectriaceae Vlrldlspora dlparletlsporaI L.euconectrla clusiae I Stllbocrea macrostoma Hypocreales Blonectrla pityrodes rL_.1 Sleohanonecirla kelthil I cfrl; arni I Bionectriaceae Hydroplspha era pezlza I Hydroplsphaera erubescensI Roumeguierieila rufula I 96 VertIclIllum Incurvum Sphaerostllbella berkeleyana 00 100 Hypocrea rufa Hypocreaceae Hypocrea lutea 100 SImplicililum lamelllcola I Simpilcililum lanosoniveum I -j------Engydontium aranearum I I '- Lecanlcilllum antillanum I 1001 cardinalls I Torrublella ratticaudata taarta lenulpes I Cordycipitaceae 90 Cordyceps milifaris I Marlannaea pruinosa I 100[j Cordyceps scarabaeicoia I L.!t- Cordyceps tubercuiata I 100'- Torrublella con fragosa I Mefacordyceps chiamydosporla I 100 Rotiferophfhora angustlspora I Metarhizium album I 100 Met acordyceps fail I 100 Metarhlzlum flavovlride I Aschersonia badia I Hypocrellasp Clavicipitaceae 7 nypU0eflda0JIuaLdL.0Iyt 100 Balansia pllulaeformls 92 100 Myrlogenospora atramentosa 100 74 Epichloe typhina 100 CIa viceps paspali 100 Clavlceps purpurea 100 CordyCeps gunnii 100 Haptocllllum balanoldes 10 Haptoclliilum zeosporum Elaphocordyceps fracta 100 100 Elaphocordyceps Japonica 100 Elaphocordyceps capitata Elaphocordyceps ophloglossoldes Ophlocordyceps net eropoda Ophiocordyceps melolonthae Ophlocordyceps raven clii 91 Ophlocordyceps varlabllls Ophiocordyci pitaceae 100 Qphlocordyceps entomorrhlza 100 Ophlocordyceps aphodll 100 Ophiocordyceps brunneapuct ala Ophiocordyceps op. NHJ12581 Ophiocordyceps Ophlocordyceps op. NHJ 12522 Ophlocordyceps unllaferalls 100 Ophlocordyceps stylophora 94 100 Ophlocordyceps sinensis 100 Ophiocordyceps agrlota 0.05 substitutIons/site Ophlocordyceps cf acicuiaris

Figure 6.1. Phylogenetic relationships among the hypocrealean fungi (74 taxa, 5,014 aligned nucleotide positions) based combined evidence of five genes (i.e. nrSSU, nrLSU, EF-la, RPBJandRPB2).The topology presented is a 50% majority consensus tree of 90,000 trees, rooted using Glomerella cingulata. The internodes that are supported with both bootstrap proportions of WP analyses (BP)? 70% and posterior probabilities (PP)95% are considered strongly supported and in a thicker line. For the strongly supported relationships, bootstrap proportions (BP)? 70% and posterior probabilities (PP)? 95% are shown above and below the internodes, respectively. 237 families (Clavicipitaceae + Ophiocordycipitaceae, Cordycipitaceae + Hypocreaceae) and excluding Ca. 300 more aligned positions in the Hypocreales data set results in diminished support for the sister-group relationships (Figure 6.1 and 6.3).

The Bayesian analyses based on the Sordariomycetes data set also resulted in an almost identical topology to the results of the WP analyses (a single tree of

21,309.48 steps) and thus the consensus tree from the Bayesian analyses is presented in Figure 6.4. The overall topology is more similar to that of 147-taxon data than with the Hypocreales data set in that it resolves the basal placement of the Bionectnaceae with respect to the Nectriaceae. These analyses strongly support the monophyly of the

Hypocreales (BP = 97%, PP = 100% in Figure 6.4) but the higher relationships are not as well supported as in the Hypocreales and 147-taxon data sets (Figure 6.4). In the ordinal level relationships, the monophyly of the Hypocreales, Microascales and

Phyllachorales is strongly supported (BP = 99 %, PP = 100% in Figure 6.4), but most of the relationships among orders in the Sordariomycetes received BP values below

70% (Figure 6.4).

Phylogenetic dating and evolutionofhost affiliation in the hypocrealean fungi

Phlogenetic dating was conducted by using NIPRS and PL smoothing methods

(Sanderson, 1997; Sanderson, 2002) with the Bayesian consensus tree from the

Hypocreales data set. Although the tree topology differs from that of 147-taxon data set in the placement of Nectriaceae and Bionectriaceae, recent dating studies demonstrated that the relationships in the basal lineages do not affect the estimation of 238 the divergence times for the terminal relationships (Sanderson and Doyle, 2001; Soltis et al., 2002). Therefore, the topological uncertainty of the Nectriaceae and

Bionectriaceae is not considered significant in estimating the age estimates of the major lineages of the hypocrealean fungi. In constructing the chronogram in Figure

6.2, we only show the age estimates from NPRS method, as age estimates (Table 6.2) from the NRPS and PL methods are not appreciably different.

The Hypocreales.The divergence time estimates from Hypocreales data set suggest that the ancestor to the Hypocreales was present in the Triassic (NPRS: 236 ±

22 MYA, PL: 247 MYA, node 1, Figure 6.2; Table 6.2) and that the major familial lineages radiated in the Jurassic. Our results also support that most of the hypocrealean families diversified in the Cretaceous, consistent with the diversification of angiosperms (Figure 6.2). Based on fossil evidence, the period of significant angiosperm diversification is generally considered being between 115 and 90 MYA

(Lidgard and Crane, 1988; Wing, 2000) with the unequivocal earliest record for an angiosperm crown group being an 125 million years old water lily in Nymphaeceae

(Friis et al., 2001; Soltis and Soltis, 2004). The most diverse angiosperm groups, such as many extant monocot and eudicot lineages, radiated in this period and affected the tenestrial diversities of other organisms by changing the terrestrial ecosystems and initiating diverse interactions with other organisms (Farrell, 1998; Labandeira and

Sepkoski, 1993; Lidgard and Crane, 1988; Strong et al., 1984; Wing, 2000). 239

USe,.,pwn pom.foms Sfrhy&,t,y St.d,yboIrys echweta StschyboUys wtv,?pe. PnthaAtfl SSM Mymffieoirn, wuc.,a Stachybofrys Clade Mymit.oum Mocotnct.,fl MyrOTh.00r,,CflCffl My,othoo,,,,o,,dum dn,nnth OM,.dns _o,tnd,ospora Nectriaceae Lwcot,*ctha cse Hypoc Sio,,.th,. pydo St.pMnonecUp. ke,tfl,, Oth,00e Bionectriaceae H)p.ph.r..- pen

Sjt..,DsIilb.1t* beskemysna Hypocreaceae HypO,ee b.bee S.rU.cm /nico!8 Si,bclle,n be,oso,,,eo,fl

L.ce,,r4âum bMe Co*cep Tweibtebe feW.)dete bt.ee tweepe Cordycipitaceae Co.4weee bt,, Mweeee.. pewf we Co,ewepe bere Cw*c.ps Wberaeqa Te.mb.eNecOthegwe Moteconfrceps chiemydoSpoea RobMqophtho,se090sOwe,e Methwum elbom

Melethffeon, 7eeffefffe bebw M,owbee Clavicipitaceae Hypocrebe f,zostechy 8&enw.pdthnfwe.s My20g20weofe ewefwe EewMlo. typeeee cbtwepep.2 o.wee pwee COe*Weps Htptocilnjm beno4es

EieØ1ecowiycep b.ffe

EJeehobOedy010b obfhkwedef C.oco.cepshefempode t*beecoe5c.ps,,.2oSonThee Ophiocordycipitaceae ceweoepf eo,wef we Cmwwe.5w.pepboth Tbt Opbtocwe,wep, W- OWl 2581 Cbwoeebyepeep NW1252 Oocceceps u25ele,ele of25cepf fey

OØ,*,00fdfc.58 eg'folf

200 150 500 50 1.088 Tflassic Jarassic Coetaceous Terloary Mesozoic Cenozoic

Figure 6.2. Chronogram based on a Bayesian consensus tree in Figure 6.1. The divergence times were estimated using the nonparametric rate smoothing (NPRS) method. The calibration (crown node of genusOphiocordyceps)based on a 100 million-year-oldHymenostiThe fossil is marked with an asterisk and the approximate ages are indicated by the length of branches. Geological times are provided below the chronogram and the Cretaceous period is highlighted by the gray rectangular box. Numbers in the circles correspond to twenty selected nodes for estimating the divergence times (Table 62) and reconstructing the ancestral states of host affiliations. Ancestral states of host affiliations were reconstructed using maximum likelihood framework(Pagel,1999), as implemented in Multistate computer program (Pagel, 1999). The host affiliations were coded in the kingdom level. The significance of the ancestral states is determined by a likelihood ratio test (Huelsenbeck and Rannala, 1997) and the branches were colored for the internodes based on the results of the likelihood ratio test (Black = ambiguous, blue = fungal. green = plant and red = animal). The proportions in a pie chart refer to the proportions of the maximum likelihood associated with each character state for the ambiguous internodes on the ancestral host states. Yellow colored branches indicate the taxa that were not included in Multistate analyses due to the insufficient host information (Table 6.1). 240

Table 6.2. Inferred age estimates on the major hypocrealean divergences based on a Bayesian consensus tree (Figure 6.1) from Hypocreales data set. The node number corresponds to the selected internodes in estimating the divergence times in Figure 6.2. The age estimates were estimated using nonparametric rate smoothing (NPRS) and penalized likelihood (PL) methods. The standard deviations are based on the NPRS analyses of Bayesian trees.

Est. ages(Myr) Est. ages(Myr) Node ±S. D. Node ± s.D. NPRS PL NPRS PL

1 236.45 247.54 ±22.54 11 160.08 156.98 ±16.65 2 225.23 230.04 ±27.43 12 145.09 148.09 ±8.43 3 210.43 209.85 ±23.32 13 142.41 147.75 ±13.56 4 200.89 210.65 ±32.32 14 138.32 140.56 ±12.23 5 192.91 201.75 ±15.34 15 133.94 140.57 ±11.34 6 182.05 190.54 ±25.54 16 108.57 101.85 ±7.54 7 177.64 179.45 ±12.65 17 109.67 105.68 ±8.2 8 150.01 155.97 ±16.76 18 92.78 99.75 ±5.4 9 162.54 167.84 ±25.43 19 101.05 104.56 ±7.1 10 174.46 178.07 ±19.43 20 84.87 70.47 ±8.9

A Cretaceous radiation of the hypocrealean fungi (Figure 6.2) is therefore consistent with the diversification of Hypocreales co-occurring with that of the angiosperms and associated biota.

In maximum likelihood host reconstruction (Figures 6.2 and 6.3), the ancestral state of the hypocrealean fungi is inferred to be plant-based nutrition with a shift in nutritional mode to more derived animal and fungal-based nutrition. Species of the

Stachybotrys dade, Bionectriaceae and Nectriaceae are primarily saprobes of plants including angiosperms and gymnosperms and many species of the Nectriaceae also function as causal agents of plant diseases, especially in their asexual forms (e.g.,

Fusarium) (Rogerson, 1970; Rossman et al., 1999).

The shift from plant-based nutrition to animal/fungal-based nutrition occurs node 5, which is the common ancestor for the Hypocreaceae, Cordycipitaceae,

Ophiocordycipitaceae and Clavicipitaceae (Figures 6.2 and 6.3). The ancestral 241 character state reconstruction of node 5 is equivocally reconstructed as either animal, fungal or plant. Thus, the earliest fungal-animal symbiosis in the hypocrealean fungi is inferred to initiate at either crown node 5 or 7, for which age estimates date to the

Jurassic (NPRS: 192 ± 15 MYA, PL: 201 MYA, node 5; NPRS: 177 ± 12 MYA, PL:

179 MYA, node 7; Table 6.2). Overall, the evolution of the fungal-animal symbioses of the hypocrealean fungi is characterized by the diversification and cladogenesis resulting in three families (Clavicipitaceae, Cordycipitaceae and

Ophiocordycipitaceae), the majority of which are pathogens of arthropods (Figures 6.2 and 6.3). Our results of age estimates support the Cretaceous radiation of many lineages in these three families and their fungal-arthropod interactions (Figure 6.2;

Table 6.2). To date, the oldest known evidence for the presence of fungal-arthropod symbioses is the Cretaceous galleries (140 MYA) of bark beetles that resulted from mutualistic relationships between bark beetles and fungi (e.g., Ophiostoma) (Berbee and Taylor, 1993; Bright, 1993; Wood, 1982). The results presented here on phylogenetic patterns of host affiliation and age estimates are consistent with a fungal association of arthropods with hypocrealean fungi initiating at least in the Cretaceous and this provides evidence for the antiquity of fungal-arthropod symbioses.

The Cordycipitaceae. The crown age of the Cordycipitaceae is inferred to be

142 ± 13 (NPRS) or 147 (PL) MYA, or the Early Cretaceous (Figure 6.2; Table 6.2), and its ancestral ecology is equivocally reconstructed as animal or fungal 242

Blonecttla OChroteuca Hydropisphaera peziza 50 Hydropisphaera erubescens Bionectriaceae 100 Roumegueriella rufula -ens. lOO Roumegueriella rufula - Ness Nectrla clnnabarina - betssaceae Cosmospora coccinea -ymenomycotes P,.udonectria rousseliana - 'ssxaceae Nectriaceae Viridis pore diparietispofa- Rosaceae Ophlonectrla trichospora VerlJcill,um incurvum - Hymenomycetes Hypocrealutea Hypocreaceae Sphaerostilbella berkeloyana - Hymenomycetes polyporinus - Hymenomycetes Simpliclhium lamellicola- Hymenomycetes I 100 Slmplicilium lanosoniveum - Urediniomycetes 100 Simplicthum lanosonivaum - Urethniomycetes Torrubielle wallace! - Engyodontium aranearum- Arachnida Torrubiella ratticau data - Arachnida Lcanicillium antillanum- etes I-Lecanicillim aranearum - Cordycipitaceae Cordyceps cardinalts sdoptera 100 Cordyceps cardinalis - pidoptera L.canicillim fusisporum - I-lymenomycetes L.canicillim psalliotae - Hymenomycetes 100 lsaria farinosa - Lepidoptera lotbane faninosa - Lepidoptera Cordyceps tuberculela - Lepidoptera Torrubiella confragosa - - era - scale insect ISO Cordyceps scarabaeicola (Be.uvsña bassiana) - Coleoptera I Cordyceps staph ylinidicola (Beauverla bassiana) - Coleoptera Cordyceps Cf. takaomontana bsania Cf. faflnosa - Cordyceps ci. ochraceostromata --sidoptera 100 bsaria tenuipes - - 100 Cordyceps bifusispora- epuoptera lot Cordyceps bifusispora -epidoptera 100 Cordycepssp.EFCC2535-Coloptera I Cordyceps takaomontata - -pidoptera I Mlcrohilum oncoperae- sptera ioo Cordyceps kyushuensis - Leprdoptera 1 Cordyceps miitanis -- yrdoptera Verticillium op. CBS 102184- Arachnida I Cordyceps op.EFCC5693- Lepidoptera SO Cordyceps sp.EFCC5197 - Lepidoptera 150 Maniannaea pruinosa - Lepidoptera Cord yceps op.NHJ10627 - Lepidopters 100 -s Cordyceps op.NHJ10684- Lepidoptes Pochonla bulbilbosa - Psnaceae . Pochonia gonioides - so atoda Tolypocladium parasiticum - Rotifera Pechonia rubescens - Metacordyceps chlamydosporia - )splopoda Rofiferophtlrora angustispora - Rotifera Metacordyceps liangshar'enais - Lepidoptera Metacordyceps Ian gshanensls - Lepidoptera Metacordyceps sp.EFCC 2131 -Lepidoptera 00 Meiacordyceps op EFCC2I3S-Lepidoptera -- Nomuraea riley! - oidoptera avicipu ceae Metarhizium album - -emiptora Metacordyceps sp.NHJ 12118-Lepidoptera t0 Metarhizium fbavoreiride vat, minus - Hemiptera L 100Metacordyceps tail - lopb ,OO Metacordyceps op. OSC number -epidoptera Metarhizium anisopliae oar. majus - Coleoptera lOG Shimizuomyces paradoxus - Smilacaceae isi' Shlmizuomyces paradoxus - Smilacaceae Ascloersonia badia -elltipterascale nsect 8) Hypocrella schlzostachyl - Hemiptera - scale insect _ 90 Aschersonia placenta - -lemiptera - scale insect Ito Hypocrelle sp.GJS89104- Hemiptera scale insect Epichloe typhina - 100 Clariceps paspali -- -IS lOt Claviceps purpurea - Poacoae 100 Clariceps purpurea - Poaceae VerticlIlium epiphytum- Urediniomycetes Thi.-.Verticillium epiphytum - Urediniomycetes lot Myriogenospora .frwm.080.a - Poaceae 100 Balansia pilulaeformis - Poaceae

