HOST SPECIFICITY and PHYLOGENETIC RELATIONSHIPS AMONG TILLETIA SPECIES INFECTING WHEAT and OTHER COOL SEASON GRASSES by XIAODO

Home , Agropyron fragile, Ryegrass bunt, Tilletia, Tilletiales, Tilletia caries, Tilletia controversa

HOST SPECIFICITY AND PHYLOGENETIC RELATIONSHIPS

AMONG TILLETIA SPECIES INFECTING WHEAT AND

OTHER COOL SEASON GRASSES

XIAODONG BAO

A dissertation submitted in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

WASHINGTON STATE UNIVERSITY Department of Plant Pathology

DECEMBER 2010

To the Faculty of Washington State University:

The members of the Committee appointed to examine the dissertation of XIAODONG BAO find it satisfactory and recommend that it be accepted.

Lori M. Carris, Ph.D., Chair

Tobin L. Peever, Ph.D.

Jack D. Rogers, Ph.D.

Scot H. Hulbert, Ph.D.

ACKNOWLEDGEMENTS

I would like to express the deepest gratitude to my major advisor Dr. Lori M. Carris, for

her persistent guidance, support and encouragement which make it possible to this dissertation.

Her enthusiasm for research and excitement in teaching continuously provide inspiration and

motivation for my academic goals. I would like to give the most sincere thanks to the members

of my dissertation committee, Drs. Tobin L. Peever, Jack D. Rogers and Scot H. Hulbert, for

their insightful suggestions to my research and critical review of the dissertation.

I am heartily thankful to Dr. Lisa A. Castlebury (USDA-ARS, Baltimore, Maryland) for generously providing sequencing facilities and guidance for my research. I appreciate Mr. Blair J.

Goates (USDA-ARS, Aberdeen, ID) for hosting our visit to National Small Grains Collections and providing a treasure of wheat bunt collections for my research. My thanks also go to Drs.

Kálmán Vánky (Herbarium Ustilaginales Vánky, Germany), Lennart Johnsson (Plant Pathology

and Biocontrol Unit, SLU, Sweden), Veronika Dumalasová (Research Institute of Crop

Production, Czech Republic) and Denis A. Gaudet (Lethbridge Research Centre, Canada), for

many of the collections used in the analyses.

I would also like to thank Dr. Martin I. Chilvers, Dr. Hajime Akamatsu, Jane E. Stewart,

Dr. Chuntao Yin, Janet G. Matanguihan and Jeremiah Dung, for their unselfish support on my

research and writing.

iii

I’m thankful to the Department of Plant Pathology in Washington State University for providing great staff and facilities to support my Ph.D. program. I thank NSF grant 0542603 for providing financial support on my research.

Special thanks to my family and all my friends for their love, understanding and support

whenever I need it.

HOST SPECIFICITY AND PHYLOGENETIC RELATIONSHIPS

AMONG TILLETIA SPECIES INFECTING WHEAT AND

OTHER COOL SEASON GRASSES

Abstract

by Xiaodong Bao, Ph. D. Washington State University December 2010

Chair: Lori M. Carris

The 140 species of Tilletia (Ustilaginomycotina, Basidiomycota) are biotrophic pathogens that

cause bunt and smut diseases of grasses (Poaceae). Among the most economically important species are the wheat bunt pathogens, T. caries, T. contraversa, and T. laevis. Most smut fungi infect relatively few hosts, but the host range of T. contraversa includes 65 species in 17 grass genera. The objective of this study was to incorporate sequence data from ITS rDNA, eukaryotic translation elongation factor 1-alpha (EF1-alpha), and the second largest subunit of RNA polymerase II (RPB2) with three anonymous loci (A13, A16 and P18) to elucidate relationships and test the host range of the wheat bunt fungi and related Tilletia species infecting cool-season grasses in the Pacific Northwestern U.S. Sorus shape and teliospore size of 60 dwarf and common bunt isolates analyzed with k-means clustering revealed two groups, one largely corresponding to common bunt and the second to dwarf bunt. Network analysis based on three anonymous loci and RBP2 region revealed three clades, but none of the clades contained isolates of all three species. Phylogenetic analyses of the three anonymous loci, EF1-alpha, ITS rDNA

v and RPB2 were used to test the conspecific status of Eurasian and North American isolates of T. contraversa infecting species of nine grass genera. Phylogenetic trees were reconstructed independently for RPB2/ITS/EF1-alpha and A16/P18/ITS/EF1-alpha loci. NeighborNet networks were estimated from A16/P18/ITS/EF1-alpha loci and compared with the NeighborNet networks constructed with “gene-jackknifing” method. All phylogenetic trees and networks suggested a wide host range for T. contraversa. The analyses failed to support dwarf bunt pathogen as genetically distinct from the common bunt pathogens, but Eurasian isolates of T. contraversa on Elymus repens and Thinopyrum intermedium were consistently placed in a clade distinct from the majority of T. contraversa isolates, suggesting that substructure exists among isolates in this group. In contrast to the broad host range supported for T. contraversa, phylogenetic analysis of Tilletia species on cool-season grasses based on ITS, EF1-alpha and

RPB2 revealed well-supported clades corresponding to host. A new species, T. puccinelliae, was proposed for a bunt infecting Puccinellia distans (alkaligrass) in Washington.

TABLE OF CONTENTS

Page ACKNOWLEDGEMENTS ...... iii

ABSTRACT ...... v

LIST OF TABLES ...... ix

LIST OF FIGURES ...... x

CHAPTER

1. GENERAL INTRODUCTION ...... 1

Literature cited ...... 5

2. TILLETIA PUCCINELLIAE, A NEW SPECIES OF RETICULATE-SPORED BUNT

FUNGUS INFECTING PUCCINELLIA DISTANS ...... 8

Abstract ...... 8

Introduction ...... 8

Results ...... 23

Taxonomy ...... 27

Discussion ...... 28

Acknowledgements ...... 32

Affiliation of Co-authors ...... 33

Literature cited ...... 33

3. A MULTILOCUS PHYLOGENETIC APPROACH TO TESTING SPECIES LIMITS IN

THE WHEAT BUNT FUNGI TILLETIA CARIES, T. CONTRAVERSA

AND T. LAEVIS ...... 36

Abstract ...... 36

vii

Introduction ...... 37

Materials and Methods ...... 41

Results ...... 57

Discussion ...... 69

Literature cited ...... 75

4. THE HOST RANGE AND SPECIES CONCEPT OF TILLETIA CONTRAVERSA ...... 83

Abstract ...... 83

Introduction ...... 84

Materials and Methods ...... 93

Results ...... 99

Discussion ...... 114

Literature cited ...... 118

ATTRIBUTION ...... 125

APPENDIX ...... 126

viii

LIST OF TABLES

CHAPTER 2

Table 2-1. Isolates and DNA sequences used in Chapter 2...... 10

Table 2-2. Host, infection type, morphology and germination of Tilletia species compared.

...... 16

CHAPTER 3

Table 3-1. Tilletia isolates used in Chapter 3...... 42

Table 3-2. Species, teliospore morphology and morphological groups determined for Tilletia

isolates...... 54

Table 3-2. Likelihood tests of alternative phylogenetic hypotheses ...... 68

CHAPTER 4

Table 4-1. Isolates used in Chapter 4...... 89

Table 4-2. Spore morphology of isolates used in phylogenetic analysis ...... 97

Table 4-3. Primers, PCR annealing temperature, amplified length and number of

polymorphic sites for loci used in study...... 101

Table 4-4. The compatibility in placement of each taxon in NeighborNet tree when

including different loci in analysis...... 106

APPENDIX

Table S-1. Host range for T. contraversa (also as T. brevifaciens) reported in US Pacific

Northwest, 1950-1958...... 130

Table S-2. Isolates studied as part of Chapter 4 for which sequence data could not be

obtained ...... 134

LIST OF FIGURES

CHAPTER2

FIG. 2-1 a, b. Tilletia puccinelliae (WSP 71469, holotype) sori in seed of Puccinellia distans

...... 15

FIG. 2-2. Tilletia spp. associated with seed of Puccinellia distans ...... 21

FIG. 2-3. Tilletia puccinelliae (WSP 71469, holotype) ...... 25

FIG 2-4. Maximum likelihood tree of 30 taxa of Tilletia spp...... 26

CHAPTER 3

FIG. 3-1. Teliospores (a-f) and sori (g-h) of wheat bunts ...... 38

FIG. 3-2. K-means clustering of 51 isolates of wheat bunt on Triticum aestivum ...... 53

FIG. 3-3. Maximum parsimony tree inferred from concatenated data of 3 anonymous loci . 61

FIG 3-4. Statistical parsimony network reconstructed from 3 anonymous loci and RPB2

region...... 64

FIG 3-5. Statistical parsimony network reconstructed with gene jackknifing and

NeighborNet network estimated from 4 loci ...... 66

CHAPTER 4

FIG. 4-1. Maximum likelihood tree of T. contraversa isolates and related species with

sequences from A16, P18, ITS and EF1-alpha loci ...... 102

FIG. 4-2. NeighborNet networks constructed from different sequences dataset with 49

common and dwarf bunt isolates ...... 104

FIG. 4-3. Maximum likelihood tree of T. contraversa isolates and related species with RPB2,

EF1-alpha and ITS loci...... 109

APPENDIX

FIG. S-1. The scatter plot distribution of morphological characters of each wheat bunt isolate..

...... 126

FIG. S-2. Maximum likelihood tree of 60 wheat bunt isolates reconstructed from three

anonymous loci or from RPB2-primer-amplified region...... 129

FIG. S-3. Parametric bootstrap tests of alternative hypothesis on RPB2, EF1-alpha and ITS

tree...... 135

FIG. S-4. Maximum likelihood tree of the T. contraversa (T. brevifaciens) isolates and

related species with individual or combined gene trees ...... 143

CHAPTER ONE

GENERAL INTRODUCTION

Approximately 140 recognized species of Tilletia (Ustilaginomycotina, Basidiomycota)

cause bunt and smut diseases of grasses (Poaceae) (Vánky 1994). Most of the species cause bunt

diseases in which the sorus, containing teliospores and sterile cells, completely or partially replaces the developing ovary in the grass host, but nine species form sori in vegetative portions

of the host. Tilletia caries, the type species of the genus, causes common bunt, one of the most

destructive wheat diseases in human history (Durán and Fischer 1961, Goates 1996). The genus was named to honor M. Mathieu Tillet by the Tulasne brothers in 1847 (Durán and Fischer 1961).

Tillet conducted a series of experiments in 1752-1753 that demonstrated wheat bunt resulted

from contamination of seed with “bunt dust”, a remarkable conclusion considering that fungi

were not yet recognized as the cause of plant disease (Tillet 1755). Prevost (1807) was the first to

make a direct connection between the spores (“dark dust”) and wheat bunt.

Tilletia caries, T. contraversa, and T. laevis are closely related, morphological species

causing wheat bunt diseases. Tilletia contraversa, recognized as the dwarf bunt pathogen of wheat,

is distinguished from T. caries and T. laevis by its ability to stunt the host, formation of reticulately

ornamented teliospores enveloped in a thick gelatinous sheath, and a requirement for low

temperature (optimum 3-8 C) and light for germination of teliospores. Germination typically

occurs in 3-6 weeks. Tilletia caries and T. laevis, the common bunt pathogens of wheat, do not

cause host stunting and have teliospores that germinate at 15C within a week in the absence of light.

Teliospores of T. caries are reticulate and lack a conspicuous gelatinous sheath, while teliospores

of T. laevis are smooth. Common bunt is present throughout wheat growing areas of the world, whereas dwarf bunt is limited to those regions with persistent snow cover (Goates 1996). The

Pacific Northwestern US is an important wheat growing center of the world, and the common and dwarf bunt pathogens caused substantial yield losses prior to the development of effective control measures for common bunt in the 1950s, and for dwarf bunt in the 1990s (Hoffmann

1982, Trione 1982, Sitton et al 1993). Many of the seminal studies on the wheat bunt fungi were conducted in the Pacific Northwest, including those that demonstrated sexual compatibility among the three wheat bunt species (Flor 1932, Silbernagel 1964, Pimentel et al 2000).

Teliospores of the common bunt pathogens are seed- or soil-borne, and infection occurs

through the coleoptile shortly after seed germination (Goates 1996). Common bunt of wheat has

largely been controlled in North America by the combined use of resistant cultivars and

fungicide seed treatments since the late 1950s (Goates 1996), but remains a problem in many wheat production areas around the world, and has reemerged as an important disease in organic wheat production in Europe. Dwarf bunt of wheat, with soil-borne teliospores and infection through tiller initials, was not effectively controlled in North America until the release of the systemic fungicide difenoconazole (Sitton et al 1993, Goates 1996). When Tilletia contraversa

was quarantined by China in 1973, the Pacific Northwest lost a major export market for wheat because of the widespread contamination of wheat with dwarf bunt spores (Trione 1982). An

agreement between the U.S. and China in 1999 established acceptable tolerance levels for T.

contraversa teliospores in wheat, and limited export of Pacific Northwest wheat to China

resumed (Carris 2010).

Species recognition in Tilletia evolved over the last century as species concepts and criteria

used to delimit species of smut fungi were refined based on a better understanding of fungal

systematics (Fischer and Shaw 1953, Durán and Fischer 1961, Vánky 1991, Russell and Mills

1994). Morphological species concept (MSC), biological species concept (BSC), and phylogenetic

species concept (PSC) are the three most commonly used concepts in fungal systematics

(Harrington and Rizzo 1999). There is no agreement so far on what is the best species concept that

should be applied to fungal systems (Harrington and Rizzo 1999, Taylor et al 2000, Hey 2006,

Giraud et al 2008). As in other fungi, most species of smut fungi were described based on

morphological characters. However, the host has also played an important role in the description of

new smut species. For example, many of the 200 species of Tilletia described prior to Durán and

Fischer’s 1961 monograph were based on their occurrence on a previously unreported host. Durán

and Fischer (1961) used a numerical taxonomic approach, and based on the examination of

hundreds of collections, recognized 72 species from the 200 that had been described. Molecular

data were not used in smut systematic studies until 1984 (Piepenbring 2003), and the first

molecular systematic study of the genus Tilletia was published in 2005 (Castlebury et al 2005).

The combined use of morphological species and phylogenetic species concepts has provided

evidence that most species of smut fungi are highly specialized to host genus or species (Begerow

et al 2004, Stoll et al 2005, Carris et al 2007). Among approximately 600 smut species listed in

European Smut Fungi (Vánky 1994), Begerow et al (2004) found that 55% of the species occur on

a single host, 86% on five or fewer hosts, and 93% on ten or fewer hosts. In Tilletia, host association between smut species and host genus/species has been supported by multilocus phylogenetic analysis based on sequence data from ITS rDNA, elongation factor 1 alpha (EF1α), and a portion of the second largest subunit of RNA polymerase II (RPB2) (Carris et al 2007). In that study, isolates of T. contraversa were placed in two distinct clades corresponding to Triticum

(wheat) or Thinopyrum (wheatgrass) as host. The dwarf bunt pathogen T. contraversa has been

reported on over 60 host species (Purdy et al 1963). In the Pacific Northwest in the 1950s, a

serious outbreak of dwarf bunt was reported on Tualatin oat grass (Arrhenatherum elatius) and

other cultivated grasses, expanding the known host range for this fungus (Hardison and Corden

1952, Hardison 1954). However, the genetic backgrounds of the dwarf bunt isolates on wheat and

other genera of grasses have not been directly compared. It is not known if the grass-infecting

isolates are genetically similar to the wheat dwarf bunt pathogen, if they represent a specialized

biotype, or a distinct species.

In brief, based on the criteria used for morphological species recognition, three closely

related but distinct species, T. caries, T. laevis and T. contraversa, co-occur on wheat without

apparent post-zygotic isolation, and a wide range of grass hosts are infected by T. contraversa.

These concepts are in conflict with the understanding gained from phylogenetic analyses that

support Tilletia species as having narrow host specificity. In Chapter 2, a bunt fungus infecting

Puccinellia distans (alkaligrass) that is closely related to the wheat bunt pathogens is supported

as a new species based on morphological characters and a multilocus phylogenetic analysis. In

Chapter 3, molecular phylogenetic approaches based on sequence data from three anonymous

loci and a region of RPB2 are used to test the hypothesis that multilocus phylogenetic trees or

networks can provide evidence to support the morphological species concept of the wheat bunt

pathogens. In Chapter 4 the host association and phylogenetic lineages within 63 dwarf bunt

isolates collected on 17 hosts throughout Eurasia and North America since 1876 are tested with

its closely related species using sequence information on three anonymous locus and ITS, RPB2

and EF1-alpha regions.

LITERATURE CITED

Begerow D, Göker M, Lutz M, Stoll M. 2004. On the evolution of smut fungi and their hosts. in:

Frontiers in Basidiomycote Mycology. pp. 81-98 p. Agerer R, Piepenbring M, Blanz P,

eds. IHW‐Verlag & Verlagsbuchhandlung, Germany.

Carris LM. 2010. Dwarf Bunt. in: Compendium of Wheat Diseases and Pests, Third Edition. 171

p. Bockus WW, Bowden RL, Hunger RM, Morrill WL, Murray TD, Smiley RW, eds.

APS Press, St. Paul, Minnesota.

———, Castlebury LA, Huang G, Alderman SC, Luo J, Bao X. 2007. Tilletia vankyi, a new

species of reticulate-spored bunt fungus with non-conjugating basidiospores infecting

species of Festuca and Lolium. Mycol Res 111:1386-1398.

Castlebury LA, Carris LM, Vánky K. 2005. Phylogenetic analysis of Tilletia and allied genera in

order Tilletiales (Ustilaginomycetes; Exobasidiomycetidae) based on large subunit

nuclear rDNA sequences. Mycologia 97:888-900.

Durán R, Fischer GW. 1961. The genus Tilletia. Washington State University.

Fischer GW, Shaw CG. 1953. A proposed species concept in the smut fungi, with application to

North American species. Phytopathology 43:181-188.

Flor HH. 1932. Heterothallism and hybridization in Tilletia tritici and T. levis. Jour Agr Res

44:49-58.

Giraud T, Refregier G, Le Gac M, de Vienne DM, Hood ME. 2008. Speciation in fungi. Fungal

Genet Biol 45:791-802.

Goates BJ. 1996. Common bunt and dwarf bunt. in: Bunt and Smut Diseases of Wheat: Concepts

and Methods of Disease Management. pp. 12-25 p. Wilcoxson RD, Saari EE, eds.

CIMMYT, Mexico D.F.

Hardison JR. 1954. New grass host records and life history observations of dwarf bunt in eastern

Oregon. Plant Disease Reporter 38:345-347.

———, Corden ME. 1952. A serious outbreak of dwarf bunt in tall oat grass in Oregon. Plant

Disease Reporter 36:343-344.

Harrington TC, Rizzo DM. 1999. Defining Species in the Fungi. in: Structure and Dynamics of

Fungal Populations. 43-70 p. Worrall JJ, ed. Kluwer Academic Pub, Dordrecht.

Hey J. 2006. On the failure of modern species concepts. Trends Ecol Evol 21:447-450.

Hoffmann JA. 1982. Bunt of wheat. Plant Dis 66:979-986.

Piepenbring M. 2003. Smut Fungi (Ustilaginomycetes P.P. and Microbotryales, Basidiomycota).

Flora Neotropica 86:1-291.

Pimentel G, Carris LM, Peever TL. 2000. Characterization of interspecific hybrids between

Tilletia controversa and T. bromi. Mycologia 92:411-420.

Prévost IB. 1807. Mémoire sur las causa immédiate de la carie ou charbon des blés et de

plusieurs autres maladies des plantes, et sur les présvatifs de la caire. Paris. 80 p. English

translation by Keitt GW. 1939. Phytopathology Classics 6:1-95.

Purdy LH, Kendrick EL, Hoffmann JA, Holton CS. 1963. Dwarf bunt of wheat. Annual Reviews

in Microbiology 17:199-222.

Russell BW, Mills D. 1994. Morphological, physiological, and genetic evidence in support of a

conspecific status for Tilletia caries, T. controversa, and T. foetida. Phytopathology

84:576-582.

Silbernagel MJ. 1964. Compatibility between Tilletia caries and T. controversa. Phytopathology

54:1117-1120.

Sitton JW, Line RF, Waldher JT, Goates BJ. 1993. Difenoconazole seed treatment for control of

dwarf bunt of winter wheat. Plant Dis 77:1148-1151.

Stoll M, Begerow D, Oberwinkler F. 2005. Molecular phylogeny of Ustilago, Sporisorium, and

related taxa based on combined analyses of rDNA sequences. Mycol Res 109:342-356.

Taylor JW, Jacobson DJ, Kroken S, Kasuga T, Geiser DM, Hibbett DS, Fisher MC. 2000.

Phylogenetic species recognition and species concepts in fungi. Fungal Genet Biol

31:21-32.

Tillet M. 1755. Dissertation on the cause of the corruption and smutting of the Kernels of wheat

in the head and on the means of preventing these untoward circumstance. Bordeaux. 150

p. Translated by Humphrey HB. 1937. Phytopathological classics 5:1-191.

Trione EJ. 1982. Dwarf bunt of wheat and its importance in international wheat trade. Plant Dis

66:1083-1089.

Vánky K. 1991. Spore morphology in the taxonomy of Ustilaginales. Trans Mycol Soc Japan

32:381-400.

———. 1994. European smut fungi. Gustav Fischer Verlag, Stuttgart, New York.

CHAPTER TWO

TILLETIA PUCCINELLIAE, A NEW SPECIES OF RETICULATE-SPORED BUNT

FUNGUS INFECTING PUCCINELLIA DISTANS

Xiaodong Bao, Lori M. Carris, Guoming Huang, Jiafeng Luo, Yueting Liu, Lisa A. Castlebury

(Published in Mycologia (2010) 102:613-623)

ABSTRACT

A shipment of Fults alkaligrass seed (Puccinellia distans) grown in Washington State containing bunted florets was intercepted by quarantine officials at China’s Tianjin Entry-Exit Quarantine and

Inspection Bureau. The bunted florets were filled with irregularly shaped, reticulately ornamented teliospores that germinated in a manner characteristic of systemically-infecting Tilletia spp. on

grass hosts in subfamily Pooideae. Based on morphological characters and a multigene

phylogenetic analysis of the ITS region rDNA, eukaryotic translation elongation factor 1 alpha and

a region of the second largest subunit of RNA polymerase II including a putative intein, the

Puccinellia bunt is genetically distinct from known species of Tilletia and is proposed as a new

species, T. puccinelliae.

INTRODUCTION

In 2004, bunted florets (smut sori) were found in a shipment of U.S. Fults alkaligrass seed,

Puccinellia distans (Pooideae, Poaceae), by scientists at the Tianjin Entry-Exit Quarantine and

Inspection Bureau in Tanggu, China. The bunted florets were filled with teliospores with

ornamentation and germination characteristic of Tilletia (Ustilaginomycotina, Basidiomycota).

None of the approx. 150 species recognized in Tilletia is known to infect Puccinellia spp.; the only

smuts reported on Puccinellia are species of Entyloma, Urocystis and Ustilago (Durán and Fischer

1961; Fischer 1953; Vánky 1994; Guo and Zhang 2004). However, two grassy weeds present in

alkaligrass seed fields, Apera interrupta (windgrass) and Bromus tectorum (downy brome; cheatgrass), are hosts for Tilletia goloskokovii and T. bromi, respectively (Boyd and Carris 1997,

1998; Boyd et al 1998). The teliospores of T. bromi and T. goloskokovii are morphologically similar to each other (Boyd et al 1998) and to the spores in the bunted florets in the Fults alkaligrass seed.

A multigene phylogenetic analysis of the Puccinellia bunt, T. bromi, T. goloskokovii and similar species on hosts in subfamily Pooideae was conducted. Results of the analysis, and comparison of morphological characters of Tilletia spp. associated with cool season grass seed crops in the Pacific Northwestern U.S.A. are presented.

TABLE 2- 1. Isolates and DNA sequences used in Chapter 2.

