The Pennsylvania State University

The Graduate School

Department of Crop and Soil Sciences

PARASITIC CASTRATION BY A STINKING REGULATES SEX

DETERMINATION AND ARCHITECTURE IN DIOECIOUS

BUFFALOGRASS

A Thesis in

Agronomy

by

Ambika Chandra

© 2007 Ambika Chandra

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2007

The thesis of Ambika Chandra has been reviewed and approved* by the following:

David R. Huff Associate Professor of Turfgrass Breeding and Genetics Thesis Adviser Chair of Committee

Dawn S. Luthe Professor of Stress Biology

Barbara J.Christ Professor and Head of Plant Pathology

Paula McSteen Assistant Professor of Biology

David M. Sylvia Professor of Soil Microbiology Head of the Department of Department of Crop and Soil Sciences

*Signatures are on file in the Graduate School.

ii

ABSTRACT

PARASITIC CASTRATION BY A STINKING SMUT REGULATES SEX

DETERMINATION AND INFLORESCENCE ARCHITECTURE IN DIOECIOUS

BUFFALOGRASS

Buffalograss is a dioecious grass with male and female separated onto two different individuals. Buffalograss presumably evolved its present day dioecious breeding system from hermaphroditic ancestors. Pistil smut infection phenotypically alters unisexual florets of dimorphic male and female buffalograss making them hermaphrodites which are morphologically indistinguishable from one another. In this way, pistil smut infection allows the phenotypic expression of a hermaphroditic ancestor, which existed back in time, in the form of a retrospective phenotype or simply a retrophenotype. Underdeveloped stamens (male sex organs) and fungal sporulation within ovaries of male and female renders hermaphroditic florets of infected buffalograss reproductively sterile, i.e. parasitically castrated. Parasitic castration is a disease affecting animals, mollusks, insects, and plants where host gonads are sterilized preventing evolution of host resistance and redirecting resource allocation. How and why fungi induce hermaphroditism and cause parasitic castration in their hosts is unknown.

Here I show that pistil smut induces hermaphroditism in male buffalograss by down- regulating a female suppressor gene homologous to Tasselseed2 (Ts2). This molecular mechanism is supported by temporal and spatial expression analyses performed using quantitative real time PCR and in situ hybridization, respectively, on a cloned full length buffalograss Ts2 homolog. In addition to inducing hermaphroditism, pistil smut infection iii

enhances overall sexual reproductive allocation (SRA) in both male and female sex forms

of buffalograss. In female plants, the induces a 12.6-fold increase in ovary production while in male plants the fungus induces a 2.3-fold increase in floret number.

Furthermore, phylogenetic analyses of morphology and nuclear 28s large subunit of

ribosomal DNA (nLSU-rDNA) sequences show that pistil smut is clearly distinct from any other species of suggesting that it may not even be a species of Tilletia. The unusually large genetic distance of pistil smut from Tilletia clade indicates an accelerated rate of evolution possibly due to pistil smut’s long term coevolutionary struggle with its host culminating in induced hermaphroditism. Therefore, a new Salmacisia is

proposed and described to accommodate pistil smut.

iv

TABLE OF CONTENTS

LIST OF FIGURES ...... viiii

LIST OF TABLES...... x

Chapter 1 INTRODUCTION...... 1

BIBLIOGRAPHY ...... 13

Chapter 2 LIFE CYCLE AND PHYLOGENY OF TILLETIA BUCHLOËANA; A FUNGAL PARASITE CAUSING INDUCED HERMAPHRODITISM IN DIOECIOUS BUFFALOGRASS (BUCHLOË DACTYLOIDES)...... 19

ABSTRACT ...... 19 INTRODUCTION...... 20 MATERIALS AND METHODS ...... 23 Isolation and maintenance of pistil smut...... 23 Host infection by pistil smut ...... 23 Scoring and analysis of pistil smut morphology ...... 24 Nucleic acid extraction and PCR amplification ...... 24 Taxa examined for phylogenetic analyses ...... 25 Taxa sequence alignments...... 26 Phylogenetic analyses ...... 27 RESULTS...... 29 Induced hermaphroditism and parasitic castration of buffalograss by pistil smut...... 29 Morphological characteristics of pistil smut fungus in vitro...... 31 Life cycle of pistil smut in vivo ...... 32 Phylogenetic analysis of pistil smut based on morphology and life cycle characteristics ...... 33 Phylogenetic placement of pistil smut based on nLSU and ITS sequence analyses ...... 35 ITS...... 39 Gamma distribution...... 39 DISCUSSION...... 40 BIBLIOGRAPHY ...... 45

Chapter 3 SALMACISIA, A NEW GENUS OF : RECLASSIFICATION OF TILLETIA BUCHLOËANA CAUSING INDUCED HERMAPHRODITISM IN BUFFALOGRASS...... 62

TAXONOMY...... 63 Salmacisia D. R. Huff & A. Chandra, gen. nov...... 63 v

Sori...... 63 Type species ...... 63 Etymology...... 64 Salmacisia buchloëana (Kellerman & Swingle) D. R. Huff & A. Chandra comb. nov...... 64 Sori...... 65 Characteristic DNA sequences ...... 65 Characteristic fixed DNA polymorphisms ...... 66 Specimen examined ...... 67 Known distribution...... 67

Chapter 4 INDUCED HERMAPHRODITISM CAUSED BY PARASITIC CASTRATION DISPLAYS AN ANCESTRAL PHENOTYPE BY REMOVING A SEX DETERMINING ONTOGENETIC LAYER FROM DIOECIOUS BUFFALOGRASS...... 69

ABSTRACT ...... 69 INTRODUCTION...... 69 MATERIALS AND METHODS ...... 74 Plant Material ...... 74 Morphological stages of buffalograss inflorescence development ...... 74 Scanning Electron Microscopy ...... 74 Isolation of full length Ts2 homolog (BdTs2) from buffalograss...... 75 Quantitative Real-Time PCR ...... 76 RNA in situ hybridization ...... 77 RESULTS...... 78 DISCUSSION...... 82 BIBLIOGRAPHY ...... 85

Chapter 5 TEN-FOLD INCREASE IN POTENTIAL SEED YIELD DUE TO FUNGAL INFECTION OF BUFFALOGRASS ...... 99

ABSTRACT: ...... 99 INTRODUCTION...... 99 MATERIALS AND METHODS ...... 102 Isolation of pistil smut...... 102 Plant Material ...... 102 Host infection by pistil smut ...... 103 Resource partitioning analysis ...... 103 Analysis of components of sexual reproductive allocation...... 103 RESULTS...... 104 DISSCUSSION ...... 107 BIBLIOGRAPHY ...... 108

Chapter 6 RATIONALE FOR FUTURE STUDIES ON BUFFALOGRASS-PISTIL SMUT INTERACTION ...... 115 vi

BIBLIOGRAPHY ...... 121

Chapter 7 SUMMARY AND CONCLUSION ...... 125

BIBLIOGRAPHY ...... 128

Appendix A PISTIL SMUT INFECTION OF BUFFALOGRASS RELATIVES BELONGING TO PACC CLADE UNDER CONTROLLED CONDITIONS OF GREENHOUSE ...... 129

Appendix B SIZE MEASUREMENT OF SEX ORGAN WITHIN UNISEXUAL MALE AND FEMALE BUFFALOGRASS FLORETS...... 130

Appendix C SOUTHERN BLOTTING TO DETERMINE TS2 COPY NUMBER IN DIPLOID BUFFALOGRASS...... 132

Appendix D BDTS2 EXPRESSION IN NODES AND LEAVES OF BUFFALOGRASS ...... 134

Appendix E PRELIMINARY METABOLOMIC ANALYSIS AND BIOASSAY TO ISOLATE THE TRIGGER COMPOUND FROM BUFFALOGRASS UPON PISTIL SMUT INFECTION.....135

Metabolomic analysis...... 135 Preliminary bioassay ...... 136

Appendix F PARASITIC CASTRATION AND THE EVOLUTION OF DIOECY IN BUFFALOGRASS...... 140

BIBLIOGRAPHY ...... 144

vii

LIST OF FIGURES

Figure 1.1 to Figure 1.5: Distribution and morphology of buffalograss...... 17

Figure 2.1 to Figure 2.8: Effects of pistil smut infection on male and female sex forms of buffalograss...... 49

Figure 2.9 to Figure 2.19: Stages of pistil smut life-cycle in vitro and in vivo...... 51

Figure 2.20: A Minimum Evolution (ME) distance-based phylogenetic tree generated by simple matching of six morphological and life cycle characteristics of pistil smut along with 40 other species of order Tilletiales...... 53

Figure 2.21 to Figure 2.26: Phylogenetic placement of pistil smut within order Tilletiales based on nLSU...... 56

Figure 2.27 to Figure 2.29: Phylogenetic placement of pistil smut within order Tilletiales based on the entire ITS region including ITS1-5.8S-ITS2 DNA sequences...... 57

Figure 2.30 to Figure 2.31: Graphical representation of the expected number of substitutions (k) of nucleotide sites plotted against the position of each site for species of Tilletiales with pistil smut (blue) and without pistil smut (pink)...60

Figure 4.1 to Figure 4.7: Pistil smut induced hermaphroditism and development of additional florets in buffalograss...... 89

Figure 4.8: Stages of male and female buffalograss inflorescence development...... 91

Figure 4.9: Scanning electron micrographs depicting the ontogenetic events of buffalograss unisexual development...... 92

Figure 4.10: Multiple sequence alignment of predicted amino acids for Ts2 gene....94

Figure 4.11: Hypervariable region (36 nt.) long, comprising of a combination of direct, inverted, present in buffalograss, maize and Tripsacum but absent in species...... 96

Figure 4.12: Role of the female suppressor gene, BdTs2, in buffalograss unisexual floret development and its regulation by pistil smut...... 98

Figure 5.1 to Figure 5.3: Effect of pistil smut infection on buffalograss resource partitioning...... 109

viii

Figure 5.4: Developmental changes in inflorescence architecture of male buffalograss induced by pistil smut...... 111

Figure 6.1: One of the most parsimonious tree displaying relationship of 58 Tasselseed2 (Ts2) and related short-chain dehydrogenase/reductase (SDR) genes...... 123

Figure C.1: Southern blot of diploid buffalograss genome with BdTs2...... 133

Figure D.1: BdTs2 expression in leaves and nodes of male and female buffalograss infected with pistil smut, as determined by quantitative real-time PCR...... 134

Figure E.1: Chromatographs with a mass range of 775-800 amu extracted from acetone extracts of non-infected (top) and infected infected males, (middle) male buffalograss plants and pure cultures of pistil smut fungus (bottom)...... 138

Figure E.2: Mass spectra of the significant peak 14.66 from the three spectra: non-infected males, infected males and pure fungal cultures...... 139

Figure F.1: Model for parasitic castration and the evolution of dioecy in buffalograss...... 145

ix

LIST OF TABLES

Table 2.1: List of taxa and GenBank accession numbers for nLSU and ITS regions of rDNA used in the phylogenetic analyses...... 61

Table 5.1: Mean resource partitioning for identical genotypes (clones) of either female or male sex forms of buffalograss that were infected or non-infected with the pistil smut fungus. Means are the averages of genotypes for two replications from 2005 summer harvest of greenhouse grown plants. Mean differences between non-infected and infected clones were tested for significance using Student’s paired t-test. Standard error = (SE)...... 112

Table 5.2: Mean resource partitioning for identical genotypes (clones) of either female or male sex forms of buffalograss that were infected or non-infected with the pistil smut fungus. Means are the averages of genotypes for two replications from 2006 summer harvest of greenhouse grown plants. Mean differences between non-infected and infected clones were tested for significance using Student’s paired t-test. Standard error = (SE)...... 113

Table 5.3: Mean estimated components of sexual reproductive allocation for identical genotypes (clones) of either female or male sex forms of buffalograss that were infected or non-infected with the pistil smut fungus Means are the averages of genotypes for two replications from 2006 summer harvest of greenhouse grown plants. Mean differences between non-infected and infected clones were tested for significance using Two-tail Student’s paired t-test. Standard error = (SE)...... 114

Table B.1: Means, standard deviations (std), and fold differences of morphological characteristics between pistil smut infected and non-infected genotypes of buffalograss...... 131

x

Chapter 1

INTRODUCTION

Evolution of sex, with the two-fold (50%) cost associated with its origin and

maintenance (Smith 1978), presents a challenging puzzle for evolutionary biologists

since early 1900s. The cost of sex is realized only in out-crossing species and not in self-

fertilizing hermaphrodites (Williams 1975). Individuals within an out-crossed population

have separation of sexes into male and female sex forms, either on the same plant

(monoecious) or on different plants (dioecious), most often in a 1:1 sex ratio

(Westergaard 1958, Bull 1983). Only female sex forms (50% of the individuals) within

an out-crossed population are capable of bearing offspring, resulting in 50% fitness

(reproductive success) of the population, and thus accounts for 50% cost of sex (Smith

1978). In addition, individuals of an out-crossed population transmit their genes at only half the rate as compared to self-fertilizing or asexually reproducing organisms, thereby, diluting the genome of their own offspring by 50% every generation (Williams 1975).