00 Balansia epichloe - Poaceae too Balansia hertningsiana - Poaceae 243

Figure 6.3. (continued)

Nomuraea atypacola - chnida 00 Paecilomyces lilacinus -"matoda jotPaecalomyces lilacinus - atoda Paecilomyces cinnamomeus - Hemiptera - scale Insect Cordyceps gunnn - Haptocillium zeosporum - Nematoda t,00 Haptocillium balanoides - 3toda 100 Haptocillium sinense- Nernatoda Cordyceps capitata - Elaphomyces Cordyceps Ion gisegmentis - Elaphomyces Cordyceps fractaElaphomyces - Cordyceps japonica - Elaphomyces Cordyceps paradoxa - flora - cicada Cordyceps inegoonsis -emiptera - cicada Cordycepsjezoensis- nip'er cicada - Cordyceps ophioglossoides- Elaphomyces h Cordyceps subsessalis - Coleoptera p iocor ycipi ceae 100 Ophiocordyceps variabilis - Diptera lot Ophiocordyceps variabilis- Diptera Ophiocordyceps melolonffiae Coleoptera 100 Ophiocordyceps gracilis - Lepidoptera 100 Ophiocordyceps gracilis - Lepidoptera tOO Ophiocordyceps heferopoda - Hemiptera - cicada 100 Ophiocordyceps heteropoda -'onhiptera - cicada 000 Ophiocordyceps konnoana "feoptera 100 Ophiocordyceps konnoaroa leoptera 0 00 -' Ophaocordyceps superficaalis Coteoptera 000 Ophiocordyceps nigrella-:pidoptera 000 Ophiocordyceps ravenelli- Coleoptera Opha'ocordyceps entomorrhiza - eoptera Ophaocordyceps aphodii - Coleoptera 100 100 Ophiocordyceps brunneapunctata - Coleoptera V 100 Ophaocordyceps sobolaf era Ophiocordyceps i Ophiocordyceps Ion gissama da 100-" Ophiocordyceps yakushimensis pte -oda 100 Ophiocordyceps bicentri - Hvs000ptere Ophiocordyceps sphecocephala - Flymenoptera . Hemiptera iooj Hymenostilbe aurantiaca - lymenoptera o! Ophiocordyceps irangeensss - Hymenopter 100 Ophiocordycopsirangiensis - Hymenopter Ophiocordyceps elongata- era 00 Ophiocordyceps sp. NI-1J12581 - Coleoptera '° Ophiocordyceps sp. NHJ12582 - Coleoptera 100 Ophiocordyceps sp. NHJ 12522 - Cole 100 Ophiocordyceps sp. NHJ 12529 - Ophiocordyceps unilateral,s - iymenoptera 100 Ophiocordyceps sp. OSC 3ymenoptera Ophiocordyceps agriota - Coleoptera Opiiiocordyceps cocc,diacola- Nemiptera - scale insect , Ophiocordyceps cf acicularis - Coleoptera 000 Ophaocordyceps cochl,diicoia- Lepidoptera 100 oo Harsutella up NHJ 12525- Coleoptera 000 72 Hirsu(ella sp NHJ 52527- Coleoptera 000 - lop(era Ophiocordyceps rObertsii - Lepidoptera Poecilomyces cicadae - tera - cicada 100 O Ophiocordyceps stylophora - -leoptera 100' Ophiocordyceps stylophora - eoptera 100 100 Ophi000rdyceps acicularis - Coteoptera 0.1 substitutionslsite 100 Ophiocordyceps acicularis -oleoptera

Figure 63. Phylogenetic relationships among the hypocrealean fungi (147 taxa, 4,926 aligned nucleotide positions) to maximize the ecological diversity of host affiliation in three entomopathogenic families (Clavicipitaceae, Cordycipitaceae and Ophiocordycipitaceae) The topology presented is from a 50% majority consensus tree of 90,000 trees based combined evidence of five genes (ie. nrSSU, nrLSU, EF-la, RPBJ andRPB2).Outgroup taxa (Glomerella cingulata and Verticillium dahliae) were excluded and only the Hypocreales portion of the tree is shown to emphasize the relationships in the Hypocreales. The internodes that are supported with both bootstrap proportions of WP analyses (BP)? 70% and posterior probabilities (PP)? 95% are considered strongly supported and in a thicker line. For the strongly supported relationships, bootstrap proportions (BP) 70% and posterior probabilities (PP)> 95% are shown above and below the internodes,, respectivety. Numbers in the circles correspond to selected nodes for estimating the divergence times and reconstructing the ancestral states of host affiliations in Figure 6.2. Ancestral states of host affiliations were 244 reconstructed using maximum likelihood framework (Pagel,1999),as implemented in Multistate computer program (Pagel,1999).The host affiliations were coded in the kingdom level and the significance of the ancestral states is determined by a likelihood ratio test (Huelsenbeck and Rannala,1997).The branches were colored for the internodes based on the results of the likelihood ratio test (Black = ambiguous, blue = fungal, green = plant and red = animal). The proportions in a pie chart refer to the proportions of the maximum likelihood associated with each character state for the ambiguous internodes on the ancestral host states. The taxonomic units (e.g., class, order and genus) of the hosts are provided to right of the species.

(Figures 6.2 and 6.3). The basal lineages of the Cordycipitaceae comprise parasites of mushrooms and rust fungi and it is sister-group to the Hypocreaceae, which primarily includes parasites or saprobes of other fungi such as mushrooms and polypores

(Rossman et al., 1999) (Figure 6.3). Although pathogens of arthropods are the dominant ecology of the Cordycipitaceae, its ancestral ecology is not confidently addressed due to the sister-group relationships of the Cordycipitaceae and

Hypocreaceae and the intermixing of animal and fungal hosts among the more basal branches of the family. As a result, the ancestral ecology of the crown group of the

Cordycipitaceae and Hypocreaceae (node 6 in Figures 6.2 and 6.3) is inferred to be either fungal or arthropod pathogens and the polarities of the interkingdom host shifts are not clearly defined by the maximum likelihood reconstruction.

Our results do, however, support the Cretaceous radiation of the arthropod pathogens of the Cordycipitaceae (Figure 6.2). The age estimate for the origin of fungal-arthropod symbioses in the Cordycipitaceae is at least in the mid Cretaceous

(NPRS: 108 ± 7 MYA, PL: 101 MYA, node 16 in Figures 6.2 and 6.3). The majority of insect pathogens in the Cordycipitaceae attack lepidopterans (Chapter 5; Kobayasi

1941, Nikoh & Fukatsu REF). This ecology is of particular interest due to intimate association of Lepidoptera with the radiation of angiosperms during the Cretaceous. 245

The earliest fossil record of the lepidopterans is from the Early Jurassic (ca. 190

MYA) in England (Whalley, 1985). The lepidopterans are considered as being the most recently radiated group of insect orders and the largest lineage of the phytophagous insects (Grimaldi and Engel, 2005). The molecular age estimates

(Gaunt and Miles, 2002), coupled with the numerous fossil records (Grimaldi and

Engel, 2005; Labandeira and Sepkoski, 1993), support the Cretaceous and Early

Tertiary radiations for the major lineages of the lepidopterans. Although all pathogens of arthropods in the Cordycipitaceae are not associated with the lepidopterans, the ecology as pathogens of lepidopterans is one of major pathogenecities in the family

(Figure 6.3). Therefore, we propose that the Cretaceous radiation of fungal-arthropod symbioses in the family correlates with the radiation of the lepidopterans.

The Clavicipitaceae. The ancestral ecology of the Clavicipitaceae is reconstructed as an animal pathogen and the age estimates of its crown node is 138 ±

12 (NPRS) or 140 (PL) MYA in the Early Cretaceous (Figures 6.2 and 6.3; Table 6.2).

The Cretaceous diversification and cladogenesis of the family resulted in four ecological distinct subclades (Figure 6.3), representing multiple interkingdom host shifts among three major kingdoms of eukaryotes (Animal, Plant and Fungi). The family includes the ergot and grass endophytes (e.g., Claviceps, Balansia and Epichloe) that are hypothesized to be derived from an ancestor of animal pathogen through interkingdom host shifts (Figures 6.2 and 6.3; Spatafora et al., submitted). In addition, the family also includes another plant-associated genus Shim izuomyces, which parasitizes the seeds of Smilax sieboldii (Smilacaceae) (Kobayasi, 1981). In maximum 246 likelihood host reconstruction, a single origin of fungal-plant symbioses is neither supported nor rejected (Figure 6.3). Thus, the independent, parallel origins of plant- fungal symbioses remain as an alternative hypothesis for the evolution of fungal-plant symbioses in the family. Regardless, maximum likelihood reconstruction does support a single origin of the grass symbionts, although the grass symbionts do not form a monophyletic group due to derivation of pathogens of rust fungi (e.g., Verticillium epiphytum) from among the grass symbionts (Figure 6.3).

The grass symbionts in the family are specifically associated with the members of the spikelet dade (e.g., genera Arundo, Bambusa, Oryza and Zea) in the grass family Poaceae (Clay and Schardi, 2002; Diehi, 1950; Pazoutova, 2001), which are characterized by two-glumed spikelets with three-staminate flowers (GPWG, 2001).

The crown node of grass symbionts here dates back to nearly 100 MYA in the

Cretaceous, suggesting that fungal-grass symbioses originated at least in the mid

Cretaceous (Figure 6.2). These results are contradictory with the recent molecular clock study (Bremer, 2002) that shows that the grass family is estimated to be originated around the K-T boundary based on the calibration of 55 million-year-old fossil record (Crepet and Feldman, 1991). On the other hand, our results are consistent with the recent discovery of a 100 million years old grass-like fossil that possesses the structure of spikelets (bambusoid-type) (Poinar, 2004). Although this fossil record is not conclusive for the early origin of fungal-grass symbioses, there are several records of the Cretaceous fossil records for monocots (Daghlian, 1981; Gandolfo et al., 2000;

Gandolfo et al., 2002). Gandolfo et al., (2002) recently discovered a series of--90 247 million years old flower fossils that are extant members of the monocot family

Triuridaceae. Molecular studies (Bremer, 2000; Davies et al., 2004) also suggest that the origin of monocots is estimated to be 134 million years old and many lineages of monocots may date back to at least 80-100 MYA, indicating that the monocot may be of Cretaceous origin and one of the earliest lineages of angiosperms. Additional dating studies with the calibration of the recently discovered fossil records (Gandolfo et al.,

2002; Poinar, 2004) will expand our understanding of the evolution of monocot and provide further data on the possibility of ancient evolutionary history of fungal-grass symbioses.

In addition to plant-associated species, the Clavicipitaceae includes pathogens of arthropods and parasites of nematodes (Figure 6.3). The Cretaceous radiation of the subclade that includes the species of Metacordyceps is characterized by the frequent intrakingdom host shifts among the animal orders including Coleoptera, Hemiptera,

Lepidoptera and Nematoda (Figure 6.4). The reconstruction of the ancestral ecology of

Metacordyceps is complicated by the diversity of hosts, although the dominant ecology of the subclade is pathogens of arthropods. In contrast, species of the

Hypocrella subclade (e.g., Aschersonia and Hypocrella) are only known as pathogens of scale insects and white flies (Figure 6.3). The age estimates for the Hypocrella dade (node 18) is 92 ± 5 (NPRS) or 99 (PL) MYA, corresponding to the Cretaceous

(Figure 6.2; Table 6.2). Its ancestral ecology is inferred to be a pathogen of scale insects, a finding further corroborated by the recently described genus Regiocrella.

Chaverri et al., (2006) erected the genus Regiocrella based on morphologies of asci 248 and ascospores (e.g., lack of ascus apex and fusiform ascospores that are unicellular) for the species of pathogens of scale insects. In molecular analyses, species of the genus were placed as the basal lineage of the Hypocrella dade (Chaverri et al., 2006).

The age estimates presented here and the recent publication of Regiocrella (Chaverri et al., 2006) support a long evolutionary history for the interactions between scale insects and the pathogens of the Hypocrella dade.

Scale insects are currently classified in the superfamily Coccoidea (suborder

Sternorrhyncha) and consist of over 7,000 species (Cook et al., 2002). Although the fossil records suggest a considerable diversity of the hemipterans in the Permian

(Cook et al., 2002; Koteja, 2000; Labandeira, 1997; Labandeira and Sepkoski, 1993), the most convincing fossil record of extant members of scale insects is 130-140 million years old in the Early Cretaceous (Koteja, 1999). There are numerous

Cretaceous fossil records for most of extant families of scale insects from Alaskan,

Burmese and New Jersey ambers, suggesting the major radiation of scale insects in the

Cretaceous (Koteja, 2004; Koteja, 1999). The fossil record also indicates that the basal lineages of coccoids are abundant in coniferous ambers (Gnmaldi, 2000; Koetja, 2004;

Koteja, 1999). This suggests that many early coccoids fed on conifers as an ancestral ecology and colonized angiosperms in the Late Cretaceous. Of particular interest, the neococcoid scale insects (e.g., family Coccidae) are the most diverse group in the superfamily Coccoidea and their radiation is considered as being a result of the angiosperm diversification in the Late Cretaceous (Koetja, 2004; Koteja, 1999; Koteja,

2000). Our findings of an ancient origin of parasitism of scale insects by Hypocrella, which dates to the Cretaceous, is consistent with the co-diversification of the scales insects and angiosperms.

In addition to the pathogens of scale insects in the Clavicipitaceae, other families (i.e., Cordycipitaceae and Ophiocordycipitaceae) also include the pathogens of hemipterans (Figure 6.3). The hemipteran pathogens in the hypocrealean fungi exhibit two distinct patterns of host affiliation (Figure 6.3). The majority of species are specifically associated with either scale insects or cicadas, the latter of which has a fossil record that extends to the Pre-Mesozoic (Labandeira, 1997). Therefore, our age estimates (NPRS: 138 ± 12 MYA, PL: 140 MYA in node 14; NPRS: 133 ± 11 MYA,

PL: 140 MYA in node 15) and phylogenetic relationships support multiple origins of the hemipteran parasitism through the frequent interordinal and interkingdom host jumps (Figure 6.3).