Taxon Voucher/Culture Year Host Origin Collector GenBank accession numbers numbers ITS1 RPB2 EF1A Tilletia H.U.V. 20.802 2004 Thinopyrum Poland K. Vánky EU257565a EU257599a EU257535a brevifaciens G. intermedium W. Fischer Barkworth & D. R. Dewey × Elymus repens (L.) Gould T. brevifaciens WSP 68945 (Vánky 1999 Thinopyrum Austria K. Vánky EU257566a EU257600a EU257536a 412) intermedium T. bromi WSP 71271 (LMC 1992 Bromus tectorum L. US L. Carris EU257555a EU257592a EU257528a (Brockmuller) 148) (Washington) Brockmuller T. bromi WSP 71272 (LMC 1992 Bromus arvensis L. US L. Carris EU257556a EU257593a EU257529a 167; CBS 123002) (Wyoming ) T. bromi WSP 71273 (LMC 1991 Bromus hordeaceus L. US L. Carris EU257557a EU257594a EU257530a 10 75; CBS 123001) ssp. hordeaceus (Wyoming) T. caries (DC.) WSP 71304 (DAR 1997- Triticum aestivum L. Australia G. Murray EU257559a EU257596a EU257532a Tul. 73302) 1999 T. caries WSP 71303 (LMC 1995 Triticum aestivum Sweden L. EU257560a EU257597a EU257533a J-19; CBS 121951) Johnsson T. contraversa WSP 71280 (LMC 1994 Triticum aestivum US (Oregon) L. Carris EU257561a EU257598a EU257534a Kuhn 282; CBS 121952) T. contraversa WSP 69062 (Vánky 1984 Triticum aestivum Germany K. Vánky EU257562a EU257588a EU257526a 528) T. elymi Diet. & WSP 71274 (LMC 1992 Elymus glaucus Buckl. US L. Carris EU257564a EU257591a EU257527a Holw. 158; CBS 123000) ssp. glaucus (Wyoming) T. fusca Ell. & WSP 71275 (LMC 1971 Vulpia microstachys US L. Carris EU257567a EU257601a EU257537a Everh. 141; CBS 122991) (Nutt.) Munro var. (Washington) microstachys T. goloskokovii WSP 69687 (LMC 1995 Apera interrupta (L.) US L. Carris EU257569a EU257603a EU257539a 315; CBS 122995) Beauv. (Washington) T. goloskokovii WSP 71281 (LMC 1993 Apera interrupta US L. Carris EU257568a EU257602a EU257538a 238-2) (Washington)

T. laevis Kuhn WSP 71278 1971 Triticum aestivum US L. Carris EU257571a EU257605a EU257541a (LMC178) T. laevis WSP 71300 (Vánky 1988 Triticum aestivum Iran K. Vánky EU257573a EU257607a EU257543a 766; CBS 121950) T. laguri G. M. HUV 16.352 1992 Lagurus ovatus L. Italy K. Vánky EU257574a EU257608a Zhang, G. X. Lin & J. R. Deng T. lolii Auers.ex WSP 71298 (Vánky 1990 Lolium rigidum Iran K. Vánky EU257575a EU257609a EU257544a Rab. 767) Gaudin T. lolioli Vánky, WSP 71305 (Vánky 1990 Loliolum subulatum Iran K. Vánky EU257576a EU257610a EU257545a Carris, Castl. & 763) (Banks & Sol.) Eig H. Scholz. T. puccinelliae WSP 71469 2004 Puccinellia distans US L. Carris EU910057b EU910063b Carris, G. HOLOTYPE (NBS (Washington) EU910051b Huang & Castl. HF4 FUL3-2 205 P7; ex-type CBS 122993) 11 T. puccinelliae WSP 71465 (NBS 2004 Puccinellia distans US L. Carris EU910058b EU910064b EU910052b HF4 FUL3 seed 2; (Washington) CBS 122994) T. puccinelliae WSP 71472 (NBS 2004 Puccinellia distans US L. Carris EU910059b EU910065b EU910053b NE4 FUL 1-2 (Washington) palea/lemma; CBS 122996) T. puccinelliae WSP 71470 2004 Puccinellia distans US L. Carris EU910060b EU910066b EU910054b (NBS NE4 FUL 4-6 (Washington) sm intact; CBS 122997) T. puccinelliae WSP 71470 (NBS 2004 Puccinellia distans US L. Carris EU910061b EU910067b EU910055b NE4 FUL4-1 lg (Washington) broken; CBS 122998) T. puccinelliae WSP 71471 (LH52 2004 Puccinellia distans US L. Carris EU910056b EU910062b EU910050b 425B-9; CBS (Washington) 122999)

T. secalis (Cda.) WSP 71279 (LMC 1993 Secale cereale L. US (Idaho) L. Carris EU257577a EU257589a EU257525a Koern. 255) T. sphaerococca culture 2003 Agrostis stolonifera L. US (Chinese G. Huang EU257578a EU257611a EU257546a (Rab.) Fisch. v Intercept) Waldh. T. togwateei WSP 71277 (LMC 1992 Poa reflexa Vasey & US L. Carris EU257580a EU257613a EU257547a 169) Scribn. ex Vasey (Wyoming) T. trabutii Jacz. WSP 71299 (Vánky 1990 Hordeum murinum L. Iran K. Vánky EU257581a EU257614a EU257548a 764; CBS 122324) ssp. glaucum (Steud.) Tzvlev T. trabutii VPRI 32106 2005 Hordeum murinum L. Australia I. Pascoe EU257582a EU257615a EU257549a ssp. leporinum (Link) Arcang. T. vankyi Carris WSP 71266 1997 Lolium perenne L. Australia L. Carris EU257587a EU257620a EU257554a & Castlebury (V21-713) T. vankyi WSP 71270 (FF1-1; 2005 Festuca rubra L. ssp. US (Oregon) S. EU257585a EU257618a EU257552a CBS 122323) fallax (Thuill.) Nyman Alderman 12 a. Cited from Carris et al 2007 b. From this study

MATERIALS AND METHODS

Isolation, maintenance and deposition of cultures and voucher specimens.—Species used in this

study are listed in TABLE 2-1. Host names and authorities are listed as in the U.S. Department of

Agriculture PLANTS Database (http://plants.usda.gov). Voucher materials were deposited in Herb.

WSP (Washington State University, Pullman, WA), Herb. BPI (Beltsville, MD), and Herb. H.U.V.

(Tübingen, Germany). Representative cultures were deposited in Centraalbureau voor

Schimmelcultures (CBS; Utrecht, The Netherlands) (see TABLE 2-1). Seed samples ranging in

size from 25-430 g from ten seedlots of Fults alkaligrass grown in Franklin County, Washington

during 2002 (one sample) and 2004 (nine samples) were examined using a modification of the seed

wash protocol used for the US National Karnal Bunt Survey (Peterson et al 2000). Twenty-five grams of seed in 100 ml tap water with two drops of Tween-20 were placed in a 500 ml flask on a

rotary shaker at 200 rev min-1 for 10 min. The seed suspension was passed through a 53 µm-pore

sieve into a clean 600 ml beaker. The suspension from the beaker was poured through a 25

µm-pore sieve, and any spores and debris remaining on the sieve were washed into a 15 ml conical

centrifuge tube and pelleted by centrifugation for 3 min at 200 rev min-1. The supernatant was discarded and the pellet was suspended in 100 µl Shear’s mounting medium [50 ml 2% (w/v) potassium acetate, 20 ml glycerol, 30 ml 95% ethyl alcohol]. The suspension was transferred to microscope slides, a 22 × 50 mm coverslip was placed on each slide, and the slides were examined

at ×125 using a compound microscope. If reticulately ornamented teliospores were observed on

the microscope slides, seed were soaked in water overnight to render the palea and lemma

translucent, and seed examined under low power (×10) magnification to detect bunted florets (FIG.

2-1). Teliospores from bunted florets were mounted in Shear’s mounting medium and examined using differential interference contrast microscopy at ×1000. Teliospore diam (including exospore),

shape, color, thickness of exospore, number of meshes per spore diam, and sterile cell shape, color and thickness were recorded for 20 spores per sample (TABLE 2-2). For germination, teliospores were surface-sterilized in 0.26% NaClO (5% commercial bleach) in a 1.5 ml Eppendorf tube for 50 s, pelleted by centrifugation at approx. 13000 g in a benchtop microcentrifuge for 10 s and rinsed twice with sterile, distilled water. Surface-sterilized teliospores were streaked on 1.5% water agar

(WA) and incubated at 5, 15 C and room temperature (20–25 C). Primary basidiospores were transferred to M-19 agar (Trione 1964) to establish colonies for nucleic acid extraction. Cultures on M-19 were maintained at 15C in the dark.

FIG. 2-1 a, b. Tilletia puccinelliae (WSP 71469, holotype) sori in seed of Puccinellia distans; seed soaked overnight in water to render palea and lemma transparent. a. Single bunted floret. b.

Three bunted florets attached to rachis. Bar = 1 mm.

TABLE 2- 2. Host, infection type, morphology and germination of Tilletia species compared.

Tilletia Tilletia bromi a Tilletia goloskokovii Tilletia sphaerococca Tilletia vankyi a puccinelliae b a a Hosts Puccinellia distans Bromus Apera interrupta Agrostis stolonifera Lolium perenne, L. tectorum multiflorum, Festuca rubra Infection type systemic systemic systemic systemic local? Sori (mm) 1–1.5 × 0.5–0.8 1.5–4 × 0.5–1 1–1.5 × 0.5–0.75 0.75–1 × 0.3–0.5 1.5–3 × 1–1.5 Teliospores: Diam (µm) 18.5–31 × 15–28 17–24.5 17.5–30 21–30 17.5–30 Exospore depth 0.5–1.5(–2) 1–2 1–3 1.5–3.5 1–2 (µm)

16 Meshes/spore 6–10 (–14) 7–9 4–8 5–10 5–8 diam Color medium to dark medium to dark pale to dark fuscous pale yellow-brown to pale to medium reddish reddish brown reddish brown brown dark brown brown Shape subglobose to globose globose globose globose angular Germination 5C (10–12 d) 5–15C (7–21 d) 5–10C (7–24 d) 5C (22–30 d) 5C (>14 d), 15C (5–7 d)

Sterile Cells: Diam (µm) 11–17.5 × 10.5–23 15–28 9.5–23 10–21 10.5–16.5 Wall (µm) 1–2.5 1–2.5 1–2.5 1–2.5 0.5–2

Primary basidiospores: Shape filiform filiform filiform filiform filiform # /Basidium 9–16 9–18 22–26 26–32 24–40 Size (µm) 70–80 × 2.5 57–88 × 2--2.5 48.5–68 × 2--2.5 68.5–75.5 × 2–2.5 53–74 × 1.5–3.5 # Nuclei/ 1 1 1 1 1 basidiospore Conjugation yes yes yes yes no

a. Cited from Carris et al 2007

b. From this study 17

Nucleic acid extraction and PCR amplification.—DNA was extracted directly from actively

growing surface mycelium scraped from M-19 plates. DNA was extracted with the DNeasy Plant

Mini Kit (Qiagen Inc., Valencia, California) according to the manufacturer’s instructions using approximately 50 mg of fresh mycelium.

All amplifications were performed in 20 μL on a GeneAmp 9700 thermal cycler (Applied

Biosystems, Foster City, CA). Three PCR amplified nuclear DNA fragments, elongation factor 1 α

(EF1A), internal transcribed spacer 1 (ITS1) and the second largest subunit of RNA polymerase II

(RPB2) were used in this study as described by Carris et al (2007). The RPB2 region amplified was shown to be informative for phylogenetic studies of Tilletiales, and is more variable in sequence than EF1A and ITS1 regions (Carris et al 2007). Recent analysis discovered motif similarity between the RPB2 amplified regions from Tilletiales and other fungal intein coding regions

(Poulter et al 2007; unpublished). The primers were used with the following combinations: 1)

EF1-526F with EF1-1567R (Rehner, S. 2001. Primers for elongation factor 1-a (EF1-a). http://ocid.nacse.org/research/deephyphae/EF1primer.pdf.); 2) ITS5 or ITS1 with ITS4 (White et al 1990) as the reverse primer; 3) RPB2-740F and RPB2-1365R (Carris et al 2007).

The PCR products were purified with ExoSAP-IT (USB, Cleveland, OH) according to the manufacturer’s instructions and amplified with respective forward and reverse PCR primers using the BigDye version 3.1 dye terminator kit (Applied Biosystems, Foster City, CA). Those products were purified again with Performa DTR gel filtration cartridges (Edge BioSystems, Gaithersburg,

MD) and sequenced on an ABI 3100 automated DNA sequencer.

Sequence analysis.— The chromatographic outputs were assembled and edited into consensus sequences with SEQUENCHER version 4.5 for Windows (Gene Codes Corporation, Ann Arbor, MI).

Eighteen sequences of 3 regions from 6 isolates were generated in this study, which were

individually aligned to existing alignments of 21 isolates from Carris et al (2007), using

CLUSTALW as implemented in BIOEDIT v7.0.5.3 for Windows (Hall 1999). The resulting individual gene alignments were manually edited in BIOEDIT and concatenated to produce the final multiple gene dataset. All gaps in the data set were treated as missing data in this study.

Two methods were used to test for conflict among the genes: the partition homogeneity test

(incongruence length difference test) as implemented in PAUP (Swofford 2002) with 1000 replicates; and a reciprocal bootstrap analysis (Reeb et al 2004). Independently inferred topologies from each partition were examined for conflict (e.g., one monophyletic and the other paraphyletic).

Reciprocal 80% ML bootstrap (ML-BS) value or 95% Bayesian posterior probability (PP) were used as criteria for strongly supported clades.

The phylogenetic trees were inferred independently from each partition and the combined alignment by maximum likelihood (ML) with PAUP 4.0b10 as well as Bayesian method with

MRBAYES v3.1.2 for Windows (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck

2003). DNA substitution models were determined by MODELTEST 3.7 (Posada and Crandall 1998).

The best fit models and parameters were chosen with Akaile information criterion (AIC) and hierarchical likelihood ratio tests (hLRTs).

For ML analysis, heuristic search was chosen with 100 starting trees obtained via stepwise, random addition of sequences. The branch-swapping algorithm is tree bisection-reconnection

(TBR) on each tree. All aligned positions were included in the analysis. All characters are unordered and have equal weight. Specific substitution models chose determined by MODELTEST were used for each dataset as follows: 1) EF1A, TrN+I+G model with BASE=(0.2338 0.3171

0.2555) NST=6 RMAT=(1.0000 1.6571 1.0000 1.0000 9.3667) RATES=gamma SHAPE=0.7539

PINVAR=0.5776; 2)ITS1, TVM+I model with BASE=(0.2443 0.2249 0.2155) NST=6 RMAT=(4.9187

4.9305 0.4291 0.6961 4.9305) RATES=equal PINVAR=0.9013; 3) RPB2, TrN+G model with

BASE=(0.1801 0.2732 0.3165) NST=6 RMAT=(1.0000 4.1486 1.0000 1.0000 6.1422) RATES=gamma

SHAPE=0.1834 PINVAR=0; 4) EF1A + ITS1 + RPB2, TrN+I+G model with BASE=(0.2190 0.2754

0.2614) NST=6 RMAT=(1.0000 3.0572 1.0000 1.0000 6.7569) RATES=gamma SHAPE=0.7073

PINVAR=0.6948. Branch support was estimated using nonparametric bootstrap analysis with 1000

replicated samples. The major rule consensus trees were generated with ML bootstrap values

higher than or equal to 80%.

The Bayesian inference of phylogeny was performed with Monte Carlo Markov Chain

(MCMC) sampling method. Several different GTR models without parameters were used as the

priors in MRBAYES. The priors for each datasets were: 1) EF1A = GTR+I+G; 2) ITS1 = GTR+I; 3)

RPB2 = GTR+G; 4) the combined gene was unlinked in STATEFREQ, REVMAT, SHAPE and PINVAR,

so that each partition (gene) could be applied with the corresponding GTR model described above.

MCMC sampling was limited to 500 000 generations. Two independent runs were started

simultaneously, with four chains in each run (one heat chain and three cold chains). The frequency

of swap was set to 1. Trees were sampled at every 100 generation. The first 25% of sampled trees

were excluded from the analysis. Well-supported clades were inferred with a minimum of 95% of

Bayesian posterior probabilities.

FIG. 2-2. Tilletia spp. associated with seed of Puccinellia distans. a–c. T. puccinelliae. a. Bunted and healthy florets of Puccinellia distans. b. Teliospores, plane view. c. Teliospores, surface view

of reticulate ornamentation. d–f. T. goloskokovii. d. Bunted and healthy florets of Apera interrupta. e. Teliospores, plane view. f. Teliospores, surface view of reticulate ornamentation. g–i. T. bromi. g.

Bunted and healthy florets of Bromus tectorum. h. Teliospores, plane view. i. Teliospores, surface view of reticulate ornamentation. a, d, g. units = 1 mm; b, c, e, f, h, i bar = 10 µm.

RESULTS

Seed Survey.—Reticulately-ornamented teliospores were found in seed washes for 9 of the 10 seed

samples available from 2004. Number of teliospores ranged from 9 to approximately 3500 per 25 g

of seed. No teliospores were found in the 2002 sample. The number of bunted florets per 25 g seed in the samples in which teliospores were present ranged from 0 (with 9 and 34 teliospores/25 g of seed) to 121 (3500 teliospores/25 g of seed). All bunted florets were completely filled with teliospores, and several remnants of inflorescences with 2 to 3 bunted florets attached to the rachis were found (FIGS. 2-1, 2-2)

Morphology.—Morphological and germination characters for the Puccinellia bunt fungus and four other Tilletia spp. that either infect cultivated cool season grasses in the Pacific Northwestern

U.S. (T. sphaerococca and T. vankyi) or are likely to occur as contaminants in P. distans based on

the known distribution of their hosts (T. bromi and T. goloskokovii) are summarized in TABLE 2-2.

Phylogenetic analysis.— The combined alignment consisted of EF1A (721 bp), ITS1 (649 bp) and

RPB2 (587 bp), with a total of 1957 characters. Of those characters, 1724 characters are constant,

86 are variable but parsimony-uninformative, and 147 are parsimony-informative. A ML tree was

inferred by combined analysis of all three loci (FIG. 2-4). A clade containing all isolates of the

Puccinellia bunt was identified from the ML tree (ML-BS: 94%, PP: 95%). The Puccinellia bunt

was distinct but, with 86% ML-BS support, grouped with two isolates of T. bromi isolates in a

well-supported clade (ML-BS: 100%, PP: 100%). Other well-supported clades include those

containing isolates of T. vankyi (ML-BS: 99%, PP: 100%), T. brevifaciens (ML-BS: 96%, PP:

100%), T. goloskokovii (ML-BS: 99%, PP: 100%). Isolates of three wheat bunt pathogens, T.

contraversa, T. caries and T. laevis, form a well supported clade (ML-BS: 83%, PP: 100%), but

without internal support to distinguish individual species.

Incongruence among the three genes was evaluated by partition homogeneity test in PAUP.

A strong conflict was indicated (P = 0.001 for all genes). One source of incongruence is that the

ML tree topologies inferred by the ITS1 locus included six different trees with the same likelihood score. Incongruence of tree topologies for the three genes was also examined by looking for conflict among well-supported clades. In general, clades inferred from EF1A and ITS genes were not as well supported as those from RPB2. Among well-supported clades (80% ML-BS and 95%

PP), none were found to be in conflict. In the clades with either 80% ML-BS or 95% PP, incongruence was found in the placement of T. secalis, which was clustered with two T.

brevifaciens isolates by EF1A gene (ML-BS: 100%, PP: 100%), while clustered with T. trabutii, T.

goloskokovii, T. sphaerococca and T. lolioli by the RPB2 gene (ML-BS: 47%, PP: 99%) (trees not

shown). This result is consistent with previous results (Carris et al 2007), and T. secalis was not

excluded from the analyses in this study. The sequence alignment and phylogenetic trees were

deposited in TreeBASE as SN4002.

FIG. 2-3. Tilletia puccinelliae (WSP 71469, holotype). a. Teliospores, plane view. b. Teliospores, surface view of reticulate ornamentation. c. Teliospores and sterile cells. d. Teliospore and basidium. e. Germinated teliospore and conjugated basidiospores. f. Conjugated basidiospores and

sporidia (secondary basidiospores). a–c, bar = 10 µm; d, bar = 10 µm; e, bar = 20 µm; f, bar = 20

µm.

BS PP

T. puccinelliae WSP 71471 T. puccinelliae CBS 122998 T. puccinelliae CBS 122997 Puccinellia distans 94 T. puccinelliae WSP 71465 95 T. puccinelliae WSP 71469 86 T. puccinelliae WSP 71472

T. bromi WSP 71272 100 Bromus spp. 100 T. bromi WSP 71271 96 T. laguri HUV16.352 Lagurus ovatus 99 Lolium rigidum 99 T. lolii WSP 71298 100 Lolium perenne 99 T. vankyi WSP 71266 100 T. vankyi WSP71270 Fustuca rubra

96 T. brevifaciens WSP 68945 Thinopyrum 100 intermedium T. brevifaciens HUV20.802 T. caries WSP 71304 T. caries WSP 71303 T. laevis WSP 71300 Triticum spp. T. controversa WSP 69062 83 100 T. controversa WSP 71280 T. laevis WSP 71302 T. trabuti WSP 71299 Hordeum spp. T. trabutii VPRI 32106

T. secalis WSP 71279 Secale cereale

100 T. goloskokovii WSP 71281 99 Apera interrupta 100 100 T. goloskokovii WSP 69687 96 98 T. sphaerococca WSP 71314 Agrostis stolonifera 100 T. lolioli WSP 71305 Loliolum subulatum T. togwateei WSP 71277 Poa reflexa

T. elymi WSP 71274 Elymus glaucus T. fusca WSP 71275 Vulpia microstachys 0.01

FIG 2-4. Maximum likelihood tree of 30 taxa of Tilletia spp. with the combination of EF1A,

ITS1 and RPB2 genes. Branch length indicates the substitution rate. Maximum likelihood bootstrap (BS) values above 80% are indicated on branches, and Bayesian posterior probabilities (PP) values above 95% are indicated under branches.

TAXONOMY

Tilletia puccinelliae Carris, Castl. & G. Huang, sp. nov. (FIGS. 2-1, 2-2a, 2-3b, 2-2c, 2-3)

MycoBank no.: MB 512086

Etym: derived from the host plant Puccinellia

Sori in ovario inclusi, ellipsoidei-citriformes, 1–1.5 × 0.5–0.8 mm; massa sporarum foetida, fusci

brunneo et pulverulenta. Cellulae steriles subglobosae vel irregulares, 11–17.5 × 10.5–16.5 µm, hyalineae, cum parietibus 1–2.5 µm crasso, levibus. Sporae globosae, subglobosae, ovoidae, ellipsoidales, rotundae subpolyedrice irregulars, plerumque cum complanatus lateribus, usque elongatae, raro cum subacutae apice, subrubeo-brunneae, 18.5–31 ×15–28 µm, exosporium subtiliter reticulatum usque incompletae reticulatum, raro cerebriformae, muri 0.5–1.5 (–2) µm alti,

6–10 (–14) in circulo aequatoriali. Basidiosporae filiformae, 9–16 per basidium, conjugatae,

uninucleatae, 70–80 × 2.5 µm.

Sori in ovaries of Puccinellia distans, enclosed by pericarp, ellipsoidal to lemon-shaped, 1–1.5 ×

0.5–0.8 mm; spore mass foetid, dark fuscous brown, powdery (FIGS. 2-1, 2-2a). Sterile cells

subglobose to irregular, 11–17.5 ×10.5–16.5 µm, hyaline, wall 1–2.5 µm thick, smooth (FIG. 3-3c).

Teliospores globose, subglobose, ovoid, ellipsoidal, irregularly rounded subpolyhedral, often with a flattened side, to elongated, rarely with a subacute tip, medium reddish-brown, close to Chestnut

(color no. 40, Rayner 1970), 18.5–31× 15–28 µm, exospore finely reticulate to incompletely

reticulate, rarely cerebriform (Vánky 1991), muri 0.5–1.5 (–2) µm high, with 6–10 (–14) meshes

per spore diameter (FIGS. 2-2b, 2-2c, 2-3a, 2-3b, 2-3c). Teliospore germination 10–12 d at 5 C on

WA; no germination at room temperature and 15 C. Basidium simple, up to 80 µm long (FIG. 2-3d, e); basidiospores 9–16 per basidium, conjugating, uninucleate, hyaline, filiform, 70–80 × 2.5 µm

(FIG. 2-3e, f). Sporidia (ballistospores) allantoid, 26–35 × 4.5–6 µm, formed from subulate

sporogenous cells formed from hyphae or primary basidiospores (FIG. 2-3f).

Holotype.— U.S.A: Washington, Franklin County: ovaries of Puccinellia distans, 2004, L. M.

Carris WSP 71469. Isotypes: BPI 878741, HUV 20962. Ex-type: CBS 122993.

Additional representative collections examined. — See TABLE 2-1.

Commentary.— Among the Tilletia spp. most likely to occur as contaminants in cool season grass

seed, the irregular shape of the teliospores in T. puccinelliae is a distinctive feature of this species,

but the size, color and ornamentation of the teliospores are similar to T. bromi and T. goloskokovii.

The sterile cells of T. puccinelliae are smaller than those of T. bromi and T. goloskokovii, approx. half the diam of the teliospores (TABLE 2-2, FIG. 2-3a – FIG. 2-3c). The size and shape of primary

basidiospores of the five species compared in TABLE 2-2 overlap, but the number of basidiospores

per basidium in the Puccinellia bunt (9–16) is more similar to T. bromi than T. goloskokovii. The

sori of T. puccinelliae are similar in size and shape to those of T. goloskokovii (FIG. 2-2a, b). Tilletia bromi sori (FIG. 2-2c), in contrast, are up to 3 times as long as those of either T. goloskokovii (FIG.

2-2b) or T. puccinelliae (FIG. 2-2a).

DISCUSSION

The genus Puccinellia Parl. belongs in tribe Poeae of the grass subfamily Pooideae and is most

closely related to Poa and Sclerochloa among the festucoid grass genera (Catalán et al 2004).

There are no known smuts on Sclerochloa. Six Tilletia species are reported to infect Poa — T.

cathcartae Durán & G.W. Fisch., T. paradoxa Jacz., T. poae Nagorny, T. sterilis Ule, T. togwateei

and T. transiliensis M.N. Kusnezow & Schwarzman. Teliospores of T. paradoxa, T. transiliensis

and T. cathcartae have echinulate or tuberculate ornamentation, and sori of T. sterilis form in leaf

sheaths and culms (Durán and Fischer 1961). Teliospores of T. poae are reticulate, similar in size to

those of T. puccinelliae but more regular in shape and with deeper exospore than those of T.

puccinelliae (Vánky 1994); no records of T. poae were found for North America. T. togwateei, an

ovary-infecting species with reticulate spores, is included in the present analysis (Fig. 2-4). T.

puccinelliae is easily distinguished from other genera of smuts on Puccinellia spp. For example,

three species of Ustilago— U. hypodytes (D. F. L. Schlechtendal) E. Fries, U. spegazzinii Hirschh.,

and U. trebouxii H. & P. Sydow — have been reported in North America (Fischer 1953), all of

which form sori in the leaves and leaf sheaths of P. distans. Entyloma dactylidis (G. Passerini) R.