These factors along with the need of male and female to search for each other to mate, makes sex a costly endeavor. During evolution, a net selective disadvantage of 1% can cause a gene to rapidly disappear from a population as a consequence of natural selection

(Smith 1978, Williams 1975). Therefore, sex would have never originated and maintained in a population unless it had any significant evolutionary advantage to offer.

Several theories have been put forward by evolutionary biologists to explain the evolution of sex but all are quiet difficult to test experimentally, and hence are speculative. Out-crossing has supposedly evolved from self-fertilizing hermaphrodites as

1

a result of mutation of at-least two genetic loci (male and female sterility) coupled with complete linkage (Westergaard 1958; Bull 1983). The most common theory for the evolution of sex from self-fertilizing or asexual progenitors is that sexual reproduction leads to genetic recombination creating genetic variation. As a result, sexual reproduction can bring together beneficial mutations into a single individual; can bring

together deleterious mutations creating several unfit individuals that are then eliminated

from the population (avoiding Muller’s ratchet); and, can also create new gene

combinations that make individuals more fit than the previously existing relatives.

Sexual reproduction also serves as means to escape inbreeding depression (Charlesworth

1978) and optimize resource allocation (Charnov 1982). Overall, sex is believed to be an

adaptation to uncertain and changing conditions (Williams 1975) including changes in

biotic and abiotic environments. The most widely accepted explanation for the evolution

of sex is that sex serves as a defense against parasites, as described in the Red Queen

hypothesis (Hamilton 1990). In asexually reproducing organisms, the only way to

generate resistance against a newly arisen parasitic virulence is through mutation, the

frequency of which is quiet low. On the other hand, sexual reproduction enables

individuals to create new gene combinations at higher rates through genetic

recombination, thereby, repeatedly generating new resistance against parasitic virulence.

In this way, sexual reproduction facilitates a coevolutionary arms race between host and

parasite, where the parasite exerts a selective force on its host to a evolve resistance mechanism and the resistance developed in the host exerts a selective force on the parasite to evolve an alternative virulence mechanism.

2

In some species, this ongoing coevolutionary arms race between host and parasite is put to an end when the parasite sterilizes host gonads (gamete-producing organs), thereby limiting the ability of the host to sexually reproduce. Loss of sexual reproduction stops the host from sexually recombining mutations to generate new resistance mechanisms against the parasitic virulence, and leads to the ultimate victory of the parasite over its host in this arms race. This intriguing illustration of coevolution is known as parasitic castration, and is a disease affecting animals, mollusks, insects and plants where host gonads are sterilized by the parasite either by preventing development or by causing atrophy of host gonads (Boudoin 1975). Loss of sexual reproduction, as a consequence of parasitic castration, yields a potential selective advantage to the parasite by limiting the host’s ability to coevolve resistance mechanisms. It may also lead to changes in host resource allocation (Clay 1991) which in turn enhances the survival and transmission of the parasite within the vegetative body of its host.

Parasitic castration is a widespread phenomenon. Examples include grasses infected by systemic fungal endophytes which are typically reproductively sterile; bunts; phyllody caused in maize by corn smut; crazy top in maize caused by Sclerospora macrospore; big blue stem infected with Sorosporium everhartii; anther smut infected plants of red and white campion; twisted-wing parasite (Strepsiptera) infected insects; and, golden crab infected with barnacles (Clay 1991; Fischer and Holton 1957; Boudoin

1975). Despite its wide occurrence in the nature, the molecular or biochemical mechanism underlying parasitic castration has yet to be discovered.

3

Originally, the term parasitic castration was coined by Alfred Giard (1888-1902) in reference to the phenomenon known as induced hermaphroditism. This interesting effect of parasitic castration occurs in some unisexual organisms where the parasite not only causes the atrophy of the attending sex but also induces hypertrophy of opposite sex organs making the host morphologically hermaphrodite (Fischer and Holton 1957). The resulting variant hermaphroditic morphs are reproductively sterile as the hypertrophied sex serves as a site for parasite reproduction, and the atrophied sex is underdeveloped and non-functional. Induced hermaphroditism has long been known, and basically three types of induced hermaphroditism have been recognized: androgene, when female sex forms are stimulated to contain male characteristics; thelygene, when female characteristics are induced in male sex forms; and amphigene, when the parasite stimulates the development of the opposite sex organs, in both sex forms (Fischer and

Holton 1957). For instance, an example of androgene is anther smut fungus induced male sex organs in female plants of white campion (Uchida et al. 2003), and an example of thelygene is rhizocephalans induced female sex organs in male golden king crabs

(Shukalyuk et al. 2005).

A rare example of induced hermaphroditism in grasses was observed by

Kellerman and Swingle (1889) who discovered a smut fungus (Tilletia buchloëana Kell

& Swing.; recently proposed as Salmacisia buchloëana (Kell & Swing.) Huff & Chandra) whose “monstrosity lies solely in its ability to produce ovaries in of male buffalograss”. Due to its pistil inducing ability (female sex organs), we refer to this smut fungus as pistil smut. To date, the buffalograss-pistil smut interaction has never been the subject of scientific study beyond the observation of infection on male plants in the

4 original taxonomic description of the fungus by Kellerman and Swingle (1889). The only documentation that female plants were even capable of being infected with pistil smut was illustrated by Duran (1987), but he did not describe the effect of fungal infection on the floral organs. This lack of scientific information is most likely due to the rarity of the fungus in nature.

In the present study, we show that pistil smut induces hermaphroditism in not only florets (the modified flowers of grasses) of male buffalograss but also in florets of female buffalograss and thus, is an example of amphigene type of induced hermaphroditism and represents a unique form of parasitic castration. In addition to induced hermaphroditism, pistil smut infection alters the inflorescence architecture of buffalograss at the level of meristem determinacy. Since little to no knowledge exists regarding the basic biology of buffalograss-pistil smut interaction, a description of the parasite and the host seems warranted.

Pistil smut was first collected from plants of male buffalograss growing in Trego and Ford County of Kansas State in 1888 by William A. Kellerman, and his then student

Walter T. Swingle (Kellerman and Swingle 1889). Pistil smut has a very narrow host range (more details in Appendix A); in addition to buffalograss, under natural conditions, it has been reported to infect Cathestecum erectum, Hilaria belangeri and Muhlenbergia distichophylla. Whether or not pistil smut infection causes induced hermaphroditism of these host species is yet unknown.

Buffalograss, Buchloë dactyloides (Nutt.) Engelm., Syn. Bouteloua dactyloides

(Nutt.) Columbus, is derived from the Greek word "boubalos" (buffalo) and "chloe"

5

(grass), because it once served as a primary food source for the enormous herds of

American bison (Bison bison L.) and continues to serve as an important species for

grazing lands and low maintenance turf in Midwestern states. Buffalograss is an

excellent example of the “Foliage is the Fruit” (FF) hypothesis (Quinn et al. 1994) which

suggests that large grazing animals, like buffalo, ingest and later disperse seeds of the

grass while consuming their foliage. Because of its excellent ground covering ability

with aggressive spreading (stoloniferous growth habit) under heavy use, buffalograss was widely planted to prevent wind erosion of soil after the Dust Bowl days of the late 1930s.

Buffalograss also provided the from which early Midwestern settlers built their houses in early 1800s (Harper 1938).

Buffalograss originated in Central Mexico and is distributed throughout the Great

Plains of North America covering north to south from Canada to Mexico and west to east from the eastern slope of the Rocky Mountains to the Mississippi Valley (Hitchcock

1951) (FIG. 1.1). Its present day distribution contains three autopolyploid chromosomal

races (2n=2x, 4x, or 6x, where x= 10 chromosomes (Gould 1975), and likely contains

clonal plants of genotypes whose original establishment date back to the era of

Pleistocene glaciations (Stebbins 1975). Buffalograss is native to the short-grass of North America and is the only native grass used for landscape and utility turf, requiring the least amount of annual precipitation (38 cm) to develop a solid stand of turf

(Keen 1969). It is also tolerant of extreme heat, cold and drought. Buffalograss is a low growing grass with low mowing requirements, and hence is ideally suited for establishing a low maintenance turf stand.

6

Buffalograss is a warm season (C4) perennial grass and exhibits a wide variety of

sexual breeding systems. Buffalograss predominantly exhibits an environmentally stable,

dioecious breeding system (FIG. 1.2) with approximate 1:1 male to female sex ratios

(Huff and Wu 1987, 1992). Buffalograss most likely evolved dioecious sex expression

from a hermaphroditic ancestor. Hermaphroditic florets containing both male and female

sex organs within the same floret have, on rare occasion, been observed in male spikelets

(Wenger 1940) (FIG. 1.3). These developmental variants (hermaphroditic florets) are

fully capable of setting viable seeds. Populations of buffalograss may also contain

monoecious individuals (FIG. 1.4) whose frequencies are negatively correlated with plant densities (Huff and Wu 1992). Because of the presence of low frequency (0-10 %) of

monoecious individuals, in addition to dioecious individuals, buffalograss is most

accurately termed sub-dioecious.

The structural difference in inflorescence architecture between male and female

plants is so striking that buffalograss serves as the definition of dimorphic, dioecious sex

expression among North American grass species (FIG. 1.5). Male inflorescences are elevated well above the foliage with two to three spikes branched off the main culm axis

(FIG. 1.5). Each male spike contains three to eight spikelets. The male spikelet is

comprised of a pair of papery outer glumes and two staminate florets (termed here,

primary and secondary). Each male floret contains a pair of inner glumes (lemma and

palea) and three stamens (anther and filament) (FIG. 1.5). Female inflorescences are

much shorter than males and remain hidden in the foliage (FIG. 1.5). The outer glumes of

female spikelets become thickened and fused during development forming a burr-like

7

structure that permanently encloses the seed (caryopsis). Female burrs (analogous to a

male spike) contain two to six spikelets. Each female spikelet contains a single floret

(termed here, primary). Each female floret contains a pair of inner glumes (lemma and palea), a pair of lodicules, and a central pistil (ovary, style and purple feathery stigma)

(FIG. 1.5).

Buffalograss represents an example of sexual dimorphism in the form of

dioecious sex expression, however, the pathway involved in the separation of sexes

during the development of buffalograss unisex florets is not known. Unisexuality

(separate sexes) is a common phenomenon in animals however, about 90 % of the

flowering plants (angiosperms) are hermaphrodites (bisexual) while the remaining 10 %

are either monoecious or dioecious (Dellaporta and Calderon-Urrea 1993). Flowering

plants exhibit a wide variety of sexual reproductive systems (Barrett 2002). Several sex

determining mechanisms have been identified in other organisms, for example,

environmental sex determination as in most turtles and crocodiles, male and female

heterogamety (X/Y system of sex determination) as in Silene latifolia, humans and other

mammals, polyfactorial sex determination as in swordtails, and haplo-diploidy, where

females are produced from the fertilized eggs and males are produced from unfertilized eggs, as in thrips, mites and ticks (Bull 1983). Sex determination has also been shown to be under the genetic control of some sex determining loci, a minimum of two (male suppressor and female suppressor) are needed for the separation of sexes to evolve

(Westergaard 1958).

In maize, a female suppressor genetic loci, Ts2, has been shown to play an important role in the formation of staminate florets (DeLong et al. 1993). The

8

development of pistillate florets is under the control of a different genetic mechanism and

is decoupled with the Ts2-mediated development of the staminate florets (Malcomber and

Kellogg 2005). Mutations that effect gibberellin biosynthesis in maize like anther ear1

and dwarf show defects in stamen abortion in the developing ear of a maize plant,

indicating that GAs play a key role in the development of pistillate florets (Phinney 1961,

1984).

A major advance in our research was the realization that the overall effect of pistil

smut infection on the sexuality and development of buffalograss unisex florets

phenotypically mimics maize tasselseed mutants. In Zea mays L. (maize), Tripsacum dactyloides L. (eastern gamagrass) and in many other grass species, florets start

development as hermaphroditic precursors, initiating both pistil and stamen primordia

(Le Roux and Kellogg 1999; Mitchell and Diggle 2005). Unisexual florets are then

formed by the selective elimination of the opposite sex organs in the developing bisexual

florets (Mitchell and Diggle 2005; Cheng et al. 1983). In maize, pistil elimination in lower and upper male (tassel) staminate florets requires a cell death pathway protein

encoded by Tasselseed2 (Ts2) (Delong et al. 1993). In female (ear) pistillate florets of

maize, Ts2 activity is blocked (presumably by Sk, silkless) (Calderon-Urrea and

Dellaporta 1999). In ts2 mutant maize plants and gsf1 (an orthologue of Ts2) in eastern

gamagrass, pistils in lower and upper florets of male spikelets fail to abort resulting in

viable ovaries in otherwise male staminate florets (Irish and Nelson 1993; Li et al. 1997).

In female florets of ts2 mutant plants, upper pistils fail to abort resulting in two (lower

and upper) pistillate florets (Calderon-Urrea and Dellaporta 1999; Irish and Nelson

1993).