The Ophiocordycipitaceae. The ancestral ecology of the Ophiocordycipitaceae is unequivocally reconstructed as an animal pathogen (Figure 6.2 and 6.3) and the age estimates of its crown group are in the Early Cretaceous (NPRS: 133 ± 11 MYA, PL:

140 MYA in node 15). The family includes the pathogens of arthropods, most of which are currently members of the genusOphiocordyceps.Although the monophyly of the genus receives less than 70% in the 147-taxon analyses (Figure 6.3), previous studies with sequence data from additional genes (Chapter 4; Chapter 5) demonstrated that the genus is a monophyletic group with good statistical support. Species of the genus parasitize insects in the orders Coleoptera, Hemiptera, Hymenoptera and

Lepidoptera (Figure 6.3) and collectively represent the most diverse dade of the 250 pathogens of arthropods in the Hypocreales (Chapter 5). Most of species in the family parasitize larval, nymph and pupal stages of subterranean and wood-inhabiting insects.

This ecology is in direct contrast to the species in the Cordycipitaceae that parasitize arthropods in relatively exposed settings in the forest, such as leaf litter (Chapter 5).

Because the Cretaceous radiation of pathogens of arthropods is intimately related with the diversification of angiosperms (Figure 6.2), the nearly simultaneous origins we observe for these two families indicate that the diversification of these two families may occurred in different niches, i.e., litter-inhabiting hosts for the Cordycipitaceae and soil and wood-inhabiting hosts for the Ophiocordycipitaceae, of paleoforests.

Although the dominant ecology of the Ophiocordycipitaceae as pathogens of arthropods, other ecologies represented in family include the parasites of nematodes and parasites of the truffle genus Elaphomyces (Figure 6.3). The nematode parasitism in the family is primarily represented by asexual forms that are closely related to

Cordyceps and Torrubiella (e.g., Haptocillium and Paecilomyces) (Hywel-Jones,

1993; Zare and Gams, 2001b). The sexual forms of nematode parasites often parasitize arthropods (e.g., genus Podocrella) and there is increasing evidence for the presence of disparate host affiliations between sexual and asexual forms of the same species

(Chaverri et al., 2005; Hywel-Jones, 1993). This trend is also found in the distantly related species (e.g., P. chiamydosporia) in the Clavicipitaceae (Figure 6.4); the sexual form of a nematophagous species, P. chiamydosporia, was linked to Metacordyceps chiamydosporia (=Cordyceps chiamydosporia), which parasitizes mollusc eggs (Zare et al., 2001). Interestingly, nematophagous species of Haptocillium in the 251

Ophiocordycipitaceae differ from Pochonia species of the Clavicipitaceae in host preference (free-living nematodes vs. nematode cysts and eggs) and some have proposed that nematode parasitism may be independently evolved from fungal- arthropod symbioses (Zare and Gams, 2001b; Zare et al., 2001).

In addition to nematode parasites, the Ophiocordyciptaceae includes the parasites of the truffle genus Elaphomyces (Kobayasi and Shimizu, 1960; Mains,

1957). Species of the parasites are currently classified in the genus Elaphocordyceps, which also includes parasites of cicadas (e.g., E. paradoxa and E. inegoensis) and beetles (e.g., E. subsessilis) (Chapter 5). Our results support an interkingdom host shift from animal to the truffles of the genus Elaphomyces in the Late Cretaceous (NPRS:

84 ± 8 MYA, PL: 70 MYA in node 15) (Figure 6.2). This estimate is older than the previous estimate (43 MYA) based on nucleotide substitutional rates in small subunit rDNA (Nikoh and Fukatsu, 2000). Within the genus Elaphocordyceps, the polarity of the host affiliation between arthropods and truffles is not apparent due to the low support stemming from small amounts of molecular differences among the species of

Elaphocordyceps (Figure 6.4). All fungal parasites of Elaphocordyceps are restricted to a single host genus, although numerous other truffle taxa co-exist in these environments, but arthropod hosts include two distantly related taxa. The morphological and molecular evidence (Chapter 5; Figure 6.4) are consistent, therefore, in demonstrating subsequent reversals from the truffles to beetle and cicadas. 252

All three entomopathogenic families (i.e., Clavicipitaceae, Cordycipitaceae and

Ophiocordycipitaceae) include species formerly classified in the genus Cordyceps

(Chapter 5) and its asexual allies (e.g., Lecanicillium, Pochonia, Hirsutella and

Hymenostilbe). Cordyceps was distinguished from other hypocrealean fungi based on the production of stipitate stromata and pathogenicity of arthropods. Here we demonstrate that pathogenicity of animalsarthropods in particularis an ancestral ecology for the Clavicipitaceae and Ophiocordycipitaceae, and possibly for the

Cordycipitaceae and that the current phylogenetic distribution of host affiliation is the product of frequent inter and intrakingdom host shifts (Figure 6.3). In total eight orders of arthropods serve as hosts for the species of Cordyceps, including Arachnida,

Coleoptera, Diptera, Hemiptera, Hymenoptera, Isoptera, Lepidoptera and Orthoptera

(Kobayasi, 1941; Kobayasi, 1982). Based on the estimate from the monographic studies of Cordyceps (Kobayasi, 1941; Kobayasi, 1982; Mains, 1958), more than 75 % species of Cordyceps are associated with insects in the orders Coleoptera, Hemiptera and Lepidoptera, which collectively represent the great diversity of the phytophagous insects (Grimaldi and Engel, 2005; Moran and Southwood, 1982). The radiations of the phytophagous insects are intimately associated with the angiosperm diversification in the Cretaceous (Farrell, 1998; Gaunt and Miles, 2002; Moran and Southwood,

1982) and these results support the Cretaceous radiation of many lineages of insect pathogens in three entomopathogenic families of the Hypocreales. Today, more than

25% of all extant arthropod species are known as phytophagous insects and occupy 50

% of the biomass and 75 % of the individuals in temperate deciduous forest (Grimaldi 253 and Engel, 2005; Moran and Southwood, 1982). Therefore, the diversity of phytophagous insect pathogens of the Hypocreales may be directly related to ecological dominance of the phytophagous insects and their coevolutionary interactions with angiosperms during the Cretaceous.

Internal consistencyofage estimates between Hypocreales and Sordariomycetes data sets

The divergence times of the major lineages of the Hypocreales were estimated based on the single calibration using the minimum age (100 MYA) of the Hymenostilbe fossil (Figure 6.2; Table 6.2). To verify the results of the age estimates from the

Hypocreales data set, we estimated the divergence times of the major lineages of the

Sordariomycetes by calibrating the Bayesian consensus tree (Figure 6.4) from the

Sordariomycetes data set based on the minimum ages of 100 and 400 MYA of the

Hymenostilbe and Paleopyrenomycites fossils, respectively. Because many of the higher relationships are not strongly supported (Figure 6.4), eight nodes that receive strong statistical supports were selected to estimate the divergence times in the

Sordariomycetes and compare the age estimates with those from the Hypocreales data set (Figures 6.4 and 6.5; Table 6.3). Three independent dating analyses were performed using both NPRS and PL methods due to the uncertainty of the calibration point of the Sordariomycetes fossil (see materials and methods).

The dating analyses based on the single calibration of the Hymenostilbe fossil resulted in two age estimates from NPRS and PL methods that differ somewhat for the 254

Kylaria sp. XylarialeS Lindra thalassiae Luiworthia grandispora Luiworthiales - Gnomonia gnomonI Dlaportheeres ijiapui UIdIe - Camaropssp. IBoliniales Lanspora coronata - Papulosa amerospora Halosphaeriales Ophiostoma pillferum Ophiostomatales Sordariomycetes Thyridium sp. - Glomerella clngulaia Phyllachorales Verticillium dahllae Cirrenalia macrocephala Petriella setffera Microascales Microascus trigonosporus Stachyfr toys subslmple.x - Myrotheclum roridum Peethambara spirostriata - Pseudoneciira rousseliana %/lridispora diparietispora Ophionectria trichoSpOra HydropiSphaera peziza erubescens Roumegueriellarufula Hypocreales Verticillium incurvum [ Hypocrea lutea Sphaerostllbella berkeleyana 100 SimplicIllium lameilicola SImplicillium lanosoniveum E ' Cordyceps militaris 100 100 Isaria tenulpes cardinalls Tarrubiella ratticaudata ez Pochonia gonioides tail Rotiferophthosa angustlspora F Ilsctiersonia badia iso Hypocrella schizostachyi 100 BalanSia heioningsiana Epichloe iyphina 01 Clavlceps paspali los Cavlceps purpurea ioor Cordyceps gu,lnhl balanoldes 100 Haptocillium zeosporum Elaphocordyceps capitata G °° ElaphocordycepS ophiaglossoldes Elaphocordyceps fracta Elaphocordyceps Japonica Ophiocordyceps stylophora 01 Ophlocordyceps agriota los Ophiocordyceps Cf. acicularis Ophlocordyceps heteropoda 15 Ophiocordyceps melolonthae Ophlocordyceps revenelli Ophiocordyceps '5 Ophlocordyceps aphodli - 0.05subititutloi1/lte iso Ophiocordyceps brunneapunctata

Figure 6.4. Phylogenetic relationships among the fungi in Sordariomycetes (60 taxa, 5,110 aligned nucleotide positions) based combined evidence of five genes (i.e. nrSSU, nrLSU, EF- la,RPBJandRPB2).The topology presented is a 50% majority consensus tree of 90,000 trees, rooted using Cochliobolus heterostrophus, and only Sordariomycetes portion of the tree is shown to emphasize the relationships. The internodes that are supported with both bootstrap proportions of WP analyses (BP)70% and posterior probabilities (PP)?: 95% are considered strongly supported and in a thicker line. For the strongly supported relationships, bootstrap proportions (BP)? 70% and posterior probabilities (PP)? 95% are shown above and below the internodes, respectively. Alphabets in the circles correspond to eight selected intemodes for estimating the divergence times in Figure 6.5 and Table 6.3. 255 crown group of the Sordariomycetes (node A in Figure 6.4.A; Table 6.3). The age estimate from NPRS method is 355 ± 34 MYA, which is younger than the minimum age (400 MYA) of the Sordariomycetes fossil (Figure 6.5.A; Table 6.3). The results from PL methods, however, infer the crown node of the Sordariomycetes to be approximately 424 million years old (Figure 6.5.A; Table 6.3), consistent with

Paleopyrenomycites as a member of the crown group of the Sordariomycetes. The results from PL method are also consistent with a previous dating study (Padovan et al., 2005), which concluded that the crown node of the Sordariomycetes is older than

400 MYA. The age estimates of the crown node of the genus Ophiocordyceps (node H in Figure 6.4.B; Table 6.3) based on Paleopyrenomycites are 94 and 115 MYA from

PL and NPRS methods, respectively. Although the genus Hymenostilbe is one of the derived lineages in the genus Ophiocordyceps (Chapters 4 and 5), the results of the age estimates on the genus Ophiocordyceps are congruent with the minimum age (100MYA) of the Hymenostilbe fossil, thus verifying an internal consistency between the calibration points based on the Hymenostilbe and Paleopyrenomycites fossils.

In recent molecular dating studies (Heckman et al., 2001; Hedges et al., 2004;

Padovan et al., 2005), the Paleopyrenomycites fossil (Taylor et al., 1999; Taylor et al.,

2005) was used to estimate divergence times of the major fungal lineages. The minimum age of the fossil was, for example, used by Heckman et al. (1999) in interpreting the Pre-Cambrian divergence (670 MYA) of the and

Sordariomycetes, and Padovan et al. (2005) suggested that the crown group age of 256

(A) W/ Hymenostilbe fossil

Corciycipitaceae

Clavicipitaceae

Ophiocordycipitaceae

Cordycipitaceae

Clavicipitaceae

Ophiocordyciprtaceae * (B) W/ Sordariomycetes fossil

Cordycipitaceae NP CIavicipitaceae

Ophiocordycipitaceae

Cordycipitaceae PL Clavicipitaceae

Ophiocordycipitaceae

(C) W/Hymenostibe+Sordariomycetes fossils

Cordycipitaceae

Clavicipitaceae

Ophiocordycipitaceae

Cordycipitaceae PL Ciavicipitaceae

Ophiocordycipitaceae *

ACfl Ann 16fl Inn ICfl .. Inn 100 50 MYA

I0rdovcid1I Silurian CarboniferousPermianTriassic Jurassicj Cretaceous Tertiary I I IPaleozoic Mesozoic Cenozoic

Figure 6.5. Schematic diagrams of chronograms based on a Bayesian consensus tree (Figure 6.4) from Sordariomycetes data set. The divergence times were estimated nonparametric rate smoothing (NPRS) and penalized likelihood (PL) methods and the estimated ages (Table 6.3) are indicated by the length of branches for eight selected internodes (nodes A-El in Figure 6.4) 257

that are strongly supported. Geological times are provided below the chronograms and the Cretaceous period is highlighted by the gray rectangular box. The calibrations were applied with the minimum ages (100 MYA and 400 MYA) of the Hymenostilbe and Paleopyrenomycites fossils to the crown nodes of Ophiocordyceps (node H in Figure 6.4) and the Sordariomycetes (node A in Figure 6.4), respectively. Three independent dating analyses were performed and their corresponding calibrations are marked with asterisks(*). (A) The minimum age (100 MYA) of the Hymenostilbe fossil was used as a calibration for the crown node of Ophiocordyceps (node H in Figure 6.4) to estimate the divergence times of the crown node of Sordariomycetes (node A in Figure 6.4). (B) The minimum age (400 MYA) of the Paleopyrenomycites fossil was applied to the crown node of Sordariomycetes (node A in Figure 6.4) to estimate the divergence times of the crown node of Ophiocordyceps (node H in Figure 6.4). (C) Two reference fossils were simultaneously used in calibrations to confirm the age estimates from the Hypocreales data set for the crown nodes (nodes E, F and G in Figure 6.4) of three entomopathogenic families (Clavicipitaceae, Cordycipitaceae and Ophiocordycipitaceae).

Sordariomycetes is older than the minimum age of the fossil record. These studies rely

on distant external calibrations with the closest calibration being the crown node of

animals, plants, and fungi (1600 MYA) based on extrapolations from a molecular

dating study of vertebrates (Heckman et al., 2001; Hedges et al., 2004; Padovan et al.,

2005). Nonetheless, the results from these studies are consistent with the age estimates

of the Sordariomycetes from this study based on internal calibrations using the

Hymenostilbe fossil and corroborating the Paleopyrenomycites fossil as a minimum

age of the crown group of the Sordariomycetes.