Ciferri, the only known species in this genus on Puccinellia, forms sori in the leaves and leaf

sheaths with densely packed, smooth teliospores (Vánky 1994). Urocystis puccinelliae L.Guo & H.

C. Zhang (2004) is reported on Puccinellia tenuiflora (Griseb.) Scribn. & Merr. in China. Species

of Urocystis are ready distinguished from species of Tilletia by formation of spore balls in sori.

The ca. 25 species in the Puccinellia occur in saline soils in coastal and inland habitats

(Choo et al 1994). P. distans is a Eurasian perennial species that has become established throughout much of North America (Hitchcock 1950). Increased salinity due to the use of effluent and low quality water for irrigation of golf courses and other types of turf has resulted in a demand for salt tolerant turfgrasses, and Fults alkaligrass is one of the most salt tolerant cool season grasses

(Alshammary et al 2004). Approximately 121 ha of Fults alkaligrass are grown for seed production

in Washington. Tilletia puccinelliae is known only from individual bunted florets detected in eight

seedlots of alkaligrass cv. Fults from Franklin County, Washington produced during the 2004

growing season. Infected plants have not been observed in the field, and the manner in which T. puccinelliae infects, e.g., systemically or locally through developing florets, is not known.

However, T. puccinelliae is closely related to other systemically-infecting bunts with hosts in the

grass subfamily Pooideae, and it likely infects at the seedling stage and grows systemically within

the host as do other bunts in this group. The complete replacement of the seed with teliospores, and

the presence of several bunted florets attached to the rachis (FIG. 2-1b) supports the systemic

nature of this bunt; local infection through the developing floret typically results in a partially

bunted seed, and typically only a few florets of the inflorescence are bunted (Carris et al 2007). We

know little about the ability of systemically infecting smuts to survive year to year in perennial

grass hosts such as P. distans, as most of the well-studied Tilletia spp. infect annual grasses. In

several regards, T. puccinelliae is similar to T. vankyi, a recently described bunt fungus infecting cultivated Festuca rubra (fine fescue) and Lolium perenne (perennial ryegrass) (Carris et al 2007).

Both species were intercepted as bunted florets in turfgrass seed by scientists at the Tianjin

Entry-Exit Quarantine and Inspection Bureau, and the mode of infection is not known with certainty for either species. Bunted florets of T. vankyi found in seedlots of fine fescue and perennial ryegrass were completely filled with teliospores, as with T. puccinelliae. Greenhouse studies conducted with T. vankyi demonstrated that inoculation methods appropriate for both seedling and floret-infecting species resulted in infection in L. perenne and L. perenne var. multiflorum (annual ryegrass) (Carris et al 2007).

Tilletia spp. on other hosts in grass tribe Poeae (Lolium, Festuca, Loliolum, Vulpia and Poa) do not form a monophyletic group based on the multigene analysis in Carris et al (2007). Tilletia puccinelliae isolates form a well-supported group in a clade with T. bromi, a species restricted to

Bromus (tribe Bromeae). Bromus tectorum was introduced into North America in the late 1800s from Eurasia, and has become one of the most abundant plant species in the Intermountain West

(Mack 1984). Bromus tectorum infected by T. bromi is common in and around cultivated grass and grain fields in Washington, and bunted florets of B. tectorum occur as contaminants in wheat and grass seed. The spore germination, size, ornamentation and color of teliospores of T. puccinelliae

and T. bromi overlap sufficiently (TABLE 2-2, FIGS. 2-2b, c, h, i) for the two taxa to be considered conspecific in the absence of molecular data, although the irregular shape of T. puccinelliae teliospores is distinct from T. bromi. Studies utilizing a molecular approach have demonstrated, in general, that Tilletia spp. have a relatively narrow host range usually restricted to a single genus, or in some cases, to a single host species (Boyd and Carris 1997; Pimentel et al 1998; Carris et al

2007). An apparent exception to the narrow host range is T. vankyi which infects Lolium and

Festuca (Carris et al 2007). However, Festuca is a large, complex genus shown to encompass species of Lolium in phylogenetic reconstructions (Catalán et al 2004).

When the genus Tilletia was monographed (Durán and Fischer 1961), 76 species were accepted from among 200 described species. Since Durán and Fischer (1961), the number of species recognized has more than doubled, with 85 new species or new combinations accepted based on the MycoBank database (http://www.mycobank.org). Only 14 of the new species were described from North America, with 10 of these from Mexico described in Durán (1987). As with many other fungi, the known distribution of Tilletia spp. clearly reflects the activities of collectors; the majority (52) of new Tilletia species have been described by Kalmán Vánky and various coauthors.

Tilletia puccinelliae is one of four new species, with T. laguri, T. vankyi, and T. walkeri, first encountered by seed inspectors and/or quarantine scientists; none of these species has been observed directly on its respective host(s) in the field. The type specimen of Tilletia laguri is based on a collection of bunted Lagurus ovatus from Italy which was intercepted in China (Zhang et al

1995); an earlier collection was intercepted in the US in 1982 in a shipment of dried flowers from

Spain, but the specimen (WSP 71282) was not identified as a new species. Tilletia vankyi, as previously noted, was intercepted by scientists as the Tianjin Entry-Exit Inspection and Quarantine

Bureau in seed of Festuca rubra from the US in 2003, and Lolium perenne from Australia in 2003,

and Germany in 2005 (Carris et al 2007). The first record of Tilletia walkeri was a bunted floret of

L. perenne from Australia’s Kangaroo Valley (New South Wales) found by a seed testing lab in

1967 and deposited as DAR 16719 as “Tilletia sp.”, although the species was not formally

described until it was found in the U.S.A. thirty years later during the National Karnal Bunt Survey

(Castlebury and Carris 1999). These examples illustrate the need for more extensive surveys of cultivated grasses, particularly those being exported as seed around the world.

ACKNOWLEDGEMENTS

PPNS 0523, Department of Plant Pathology, College of Agricultural, Natural and Human

Resource Sciences Research Project No. WNP03837, Washington State University, Pullman. We thank Orlin Reinbold, Landmark Native Seeds, Spokane, WA, and Victor Shaul, Washington State

Department of Agriculture Seed Lab, for providing alkali grass seed samples, Dr. Christian

Feuillet for assistance with the Latin description, and Dr. Kálmán Vánky, Herb. H.U.V., for input and advice. The Tianjin Entry-Exit Inspection and Quarantine Bureau project is supported by

20081K236 of AQSIQ, P.R. China.

AFFILIATION OF CO-AUTHORS

Lori M. Carris

Department of Plant Pathology, Washington State University, Pullman, WA 99164-6430, USA

Guoming Huang, Jiafeng Luo, Yueting Liu

Tianjin Entry-Exit Inspection and Quarantine Bureau, Tanggu, Tianjin 300456, China

Lisa A. Castlebury

USDA-ARS, Systematic Mycology and Microbiology Lab, 10300 Baltimore Avenue, Beltsville,

MD 20705-2350, USA

LITERATURE CITED

Alshammary SF, Qian YL, Wallner SJ. 2004. Growth response of four turfgrass species to salinity.

Agricultural Water Management 66: 97–111.

Boyd ML, Carris LM. 1997. Molecular relationships among varieties of the Tilletia fusca (T. bromi)

complex and related species. Mycol. Res. 101: 269–277.

———, ———. 1998. Evidence supporting the separation of the Vulpia- and Bromus-infecting

isolates in the Tilletia fusca (T. bromi) complex. Mycologia 90: 1031–1039.

———, ———, Gray PM. 1998. Characterization of Tilletia goloskokovii and allied species.

Mycologia 90: 1031–1039.

Carris LM, Castlebury LA, Huang G, Alderman SC, Luo J, Bao X. 2007. Tilletia vankyi, a new

species of reticulate-spored bunt fungus with non-conjugating basidiospores infecting

species of Festuca and Lolium. Mycological Research 111: 1386–1398.

Castlebury LA, Carris LM. 1999. Tilletia walkeri, a new species on Lolium multiflorum and L.

perenne. Mycologia 71:449–455.

Catalán P, Torrecilla P, López Rodríguez JA, Olmstead RG. 2004. Phylogeny of the festucoid

grasses of subtribe Loliinae and allies (Poeae, Pooideae) inferred from ITS and trnL-F

sequences. Mol. Phyl. Evol. 31: 517–541.

Choo MK, Soreng RJ, David JI. 1994. Phylogenetic relationships among Puccinellia and allied

genera of Poaceae as inferred from chloroplast DNA restriction site variation. American

Journal of Botany 81: 119–126.

Durán R. 1987. Ustilaginales of Mexico. Taxonomy, Symptomology, Spore Germination and

Basidial Cytology. Washington State University, Pullman.

Durán R, Fischer GW. 1961. The Genus Tilletia. Washington State University, Pullman.

Fischer GW. 1953. Manual of the North American Smut Fungi. The Ronald Press, New York.

Guo L, Zhang HC. 2004. A new species and two new records of Ustilaginomycetes from China.

Mycotaxon 90: 387–390.

Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program

for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41:95–98.

Hitchcock AS. 1950. Manual of the Grasses of the United States. United States Department of

Agriculture Miscellaneous Publication No. 200, Government Printing Office, Washington.

Huelsenbeck JP, Ronquist F. 2001. MRBAYES. Bayesian inference of phylogeny. Bioinformatics 17:

754–755.

Mack RN. 1984. Invaders at home on the range. Natural History 2: 40–47.

Peterson GL, Bonde MR, Phillips JG. 2000. Size-selective sieving for detecting teliospores of

Tilletia indica in wheat seed samples. Plant Disease 84: 999–1007.

Pimentel G, Carris LM, Levy L, Meyer R. 1998. Genetic variation among isolates of Tilletia

barclayana, T. indica, and allied species. Mycologia 90: 1017–1027.

Posada D, Crandall KA. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics

14: 817–818.

Poulter RTM, Goodwin TJD, Butler MI. 2007. The nuclear encoded inteins of fungi, Fungal Genet.

Biol. 44: 153–179.

Rayner RW. 1970. A Mycological Colour Chart. Commonwealth Mycological Institute, Kew.

Reeb V, Lutzoni F, Roux C. 2004. Contribution of RPB2 to multilocus phylogenetic studies of the

Pezizomycotina (euascomycetes, Fungi) with special emphasis on the lichen-forming

Acarosporaceae and the evolution of polyspory. Molecular Phylogenetics and Evolution 32:

1036–1060.

Ronquist F, Huelsenbeck JP. 2003. MRBAYES3: Bayesian phylogenetic inference under mixed

models. Bioinformatics 19:1572–1574.

Swofford DL. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods).

Version 4. Sinauer Associates, Sunderland, Massachusetts.

Trione EJ. 1964. Isolation and in vitro culture of the wheat bunt fungi Tilletia caries and T.

controversa. Phytopathology 54: 592–596.

Vánky K. 1991. Spore morphology in the taxonomy of Ustilaginales. Trans. Mycol. Soc. Japan 32:

381–400.

Vánky K. 1994. European Smut Fungi. Gustav Fischer Verlag, New York.

White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal

RNA sequences for phylogenetics. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ (eds),

PCR Protocols: a guide to methods and applications. Academic Press, San Diego, pp.

315–322.

Zhang G, Lin G, Deng J. 1995. A new species of the genus Tilletia. Acta Mycologica Sinica

14:17–19.

CHAPTER THREE

A MULTILOCUS PHYLOGENETIC APPROACH TO TESTING SPECIES LIMITS

IN THE WHEAT BUNT FUNGI TILLETIA CARIES,

T. CONTRAVERSA AND T. LAEVIS

ABSTRACT

Common and dwarf bunt of wheat are recognized as being caused by three closely related, but morphologically and physiologically distinct species. Tilletia caries and T. laevis cause common bunt, and T. contraversa causes dwarf bunt. Previous phylogenetic analyses of a limited number of isolates have presented conflicting results with three morphological species for the wheat bunt pathogens. In this study, a collection of 60 wheat bunt isolates of different geographic and temporal origin was used to compare morphological species recognition methods with phylogenetic approaches. K-means cluster analysis of sorus shape and teliospore size supported two main groups, one containing most of the common bunt isolates, and the second containing most of the dwarf bunt isolates. Phylogenetic analyses based on DNA sequence data from three sequence-characterized amplified regions (anonymous loci) and an amplified region with the second largest subunit of RNA polymerase II (RPB2) primers revealed conflict between

Maximum Likelihood (ML) trees estimated from the anonymous loci and the RPB2 region.

Statistical parsimony and NeighborNet networks were estimated from concatenated data and none of the clades in either ML trees or in the networks exclusively contained isolates of the three morphological species, nor did they correspond to the groups supported by k-means cluster analysis. Geographic origin of isolates was also not correlated to network clades. Three

morphological species of wheat bunt pathogens were not recognized by systematic analyses with

DNA sequence information in this study.

INTRODUCTION

Bunt diseases are caused by a type of smut fungus (Ustilaginomycotina, Basidiomycota) restricted

to species of Tilletia characterized by the formation of dark, powdery teliospores interspersed with

hyaline sterile cells in sori, which partially or completely replace the ovary tissue of infected

cereals or other related grasses. Tilletia caries (DC.) Tul. & C. Tul., T. contraversa Kühn, and T.

laevis Kühn are widely recognized as being closely related but morphologically and physiologically distinct species that cause destructive bunt diseases on wheat and other grasses.

Tilletia contraversa, the dwarf bunt pathogen of wheat, is distinguished from the common bunt pathogens T. caries and T. laevis by stunting and increased tillering in the infected host with hard and rounded sori (FIG. 3-1g, h), reticulately ornamented teliospores with a relatively deep exospore

enveloped in a thick gelatinous sheath (FIG. 3-1a, b), and a requirement for low temperature

(optimum 3-8 C) and light for germination of teliospores (Goates 1996). Germination typically

occurs in 3-6 weeks. T. caries and T. laevis are biologically similar to each other; they do not cause

host stunting and have teliospores that germinate at 15 C within a week in the absence of light.

Teliospores of T. caries are 14–23.5 µm diam, reticulate and lack a conspicuous gelatinous sheath

(FIG. 3-1c, d), while teliospores of T. laevis are smooth, 14–22 µm diam (Durán and Fischer 1961)

(FIG. 3-1e, f).

FIG. 3-1. Teliospores (a-f) and sori (g-h) of wheat bunts. (a-b) Tilletia contraversa (35821). (c-d). T.

caries (V1033). (e-f). T. laevis (99-3). g. common bunt (LMC178). h. dwarf bunt (D7).

Wheat bunt has been known since ancient times, and T. caries and T. laevis are widespread

on winter- and spring-sown wheat throughout the wheat-growing regions of the world (Durán and

Fischer 1961, Hoffmann 1982). In contrast, dwarf bunt of wheat was not recognized as distinct

from common bunt until the early twentieth century, perhaps because dwarf bunt is restricted to winter-sown wheat grown in areas with persistent snow cover and common bunt is the dominant pathogen when occurring in the same field (Bamberg et al 1947). Dwarf bunt was first reported to be a new race of T. caries in Montana in 1935 (Young 1935), and was formally described as T. brevifaciens G. Fisch. on intermediate wheatgrass, Thinopyrum intermedium, in 1952 (Fischer

1952). Dwarf bunt was later shown to be present in cultivated wheat in Europe and North America as early as the 19th century based on examination of herbarium specimens (Purdy et al 1963, Trione

1982). Tilletia contraversa was established by Kühn (1874) for a bunt in Germany on Elymus

repens, a widespread grass commonly known as ‘quackgrass’. In Kühn’s observations

accompanying the description, he noted that the quackgrass bunt was distinct from T. caries in

having on average 1-µm smaller (16-19 µm diam), more uniformly spherical teliospores, with a

deeper exospore and larger diameter mesh (Kühn 1874). Kühn also observed that the spores did

not germinate within 60-72 hours, as did those of T. caries, but he did not note host stunting or a

gelatinous sheath on the teliospores of T. contraversa in his description or observations. Tilletia

contraversa supplanted T. brevifaciens as the correct, earlier name for the causal agent of dwarf

bunt of wheat primarily on the basis of similar teliospore morphology (Conners 1952).

In 1973, wheat exports from the U.S. to China were abruptly halted when Chinese

quarantine officials began rejecting shipments of wheat contaminated with teliospores of T.

contraversa, a species not known to occur in China (Trione 1982, Mathre 1996). Tilletia caries and

T. laevis were both present in China and not subject to the quarantine (Trione 1982, Goates 1996).

Because of the implications for misidentification of T. caries as T. contraversa, a number of studies

were conducted to develop methods for distinguishing the two species based on teliospore

morphology (Trione and Krygier 1977), epifluorescence of exospore (Stockwell and Trione 1986),

serology (Banowetz et al 1984) and total proteins (Kawchuk et al 1988, Banowetz and Doss 1994).

None of these methods have provided a reliable means for distinguishing spores of T. contraversa

from those of T. caries (Russell and Mills 1994).

Sexual compatibility studies on this group of fungi have suggested a close evolutionary

relationship among the three wheat bunt species. T. caries and T. laevis were crossed by Flor

(1932), providing the first evidence that the wheat bunt fungi are heterothallic and that the two species are reproductively compatible. Similarly, Silbernagel (1964) showed that monosporidial lines of T. contraversa were compatible with those of T. caries, and the resulting dikaryotic mycelium was able to infect wheat. Pimentel et al (2000) showed that hybrid progenies of T.

contraversa and T. laevis were intermediate in teliospore ornamentation relative to the parental

lines and were as fit as parental lines in basidiospores fusion (dikaryon formation) and infectivity.

A phylogenetic analysis of Tilletia and related genera in the order Tilletiales utilizing a

portion of the nuclear large subunit (nLSU) rDNA gene distinguished a well-supported lineage

containing Tilletia species on hosts in the grass subfamily Pooideae (Castlebury et al 2005). The nLSU rDNA region sequenced for this study was not sufficiently variable to separate T. caries (as

T. tritici), T. contraversa, T. laevis or any other pooid-infecting species within the lineage. In a later study, a multi-locus analysis based on ITS rDNA, elongation factor 1 alpha (EF1α), and a portion of the second largest subunit of RNA polymerase II (RPB2) resolved one well-supported clade containing the common and dwarf bunt species, as well as strongly-supported clades for individual, narrow host range species of Tilletia on other pooid grass hosts (Carris et al 2007). There was no

internal branch support in the wheat bunt clade corresponding to T. caries, T. contraversa or T. laevis (Carris et al 2007).

Sequence characterized amplified regions (SCARs) (Paran and Michelmore 1993) are anonymous loci that have been shown to be useful for phylogenetic analyses (Jany et al 2002,

Castrillo et al 2003). Bailey et al (2004) used SCAR loci as the basis for a phylogenetic study on the plant genus Leucaena (Fabaceae), and showed better resolution and clade support with

SCARS than those generated from ITS sequence data. SCAR loci were also used independently or combined with other loci to reconstruct phylogenies for closely related taxa of Alternaria (Peever et al 2004, Andrew et al 2009). In the current study, a similar strategy was used to develop 3 anonymous loci for the analyses of 60 wheat bunt isolates from diverse geographic and temporal origins. In this way, we were able to: (i) determine if distinct clades can be recognized with phylogenetic approaches among a collection of wheat bunt isolates from three morphological species, and (ii) evaluate the evolutionary significance and reliability of using morphological and physiological characters as criteria to distinguish TCK from other wheat bunt isolates.

MATERIALS AND METHODS

Fungal isolates.—The 60 wheat bunt isolates collected from different locations in North America and Eurasia as three morphological species were used with other related Tilletia isolates in this study, which are listed in TABLE 3-1. Voucher materials were deposited in Herb. WSP (Washington

State University, Pullman, WA), Herb. BPI (Beltsville, MD), or Herb. HUV (Tübingen, Germany).

TABLE 3- 1. Tilletia isolates used in Chapter 3.

Speciesa Voucher Host Origin Collector Year /Culture /Source numbers T. contraversab 35812 Triticum aestivum Canada (British D. Taylor 1950 L. Columbia) T. contraversa D7 Triticum aestivum Greenhousec B. Goates 1999 T. contraversa D17 Triticum aestivum Greenhousec B. Goates 1999 T. contraversa DB36 Triticum aestivum US (Washington) B. Goates 1957 T. contraversa DB295 Triticum aestivum US (Utah) B. Goates 1980 T. contraversa? d DB572 Triticum aestivum US (Wyoming) B. Goates 1982 T. contraversa DB688 Triticum aestivum US (Colorado) B. Goates 1990 T. contraversa DB90-30 Triticum aestivum US (Oregon) B. Goates NA T. contraversa HK127 Triticum aestivum US (Montana) J. Hoffmann 1962 T. contraversa J3 Triticum aestivum Sweden L. Johnsson 1990 T. contraversa LMC177 Triticum aestivum US (Utah) L. Carris 1985 T. contraversa LMC282 Triticum aestivum US (Oregon) L. Carris 1994 T. contraversa LMC295 Triticum aestivum US L. Carris NA T. contraversa LMC446 Triticum aestivum Germany K. Vanky 1992 (V528) T. contraversa LMC520 Triticum aestivum US (Oregon) S. Alderman 2008 T. contraversab LMC522 Triticum aestivum Italy V. Grasso 1954 (Tr. vulgare) T. contraversa LMC523 Triticum spelta L. Germany M. 1996 Piepenbring T. contraversa? e LMC524 Triticum aestivum Kazakhstan B. Goates 2008 T. contraversa S90-09 Triticum aestivum US (Idaho) J. Sitton 1990 T. contraversa TK34-1 Triticum aestivum Turkey B. Goates 1979 T. contraversab WSP35784 Thinopyrum US (Idaho) E. Horning 1950 intermedium (Host) Barkworth & D.R. Dewey T. contraversab WSP35785 Triticum aestivum US (Washington) J. Meiners 1946 T. contraversab WSP63659 Triticum aestivum Germany Not known 1954 T. contraversab WSP63862 Triticum aestivum Switzerland H. Aebi 1954 T. caries 17-13 Triticum aestivum Canada (Alberta) D. Gaudet 1976 T. caries 78-49 Triticum aestivum Canada (Alberta) D. Gaudet 1978 T. cariesf HB11 Triticum aestivum China B. Goates 1993 T. caries J10 Triticum aestivum Sweden L. Johnsson 1981 T. caries J11 Triticum aestivum Sweden L. Johnsson 1988 T. caries J13 Triticum aestivum Finland L. Johnsson 1993 T. caries J17 Triticum aestivum Sweden L. Johnsson 1996 T. caries LMC360 Triticum aestivum US (Washington) Not known 1989

T. caries LMC515 Triticum aestivum US (Washington) Not known 2006 T. caries LMC516 Triticum aestivum US (Oregon) X. Chen 2007 T. caries LMC517 Triticum aestivum US (Washington) X. Chen 2007 T. caries LMC518 Triticum aestivum US (Oregon) R. Smiley 2008 T. caries LMC527 Triticum aestivum US (Oregon) B. Goates 1987 T. caries T1 Triticum aestivum Greenhousec B. Goates 1984 T. caries T2 Triticum aestivum Greenhousec B. Goates 1989 T. caries TK98-2 Triticum turgidum Turkey B. Goates 1979 L. (Tr. dicoccum) T. caries V4 Triticum aestivum Romania V. 1996 Dumalasova T. caries V12 Triticum aestivum Hungary V. 1997 Dumalasova T. caries V31 Triticum aestivum Switzerland V. 1996 Dumalasova T. caries V73 Triticum aestivum Latvia V. 1996 Dumalasova T. caries V103 Triticum aestivum Germany V. 2003 Dumalasova T. caries WSP70347 Triticum aestivum Poland K. Vanky 1994 (V1033) T. laevis 79-3 Triticum aestivum Canada (Alberta) D. Gaudet 1979 T. laevis 90-61 Triticum aestivum US (Corolado) B. Goates 1990 T. laevis 99-3 Triticum aestivum Australia G. Murray 1999 T. laevis DAR73299 Triticum aestivum Australia G. Murray 1997 T. laevis L16 Triticum aestivum Greenhousec B. Goates 1984 T. laevis LMC178 Triticum aestivum Greenhousec L. Carris 1998 T. laevis LMC519 Triticum aestivum US (Michigan) B. Goates 1989 T. laevis LMC521 Triticum aestivum Germany J. Grunzel NA T. laevis LMC525 Triticum aestivum Tajikistan B. Goates NA T. laevis TK52-1 Triticum aestivum Turkey B. Goates 1979 T. laevis TK108 Triticum sp. Turkey B. Goates 1979 T. laevis TK126-2 Triticum durum Turkey B. Goates 1979 Desf. T. laevis V3 Triticum aestivum Bulgrary V. 1996 Dumalasova T. laevis V90 Triticum aestivum Syria V. 1996 Dumalasova T. laevis WSP71300 Triticum aestivum Iran E. Pourjam 1988 (V766) T. caries/T. LMC474 Triticum aestivum US C. Peterson 1985 laevis a. Species determination as indicated by original collector b. Specimens originally determined as T. brevifaciens

c. Isolates maintained in USDA greenhouse, Aberdeen, Idaho d. Germinated at 15C e. Species determined by plant dwarfing

Morphological analysis.— Intact bunted florets (sori) were detached from smutted heads and

placed on a smooth surface, with groove side facing down to ensure all the sori were measured in

the same way. Length and width of 3-5 sori were recorded per specimen where sufficient material

was available. Teliospores were mounted in Shear’s mounting medium and examined using

differential interference contrast microscopy at ×1000. Teliospore diam (including exospore),

thickness of exospore, number of meshes (areolae) per spore were recorded for 20 spores per

sample (TABLE 3-2).