9

Traits exhibited by ts2 mutant phenotypes are similar to those we have observed for buffalograss infected with pistil smut, i.e. development of pistils in both primary

(lower) and secondary (upper) male florets and the development of a secondary floret in female spikelets. The mutant phenotypes suggested that pistil smut might, either directly or indirectly, regulate Ts2-like activity, inducing hermaphroditic florets. The Ts2 gene product shows similarity to short-chained alcohol dehydrogenases, particularly hydroxysteroid dehydrogenases (Delong et al. 1993), and thus is suggestive of being involved in either or both the gibberellin and/or brassinosteroid pathways. It is important to note that pistil smut induced hermaphroditic variants of male florets are reproductively sterile as a result of parasitic castration.

Parasitic castration, in addition to preventing host sexual reproduction, may also manipulate host resource allocation in order to increase its own survival and fitness

(Boudoin 1975). A potential adaptive strategy of the parasite is to sterilize the host sexual reproductive organs, and to use the energy and resources that were going to be used for sexual reproductive towards the vegetative growth of the host. In doing so, the increased vegetative biomass of the host can support the parasite’s own growth and reproduction and may also enhance transmission of the parasite. Numerous examples of parasitic castration by fungi have been shown to increase plant vegetative growth, vegetative biomass and survivability (Clay 1991). Smuts, in particular, have demonstrated an ability to increase the vegetative vigor of many of their hosts (wheat, sorghum, common millet, and rye) (Clay 1991), but not every smut fungus-plant host

system shows increased vegetative growth; for example, the white campion-anther smut

system (Alexander and Antonovics 1988). However, I am unaware of any example

10

where parasitic castration allocates resource towards host sexual reproduction. The

present study on buffalograss/pistil smut system will prove to be an example of parasitic

castration where pistil smut dramatically enhances sexual reproductive allocation (SRA) of its host. SRA is essentially the amount of energy and resources that an organism allocates towards the process of sexual reproduction (Charnov 1982). Evolution molds

SRA differently in different organisms depending on environmental adaptations and life history strategies (Charnov 1982). Resources allocated for sexual reproduction may be further partitioned into male versus female function (sex allocation) and is governed by a trade-off hypothesis whereby gains in fitness of one sex must be greater than the cost of reducing allocations to the opposite sex for the evolution of separate sexes to occur

(Charnov1982; Wilson 1983). When no such fitness gains are realized, then hermaphroditism (bisexual) is the favored evolutionary stable strategy (ESS) (Wilson

1983).

Coevolution among organisms provides a rich source of interesting biology to study. Examples like the pollinator mimicry of terrestrial orchids, the nitrogen-fixing bacteria associated with legumes, and the behavioral modification of hosts by parasites each provide a source of endless fascination. The present study portrays an example of parasitic castration of buffalograss by pistil smut where the parasite dramatically alters the host morphology by inducing hermaphroditism in unisexual male and female florets and also alters resource partitioning within infected plants thereby increasing host SRA.

The buffalograss-pistil smut system has much to offer as a model system in the areas of

11 reproductive biology, evolution of sexual dimorphism in plants, developmental biology, resource partitioning of SRA and improved crop seed production.

12

BIBLIOGRAPHY

Alexander H and Antonovics J. 1988. Disease spread and population dynamics of anther-

smut infection of Silene alba caused by the fungus Ustilago violacea. J Ecol 76: 91-

104.

Antonovics J, Thrall PH, Jarosz AM and Stratton D. 1994. Ecological genetics of

metapopulations: the Silene-Ustilago plant-pathogen system. In: Real, L. (ed.)

Ecological Genetics. pp. 146-170. Princeton University Press.

Barrett SCH. 2002. The evolution of plant sexual diversity. Nat Rev Gen 3: 274−284.

Boudoin M. 1975. Host castration as a parasitic strategy. Evolution 29 (2): 335-352.

Bull JJ. 1983. Evolution of Sex Determining Mechanisms. Benjamin/Cummings Publ.

Co., Menlo Park, California.

Calderon-Urrea A and Dellaporta SL. 1999. Cell death and cell protection genes

determine the fate of pistils in maize. Development. 126: 435-441.

Charlesworth B and Charlesworth DA. 1978. Model for the evolution of dioecy and

gynodioecy. Am. Nat. 112: 975-997.

Charnov EL. 1982. The Theory of Sex Allocation. Princeton University press, New

Jersey.

Cheng PC, Greyson RI and Walden DB. 1983. Organ initiation and the development of

unisexual flowers in the tassel and ear of Zea mays. Am. J. Bot. 70(3): 450-462.

Clay K. 1991. Parasitic castration of plants by fungi. Trends Ecol. Evol. 6: 162-166.

13

Dellaporta SL and Calderon-Urrea A. 1993. Sex determination in flowering plants. Plant

Cell 5: 1241-1251.

Delong A, Calderon-Urrea A and Dellaporta SL. 1993. Sex determination gene

TASSELSEED2 of maize encodes a short-chain alcohol dehydrogenase required for

stage-specific floral organ abortion. Cell 74: 757-768.

Fischer GW and Holton CS. 1957. Biology and Control of the Smut Fungi. Ronald Press,

New York.

Gould FW. 1975. The Grasses of Texas. Texas A&M University Press, College Station.

Hamilton WD, Axelrod R and Tanese R. 1990. Sexual reproduction as an adaptation to

resist parasites (a review). Proc. Natl. Acad. Sci. 87: (9) 3566-3573.

Harper RE. 1938. Homesteading in Northwestern Oklahoma territory. Chronicles of

Oklahoma. 16: 326-336.

Hitchcock AS. 1951. Manual of the grasses of the United States. Misc. Publ. No. 200.

Washington, DC: U.S. Department of Agriculture, Agricultural Research

Administration. 1051 p.

Huff DR and Wu L. 1992. Distribution and inheritance of inconstant sex forms in natural

populations of dioecious buffalograss (Buchloë dactyloides). Am. J. Bot. 79(2): 207-

215.

Huff DR and Wu L. 1987. Sex expression in buffalograss under different environments.

Crop Sci. 27: 623-626.

Irish EE and Nelson TM. 1993. Development of TASSELSEED2 inflorescence in maize.

Am. J. Bot. 80(3): 292-299.

14

Keen R. 1969. Turfgrasses under semi-arid and arid conditions. In: Turfgrass Science,

A.A. Hanson and F.V. Juska (eds.). Amer. Soc. Agrom. Monograph No. 14.

Le Roux LG and Kellogg EA. 1999. Floral development and the formation of unisexual

spikelets in the Andropogoneae (). Am. J. Bot. 86: 354-368.

Li D, Blakey CA, Dewald C and Dellaporta SL. 1997. Evidence for a common sex

determining mechanism for pistil abortion in maize and its wild relative Tripsacum.

Proc Natl. Acad. Sci. 94: 4217-4222.

Malcomber ST and Kellogg EA. 2006. Evolution of unisexual in grasses (Poacea) and the

putative sex determining gene TASSELSEED2 (Ts2) New Phytologist 170:885-899.. 9

Maynard SJ. 1978. The Evolution of Sex. Cambridge University Press.

Mitchell CH and Diggle PK. 2005. The evolution of unisexual flower: Morphological and

functional convergence results from diverse developmental transition. Am. J. Bot.

92(7): 1068–1076.

Phinney BO. 1961. Dwarfing genes in Zea mays and their relation to the gibberellins. In:

Klein RM, ed. Plant growth regulation. Ames, IA, USA: Iowa State University Press,

489–501.

Phinney BO. 1984. Gibberellin A1, dwarfism and the control of shoot elongation and

metabolism in higher plants. In: Crozier A, Hillman JR, eds. The biosynthesis and

metabolism of plant hormones. Cambridge, UK: Cambridge University Press, 17–41.

Quinn JA, Mowrey DP, Emanuele SM and Whalley RDB. 1994. The “Foliage is the

fruit” Hypothesis: Buchloe dactyloides (Poaceae) and the Shortgrass of North

America. Am J Bot. 81 (12): 1545-1554.

15

Shukalyuk AI, Isaeva VV, Pushchin II and Dolganov SM. 2005. Effects of the

Briarosaccus callosus Infestation on the Commercial Golden King Crab Lithodes

aequispina. J. Parasitology 91 (2): 1502-1504.

Stebbins GL. 1975. The role of polyploid complexes in the evolution of North American

grasslands. Taxon. 24(1): 91-106.

Uchida W, Matsunaga S, Sugiyama R, Kazama Y and Kawano S. 2003. Morphological

development of anthers induced by the dimorphic smut fungus Microbotryum

violaceum in female flowers of the dioecious plant Silene latifolia. Planta. 218: 240-

248.

Wenger LE. 1940. Inflorescence variations in buffalograss, Buchloë dactyloides. J Am.

Soc. Agron. 32: 274-277.

Westergaard M. 1958. The mechanism of sex determination in dioecious flowering

plants. Adv. Genet. 9:217-281.

Williams GC. 1975. Sex and Evolution. Princeton University Press, Princeton, N.J.

Wilson MF. 1983. Plant reproductive Ecology. John Wiley & Sons, New York.

16

1.1 1.2

1.3 Unisex Inflorescence flower 1.5 Pistil

Male Anther Stamen Anthers Filament

1.4 Stigma Male Inflorescence Female Style Pistil

Female Inflorescence Ovary

Figure 1.1 to Figure 1.5: Distribution and morphology of buffalograss.

1.1, Distribution of the three autopolyploid races within North America along with metaphase (6x) and telophase (2x) chromosome preps. 1.2, Drawing of female (top) and male (bottom) sex forms of buffalograss (from Hitchcock 1951). 1.3, Naturally occurring rare hermaphroditic variant of male floret containing both stamens and a pistil.

1.4, Monoecious plant bearing both male and female inflorescences. 1.5, Male inflorescence, typically bearing 3-4 spikes and extend above the plant foliage. Each male

17

floret contains three functional stamens (anther and filament). Female inflorescence, typically bearing 2-3 spikes (seed dispersal capsules) and remains low within the foliage.

Each female floret contains a functional pistil (ovary, style, and a pair of purple feathery stigmas).

18

Chapter 2

LIFE CYCLE AND PHYLOGENY OF TILLETIA BUCHLOËANA; A FUNGAL PARASITE CAUSING INDUCED HERMAPHRODITISM IN DIOECIOUS BUFFALOGRASS (BUCHLOË DACTYLOIDES)

ABSTRACT: A fascinating illustration of parasitic castration is the fungal induced sex alteration of plants known as induced hermaphroditism. For 118 yrs, Tilletia buchloëana has been known to induce female sex organs (pistils) in male plants of buffalograss, making them hermaphrodite. Here we show that T. buchloëana induces hermaphroditism in not only male sex forms of buffalograss by inducing the development of otherwise vestigial pistils, but also in female sex forms by inducing hypertrophy of otherwise vestigial stamens (male sex organs). Both male and female sex organs within these induced hermaphroditic florets are parasitically castrated but apparently for different sex-specific reasons. In addition to inducing hermaphroditism, the fungus also induces the development of additional pistillate florets in both infected male and female plants. Due to its pistil inducing effects, we refer to T. buchloëana as pistil smut. We also report the life cycle characteristics and phylogeny of pistil smut based on morphology, nuclear large subunit (nLSU), and internal transcribed spacer (ITS) rDNA sequences in order to accurately place pistil smut within order Tilletiales. We describe in vitro and in vivo life cycle of pistil smut, and that pistil smut exhibits a set of morphological and life cycle characteristics that are unique among species of order

Tilletiales. Phylogenetic analyses based on maximum parsimony, maximum likelihood and genetic distance of rDNA sequences show that pistil smut exhibits an elevated rate of

19

nucleotide substitution and is as, or more, distant from Tilletia species than the basal group patelli. As such, these morphological and molecular analyses place

pistil smut outside the current taxonomic circumscription of genus Tilletia. Therefore, we conclude that pistil smut is unlike any other species of Tilleitia and that insufficient evidence was displayed to accommodate pistil smut within the genus Tilletia.

Key words: Buchloë dacyloides, ITS, life cycle, molecular phylogenetics, morphology, nLSU, parasitic castration, smut fungi

INTRODUCTION

One of the most interesting examples of coevolutionary biology is the alteration of sex in a host by a parasite. In 1889, William A. Kellerman, an early American mycologist and founder of the Journal of Mycology (later renamed Mycologia), and his then student Walter T. Swingle, who subsequently became a legendary agricultural botanist (Seifriz 1953), discovered a smut fungus that induces the production of ovaries in florets of otherwise male plants of buffalograss [Buchloë dactyloides (Nutt.) Engelm.]

(Kellerman and Swingle 1889). They placed the fungus within the genus Tilletia

(Phylum: , Order: Tilletiales) based on teliospore characteristics, naming it

Tilletia buchloëana after its host. The phenomenon of inducing opposite sex organs in individuals which would otherwise bear only a single sex is known as induced hermaphroditism and occurs in a wide variety of organisms either via biological agents

(ex. parasites) (Fisher and Holton 1957) or exogenous chemicals (Hayes et al. 2002).

20

Parasites that induce hermaphroditism also cause sterility of their host reproductive

structures as a consequence of the parasite’s own reproduction thereby limiting the ability of the host to sexually reproduce and evolve resistance mechanisms (Clay 1991). The sterilization of host gonads by a parasite is known as parasitic castration (see reviews by

Fisher and Holton 1957; Clay 1991; Boudoin 1975; Shukalyuk et al. 2005) and thus, T. buchloëana induced ovary development in male buffalograss is an example of induced hermaphroditism and represents a form of parasitic castration.