Based on the results of the dating analyses with single calibration point (Figure

6.5.A and B), the minimum ages of the Hymenostilbe and Sordariomycetes fossils were simultaneously used to estimate the divergence times of the major lineages in the

Hypocreales. The age estimates from NPRS and PL methods (Figure 6.4.C; Table 6.3) are not significantly different as in the dating analyses based on single calibration of the Hymenostilbe fossil (Figure 6.4.A). The crown group of the Hypocreales (node C in Figures 6.4 and 6.5) is estimated to be 251 or 278 million years old. 258

Table 6.3. Inferred age estimates on the major divergences in the Sordariomycetes based on a Bayesian consensus tree (Figure 6.4) from Sordariomycetes data set. The node number corresponds to the selected internodes in estimating the divergence times in Figure 6.4. The age estimates were estimated using nonparametric rate smoothing (NPRS) and penalized likelihood (PL) methods. The standard deviations are based on the NPRS analyses of Bayesian trees. Three independent dating analyses were performed and their corresponding calibration points are marked with asterisks (*).

wi Hymenostilbe fossil WI Sordariomycetes fossil wi Hymeno.+ Sordar. fossil Node Est. ages (Myr) Est. ages (Myr) Est. ages(Myr) ± D ± D ± D NPRS PL NPRS PL NPRS PL A 355.56424.56 ±34.41 400* 400* +0.0 400* 400* +0.0 B 283.46354.85 +44.54 325.43 334.43 ±38.46 325.65 337.54±28.54 C 246.21 290.45 ±35.14 278.95 273.64±38.56 279.98 251.43 ±24.56 D 188.86220.16 +28.43 218.54 207.42 +29.56 223.75 214.54 ±18.65 E 140.54 169.45 +32.21 158.65 159.64 +39.45 150.43 143.54 ±12.43 F 137.56 158.84 +23.32 156.73 149.65 +29.32 138.87 134.56 ±18.54 G 131.43 151.46 +20.43 148.43 142.69 ±23.45 134.54 132.32 ±9.43 H 100* 100* ±0.0 115.75 94.21 ±14.65 100* 100* ±0.0

These age estimates are slightly older than those from the Hypocreales data set (NPRS

236 MYA, PL 247 in Table 6.2). However, the age estimates of three entomopathogenic families (Cordycipitaceae: NPRS 150 MYA, PL143 MYA;

Clavicipitaceae: NPRS 138 MYA, PL 134 MYA; Ophiocordycipitaceae: NPRS 134

MYA, PL 132 MYA) are consistent with those (Cordycipitaceae: NPRS 142 MYA,

PL147 MYA; Clavicipitaceae: NPRS 138 MYA, PL 140 MYA;

Ophiocordycipitaceae: NPRS 133 MYA, PL 140 MYA) from the Hypocreales data set

(Tables 6.2 and 6.3). These results indicate that the Cretaceous radiation of fungal- arthropod symbioses in these three families is a robust hypothesis for the evolution of the Hypocreales and is consistently resolved across multiple data sets and calibration strategies. 259

Conclusion

The Cretaceous radiation of angiosperms fundamentally changed terrestrial ecosystems and affected the terrestrial biodiversity of other organisms. Phylogenetic dating and reconstruction of the early evolutionary history of arthropods, coupled with the numerous fossil records (Barret and Willis, 2000; Dmitriew and Ponomarenko,

2002; Labandeira et al., 2002; Labandeira and Sepkoski, 1993), have demonstrated that the diversification of numerous lineages in arthropods is intimately associated with the radiation of angiosperms (Dmitriew and Ponomarenko, 2002; Strong et al.,

1984). Here, we report a dating study that provides fungal examples of the Cretaceous radiation and documents the Cretaceous radiation of fungal-arthropod symbioses by a dynamic process of inter- and intra-kingdom host jumping.

References

Barret, P. M., and K. J. Willis. 2000. Did dinosaurs invent flowers? Dinosaur- angiosperm coevolution revisited. Biol. Rev. Camb. Philos. Soc. 76:411-447.

Berbee, M. L., and J. W. Taylor. 1993. Dating the evolutionary radiations of the true fungi. Canadian Journal of Botany 71:1114-1127.

Bremer, K. 2000. Early Cretaceous lineages of monocot flowering plants. PNAS 97:4707-47 1 1.

Bremer, K. 2002. Gondwanan evolution of the grass alliance of families (). Evolution 56:1374-1387. Bright, D. E. 1993. Systematics of bark beetles. Pages 23-36 in Beetle-pathogen interactions in conifer forests. (T. D. Schowalter, and G. M. Filip, eds.). Academic press, London.

Castlebury, L. A., A. Y. Rossman, G.-H. Sung, A. S. Hyten, and J. W. Spatafora. 2004. Multigene phylogeny reveals new lineage for Stachybotrys chartarum, the indoor air fungus. Mycological Research 108:864-872.

Chaverri, P., J. F. Bischoff, H. C. Evans, and K. T. Hodge. 2006. Regiocrella, a new entomopathogenic genus with a pycnidial anamorph and its phylogenetic placement in the Clavicipitaceae. Mycologia in press.

Chaverri, P., G. J. Samuels, and K. T. Hodge. 2005. The genus Podocrella and its nematode-killing anamorph Harposporium. Mycologia 97:433-443.

Clay, K., and C. L. Schardi. 2002. Evolutionary origins and ecological consequences of endophyte symbiosis with grasses. The American Naturalist 160:S99-S 127.

Cook, L. G., P. J. Gullan, and H. E. Trueman. 2002. A preliminary phylogeny of the scale insects (Hemiptera: Sternorrhyncha: Coccoidea) based on nuclear small- subunit ribosomal DNA. Molecular Phylogenetics and Evolution 25:43-52.

Crepet, W. L., and G. D. Feldman. 1991. The earliest remains of grasses in the fossil record. American Journal of Botany 78:1010-1014.

Cruickshank, R. D., and K. Ko. 2003. Geology of an amber locality in the Hukawng Valley, northern Myanmar. Journal of Asian Earth Sciences 21:441-455.

Currie, C. R., B. Wong, A. E. Stuart, T. R. Schultz, S. A. Rehner, U. G. Mueller, G.-H. Sung, J. W. Spatafora, and N. A. Straus. 2003. Ancient tripartite coevolution in the attine ant-microbe symbiosis. Science 299:386-388.

Daghlian, C. P. 1981. A review of the fossil record of . Bot. Rev. 47:517-555.

Davies, T. J., T. G. Barraclough, M. W. Chase, P. S. Soltis, D. E. Soltis, and V. Savolainen. 2004. Darwin's abominable mystery: Insights from a supertree of the angiosperms. PNAS 101:1904-1909.

Diehi, W. W. 1950. Balansia and the Balansiae in America. Agricultural monograph No.4. U. S. D. A. Washington. 261

Dmitriew, V. Y., and A. G. Ponomarenko. 2002. Dynamics of insect taxonomic diversity. Pages 517 in History of insects. (A. P. Rasnitsyn, and D. L. J. Quicke, eds.). Kiuwer Academic, Dordrecht, the Netherlands.

Farrell, B. D. 1998. "Inordinate fondness" explained: Why are there so many beetles? Science 281:555-559.

Felsenstein, J. 1981. A likelihood approach to character weighting and what it tells us about parsimony and compatibility. Bio. J. Linn. Soc. 16:183-196.

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

Friis, E. M., K. R. Pedersen, and P. R. Crane. 2001. Fossil evidence of water lilies in the Early Cretaceous. Nature 410:357-360.

Gandolfo, M. A., K. C. Nixon, and W. L. Crepet. 2000. Monocotyledons: A review of their early Cretaceous record.in Monocots: systematics and evolution. (K. L. Wilson, and D. A. Morrison, eds.). CSIRO, Melbourne.

Gandolfo, M. A., K. C. Nixon, and W. L. Crepet. 2002. Triuridaceae fossil flowers from the Upper Cretaceous of New Jersey. American Journal of Botany 89: 1940-1957.

Gaunt, M. W., and M. A. Miles. 2002. An insect molecular clock dates the origin of the insects and accords with palaeontological and biogeographic landmarks. Molecular Biology and Evolution 19:748-761.

GPWG. 2001. Phylogeny and subfamilial classfication of the grasses (Poaceae). Ann. Mo. Bot. Gard. 88:373-457.

Grimaldi, D., and M. S. Engel. 2005. Evolution of the insects. Cambridge Univ. Press, Hong Kong, China.

Grimaldi, D. A. 2000. A diverse fauna of Neuropterodea in amber from the Cretaceous of New Jersey. Pages 259-303 in Studies on fossils in amber, with particular refenence to the Cretaceous of New Jersey. (D. A. Grimaldi, ed.) Backhuys Publishers, Leiden, the Netherlands.

Heckman, D. S., D. M. Geiser, B. R. Eidell, R. L. Stauffer, N. L. Kardos, and S. B. Hedges. 2001. Molecular evidence for the early colonization of land by fungi and Plants. Science 293:1129-1133. 262

Hedges, S., J. Blair, M. Venturi, and J. Shoe. 2004. A molecular timescale of evolution and the rise of complex multicellular life. BMC Evolutionary Biology 4:2-10.

Hodge, K. T. 2003. Clacicipitaceous anamorphs. Pages 75-123 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Huelsenbeck, J. P., and B. Rannala. 1997. Phylogenetic methods corn of age: Testing hypotheses in an evolutionary context. Science 276:227-232.

Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.

Huelsenbeck, J. P., F. Ronquist, R. Nielsen, and J. P. Boilback. 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294:2310-2314.

Hywel-Jones, N. L. 1993. Torrubiella luteorostrata: a pathogen of scale insects and its association with Paecilomyces cinnamomeus with a note on Torrubiella tenuis. Mycological Research 97:1126-1130.

Hywel-Jones, N. L. 1995a. Cordyceps sphecocephala and a Hymenostilbe sp. infecting wasps and bees in Thailand. Mycological Research 99:154-158.

Hywel-Jones, N. L. 1995b. Hymenostilbe ventricosa sp. nov. a pathogen of cockroaches in Thailand. Mycological Research 99:1201-1204.

Hywel-Jones, N. L., and H. C. Evans. 1993. Taxonomy and ecology of Hypocrella discoidea and its anamorph, Ashersonia samoensis. Mycological Research 97:871-876.

Hywel-Jones, N. L., and G. J. Samuels. 1998. Three species of Hypocrella with large stromata pathogenic on scale insects. Mycologia 90:36-46.

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku (Section B, no. 84) 5:53-260.

Kobayasi, Y. 1981. Revision of the genus Cordyceps and its allies 1. Bulletin of the National Science Museum, Tokyo, series B 7:1-13.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364. 263

Kobayasi, Y., and D. Shimizu. 1960. Monographic studies of Cordyceps 1. Group parasitc on Elaphomyces. Bulletin of the National Science Museum, Tokyo, Ser. B. 5:69-85.

Koetja, J. 2004. Scale insects (Hemiptera: Coccinea) from Cretaceous Myanmar (Burmese) amber. Journal of Systematic Paleontology 2:109-1 14.

Koteja, A. 1999. Eomatsucoccus andrewi sp. nov. (Hemiptera: Sternorrhyncha: Coccinea) from the Lower Cretaceous of southern England. Cretaceous Research 20:863-866.

Koteja, A. 2000. Advances in the study of fossil coccids (Hemiptera: Coccinea). Polskie Pismo Entomologiczne 69:187-2 18.

Labandeira, C. C. 1997. Insect mouthparts: Acertaining the paleobiology of Insect feeding stratagies. Annu. Rev. Ecol. Syst. 28:153-193.

Labandeira, C. C., K. R. Johnson, and P. Wilf. 2002. Impact of the terminal Cretaceous event on plant-insect associations. PNAS 99:2061-2066.

Labandeira, C. C., and J. J. Sepkoski. 1993. Insect diversity in the fossil record. Science 261:310-315.

Lidgard, S., and P. R. Crane. 1988. Quantitative analysis of early angiosperm radiation. Nature 331:344-346.

Lutzoni, F., F. Kauff, C. J. Cox, D. McLaughlin, G. Celio, B. Dentinger, M. Padamsee, D. Hibbett, T. Y. James, E. Baloch, M. Grube, V. Reeb, V. Hofstetter, C. Schoch, A. B. Arnold, J. Miadlikowska, J. Spatafora, D. Johnson, S. Hambleton, M. Crockett, R. Shoemaker, G.-H. Sung, R. Lucking, T. Lumbsch, K. O'Donnell, M. Binder, P. Diederich, D. Ertz, C. Gueidan, K. Hansen, R. C. Harris, K. Hosaka, Y. -W. Lim, B. Matheny, H. Nishida, D. Pfister, J. Rogers, A. Rossman, I. Schmitt, H. Sipman, J. Stone, J. Sugiyama, R. Yahr, and R. Vilgalys. 2004. Assembling the fungal tree of life: progress, classification, and evolution of subcellular traits. American Journal of Botany 91:1446-1480.

Maddison, D. R., and W. P. Maddison. 2000. MacClade 4: Analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts.

Mains, E. B. 1957. Species of Cordyceps parasitic on Elaphomyces. Bulletin of The Torrey Botanical Club. 84:243-251. 264

Mains, B. B. 1958. North American entomogenous species of Cordyceps. Mycologia 50: 169-222.

Moran, V. C., and T. R. C. Southwood. 1982. The guild composition of arthropod commumities in trees. Journal of Animal Ecology 5 1:289-301.

Nikoh, N., and T. Fukatsu. 2000. Interkingdom host jumping underground: Phylogenetic analysis of entomopathogenic fungi of the genus Cordyceps. Molecular Biology and Evolution 17:629-638.

Padovan, A. C. B., G. F. 0. Sanson, A. Brunstein, and M. R. S. Briones. 2005. Fungi evolution revisited: application of the penalized likelihood method to a bayesian fungal phylogeny provides a new perspective on phylogenetic relationships and divergence dates of ascomycota groups. Journal of Molecular Evolution 60:726-735.

Pagel, M. 1999. Inferring the historical patterns of biological evolution. Nature 401:877-884.

Page!, M., A. Meade, and D. Barker. 2004. Bayesian estimation of ancestral character states on phylogenies. Systematic Biology 53:673-684.

Pazoutova, S. 2001. The phylogeny and evolution of the genus Claviceps. Mycological Research 105:275-283.

Percy, D. M., R. D. E. Page, and Q. C. B. Cronk. 2004. Plant-insect interactions: Double-dating associated insect and plant lineages reveals asynchronous radiations. Systematic Biology 53:120-127.

Petch, T. 1921. Studies in entomogenous fungi. II. The genera of Hypocrella and Aschersonia. Ann. R. Bot. Gard (Peradeniya) 7:167-278.

Poinar, G. 0. J. 2006. Early fossil evidence of fungal parasitism of insects. Mycologia in preparation.

Poinar, G. 0. J., and G. M. Thomas. 1984. A fossil entomogenous fungus from Dominican Amber. Experientia 40:578-579.

Poinar, J. G. 0. 2004. Programinis burmitis gen. et sp. nov., and P. laminatus sp. nov., early cretaceous grass-like monocots in Burmese amber. Australian Systematic Botany 17:497-504.

Posada, D., and K. A. Crandall. 1998. ModelTest: Testing the model of DNA substitution. Bioinformatics 14:817-818. 265

Reeb, V., F. Lutzoni, and C. Roux. 2004. Contribution of RPB2 to multilocus phylogenetic studies of the euascomycetes (Pezizomycota, Fungi) with special emphasis on the lichen-forming Acarosporaceae and evolution of polyspory. Molecular Phylogenetics and Evolution 32:1036-1060.

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

Rossman, A. Y., G. J. Samuels, J. S. Rogers, and R. Lowen. 1999. Genera of Bionectriaceae, Nectriaceae and Hypocreaceae (Hypocreales, Ascomycetes). Studies in Mycology No. 42.

Samson, R. A., H. C. Evans, and J.-P. Latge. 1988. Atlas of entomopathogenic fungi. Springer-Verlag, Berlin Heidelberg New York.

Sanderson, M. 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Molecular Biology and Evolution 14:1218-123 1.

Sanderson, M. J. 2002. Estimating absolute rates of molecular evolution and divergence Times: A penalized likelihood approach. Molecular Biology and Evolution 19:101-109.