The teliospore size/sorus shape and exospore morphology for each specimen were

re-examined in this study and used to classify the collections into subgroups which correspond to

common (T. caries, T. laevis) or dwarf bunts (T. contraversa). Teliospore size and sorus shape

were analyzed using k-means clustering method in R 2.10.0 with Lloyd's algorithm (MacQueen

1967, Lloyd 1982, Newcombe et al 2000). K-means clustering partitions the observations into k

clusters to minimize the within-cluster sum of squares (MacQueen 1967). To determine how

many clusters (k) would be selected, the inter-cluster variations (σ2) were compared in this study

when observations of wheat bunt collections were grouped into k = 2 to k = 10 clusters with teliospore size and sorus shape as 2-dimensional data. The k with highest variations was used as the best model. Teliospore morphology was used to classify collections into one of three groups based on presence or absence of a reticulated exospore, depth of the exospore reticulation and presence or absence and thickness of the gelatinous sheath: smooth (S) exospore, shallow (≤

1.5μm) exospore reticulation (R) with no or thin (≤ 2μm) gelatinous sheath, and deep (≥ 1.5μm) exospore reticulation (D) with thick (≥ 2μm) gelatinous sheath. The depth of the exospore and presence of a gelatinous sheath were combined in the classification because a strong correlation between these two characters was observed among different collections. Collectors’

morphological species determination, k-means clustering and exospore morphology are shown in

TABLE 3-2 and the spore morphology and the geographic origins of each isolate were mapped onto the network (FIG. 3-4). To test the power of using limited numbers of morphological

characters in species determination, the results were compared with the species determinations

made by the collectors. Statistical similarities between any two of three morphological species

recognition methods were tested by the marginal homogeneity (MH) method (Ireland et al 1969).

The null hypothesis of MH test assumes the outcomes from two different grouping methods on the same dataset are not significantly different. ANOVA tests were used in this study to analyze if the dwarf bunt isolates are morphologically different from the common bunt isolates with spore size or sorus shape.

Nucleic acid extraction.—DNA was extracted directly from actively growing surface mycelium scraped from M-19 agar plates (Trione 1964) or directly from teliospores. In order to establish cultures, teliospores were surface-sterilized in 0.26% NaClO in a 1.5 ml Eppendorf tube for 50 s, pelleted by centrifugation at approx. 13000 g in a benchtop microcentrifuge for 10 s and rinsed twice with sterile, distilled water. Surface-sterilized teliospores were streaked onto 1.5% water agar and incubated at 5 C with 8 h of light per day or at 15 C in the dark. Primary basidiospores were transferred to M-19 agar to establish colonies. Cultures on M-19 were maintained at 15 C in the dark. Actively growing mycelium was scraped from agar, sliced into small pieces and placed in a 1.5 Eppendorf tube. The mycelia were immediately frozen at -80 C, lyophilized for 24 h, and stored at -20 C until use. DNA was extracted using approximately 5 mg of lyophilized and ground mycelium with DNeasy Plant Mini Kit (Qiagen Inc., Valencia, California) according to the manufacturer’s instructions. For DNA extraction from teliospores, 5 mg of teliospores were suspended in a 1.5 mL Eppendorf microcentrifuge tube with extraction buffer [50mM Tris-HCl

pH8.0, 150 mM NaCl, 100mM EDTA] (Shi et al 1996). Up to 200 μl Zirconia beads were added to each tube and shaken with extraction mixture in a Mini-Beadbeater (BioSpec Inc., Bartlesville,

Oklahoma) for 3 min. In each tube, 50 μl 20% sodium dodecyl sulfate (SDS), 100 μl 5 M NaCl and

120 μl 10% poly-vinyl pyrrolidone (PVP40) were added to the mixture and shaken gently at 55 C overnight, followed by routine phenol/chloroform extraction and ethanol precipitation. DNA concentrations were determined using a NanoDrop ND1000 (Thermo Scientific, Wilmington,

Delaware) and adjusted to 30 ng/μL for amplification with single RAPD (Random Amplification of Polymorphic DNA) 10 bp primers, or 10 ng/μL for specific PCR reactions.

Development of anonymous loci.—All PCR amplifications were performed in 20-μL reactions on a GeneAmp 9700 thermal cycler (Applied Biosystems, Foster City, California). Random 10 bp primers from kits OPA, OPB, OPE and OPP (Operon Technologies, Alameda, California) were used to amplify DNA extracted from cultures of two collections of T. contraversa (LMC295, J3), one collection of T. caries (T1), and one collection of T. laevis (99-3). Amplified products were visualized with 1.5% agarose gel electrophoresis. Primers yielding strong and uniformly amplified bands for all four isolates were selected for further cloning and sequencing. Amplified products were precipitated in PCR tubes with 70% EtOH, and purified products were adjusted to recommended concentrations for direct cloning using pGEM Easy Vector (Promega, Madison,

Wisconsin) and One Shot TOP10 Competent Cells (Invitrogen, Carlsbad, California). The clones were screened with PCR-RFLP methods using tetra-nucleotide recognition restriction endonucleases (MspI, HaeIII and TaqI). Cloned products with the same restriction digest patterns for all four isolates were selected as possible homologs for sequencing. The PCR products were purified with ExoSAP-IT (USB, Cleveland, Ohio) according to the manufacturer’s instructions and amplified with respective forward and reverse PCR primers using the BigDye version 3.1 dye

terminator kit (Applied Biosystems). Sequencing reactions were purified with Performa DTR gel filtration cartridges (Edge BioSystems, Gaithersburg, Maryland) and sequenced on an ABI 3100 automated DNA sequencer. Specific primers were designed by using Primer3 (Rozen and

Skaletsky 2000) with manual modifications from the amplicons that have 95–100% sequence identity. All primers were designed to have similar annealing temperatures. Primers were tested on

16 additional isolates with PCR-based sequencing. Based on this approach, three anonymous loci were developed with the following primer pairs for use in PCR amplifications with the following

combinations: (i) A13-1AR874 (5′-GCAAGCGTGGGCCTCARTA-3′) and A13-1AR1439

(5′-TCAAGTAAGTACGGTAGAAGGGTGC-3′), (ii) A16-2Fb

(5′-AGGAGGGTTGTAGCAGCGAGCGA-3′) and A16-2R677

(5′-CCCCATTATTATTGCTGCTGACTG-3′), (iii) P18-1F1

(5′-CGCCCTTTTCCTGGTAGTTCA-3′) and P18-1R535 (5′-GCGGAGCATCCCACCAACT-3′).

All PCR amplifications were performed at 64 C as annealing temperature. Genomic DNA extracted from teliospores was used as templates. Strong and specific amplicons were selected with 1.5% agarose electrophoresis for sequence analysis. A region of RPB2 (the second largest subunit of RNA polymerase II) used in previous studies (Carris et al 2007, Bao et al 2010) was also amplified in this study as comparison, with 59 C as annealing temperature and the following primers: RPB2-740F (5′-GATGGACGCGGTTTGTAATG-3′) and RPB2-1365R

(5′-TCGAAGAGCYAACACTGAGACG-3′).

Sequence analysis.—Raw sequences of each amplicon from both directions were assembled using BIOEDIT v7.0.5.3 for Windows (Hall 1999). Sequence data from each individual locus was

generated in BIOEDIT and edited into alignments using CLUSTALW as implemented in BIOEDIT

with manual correction. Conflict between each of two datasets (loci) was evaluated by the

partition homogeneity test (incongruence length difference test) as implemented in PAUP

(Swofford 2002) with 1000 replicates. Independently inferred topologies from each partition were

examined for conflict (e.g., one monophyletic and the other paraphyletic). Reciprocal 70% ML

bootstrap (ML-BS) value and 95% Bayesian posterior probability (PP) were used as criteria for

strongly supported clades (Zharkikh & Li 1992, Ronquist & Huelsenbeck 2003).

To further address the incongruence among 4 loci, sequence alignments of 3 anonymous

loci and the RPB2 primer-amplified region were combined into concatenated sequences for all

taxa using MESQUITE 2.72 (Maddison and Maddison 2006). An isolate of T. vankyi was used as outgroup in the combined anonymous loci dataset according to results from previous multilocus phylogenetic analyses of Tilletia spp., which were based on ITS rDNA, EF1-alpha and a region of

RPB2 and showed T. vankyi was closely related to the wheat bunt clade (Carris et al 2007).

Phylogenetic trees were inferred independently from the concatenated anonymous loci alignment

by maximum parsimony (MP) or maximum likelihood (ML) with PAUP 4.0b10 as well as

Bayesian inference with MRBAYES v3.1.2 for Windows (Ronquist and Huelsenbeck 2003). DNA

substitution models were determined by MODELTEST 3.7 for ML method (Posada and Crandall

1998). The best fit models and parameters were chosen with Akaike information criterion (AIC).

For MP and ML analysis, heuristic search was chosen with 100 starting trees obtained via stepwise,

random addition of sequences using the tree bisection-reconnection (TBR) branch-swapping

algorithm. All aligned positions were included in the analyses. All characters were unordered and have equal weight. Monte Carlo Markov Chain (MCMC) sampling method was used for the

Bayesian inference and limited to 10 000 000 generations. The combined partitions were unlinked in STATEFREQ, REVMAT, SHAPE and PINVAR. In each analysis two independent runs were started

simultaneously, with four chains in each run (one heated chain and three cold chains). The MCMC

analyses were repeated three times. Trees were sampled at a frequency of every 1000 generation.

The first 25% of sampled trees were excluded from the final analysis. A minimum of 95% of

Bayesian posterior probabilities was set as the criterion for inferring clades. Partitioned Bremer values were calculated using TREEROT v. 3 (Sorenson & Franzosa 2007) to detect conflict of

each anonymous locus on one of the maximum parsimony trees referred from concatenated data.

To avoid the uncertain influence of missing data on incongruence test, a subset of concatenated

data with 36 taxa with full sequencing information for each locus was analyzed for compatibility

among polymorphism sites. The sequence data were collapsed into haplotypes from which the

site compatibility matrix was generated and visualized using SNAP WORKBENCH v. 2.0 (Price &

Carbone 2005, Aylor 2006).

A network was reconstructed from the concatenated data using TCS v. 1.21 (Clement et al

2000) which implements the statistical parsimony method (Templeton et al 1992). The network

method can be used to estimate data at the population genetics level when phylogenies indicate a

low level of divergence (Templeton et al 1992). TCS was used to collapse concatenated sequence

from 4 loci into haplotypes and link into a network with 95% probability using parsimony

criterion. The network was manually edited and visualized graphically using GHOSTSCRIPT v.

7.0.6 implemented in SNAP WORKBENCH (http://www.cs.wisc.edu/~ghost/). The network was

analyzed as a nested design by pooling haplotypes into 3-step clades as described by Carbone

and Kohn (2004). The information on teliospore morphology and location of each collection was

manually mapped on the network. A “gene jackknifing” method (Hackett et al 2008) was used to

detect whether the network structure was strongly impacted by individual loci by excluding one

of the loci from the concatenated sequence one at a time, and networks estimated with the

remaining three loci in TCS as described above. A NeighborNet network (Bryant & Moulton,

2004) was constructed from concatenated sequence of 4 loci to compare with the parsimony network. SPLITSTREE v. 4.11 (Huson & Bryant 2006) was also used to estimate a network with

variance estimated from original least-squares using maximum four dimensions as filter, and

bootstrap values calculated with 1000 replicates. The recombination among sequence data was

detected using RDP v. 3.44 (Martin et al 2005) with following methods: Gene conversion

(GENECONV) (Sawyer 1989), Maximum chi-square test (MAXCHI) (Smith 1992), maximum

mismatch chi-square (CHIMAERA) (Posada & Crandall 2001), sister-scanning (SISCAN) (Gibbs et

al 2000), and 3SEQ (Boni et al 2007).

Parametric bootstrapping (Huelsenbeck et al 1996) and non-parametric tests including

Shimodaira-Hasegawa (SH), Kishino-Hasegawa (KH), and Approximately Unbiased (AU) tests

(Shimodaira 2002, Shimodaira and Hasegawa 1999, Kishino and Hasegawa 1989) were used to

test the incongruence between the maximum likelihood trees from the anonymous loci and

amplified RPB2 region. The parametric test (SOWH test) was conducted on PAUP with 200 sets

of simulated data generated from SEQ-GEN (Rambaut & Grass 1997). The non-parametric tests were performed with PAUP and CONSEL 0.1j (Shimodaira & Hasegawa 2001). The hypotheses of

3 morphological species among the wheat bunt isolates in this study were also compared with the

maximum likelihood trees with parametric bootstrapping and non-parametric tests. Constraint trees were reconstructed with the Maximum Likelihood method with the sequence data set of 3 anonymous loci or RPB2 region. Three constraints were used to limit the tree topologies with morphological species assumption. The first constraint grouped wheat bunt collections into three groups based on the species recognition by the collectors. The second constraint grouped the wheat bunt collections into three groups based on the depth of spore reticulation and thickness of gelatinous sheath. The third constraint grouped the wheat bunt collections into common bunt and

dwarf bunt groups defined by the K-clustering method. Some wheat bunt collections with missing morphological data were not forced into constraints. The constraints were also illustrated in the footnotes of TABLE 3-3. The morphological species recognition methods were used to

constrain tree topologies from anonymous loci or RPB2 region, so that parametric and

non-parametric tests could be applied to compare the morphological species recognition with

phylogenetic species recognition in this study.

Sorus shape

FIG. 3-2. K-means clustering of 51 isolates of wheat bunt on Triticum aestivum based on teliospore

size and sorus shape. Standard deviation of sorus shape (x-axis) and teliospore size (y-axis) from overall means indicated. These two components explain 100% of point variability. Each data point represent a single isolate. Circles: Tilletia caries and T. laevis (common bunt) isolates; triangles: T. contraversa (dwarf bunt) isolates. Also see TABLE 3-2.

TABLE 3- 2. Species, teliospore morphology and morphological groups determined for Tilletia isolates

Voucher Teliospore morphology Sterile Sorus Morphological Classification /isolatenumbers cell shape Diam Avg Reticulation Sheath Meshes Diam Avg Collector Spore K-means (μm) Diam depth depth /spore (μm) Ratio assigned ornamentation clustering with (μm) (μm) (μm) diam species and sheath spore size and sorus shapea 35812 23–26 25.1 2–5 2–7 4–5 20.5 1.3 TCKa TCK Dwarf bunt 17-13 17–20 18.5 0.5–1 NA 5–7 13.5 1.4 T. caries T. caries Common bunt 78-49 17–21 19 0.5–1 NA 5–6 15 1.42 T. caries T. caries Common bunt 79-3 16–23 18 NA NA NA NA 1.68 T. laevis T. laevis Common bunt 90-61 18–22 19 NA NA NA 19.5 1.55 T. laevis T. laevis Common bunt 99-3 18–23 19.5 NA NA NA NA 1.56 T. laevis T. laevis Common bunt D7 22.5–25 24 2.5–4 2.5–4 4–7 20 1.23 TCK TCK Dwarf bunt D17 21–24 22 1–2 1–3 5–7 14 1.24 TCK TCK Dwarf bunt b c 54 DAR73299 21–25 22.3 0.5–3.5 0.5–4 5–8 18 1.43 T. laevis TCK Dwarf bunt DB36 22–24.5 23.4 2.5–4 2.5–4.54–5 23 1.46 TCK TCK Dwarf bunt DB295 22–25.5 24 2–4.5 2–5 5–6 16 1.35 TCK TCK Dwarf bunt DB572 20–22 21 1.5–2.5 1.5–2.55–6 16 1.7 TCK TCK Common bunt c DB688 23–25 24 2.5–3.5 2.5–3.55–6 20 1.28 TCK TCK Dwarf bunt DB9030 23–25.5 24.6 3–4.5 3–6 4–7 23 1.14 TCK TCK Dwarf bunt HB11 22–24.5 23.4 1.5–5 1.5–5 4–5 14 1.3 T. caries TCK b Dwarf bunt c HK127 22–24.5 24 2–4 2–5 5–7 17 1.54 TCK TCK Dwarf bunt J10 16–21 17.3 0.5–1 NA 5–7 14.5 1.37 T. caries T. caries Common bunt J11 20–22 20.5 0.5–2 1–2 4–7 13 1.39 T. caries T. caries Common bunt J13 16–19 17 0.5–1 NA 5–8 14 1.6 T. caries T. caries Common bunt J17 17–20 17.7 0.5–2 0.5–2 5–7 14 1.66 T. caries T. caries Common bunt L16 17–20 18.2 NA NA NA 14 1.7 T. laevis T. laevis Common bunt LMC177 23–27 24.6 2–5 2–7 5–7 19 1.15 TCK TCK Dwarf bunt LMC178 20–22.5 21.2 NA NA NA 17 2.16 T. laevis T. laevis Dwarf bunt c

LMC282 22–28 25.1 2–4 3–5 5–6 16 1.7 TCK TCK Dwarf bunt LMC295 20–25 23.4 2–2.5 2–3 6–7 13–18 1.39 TCK TCK Dwarf bunt LMC360 19–23 20.6 1–2 NA 5–7 NA 1.48 T. caries T. caries Common bunt LMC446 23–28 25.1 3–5 3–5 5–6 22 1.07 TCK TCK Dwarf bunt LMC474 18–23 21.2 1–2 NA 5–6 15 1.3 T. caries T. caries Dwarf bunt c LMC515 21–24 22.3 1–2 NA 5–7 18 1.59 T. caries T. caries Dwarf bunt c LMC516 18–21 18.9 1–2 NA 5–6 16–18 1.37 T. caries T. caries Common bunt LMC517 17–21 18.9 0.5–1 NA 5–7 15 1.42 T. caries T. caries Common bunt LMC518 20–24 21.8 0.5–1.5 NA 5–7 20 1.44 T. caries T. caries Dwarf bunt c LMC519 17–22 20.1 NA NA NA 16 1.32 T. laevis T. laevis Common bunt LMC520 20–24 22 1–2 1.5–2 5–7 15 1.37 TCK TCK Dwarf bunt LMC521 17–22 19 NA NA NA NA 1.68 T. laevis T. laevis Common bunt LMC522 22–27 25.1 2–3 2–3 5–6 19 1.23 TCK TCK Dwarf bunt LMC523 19–21 20.2 1.5–2.5 2–2.5 5–6 15 NA TCK TCK NA LMC524 16–20 18 NA NA NA NA 1.81 TCK T. laevis b Common bunt c LMC525 17–21 19 NA NA NA 15 2 T. laevis T. laevis Common bunt 55 LMC527 18–20 18.4 NA NA NA NA 1.81 T. caries T. laevis b Common bunt c S90-09 21–25 23.4 2–3.5 2–3.5 5–7 30 1.61 TCK TCK Dwarf bunt T2 18–22 19.5 0.5–1 NA 5–7 18 2.05 T. caries T. caries Common bunt TK34-1 21–24 22.3 2–4 2–4 4–7 33 1.32 TCK TCK Dwarf bunt TK52-1 NA NA NA NA NA NA NA T. laevis NA NA TK98-2 17–20 18.8 0.5–1 NA 5–7 13.5 NA T. caries T. caries NA TK108 15–20 16.5 NA NA NA 14 NA T. laevis T. laevis NA TK126-2 18–21 19 NA NA NA 14 1.94 T. laevis T. laevis Common bunt V3 17–20 18 NA NA NA NA 2.49 T. laevis T. laevis Common bunt V4 16–19 17 NA NA NA 15 1.84 T. caries T. caries Common bunt V12 16–20 18.8 0.5–1 NA 5–7 16 1.9 T. caries T. caries Common bunt V31 NA NA NA NA NA NA 1.83 T. caries NA NA V73 17–20 18.7 0.5–1.5 0.5–1.55–6 NA 2.21 T. caries T. caries Common bunt V90 15–19 17 NA NA NA 16 1.99 T. laevis T. laevis Common bunt V103 18–22 19.2 0.5–1 0.5–1.54–8 15 2.38 T. caries T. caries Common bunt

WSP35784 21–23 21.5 0.5–1.5 1.5–3 4–6 12.5 NA TCK TCK NA WSP35785 19–23 21 0.5–2 1–2.5 4–6 15–17 NA TCK TCK NA WSP63659 21–25 23.5 1–2.5 1.5–3 4–6 14–16 NA TCK TCK NA WSP63862 19–22 20.5 1–2.5 1–3 3–5 11–14 NA TCK TCK NA T. WSP70347 21–24 22.9 0.5–2.5 0.5–3.5 6–7 18 1.42 TCK b Dwarf bunt c caries WSP71300 32–37 19.5 NA NA NA NA 2.04 T. laevis T. laevis Common bunt

a. T. contraversa (TCK) isolates from hosts other than Triticum aestivum were not compared

b. Spore morphology determined to be different than species determination of collector

c. Results from k-means clustering method in conflict with species determination of collector 56

RESULTS

Morphology.—Teliospore morphology and sorus shape of specimens are summarized in TABLE

3-2. Disagreement between assessment of teliospore morphology and species assignment made by the original collector was found for five isolates (TABLE 3-2). Marginal homogeneity (MH) test

suggested no significant difference between the two grouping methods based on teliospore

morphology and species assignment by collectors (W = 1.84, p = 0.4). Isolates collected from

hosts other than Triticum aestivum were not included in the K-cluster analysis. In K-means

clustering analyses, two groups were selected as the best model based on inter-cluster variance

2 2 2 (σ k=2 = 188.12 > σ k=3 = 152.47 > σ k=4, 5…) as shown graphically in FIG. 3-2 (comparewith

original scatter plot in FIG. S-1). Cluster 1 was not significantly different from T. caries and T.

laevis identified by the collectors and is indicated as “common bunt” in TABLE III-II; cluster 2

was not significantly different from T. contraversa identified by the collectors and is indicated as

“dwarf bunt” in TABLE III-II (MH test: W = 2.76, p = 0.1) (also refer to FIG. 3-2). Placement for

five isolates by K-means cluster analysis was not significantly different from species identification based on teliospore ornamentation and presence/absence of sheath (MH test: W =

1.80, p = 0.2) (TABLE 3-2).

Anonymous loci and amplified RPB2 region.—Four wheat bunt isolates collected from different

geographic origins and identified by collectors as representing different morphological species

were selected for marker screening (TABLE 3-1: 99-3, J3, LMC295, T2). Thirty 10-mer primers

from Operon kits OPA, OPB, OPE and OPP were used to amplify DNA from the four isolates

(data not shown). Strongly amplified PCR products obtained with 15 primers were cloned and analyzed with PCR-RFLP. Cloned PCR fragments with polymorphic restriction digestion

patterns among four tester isolates were sequenced; 17 of the sequenced loci had 95–100%

sequence identity and were determined to be homologous. Specific primers were designed to

amplify these 17 loci, and three primer pairs that consistently amplified loci with more than 1% polymorphic nucleotide sites among the four isolates were selected for phylogenetic analysis.

The amplified anonymous loci were designated as A13, A16 and P18 based on the RAPD

primers used in the first step. BLASTX searches against non-redundant protein sequences in the database with the three amplified loci revealed no match with E-value threshold as 0.01.

Sequence data for the 3 anonymous loci from 60 isolates of T. caries, T. contraversa and T. laevis

and one isolate of T. vankyi as the outgroup taxon were analyzed. Locus A13 had 13 variable

sites from the 705 bp region sequenced; locus A16 had 10 variable sites from the 535 bp region

sequenced; and locus P18 had 11 variable sites from the 403 bp region sequenced. In locus P18,

three ambiguous sites that presented different nucleotide readings from forward and reverse

sequencing reactions were removed from the analysis. The PRB2 region amplified was 446 bp

with 18 variable sites among wheat bunt isolates and three outgroups. BLASTX searches

revealed no matches similar to this amplified region in the database with 0.01 as E-value

threshold except RPB2-primer-amplified sequences of Tilletia spp. from previous studies (Carris

et al 2007, Bao et al 2010).

Phylogenetic analysis.—Partition homogeneity tests of combined data to each partition as three

anonymous loci and amplified RPB2 region indicated incongruence (P = 0.001) among data.

Pairwise homogeneity tests between any two partitions revealed strong incongruence between

each of the A13, P18 and RPB2 loci. The A16 locus was compatible to the other three loci, which

also reflects the low substitution rate in this locus. Reciprocal 70% ML BS analyses on individual

topology of each partition revealed two well supported clades in the A13 phylogeny, one well

supported clade in the RPB2 phylogeny and no well supported clades in the A16 and P18

analyses. Incongruence was detected in the 70% majority rule consensus trees between A13 and

RPB2 regions. No conflict was detected in the 70% majority rule consensus trees between A13

tree and combined A13, A16 and P18 tree.