The most widely known example of parasitically induced hermaphroditism in plants occurs in the host white campion (Silene latifolia Poir. ssp. alba (P. Mill.) Greuter

& Burdet) infected by the anther smut fungus, Microbotryum violaceum (Pers) G. Deml

& Oberw, (Class: Uredinomycetes) (Deml and Oberw 1982). Anther smut induces the development of male sex organs (stamens) in female plants of white campion and sporulates only within the anthers of infected florets (Uchida et al. 2003). Hence, the male sex organ inducing ability of anther smut is in sharp contrast to the female sex organ

(pistil) inducing ability of T. buchloëana; both examples illustrate induced hermaphroditism but for the opposite sexes. Due to its pistil inducing effect, we will refer to T. buchloëana as pistil smut.

Despite the remarkable sex-altering ability of pistil smut, surprisingly limited knowledge exists regarding its basic biology. Kellerman and Swingle (1889) observed that, unlike other smut fungi, ‘the monstrosity of T. buchloëana consists solely in its ability to produce ovaries in male plants’. However, Kellerman and Swingle (1889) were unable to detect the effects of pistil smut infection on female plants of buffalograss and

their attempts, as well as those of Norton (1896), to germinate teliospores in vitro failed.

21

Duran (1987) observed both male and female sex forms of buffalograss infected with

pistil smut and showed teliospore up to the stage of conjugation, however, he did not describe the effects of infection on either host sex form or the later stages of pistil smut life cycle.

To our knowledge, pistil smut is the only species within order Tilletiales known to induce hermphroditism in its host, however, its evolutionary relationship within the order

is unknown. Species within Tilletiales are commonly identified by the color and surface

ornamentation of their teliospores which, for most species, are produced within host

ovaries. Castlebury et al. 2005 performed the most comprehensive study to date

regarding Tilletiales phylogenetics and showed that spore morphology, germination

pattern, infection type, host sub-family, and large subunit nuclear (nLSU) rDNA

sequences were useful characteristics for determining the lineage structure within order

Tilletiales. In an effort to make the sex altering effects of pistil smut more widely known for study, and to establish its evolutionary relatedness among species of order Tilletiales, we conducted a study on the life cycle and phylogeny of pistil smut. Here we show the in vitro and in vivo life cycle of pistil smut, that pistil smut induces hermaphroditism not only in male sex forms of buffalograss but also in female sex forms, and that pistil smut has an evolutionary phylogeny unlike any other species of order Tilletiales.

22

MATERIALS AND METHODS

Isolation and maintenance of pistil smut.— The pistil smut fungus used in this study was a single source originally collected from an individual male buffalograss plant growing in a short grass prairie located in Kingfisher County, Oklahoma in 1984 (Huff et al. 1987). Five to eight mm long explant tissues including leaves, nodes, and stem segments from infected plants were cultured on PDA (potato dextrose agar) (Difco,

Detroit, Michigan) media at room temperature (20–25° C) in the dark. Fungal growth was observed from the cut-ends of the explants after about a month of culture which eventually gave rise to mycelia and secondary sporidia. Fungal mycelia were maintained through sub-culturing every three months under axenic conditions. Teliospore filled host ovaries (smut balls) were collected from infected plants and surface sterilized in 70% ethanol for 1 min followed by 0.26% NaClO (5% v/v commercial bleach) for 3 min before soaking in distilled water for 2 days. Teliospore germination was attempted on

2% agar at 5° C, 25° C and 37° C under light and dark aseptic conditions.

Host infection by pistil smut.— Sixty genotypes of Mexican diploid (2n=20)

(Huff et al. 1993) buffalograss were grown in plastic pots (15.25 cm diameter) containing potting soil (Promix®, Premier Horticulture, Inc., Quakertown, Pennsylvania) in the greenhouse (26° C day/21° C night) under natural day length conditions. Each genotype was vegetatively propagated, at the 4 to 6 tiller stage, into four clonal replicate plants, two of which were left uninfected and two of which were inoculated by embedding teliospores into the soil surface close to the base of vegetative shoots. Inoculated plants were saturated with water and kept sealed in clear plastic bags in order to maintain high

23 humidity conditions, for approximately 6 weeks to allow teliospores to germinate and the fungus to enter the plant.

Scoring and analysis of pistil smut morphology.— Six morphological and life cycle characteristics documented for pistil smut were scored as qualitative parameters, namely, teliospore surface ornamentation, number of primary produced, germination pattern, infection type, host subfamily, and ability to induce hermaphroditism. These same characteristics were obtained from Castlebury et al.

(2005) for 39 Tilletia and Tilletia-like species, and Erratomyces patelli. Scoring of these qualitative characters were based on the following criteria: teliospore ornamentation (0 for reticulate, 1 for tuberculate/verrucose, 2 for ridged, and 3 for foveolate), number of primary basidiospores produced (0 for ≤30, and 1 for >30), germination pattern (0 for non-conjugating, and 1 for conjugating), infection type (0 for systemic, and 1 for local), host subfamily (1 for Pooideae, 2 for Panicoideae, 3 for , 4 for

Ehrhartoideae, 5 for Arundinoideae subfamilies of the grass family Poaceae, and 6 for

Fabaecea family) and ability to induce hermaphroditism (0 for not inducing hermaphroditism, and 1 for inducing hermaphroditism). PAUP 4.10b (Swafford 1998) was used to generate a genetic distance matrix based on simple matching of these qualitative scores. The minimum evolution (ME) distance-based phylogenetic tree obtained was viewed and labeled using MEGA 3.1 (Kumar et al. 2004).

Nucleic acid extraction and PCR amplification.— Actively growing mycelia was scrapped from PDA followed by vigorous maceration using liquid nitrogen. DNA was extracted using CTAB method (Doyle and Doyle 1990). The nuclear LSU (nLSU) and the entire ITS region including ITS1-5.8S-ITS2 of rRNA gene were amplified in 25 µL

24

reactions with the following reaction conditions: 10–15 ng of genomic DNA, 1 X buffer,

2 mM MgCl2, 0.25 mM dNTPs, 2.5 units Taq-polymerase, 10 µM each of primers LR0R

and LR7 for nLSU region, and ITS1 and ITS4 for the entire ITS region of rRNA gene

(Vilgalys and Hester 1990, Rehner and Samuels 1994). The thermal cycler program was:

initial denaturation 95° C for 10 min, followed by 35 cycles of 94° C for 30 sec, 55° C for

30 sec, 72° C for 1 min with a final extension 72° C for 10 min. The resulting amplicons,

containing either a portion of nLSU or ITS1-5.8S-ITS2, were gel-purified using

QIAquick Gel Extraction Kit (Qiagen Science, Inc., Germantown, Maryland) and were

cloned in pCR®II-TOPO® (Invitrogen, Carlsbad, California). Cloned rDNA fragments

were sequenced with the ABI Hitachi 3730XL DNA Analyzer (Applied Biosystems,

Foster City, California). Four clones of 1292bp region of nLSU with accession numbers

DQ659921–DQ659924, and two clones of 684bp long entire ITS region with accession

numbers EF204935–EF204936 were submitted to the GenBank.

Taxa examined for phylogenetic analyses.— BLASTn (Basic Local Alignment

Search Tool) program was used to extract species from the national public database,

NCBI (http://www.ncbi.nih.gov/), that were found to be related to pistil smut based on

rDNA sequence homology. The resulting data set for conducting nLSU phylogenetic

analyses of order Tilletiales included four clones derived from pistil smut along with 49

additional taxa comprising 33 Tilletia species, four Tilletia-like species from three allied

genera proposed to be synonyms of Tilletia (namely, Ingoldiomyces, , and

Conidiosporomyces; Castlebury et al. 2005), one basal group to genus Tilletia

(Erratomyces patelii), and two distantly related outgroup species Exobasidium

rhododendri (order Exobasidiales) and Ustilago tritici (order Ustilaginales) (TABLE I).

25

The data set for ITS phylogenetic analyses included two clones from pistil smut along with 19 Tilletia species, 2 Tilletia-like species (I. hyalosporus, N. iowensis), and E. patelii (TABLE I).

Taxa sequence alignments.— Multiple sequence alignment was conducted using

CLUSTALW program of MEGA version 3.1 (Kumar et al. 2004) using parameters that

allowed optimization of gaps including gap opening (GO) penalty of 10, gap extension

(GE) penalty of 4.44 (44% of GO penalty), and transition weight of 0.9 (Terry and

Whiting 2005). For nLSU analyses with the inclusion of outgroup taxa, nucleotide sequences were aligned for 1456 positions. Missing data or regions with ambiguous

alignment were removed from the analyses (253 positions). Of the remaining 1203

characters, 942 were conserved, 261 were variable with 122 singleton sites and 139

parsimony informative sites. Transition/transversion ratio was 2.4 with 25 transitional

pairs and 11 transversional pairs. For nLSU analyses with the exclusion of outgroup

taxa, nucleotide sequences were aligned for 1340 positions. Missing data or regions with

ambiguous alignment were removed from the analyses (136 positions). Of the remaining

1204 characters, 1065 were conserved, 139 were variable with 43 singleton sites and 96

parsimony informative sites. Transition/transversion ratio was 2.6 with 20 transitional

pairs and 8 transversional pairs. For ITS phylogenetic analyses, nucleotide sequences

were aligned across taxa for 734 positions. Missing data or regions with ambiguous

alignment were removed from the analyses (204 positions). Of the remaining 530

characters, 233 were conserved, 297 were variable with 31 singleton sites and 266

parsimony informative sites. Transition/transversion ratio was 0.7 with 54 transitional

26

pairs and 74 transversional pairs. All sequence alignments were deposited in TreeBase

(SN3366-14956–SN3366-14960).

Phylogenetic analyses.— Modeltest 3.7 (Posada and Crandall 1998) with the

Akaike Information Criterion (AIC) was used to estimate the optimal model of evolution for conducting phylogenetic analyses of both nLSU and ITS regions with and without

inclusion of outgroup taxa, and also for determining the parameters for likelihood

assumptions. For the nLSU region, the General Time Reversible (GTR) model with a

gamma shape parameter and a proportion of invariable sites (referred to as GTR+I+G)

fitted the data set the best. For nLSU sequence analyses with inclusion of outgroup taxa,

the estimated base composition frequencies were A, 0.2729; C, 0.1957; G, 0.2930; T,

0.2385; the gamma shape parameter was 0.6218; the proportion of invariable sites was

0.6043, and the substitution rate matrix was A-C, 0.5163; A-G, 3.0595; A-T, 0.9471; C-

G, 0.1098; C-T, 5.8501; G-T, 1.0000. For nLSU sequence analyses with exclusion of

outgroup taxa, the estimated base composition frequencies were A, 0.2732; C, 0.1852; G,

0.2847; T, 0.2569; the gamma shape parameter was 0.5417; the proportion of invariable

sites was 0.8014, and the substitution rate matrix was A-C, 1.0000; A-G, 4.8100; A-T,

1.0000; C-G, 1.0000; C-T, 9.5103; G-T, 1.0000. For the ITS region, Modeltest 3.7

predicted that the TVM+I+G (Transversion model) was the best fitting model however,

the GTR+I+G model of evolution was not significantly different from TVM+I+G (delta =

0.6978). In order for us to perform comparative Mantel correlations between nLSU and

ITS regions, we elected to use the GTR+I+G model for phylogenetic analyses of ITS

region. For ITS sequence region, the estimated base composition frequencies were A,

0.2434; C, 0.2306; G, 0.2144; T, 0.3115; the gamma shape parameter was 1.2897; the

27

proportion of invariable sites was 0.3732, and the substitution rate matrix was A-C,

1.0829; A-G, 2.2476; A-T, 0.9318; C-G, 0.4594; C-T, 2.2476; G-T, 1.0000.

The estimated optimal model of evolution was utilized to construct phylogenetic

trees, maximum parsimony (MP) and maximum likelihood (ML), by PAUP 4.01b. Mega

3.1 was used to view and label trees generated from PAUP 4.10b. All MP and ML trees

were inferred using the heuristic search option with the random addition of trees and the

branch swapping (Tree Bisection-Reconnection, TBR) options of PAUP 4.10b. For MP

analyses, 10000 replications and for ML analyses 20 replications were run with

automatically increasing the Maxtree (maximum numbers of trees stored) by 100. From

all of the trees generated, the most parsimonious tree (with minimum number of steps)

and best ML tree (highest log likelihood value, –lnL) were filtered along with branch

lengths which are indicative of the number of base substitutions site-1 (ML) or sequence-1

(MP). To evaluate branch robustness in MP and ML trees, bootstrap (Felsentein 1985) analyses were conducted. Bootstrap support was estimated using 1000 bootstrap replicates, each replicate consisting of 10 heuristic searches and random addition sequences with no branch swapping and multree option off.

GTR based genetic distances for both nLSU and ITS sequences were generated using PAUP 4.10b and analyzed as follows using NTSYS version 2.2 (Rohlf 2005). The obtained genetic distance matrices were transformed to scalar product form using subprogram DCEN in order to compute eigenvalues and eigenvectors (Gower 1966).