Sanderson, M. J. 2003. r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinformatics 19:301- 302.

Sanderson, M. J., and J. A. Doyle. 2001. Sources of error and confidence intervals in estimating the age of angiosperms from rbcL and 1 8S rDNA data. American Journal of Botany 88:1499-15 16.

Soltis, P. 5., and D. B. Soltis. 2004. The origin and diversification of angiosperms. American Journal of Botany 91:1614-1626.

Soltis, P. 5., D. E. Soltis, V. Savolainen, P. R. Crane, and T. G. Barraclough. 2002. Rate heterogeneity among lineages of tracheophytes: Integration of molecular and fossil data and evidence for molecular living fossils. PNAS 99:4430-4435.

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes I. the Clavicipitales-Hypocreales connection. Mycologia 85:912-922. 266

Spatafora, J. W., G.-H. Sung, J.-M. Sung, N. L. Hywel-Jones, and J. F. White, Jr. 2006. Phylogenetic evidence for an animal pathogen origin for ergot and the grass endophytes. PNAS submitted.

Spatafora, J. W., B. Volkmann-Kohlmeyer, and J. Kohlmeyer. 1998. Independent terrestrial origins of the Halosphaeriales (marine Ascomycota). American Journal of Botany 85:1569-1580.

Strong, D. R., J. H. Lawton, and T. R. E. Southwood. 1984. Insects on plants. Harvard Univ. Press, Cambridge, MA.

Suh, S.-O., H. Noda, and M. Blackwell. 2001. Insect symbiosis: Derivation of yeast- like endosymbionts within an entomopathogenic filamentous lineage. Molecular Biology and Evolution 18:995-1000.

Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.OblO. Sinauer, Sunderland, Massachusetts.

Taylor, T. N., H. Hass, and H. Kerp. 1999. The oldest fossil ascomycetes. Nature 399:648.

Taylor, T. N., H. Hass, H. Kerp, M. Krings, and R. T. Hanlin. 2005. Perithecial ascomycetes from the 400 million year old Rhynie chert: an example of ancestral polymorphism. Mycologia 97:269-285.

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673-4680.

Thorne, J. L., and H. Kishino. 2002. Divergence time and evolutionary rate estimation with multilocus data. Systematic Biology 51:689-702.

Whalley, P. E. S. 1985. The systematics and palaeogeography of teh Lower Jurrasic insects of Dorset, England. Bulletin of the British Museum (Natural History) (Geology) 39:107-189.

Wheeler, W. C. 1990. Combinatorial weights in phylogenetic anlaysis: a statistical parsimony procedure. Cladistics 6:269-275.

Wing, S. L. 2000. Evolution and expansion of flowering plants. Pages 308 in Phanerozoic terrestrial ecosystems (R. A. Gastaldo, and W. A. DiMichele, eds.). The Paleontological society, Pittsburgh. 267

Wood, S. L. 1982. The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Nat. Mem. 6:37-39.

Yang, Z., and A. D. Yoder. 2003. Comparison of likelihood and Bayesian methods for estimating devergence times using multiple gene loci and calibration points with application to a radiation of cute-looking mouse lemur species. Systematic Biology 52:705-716.

Zare, R., and W. Gams. 2001. A revision of Verticillium sect. Prostrata. VI. The genus Haptocillium. Nova Hedwigia 73 :271-292.

Zare, R., W. Gams, and H. C. Evans. 2001. A revision of Verticillium sect. Prostrata. V. The genus Pochonia, with notes on Rotiferophthora. Nova Hedwigia 73:5 1- 86. Chapter 7

CONCLUSIONS

In this dissertation, I have tested hypotheses of the systematics of Cordyceps and the Clavicipitaceae based on multi-gene phylogenetic analyses. The analyses of the combined sequences of nrSSU, nrSSU, /3-tubulin, EF-la, RPBJ, RPB2 and mtATP6 in Chapter 4 indicate that the Clavicipitaceae comprises of three strongly supported clades. This inferred 3-dade hypothesis for the family supports the non- monophyly of both Cordyceps and the Clavicipitaceae. The Clavicipitaceae is paraphyletic because of the monophyly of the Hypocreaceae and one of three clavicipitaceous clades (Clavicipitaceae dade C in Figure 4.2 and 4.3). These results are also not consistent with the subfamilial classification by Diehi (1950), who divided the family into three subfamilies (i.e., Clavicipitoideae, Cordycipitoideae and

Oomycetoideae) with considerable emphasis on host affiliation. Although the monophyly of the subfamily Clavicipitoideae is strongly supported, all three clavicipitaceous clades include the species of Cordyceps, which is the representative genus of the subfamily Cordycipitoideae.

The phylogenetic analyses in Chapter 5 were performed with the most extensive sampling of Cordyceps and the Clavicipitaceae based on the combined sequences of nrSSU, nrSSU, EF-la, RPBJ and RPB2. The inferred relationships are consistent with the 3-dade hypothesis and reject a subgeneric classification of 269

Cordyceps that consists of four subgenera, i.e., C. subg. Bolacordyceps, C. subg.

Cordyceps, C. subg. Neocordyceps and C. subg. Ophiocordyceps. Therefore, a new classification of Cordyceps and the clavicipitaceous fungi was proposed to be consistent with the phylogeny (Chapter 5). One new family Ophiocordycipitaceae was proposed and two families (Clavicipitaceae and Cordycipitaceae) were emended to reflect three clavicipitaceous clades. Species of Cordyceps were reclassified into

Cordyceps sensu stricto, Elaphocordyceps, Metacordyceps and Ophiocordyceps. The

Cordycipitaceae includes the species of Cordyceps sensu stricto that are characterized by possessing fleshy stromata that are pallid or brightly colored. Metacordyceps was proposed for the species of Cordyceps sensu lato that are closely related with the grass symbionts in the Clavicipitaceae s. s. In addition, two new genera Elaphocordyceps and Ophiocordyceps were proposed for the species of Cordyceps sensu lato in

Ophiocordycipitaceae. Elaphocordyceps is primarily represented by pathogens of

Elaphomyces and darkly pigmented stromata with olivaceous tints. Ophiocordyceps is the most diverse genus in terms of number of species. The genus is characterized by possessing stromata that are fibrous to tough to pliant in texture and darkly pigmented in at least its some part of the stroma. A total of 147 new combinations were proposed based on the new classification system.

In addition to improving the phylogenetic classification, the analyses in

Chapters 2, 3 and 5 provide information on the utility of anamorphic forms in the classification of Cordyceps. As in many other genera in Ascomycota, anamorphic forms play important roles in the life cycles of Cordyceps sensu lato and more than 25 270 asexual genera have been linked to Cordyceps sensu lato (Hodge, 2003). The analyses in Chapter 2 and 5 included the 18 most diverse anamorphic genera and demonstrated that several anamorphic forms (e.g., Aschersonia, Beauveria, Hirsutella, Hymenostilbe and Metarhizium) are phylogenetically informative in characterizing the species of

Cordyceps sensu lato. As a result, the analyses presented in Chapter 5 enable us to integrate the following maj or anamorphic genera with their teleomorphic genera in the new classification system (teleomorphic genera: anamorphic genera): Cordyceps:

Beauveria, Lecanicillium, Isaria; Hypocrella: Aschersonia; Metacordyceps:

Metarhizium, Pochonia; Ophiocordyceps: Hirsutella, Hymenostilbe.

The representative taxa in major families of the Hypocreales were selected in

Chapter 6 to better understand the evolution of host affiliation. In addition to the multi-gene phylogenetic analyses, ancestral states of host affiliation were determined in a maximum likelihood framework and phylogenetic dating was conducted for major lineages in the Hypocreales. These analyses support a Triassic or early Jurassic origin of the hypocrealean fungi and provide evidence for their Cretaceous radiation.

Ancestral nutritional mode of the hypocrealean fungi is inferred to be plant-based nutrition (e.g., decomposers of plant debris and plant pathogens). The inferred dates and patterns of host affiliation support the Jurassic origin of fungal-animal symbioses in the hypocrealean fungi and the Cretaceous radiation of the pathogens of arthropods in three entomopathogenic families, i.e., Clavicipitaceae, Cordycipitaceae and

Ophiocordycipitaceae. A total of 9-14 interkingdom host shifts are hypothesized in three entomopathogenic families, including 5-7 onto fungi, 2-4 onto animals, and 2-3 271 onto plants. These results reject the previous hypotheses that the grass symbionts were derived from plant pathogens (Carroll, 1988; Saikkonen et al., 1998). Instead, the grass symbionts are hypothesized to have originated from the pathogens of arthropods through interkingdom host shift in the Cretaceous. The total evidence of the

Cretaceous radiation presents the hypothesis that the diversification of fungal- artbropod symbioses is to some extent associated with the diverse interaction between arthropods and angiosperms in the Cretaceous.

The results from this study support a long and diverse evolutionary history of the fungal-arthropod symbioses in the hypocrealean fungi and provide a foundation for understanding character state evolution that is both more reflective of phylogeny and more useful in taxonomy. As a result, the inferred relationships have resulted in a new classification ofCordyceps sensu latoto be consistent with the phylogeny. Although this study included the most extensive sampling ever forCordyceps s.1., the diversity ofCordyceps s. 1. isconsidered largely under-represented in the phylogenetic analyses. The taxonomy and systematics ofCordyceps s. 1.remain to be further clarified due to the large number of species (over 400 species) and inability to locate the numerous type specimens (e.g., over 100 species of the collection of Y. Kobayasi)

(Mains, 1957; Kobayasi, 1982). Further studies with the combination of morphological and molecular phylogenetic approaches are necessary to better understand the species boundary and the diversity of the entomopathogenic fungi of the Hypocreales. 272

In fungi, approximately 70,000 species have been described (Hawksworth,

1991; Hawksworth and Rossman, 1997). This number only represents 5-10% of the fungal diversity in the world based on the estimation from fungal flora associated with plants in well-documented regions of temperate ecosystems (Hawksworth, 1991;

Hawksworth and Rossman, 1997). Although this estimation did not take account of the fungal diversity associated with arthropods, arthropod-associated fungi are undoubtedly one of the under-documented repositories of fungal diversity.(Spatafora,

2002). As additional taxa are incorporated in further molecular phylogenetic studies, a better understanding of the fungal diversity will be achieved for the fungal-arthropod symbioses in the hypocrealean fungi.

References

Carroll, G. C. 1988. Fungal endophytes in stems and leaves: from latent pathogen to mutualistic symbiont. Ecology 69:2-9.

Diehi, W. W. 1950.Balansiaand the Balansiae in America. Agricultural monograph No.4. U. S. D. A. Washington.

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

Hawksworth, D. L., and A. Y. Rossman. 1997. Where are all the undescribed fungi? Phytopathology 87:888-891.

Hodge, K. T. 2003. Clacicipitaceous anamorphs. Pages 75-123inClavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York. 273

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku (Section B, no. 84) 5:53-260.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364.

Mains, E. B. 1958. North American entomogenous species of Cordyceps. Mycologia 50:169-222.

Saikkonen, K., S. H. Faith, M. Helander, and T. J. Sullivan. 1998. Fungal endophytes: A continuum interactions with host plants. Annu. Rev. Ecol. Syst. 19:319-343.

Spatafora, J. W. 2002. Evolution of Ascomycota-arthropoda symbioses. Pages 589- 609 in Symbiosis (J. Seckbach, ed.) Kiuwer, Dordrecht. 274

BIBLIOGRAPHY

Akaike, H. 1974. A new look at the statistical model identification. IEEE Trans. Automatic Control 19:716-723.

Alfaro, M. E., S. Zoller, and F. Lutzoni. 2003. Bayes or bootstrap? A simulation study comparing the performance of Bayesian markov chain monte carlo sampling and bootstrapping in assessing phylogenetic confidence. Molecular Biology and Evolution 20:255-266.

Artjariyasripong, S., J. I. Mitchell, N. L. Hywel-Jones, and E. B. Garath Jones. 2001. Relationships of the genus Cordyceps and related genera, based on parsimony and spectral analysis of partial 1 8S and 28S ribosomal gene sequences. Mycoscience 42:503-517.

Baker, R. H., and R. DeSalle. 1997. Multiple sources of character information and the phylogeny of Hawaiian drosophilids. Systematic Biology 46:654-673.

Baker, R. H., X. Yu, and R. Desalle. 1998. Assessing the relative contribution of molecular and morphological characters in simultaneous analysis trees. Molecular Phylogenetics and Evolution 9:427-436.

Barker, F. K., and F. Lutzoni. 2002. The utility of the incongruence length difference test. Systematic Biology 51:625-637.

Barret, P. M., and K. J. Willis. 2000. Did dinosaurs invent flowers? Dinosaur- angiosperm coevolution revisited. Biol. Rev. Camb. Philos. Soc. 76:411-447.

Barrett, M., M. Donoghue, and E. Sober. 1991. Against consensus. Systematic Zoology 40:486-493.

Barron, G. L., and A. H. S. Onions. 1966. Verticillium chiamydosporium and its relationships to Diheterospora, Stemphyliopsis, and Paecilomyces. Canadian Journal of Botany 44:861-869.

Barron, G. L. 1977. Predators and parasites of microscopic animals. Pages 167-200 In Biology of conidial fungi, Vol. 2 (Cole, G. T. & B. Kendrick, eds). Academic Press, New York.

Barron, G. L. 1985. Fungal parasites of bdelloid rotifers: Diheterospora. Canadian Journal of Botany 63:211-222. 275

Barron, G. L. 1991. A new genus, Rotferophthora, to accommodate the Diheterospore-likeendoparasites of rotifers. Canadian Journal of Botany 69: 494-502.

Broughton, R. E., S. B. Stanley, and R. T. Durett. 2000. Quantification of homoplasy for nucleotide transitions and transversions and a reexamination of assumptions in weighted phylogenetic analysis. Systematic Biology 49: 617- 627.

Berbee, M. L., and J. W. Taylor. 1993. Dating the evolutionary radiations of the true fungi. Canadian Journal of Botany 71:1114-1127.

Bichoff, J. F., P. Chaverri, and J. F. White, Jr. 2005. Clarification of the host substrate of Ascopolyporus and description of Ascopolyporus philodendrus sp. nov. Mycologia in press.

Bissett, J. 1983. Notes on Tolypocladium and related genera. Canadian Journal of Botany 61:1311-1329.

Blackwell, M., and R. L. Gilbertson. 1984. New information on Cordycepioideus bisporus and Cordycepioideus octosporus. Mycologia 76:763-765.

Braum, E. L., R. K. Jansen, B. Oxelman, M. Backlund, H. Lantz, and K.-J. Kim. 1994. Examining basal avian divergences with mitochondrial sequences: model complexity, taxon sampling, and sequence length. Systematic Biology 5 1:614- 625.

Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42:795-803.

Bremer, K. 1994. Branch support and tree stability. Cladistics 10:295-304.

Bremer, K. 2000. Early Cretaceous lineages of monocot flowering plants. PNAS 97:4707-4711.

Bremer, K. 2002. Gondwanan evolution of the grass alliance of families (Poales). Evolution 56:1374-1387.

Bright, D. E. 1993. Systematics of bark beetles. Pages 23-36 in Beetle-pathogen interactions in conifer forests. (T. D. Schowalter, and G. M. Filip, eds.). Academic press, London.