A concatenated sequence matrix with a total of 2083 characters was constructed from A13,

A16, P18 and RPB2 loci. For Maximum Parsimony method, 60 wheat bunt taxa with 49 variable

sites and 26 parsimony-informative sites were collapsed into 38 haplotypes. Five thousand

equally parsimony trees with 105 steps and CI=0.724 were saved. Partitioned Bremer support

values were calculated for each locus from one of the maximum parsimony trees. Positive

Bremer values indicate support for that branch from each locus. Negative Bremer values indicate

conflict at a locus on that branch. If both large positive and negative Bremer values were

suggested from different loci on the same branch, incongruence was indicated between those loci.

The partitioned Bremer support values calculated from concatenated data in this study supported

the majority of incongruence was between A13 and RPB2 loci (FIG. 3-3). Maximum likelihood

trees were estimated from concatenated, incongruent data to compare result with MP tree and

determine if different tree topologies are reconstructed by the different methods; ML trees are not

shown. The ML trees were estimated using HKY+I+G model (LSET BASE= (0.2123 0.2516 0.3237)

NST=2 TRATIO= 1.3517 RATES=GAMMA SHAPE= 0.9164 PINVAR= 0.9022) selected by the AIC criterion in MODELTEST. Support for clades measured as Bayesian posterior probabilities were

compared with ML bootstrap values on the maximum likelihood tree. Statistical parsimony

networks were estimated from concatenated data with TCS and nested into three 3-step clades

(FIG. 3-4).

Comparison among the phylogenetic genealogies consistently revealed two clades in the

MP trees and ML tree with 95–97% Bayesian support, but without strong bootstrap support (data

not shown). The large clade in the phylogenies completely corresponded to Clade I in the parsimony network; and the small clade in the phylogenies corresponded to the Clade II and III

in the parsimony network. The three clades constructed in the parsimony network formed a loop

structure, with 6 to 7 steps (missing haplotypes) separating each clade. A NeighborNet network

was compared with the parsimony network, and the result also indicated two large groups well

supported by bootstrap values. These two clades completely corresponded to the two clades

identified in the phylogenetic analyses. “Gene jackknifing” method was used to determine if a

single locus contributed most of the signal to the structure in the parsimony network. Four

additional networks were estimated using the parsimony method by excluding each locus one at

a time. The result showed both A13 and RPB2 loci are sources of the large loop structure in the

network. A13 was the major factor in the 2 clade structure of the network.

Comparing the structure recognized in the phylogenies and networks constructed using

different data set and algorithms, no clade exclusively corresponded to a morphological species of

wheat bunt (FIG. 3-4, FIG. S-2), although in FIG. 3-4, only three isolates with smooth spore

morphology were in Clade I. No association between clade, sorus shape and teliospore

morphology was observed in the analyses. Statistical parsimony network were also mapped with

collection locations (FIG. 3-4.b) and no clade was revealed to associate with location.

FIG. 3-3. Maximum parsimony tree inferred from concatenated data of 3 anonymous loci.

Incompatibility among loci on branches estimated by partitioned Bremer support; where large positive and negative values occurred on the same node, these are indicated by the values on the branches.

FIG 3-4. Statistical parsimony network reconstructed from 3 anonymous loci and RPB2 region.

Each haplotype is represented by a circle. Small solid circles represent inferred missing

haplotypes. The size of the circles represents the frequency of the represented haplotypes.

Isolates are indicated beside the corresponding circle or by arrows. a. Shading on each haplotype

indicates sampled frequencies of the isolates with 3 teliospore morphologies. b. The geographic origins of taxa are indicated by shading within circles, and size of circles represents frequency of haplotype.

FIG 3-5. a-d. Statistical parsimony networks reconstructed from 3 anonymous loci and RPB2

region by removing individual loci (gene jackknifing). a. Statistical parsimony network without

RPB2 region, b. Statistical parsimony network without A13 locus, c. Statistical parsimony network without A16 locus, d. Statistical parsimony network without P18 locus. e. NeighborNet

network estimated from 3 anonymous loci and RPB2 region, with bootstrap value indicated on

major branches.

TABLE 3- 3. Likelihood tests of alternative phylogenetic hypotheses

Tested Tree Topologies Alternative Tree Topologiesa AUb KHc SHd SOWHe

Maximum likelihood tree 1. Collectors' morphological species: 0.000 0.001 0.001 <0.01

referred from concatenated 2. Teliospore morphology 0.000 0.001 0.001 <0.01

data. 3. K-means clustering 0.000 0.001 0.001 <0.01

a. Constrained trees:

1. ((T. contraversa 1,…,n), (T. caries 1,…,n), (T. laevis 1,…,n), T. vankyi), species as assigned by collectors;

2. ((T. contraversa 1,…,n), (T. caries 1,…,n), (T. laevis 1,…,n), NA 1,…,n, T. vankyi), species as identified with spore 68 ornamentation and sheath;

3. ((Dwarf bunt 1,…,n), (Common bunt 1,…,n), NA 1,…,n, T. vankyi), groups as clustered with spore size and sorus shape;

b. Approximately unbiased test

c. Kishino-Hasegawa test

d. Shimodaira-Hasegawa test

e. Parametric bootstrap analysis (Swofford-Olsen-Waddell-Hillis test)

Alternative trees constrained by morphological species assignment from the original collectors, teliospore morphology and K-clustering result were compared with the ML trees reconstructed from 3 anonymous loci and RPB2-primer-amplified region. The hypothesis that alternative trees were an equally likely explanation of the data using AU, SH and KH tests was rejected (P < 0.01) when compared with both ML trees. Morphological species recognition methods were significantly different to the phylogenetic relatedness revealed by ML tree topologies from this study (TABLE 3-3).

DISCUSSION

Multilocus phylogenetic analysis and compatibility tests of sequence data among 60 wheat bunt

isolates from three anonymous regions and RPB2 region revealed significant incongruence

among four loci. Both ML consensus tree and MP consensus tree identified two lineages, which

were supported by Bayesian inference but not significantly supported by bootstrap values.

Neither lineage was exclusively associated with the morphological species recognized for the

common and dwarf bunt pathogens of wheat. Parsimony network analysis estimated a reticulated

structure among wheat bunt isolates.“Gene bootstrap” analysis indicated that conflict between

A13 and RPB2 regions contributed to the reticulated structure in the network. A reticulate

network structure may also be due to recombination, but the tests that were used failed to detect

recombination among the sequence datasets. Three clades were recognized by nested design

method, and none of the clades was exclusively associated with the wheat bunt morphological species. Overall, phylogenetic species recognition suggests a different evolution history than the recognition of three morphological species, T. caries, T. contraversa and T. laevis, among the geographically diverse set of isolates used in this study, although support for some substructure within the wheat bunt pathogen clade is indicated as identified in previous studies (Carris et al

2007, Bao et al 2010).

The lack of correlationbetween phylogenetic species and morphological species is supported by previous studies examining the relationship among the wheat bunt pathogens and related species. Russell and Mills (1993, 1994) compared wheat bunt isolates from Oregon

(collection M1), Pakistan and Turkey (collections M2–M7) and observed that two Pakistan collections were intermediate between common bunt and dwarf bunt in taxonomic criteria used for species determination. Shi et al (1996) used RAPD loci to analyze the genetic relationships among

66 isolates of wheat bunt fungi, and their results showed most of the dwarf bunt isolates fell into one group and a mixture of dwarf and common bunt isolates in the second group; neither group, however, was supported by bootstrap analysis. Shi et al (1996) also found two distinct restriction digest patterns in 5.8s and ITS rDNA among wheat bunt isolates. One haplotype was associated with 85% of the dwarf bunt isolates, and the second haplotype was associated with all of the common bunt isolates and four dwarf bunt isolates.

Partition homogeneity tests in this study suggested strong incongruence between any pair combination of A13, A16 and the amplified RPB2 region. However, the partition homogeneity test has been shown to be problematic when analyzed loci exhibit low divergence or unequal substitution rates occur among different loci (Barker & Lutzoni 2002, Darlu & Lecointre 2002).

In this study, individual trees from A16 and P18 loci did not contain any well-supported clades.

Bremer support values showed strong conflict between A13 and RPB2 loci in the concatenated data phylogeny and less conflict between A13 and P13 loci. The compatibility matrix generated from 36 fully sequenced taxa support that the major source of incongruence was between A13 and RPB2. Because of incongruence among loci, a more appropriate analysis of wheat bunt collections is based on network methods which assume recombination and incomplete lineage

sorting. The large loop structure revealed by the parsimony network method was not supported by NeighborNet network. “Gene jackknifing” networks based on different subsets of data

suggest that the loop structure was related to the A13 and RPB2 loci in the dataset.

The network analysis of data in this study provides preliminary evidence for geographic

association with the 3 clades in parsimony network of the wheat bunt collections (FIG. 3.4b). All

North American isolates of T. laevis were exclusively in Clade II with four isolates sharing the

same haplotype (Hp2, FIG. 3.4b), and all the North American isolates of T. contraversa except

LMC520 were in Clade I with nine isolates from the North American Pacific Northwest (PNW –

Washington, Oregon, Idaho and British Columbia) represented by the same haplotype (Hp1, FIG.

3.4b). The type and paratype specimens of T. brevifaciens (WSP35785, WSP35784) shared haplotype 1 with other 7 PNW dwarf bunt isolates. The relatively high frequency of these two haplotypes indicates a potential oversampling of this area, or a potential founder effect.

Teliospore morphology and physiological traits relating to spore germination and host stunting are used as key criteria to distinguish the common and dwarf bunt pathogens. However, these characters are continuously distributed and the lack of a clear species boundary has been debated since Fischer’s (1952) description of the dwarf bunt pathogen as T. brevifaciens. Holton and Kendrick (1956) stated “virtually every characteristic used in separation of dwarf bunt (T. contraversa) from common bunt (T. caries) is one of degree only”. Although considered an important identifier of dwarf bunt, host stunting is not exclusively associated with T. contraversa.

Stunting can be a symptom of some races of T. caries or T. laevis on certain varieties of wheat

(Holton and Rodenhiser 1942). Teliospore ornamentation and thickness of gelatinous sheaths are also highly variable between isolates, and among teliospores within isolates (Holton and Kendrick

1956). Teliospore germination requirements are more difficult to ascertain, requiring 6 wk or

longer for T. contraversa, and cannot be determined for old or otherwise nonviable collections.

Moreover, the range of germination conditions between common bunt and dwarf bunt were shown to be overlapping (Russell and Mills 1994). For example, light is not necessary for the germination of some isolates of dwarf bunt (Meiners and Waldher 1959).

In our study, wheat bunt isolates were compared using three methods of morphological species determination. There were no statistically significant differences between any two methods, and the discrepancies indicated that with some specimens, the original collectors’

identification was not based on microscopic features. For example, LMC524 was identified as

dwarf bunt by host stunting, but the teliospores are smooth. Chinese isolate HB11 was identified

as T. caries and germinated at 15 C, but has teliospore morphology more typical of T. contraversa.

Inoculation studies on winter wheat revealed low infection rates with the HB11 strain relative to

typical common bunt strains (Blair Goates, personal communication).

Teliospore size and sorus shape have not been emphasized as characters in wheat bunt

identification in previous studies. Kühn (1874) observed that spores of T. contraversa were on

average 1 µm smaller than those of T. caries, and Fischer (1952) noted the sori of T. brevifaciens

were “hard and compact” relative to those of T. caries. In the present study, a k-means clustering

test was used to classify wheat bunt isolates in subgroups using spore size and sorus shape (FIG.

3.2, TABLE 3-2). The results supported that the recognition of two subgroups explained more

intergroup variation than three groups. Spore ornamentation was not considered in this test.

The grouping results based on spore ornamentation, k-means clustering and collector’s designation were compared with the results from the phylogenetic analyses by the parametric and non-parametric tests on alternative tree topologies in TABLE 3-3, and none of the alternative

trees was supported by ML phylogeny. Although some previous studies have provided evidence

supporting the recognition of three morphological species of wheat bunt pathogens, their

conclusions must be assessed in the context of number and diversity of the isolates that were

analyzed (Stockwell and Trione 1986, Liang et al 2005, Liu et al 2009). For example, Liu et al

(2009) developed an anonymous marker that purportedly distinguished T. contraversa from

common bunt species. The authors tested the marker using 15 common bunt isolates obtained from

Chinese collaborators, without indicating the geographic origin of the isolates, and T. contraversa

isolates representing physiological races D1-D17, obtained from Blair Goates (USDA ARS,

Aberdeen, Idaho). Number of isolates and the selection of the loci are two common problems in

molecular phylogenetic studies that may produce bias in the results (Leigh et al 2003).

In this study, incongruence was found among anonymous loci and amplified RPB2 region,

which suggests different evolutionary histories for each individual region. Recombination among

tested loci could be a reason of the incongruence among data from wheat bunt isolates in this

study. The potential for recombination among the three morphological wheat bunt species was

shown by a series of studies based on artificial inoculation (Flor 1932, Silbernagel 1964,

Pimentel et al 2000). Tests used in this study failed to detect reconbination events among the concatenated sequence data from four loci. The RPB2 region sequenced in this study does not show homology to other fungal RPB2 data available in GenBank except those from Tilletia spp.

(data not shown); however, it has been shown to have similarity to other fungal intein coding regions (Poulter et al 2007). Inteins are genetic elements that are transcribed and translated with their host protein, but have a self-excising mechanism (Poulter et al. 2007). The putative RPB2 intein appears to be widely distributed among the Pooid-infecting species of Tilletia (Carris et al.

2007, Bao et al 2010). But it is still unknown that whether recent horizontal transfer has played role in this region so its evolutionary history is different from other regions of the genome.

Phylogenetic analyses on Tilletia and related genera utilizing the nuclear large subunit rDNA distinguished a well-supported clade associated with subfamily Pooideae hosts, but with no internal support within the clade (Castlebury et al 2005). In contrast, the combined sequence analysis of ITS rDNA, elongation factor 1 alpha (EF1α), and RPB2 revealed well-supported clades corresponding to distinct species of Tilletia with narrow host ranges, but did not support recognition of T. caries, T. contraversa, and T. laevis as distinct phylogenetic species (Carris et al

2007).

Conflict among recognition of phylogenetic species, biological species, and morphological species is common in systematic studies of fungi, but more typically additional phylogenetic species are recognized within biological or morphological species (Dettman et al

2003). For instance, multilocus phylogenetic analyses supported distinct, host specific clades within the morphological species Fusarium solani (O'Donnell 2000) and Microbotryum violaceum (Gac et al 2007). In the small-spored Alternaria species complex, morphological and phylogenetic lineages were not found to be associated (Andrew et al 2009). In the present study, clades identified from nested network analysis and lineages identified in multilocus phylogenies were not found to be exclusively associated with morphological characters or geographic origin of the isolates. The wheat bunt pathogens may represent a single evolutionary species, or a recently diverged species complex in which gene flow occurs among taxa, but analysis of additional loci and/or additional isolates is necessary to resolve the conflict between morphological species and phylogenetic lineages. Although no definitive conclusions can be drawn from the topologies generated in this study, the analyses consistently fail to support the morphological species concept that has been widely applied to recognize three wheat bunt pathogens.

Morphological species recognition of the wheat bunt pathogens does not support the

evolutionary history of these organisms, and additional study is needed to determine what species

epithet(s) should be applied if more than one lineage is supported by additional analyses. Tilletia

caries, the type species of the genus, with the type specimen on infected Triticum aestivum from

France, would be the valid name for a single lineage of wheat bunt. Interestingly, Vánky (1994) reported that the isotype specimens of T. laevis in Herb. HUV and UPS contain both smooth-walled and reticulate teliospores within the same sori. Tilletia contraversa, as previously discussed, is based on infected Elymus repens from Germany, and has not been established as conspecific with the dwarf bunt pathogen of wheat by molecular phylogenetic study. A companion study in Chapter 4 addresses the conspecificity of the dwarf bunt pathogen of wheat and T. contraversa on Elymus repens. The ages of type specimens of T. caries, T. contraversa and T. laevis preclude their use for molecular analyses. The status of T. contraversa as a quarantined pathogen makes the unresolved issue of one, two or three species of wheat bunts an economically important question warranting further study.

LITERATURE CITED

Andrew M, Peever TL, Pryor BM. 2009. An expanded multilocus phylogeny does not resolve

morphological species within the small-spored Alternaria species complex. Mycologia

101:95-109.

Aylor DL, Price EW, Carbone I. 2006. SNAP: Combine and Map modules for multilocus

population genetic analysis. Bioinformatics 22:1399-1401.

Bailey CD, Hughes CE, Harris SA. 2004. Using RAPDs to identify DNA sequence loci for species

level phylogeny reconstruction: an example from Leucaena (Fabaceae). Syst Bot 29:4-14.

Bamberg RH, Holton CS, Rodenhiser HA, Woodward RW. 1947. Wheat dwarf bunt depressed by

common bunt. Phytopathology 37:556-560.

Banowetz GM, Doss RP. 1994. A comparison of polypeptides from teliospores of Tilletia

controversa (Kuhn) and Tilletia tritici (Bjerk) Wint. Journal of Phytopathology (Berlin)

140:285-292.

Banowetz GM, Trione EJ, Krygier BB. 1984. Immunological comparisons of teliospores of two

wheat bunt fungi, Tilletia species, using monoclonal antibodies and antisera. Mycologia

76:51-62.

Bao X, Carris LM, Huang G, Luo J, Liu Y, Castlebury LA. 2010. Tilletia puccinelliae, a new

species of reticulate-spored bunt fungus infecting Puccinellia distans. Mycologia

102:613-623.

Barker FK, Lutzoni FM. 2002. The utility of the incongruence length difference test. Syst Biol

51:625-637.

Boni MF, Posada D, Feldman MW. 2007. An exact nonparametric method for inferring mosaic

structure in sequence triplets. Genetics 176 :1035-1047.

Bryant D, Moulton V. 2004. Neighbor-net: an agglomerative method for the construction of

phylogenetic networks. Mol Biol Evol 21:255-265

Carbone I, Kohn L. 2004. Inferring Process from Pattern in Fungal Population Genetics. Applied

Mycology and Biotechnology 4:29-58

Carpenter-Boggs L, Carris LM, Castlebury LA. 2003. Multi-locus phylogenetic analysis of Tilletia

species infecting hosts in grass subfamily Pooideae (Abstract). Inoculum 54:15.

Carris LM, Castlebury LA, Huang G, Alderman SC, Luo J, Bao X. 2007. Tilletia vankyi, a new

species of reticulate-spored bunt fungus with non-conjugating basidiospores infecting

species of Festuca and Lolium. Mycol Res 111:1386-1398.

Castlebury LA, Carris LM, Vánky K. 2005. Phylogenetic analysis of Tilletia and allied genera in

order Tilletiales (Ustilaginomycetes; Exobasidiomycetidae) based on large subunit nuclear

rDNA sequences. Mycologia 97:888-900.

Castrillo LA, Vandenberg JD, Wraight SP. 2003. Strain-specific detection of introduced Beauveria

bassiana in agricultural fields by use of sequence-characterized amplified region markers.

J Invertebr Pathol 82:75-83.

Clement M, Posada D, Crandall KA. 2000. TCS: A computer program to estimate gene

genealogies. Mol Ecol 9:1657-1659.

Darlu P, Lecointre G. 2002. When does the incongruence length difference test fail? Mol Biol

Evol 19:432-437.

Dettman J, Jacobson D, Taylor J. 2003. A multilocus genealogical approach to phylogenetic

species recognition in the model eukaryote Neurospora. Evolution 57:2703-2720.

Durán R, Fischer GW. 1961. The genus Tilletia. Washington State University.

Fischer GW. 1952. Tilletia brevifaciens sp. nov., causing dwarf bunt of wheat and certain grasses.

Res Stud State Coll Wash 20:11-14.

Flor HH. 1932. Heterothallism and hybridization in Tilletia tritici and T. levis. Jour Agr Res

44:49-58.

Gac ML, Hood ME, Fournier E, Giraud T. 2007. Phylogenetic evidence of host-specific crypic

species in the anther smut fungus. Evolution 61:15-26.

Gibbs MJ, Armstrong JS, Gibbs AJ. 2000. Sister-scanning: a Monte Carlo procedure for assessing

signals in recombinant sequences. Bioinformatics 16:573-582.

Goates BJ. 1996. Common bunt and dwarf bunt. in: Bunt and Smut Diseases of Wheat: Concepts

and Methods of Disease Management. pp. 12-25 p. Wilcoxson RD, Saari EE, eds.

CIMMYT, Mexico D.F.

Hackett SJ, Kimball RT, Reddy S, Bowie RCK, Braun EL, MJ Braun MJ, Chojnowski JL, Cox

WA, Han K-L,Harshman J, Huddleston CJ, Marks BD, Miglia KJ, Moore WS, Sheldon

FH, Steadman DW, Witt CC, Yuri T. 2008. A Phylogenomic Study of Birds Reveals Their

Evolutionary History. Science 320:1763-1768.

Hall T. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program

for Windows 95/98/NT. Nucleic Acids Symposium Series 41:95-98.

Hoffmann JA. 1982. Bunt of wheat. Plant Dis 66:979-986.

Holton CS, Kendrick EL. 1956. Problems in the delimitation of species of Tilletia occurring on

wheat. Res Stud State Coll Wash 24:318-325.

———, Rodenhiser HA. 1942. New physiologic races of Tilletia tritici and T. levis

Phytopathology 32:117-129.

Huelsenbeck JP, Hillis DM, Jones R. 1996. Parametric bootstrapping in molecular phylogenetics:

applications and performance. Molecular zoology: advances, strategies, and protocols

3:19-45.

Huson DH, Bryant D. 2006. Application of phylogenetic networks in evolutionary studies. Mol

Biol Evol. 23:254-267.

Ireland CT, Ku HH, Kullback S. 1969. Symmetry and marginal homogeneity of an r × r

contingency table. J Amer Statistical Assoc 64:1323-1341.

Jany J-L, Garbaye J, Martin F. 2002. Cenococcum geophilum populations show a high degree of

genetic diversity in beech forests. New Phytol 154:651-659.

Kühn J. 1874. Tilletia contraversa. Hedwigia 13:188-189. Rabenhorst, L. Fungi europaei exs.

1801-1900. Dresden.

Kawchuk LM, Kim WK, Nielsen J. 1988. A comparison of polypeptides from the wheat bunt fungi

Tilletia laevis Tilletia tritici and Tilletia controversa. Canadian Journal of Botany

66:2367-2376.

Leigh J, Seif E, Rodriguez N, Jacob Y, Lang B. 2003. Fungal evolution meets fungal genomics. in:

Handbook of fungal biotechnology. 145-162 p. Arora DK, Bridge PD, Bhatnagar D, eds.

Marcel Dekker Inc., New York City.

Liang H, Zhang G, Chen W, Liu T. 2005. Sequence analysis of rDNA-ITS of Tilletia controversa

and its related species. Acta Phytopathologica Sinica 35:181-183.

Liu JH, Gao L, Liu TG, Chen WQ. 2009. Development of a sequence-characterized amplified

region marker for diagnosis of dwarf bunt of wheat and detection of Tilletia controversa

Kühn. Lett Appl Microbiol 49:235-240.

MacQueen J. 1967. Some methods for classification and analysis of multivariate observations.

Proceedings of 5th Berkeley Symposium on Mathematical Statistics and Probability,

Berkley, California, University of California Press, 1:281-297.

Maddison WP, Maddison DR. 2006. Mesquite: a modular system for evolutionary analysis.

Version 2.72.

Martin DP, Williamson C, Posada D. 2005. RDP2: recombination detection and analysis from

sequence alignments. Bioinformatics 21:260-262.

Mathre DE. 1996. Dwarf bunt: Politics, identification, and biology. Annu Rev Phytopathol

34:67-85.

Meiners JP, Waldher JT. 1959. Factors affecting spore germination of twelve species of Tilletia

from cereals and grasses. Phytopathology 49:724-728.

Newcombe G, Stirling B, McDonald S, Bradshaw Jr HD. 2000. Melampsora ×columbiana, a

natural hybrid of M. medusae and M. occidentalis. Mycol Res 104:261-274.

O'Donnell K. 2000. Molecular phylogeny of the Nectria haematococca-Fusarium solani species

complex. Mycologia 92:919-938.

Paran I, Michelmore R. 1993. Development of reliable PCR-based markers linked to downy

mildew resistance genes in lettuce. Theor Appl Genet 85:985-993.

Peever TL, Su G, Carpenter-Boggs L, Timmer LW. 2004. Molecular systematics of

citrus-associated Alternaria species. Mycologia 96:119-134.

Pimentel G, Carris LM, Peever TL. 2000. Characterization of interspecific hybrids between Tilletia

controversa and T. bromi. Mycologia 92:411-420.

Posada D, Crandall KA. 1998. MODELTEST: Testing the model of DNA substitution.

Bioinformatics (Oxford) 14:817-818.

———. ———. 2001. Evaluation of methods for detecting recombination from DNA sequences:

computer simulations. PNAS 98:13757-13762.

Price EW, Carbone I. 2005. SNAP: workbench management tool for evolutionary population

genetic analysis. Bioinformatics 21:402-404.

Purdy LH, Kendrick EL, Hoffmann JA, Holton CS. 1963. Dwarf bunt of wheat. Annual Reviews

in Microbiology 17:199-222.

Rambaut A, Grass NC. 1997. Seq-Gen: an application for the Monte Carlo simulation of DNA

sequence evolution along phylogenetic trees. Bioinformatics 13:235-238.

Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed

models. Bioinformatics 19:1572-1574.