The first three eigenvectors were plotted as principal coordinate axes representing the 3- dimensions of principle coordinate analysis (PCoA). Cophenetic goodness-of-fit tests were performed for each PCoA. In addition, Mantel matrix correlation tests were

28

performed, using NTSYS version 2.2, between morphological and nLSU distance

matrices and between nLSU and ITS distance matrices to determine their level of

correspondence.

GZ-GAMA (Gu and Zhang 1997) was used to estimate the expected number of

substitutions (k) of each nucleotide site for both nLSU and ITS sequences. This program

uses the nucleotide sequences without any gaps and ambiguous sites along with the tree topology in PHYLIP format in order to estimate k by using maximum likelihood approach under the Jukes Cantor model. The profile of rate variability with sites was generated graphically by plotting k against the position of sites for both nLSU and ITS region using Microsoft Excel.

RESULTS

Induced hermaphroditism and parasitic castration of buffalograss by pistil smut.— Pistil smut infection of buffalograss plants under controlled conditions exhibited an 93% rate of infection with 112 plants (both replicates of 33 male and

23 female genotypes) becoming infected out of the 120 plants inoculated with teliospores. The most conspicuous manifestation of pistil smut infection is the presence of purple feathery stigma in florets of infected male plants, indicating the presence of ovaries, as compared to the bright yellow anthers normally produced in non-infected male plants (FIG. 2.1: infected left; non-infected right). Infection of

female plants is visibly less noticeable because female florets normally display

purple feathery stigmas and produce ovaries which are permanently enclosed within

29

thickened outer glumes forming burr-like seed capsules obscuring the effects of infection on floral structures (FIG. 2.2: infected left; non-infected right). One

noticeable symptom of infection in female plants is the increased number of florets

per spikelet (two florets per spikelet in infected versus one floret per spikelet in

non-infected) (FIG. 2.4: infected left; non-infected right). Similarly, infected males

also show an increased number of florets per spikelet (three florets per spikelet in

infected versus two florets per spikelet in non-infected; FIG. 2.3: infected left; non-

infected right). These secondary effects of pistil smut infection on meristem

determinacy produce visual crowding of both male and female spikes.

Florets of infected male plants contain a fully developed pistil (stigma, style,

and ovary) and stamens that are reduced in size (FIG. 2.5) compared to the vestigial

pistil and fully developed and functional (pollen producing) stamens of non-infected male plants (FIG. 2.6). Florets of infected female plants retain a fully developed pistil but possess enlarged stamens (FIG. 2.7) compared to the vestigial stamens of

non-infected females (FIG. 2.8). Comparisons between florets of infected male and infected female plants suggest that little, if any, differences exist in their floret structure and composition (FIG. 2.5 vs. FIG. 2.7). Hence, pistil smut infection

renders the unisexual florets of both male and female buffalograss morphologically

hermaphrodite containing both sex organs within the same floret. Overall, we observed that 95% of the florets of infected male plants (827 out of 870) and all of

the florets of female plants (167 out of 167) examined exhibited these symptoms of

induced hermaphroditism. We therefore conclude that pistil smut induces

hermaphroditism in both male and female sex forms of buffalograss.

30

We found the sex organs within these induced hermaphroditic florets to be reproductively sterile as a result of infection, or in other words, parasitically castrated. Ovaries are sterile because they become supplanted with teliospores, forming smut balls (FIG. 2.5, 2.7, 2.9) that are incapable of setting viable seed.

However, anthers within induced hermaphroditic florets show no signs of teliospore

production but are nonetheless sterile because they are underdeveloped in size

(structure) and maturity (function). Thus, even though anthers of infected male

plants are reduced in size while those of infected female plants are hypertrophied,

both are similarly underdeveloped in appearance by being typically thin, white (FIG.

2.5 and 2.7) and attached to filaments which do not exert out of the floret.

Therefore, stamens of infected plants are incapable of producing and liberating

pollen. Hence, we conclude that both male and female reproductive organs within

induced hermaphroditic florets are parasitically castrated but apparently for

different sex specific reasons.

Morphological characteristics of pistil smut fungus in vitro.— The ability to

grow pistil smut under aseptic conditions and to infect young plants of buffalograss under

controlled conditions, enabled us to perform Koch’s postulates verifying pistil smut as

the causal agent of induced hermaphroditism in buffalograss. Mature ovaries (induced or

not) of florets of infected plants (male or female) become filled with fungal teliospores

(FIG. 2.9). Teliospore masses emit a powerful decaying fish odor and thus pistil smut

belongs to a group of smuts known as the stinking smuts which produce the odorous

compound trimethylamine (Hanna et al. 1932). Teliospores of pistil smut exhibit obscure

tuberculate ornamentation (FIG. 2.9, 2.10), generally range from pale yellow to light-

31

chocolate brown in color, mostly spherical in shape and are enclosed within a hyaline sheath (FIG. 2.11). Teliospore germination was only observed under room temperature

conditions (25° C) and showed a higher rate of germination under light versus dark

conditions (1.5% ± 1.17 light vs 0.7% ± 0.24 dark). Pistil smut completed its life cycle

within approximately three months following a sequence of developmental stages (FIG.

2.12–2.17). Germinating teliospores produce a promycelium (basidium) bearing whorls of finger-like, mostly monokaryotic but rarely dikaryotic (Duran 1987), primary basidiospores (FIG. 2.12). Individual pistil smut teliospores typically produce > 30 primary basidiospores which mostly conjugate at the base but sometimes conjugate to form an “H-bridge” structure (FIG. 2.13). The conjugated basidiospores produce fungal

hyphae and initiate secondary basidiospore production. Pistil smut was primarily observed to produce a blastoform-type of dikaryotic secondary basidiospores which are usually fusiform in shape (FIG. 2.14). Consistency of cultures at this stage is

predominantly sporidial giving a “yeast-like” appearance including spike-like structures

resembling “suchfaden” (Fisher and Holton 1957) (FIG. 2.15). As the culture ages, its

consistency becomes more mycelial (FIG. 2.16). Finally, as the mycelium proliferates,

teliospores are produced from the forward portion of the mycelium (FIG. 2.17),

completing the life cycle of pistil smut in vitro.

Life cycle of pistil smut in vivo.— Pistil smut is a soil-borne fungus. Ruptured smut balls release teliospores into the soil which, upon germination, produce dikaryotic secondary basidiospores, presumably by following the same sequence of developmental stages as in case of in vitro life cycle. Secondary basidiospores give rise to the infection hyphae that enable the fungus to enter the plant. Buffalograss plants infected with soil-

32

borne inoculum represent an infectious horizontal mode of disease transmission. Once

inside the plant, the fungal mycelium grows intercellularly producing a systemic infection

(FIG. 2.18) within the basal meristems (crowns) of vegetative shoots. Any new plant growth arising from these infected basal meristems will contain fungal mycelium through a vertical mode of disease transmission by mycelial propagation from parent to daughter shoots (FIG. 2.19). When infected plants begin to initiate a flowering apex, the fungus

infects the developing inflorescence and induces hermaphroditism in the florets of male

and female buffalograss. Pistil smut causes parasitic castration as it limits host sexual reproduction, however, pistil smut does not kill the host plant under greenhouse

conditions but rather co-exists with its host by parasitizing the perennial production of

vegetative shoot meristems.

Phylogenetic analysis of pistil smut based on morphology and life cycle

characteristics.— Pistil smut’s morphological and life cycle characteristics were

compared with 40 other species within the order Tilletiales which have previously been described by Castlebury et al. 2005 including 39 Tilletia and Tilletia-like species along

with the basal group E. patelii. The six qualitative characteristics of pistil smut comprise

of teliospores that exhibit tuberculate surface ornamentation, which upon germination

give rise to >30 primary basidiospores that conjugate to produce systemic infection

causing induced hermaphroditism in its Chloridoideae host (FIG. 2.20). This

combination of qualitative characteristics is unique for pistil smut as no other species of

order Tilletiales used in our analysis exhibit the same set of morphological

characteristics. A similar topology was obtained without (tree not shown) the inclusion

of the sixth character, induced hermaphroditism. Cluster analysis identified 17 unique

33

morphological categories of which 10 categories contain only a single species including the unique category for pistil smut (Cat 1) (FIG. 2.20). All other remaining categories

contain two or more species. Pistil smut shares a maximum of three of the six characters

with T. boutelouae and T. savilei (Cat 4), T. eremopoae (Cat 14), T. asperfolia (Cat 13),

T. rugispora (Cat 3), and the basal group, E. patelii (Cat 2) (FIG. 2.20 bold). Pistil smut

shares tuberculate ornamentation, spore number and germination pattern with E. patelii

and T. rugispora but each of these species produce local infection and occur on a host

subfamily different than Chloridoideae. Pistil smut shares tuberculate ornamentation and

spore number with T. boutelouae and T. savilei which also infect Chloridoideae but these

species produce non-conjugating basidiospores and local infection. Lastly, pistil smut

shares three of the six qualitative characters with two species possessing reticulate spore

ornamentation, namely, T. asperfolia and T. eremopoae. At the opposite extreme, pistil

smut was found to share none of these six characteristics with N. iowensis (Cat 10) (FIG.

2.20). Pistil smut shared a low level of match (16–33%) with the remaining species used in our analysis. Given these differences, pistil smut exhibited the longest branch length of any of the species in our phylogenetic analysis. The next longest branch length observed belonged to N. iowensis which also shared a maximum of three of the six qualitative characters with other species but, on average, shared more characteristics with the other categories than did pistil smut. All other species in the analysis shared a maximum of five characters with at least one of the remaining species. Therefore, based on the phylogeny of these qualitative charateristics, we found that pistil smut is morphologically the most distant species within order Tilletiales. In order to confirm these results and determine a more accurate placement of pistil smut within order

34

Tilletiales, we conducted phylogenetic analyses based on nLSU and ITS regions of rDNA

sequences.

Phylogenetic placement of pistil smut based on nLSU and ITS sequence

analyses.— The nLSU sequences obtained from pistil smut were analyzed along with 47

taxa representing 37 of the Tilletia and Tilletia-like species along with the basal group E. patelii used in the morphological analysis. In the alignment of these sequences

(TreeBase: SN3366-14956), pistil smut displayed a cluster of gaps (base positions 406,

410–419, 423–436, 451) not present in any other taxa used in the analyses. Due to the size of these gaps and the uneven length of taxa sequences, all gaps were removed for subsequent analyses. In addition, the outgroup species Exo. rhododendri and U. tritici were included to provide phylogenetic polarity. Each nLSU phylogenetic analysis (MP,

ML, and genetic distance) displayed the same lineage structure of Tilletia and Tilletia- like species as originally described by Castlebury et al. (2005). This structure comprises

4 lineages, L-I through L-IV (see TABLE I) and four Tilletia species not fitting within any

lineage (unresolved) namely, T. ehrhartae, T. setariae, T. horrida and T. rugispora. We will refer to this structure as the Tilletia clade (FIG. 2.21–2.24 and FIG. 2.27–2.28; see ●).

The correspondence between our phylogenetic analyses and those of Castlebury et al.

(2005) allowed us to simply collapse each of the four lineages within the resulting

phylogenetic trees. In addition, we found that E. patelii acquired a basal position to the

Tilletia clade supporting Castlebury et al.’s (2005) contention of E. patelii as a basal

group to Tilletia.

MP analysis yielded 15 equally parsimonious trees of length 457 steps [(CI) =

0.6696, (RI) = 0.8409, (RC) = 0.5630]. Figure 2.21 displays one of these 15 most-

35

parsimonious trees. To our surprise, pistil smut was found to occupy a well supported

position (>95% boostrap support) outside of the Tilletia clade (FIG. 2.21). A prominent

feature of this representative MP tree is that pistil smut resides on a branch length (37

base substitutions sequence–1) that is remarkably longer than any other species within the

Tilletia clade. Moreover, pistil smut was slightly further away from the Tilletia clade (50

–1 –1 base substitutions sequence ) than E. patelii (48 base substitutions sequence ) (FIG.

2.21). Long branches are known to influence species placement within phylogenetic tree

space as a result of a phenomenon known as long branch attraction (LBA) (Bergsten

2005). External placement of pistil smut to the Tilletia clade could be real or an artifact

resulting from the association between the long branch of pistil smut and that of outgroup taxa. In order to examine the possible effects of LBA, we followed Bergsten (2005) who prescribed several tests and remedies for LBA. One approach followed was to reanalyze the data after removing the outgroup taxa, Exo. rhododendri and U. tritici. MP analysis

yielded 10 equally parsimonious trees of length 259 steps [(CI) = 0.5602, (RI) = 0.8855,

(RC) = 0.5607]. Figure 2.22 displays one of these 10 most-parsimonious trees. The

topological placement of pistil smut remained essentially the same, being external to the

Tilletia clade, after exclusion of outgroup taxa (FIG. 2.22) indicating little to no effect of

LBA.