Buckley, T. R. 2002. Model misspecification and probabilistic tests of topology: Evidence from empirical data sets. Systematic Biology 51:509-523. 276

Bull, J. J., J. P. Huelsenbeck, C. W. Cunningham, D. L. Swofford, and P. J. Waddell. 1993. Partitioning and combining data in phylogenetic analysis. Systematic Biology 42:384-397.

Cameron, K. M., M. W. Chase, W. M. Whitten, P. J. Kores, D. C. Jarrell, V. A. Albert, T. Yukawa, H. G. Hillis, and D. H. Goldman. 1999. A phylogenetic analysis of the Orchidaceae: evidence from rbcL nucleotide sequences. Am. J. Bot. 86:208-224.

Carroll, G. C. 1988. Fungal endophytes in stems and leaves: from latent pathogen to mutualistic symbiont. Ecology 69:2-9.

Castlebury, L. A., A. Y. Rossman, G.-H. Sung, A. S. Hyten, and J. W. Spatafora. 2004. Multigene phylogeny reveals new lineage for Stachybotrys chartarum, the indoor air fungus. Mycological Research 108:864-872.

Chaverri, P., G. J. Samuels, and K. T. Hodge. 2005. The genus Podocrella and its nematode-killing anamorph Harposporium. Mycologia 97:433-443.

Chaverri, P., J. F. Bischoff, H. C. Evans, and K. T. Hodge. 2006. Regiocrella, a new entomopathogenic genus with a pycnidial anamorph and its phylogenetic placement in the Clavicipitaceae. Mycologia in press.

Chippindale, P. T., and J. J. Wiens. 1994. Weighting, partitioning, and combining characters in phylogenetic analysis. Systematic Biology 43:278-287.

Clay, K. 1988. Clavicipitaceous endophytes of grasses. Coevolution and the change from parasitism to mutualism. Pages 79-105 In Coevolution of Fungi with Plants and Animals (Hawksworth, D. L. & K. Pirozynski, eds). Academic Press, London.

Clay, K., and C. L. Schardl. 2002. Evolutionary origins and ecological consequences of endophyte symbiosis with grasses. The American Naturalist 160:S99-S 127.

Cook, L. G., P. J. Gullan, and H. E. Trueman. 2002. A preliminary phylogeny of the scale insects (Hemiptera: Sternorrhyncha: Coccoidea) based on nuclear small- subunit ribosomal DNA. Molecular Phylogenetics and Evolution 25:43-52.

Coyle, F. A., P. A. Goloboff, and R. A. Samson. 1990. Actinopus trapdoor spiders (Araneae, Actinopodidae) killed by the fungus, Nomuraea atypicola (Deuteromycotina). Acta. Zool. Fenri. 190:89-93. 277

Crepet, W. L., and G. D. Feldman. 1991. The earliest remains of grasses in the fossil record. American Journal of Botany 78:1010-1014.

Cruickshank, R. D., and K. Ko. 2003. Geology of an amber locality in the Hukawng Valley, northern Myanmar. Journal of Asian Earth Sciences 21:441-455.

Cummings, M. P., S. A. Handley, D. S. Myers, D. L. Reed, A. Rokas, and K. Winka. 2003. Comparing bootstrap and posterior probability values in the four-taxon case. Systematic Biology 52:477-487.

Currie, C. R., B. Wong, A. E. Stuart, T. R. Schultz, S. A. Rehner, U. G. Mueller, G.-H. Sung, J. W. Spatafora, and N. A. Straus. 2003. Ancient tripartite coevolution in the attine ant-microbe symbiosis. Science 299:386-388.

Daghlian, C. P. 1981. A review of the fossil record of monocotyledons. Bot. Rev. 47:517-555.

Danforth, B. N., S. G. Brady, S. D. Sipes, and A. Pearson. 2004. Single-copy nuclear genes recover Cretaceous-age divergences in bees. Systematic Biology 53:309- 326.

Darlu, P., and G. Lecointre. 2002. When does the incongruence length difference test fail? Molecular Biolology Evolution 19:432-437.

Davies, T. J., T. G. Barraclough, M. W. Chase, P. S. Soltis, D. E. Soltis, and V. Savolainen. 2004. Darwin's abominable mystery: Insights from a supertree of the angiosperms. PNAS 101:1904-1909.

De Hoog, G. S. 1972. The genera Beauveria, Isaria, Tritirachium and Acrodontium gen. nov. Studies in Mycology No 1. de Queiroz, A. 1993. For consensus (sometimes). Systematic Biology 42:368-372.

Diehl, W. W. 1950. Balansia and the Balansiae in America. Agricultural monograph No.4. U. S. D. A. Washington.

Dmitriew, V. Y., and A. G. Ponomarenko. 2002. Dynamics of insect taxonomic diversity. Pages 517 in History of insects. (A. P. Rasnitsyn, and D. L. J. Quicke, eds.). Kluwer Academic, Dordrecht, the Netherlands.

Doaudy, C. J., F. Delsuc, Y. Boucher, W. F. Doolittle, and E. J. P. Douzery. 2003a. Comparison of the Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Molecular Biology and Evolution 20:248-254. 278

Doaudy, C. J., M. Dosay, M. S. Shivji, and M. J. Stanhope. 2003b. Molecular phylogenetic evidence refuting the hypothesis of Batoidea (rays and skates) as derived sharks. Molecular Phylogenetics and Evolution 26:215-221.

Dolphin, K., R. Beishaw, D. L. C. Orme, and D. L. J. Quicke. 2000. Noise and incongruence: Interpreting results of the incongruence length difference test. Molecular Phylogenetics and Evolution 17:401-406.

Dowton, M., and A. D. Austin. 2002. Increased congruence dose not necessarily indicate increased phylogenetic accuracy-The behavior of the incongruence length difference test in mixed-model analyses. Systematic Biology 51:19-31.

Driver, F., R. J. Milner, and J. W. H. Trueman. 2000. A taxonomic revision of Metarhiziumbased on a phylogenetic analysis of rDNA sequence data. Mycological Research 104:134-150.

Earle, F. S. 1901. Collections of Alabama fungi. Pages 150-263inPlant life of Alabama. Contr. U. S. Nat!. Herb. (C. Mohr, ed.).

Engstrom, T. N., H. B. Shaffer, and W. P. McCord. 2004. Multiple data sets, high homoplasy, and the phylogeny of softshell turtles. Systematic Biology 53:693- 710.

Eriksson, 0. 1982. Outline of the ascomycetes-1982. Mycotaxon 15:203-248.

Eriksson, 0. E., and D. L. Hawksworth. 1986. Notes on ascomycete systematics. Nos 1-224. Systema Ascomycetum 5:114-174.

Erixon, P., B. Svennblad, T. Britton, and B. Oxelman. 2003. Reliability of bayesian probabilities and bootstrap frequencies in phylogenetics. Systematic Biology 52:665-673.

Evans, H. C. 2003. Use of clavicipitalean fungi for the biological control of arthropod pests. Pages 5 17-548inClavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Evans, H. C., and R. A. Samson. 1982. Entomogenous fungi from the Galapagos islands. Canadian Journal of Botany 60:2325-233 3.

Farrell, B. D. 1998. "Inordinate fondness" explained: Why are there so many beetles? Science 281:555-559. 279

Farris, J. S., A. G. Kallersjo, A. G. Kiuge, and C. Bult. 1994. Testing significance of incongruence. Cladistics 10:315-319.

Felsenstein, J. 1981. A likelihood approach to character weighting and what it tells us about parsimony and compatibility. Bio. J. Linn. Soc. 16:183-196.

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

Friis, E. M., K. R. Pedersen, and P. R. Crane. 2001. Fossil evidence of water lilies in the Early Cretaceous. Nature 4 10:357-360.

Gams, W. 1971.Cephalosporium-artigeSchmmelpilze (Hyphomycetes). G. Fischer, Stuttgart.Gams, W. 1988. A contribution to the knowledge of nematophagous species of Verticillium. Neth. J. P1. Pathol. 94:123-148.

Gams, W and A. V. Zaayen. 1982. Contribution to the taxonomy and pathogenicity of fungicolous Verticillium species. I. Taxonomy. Neth. J. P1. Path. 88:57-78.

Gams, W., and R. Zare. 2001. A revision of Verticillium sect. Prostrata. III. Generic classification. Nova Hedwigia 72:329-337.

Gams, W., and R. Zare. 2003. A taxonomic review of the clavicipitaceous anamorphs parasitizing nematodes and other microinvertebrates. Pages 17-73 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Gams, W., K. O'Donnell, H.-J. Schroers, and M. Christensen. 1998. Generic classification of some more hyphomycetes with solitary conidia borne on phialides. Canadian Journal of Botany 76:1570-1583.

Gams, W., K. T. Hodge, R. A. Samson, R. P. Korf, and K. A. Seifert. 2005. Proposal to conserve the name Isaria (anamorphic fungi) with a conserved type. Taxon 54:537.

Gandolfo, M. A., K. C. Nixon, and W. L. Crepet. 2000. Monocotyledons: A review of their early Cretaceous record.in Monocots: systematics and evolution. (K. L. Wilson, and D. A. Morrison, eds.). CSIIRO, Melbourne.

Gandolfo, M. A., K. C. Nixon, and W. L. Crepet. 2002. Triuridaceae fossil flowers from the Upper Cretaceous of New Jersey. American Journal of Botany 89:1940-1957. Gardes, M and T. D. Bruns. 1993. ITS primers with enhanced specificity for basidiomycetes-application to the identification of mycorrhizae and rusts. Molecular Ecology Ecology 2:113-118.

Gatesy, J., G. Amato, M. Norell, R. DeSalle, and C. Hayashi. 2003. Combined support for wholesale taxic atavism in gavialine crocodylians. Systematic Biology 52:403-422.

Gatesy, J., P. O'Grady, and R. H. Baker. 1999. Corroboration among data sets in simultaneous analysis: Hidden support for phylogenetic relationships among higher-level artiodactyl taxa. Cladistics 15:271-3 13.

Gaunt, M. W., and M. A. Miles. 2002. An insect molecular clock dates the origin of the insects and accords with palaeontological and biogeographic landmarks. Molecular Biology and Evolution 19:748-761.

Gernandt, D. S., and J. K. Stone. 1999. Phylogenetic analysis of nuclear ribosomal DNA places the nematode parasite, Drechmeria coniospora, in Clavicipitaceae. Mycologia 9 1:993-1000.

Glenn, A. E., C. W. Bacon, R. Price, and R. T. Hanlin. 1996. Molecular phylogeny of Acremonium and its taxonomic implications. Mycologia 88:369-383.

Gontcharov, A. A., B. Mann, and M. Melkonian. 2004. Are combined analyses better than single gene phylogenies? A case study using SSU rDNA and rbcL sequence comparisons in the Zygnematophyceae (Streptophyta). Molecular Biology and Evolution 21:612-624.

GPWG. 2001. Phylogeny and subfamilial classfication of the grasses (Poaceae). Ann. Mo. Bot. Gard. 88:373-457.

Graybeal, A. 1998. Is it better to add taxa or characters to a difficult phylogenetic problem? Systematic Biology 47:9-17.

Greuter, W., J. McNeill, F. R. Barrie, H.-M. Burdt, V. Demoulin, T. S. Filgueiras, D. H. Nicolson, P. C. Silva, J. E. Skog, P. Trehane, N. J. Turland, and D. L. Hawksworth. 2000. Internationsl Code of Botanical Nomenclature (St Louis Code). Konigstein: Koeltz Scientific Books.

Grimaldi, D. A. 2000. A diverse fauna of Neuropterodea in amber from the Cretaceous of New Jersey. Pages 259-303 in Studies on fossils in amber, with particular referrence to the Cretaceous of New Jersey. (D. A. Grimaldi, ed.) Backhuys Publishers, Leiden, the Netherlands. 281

Grimaldi, D., and M. S. Engel. 2005. Evolution of the insects. Cambridge Univ. Press, Hong Kong, China.

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

Hawksworth, D. L., and A. Y. Rossman. 1997. Where are all the undescribed fungi? Phytopathology 87:888-891.

Heckman, D. S., D. M. Geiser, B. R. Eidell, R. L. Stauffer, N. L. Kardos, and S. B. Hedges. 2001. Molecular evidence for the early colonization of land by fungi and Plants. Science 293:1129-1133.

Hedges, S., J. Blair, M. Venturi, and J. Shoe. 2004. A molecular timescale of eukaryote evolution and the rise of complex multicellular life. BMC Evolutionary Biology 4:2-10.

Hillis, D. M., and J. J. Bull. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Systematic Biology 42:182-192.

Hodge, K. T. 1998. Revisionary studies in Hirsutella (Anamorphic Hypocreales: Clavicipitaceae). Ph. D thesis in Cornell University, Ithaca, NY.

Hodge, K. T. 2003. Clacicipitaceous anamorphs. Pages 75-123 in Clavicipitalean fungi: Evolutionary biology, Chemistry, Biocontrol and Cultural impacts (J. F. White, Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora, eds.). Marcel Dekker, Inc., New York.

Hodge, K. T., S. B. Krasnoff, and R. A. Humber. 1996. Tolypocladium inflatum is the anamorph of Cordyceps subsessilis. Mycologia 88:715-719.

Huelsenbeck, J. P., and B. Rannala. 1997. Phylogenetic methods corn of age: Testing hypotheses in an evolutionary context. Science 276:227-232.

Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.

Huelsenbeck, J. P., B. Larget, and D. Swofford. 2000. A compound poisson process for relaxing the molecular clock. Genetics 154:1879-1892.

Huelsenbeck, J. P., B. Larget, R. B. Miller, and F. Ronquist. 2002. Potential applications and pitfalls of bayesian inference of phylogeny. Systematic Biology 5 1:673-688. 282

Huelsenbeck, J. P., F. Ronquist, R. Nielsen, and J. P. Boilback. 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294:2310-2314.

Huelsenbeck, J. P., F. Ronquist, R. Nielsen, and J. P. Bollback. 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294:2310-2314.

Huelsenbeck, J. P., J. J. Bull, and 119. 1996. A likelihood ratio test to detect conflicting phylogenetic signal. Systematic Biology 45:92-98.

Hywel-Jones, N. L. 1993. Torrubiella luteorostrata: a pathogen of scale insects and its association with Paecilomyces cinnamomeus with a note on Torrubiella tenuis. Mycological Research 97:1126-1130.

Hywel-Jones, N. L. 1994. Cordyceps khaoyaiensis and C. pseudomilitaris, two new pathogens of lepidopteran larvae from Thailand. Mycological Research 98:939-942.

Hywel-Jones, N. L. 1995a. Cordyceps sphecocephala and a Hymenostilbe sp. infecting wasps and bees in Thailand. Mycological Research 99:154-158.

Hywel-Jones, N. L. 1 995b. Hymenostilbe ventricosa sp. nov. a pathogen of cockroaches in Thailand. Mycological Research 99:1201-1204.

Hywel-Jones, N. L. 1996. Cordyceps myrmecophila-like fungi infecting ants in the leaf litter of tropical forest in Thailand. Mycological Research 100:613-619.

Hywel-Jones, N. L. 2002. Multiples of eight in Cordyceps ascospores. Mycological Research 106:2-3.

Hywel-Jones, N. L., and G. J. Samuels. 1998. Three species of Hypocrella with large stromata pathogenic on scale insects. Mycologia 90:36-46.

Hywel-Jones, N. L., and H. C. Evans. 1993. Taxonomy and ecology of Hypocrella discoidea and its anamorph, Ashersonia samoensis. Mycological Research 97:871-876.