Rozen S, Skaletsky H. 2000. Primer3 on the WWW for general users and for biologist

programmers. Methods Mol Biol 132:365-386.

Russell BW, Mills D. 1993. Electrophoretic karyotypes of Tilletia caries, T. controversa, and their

F1 progeny: Further evidence for conspecific status. Mol Plant-Microbe Interact 6:66-74.

———, ———. 1994. Morphological, physiological, and genetic evidence in support of a

conspecific status for Tilletia caries, T. controversa, and T. foetida. Phytopathology

84:576-582. Sawyer SA. 1989. Statistical tests for detecting gene conversion. Mol Biol Evol 6:526-538.

Shi YL, Loomis P, Christian D, Carris LM, Leung H. 1996. Analysis of the genetic relationships

among the wheat bunt fungi using RAPD and ribosomal DNA markers. Phytopathology

86:311-318.

Shimodaira, H. & Hasegawa, M. 2001. CONSEL: for assessing the confidence of phylogenetic

tree selection. Bioinformatics 17:1246-1247.

Silbernagel MJ. 1964. Compatibility between Tilletia caries and T. controversa. Phytopathology

54:1117-1120.

Smith JM. 1992. Analyzing the mosaic structure of genes. J Mol Evol 35:126-129.

Sorenson MD, Franzosa EA. 2007. TreeRot, v. 3. Boston MA: Boston University.

Stockwell VO, Trione EJ. 1986. Distinguishing teliospores of Tilletia controversa from those of T.

caries by fluorescence microscopy. Plant Dis 70:924-926.

Swofford DL. 2002. PAUP*: phylogenetic analysis using parsimony (* and other methods),

version 4.0 b10. Sinauer Associates, Sunderland, MA.

Templeton AR, Crandall KA, Sing CF. 1992. A cladistic analysis of phenotypic associations with

haplotypes inferred from restriction rndonuclease mapping and DNA sequence data. III.

Cladogram estimation. Genetics 132: 619-633

Trione E, Krygier B. 1977. New tests to distinguish teliospores of Tilletia controversa, the dwarf

bunt fungus, from spores of other Tilletia species. Phytopathology 67:1166-1172.

Trione EJ. 1964. Isolation and in vitro culture of the wheat bunt fungi Tilletia caries and T.

controversa. Phytopathology 54:592-596.

———. 1982. Dwarf bunt of wheat and its importance in international wheat trade. Plant Dis

66:1083-1089. Young PA. 1935. A new variety of Tilletia tritici in Montana. Phytopathology 25:40.

Zharkikh A, Li W-H. 1992. Statistical properties of bootstrap estimation of phylogenetic

variability from nucleotide sequences. I. Four taxa with a molecular clock. Mol Biol Evol

9:1119-1147.

CHAPTER FOUR

THE HOST RANGE AND SPECIES CONCEPT OF TILLETIA CONTRAVERSA

ABSTRACT

Tilletia contraversa, widely recognized as causing dwarf bunt of wheat, was originally described

by Julius Kühn from Elymus repens in Germany in 1862 and not recognized as a wheat pathogen

until the mid-1950s. Most smut fungi are host-specific with narrow host ranges, but T.

contraversa is exceptional in having a broad host range that includes 65 species of grasses

(family Poaceae). A number of grass species were added to the recognized host range for T.

contraversa in the 1950s after outbreaks of dwarf bunt in the Pacific Northwestern U.S. The

objective of this study was to determine whether collections identified as T. contraversa infecting

wheat and other grasses in subfamily Pooideae are conspecific. DNA was extracted and

amplified from 65 isolates of T. contraversa and common bunt species T. caries and T. laevis

collected 1945–2006 from Eurasia and North America. The collection included 15 isolates on

species of Agropyron, Arrhenatherum, Bromus, Festuca, Secale, Thinopyrum and Triticum from dwarf bunt outbreaks in Idaho, Oregon, and Washington in the 1950s. Phylogenetic analyses based on three anonymous loci (A13, A16 and P18), ITS rDNA, elongation factor 1 alpha

(EF1-alpha), and a portion of the second largest subunit of RNA polymerase II (RPB2) are presented. Maximum likelihood phylogenetic trees were estimated independently for

RPB2/ITS/EF1-alpha loci, and A16/P18/ITS/EF1-alpha loci. NeighborNet networks were estimated from A16/P18/ITS/EF1-alpha loci and all 6 loci with “gene jackknifing” method. All the phylogenetic trees and the networks supported a wide host range for T. contraversa, with

identical anonymous loci genotypes occurring in North American isolates on five different host genera. The analyses failed to support a lineage corresponding to the dwarf bunt pathogen of wheat that was distinct from the common bunt pathogens of wheat, T. caries and T. laevis. Five

Eurasian isolates of T. contraversa on Elymus repens and Thinopyrum intermedium were consistently placed in a clade distinct from the majority of T. contraversa isolates in the phylogenetic trees and networks, indicating that substructure exists within dwarf +common bunt assemblage and leaving open the possibility that T. contraversa represents a distinct genotype, but one that does not correspond to the accepted concept applied to the dwarf bunt pathogen of wheat.

INTRODUCTION

The nutritional strategies of biotrophic fungi are highly specialized with their plant hosts

(Mendgen and Hahn 2002). The complex mechanisms of biotrophy require specific associations between fungi and plant, and these associations can be a reflection of their evolutionary history

(Savile 1979). As biotrophic pathogens, most of the ca 1500 species of smut fungi (Basidiomycota;

Ustilaginomycotina) have a relatively narrow host range, with host specificity evident at the genus level (Stamatakis et al 2007). Molecular studies also support a narrow species concept that encompasses a few or even single host species (Begerow et al 2002, Stoll et al 2003, Carris et al

2007). Based on a quantitative analysis on the 600 smut species included in European Smut Fungi

(Vánky 1994), Begerow et al (2004) found 55% of the species occur on a single host, 86% on five or less hosts, and 93% on fewer than 10 hosts. Only 11 of the 600 smut species were reported to occur on more than 20 hosts (Begerow et al 2004). Tilletia contraversa Kühn is one of the exceptional, broad host range smut fungi, recognized as infecting species in 16 grass genera

(Vánky 1994).

Tilletia contraversa is widely recognized as the causal agent of dwarf bunt of wheat, a pathogen restricted to winter-sown wheat in areas with persistent snow cover (Trione 1982).

However, T. contraversa was originally described from infected ovaries of Elymus repens (L.)

Gould (‘quackgrass’) in Germany (Kühn 1874). In his comments on the new species, Kühn compared the quackgrass bunt with the wheat bunt pathogen T. caries (DC) Tul., observing that the spores of T. contraversa are more uniformly spherical and on average 1 µm larger than the spores of T. caries. He also noted that the spores of T. contraversa had deeper exospore ornamentation and wider diameter areolae, and when placed under the same conditions as T. caries, the spores did not germinate within 60-72 hours as typical of the wheat bunt species (Kühn 1874).

Tilletia caries and its close relative Tilletia laevis Kühn cause a disease known as common bunt of wheat; both species infect winter- and spring-sown wheat throughout wheat-growing regions of the world (Hoffmann 1982). Dwarf bunt of wheat was not recognized as a distinct disease until 1935, when Young (1935) described a new type of wheat bunt in Montana characterized by teliospores with prominent reticulations, bunt balls (sori) that were hard and solid instead of loose and powdery as in common bunt, excess tillering and extreme dwarfing of infected plants. One year later, Holton and Heald (1936) reported a similar race had been prevalent in some regions of the Pacific Northwestern U.S. (PNW) for years and locally recognized as “short smut”.

Subsequent studies indicated that this pathogen was present and well-established in the northeastern U.S. by 1931 (Fischer and Tyler 1952), and in Germany as early as 1930 (Warmbrunn

1952). As part of investigations on the dwarf bunt pathogen, Holton (1941) demonstrated that this new “race” of wheat bunt was able to hybridize with T. caries and T. laevis and presented evidence for segregation of teliospore morphology and host dwarfing among the hybrid lines. In 1952, G. W.

Fischer formally described the dwarf bunt pathogen as a new species, T. brevifaciens Fischer, and

designated the holotype specimen on Thinopyrum intermedium and paratype on Triticum aestivum

(Fischer 1952). In addition to differences in host symptomology, teliospore exospore and

germination noted by previous authors, Fischer also emphasized the presence of a 1.5-4 µm thick,

hyaline gelatinous sheath enveloping teliospores as a distinctive feature of T. brevifaciens but absent in collections of T. caries (Fischer 1952). Conners (1954) showed that spores with intermediate morphology between those in the types of T. contraversa and T. brevifaciens were

present in collections on Elymus repens (as Agropyron repens), Thinopyrum intermedium (as

Agropyron glaucum and Triticum glaucum) and Th. ponticum (as Agropyron rigidum var.

tomentosum), and concluded that T. brevifaciens was not sufficiently distinct from T. contraversa

to warrant recognition as a distinct species. However, Conners (1954) observed that T. brevifaciens

and T. contraversa could be distinguished because “the spores of T. contraversa in the type

collection are paler and slightly smaller than those of T. brevifaciens”. He also noted that a wide

host range had been reported for T. brevifaciens by several authors, including Fischer (1952),

Fischer & Tyler (1952), and Warmbrunn (1952). Durán and Fischer (1956) accepted T.

brevifaciens as a synonym of T. contraversa, and developed complex criteria to delimit T.

contraversa and distinguish it from T. caries. Amended slightly from Durán and Fischer (1956), in

brief, T. contraversa is distinguished from T. caries by host stunting, excessive tillering of

infected plants, hard and rounded sori, deeper exospore ornamentation with larger diameter

areolae, a thick gelatinous sheath enveloping teliospores, low temperature (3-8 C) and light

requirements for spore germination (Goates 1996). Teliospores of T. caries have a relatively thin

exospore reticulum with narrower areolae, lack a conspicuous gelatinous sheath and germinate at

5-15 C without light. Tilletia laevis is similar to T. caries in host symptoms and spore

germination, but the teliospores are smooth rather than reticulate.

After the dwarf bunt fungus was recognized as a distinct species from T. caries, a number of

studies were published that expanded the known host range of T. contraversa. Collections of

Tilletia spp. on different grass hosts were reexamined, and additional species of Tilletia were

placed in synonymy with T. contraversa including T. calospora and T. elymicola on Alopecurus

and Elymus, respectively; T. hordei, T. hordeina, T. pancicii, and T. trabutii on species of Hordeum;

and collections of T. tritici, T. secalis and T. lolii occurring on hosts other than wheat were transferred to T. contraversa (Durán and Fischer 1956, Fischer and Durán 1956). New host records

were also published based on infected hosts in grass seed production fields and wheat bunt

screening nurseries (Hardison and Corden 1952, Woodward et al 1952, Hardison 1954, Hardison

et al 1955, Hardison and Meiners 1956, Meiners et al 1956, Meiners and Hardison 1957, Dewey

and Tingey 1958, Hardison et al 1959). By 1963, 65 grass species representing 4 tribes in family

Pooideae were reported as hosts for T. contraversa, including species of Aegilops, Agropyron,

Bromus, Elymus, Festuca, Hordeum, Lolium, and Triticum (Purdy et al 1963). Notably, 40 host

records were recognized in the PNW in the 1950s (APPENDIX TABLE S-1). A serious outbreak of

dwarf bunt was reported on Tualatin oat grass (Arrhenatherum elatius) and other cultivated forage

grasses including Agropyron cristatum, Thinopyrum ponticum (as Agropyron elongatum), and

Festuca rubra in the Elgin area (Union County, Oregon) (Hardison and Corden 1952, Hardison

1954). Additional grass species were later added to the host range list for T. contraversa, most of

which were based on experiments in bunt screening nurseries with artificially infested soil or spore

inoculations (APPENDIX TABLE S-1). Most bunt collections were identified based on host

symptoms and/or teliospore morphology in studies expanding the host range for T. contraversa.

Several grasses reported as hosts for T. contraversa did not become infected when inoculated in

later studies (Hardison et al 1959). For example, Meiners (1955) conducted cross inoculation

studies using three dwarf bunt collections from Arrhenatherum elatius, Thinopyron intermedium and volunteer Elgin wheat collected in the Elgin, Oregon area in 1952. Six to forty-two percent of

Elgin wheat plants were infected when inoculated with teliospores from all three collections, but no infection resulted in inoculated A. elatius and Th. intermedium regardless of the source of the inoculum (Meiners 1955).

Host specificity of Tilletia species has been supported by a multilocus phylogenetic study based on sequence data from ITS rDNA, elongation factor 1 alpha (EF1-alpha), and a portion of the second largest subunit of RNA polymerase II (RPB2) (Carris et al 2007, Bao et al 2010). In those studies, bunt isolates were placed into distinct clades with specific association to host.

Several collections identified as T. contraversa based on teliospore morphology and infecting

Hordeum spp., Secale cereale, and Th. intermedium were shown to be genetically distinct from the dwarf bunt pathogen of wheat and were recognized as distinct species T. trabutii, T. secalis and T. brevifaciens, respectively (Carris et al 2007, Bao et al 2010). Similarly, collections of T. caries, T. contraversa and T. laevis from wheat formed one well-supported clade (Carris et al 2007, Bao et al

2010).

TABLE 4-1. Isolates used in Chapter 4.

Taxon Voucher/ Years Hosts Origins Collectors Culture numbers T. bromi WSP 71271 1992 Bromus tectorum US L. Carris (Brockmuller) (LMC 148) L. (Washington) Brockmuller T. bromi WSP 71272 1992 Bromus arvensis US (WY) L. Carris (LMC 167) L. T. caries WSP 71304 1997 Triticum aestivum Australia G. Murray (DAR -1999 73302) T. caries J-10 1981 Triticum aestivum Sweden L. Johnsson T. caries J-11 1987 Triticum aestivum Sweden L. Johnsson T. caries J-17 1996 Triticum aestivum Sweden L. Johnsson T. caries LMC516 2006 Triticum aestivum US (Oregon) X. Chen T. caries WSP 41412 1953 Triticum aestivum US S. (Washington) Hashimoto T. caries WSP 47237 1953 Triticum aestivum Finland L. Kari T. caries WSP 47286 1936 Triticum aestivum Finland L. Kari T. caries WSP 49568 1959 Triticum aestivum US T. Filer (Washington) T. caries WSP 71303 1995 Triticum aestivum Sweden L. Johnsson (LMC J-19) T. contraversaa HUV 20.802 2004 Thinopyrum Poland K. Vánky intermedium × Elymus repens T. contraversaa LMC 514 1952 Th. ponticum US (Oregon) J. Hardison T. contraversaa WSP 35764 1952 Secale cereale US (Utah) R. Woodward T. contraversaa WSP 35784 1950 Th. intermedium US (Idaho) E. Horning T. contraversaa WSP 35785 1946 Triticum aestivum US J. Meiners L. (Washington) T. contraversaa WSP 35786 1952 Arrhenatherum US (Oregon) R. Wendel elatius T. contraversaa WSP 35787 1952 Th. intermedium US (Oregon) R. Wendel T. contraversaa WSP 40312 1945 Elymus US (Idaho) G. Fischer trachycaulus T. contraversaa WSP 40406 1950 Th. intermedium US (Idaho) E. Horning T. contraversaa WSP 40667 1952 Arrhenatherum US (Oregon) R. Wendel elatius T. contraversaa WSP 40685 1952 Th.intermedium US (Oregon) R. Wendel T. contraversaa WSP 41411 1953 Triticum aestivum US S. (Washington) Hashimoto

T. contraversaa WSP 63652 NA Hordeum Turkey NA bulbosum L. T. contraversaa WSP 63656 1953 Triticum sp. Turkey NA T. contraversaa WSP 63659 1954 Triticum aestivum Germany NA T. contraversaa WSP 63665 1954 Agropyron US (Idaho) J. Meiners cristatum T. contraversaa WSP 63667 1954 Agropyron US (Idaho) J. Meiners desertorum T. contraversaa WSP 63680 1956 Bromus US J. Meiners marginatus (Washington) T. contraversaa WSP 63683 1955 Hordeum sp. Iran G. Fischer T. contraversaa WSP 63687 1956 Secale cereale US (NY) L. Tyler T. contraversaa WSP 68945 1999 Thinopyrum Austria K. Vánky (Vanky 412) intermedium

T. contraversa WSP 71280 1994 Triticum aestivum US (Oregon) L. Carris (LMC 282) T. contraversa DB36 1957 Triticum aestivum US B. Goates (Washington) T. contraversa DB90-30 ? Triticum aestivum US (Oregon) B. Goates T. contraversa J-1 1991 Elymus repens Sweden L. Johnsson T. contraversa J-8 1984 Elymus repens Sweden L. Johnsson T. contraversa S9009 1990 Triticum aestivum US (Idaho) J. Sitton T. contraversa WSP 44943 1957 Festuca US (Oregon) J. Meiners idahoensis T. contraversa WSP 49569 1959 Triticum aestivum US (Idaho) J. Hoffmann T. contraversa WSP 55699 1965 Triticum aestivum US Purdy (Washington) T. contraversa WSP 63766 1953 Elymus repens Turkey Gassner T. contraversa WSP 63862 1954 Triticum aestivum Switzerland H. Aebi T. contraversa WSP 69062 1984 Triticum aestivum Germany K. Vánky (Vanky 528) T. contraversa WSP 70123 1999 Elymus repens Austria A. Drescher T. contraversa DB460 1978 Thinopyrum US B. Goates intermedium (Washington) T. contraversa TK52-2 1979 Thinopyrum sp. Turkey B. Goates T. contraversa TK83-2 1979 Elymus repens? Turkey B. Goates T. elymi Diet. WSP 71274 1992 Elymus glaucus US (WY) L. Carris & Holw. (LMC 158) ssp. glaucus T. fusca Ell. & WSP 71275 1971 Vulpia US L. Carris Everh. (LMC 141) microstachys (Washington) (Nutt.) Munro var. microstachys T. goloskokovii WSP 69687 1995 Apera interrupta US L. Carris Schwarzman (LMC 315) (L.) Beauv. (Washington)

T. goloskokovii WSP 71281 1993 Apera interrupta US L. Carris (LMC (Washington) 238-2) T. laevis Kuhn WSP 71278 1971 Triticum aestivum US L. Carris (LMC178) T. laevis TK108 1979 Triticum sp. Turkey T. laevis TK126-2 1979 Triticum durum Turkey Desf. T. laevis TK52-1 T. laevis WSP 71300 1988 Triticum aestivum Iran K. Vánky (Vanky 766) T. laguri G. M. HUV 16.352 1992 Lagurus ovatus L. Italy K. Vánky Zhang, G. X. Lin & J. R. Deng T. lolii WSP 71298 1990 Lolium rigidum Iran K. Vánky Auers.ex Rab. (Vanky 767) Gaudin T. lolioli WSP 71305 1990 Loliolum Iran K. Vánky Vanky, Carris, (Vanky 763) subulatum (Banks Castl. & H. & Sol.) Eig Scholz. T. puccinelliae WSP 71471 2004 Puccinellia US L. Carris Carris, G. (Holotype) distans (Jacq.) (Washington) Huang & Castl. Parl. T. puccinelliae WSP 71469 2004 Puccinellia US L. Carris distans (Washington) T. secalis WSP 71279 1993 Secale cereale US (Idaho) L. Carris (Cda.) Koern. (LMC 255) T. secalis G93 ? Secale cereale US (Idaho) L. Carris T. WSP 71314 2003 Agrostis US (Chinese G. Huang sphaerococca stolonifera L. Intercept) (Rab.) Fisch. v Waldh. T. togwateei WSP 71277 1992 Poa reflexa Vasey US (WY) L. Carris Guillem. (LMC 169) & Scribn. ex Vasey T. trabutii Jacz. WSP 71299 1990 Hordeum Iran K. Vánky (Vanky 764) murinum L. ssp. glaucum (Steud.) Tzvlev T. trabutii VPRI 32106 2005 Hordeum Australia I. Pascoe murinum L. ssp. leporinum (Link) Arcang.

T. vankyi WSP 71266 1997 Lolium perenne Australia L. Carris Carris & (V21-713) Castlebury T. vankyi WSP 71270 2005 Festuca rubra ssp. US (Oregon) S. Alderman (FF1-1) Fallax Isolates identified as T. brevifaciens.

The objective of this study was to test the hypothesis that T. contraversa is a broad host range

smut fungus using a phylogenetic approach based on sequence data from a collection of geographically and temporally diverse isolates of T. contraversa, including the type specimen of

T. brevifaciens, other historical specimens collected from the Pacific Northwestern U.S in the

1950s, and Eurasian collections from the early to late 1900s.

MATERIALS AND METHODS

Fungal isolates.—Specimens used in this study are listed in TABLE 4-1 and APPENDIX TABLE S-2.

Voucher specimens were deposited in herb. WSP (Washington State University, Pullman, WA) and herb. BPI (Beltsville, MD). Teliospores from bunted florets were soaked in sterile distilled water for 30 min at room temperature and mounted on a microscope slide in Shear’s mounting medium.

Spore morphology including teliospore diam with gelatinous sheath (Fischer 1952, Durán and

Fischer 1961), thickness of exospore, number of meshes (areolae) per spore diam, and sterile cell diam were recorded for 20 spores per sample using differential interference contrast microscopy at

×1000 (TABLE 4-2).

Nucleic acid extraction and PCR amplification.—DNA was extracted directly from teliospores by suspending 5 mg of teliospores in extraction buffer [50 mM Tris-HCl pH 8.0, 150 mM NaCl, 100

mM EDTA] (Shi et al 1996) in a 1.5 mL Eppendorf microcentrifuge tube with 200 μl Zirconia

beads. The spores were broken by shaking the tubes in a Mini-Beadbeater (BioSpec Inc.,

Bartlesville, Oklahoma) for 3 min, then 50 μl of 20% SDS (sodium dodecyl sulfate), 100 μl 5 M

NaCl, and 120 μl 10% PVP-40 (poly-vinyl pyrrolidone) were added followed by gentle shaking at

55 C overnight. Genomic DNA was purified by phenol/chloroform extraction and precipitated in

70% ethanol. DNA concentrations were determined using a NanoDrop ND1000 (Thermo

Scientific, Wilmington, Delaware) and adjusted to 10 ng/μL. Before PCR amplification, DNA

samples extracted from specimens more than 30 years old were further treated with GeneReleaser

(BioVentures, Inc. Murfreesboro, Tennessee) following manufacturer’s instructions. All PCR

amplifications were performed in 20 μL on a GeneAmp 9700 thermal cycler (Applied Biosystems,

Foster City, California). Primers for PCR amplification of three anonymous loci were derived from

a previous study (Bao et al, unpublished). Three nuclear loci, elongation factor 1 alpha (EF1A),

internal transcribed spacer 1 (ITS) and a portion of the second largest subunit of RNA polymerase

II (RPB2) were amplified as described in Carris et al (2007). All PCR primers used for this study

are listed in TABLE 4-3. PCR products were purified with ExoSAP-IT (USB, Cleveland, Ohio)

according to the manufacturer’s instructions and amplified with respective forward and reverse

PCR primers using the BigDye version 3.1 dye terminator kit (Applied Biosystems). Those products were purified again with Performa DTR gel filtration cartridges (Edge BioSystems,

Gaithersburg, Maryland) and sequenced on an ABI 3100 automated DNA sequencer.

Sequence analysis.—Sequence data were aligned and manually edited for each locus using

CLUSTALW as implemented in BIOEDIT v7.0.5.3 for Windows (Hall 1999). Concatenated

sequences were generated from three anonymous loci or from RPB2, ITS and EF1-alpha regions using MESQUITE 2.72 (Maddison and Maddison 2006). Gene incongruence tests were performed

in PAUP using partition homogeneity method (incongruence length difference test) (Swofford

2002) with 1000 replicates. Independently inferred topologies from each partition were examined for conflict (e.g., one monophyletic and the other paraphyletic). Reciprocal 70% ML bootstrap

(ML-BS) values or 95% Bayesian posterior probability (PP) were used as criteria for strongly supported clades. Phylogenetic trees were inferred independently from two combined alignments by maximum likelihood (ML) with PAUP 4.0b10 and Bayesian methods with MRBAYES v3.1.2 for

Windows (Ronquist and Huelsenbeck 2003).

The sequencing focused on the three anonymous loci from 49 common and dwarf bunt

isolates. A subset of isolates was sequenced for RPB2, ITS and EF1-alpha regions for comparison

with previous results (Bao et al 2010). Incongruence tests indicated that A13 and RPB2 were not

compatible with the other loci. To address the incongruence, a combined alignment from four

compatible loci, A16, P18, ITS and EF1-alpha, was generated for a subset of 22 isolates. Another

combined alignment was generated for the RPB2 amplified region, ITS and EF1-alpha, also for

comparison with previous results although results indicated that the RPB2 region was incongruent with the other loci. DNA substitution models for ML tree were determined by

MODELTEST 3.7 (Posada and Crandall 1998). The best fit models and parameters were chosen with the Akaike information criterion (AIC). Heuristic search was chosen with 100 starting trees

obtained via stepwise, random addition of sequences. The branch-swapping algorithm is tree

bisection-reconnection (TBR) on each tree. All aligned positions were included in the analysis. All

characters are unordered and have equal weight. Specific substitution models determined by

MODELTEST were used for each dataset as follows: A16, P18, ITS and EF1-alpha with TrN+I+G

model: LSET BASE=(0.2287 0.2602 0.2892) NST=6 RMAT=(1.0000 2.1005 1.0000 1.0000 4.6563)

RATES=EQUAL PINVAR=0.9392; RPB2, ITS and EF1-alpha with TrN+I+G model, LSET

BASE=(0.2166 0.2767 0.2610) NST=6 RMAT=(1.0000 2.8050 1.0000 1.0000 5.9225)

RATES=GAMMA SHAPE=0.7395 PINVAR=0.6895.