Another approach utilized to assess any possible presence of LBA was to

reanalyze the nLSU data using the ML method of cluster analysis which is considered to

be less sensitive to the effects of LBA (Bergsten 2005). The best ML tree with inclusion

of outgroup taxa (–lnL = 4070.03249), yielded a similar topology as did the MP analysis

exhibiting a well supported (>95% bootstrap support) external placement of pistil smut to

36

the Tilletia clade but with the exception that pistil smut’s branch length from the Tilletia

clade (0.06305 substitutions site–1) was much longer than that of E. patelii (0.04496

–1 substitutions site ) (FIG. 2.23). However, after excluding outgroup taxa, the best ML

tree (–lnL = 3169.65783) exhibited an altered topology such that the placement of pistil

smut was internal to the Tilletia clade but resided on an even longer branch length

–1 (0.07951 substitutions site ) (FIG. 2.24). Thus, although ML analysis excluding

outgroup taxa would provide some evidence for the presence of LBA in our previous

phylogenetic analyses, the branch robustness of this best ML tree was undefined because

only a single tree was obtained that showed an internal placement of pistil smut and

therefore, no bootstrap supports could be generated for this single best ML tree (FIG.

2.24). The second most likely tree and one for which bootstrap support could be generated showed the external placement of pistil smut to the Tilletia clade (tree not shown). Nonetheless, the best ML tree showed that the taxonomic placement of pistil smut is best associated with species of L-IV within the Tilletia clade (FIG. 2.24).

However, significant differences were found in the evolutionary rates of base

substitutions between pistil smut and the three species within L-IV compared with E. patelii as the outgroup (Tajima’s relative rate test: p-values ranged from 0.06 to 0.01 across the three species). Thus, although pistil smut clustered with L-IV, it does not imply that pistil smut is closely related to L-IV.

In order to have a better visual representation of the taxonomic position of pistil smut with respect to species of order Tilletiales, we plotted the GTR genetic distance matrix 3-dimensionally, utilizing a principle coordinates analysis (PCoA; cophenetic

correlation coefficient goodness-of-fit test, r =0.99). PCoA of order Tilletiales with

37

inclusion of outgroup taxa shows pistil smut to be genetically more distant from the

Tilletia clade (mean =0.055879; range =0.048237–0.063641) than E. patelii (mean

=0.048175; range =0.039481–0.054768) (FIG. 2.25). Exclusion of outgroup taxa from

the analysis provides a magnified view of order Tilletiales but with reversed polarity

(FIG. 2.25 vs 2.26). This reversed polarity was manifested by pistil smut acquiring more

negative values along dimension-1 and more positive values along dimension-2 in

relation to the Tilletia clade. Pistil smut appears to be genetically isolated but least

distant to taxa belonging to L-IV (mean =0.051307; range =0.049109–0.052745) (FIG.

2.26) which may explain why, among all species of the Tilletia clade, pistil smut

clustered with L-IV in the ML analysis without outgroup taxa. However, pistil smut is

still twice the distance from L-IV than any two species within the Tilletia clade are from

one another (FIG. 2.26). Thus, the genetic distance method of analysis of nLSU shows a

similarly large separation between pistil smut and other species of Tilletia as does MP

and ML methods of analyses. We therefore conclude that pistil smut and E. patelii represent extremes of the phylogenetic tree space within order Tilletiales for nLSU and that the Tilletia clade occupies an intermediate position between these two extremes.

A Mantel matrix correlation test between morphological characteristics and

nLSU genetic distance matrices demonstrated that the two data sets were

significantly correlated (r = 0.40, P < 0.001). Based on the congruence between the

separate analyses of morphology and nLSU sequence, pistil smut would be closest

to T. rugispora (unresolved), and T. savilei (Lineage II) sharing 50% matching for

morphological characteristics, and displaying a relatively small genetic distance

(0.049 substitutions site–1). In order to add further support to our morphological

38

and nLSU sequence analyses and in an effort to enhance resolution among taxa

within order Tilletiales, we conducted a phylogenetic analysis on a more variable

DNA sequence, the ITS region.

ITS. We compared the entire ITS sequence of pistil smut with a subset of 21

Tilletia and Tilletia-like species along with the basal group E. patelii. All three analyses,

MP, ML, and PCoA genetic distance for ITS were found to be in close correspondence

with the nLSU phylogenetic analyses without outgroups. MP analysis yielded one best

parsimonious tree of length 918 steps [(CI) = 0.6144, (RI) = 0.7602, (RC) = 0.4670]

displaying an external placement of pistil smut with respect to the Tilletia clade (FIG.

2.27). The best ML tree (–lnL = 4299.22231) of ITS region displayed an internal placement within the Tilletia clade (FIG. 2.28). In both MP and ML analyses, pistil smut resides on the longest branch (109 substitutions sequence–1 for MP and 0.984

substitutions site–1 for ML). PCoA also shows that pistil smut is genetically more distant

from the Tilletia clade than E. patelii (FIG. 2.29). A Mantel matrix correlation test

between nLSU and ITS genetic distance matrices demonstrated that the two data sets

were significantly correlated (r=0.90, p<0.001).

Gamma distribution. The large genetic distance of pistil smut’s nLSU and ITS

sequences from other species within order Tilletiales may represent real evolutionary

differences or may arise as a result of these sequences being rouge i.e. ones that are not

the intended sequences. To assess the validity of our nLSU and ITS1-5.8S-ITS2

sequences, we constructed a distribution of estimated substitutions per site (k) using the

GZ-GAMA program. The distribution shows that pistil smut’s nLSU and ITS sequences

are variable at the same sites as other species used in the analyses and that pistil smut

39

adds more diversity to those sites than would be observed without its inclusion (FIG. 2.30 and 2.31). We therefore conclude that the pistil smut sequences analyzed for this study were not rogue sequences and that the genetic and evolutionary distances between pistil smut and the other species are genuine.

Overall, based on our analysis of morphology and rDNA sequences, we conclude that pistil smut is unable to be accommodated within the Tilletia clade and is as, or more, distant to the Tilletia clade than the basal group E. patelii. As such, we conclude that pistil smut is phylogenetically separate and distinct from all other genera within

Tilletiales.

DISCUSSION

In 1859, George Engelmann, the famous 19th century American physician-

botanist and a founding member of the US National Academy of Science,

recognized a rare form of buffalograss containing both male and female

inflorescences on the same plant (monoecious). His practiced judgment enabled him to discover that the dimorphic male and female sex forms of dioecious buffalograss, previously considered as separate species, were actually members of a

single species, which he placed in a monotypic genus, Buchloë (Engelmann 1859).

Thirty years thereafter, Kellerman and Swingle (1889), keenly observed that T.

buchloëana (pistil smut) infection restores the development of ovaries in male

plants of buffalograss. In the current study, we show that pistil smut not only

induces hermaphroditism in males but also in female plants of buffalograss. Sex

40

expression of buffalograss has been shown to be environmentally stable and

genetically controlled (Huff and Wu 1987, 1992) yet pistil smut is capable of

altering the sex expression of buffalograss in some unknown fundamental way.

Perhaps pistil smut infection perturbs the host’s sex determining mechanism thereby

stimulating the development of a sex which the plant is genetically programmed not

to exhibit. Smuts are an important group of disease causing agents due to their

devastating impact on food grain crops worldwide. However, increasing our

knowledge at the molecular, biochemical, and evolutionary levels regarding the

secondary effects of pistil smut infection, including its ability to induce ovaries and additional florets, would ultimately enhance our ability to increase seed production

of agronomically important grasses.

Pistil smut also causes parasitic castration of both male and female

reproductive organs within induced hermaphroditic florets, but for different sex-

specific reasons. We speculate that this sex-specific pattern of parasitism likely

involves differential resource allocation between floral sex organs (see Charnov

1982). However, the mechanisms for induced hermaphroditism or parasitic

castration, for any organism, including the well studied example of anther smut

infected white campion, has yet to be discovered.

Hermaphroditism is considered to be the ancestral form of sex expression in

dioecious species like buffalograss as evidenced from the vestigial remains of

opposite sex organs within male and female florets (Charlesworth 2002). Thus,

pistil smut infection reverts the present day dioecious sex expression of buffalograss

back to a more primitive form, and in doing so, recovers a phenotype that existed

41 back in time in the form of an induced hermaphrodite. Hence, fungal induced hermaphrodites represent a retrospective phenotype, or simply, a retrophenotype of an ancient hermaphrodite. Fungal induced retrophenotypes raise interesting coevolutionary biology questions: For example, do such patterns of parasitism happen by chance or have parasites like pistil smut played a role in the evolution of separate sexes in their hosts? (Wilson 1982).

No species within order Tilletiales is known to induce hermaphroditism in their host, other than pistil smut. Tilletia is the predominant genus within the order and contains over 140 species that only infect hosts belonging to the grass family

Poaceae (Castlebury et al. 2005). Erratomyces is the only genus within Tilletiales known to infect a non-grass host (; Piepenbring and Baur 1997) and is considered to be the evolutionary basal group to Tilletia genus (Castlebury et al.

2005). The evolutionary relationship of pistil smut to species of known genera within Tilletiales is unknown. Based on our phylogenetic study of rDNA sequences, we demonstrate that pistil smut exhibited the highest number of base substitutions, and as such, appears to be genetically isolated within order Tilletiales.

Additional support of pistil smut’s evolutionary isolation is provided from our morphological and life cycle characteristic analysis. Pistil smut shares the least number of morphological and life cycle characteristics with any other species within order Tilletiales. Based on this morphological and molecular evidence, we conclude that the phylogenetic placement of pistil smut lies outside of the present day taxonomic circumscription of genus Tilletia. Furthermore, with respect to the

Tilletia clade, pistil smut inhabits an opposite polarity within the phylogenetic tree

42

space than E. patelli. This suggests that pistil smut may have once been a species of

Tilletia but is not such anymore.

Pistil smut clearly enhances rDNA sequence diversity and morphological

variability within order Tilletiales. However, the 37 species of Tilletia and Tilletia-

like species examined in the present study, only represents approximately 1/4 of the

known species of Tilletia. It would be interesting to know whether the phylogenetic

isolation of pistil smut with respect to other Tilletia species could possibly be

resolved with better sampling of taxa and/or genes (Rokas and Carroll 2005) across

order Tilletiales; or, does it provide the true evolutionary picture. A previous

phylogenetic analysis of Tilletiales by Castlebury et al. (2005) concluded that four

species from three allied genera, namely, ayresii, C.

verruculosus, Ingoldiomyces hyalosporus, and Neovossia iowensis should be

considered synonyms of the genus Tilletia. These same species were included in

the present study as Tilletia-like taxa and tightly clustered with Tilletia species

within the Tilletia clade. As such, the Tilletia clade of the present study represents

a wide range of morphological and molecular diversity. Therefore, the degree to

which pistil smut lies outside of the Tilletia clade suggests that increased taxa

sampling would less likely be able to accommodate pistil smut within the genus

Tilletia. Moreover, the taxa sampling necessary to include pistil smut within the

genus Tilletia would likely engulf E. patelii as well. Nevertheless, the generic level

character distinguishing Erratomyces is host family, while that of pistil smut is its

ability to induce hermaphroditism; both characters can be viewed as pivotal in

43 either’s evolutionarily divergence. Thus, the origin of pistil smut’s genetic isolation likely resides in its ability to induce hermaphroditism in its host.

44

BIBLIOGRAPHY

Bergsten J. 2005. A review of long-branch attraction. Cladistics 21(2):163–193.

Boudoin M. 1975. Host castration as a parasitic strategy. Evolution 29(2):335–352.

Castlebury LA, Carries LM, Vanky K. 2005. Phylogeneic analysis of Tilletia and allied

genera in order Tilletiales (Ustilaginomycetes; Exobasidiomycetidae) based on large

subunit nuclear rDNA sequences. Mycologia 97(4):888–900.

Charlesworth D. 2002. Plant sex determination and sex chromosomes. Heredity 88:94–

101.

Charnov EL. 1982. The Theory of Sex Allocation. Princeton University press, New

Jersey.

Clay K. 1991. Parasitic castration of plants by fungi. Trends Ecol Evol 6:162–166.

Deml G, Oberwinkler F. 1982. Studies in Heterobasidiomycetes. Part 24. On Ustilago

violacea (Pers.) Rouss. from Saponaria officinalis L. Phytopath. Zeitschr. 104:345–

356.

Doyle JJ, Doyle JL. 1990. Isolation of plant DNA from fresh tissue. Focus 12:13–15.

Duran R. 1987. Ustilaginales of Mexico: , Symptomatology, Spore

Germination, and Basidial Cytology. Washington State Univ. Pullman, Washington.

Engelmann G. 1859. Two new dioecious grasses of the United States. Trans St Louis

Acad Sci 1:431–443.

45

Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap.

Evolution 6:227–242.

Fischer GW, Holton CS. 1957. Biology and Control of the Smut Fungi. Ronald Press,

New York.

Gower JC. 1966. Some distance properties of latent root and vector methods used in

multivariate analysis. Biometrika 53:325–338.

Gu X, Zhang JJ. 1997. A simple method for estimating the parameter of substitution rate

variation among sites. Mol Biol Evol 14:1106–1113.

Hanna WF, Vickery HB, Pucher GW. 1932. The isolation of trimethylamine from spores

of Tilletia levis, the stinking smut of wheat. J Biol Chem 83 (1):351–357.