Hywel-Jones, N. L., and S. Sivichai. 1995. Cordyceps cylindrica and its association with Nomuraea atypicola in Thailand. Mycological Research 99:809-8 12.

Jun, Y., P. D. Bridge, and H. C. Evans. 1991. An integrated approach to the taxonomy of the genus Verticillium. Journal of General Microbiology 137:1437-1444. 283

Kamyschko, 0. P. 1962. De monilialibus terrestribe novis notula. Bot. Mater. Notul. Syst. Sect. Cryptogram. Inst. Bot. Acad. Sci. SSSR, 15:137-141.

Kauff, F., and F. Lutzoni. 2002. Phylogeny of the Gyalectales and Ostropales (Ascomycota, Fungi): among and within order relationships based on nuclear ribosomal RNA small and large subunits. Molecular Phylogenetics and Evolution 25:138-156.

Kerry, B. R. and D. H. Crump. 1977. Observations on fungal parasites of females and eggs of the cereal cyst-nematode, Heterodera avenae, and other cyst nematodes. Nematologica 23:193-20 1.

Kluge, A. G. 1989. A concern for evidence and a phylognetic hypothesis of relationships among Epicrates (Boidea, Serpentes). Systematic Zoology 38:7- 25.

Kluge, A. G. 1998. Total evidence or taxonomic congruence: cladistics or consensus classification. Cladistics 14:151-158.

Kluge, A. G., and A. J. Wolf. 1993. Cladistics: What's in a word? Cladistics 9:183- 199.

Kobayasi, Y and D. Shimizu 1982. Cordyceps species from Japan 5. Bull. Natn. Sci. Mus., Tokyo, Ser. B. 8:111-123.

Kobayasi, Y and D. Shimizu 1983. Cordyceps species from Japan 6. Bull. Natn. Sci. Mus., Tokyo, Ser. B. 9:1-21.

Kobayasi, Y. 1941. The genus Cordyceps and its allies. Science reports of the Tokyo Bunrika Daigaku (Section B, no. 84) 5:53-260.

Kobayasi, Y. 1981. Revision of the genus Cordyceps and its allies 1. Bulletin of the National Science Museum, Tokyo, series B 7:1-13.

Kobayasi, Y. 1982. Keys to the taxa of the genera Cordyceps and Torrubiella. Transactions of the Mycological Society of Japan 23:329-364.

Kobayasi, Y., and D. Shimizu. 1960. Monographic studies of Cordyceps 1. Group parasitc on Elaphomyces. Bulletin of the National Science Museum, Tokyo, Ser. B. 5:69-85. Kobayasi, Y., and D. Shimizu. 1963. Monographic studies of Cordyceps 2. Group parasitic on Cicadae. Bulletin of the National Science Museum, Tokyo, Ser. B. 6:286-314.

Kobayasi, Y., and D. Shimizu. 1982. Monograph of the genus Torrubiella. Bulletin of the National Science Museum, Tokyo, Ser. B. 8:43-78.

Koepfli, K.-P., and R. K. Wayne. 2003. Type I STS markers are more informative than cytochrome b in phylogenetic reconstruction of the Mustelidae (Mammalia: Carnivora). Systematic Biology 52:571-593.

Koetja, J. 2004. Scale insects (Hemiptera: Coccinea) from Cretaceous Myanmar (Burmese) amber. Journal of Systematic Paleontology 2:109-114.

Kolaczkowski, B., and J. W. Thorton. 2004. Performance of maximum parsimony and likelihood phylogenetics when evolution is heterogeneous. Nature 431:980- 984.

Koteja, A. 1999. Eomatsucoccus andrewi sp. nov. (Hemiptera: Sternorrhyncha: Coccinea) from the Lower Cretaceous of southern England. Cretaceous Research 20:863-866.

Koteja, A. 2000. Advances in the study of fossil coccids (Hemiptera: Coccinea). Polskie Pismo Entomologiczne 69:187-218.

Labandeira, C. C. 1997. Insect mouthparts: Acertaining the paleobiology of Insect feeding stratagies. Annu. Rev. Ecol. Syst. 28:153-193.

Labandeira, C. C., and J. J. Sepkoski. 1993. Insect diversity in the fossil record. Science 261:310-315.

Labandeira, C. C., K. R. Johnson, and P. Wilf. 2002. Impact of the terminal Cretaceous event on plant-insect associations. PNAS 99:2061-2066.

Landvik, S., 0. E. Eriksson, and M. L. Berbee. 2001. Neolecta-a fungal dinosaur? Evidence from beta-tubulin amino acid sequences. Mycologia 93:1151-1163.

Larget, B., and D. Simon. 1999. Markov chain Monte Carlo algorithms for the Bayesian analysis of phylogenetic trees. Mol Biol Evol 16:750-759.

Lemmon, A. R., and E. C. Moriarty. 2004. The importance of proper model assumption in Bayesian pylogenetics. Systematic Biology 53:265-277. 285

Levasseur, C., and F.-J. Lapointe. 2001. War and peace in phylogenetics: A rejoinder on total evidence and consensus. Systematic Biology 50:881-891.

Leuchtmann, D. Chung, D. Penny, and M. R. Siegel. 1997. Coevolution by common descent of fungal symbionts (Epichloe spp.) and grass hosts. Molecular Biology and Evolution 14:133-143.

Li, Z.-Z., C.-R. Li, B. Huang, and M.-Z. Fan. 2001. Discovery and demonstration of the teleomorph of Beauveria bassiana (Bals.) Vuill., an important entomopathogenic fungus. Chinese Science Bulletin 9:751-753.

Liang, Z.-Q., A.-Y. Liu, and J.-L. Liu. 1991. A new species of the genus Cordyceps and its Metarhizium anamorph. Acta Mycologica Sinica 10:257-262.

Lidgard, S., and P. R. Crane. 1988. Quantitative analysis of early angiosperm radiation. Nature 331:344-346.

Lim, T. K and Z. W. N, Wan. 1983. Mycoparasitism of the coffee rust pathogen, Hemileia vastatrix, by Verticillium psalliotae in Malaysia. Pertanika. 6:23-25.

Liu, Z. Y., Z. Q. Liang, A. J. S. Whalley, Y.-J. Yao, and A. Y. Liu. 2001. Cordycpes brittlebankisoides, a new pathogen of grubs and its anamorph, Metarhizium anisopliae var. majus. Journal of Invertebrate Pathology 78:178-182.

Liu, Z.-Y., Z.-Q. Liang, A.-Y. Liu, Y.-J. Yao, K. D. Hyde, and Z.-N. Yu. 2002. Molecular evidence for teleomorph-anamorph connections in Cordyceps based on ITS-5.8S rDNA sequences. Mycological Research 106:1100-1108.

Lockhart, P. J., M. A. Stell, M. D. Hendy, and D. Penny. 1994. Recovering evolutionary trees under a more realistic model of sequence evolution. Molecular Biology and Evolution 11:605-612.

Luangsa-ard, J. J., N. L. Hywel-Jones, and R. A. Samson. 2004. The polyphyletic nature of Paecilomyces sensu lato based on 18S-generated rDNA phylogeny. Mycologia 96:773-780.

Luangsa-Ard, J. J., N. L. Hywel-Jones, L. Manoch, and R. A. Samson. 2005. On the relationships of Paecilomyces sect. Isarioidea species. Mycological Research 109:58 1-5 89.

Luttrel, E. 5. 1955. The ascostromatic ascomycetes. Mycologia 47:511-532.

Lutzoni, F., F. Kauff, C. J. Cox, D. McLaughlin, G. Celio, B. Dentinger, M. Padamsee, D. Hibbett, T. Y. James, E. Baloch, M. Grube, V. Reeb, V. Hofstetter, C. Schoch, A. E. Arnold, J. Miadlikowska, J. Spatafora, D. Johnson, S. Hambleton, M. Crockett, R. Shoemaker, G.-H. Sung, R. Lucking, T. Lumbsch, K. O'Donnell, M. Binder, P. Diederich, D. Ertz, C. Gueidan, K. Hansen, R. C. Harris, K. Hosaka, Y.-W. Lim, B. Matheny, H. Nishida, D. Pfister, J. Rogers, A. Rossman, I. Schmitt, H. Sipman, J. Stone, J. Sugiyama, R. Yahr, and R. Vilgalys. 2004. Assembling the fungal tree of life: progress, classification, and evolution of subcellular traits. Am. J. Bot. 91:1446-1480.

Macleod, D. M. 1954. Investigations on the genera Beauveria Vuill. and Tritirachium Limber. Canadian Journal of Botany 32:818-890.

Maddison, D. R., and W. P. Maddison. 2000. MacClade 4: Analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts.

Mains, E. B. 1947. New and interesting species of Cordyceps. Mycologia 39:535-545.

Mains, B. B. 1949. New species of Torrubiella, Hirsutella and Gibellula. Mycologia 41:303-310.

Mains, E. B. 1950. Entomogenous species of Akanthomyces, Hymenostilbe and Insecticola in North America. Mycologia 42:566-589.

Mains, E. B. 1957. Species of Cordyceps parasitic on Elaphomyces. Bulletin of The Torrey Botanical Club. 84:243-251.

Mains, E. B. 1958. North American entomogenous species of Cordyceps. Mycologia 50:169-222.

Mains, B. B. 1959. North American species of Aschersonia parasiticon Aleyrodidae. J. Insect. Pathol 1:43-47.

Mankau, R. 1980. Biocontrol: fungi as nematode control agents. Journal of Nematology. 12:213-232.

Mason-Gamer, R. J., and E. A. Kellogg. 1996. Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Systematic Biology 45:524-545.

Massee, G. 1895. A revision of the genus Cordyceps. Annals of Botany 9:1-44.

Messmer, R., W. Schweigkofler, M. Ibi, G. Berg, and H. Prillinger. 1996. Molecular characterizattion of the plant pathogen Verticillium dahliae Kieb. using RAPD- PCR and sequencing of the I 8SrRNA-gene. Journal of Phytopathology 144:347-3 54. Mitchell, A., C. Mitter, and J. C. Regier. 2000. More taxa or more characters revisited: combining data from nuclear-protein encoding genes for phylogenetic analyses of Noctuoidea (Insecta: Lepidoptera). Systematic Biology 49:202-224.

Miyamoto, M. M., and W. M. Fitch. 1995. Testing species phylogenies and phylogenetic methods with congruence. Systematic Biology 44:64-76.

Moncalvo, J.M., F.M. Lutzoni, S.A. Rehner, J. Johnson & R. Vilgalys. 2000. Phylogenetic relationships of agaric fungi based on nuclear large subunit ribosomal DNA sequences. Systematic Biology 49:278-306.

Moran, V. C., and T. R. C. Southwood. 1982. The guild composition of arthropod commumities in trees. Journal of Animal Ecology 5 1:289-301.

Morgans-Jones, G. and W. Gams. 1982. Notes on Hyphomycetes XLI. An endophyte of Festuca arundinacea and the anamorph of Epichloe typhina, new taxa in one of two new sections of Acremonium. Mycotaxon. 15:311-318.

Mugnai, L., P. D. Bridge, and H. C. Evans. 1989. A chemotaxonomic evaluation of the genus Beauveria. Mycological Research 92:199-209.

Mullis, K. B. and F. A. Faloona. 1987. Specific synthesis of DNA in vitro viaa polymerase-catalyzed chain reaction. Methods in Enzymology 155:335-350.

Nannfeldt, J. A. 1932. Studien uber die morphologie und systematik der nicht- lichenisierten, inoperculaten Discomyceten. Nova Acta R. Soc. Scient. upsal. ser 6 8:1-368.

Nikoh, N., and T. Fukatsu. 2000. Interkingdom host jumping underground: Phylogenetic analysis of entomopathogenic fungi of the genus Cordyceps. Molecular Biology and Evolution 17:629-63 8.

Nylander, J. A. A., F. Ronquist, J. P. Huelsenbeck, and N.-A. J. Luis. 2004. Bayesian phylogenetic analysis of combined data. Systematic Biology 53:47-67.

Obornik, M., M. Jirku, and D. Dolezel. 2001. Phylogeny of mitosporic entomopathogenic fungi: Is the genus Paecilomyces polyphyletic? Canadian Journal of Microbiology 47:813-819.

Ochiel, G. S., H. C. Evans, and J. Eilenberg. 1997. Cordycepioideus,a pathogen of termites in Kenya. Mycologist 11:7-9. Pacioni, G., and G. Frizzi. 1978. Paecilomycesfarinosus, the conidial state of Cordyceps memorabilis. Canadian Journal of Botany 56:391-394.

Padovan, A. C. B., G. F. 0. Sanson, A. Brunstein, and M. R. S. Briones. 2005. Fungi evolution revisited: application of the penalized likelihood method to a bayesian fungal phylogeny provides a new perspective on phylogenetic relationships and divergence dates of ascomycota groups. Journal of Molecular Evolution 60:726-735.

Pagel, M. 1999. Inferring the historical patterns of biological evolution. Nature 401:877-884.

Pagel, M., A. Meade, and D. Barker. 2004. Bayesian estimation of ancestral character states on phylogenies. Systematic Biology 53:673-684.

Pazoutova, S. 2001. The phylogeny and evolution of the genus Claviceps. Mycological Research 105:275-283.

Percy, D. M., R. D. E. Page, and Q. C. B. Cronk. 2004. Plant-insect interactions: Double-dating associated insect and plant lineages reveals asynchronous radiations. Systematic Biology 53:120-127.

Petch, T. 1921. Studies in entomogenous fungi. II. The genera of Hypocrella and Aschersonia. Ann. R. Bot. Gard (Peradeniya) 7:167-278.

Petch, T. 1931. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 16:55-75.

Petch, T. 1932. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 17:209-245.

Petch, T. 1933. Notes on entomogenous fungi. Trans. Br. Mycol. Soc. 18:48-75.

Phylippe, H., E. A. Snell, E. Bapteste, P. Lopez, P. W. H. Holland, and D. Casane. 2004. Phylogenomics of eukaryotes: impact of missing data on large alignments. Molecular Biology and Evolution2 1:1740-1752.

Poinar, G. 0. J. 2006. Early fossil evidence of fungal parasitism of insects. Mycologia in preparation.

Poinar, G. 0. J., and G. M. Thomas. 1984. A fossil entomogenous fungus from Dominican Amber. Experientia 40:578-579.

Poinar, J. G. 0. 2004. Programinis burmitis gen. et sp. nov., and P. laminatussp. nov., early cretaceous grass-like monocots in Burmese amber. Australian Systematic Botany 17:497-504. Posada, D., and K. A. Crandall. 1998. ModelTest: Testing the model of DNA substitution. Bioinformatics 14:817-818.

Qian, H. and R. E. Ricklefs. 2000. Large-scale processes and the Asian bias in species diversity of temperate plants. Nature 407:180-1 82.

Reeb, V., F. Lutzoni, and C. Roux. 2004. Contribution of RPB2 to multilocus phylogenetic studies of the euascomycetes (Pezizomycota, Fungi) with special emphasis on the lichen-forming Acarosporaceae and evolution of polyspory. Molecular Phylogenetics and Evolution 32:1036-1060.

Rehner, S. A., and E. Buckley. 2005. A Beauveria phylogeny inferred from nuclear ITS and EF1-a sequences: evidence for cryptic diversification and links to Cordyceps teleomorphs. Mycologia 97:84-98.

Rehner, S. A., and G. J. Samuels. 1995. Molecular systematics of the Hypocreales:a teleomorph gene phylogeny and the status of their anamorphs. Canadian Journal of Botany 73 (Suppi. 1):S816-S823.