Monte Carlo Markov Chain (MCMC) sampling method was used for the Bayesian inference and limited to 10 000 000 generations. The combined partitions were unlinked in STATEFREQ,

REVMAT, SHAPE and PINVAR. In each analysis two independent runs were started simultaneously,

with four chains in each run (one heat chain and three cold chains). The analyses were repeated

three times. Trees were sampled every 1000 generations. The first 25% of sampled trees were

excluded from the final analysis.

With a large partition of missing data (27 out of 49 taxa missing sequence information

from RPB2, ITS and EF1-alpha regions), networks were reconstructed from concatenated

sequence of A13, A16, P18, RPB2, ITS and EF1-alpha regions with statistical parsimony method

(Templeton et al 1992) using TCS v. 1.21 (Clement et al 2000). The parsimony network method

provides estimation for the relationships among organisms at the population level when

phylogenetic analyses are not appropriate because of a limited number of variable characters and

possible existence of recombination (Templeton et al 1992). TCS was used to collapse

concatenated sequence from 4 loci into haplotypes and link into a network with 95% of

probability by the parsimony criterion. The script of network was manually edited and visualized

as graphics by using GHOSTSCRIPT v. 7.0.6 (http://www.cs.wisc.edu/~ghost/) implemented in

SNAP WORKBENCH (Price & Carbone 2005, Aylor 2006). The network was analyzed with nested

design by pooling haplotypes into 1-step to 3-step clades as described by Carbone and Kohn

(2004). A “gene jackknifing” method (Hackett et al 2008) was used to detect whether the

network structure was strongly impacted by individual loci, by excluding the loci from concatenated sequence one at a time and estimating NeighborNet networks (Bryant & Moulton,

2004). SPLITSTREE v. 4.11 (Huson & Bryant 2006) was used to estimate the NeighborNet

networks, with variance estimated from original least squares using maximum 4 dimensions as

filter, and bootstrap values calculated with 1000 replicates. The networks were estimated from

the concatenated sequence from all 6 loci or any possible 5 loci, or from the 4 congruent loci

(A16, P18, ITS and EF1-alpha). The networks consistently grouped the taxa into 2 similar clades

as summarized in Table 4-4.

TABLE 4- 2. Spore morphology of isolates used in phylogenetic analysis

Taxon Voucher Teliospore Avg. Exospore Sheath Number of numbers Diam Diam Depth Thickness Meshes per (μm) (μm) (μm) (μm) Spore Diam T. contraversa HUV 20.802 20-24 23.0 1-2.5 1-4 5-6 T. contraversa LMC514 21-25 22.8 1-3 2-3.5 4-6 T. contraversa WSP35764 21-23 21.2 0.5-2.5 1.5-3.5 4-6 T. contraversa WSP35784 21-23 21.5 0.5-1.5 1.5-3 4-6 T. contraversa WSP35785 19-23 21.0 0.5-2 1-2.5 4-6 T. contraversa WSP35786 20-22 21.3 0.5-2 1-2.5 5-6 T. contraversa WSP35787 21-24 22.9 0.5-2.5 2-4 4-6 T. contraversa WSP40312 20-23 21.0 1-2 1-3 4-7 T. contraversa WSP40406 T. contraversa WSP40667 19-23 22.5 0.5-2 1-3 4-6 T. contraversa WSP40685 T. contraversa WSP41411 20-22 20.9 0.5-2 1-2.5 5-6 T. contraversa WSP63652 22-25 23.9 1-2 2-3.5 4-6 T. contraversa WSP63656 20-22 21.0 1-2 1.5-3 5-6 T. contraversa WSP63659 21-25 23.5 1-2.5 1.5-3 4-6 T. contraversa WSP63665 21-24 22.3 0.5-1.5 1-3 5-7 T. contraversa WSP63667 22-24 23.1 1-2 1-3 T. contraversa WSP63680 22-23 22.5 0.5-1.5 1-2.5 4-7 T. contraversa WSP63683 21-22 21.5 1-2 2-3 4-6 T. contraversa WSP63685 20-21.5 20.8 1-2 1-2.5 4-6 T. contraversa WSP63687 18-23 20.4 0.5-2 1-2 5-7 T. contraversa DB36 22-24.5 23.4 2.5-4 2.5-4.5 4-5 T. contraversa DB9030 23-25.5 24.6 3-4.5 3-6 4-7 T. contraversa J1 21-25 22.0 1-3 1-3 4-6 T. contraversa LMC282 22-28 25.1 2-4 3-5 5-6 T. contraversa S9009 21-25 23.4 2-3.5 2-3.5 5-7 T. contraversa TK69B 19-21 20.5 0.5-1 0.5-1.5 5-7 T. contraversa WSP49569 20-22 21.5 0.5-2 1-2.5 4-6 T. contraversa WSP59699 T. contraversa WSP63766 23-26 24.8 1-3 2-5 4-6 T. contraversa WSP63862 19-22 20.5 1-2.5 1-3 3-5 T. contraversa WSP63890 T. contraversa WSP69062 23-28 25.1 3-5 3-5 5-6 T. contraversa WSP70123 19-21 20.5 0.5-2 0.5-3 4-7 T. contraversa DB460 18-25 22.5 1-2 2-5 4-5 T. contraversa TK52-2 19-21 19.8 1-2 1-3 4-6 T. contraversa TK83-2 20-25 22.0 1-2 1-3.5 5-6 T. contraversa WSP68945 21-24 22.0 1-2 1.5-3.5 4-6

T. caries J10 16-21 17.3 0.5-1 5-7 T. caries J11 20-22 20.5 0.5-2 1-2 4-7 T. caries J17 17-20 17.7 0.5-2 0.5-2 5-7 T. caries LMC516 18-21 18.9 1-2 5-6 T. caries WSP11821 17-21 18.5 0.5-1 5-7 T. caries WSP41412 19-20 19.7 0.5-1.5 0.5-1.5 5-7 T. caries WSP47237 17-20 18.2 0.5-1 5-6 T. caries WSP47286 18-20 19.5 0.5-2 4-6 T. caries WSP49568 20-23 21.8 0.5-2 1-3 4-7 T. caries WSP63716 T. laevis LMC178 20-22.5 21.2 T. laevis TK108 15-20 16.5 T. laevis TK126-2 18-21 19.0 T. laevis WSP 71300 32-37 19.5

Approximately Unbiased (AU), Shimodaira-Hasegawa (SH) and Kishino-Hasegawa (KH)

tests were used to test the hypotheses of morphological species concepts with ML trees with

CONSEL 0.1j (Kishino and Hasegawa 1989, Shimodaira and Hasegawa 1999, 2001, Shimodaira

2002).

RESULTS

The spore morphology recorded for the isolates used in this study are summarized in TABLE 4-2.

Only those isolates for which sequence data could be obtained were analyzed. The results of the morphological analysis generally agree with the collectors’ identification of isolates, with the exception of WSP 49568, identified by the collector as T. caries, but determined to have spore morphology characteristic of T. contraversa.

A total of 100 isolates of Tilletia species were initially included in this study. Sequence data for 20 isolates were from previous studies (Carris et al 2007, Bao et al 2010) (TABLE 4-1, 4-2,

APPENDIX TABLE S-2). Most of the isolates used in this study were over 30 years old, and good

quality sequence data were obtained from specimens as old as 1945 (TABLE 4-1). However, poor

quality sequence data or poor PCR amplification occurred with many of the older specimens

(APPENDIX TABLE S-2). All but one of the isolates collected after 1960 yielded good quality

sequence data, 24 of 35 isolates (69%) from the 1950s were successful sequenced, and 3 of 20

isolates (15%) collected before 1950 were successfully sequenced. None of the isolates collected

before 1921 could be amplified (APPENDIX TABLE S-2). A treatment with GeneReleaser was used

for amplification of DNA from isolates collected before 1980 to generate stronger amplicons for

sequencing; without this treatment, only 20-30% of the DNA from isolates from the 1950s could

be sequenced. Forward and reverse sequence data could not be obtained for all 6 loci for 14

isolates, even after repeated attempts. Only those isolates with high quality sequence data for at

least two anonymous loci or two of RPB2, ITS or EF1-alpha loci were included into the analyses.

In the sequence data matrix used for phylogenetic analyses, about 3% of nucleotides were missing among 3 anonymous loci, and 6% of nucleotide data were missing among RPB2, ITS and

EF1-alpha loci.

100

TABLE 4- 3. Primers, PCR annealing temperature, amplified length and number of polymorphic sites for loci used in study

Loci Primers used Primer sequences PCR Amplified Number Primer annealing length of poly- Source temp. (C) morphic sites a A13-1AR874 5′-GCAAGCGTGGGCCTCARTA-3′ A13 64 705 19 Chapter 3 A13-1AR1439 5′-TCAAGTAAGTACGGTAGAAGGGTGC-3′ A16-2Fb 5′-AGGAGGGTTGTAGCAGCGAGCGA-3′ A16 64 414 6 Chapter 3 A16-2R677 5′-CCCCATTATTATTGCTGCTGACTG-3′ P18-1F1 5′-CGCCCTTTTCCTGGTAGTTCA-3′ P18 64 403 12 Chapter 3 P18-1R535 5′-GCGGAGCATCCCACCAACT-3′ RPB2-740F 5′- GATGGACGCGGTTTGTAATG-3′ Carris et al RPB2 58 470-514 19 RPB2-1365R 5′- TCGAAGAGCYAACACTGAGACG-3′ 2007 ITS1-F 5′-CTTGGTCATTTAGAGGAAGTAA-3′ Gardes and 101 Bruns 1993 ITS1 59 564-637 16 ITS4 5′- TCCTCCGCTTATTGATATGC-3′ White et al 1990 EF1-636F 5′- TCAACGTCGTYGTYATCGG-3′ Carris et al 2007 EF1α 55 713-719 17 EF1-1567R 5′- CHGTRCCRAT ACCACCRATCTT-3′ Rehner, S. 2001b Includes only isolates of T. contraversa (T. brevifaciens) and closed related species – T. caries, T. laevis, T. trabutii and T. secalis (48 isolates for the SCAR phylogeny, and 29 isolates for the 3 gene phylogeny) Primer sequence for elongation factor 1-a (EF1-a) available from http://ocid.nacse.org/research/deephyphae/EF1primer.pdf

FIG. 4-1. Maximum likelihood tree of T. contraversa isolates and related species with sequences

from A16, P18, ITS and EF1-alpha loci. Tree is mid-point rooted, with Maximum Likelihood

bootstrap (ML-BS) values and Bayesian Posterior Probabilities (MB-PP) values as support.

102

Branch length indicates the substitution rate. Species determination, accession number, and host species indicated.

103

FIG. 4-2. NeighborNet networks constructed from different datasets from 49 common and dwarf bunt isolates, with bootstrap values on the branches connecting two major groups. a. using A16,

104

P18, ITS and EF1-alpha loci, b. using A13, A16, P18, RPB2, ITS and EF1-alpha loci, c. using

A13, A16, P18, RPB2 and ITS loci, d. using A16, P18, RPB2, ITS and EF1-alpha loci, e. using

A13, A16, P18, ITS and EF1-alpha loci.

105

TABLE 4- 4. The compatibility in placement of each taxon in NeighborNet tree when including different loci in analysis.

Voucher Taxon HOST ALLa NO NO NO NO NO NO NO A13,

numbers A13 A16 P18 RPB2 ITS EF RPB2

WSP63862 T. contraversa Triticum aestivum A A A A A A A A

WSP70123 T. contraversa Elymus repens A A A A A A A A

WSP68945 T. contraversa Thinopyrum intermedium A A A A A A A A

HUV20.802 T. contraversa Thinopyrum intermedium A A A A A A A A

Ⅹ Elymus repens

J1 T. contraversa Elymus repens A A A A A A A A

106 J8 T. contraversa Elymus repens A A A A A A A A

TK52-2 T. contraversa Thinopyrum sp. A A A A A A A A

WSP63659 T. contraversa Triticum aestivum A A A A A A A A

WSP63766 T. contraversa Elymus repens A NA A A A A A NA

TK52-1 T. laevis Triticum aestivum A NA A A A A A NA

WSP63652 T. contraversa Hordeum bulbosum A B A A A A A B

WSP71278 T. laevis Triticum aestivum A B A A A A A B

TK108 T. laevis Triticum sp. A B A A A A A B

WSP63683 T. contraversa Hordeum sp. A B A A A A A B

TK83-2 T. contraversa Elymus repens A B A A A A A B

WSP35785 T. contraversa Triticum aestivum B B B B B B B B

WSP35784 T. contraversa Thinopyrum intermedium B B B B B B B B

WSP40312 T. contraversa Elymus trachycaulus B B B B B B B B

WSP63665 T. contraversa Agropyron cristatum B B B B B B B B

WSP63687 T. contraversa Secale cereale B B B B B B B B

WSP71280 T. contraversa Triticum aestivum B B B B B B B B

WSP71300 T. laevis Triticum aestivum B B B B B B B B

J10 T. caries Triticum aestivum B B B B B B B B

107 J11 T. caries Triticum aestivum B B B B B B B B

J17 T. caries Triticum aestivum B B B B B B B B

TK126-2 T. laevis Triticum durum B B B B B B B B

WSP35786 T. contraversa Arrhenatherum elatius B B B B B B B B

WSP35787 T. contraversa Thinopyrum intermedium B B B B B B B B

WSP40667 T. contraversa Arrhenatherum elatius B B B B B B B B

WSP40685 T. contraversa Thinopyrum intermedium B B B B B B B B

WSP63656 T. contraversa Triticum aestivum B B B B B B B B

WSP35764 T. contraversa Triticum aestivum B B B B B B B B

WSP40406 T. contraversa Thinopyrum intermedium B B B B B B B B

WSP41411 T. contraversa Triticum aestivum B B B B B B B B

WSP63667 T. contraversa Agropyron desertorum B B B B B B B B

WSP63680 T. contraversa Bromus marginatus B B B B B B B B

WSP49569 T. contraversa Triticum aestivum B B B B B B B B

WSP59699 T. contraversa Triticum aestivum B B B B B B B B

WSP41412 T. caries Triticum aestivum B B B B B B B B

WSP49568 T. caries Triticum aestivum B B B B B B B B

WSP44943 T. contraversa Festuca idahoensis B B B B B B B B

108 WSP47237 T. caries Triticum aestivum B B B B B B B B

WSP69062 T. contraversa Triticum aestivum B B B B B B B B

DB36 T. contraversa Triticum aestivum B B B B B B B B

DB460 T. contraversa Thinopyrum intermedium B B B B B B B NA

DB9030 T. contraversa Triticum aestivum B B B B B B B B

LMC516 T. caries Triticum aestivum B B B B B B B B

S9009 T. contraversa Triticum aestivum B B B B B B B B

LMC514 T. contraversa Thinopyrum ponticum B NA NA NA B B B NA

a. Capital letters “A” and “B” used to designate the placement of isolates in two major clades in networks with “gene jackknifing”

FIG. 4-3. Maximum likelihood tree of T. contraversa isolates and related species with RPB2,

109

EF1-alpha and ITS loci. Mid-point rooted, with maximum likelihood bootstrap (ML-BS) values and

Bayesian posterior probabilities (MB-PP) values as support. Branch length indicates the substitution

rate. Species determination, accession number, and host species indicated.

110

Forty-nine common and dwarf bunt isolates were used for PCR-based sequencing with the 3

anonymous loci for phylogenetic analysis. Twenty-two isolates representing different haplotypes

in anonymous loci were sequenced with RPB2, ITS and EF1-alpha loci and combined with the

sequences from 21 isolates of wheat bunt and other Tilletia spp. used in Chapter 2 (TABLE 4-3).

The sequence length and number of variable sites for each locus are listed in Table 4-3. Strong incongruence (P < 0.01) was found between the concatenated dataset of the three anonymous loci and each individual anonymous locus, and in the combined data from ITS, RPB2 and EF1-alpha loci and each of three loci as partitions. Incongruence length difference (ILD) tests were performed on all possible pairs of 6 partitions of anonymous loci and ITS, RPB2 and EF1-alpha loci with available taxa. Partitions A13 and RPB2 were identified as the major source of incongruence.

Maximum likelihood trees were reconstructed from 22 taxa which sequence information was available from at least three out of the four partitions A16, P18, ITS and EF1-alpha (ILD test: p =

0.73) (FIG. 4-1).

Sequence information from A16 and P18 loci of 49 taxa, together with ITS and EF1-alpha

regions of 22 taxa were combined and estimated with statistical parsimony network, resulting in

a highly reticulated network (data not shown). Better resolved networks with 2 clades were

constructed with SPLITSTREE with NeighborNet method (FIG. 4-2 a.). Although these 2 clades

were not strongly supported by bootstrap value (58.8%), a similar 2-clade structure was

consistently revealed in other NeighborNet networks constructed from 6 loci or different

combinations of 5 loci with the “gene jackknifing” method. Each isolate was coded as “A” or

“B” to indicate which clade it was placed in the network analyses based on different datasets

(TABLE 4-4). For the 49 isolates used for network construction, conflict in placement within

clades among datasets was found for only 5 isolates (TABLE 4-4, isolates indicated by shading).

111

Incongruence in the concatenated data of ITS, RPB2 and EF1-alpha loci may be caused by

the limited number of variable sites among the closely related isolates used in this analysis. The

concatenated data were used in previous analyses and resolved host-specific species of Tilletia

into well-supported clades (e.g., Bao et al 2010). Network analyses in the present study revealed

reticulated structures indicating possible recombination among common and dwarf bunt isolates,

but not among other Tilletia spp (data not shown). For comparison with previous studies,

maximum likelihood trees were reconstructed from combined data (FIG. 4-3).

In the A16, P18, ITS and EF1-alpha tree (FIG. 4-1), seven isolates of T. contraversa on E.

repens, Th. intermedium, Triticum aestivum and Thinopyrum sp. were found in a well-supported clade (99% ML-BS, 100% MB-PP). All seven isolates were Eurasian in origin, six from Europe and one from Turkey. Fourteen other isolates of T. contraversa, T. caries and T. laevis collected on

Agropyron, Bromus, Elymus, Secale, Thinopyrum and Triticum from Eurasia and North America were unresolved.

In networks estimated with different datasets, no association between clade and host was detected (TABLE 4-4). The 7 isolates in the well-supported clade in the A16, P18, ITS and

EF1-alpha tree (FIG. 4-1) were consistently placed in the same clade among all the networks

constructed from different subsets of sequence data (TABLE 4-4, A clades) as was WSP63659

from wheat, which was not included in the phylogenetic analyses. Isolates from Elymus,

Thinopyrum and Triticum were present in both clades, but isolates from Agropyron,

Arrhenatherum, Bromus, Festuca, and Secale were found only in the B clades. The two isolates

from Hordeum, WSP68683 and WSP63652, were among the set of isolates that were

inconsistent in their placement into clades in the network analyses based on different

combinations of loci (TABLE 4-4).

112

Twenty-six common and dwarf bunt isolates representing different genotypes either in the

anonymous loci phylogenetic analysis or in previous studies (Carris et al 2007, Bao et al 2010)

were used in a multilocus phylogenetic analysis based on partial RPB2, ITS and EF1-alpha loci

(FIG. 4-3). Four isolates (HUV20.802, WSP70123, WSP 68945 and TK52-2) on Elymus and

Thinopyrum were in a well-supported clade (72% ML-BS, 100% MB-PP) (FIG. 4-3). Two Swedish dwarf bunt isolates, J1 and J8, were grouped (74% ML-BS, 99% MB-PP) as a sibling clade (FIG.

4-3). The average spore size of seven isolates (HUV20.802, J1, J8, TK52-2, TK83, WSP 68945

and WSP70123) was 21.55±1.77 μm, significantly larger than T. caries isolates from wheat

(19.12±1.42 μm, p = 0.002), and smaller but not significantly different than other T. contraversa

isolates (22.33±1.50 μm, p = 0.156) used in this study. Eighteen other isolates of T. contraversa, T.

caries and T. laevis from Agropyron, Bromus, Elymus, Secale, Thinopyrum and Triticum were

unresolved (FIG. 4-3).

Eighteen T. contraversa isolates collected on wheat and grasses hosts in PNW regions in

1946-1959 were used in the analyses to represent putative genotypes involved in the 1950s

outbreak of dwarf bunt. All the isolates are closely related based on the sequenced loci, and none

of those isolates was placed in a clade distinct from the others in the analyses presented (FIGS. 4-1,

4-2, 4-3, TABLE 4-4).

Tilletia secalis, T. trabutii, T. goloskokovii, T. lolioli and T. sphaerococca are in an unresolved

clade, with internal support for T. goloskokovii and T. trabutii lineages (FIG. 4-3). Other Tilletia

species included in the analyses formed well-supported clades as in previous studies (Carris et al

2007, Bao et al 2010). The alternate hypotheses, three monophyletic, distinct species of wheat bunts (T. caries, T. contraversa, T. laevis), versus paraphyletic grouping in ML tree were tested with parametric bootstrapping method and non-parametric tests. The SOWH test (parametric

113

bootstrapping) rejected alternative assumption with p = 0.005 (APPENDIX FIG. S-3). The non-parametric test did not reject (KH = 0.226, SH = 0.226, BP = 0.241) the hypothesis.

DISCUSSION

In this study, phylogenetic analyses and network estimations based on six genetic loci provide evidence that bunts infecting species of Agropyron, Bromus, Elymus, Secale, Thinopyrum and

Triticum corresponding to the morphological species concept of T. contraversa do not form host-specific clades, and do not form a lineage distinct from the common bunt pathogens T. caries and T. laevis. However, spore morphology of these isolates was in general agreement with the morphological species concept of T. contraversa.

Five Eurasian isolates of T. contraversa on Elymus repens and Thinopyrum intermedium were consistently placed in a clade distinct from the majority of T. contraversa isolates in the phylogenetic analyses presented, indicating that substructure exists within dwarf +common bunt assemblage and leaving open the possibility that T. contraversa represents a distinct genotype, but one that does not correspond to the accepted concept applied to the dwarf bunt pathogen of wheat. Thinopyrum intermedium and Elymus repens are genetically related, allopolyploid wheatgrasses (Mahelka et al 2007). The average spore size of seven T. contraversa isolates (HUV20.802, J1, J8, TK52-2, TK83, WSP 68945 and WSP70123) on Th. intermedium and E. repens was smaller but not significantly different than other T. contraversa isolates used in this study (p = 0.156). A larger number of samples from these hosts are needed to determine if there are consistent differences in spore morphology, spore germination or germination requirements. Morphological and phylogenetic evidence presented support that a number of closely related genotypes were involved in historical outbreaks of dwarf bunt on

114

wheat and other pooid grasses in the PNW. Fischer and Shaw (1953) proposed a species concept

for smut fungi that would recognize collections with similar morphological and physiological

characters infecting hosts in the same family as a single species, and their concept is supported

in part by the results of this study, at least in regards to the dwarf bunt pathogen.

A number of the herbarium specimens included in this study were important historical

collections that provided the basis for expanding the species concept and host range of T.

contraversa (TABLE 4-1, APPENDIX TABLE S-1). For example, WSP35784 and WSP35785 were

the type and paratype specimens, respectively, for T. brevifaciens (Fischer 1952). Although

Fischer’s decision to designate the holotype for T. brevifaciens, which he clearly considered to

be the wheat dwarf bunt pathogen, on Th. intermedium rather than on wheat is puzzling, the

present study demonstrates that the Th. intermedium and wheat bunts are genetically similar and

belong to the predominant genotype as other PNW collections in the anonymous loci analysis.

Previous studies by Carris et al (2007) and Carris and Castlebury (2008) provided evidence for

recognition of a genetically distinct bunt on Thinopyrum intermedium, and the authors referred

to the lineage as T. brevifaciens, but those analyses were based solely on European collections, and did not include the type specimens of T. brevifaciens. The present study demonstrates that T. brevifaciens is conspecific with the wheat bunt pathogens, and the European bunt infecting

Thinopyrum intermedium is genetically distinct from the bunt infecting this host in North

America. Forty new host records were reported for the dwarf bunt pathogen during the PNW dwarf bunt outbreaks in the 1950s (APPENDIX TABLE S-1). Collections WSP40406 and

WSP40685 on Arrhenatherum elatius (tribe Aveneae), WSP63680 on Bromus (tribe Bromeae),

and WSP44943 on Festuca (tribe Poeae), were part of these outbreaks (Hardison and Corden

1952, Meiners and Hardison 1957). In this study, 33 dwarf bunt isolates on 11 host species

115

representing 8 host genera, and all dwarf bunt isolates of wheat collected in the PNW during the

1950s were in the same clade in different network analyses (TABLE 4-4). These results support

that the dwarf bunt pathogen of wheat was responsible for outbreaks on other grasses during this

time.