Hayes TB, Collins A, Lee M, Mendoza M, Noriega N, Stuart AA, Vonk A. 2002.

Hermaphroditic, demasculinized frogs after exposure to the herbicide atrazine at low

ecologically relevant doses. Proc Natl Acad Sci USA 99:5476–5480.

Huff DR, Peakall R, Smouse PE. 1993. RAPD variation within and among natural

populations of outcrossing buffalograss. Theor Appl Genet 86:927 934.

Huff DR, Wu L. 1987. Sex expression in buffalograss under different environments. Crop

Sci 27:623–626.

Huff, DR, Wu L. 1992. Distribution and inheritance of inconstant sex forms in natural

populations of dioecious buffalograss (Buchloë dactyloids). Am J Botany 79: 207-215.

Kellerman WA, Swingle WT. 1889. New species of Kansas fungi. J of Mycology 5:11–

Kumar S, Tamura K, Nei M. 2004. MEGA3: Integrated software for Molecular

Evolutionary Genetics Analysis and sequence alignment. Briefings in Bioinformatics

5:150–163.

46

Norton JBS. 1896. A study of the Kansas Ustilagineae especially with regard to their

germination. Trans Acad Sci St. Louis 7:229–241.

Pipenbring M, Bauer R. 1997. Erratomyces, a new genus of Tilletiales with species on

leguminosae. Mycologia 89:924–936.

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

Bioinformatics 14(9):817–818.

Rehner S, Samuels GJ.1994. Taxonomy and phylogeny of Gliocladium analysed from

nuclear large subunit ribosomal DNA sequences. Mycol Res 98:625–634.

Rohlf FJ. 2005. NTSYS-pc: Numerical Taxonomy and Multivariate Analysis System,

ver. 2.2. Exeter Software, Setauket, NY.

Rokas A, Carroll SB. 2005. More genes or more taxa? The relative contribution of gene

number and taxon number to phylogenetic accuracy. Mol Biol Evol 22(5):1337–1344.

Seifriz W. 1953. Walter T. Swingle; 1871–1952. Science 118(3063):288–9.

Shukalyuk AI, Isaeva VV, Pushchin II, Dolganov SM. 2005. Effects of the Briarosaccus

callosus Infestation on the Commercial Golden King Crab Lithodes aequispina. J

Parasitology 91(2):1502–1504.

Swofford DL. 1998. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other

Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts.

Terry MD, Whiting MF. 2005. Comparison of two alignment techniques within a single

complex data set: POY versus Clustal. Cladistics 21:272–281.

Uchida W, Matsunaga S, Sugiyama R, Kazama Y, Kawano S. 2003. Morphological

development of anthers induced by the dimorphic smut fungus Microbotryum

47

violaceum in female flowers of the dioecious plant Silene latifolia. Planta 218:240–

248.

Vilgalys R, Hester M. 1990. Rapid genetic identification and mapping of enigmatically

amplified ribosomal DNA from several Crytococcus species. J Bacteriol 172:4238–

4246.

Wilson MF. 1983. Plant reproductive Ecology. John Wiley & Sons, New York.

48

2.1 2.2

3.7mm 3.7mm 2.3 2.4 Flowers Flowers 3

3mm 5mm Flowers 2.5 2.6 2.7 2.8

600µm 1.5mm 600µm 800µm

Figure 2.1 to Figure 2.8: Effects of pistil smut infection on male and female sex forms of buffalograss.

2.1, Spike of infected male plant exerting purple feathery stigma indicating the presence of ovaries (left) and spike of non-infected male plant with functional, pollen producing anthers protruding out from within the florets (right). 2.2, Spike of infected female plant

49

(left) and spike of non-infected female plant (right), both containing purple feathery

stigmas. 2.3, Spikelet of infected male plant with three florets (left) vs. spikelet of non-

infected male plant with two florets (right). 2.4, Spikelet of infected female plant with two florets (left) vs. spikelet of non-infected female plant with one floret (right). 2.5,

Floret of infected male plant with a fully developed pistil (stigma, style and an ovary) and

three underdeveloped stamens (anther and filament). 2.6, Floret of non-infected male

plant with three developed stamens. 2.7, Floret of infected female plant with a fully

developed pistil and three underdeveloped stamens. 2.8, Floret of non-infected female

plant with a fully developed pistil.

50

2.9 2.10 2.11

50µm 20µm

2.12 2.13 2.14

10µm 3µm 3µm 2.15 2.16 2.17

5mm 1cm 10µm

2.1818 2.19

50µm 6.35cm

Figure 2.9 to Figure 2.19: Stages of pistil smut life-cycle in vitro and in vivo.

51

2.9, Cross-section of a teliospore filled ovary (smut ball). 2.10, Teliospores of pistil smut exhibiting obscure tuberculate surface ornamentation. 2.11, Teliospore color ranges from pale yellow to light chocolate brown. 2.12, Germinating teliospore with promycelium bearing >30 primary basidiospores. 2.13, Conjugating primary basidiospores forming H- shaped bridge structure. 2.14, Blastospore-type secondary sporidia, fusiform in shape.

2.15, Young culture comprising primarily of secondary sporidia diplaying structures resembling “suchfaden”. 2.16, Old cultures displaying mycelial growth. 2.17, Teliospores produced from tips of mycelium growing in culture. 2.18, Intercellular growth of pistil smut hyphae within buffalograss vegetative meristem. 2.19, Vertical mode of disease transmission from parent to daughter tiller.

52

Telio Spore Germ Inf Host 2.20 ornam no. pattern type subfam IH

PistilT. buchloeana smut† Cat1 Tub >30 Conj Syst Chloridoid Yes 2.06 E. patelii Cat2 Tub >30 Conj Local Fabaceae No 0.65 0.48 T. rugispora Cat3 Tub >30 Conj Local Panicoid No 0.35 T. boutelouae Cat4 Tub >30 Nonconj Local Chloridoid No 0.47 T. savilei 0.27 0.21 T. obscura-reticulata]5 Ret >30 Nonconj Local Chloridoid No 0.53 T. indica Cat6 Tub >30 Nonconj Local Pooid No 0.36 T. walkeri 0.20 T. barclayana T. opaca 0.06 T. trachypogonis Cat7 0.41 T. setariae Tub >30 Nonconj Local Panicoid No 0.57 T. sumatii 0.20 T. whitechloae T. horrida Cat8 Tub >30 Nonconj Local Ehrhartoid No 0.59 C. ayresii C. verruculosus 0.55 T. chinoachnes Cat9 Tub <30 Nonconj Local Pancoid No 0.57 T. ixophori T. kimberleyensis 0.11 T. vittata Fov <30 Nonconj Local Arundinoid No N. iowensis Cat10 1.26 T. ehrhartae Cat11 0.74 Tub <30 Nonconj Syst Ehrhartoid No T. bromi T. fusca T. laevis

T. polypogonis

T. anthoxanthi T. controversa Cat12 0.50 Ret <30 Conj Syst Pooid No T. goloskokovii T. holci

T. menieri 0.08 T. tritici T. olida 0.38 T. asperfolia Cat13 Ret <30 Conj Syst Chloridoid No 0.50 Ret >30 Conj Syst Pooid No 0.41 T. eremopoae Cat14 0.51 T. cerebrina Cat15 Ret <30 Nonconj Syst Pooid No 0.20 T. sterilis I. hyalosporus Cat16 Ridg <30 Nonconj Syst Pooid No 0.80 0.74 T. aegopogonis Cat17 Ret <30 Nonconj Syst Chloridoid No 0.60 T. lycuroides

0.5

† Tilletia buchloeana

Figure 2.20: A Minimum Evolution (ME) distance-based phylogenetic tree generated by

simple matching of six morphological and life cycle characteristics of pistil smut along

with 40 other species of order Tilletiales.

53

Qualitative characters are labeled as follows: teliospore ornamentation: Ret = reticulate spores, Tub = tuberculate/verrucose spores, Rigd = ridged, Fov = foveolate; number of primary basidiospores: ≤30 = less than or equal to 30, >30 = greater than 30; germination pattern: Conj = conjugating primary basidiospores, Nonconj = non-conjugating primary basidiospores; infection type: Local = local infecting, Syst = systemic infecting; host subfamily: as indicated; ability to induce hermaphroditism: as indicated. Species grouped into 17 unique categories of qualitative characters (Cat1–17). Qualitative characteristics shared with pistil smut are indicated in boldface. Numbers below branches indicate branch lengths.

54

2.21 2.22 U. tritici (Ustilaginales) 64.2 E. rhododendri (Exobasidiales) 62.0 E. patelli E. patelii 9.0 24.8 T. buchloeana T. buchloeana 27.8 T. buchloeana 100 T. buchloeana 100 100 Pistil smut Pistil smut 37.0 T. buchloeana 35.0 T. buchloeana 42.0 T. buchloeana T. buchloeana 98 Lineage IV (3 Tilletia sp./3 taxa) Lineage IV (3 Tilletia sp./3 taxa) 12.0 10.2 9.0 96 T. rugispora T. ehrhartae 26.0 9.0 T. rugispora 75 T. setariae Tilletiales T. horrida 12.0 58 Lineage II (10 Tilletia sp./12 taxa) T. horrida 71 72 Lineage III (3 Tilletia sp./4 taxa) 13.0 Lineage III (3 Tilletia sp./4 taxa) T. horrida T. setariae T. horrida T. ehrhartae 9.0 76 Lineage II (10 Tilletia sp./12 taxa) Lineage I (17 Tilletia sp./22 taxa)

Lineage I (17 Tilletia sp./22 taxa) T. rugispora 11.0 T. rugispora

10 substitutions/ sequence 5 substitutions/ sequence

2.23 2.24 U. tritici (Ustilaginales) 0.18 E. rhododendri (Exobasidiales) 0.09 E. patelli E. patelii 0.04 T. buchloeana 0.04 T. horrida 100 T. buchloeana T. horrida 100 Pistil smut 0.05 T. buchloeana T. setariae 0.05 T. buchloeana T. ehrhartae T. horrida 96 Lineage I (17 Tilletia sp./22 taxa) T. horrida 0.01 0.02 Lineage IV (3 Tilletia sp./3 taxa) 0.02 Lineage II (10 Tilletia sp./12 taxa) Tilletiales Lineage III (3 Tilletia sp./4 taxa) Lineage III (3 Tilletia sp./4 taxa) T. rugispora 74 Lineage IV (3 Tilletia sp./3 taxa) T. rugispora T. buchloeana T. setariae T. buchloeana T. ehrhartae Pistil smut 0.08 T. buchloeana Lineage II (10 Tilletia sp./12 taxa) T. buchloeana Lineage I (17 Tilletia sp./22 taxa) T. rugispora T. rugispora

0.01 Substitution/site 0.005 substitutions/site

2.25 2.26

E. patelii E. patelli U. tritici Pistil smut I. hyalosporus 0.03 0.02 Tilletia clade 0.00 Dim-1 (59.71%) 0.01 0.00 0.01 Dim-1(45.10%) -0.03 Dim-3 -0.00-0.00 -0.02 0.00 Dim-3 -0.03 Dim-3 -0.02 -0.03 -0.03 -0.05-0.00-0.00 0.03 -0.01 -0.05-0.05 -0.09 -0.12-0.12 -0.05-0.05 (17.07%) -0.05 (9.63%) -0.01 -0.09 -0.01-0.03-0.01-0.03 -0.08-0.08-0.04-0.04

-0.03-0.03 -0.02-0.02

-0.01-0.01 Dim-2 (23.24%) -0.01-0.01 Dim-2 0.00 (16.21%) 0.00 0.01 Pistil smut Tilletia clade 0.01 E. rhododendri N. iowensis L-IV (Condiosporomyces spp.)

55

Figure 2.21 to Figure 2.26: Phylogenetic placement of pistil smut within order Tilletiales

based on 28S nLSU.

Order Tilletiales is represented by 48 taxa comprising 37 Tilletia and Tilletia-like species

(Tilletia clade) and Erratomyces patelii. Outgroup taxa are from order Exobasidiales (E.

rhododendri), and order Ustilaginales (U. tritici). The Tilletia clade (●) comprises 4 lineages, L-I through L-IV and four species not fitting within any of the four lineages namely, T. ehrhartae, T. setariae, T. horrida and T. rugispora. Each of the four lineages of the Tilletia clade are collapsed into solid triangles (area of the triangle represent the number of taxa contained in that lineage). Numbers below tree branches indicate branch lengths (number of base substitutions per site or sequence). Numbers above branches indicate bootstrap supports. Thickend branches indicate >95% bootstrap support for the placement of pistil smut outside the Tilletia clade. 2.21, One of the 15 most-parsimonious trees of length 457 steps generated with the inclusion of outgroup taxa based on

GTR+I+G model of evolution. 2.22, One of the 10 most-parsimonious trees of length 259 steps generated with the exclusion of outgroup taxa based on GTR+I+G model of evolution. 2.23, Single best ML tree (–lnL = 4070.03249) generated with the inclusion of

outgroup taxa. 2.24, Single best ML tree (–lnL = 3169.65783) generated with the

exclusion of outgroup taxa. 2.25, GTR model of genetic distance with the inclusion of

outgroup taxa represented 3-dimensionally using principle coordinate analysis (PCoA).