Reynolds, D. R., and J. W. Taylor (eds) 1993. The fungal homomorph: Mitotic, meitic and pleomorphic speciation in fungal systematics. CAB International, Wallingford, UK.

Rodrigo, A. G., M. Kelly-Borges, P. R. Bergquist, and P. L. Bergquist. 1993. A randomization test of the null hypothesis that two cladograms are sample estimates of a parametric phylogenetic tree. N. Z. J. Bot. 31:257-268.

Rogerson, C. T. 1970. The hypocrealean fungi (Ascomycetes, Hypocreales). Mycologia 62:865-9 10.

Rokas, A., and S. B. Carroll. 2005. More Genes or More Taxa? The Relative Contribution of Gene Number and Taxon Number to Phylogenetic Accuracy. Molecular Biology and Evolution 22:1337-1344.

Rombach, M. C., R. A. Humber, and D. W. Roberts. 1986. Metarhiziumfiavoviride var. minus, var. nov., a pathogen of plant-and leathoppers on rice in the phillippines and solomon islands. Mycotaxon 27:87-92.

Rosenberg, M. S., and S. Kumar. 2001. Incomplete taxon sampling is nota problem for phylogenetic inference. PNAS 98:10751-10756. 290

Rossman, A. Y., G. J. Samuels, J. S. Rogers, and R. Lowen. 1999. Genera of Bionectriaceae, Nectriaceae and Hypocreaceae (Hypocreales, Ascomycetes). Studies in Mycology No. 42.

Saikkonen, K., S. H. Faith, M. Helander, and T. J. Sullivan. 1998. Fungal endophytes: A continuum interactions with host plants. Annu. Rev. Ecol. Syst. 19:319-343.

Saksirirat, W., and Hoppe, H.-J. 1991. Secretion of extracellular enzymes by Verticillium psalliotae Treschow and Verticillium lecanii (Zimmi.) Viégas during growth on uredospores of the soybean rust fungus (Phakopsora pachyrhizi Syd.) in liquid cultures. Journal of Phytopathology 13 1:161-173.

Samson, R. A. 1974. Paecilomyces and some allied hyphomycetes. Studies in Mycology No 6.

Samson, R. A., and H. C. Evans. 1975. Notes on entomogenous fungi from Ghana III. The genus Hymenostilbe. Proc. Kon. Ned. Akad. Wetensch. C 78:73-79.

Samson, R. A., E. S. Van Reenen-Hoekstra, and H. C. Evans. 1989. New species of Torrubiella (Ascomycotina: Clavicipitales) on insects from Ghana. Studies in Mycology 3 1:123-132.

Samson, R. A., H. C. Evans, and J.-P. Latge. 1988. Atlas of entomopathogenic fungi. Springer-Verlag, Berlin Heidelberg New York.

Samuels, G. J. 1988. Fungicolous, lichenicolous, and myxomyceticolous species of Hypocreopsis, Nectriopsis, Nectria, Peristomialis, and Trichonectria. Mem. N. Y. Bot. Gdn. 48:1-78.

Sanderson, M. 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Molecular Biology and Evolution 14:1218-123 1.

Sanderson, M. J. 2002. Estimating absolute rates of molecular evolution and divergence Times: A penalized likelihood approach. Molecular Biology and Evolution 19:101-109.

Sanderson, M. J. 2003. r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinformatics 19:301- 302.

Sanderson, M. J., and J. A. Doyle. 2001. Sources of error and confidence intervals in estimating the age of angiosperms from rbcL and 1 8S rDNA data. American Journal of Botany 88:1499-15 16. 291

Schroers, H.-J. 2001. A monograph of Bionectria (Ascomycota, Hypocreales, Bionectriaceae) and its Clonostachys anamorphs. Studies in Mycology No 46.

Schuler, T. M., Hommes, H.-P. Plate and G. Zimmermann. 1991. Verticillium lecanii (Zimmermann) Viégas (Hyphomycetales: Moniliaceae): Geschichte, Systematik, Verbreitung, Biologie und Anwendung im Pflanzenschutz. Biologische Bundesanstalt, Braunschweig.

Shimazu, M., W. Mitshhashi, and H. Hashimoto. 1988. Cordyceps brongniartii sp. nov., the teleomorph of Beauveria brongniartii. Trans. Br. Mycol. Soc. 29:323-330.

Simmons, M. P., K. M. Pickett, and M. Miya. 2004. How meaningful are Bayesian support values? Molecular Biology and Evolution 21:188-199.

Soltis, P. S., and D. E. Soltis. 2004. The origin and diversification of angiosperms. American Journal of Botany 91:1614-1626.

Soltis, P. S., D. E. Soltis, V. Savolainen, P. R. Crane, and T. G. Barraclough. 2002. Rate heterogeneity among lineages of tracheophytes: Integration of molecular and fossil data and evidence for molecular living fossils. PNAS 99:4430-4435.

Sorenson, M. D. 1998. TreeRot, version 2. Univ. Michigan, Ann Arbor.

Spatafora, J. W. 2002. Evolution of Ascomycota-arthropoda symbioses. Pages 589- 609 in Symbiosis (J. Seckbach, ed.) Kiuwer, Dordrecht.

Spatafora, J. W., and M. Blackwell. 1993. Molecular systematics of unitunicate perithecial ascomycetes I. The Clavicipitales-Hypocreales connection. Mycologia 85:912-922.

Spatafora, J. W., B. Volkmann-Kohlmeyer, and J. Kohlmeyer. 1998. Independent terrestrial origins of the Halosphaeriales (marine Ascomycota). American Journal of Botany 85:1569-1580.

Spatafora, J. W., G.-H. Sung, J.-M. Sung, N. L. Hywel-Jones, and J. F. White, Jr. 2006. Phylogenetic evidence for an animal pathogen origin for ergot and thegrass endophytes. PNAS submitted.

Stensrud, 0., N. L. Hywel-Jones, and T. Schumacher. 2005. Towardsa phylogenetic classification of Cordyceps: ITS nrDNA sequence data confirm divergent lineages and paraphyly. Mycological Research 109:41-56. 292

Strong, D. R., J. H. Lawton, and T. R. E. Southwood. 1984. Insects on plants. Harvard Univ. Press, Cambridge, MA.

Suh, S.-O., H. Noda, and M. Blackwell. 2001. Insect symbiosis: Derivation of yeast- like endosymbionts within an entomopathogenic filamentous lineage. Molecular Biology and Evolution 18:995-1000.

Suh, S.-O., J. W. Spatafora, G. R. S. Ochiel, H. C. Evans, and M. Blackwell. 1998. Molecular phylogenetic study of a termite pathogen Cordycepioideus bisporus. Mycologia 90:611-617.

Sullivan, J. 1996. Combining data with different distributions of among-site variation. Systematic Biology 45:375-3 80.

Sullivan, R. F., G. F. Bills, N. L. Hywel-Jones, and J. F. White, Jr. 2000. Hyperdermium: a new clavicipitalean genus for some tropical epibionts of dicotyledonous plants. Mycologia 92:908-918.

Sullivan, R., M. S. Bergen, R. Patel, G. F. Bills, S. C. Alderman, J. W. Spatafora, and J. F. White, Jr. 2001. Neoclaviceps monostza: features of an enigmatic clavicipitalean fungus and the phyletic status of the anamorphic genus Ephelis. Mycologia 93:90-99.

Sung, G.-H., and J. W. Spatafora. 2004. Cordyceps cardinalis sp. nov., a new species of Cordyceps with an east Asian-eastern North American distribution. Mycologia 96:658-666.

Sung, G.-H., J. W. Spatafora, R. Zare, K. T. Hodge, and W. Gams. 2001. A revision of Verticillium sect. Prostrata. II. Phylogenetic analyses of SSU and LSU nuclear rDNA sequences from anamorphs and teleomorphs of the Clavicipitaceae. Nova Hedwigia 72:311-328.

Sung, J. 1996. The insects-born fungus of Korea in color. Kyohak Publ. Co., Seoul.

Suzuki, Y., G. V. Glazko, and M. Nei. 2002. Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. PNAS 99:16138-16143.

Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.Ob 10. Sinauer, Sunderland, Massachusetts.

Tanaka, E., C. Tanaka, A. Gafur, and M. Tsuda. 2002. Heteroepichloe, gen. nov. (Clavicipitaceae; Ascomycotina) on bamboo plants in East Asia. Mycoscience 43:87-93. 293

Taylor, J. W. 1993. A comtemporary view of the holomorph: nucleic acid sequence and computer databases are changing fungal classification. Pages 3-13 In The Fungal Holomorph: Mitotic, Meiotic and Pleomorphic Speciation in Fungal Systematics (Reynolds, D.R. & J.W. Taylor, eds): 3-13. CAB International, Wallingford, UK.

Taylor, J. W. 1995. Making the Deuteromycota redundant: a practical integration of mitosporic and meiosporic fungi. Canadian Journal of Botany 73:S754-S759.

Taylor, T. N., H. Hass, and H. Kerp. 1999. The oldest fossil ascomycetes. Nature 399:648.

Taylor, T. N., H. Hass, H. Kerp, M. Krings, and R. T. Hanlin. 2005. Perithecial ascomycetes from the 400 million year old Rhynie chert: an example of ancestral polymorphism. Mycologia 97:269-285.

Tehler, A., J. S. Farris, D. L. Lipscomb, and M. Kallersjo. 2000. Phylogenetic analyses of the fungi based on large rDNA data sets. Mycologia 92:459-474.

Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673-4680.

Thorne, J. L., and H. Kishino. 2002. Divergence time and evolutionary rate estimation with multilocus data. Systematic Biology 51:689-702.

Thornton, J. W., and R. DeSalle. 2000. A new method to localize and test the significance of incongruence: Detecting domain shuffling in the nuclear receptor superfamily. Systematic Biology 49:183-201.

Tiffney, B. 1985. Perspectives on the origin of the floristic similarity between eastern Asia and eastern North America. J. Arnold Arboretum 66: 73-94.

Verhaar, M. A., T. Hijwegen, and J. C. Zadoks. 1996. Glasshouse experiments on biocontrol of cucumber powdery mildew (Sphaerothecafuliginea) by the mycoparasites Verticillium lecanii and Sporothrix rugulosa. Biol. Control 6:353-360.

Vestergaard, S. A., T. Gillespie, T. M. Butt, G. Schreiter, and J. Eilenberg. 1995. Pathogenicity of the hyphomycete fungi Verticillium lecanii and Metarhizium anisopliae to the western flower thrips, Franklin iella occidentalis. Biocontrol Sci. Technol. 5:185-192. 294

Vilgalys, R., and B. L. Sun. 1994. Ancient and recent patterns of geographic speciation in the oyster mushroom Pleurotus revealed by phylogenetic analysis of ribosomal DNA sequences. PNAS 91:4599-4603.

Wehmeyer, L. E. 1976. The pyrenomycetes fungi. Mycol. Mem. 6. Cremer, Lehre.

Wen, J. 1999. Evolution of eastern Asian and eastern North American disjunct distributions in flowering plants. Annu. Rev. Ecol. Syst. 30:421-455.

Whalley, P. E. S. 1985. The systematics and palaeogeography of teh Lower Jurrasic insects of Dorset, England. Bulletin of the British Museum (Natural History) (Geology) 39:107-189.

Wheeler, W. C. 1990. Combinatorial weights in phylogenetic anlaysis: a statistical parsimony procedure. Cladistics 6:269-275.

White, J. F., Jr., C. W. Bacon, N. L. Hywel-Jones, and J. W. Spatafora. 2003. Historical perspectives: Human interactions with Clavicipitalean fungi. Pages 1-16 in Clavicipitalean Fungi (J. F. White, Jr., M. Backlund, N. L. Hywel- Jones, and J. W. Spatafora, eds.). Marcel Decker, New York.

White, J. J. F., and P. V. Reddy. 1998. Examination of structure and molecular phylogenetic relationships of some graminicolous symbionts ingenera Epichloëand Parepichloe. Mycologia 90:226-234.

Whittingham, L. A., B. Slikas, D. W. Winkler, and F. H. Sheldon. 2002. Phylogeny of the tree swallow genus, Tachycineta (Ayes: Hirundinidae), by Bayesian analysis of mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 22:430-441.

Wiens, J. J. 1998. Combining data sets with different phylogenetic histories. Systematic Biology 47:568-58 1.

Wiens, J. J. 2003. Missing data, Incomplete taxa, and phylogeneticaccuracy. Systematic Biology 52:528-538.

Wiens, J. J. 2003 a. Incomplete taxa, incomplete characters, and phylogenetic accuracy: Is there a missing data problem? Journal of Vertebrate Paleontology 23: 297-3 10.

Wiens, J. J. 2003b. Missing data, Incomplete taxa, and phylogeneticaccuracy. Systematic Biology 52:528-538. 295

Wing, S. L. 2000. Evolution and expansion of flowering plants. Pages 308 in Phanerozoic terrestrial ecosystems (R. A. Gastaldo, and W. A. DiMichele, eds.). The Paleontological society, Pittsburgh.

Wood, S. L. 1982. The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Nat. Mem. 6:37-39.

Wu, Q.-X. and G. M. Mueller. 1997. Biogeographic relationships between the macrofungi of temperate eastern Asia and eastern North America. Canadian Journal of Botany 75: 2108-2116.

Wu, Q.-X., G. M. Mueller, F. M. Lutzoni, Y.-Q. Huang, and S.-Y. Guo. 2000. Phylogenetic and biogeographic relationships of eastern Asian and eastern North American disjunct Suillus species (Fungi) as inferred from nuclear ribosomal RNA ITS sequences. Molecular Phylogenetics and Evolution 17:37-47.

Yang, Z., and A. D. Yoder. 2003. Comparison of likelihood and Bayesian methods for estimating devergence times using multiple gene loci and calibration points with application to a radiation of cute-lookingmouse lemur species. Systematic Biology 52:705-7 16.

Yoder, A. D., J. A. Irwin, and B. A. Payseur. 2001. Failure of the ILD to determine data combinability for slow loris phylogeny. Systematic Biology 50:408-424.

Zaayen, A. V., and W. Gams. 1982. Contribution to the taxonomy and pathogenicity of fungicolous Verticillium species. II. Pathogenicity. Neth. J. P1. Pathol. 88: 143-154.

Zang, M., and N. Kinjo. 1998. Notes on the alpine Cordyceps of China and nearby nations. Mycotaxon 66:215-229.

Zare, R., and W. Gams. 2001a. A revision of Verticillium sect. Prostrata.IV. The genera Lecanicillium. Nova Hedwigia 73:1-50.

Zare, R., and W. Gams. 200 lb. A revision of Verticillium sect. Prostrata. VI.The genus Haptocillium. Nova Hedwigia 73:271-292.

Zare, R., and W. Gams. A revision of Verticillium sect. Prostrata IV. Thegenera Lecanicillium and Simplicillium gen. nov. Nova Hedwigia 73:51-86.

Zare, R., W. Gams, and A. Culham. 2000. A revision of Verticilliumsect. Prostrata. I. Phylogenetic studies using ITS sequences. Nova Hedwigia 71:465-480. 296

Zare, R., W. Gams, and H. C. Evans. 2001. A revision of Verticillium sect. Prostrata. V. The genus Pochonia, with notes on Rotferophthora. Nova Hedwigia 73:5 1- 86.

Zwickl, D. J., and D. M. Hillis. 2002. Increased taxon sampling greatly reduces phylogenetic error. Systematic Biology 51:588-598.

Top View