Previous studies that included a limited number of wheat bunt isolates in multilocus analyses

showed that the wheat bunts formed a single, well-supported lineage (Carris et al 2007, Carris and

Castlebury 2008, Bao et al 2010). The present study is the first to include a large number of common and dwarf isolates in gene genealogical analyses. The results do not provide support for the wheat bunt pathogens as comprising a single lineage. A confounding factor in this study was the strong incongruence detected among data sets. The major source of incongruence was detected between A13 and RPB2 (APPENDIX FIG. S-2). The incongruence can be the result of

recombination among loci or incomplete lineage sorting, or random errors produced by a low

level of variability in the loci (Planet 2006). Network methods appear to be more appropriate for

these data than bifurcated trees estimated by phylogenetic analysis; networks resolve questions at

the population genetics level, assume recombination can occur (Templeton et al 1992), and allow

reticulations and multifurcations in structure for closely related isolates in the analyses (Cassens

et al 2005).

The type specimen of T. contraversa on Elymus repens was described as having spores that

were on average 1 µm larger than the spores of T. caries (Kühn 1874). The average spore size of

seven Eurasian T. contraversa isolates on Th. intermedium and E. repens (HUV20.802, J1, J8,

TK52-2, TK83, WSP 68945 and WSP70123) forming a well-supported clade (FIG. 4-1) was

significantly larger than the T. caries isolates from wheat included in this study. In contrast, PNW

isolates on Thinopyrum and Elymus (WSP35784, WSP40312) were not clustered with the

116

Eurasian Elymus and Thinopyrum isolates (FIGS. 4-1, 4-2, 4-3), suggesting that there are at least

two genetically distinct bunt fungi that infect Elymus and Thinopyrum. Therefore, it is possible that T. contraversa on Elymus repens in Europe is a distinct species from what has been recognized as the dwarf bunt pathogen. Analysis of additional T. contraversa collections on E. repens, Th. intermedium and Triticum from Eurasia is needed to resolve this question.

This study and a companion study (Chapter 3) failed to provide support for three morphological species, T. caries, T. contraversa and T. laevis as recognized for the common and dwarf bunt pathogens. Isolates such asWSP49568 and WSP49569, collected at the same time from one site, have morphologically distinct teliospores but are of the same or similar haplotype. The dwarf and common bunt pathogens are widely considered to have different requirements for teliospore germination and infection on wheat, but these requirements have not been assessed across a large number of isolates infecting wheat and other grass hosts. The dwarf bunt pathogen is restricted to winter-sown wheat in areas with persistent snow cover, with the highest infection rates on wheat planted in late September to early October (Hoffmann and Purdy 1967). Some species included in the host range of the dwarf bunt pathogen, including E. repens, Th. intermedium and

Arrhenatherum elatius, are perennial grasses (Weintraub 1953, Palmer and Sagar 1963, Barkworth and Dewey 1985), and the timing of infection of bunt fungi on perennial hosts has not been studied.

In the dwarf bunt outbreak in eastern Oregon in 1952, most of the grasses were planted during the spring (Hardison 1954). Dwarf bunt isolates infecting wheat and other grasses have been shown to differ in their germination requirements. For example, isolates from grasses and rye (Secale cereale) germinated more rapidly than those from wheat (Meiners and Waldher 1959). Similarly, two isolates used in the present study, J1 and J8, collected on Elymus repens in Sweden, were morphologically consistent with T. contraversa, but germinated at 15C in the absence of light (L.

117

Carris, unpublished).

This study provides the first genetic evidence in support of a broad host range for the dwarf

bunt pathogen. This is in contrast to the strong association between non-cultivated grass hosts and

phylogenetic lineage for closely related Tilletia species shown in previous studies (Carris et al

2007, Bao et al 2010). Savile (1979) similarly showed that rust pathogens infecting cultivated

hosts may have unusually broad host ranges, although related species of rusts on wild hosts typically exhibit narrow host ranges. The study of additional smut species reported to have broad host ranges, including those such as Ustilago bullata that infect non-cultivated hosts, may provide insight into the potential role agronomic practices may have played in the evolution of smut fungi.

LITERATURE CITED

Aylor DL, Price EW, Carbone I. 2006. SNAP: Combine and Map modules for multilocus

population genetic analysis. Bioinformatics 22:1399-1401.

Bao X, Carris LM, Huang G, Luo J, Liu Y, Castlebury LA. 2010. Tilletia puccinelliae, a new

species of reticulate-spored bunt fungus infecting Puccinellia distans. Mycologia

102:613-623.

Barker FK, Lutzoni FM. 2002. The utility of the incongruence length difference test. Syst Biol

51:625-637.

Barkworth ME, Dewey DR. 1985. Genomically based genera in the perennial Triticeae of North

America: Identification and membership. Am J Bot 72:767-776.

Begerow D, Göker M, Lutz M, Stoll M. 2004. On the evolution of smut fungi and their hosts. in:

Frontiers in Basidiomycote Mycology. pp. 81-98 p. Agerer R, Piepenbring M, Blanz P, eds.

IHW‐Verlag & Verlagsbuchhandlung, Germany.

118

Bryant D, Moulton V. 2004. Neighbor-net: an agglomerative method for the construction of

phylogenetic networks. Mol Biol Evol 21:255-265

———, Lutz M, Oberwinkler F. 2002. Implications of molecular characters for the phylogeny of

the genus Entyloma. Mycol Res 106:1392-1399.

Carris LM, Castlebury LA. 2008. Is rye bunt, Tilletia secalis, present in North America? North

American Fungi 3:147-159.

———, ———, Huang G, Alderman SC, Luo J, Bao X. 2007. Tilletia vankyi, a new species of

reticulate-spored bunt fungus with non-conjugating basidiospores infecting species of

Festuca and Lolium. Mycol Res 111:1386-1398.

Cassens I, Mardulyn P, Milinkovitch MC. 2005. Evaluating intraspecific “network” construction

methods using simulated sequence data: Do existing algorithms outperform the global

Maximum Parsimony approach? Syts Biol 54:363-372.

Clement M, Posada D, Crandall KA. 2000. TCS: A computer program to estimate gene

genealogies. Mol Ecol 9:1657-1659.

Conners IL. 1954. The organism causing dwarf bunt of wheat. Canadian Journal of Botany

32:426-431.

Darlu P, Lecointre G. 2002. When does the incongruence length difference test fail? Mol Biol Evol

19:432-437.

Dewey WG, Tingey DC. 1958. A new grass host for dwarf bunt. Plant Disease Reporter 42:18-19,

28.

Durán R, Fischer GW. 1956. Further studies on the synonymy and host range of the dwarf bunt

fungus, Tilletia contraversa. Res Stud State Coll Wash 24:259-266.

———, ———. 1961. The genus Tilletia. Washington State University.

119

Felsenstein J. 2004. Inferring phylogenies. Sinauer Associates, Sunderland, Massachusetts.

Fischer GW. 1952. Tilletia brevifaciens sp. nov., causing dwarf bunt of wheat and certain grasses.

Res Stud State Coll Wash 20:11-14.

———, Durán R. 1956. Some new distribution and early incidence records of dwarf bunt of wheat

and rye. Res Stud State Coll Wash 24:267-274.

———, Shaw CG. 1953. A proposed species concept in the smut fungi, with application to North

American species. Phytopathology 43:181-188.

———, Tyler LJ. 1952. ls there an earlier record of dwarf bunt than 1931. Plant Disease Reporter

36:445-447.

Goates BJ. 1996. Common bunt and dwarf bunt. in: Bunt and Smut Diseases of Wheat: Concepts

and Methods of Disease Management. pp. 12-25 p. Wilcoxson RD, Saari EE, eds. CIMMYT,

Mexico D.F.

Hall T. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program

for Windows 95/98/NT. Nucleic Acids Symposium Series 41:95-98.

Hackett SJ, Kimball RT, Reddy S, Bowie RCK, Braun EL, MJ Braun MJ, Chojnowski JL, Cox

WA, Han K-L,Harshman J, Huddleston CJ, Marks BD, Miglia KJ, Moore WS, Sheldon

FH, Steadman DW, Witt CC, Yuri T. 2008. A Phylogenomic Study of Birds Reveals Their

Evolutionary History. Science 320:1763-1768.

Hardison JR. 1954. New grass host records and life history observations of dwarf bunt in eastern

Oregon. Plant Disease Reporter 38:345-347.

———, Corden ME. 1952. A serious outbreak of dwarf bunt in tall oat grass in Oregon. Plant

Disease Reporter 36:343-344.

———, Fenwick HS, Meiners JP. 1955. Additional grass hosts and a revised host list for dwarf

120

bunt. Plant Disease Reporter 39:685-687.

———, Meiners JP. 1956. New host records for dwarf bunt in the Pacific Northwest. Plant Disease

Reporter 40:1058-1060.

———, ———, Hoffmann JA, Waldher JT. 1959. Susceptibility of Gramineae to Tilletia

contraversa. Mycologia 51:656-664.

Hoffmann JA. 1982. Bunt of wheat. Plant Dis 66:979-986.

———, Purdy LH. 1967. Effect of stage of development of winter wheat on infection by Tilletia

controversa. Phytopathology 57:410-413.

Holton CS. 1941. Preliminary investigations on dwarf bunt of wheat. Phytopathology 31:74-82.

———, Heald FD. 1936. Studies on the control and other aspects of bunt of wheat. Wash Agr Exp

Sta Bul 339:1-35.

Huson DH, Bryant D. 2006. Application of phylogenetic networks in evolutionary studies. Mol

Biol Evol. 23:254-267.

Kühn J. 1874. Tilletia contraversa. Hedwigia 13:188-189. Rabenhorst, L. Fungi europaei exs.

1801-1900. Dresden.

Kishino H, Hasegawa M. 1989. Evaluation of the maximum likelihood estimate of the

evolutionary tree topologies from DNA sequence data, and the branching order in

Hominoidea. J Mol Evol 29:170-179.

Maddison WP, Maddison DR. 2006. Mesquite: a modular system for evolutionary analysis.

Version 2.72.

Mahelka V, Fehrer J, Krahulec F, Jarolímová V. 2007. Recent natural hybridization between two

allopolyploid wheatgrasses (Elytrigia, Poaceae): Ecological and evolutionary implications.

Ann Bot 100:249-260.

121

Meiners JP. 1955. Preliminary report on cross-inoculations with dwarf bunt on wheat and grasses.

Plant Disease Reporter 39:161-162.

———, Hardison JR. 1957. New host records for dwarf bunt in the Pacific Northwest. II. Plant

Disease Reporter 41:983-985.

———, Waldher JT. 1959. Factors affecting spore germination of twelve species of Tilletia from

cereals and grasses. Phytopathology 49:724-728.

———, ———, Hardison JR, Fenwick HS. 1956. Additions to the host range of dwarf bunt. Plant

Disease Reporter 40:26-27.

Mendgen K, Hahn M. 2002. Plant infection and the establishment of fungal biotrophy. Trends

Plant Sci 7:352-356.

Palmer JH, Sagar GR. 1963. Agropyron Repens (L.) Beauv.(Triticum repens L.; Elytrigia repens

(L.) Nevski). The Journal of Ecology 51:783-794.

Planet PJ. 2006. Tree disagreement: Measuring and testing incongruence in phylogeies. J Bioed

Inf 39:86-102.

Posada D, Crandall KA. 1998. MODELTEST: Testing the model of DNA substitution.

Bioinformatics (Oxford) 14:817-818.

Poulter RTM, Goodwin TJD, Butler MI. 2007. The nuclear encoded inteins of fungi, Fungal Genet.

Biol. 44: 153–179.

Price EW, Carbone I. 2005. SNAP: workbench management tool for evolutionary population

genetic analysis. Bioinformatics 21:402-404.

Purdy LH, Kendrick EL, Hoffmann JA, Holton CS. 1963. Dwarf bunt of wheat. Annual Reviews

in Microbiology 17:199-222.

Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed

122

models. Bioinformatics 19:1572-1574.

Savile DBO. 1979. Fungi as AIDS in higher plant classification. Bot Rev 45:377-503.

Shi YL, Loomis P, Christian D, Carris LM, Leung H. 1996. Analysis of the genetic relationships

among the wheat bunt fungi using RAPD and ribosomal DNA markers. Phytopathology

86:311-318.

Shimodaira H. 2002. An approximately unbiased test of phylogenetic tree selection. Syst Biol

51:492-503.

———, Hasegawa M. 1999. Multiple comparisons of log-likelihoods with applications to

phylogenetic inference. Mol Biol Evol 16:1114-1116.

———, ———. 2001. CONSEL: for assessing the confidence of phylogenetic tree selection.

Bioinformatics 17:1246.

Stamatakis A, Auch AF, Meier-Kolthoff J, Göker M. 2007. AxPcoords & parallel AxParafit:

statistical co-phylogenetic analyses on thousands of taxa. BMC Bioinformatics 8:405.

Stoll M, Piepenbring M, Begerow D, Oberwinkler F. 2003. Molecular phylogeny of Ustilago and

Sporisorium species (Basidiomycota, Ustilaginales) based on internal transcribed spacer

(ITS) sequences. Can J Bot 81:976-984.

Swofford DL. 2002. PAUP*: phylogenetic analysis using parsimony (* and other methods),

version 4.0 b10. Sinauer Associates, Sunderland, MA.

Tajima F. 1983. Evolutionary relationship of DNA sequences in finite populations. Genetics

105:437-460.

Templeton AR, Crandall KA, Sing CF. 1992. A cladistic analysis of phenotypic associations with

haplotypes inferred from restriction rndonuclease mapping and DNA sequence data. III.

Cladogram estimation. Genetics 132: 619-633

123

Trione EJ. 1982. Dwarf bunt of wheat and its importance in international wheat trade. Plant Dis

66:1083-1089.

Vánky K. 1994. European smut fungi. Gustav Fischer Verlag, Stuttgart, New York.

Warmbrunn K. 1952. Research on dwarf bunt. Phytopathology 19:441-482.

Weintraub FC. 1953. Grasses introduced into the United States. US Dept. of Agriculture, Forest

Service.

Woodward RW, Holton CS, Vogel OA. 1952. Dwarf bunt of rye. Plant Disease Reporter 36:434.

Young PA. 1935. A new variety of Tilletia tritici in Montana. Phytopathology 25:40.

124

ATTRIBUTION

Chapter 2. This chapter was published in Mycologia (2010) 102:613-623. Xiaodong Bao

contributed to the study design, experiment execution, and manuscript writing. Dr. Lori M.

Carris assisted in study design and writing. Dr. Lisa A. Castlebury contributed to the sequencing and phylogenetic analyses. Mr. Guoming Huang, Ms. Jiafeng Luo, and Ms. Yueting Liu, at the

Tianjin Quarantine Bureau, were responsible for first identifying the Puccinellia bunt in seed samples that provided the basis for the study.

Chapter 3 & Chapter 4. These chapters were written in manuscript format and will be submitted to Mycologia or similar journal. The co-authors on both manuscripts will be Drs. Lori

M. Carris and Lisa A. Castlebury. Xiaodong Bao was responsible for the study design, experiment execution, and manuscript writing. Dr. Carris directed the research, provided advice and assistance in designing the experiments and in editing the manuscripts. Many of the isolates used in the studies were collected by Dr. Carris or acquired through collaboration with the original collectors. Dr. Castlebury provided advice and guidance for the phylogenetic analyses, and sequencing support for most of the isolates used in the analysis.

125

APPENDIX

FIG. S-1. The scatter plot distribution of morphological characters of each wheat bunt isolate.

X-axis: the average diameter of teliospores. Y-axis: the average ratio of sorus length to width. The

basis for recognition of three morphological species is spore ornamentation (smooth vs. reticulate) and presence or absence of gelatinous sheath enveloping spores.

126

127

128

FIG. S-2. a. Maximum likelihood tree of 60 wheat bunt isolates reconstructed from three

anonymous loci. b. Maximum likelihood tree of 53 wheat bunt isolates reconstructed from

RPB2-primer-amplified region. Trees are unrooted, with maximum likelihood bootstrap (ML-BS) values and Bayesian posterior probabilities (MB-PP) values as support. Branch length indicates the substitution rate. Collectors’ species determination indicated before each taxon. Teliospore morphology, k-means clustering, and geographic origin of each sample are listed on the right side.

The first row indicates teliospore morphology: blue – T. contraversa type spore, red – T. caries type spore and green – T. laevis type spore. The second row indicates k-means clustering result: blue – dwarf bunt group, yellow– common bunt group, grey– data not available. The third row indicates geographic originations: Yellow – Oceania, Blue – Europe, Red – North American,

Green – Asia.

129

TABLE S-1. Host range for T. contraversa (also as T. brevifaciens) reported in US Pacific Northwest, 1950-1958

Host Original host Characters used Year Location Resource Reference (See species in identification Chapter 4) Aegilops cylindrica Symptons, spore 1955 Nephi, Utah Natural stand Dewey & Host morphology near wheat field Tingey 1958 Agropyron cristatum N/A 1954 Estern Oregon; Infested field or Hardison et al (L.) Gaertn. Worley, Idaho nursery 1955 Agropyron desertorum N/A 1954 Estern Oregon; Infested field or Hardison et al (Fisch. ex Link) Worley, Idaho nursery 1955 Schult. Agropyron fragile Agropyron N/A, Symptons 1954, Worley, Idaho Nursery, Spore Hardison et al (Roth) P. Candargy sibiricum, A. 1958 inoculation 1955, Hardison mongolicum et al 1959 Arrhenatherum elatius Dwarfing, 1952 Union Co, Grass fields Hardison & P. Beauv. ex J. Presl spores, Oregon Corden 1952 130 & C. Presl germination, Bromus carinatus Symptons 1958 Oregon; Spore Hardison et al Hook. & Arn. Washington inoculation 1959 Bromus ciliatus L. Symptons 1958 Washington Spore Hardison et al inoculation 1959 Bromus erectus Huds. Symptons 1958 Oregon Spore Hardison et al inoculation 1959 Bromus marginatus N/A 1955 Elgin, Oregon Natrual stand Meiner et al Nees ex Steud. 1956 1956 Bromus tomentellus Symptons 1958 Oregon Spore Hardison et al Boiss. inoculation 1959 Dactylis glomerata L. N/A 1954 Eatern Oregon Test nursery Hardison et al 1955 Elymus arizonicus A. arizonicum Symptons 1958 Oregon Spore Hardison et al (Scribn. & J.G. Sm.) inoculation 1959 Gould

Elymus canadensis L. Symptons 1956 Eastern Infested field Hardison & Oregon; Meiners 1956 Northern Idaho Elymus caninus (L.) A. caninum Symptoms 1958 Idaho; Oregon; Spore Hardison et al L. Washington inoculation 1959 Elymus ciliaris (Trin.) A. ciliare Symptons 1958 Washington Spore Hardison et al Tzvel. ssp. ciliaris inoculation 1959 Elymus ciliaris (Trin.) A. amurense Symptons 1957 Pullman, Spore Meiners & Tzvelev subsp. Washington inoculation Hardison 1957 amurensis (Drobow) Tzvelev Elymus glaucus Symptons 1957 Pullman, Spore Meiners & Buckley Washington inoculation Hardison 1957 Elymus lanceolatus A. dasystachyum, Symptons 1957 Oregon; Spore Meiners & (Scribn. & J.G. Sm.) A. riparium Pullman, inoculation, Hardison 1957 131 Gould ssp. Washington; Experimental Lanceolatus Planting Elymus repens (L.) A. repens Symptons 1957 Union Co, Natural stand Meiners & Gould Oregon near exp. Hardison 1957 planting Elymus sibiricus L. Symptons 1956 Eastern Oregon Infested field Hardison & Merners 1957 Elymus trachycaulus A. trachycaulum N/A 1955 Worley, Idaho Infested nusery Meiner et al (Link) Gould ex 1956 Shinners ssp. Trachycaulus Festuca brachyphylla Festuca ovina Symptons 1958 Idaho Spore Hardison et al Schult. ex Schult. & inoculation 1959 Schult. f. ssp. brachyphylla Festuca brevipila F. ovina var. Germination, 1952 Union Co, Seed samples, Hardison 1954, Tracey duriuscula spore 1958 Oregon Spore Hardison et al

morphology inoculation 1959 Festuca idahoensis Symptoms 1957 Oregon Experimental Meiners & Elmer planting Hardison 1957 Festuca rubra L. Germination, 1952 Union Co, Seed samples Hardison 1954, spore 1958 Oregon; Idaho Hardison et al morphology 1959 Festuca rubra L. ssp. F. rubra var. Symptoms 1958 Idaho Spore Hardison et al fallax (Thuill.) Nyman commutata inoculation 1959 Koeleria cristata Symptoms 1957 Oregon Experimental Meiners & (Ledeb.) Schult. planting Hardison 1957 Leymus triticoides Elymus Symptoms 1957 Oregon Experimental Meiners & (Buckley) Pilg. triticoides planting Hardison 1957 Lolium perenne L. Symptoms 1958 Washington Spore Hardison et al inoculation 1959 Lolium perenne L. ssp. Lolium N/A 1956 Northern IdahoExperimental Hardison & multiflorum (Lam.) multiflorum planting Meiners 1957 132 Husnot Poa palustris L. Symptoms 1958 Washington Spore Hardison et al inoculation 1959 Poa pratensis L. Germination, 1952 Union Co, Seed samples Hardison 1954 spore Oregon morphology Pseudoroegneria A. inerme N/A 1955 Union Co, Grass field Meiners et al spicata A. Löve ssp. Oregon 1955 inermis (Scribn. & J.G. Sm.) A. Löve Schedonorus phoenix F. arundinacea Germination, 1952 Union Co, Seed samples Hardison 1954 (Scop.) Holub spore Oregon morphology Schedonorus pratensis F. elatior Symptoms 1956 Eastern Oregon Infested field Hardison & (Huds.) P. Beauv. Meiners 1957 Seacle cereale L. Dwarfing, tillers, 1952 UT & CO Wheat fields Woodward et al sori, spores 1952

Thinopyrum A. intermedium, Type, 1950 Idaho, Oregon, Seed samples, Fischer 1952, intermedium (Host) A. trichophorum Germination, 1952, Washington grass fields Hardison 1954, Barkworth & D.R. spore 1955 Meiners 1956 Dewey morphology Thinopyrum ponticum A. elongatum N/A 1952 Eastern OR, Seed fields Hardison 1954, (Podp.) Z.-W. Liu & 1958 Washington Hardison et al R.-C. Wang 1959 133

TABLE S-2. Isolates studied as part of Chapter 4 for which sequence data could not be obtained

I. Isolates not used because of poor sequence data T. contraversa WSP 63685 1954 Triticum aestivum US Meiners (Washington) T. contraversa WSP 63890 1953 Secale cereale Turkey Gassner T. caries WSP 11821 1923 Triticum aestivum T. caries WSP 63716 1952 Triticum aestivum Romania Corney T. contraversa TK69B 1979 Aegilops cylindrica Turkey T. caries WSP 40527 1950 Bromus MT Fischer marginatus T. contraversa WSP 44944 1957 Elymus repens OR Hardison T. contraversa WSP 63590 1955 Lolium perenne NY Dewey T. contraversa WSP 63658 1954 Triticum aestivum Switzerland H. Aebi T. contraversa WSP 63682 1956 Schedonorus OR Meiners pratensis T. contraversa WSP 63741 1926 Thinopyrum Chirnyanovsk intermedium i T. contraversa WSP 63769 1936 Elymus repens Turkey Gassner T. contraversa WSP 63885 1953 Hordeum Turkey Gassner bulbosum T. contraversa WSP 63887 1954 Secale cereale Austria Niemann

II. Isolates that could not amplified by PCR T. contraversa BPI 172708 1878 Elymus repens T. contraversa WSP 64107 1876 Agropyron sp Germany Kunze T. contraversa WSP 63868 1890 Elymus repens Bohemia Bubak T. contraversa WSP 63871 1904 Elymus repens Denmark Lind T. contraversa WSP 63872 1891 Elymus repens Germany Oertel (BPI 172707) T. contraversa WSP 63873 1891 Elymus repens Germany Oertel (BPI 172717) T. contraversa WSP 63876 1905 Elymus repens Denmark Lind T. contraversa WSP 63879 1906 Elymus repens Unknown Buvalovia T. contraversa WSP 51956 1959 Thinopyrum Unknown Vánky intermedium T. caries WSP 63705 1943 Aegilops umbellata Turkey Ozkan T. contraversa WSP 63768 1947 Elymus repens Yugoslavia Lindtner T. contraversa WSP 63777 1933 Aegilops triuncialis Iran Zohary T. contraversa WSP 63842 1890 Agropyron sp Germany Oertel T. contraversa WSP 63844 1948 Agropyron Yugoslavia Lindtner cristatum T. contraversa WSP 63893 1921 Elymus repens Czechoslovakia Picbauer

134

FIG. S-3. Parametric bootstrap tests of alternative hypothesis on RPB2, EF1-alpha and ITS tree.

Constraint tree = all T. contraversa, T. caries and T. laevis isolates in a distinct monophyletic group. Five black bars indicate the frequency of likelihood score difference between constraint tree and unconstrained tree with 200 sets of simulated data. The value 4.2 is the likelihood score difference between constrained and unconstrained trees based on actual sequence data from

RPB2, EF1-alpha and ITS regions. The arrow indicates that the likelihood score difference was greater than 99.5% of simulated data (p=0.005).

135

136

137

138

139

140

141

142

FIG. S-4. Maximum likelihood tree of the T. contraversa (T. brevifaciens) isolates and related species with individual or combined gene trees. a. A13, b. A16, c. P18, d. A16 and P18 combined, e. RPB2, f. ITS, g. EF1-alpha. Trees are mid-point rooted, with maximum likelihood bootstraps

(ML-BS) values and Bayesian posterior probabilities (MB-PP) values as support. Branch length indicates the substitution rate. Species determination indicated before each taxon, and host species indicated after each taxon.

143