2.26, GTR model of genetic distance with the exclusion of outgroup taxa represented 3-

dimensionally using PCoA.

56

E. patelli 73.3 2.27 T. buchloeana 100 Pistil smut 109.0 T. buchloeana 99 18.7 Lineage II (5 Tilletia sp./5 taxa) 55 40.0 94 Lineage I (13 Tilletia sp./13 taxa) 57.0 86 51.0 T. barclayana (Lineage II) 46.0 91 33.0 32.0 50 T. ehrhartae 45.0 18.0 T. horrida 43.0

20 substitutions/ sequence

2.28 E. patelii 0.56 Lineage II (5 Tilletia sp./5 taxa) Lineage I (13 Tilletia sp./13 taxa) 0.23 T. horrida 0.22 T. ehrhartae 0.10 0.11 T. barclayana (Lineage II) 0.10 0.20 T. buchloeana Pistil smut 0.98 T. buchloeana

0.1 substitutions/site

N. iowensis 2.29 0.18 I. hyalosporus 0.07 Dim-1 (43.04%) -0.04-0.04 Dim-3 0.16 0.05 -0.05-0.05 -0.16-0.16 -0.27-0.27 -0.15-0.15 (11.71%) -0.26-0.26-0.13-0.13

E. patelii -0.00-0.00

0.12 Dim-2 (25.30%)

0.25 Pistil smut

0.38

Figure 2.27 to Figure 2.29: Phylogenetic placement of pistil smut within order Tilletiales based on the entire ITS region including ITS1-5.8S-ITS2 DNA sequences.

57

Order Tilletiales is represented by 21 Tilletia and Tilletia-like species (Tilletia clade, ●) and Erratomyces patelii. Lineages of the Tilletia clade are collapsed into solid triangles

(area of the triangle represent the number of species contained in that lineage). Numbers below tree branches indicate branch lengths (number of base substitutions per site or sequence). Numbers above branches indicate bootstrap supports. 2.27, Single most-

parsimonious tree of length 918 generated based on GTR+I+G model of evolution. 2.28,

Single best ML tree (–lnL = 4299.22231) based on GTR+I+G model of evolution. 2.29,

GTR model of genetic distance represented 3-dimensionally using PCoA.

58

24 2.30

20

16

12

8 Expected number of substitutions (k) (k) substitutions of number Expected

4

0 1 51 101 151 201 251 301 351 401 451 501 551 601 651 701 751 801 851 901 951 1001 1051 1101 1151 1201 Sites

16 2.31

12

8 Expected number of substitutions (k) (k) substitutions of number Expected 4

0 1 51 101 151 201 251 301 351 401 451 501 Sites

ITS1 5.8S ITS2

59

Figure 2.30 to Figure 2.31: Graphical representation of the expected number of substitutions (k) of nucleotide sites plotted against the position of each site for species of

Tilletiales with pistil smut (blue) and without pistil smut (pink).

2.30, 28S nLSU region of rDNA and 2.31, The entire ITS region, including ITS1-5.8S-

ITS2.

60

Table 2.1: List of taxa and GenBank accession numbers for 28S nLSU and ITS regions of rDNA used in the phylogenetic analyses.

GenBank accession number

Taxon nLSU ITS

Lineage I Ingoldiomyces hyalospora AY818976 AF399891 Tilletia anthoxanthi AY819009 AF398458 T. bromi AY819001 AF398461 T. bromi AY818992 - T. cerebrina AY818994 AF310188 T. controversa AY818995 AF398453 T. eremopoae AY819016 - T. fusca AY818996 AF398455 T. fusca AY818997 - T. goloskokovii AY818998 DQ832248 T. goloskokovii AY818999 - T. holci AY819008 AF398459 T. indica AY818977 AF398434 T. laevis AY819004 AF398445 T. laevis AY819005 - T. menieri AY819002 AF398456 T. olida AY819000 - T. polypogonis AY819015 - T. sterilis AY819003 - T. tritici AY819006 AF398447 T. tritici AY819007 - T. walkeri AY818978 DQ143991 Lineage II Neovossia iowensis AY818988 DQ832253 T. barclayana AY818970 AF399894 T. barclayana AY818971 - T. boutelouae AY818973 - T. chionachnes AY818990 - T. ixophori AY819010 - T. kimberleyensis AY818979 - T. opaca AY818981 AF399884 T. savilei AY819018 AF399885 T. sumatii AY818986 AF399886 T. sumatii AY818987 - T. whiteochloae AY818989 - Lineage III T. aegopogonis AY818967 - T. asperfolia AY818968 - T. asperfolia AY818969 - T. obscuroreticulata AY819011 - Lineage IV Condiosporomyces ayresii AY819017 - C. verruculosus AY818984 - T. vittata AY818985 - Unresolved T. ehrhartae AY819013 AY770433 T. horrida AY818974 DQ827714 T. horrida AY818975 - T. rugispora AY818982 - T. rugispora AY818983 - T. setariae AY819014 - Pistil smut S. buchloëana (syn. Tilletia buchloëana) DQ659921 EF204935 S. buchloëana DQ659922 EF204936 S. buchloëana DQ659923 - S. buchloëana DQ659924 - Outgroups Erratomyces patelii AY818966 DQ663692 Exobasidium rhododendri DQ667151 DQ667153 Ustilago tritici DQ094784 DQ846894

61

Chapter 3

SALMACISIA, A NEW GENUS OF TILLETIALES: RECLASSIFICATION OF TILLETIA BUCHLOËANA CAUSING INDUCED HERMAPHRODITISM IN BUFFALOGRASS

Induced hermaphroditism is known to be caused by smuts from different orders and on a range of host species (Fisher and Holton 1957). To our knowledge, the pistil smut-buffalograss system of induce hermaphroditism is the only one recognized in fungal order Tilletiales or the grass family Poaceae indicating a highly specialized coevolutionary relationship. Fossil evidence for parasitically induced sex change in other organisms suggests a potential for long term coevolutionary relationships between hosts and castrators (Feldmann 1998). We speculate that pistil smut does not fit in any known genus of smut fungi, having rapidly evolved from its ancestral Tilletia genus as a result of an evolutionary arms- race with its host, which has culminated in induced hermaphroditism. We therefore propose a new Latin binomial combination for pistil smut, Salmacisia buchloëana, to better reflect its unique evolutionary history. It will be interesting to see if

Salmacisia remains a monotypic genus within order Tilletiales (Basidiomycota,

Ustilaginomycetes, Exobasidiomycetidae).

62

TAXONOMY

Salmacisia D. R. Huff & A. Chandra, gen. nov.

Sori tantum intus ovariis of aeger planta, agglutinated sporarum massa, teliospores ornatus intus hyalinae gelatinoid theca, basidiosporas unus nucleate iunctum vel duo nucleate, dikaryon producto secundus basidiosporas; formalis consimilis ex

Tilletia Tulasne & Tulasne, dissimilis nLSU-rDNA > 0.0375 nucleotide substitutions quisque positus intus Tilletia et < 0.0360 nucleotide substitutions quisque positus ex

Salmacisia buchloëana.

Sori only in ovaries of infected plants containing agglutinated spore masses.

Ornamented teliospores arise from terminal cells of sporogenous mycelia, frequently encased in hyaline gelatinoid sheath, germinating by means of continuous promycelium bearing terminal primary basidiospores that are either mononucleate which conjugate or binucleate, giving rise to secondary basidiospores. Indistinguishable from Tilletia

Tulasne & Tulasne for individual morphological characteristics but is clearly distinct from species of genera Tilletia, Conidiosporomyces, Ingoldiomyces, Neovossia, and

Erratomyces within Tilletiales by divergent nLSU- rDNA sequence exhibiting greater than 0.0375 nucleotide substitutions per site from any other genera within order

Tilletiales and less than 0.0360 substitutions per site from the typic species, Salmacisia buchloëana

Type species. Salmacisia buchloëana (Kellerman & Swingle) D.R. Huff &

A. Chandra.

63

Etymology. The genus Salmacisia is derived from Salmacis (pronounced săl-MĀ-sĭs) and refers to the ability of the type species of this genus to induce hermaphroditism in its host. According to Greek mythology, Salmacis was the determined water nymph responsible for transforming a remarkably handsome boy named Hermaphroditus into an intersexual individual possessing both male and female characteristics as a result of her divine union with him.

Salmacisia buchloëana (Kellerman & Swingle) D. R. Huff & A. Chandra comb. nov.

≡ Tilletia buchloëana Kellerman & Swingle, J of Mycology 5: 11. 1889.

Paratype on Buchloë dactyloides. USA. KANSAS: Trego and Ford County, 26 May

1888. BPI172553.

= Ustilago cathesteci P.Henn. Hedw. 35: 212. 1896.

≡ Tilletia cathesteci (P. Henn) G. P. Clint. J Myc. 8: 149. 1902.

Sori tantum intus ovariis of aeger planta, atrum brunneolae ut atrum puniceus brunneolae, agglutinated sporarum massa, cariosus nidor, teliosporas subglobosae, globosae, ovate vel elongate, pallens ut chocolate brunneolae, 13–26

µm diametro intus gelatinous hyalinae theca 1.5–3.5 µm creber, ornatus rotundus mons 1.5–2.0 µm altus, infecundus cells subglobosae vel ovoid encased in hyalinae theca, teliosporas producto simplex vel furca multinucleatae promycelia producto primary duo nucleate vel unus nucleate basidiosporas, unus nucleate primary basidiosporas iunctum, dikaryotic basidiosporas producto blastoform-typus dikaryotic secundus basidiosporas quod es usitas fusiform in forma, producto

64

intercellularibus contagio in aeger planta. visual frequentatio of spicas. Proprius

nucleotide varietas quisque positus: LSU: 33 proprius nucleotide varietas quisque

1340 positus; ITS1-5.8S-ITS2: 67 proprius nucleotide varietas quisque 734 positus.

Sori. In ovaries of staminate and/or pistillate plants; visual crowding of

inflorescences (FIGS. 2.3, 2.4 left); sori covered by the delicate pericarp and

concealed by floral bracts, pericarp easily punctured exposing the dark brown to

dark reddish brown agglutinated spore masses that give off fetid odor. Spores

subglobose, globose, ovate or elongate, pale yellow to light chocolate brown in

color (FIG. 2.11), 13–26 µm in diameter usually embedded in gelatinous hyaline

sheath 1.5–3.5 µm thick (FIG. 2.11) and exhibit obscure 1.5–2.0 µm tuberculate

ornamentation (FIG. 2.10). Single teliospore formed from terminal cells of

sporogenous mycelia. Teliospores are intermingled with sterile cells that are

subglobose or ovoid, encased in hyaline sheath, sometimes with a hyphal fragment

attached, often multilaminated walls and a comparatively small central lumen.

Germination of teliospores produces simple or branched multinucleate promycelia

(FIG. 2.12), giving rise to >30 primary binucleate or mononucleate basidiospores; mononucleate primary basidiospores conjugate (FIG. 2.13). Dikaryotic basiospores

give rise to blastoform-type of dikaryotic secondary basidiospores which are usually

fusiform in shape (FIG. 2.14) and produces systemic infection in its host (FIG. 2.18).

Characteristic DNA sequences: BLASTn results of the 28S LSU sequence of Salmacisia buchloëana (GENBANK DQ659921–DQ659924) had E-values of

0.0 (90% to 94% identical) and 1% to 3% gaps with the 28S LSU sequences from species of genera Tilletia, Conidiosporomyces, Ingoldiomyces, Neovossia, and

65

Erratomyces (TABLE I). BLASTn results of the ITS1-5.8S-ITS2 sequence of

Salmacisia buchloëana (GENBANK EF204935–EF204936) had E-values ranging

from 2e–107 to 9e–74 (78% to 87% identical) and 0% to 7% gaps with the ITS1-

5.8S-ITS2 sequences from species of genera Tilletia, Ingoldiomyces, Neovossia,

and Erratomyces (TABLE I). In the alignment of 28S rDNA sequences (TreeBase:

SN3366-14956), Salmacisia buchloëana displayed a cluster of gaps (base positions

406, 410–419, 423–436, 451) not present in species of genera Tilletia,

Conidiosporomyces, Ingoldiomyces, Neovossia, and Erratomyces.

Characteristic fixed DNA polymorphisms. Characteristic fixed DNA polymorphisms used in the description of Salmacisia buchloëana were determined

based on the sequence alignments used in the phylogenetic analyses. Only those

characters present in all Salmacisia buchloëana clones and absent in species sequences of Tilletia, Conidiosporomyces, Ingoldiomyces, Neovossia, and

Erratomyces are included. These fixed DNA polymorphisms are indicated with capital letters (nucleotides) with the alignment position given. Nucleotides separated by forward slashes indicate alternate base composition at that site. Gap openings and gap extensions resulting from sequence alignment are in brackets.

LSU (aligned sequences with gaps, 1340 bp): aatTtcgagaagCatt @ 104, 113; gctTatg @ 130; catAaTGtcc @ 188, 190–191; tt/ctAAgatAtgc @ 203–204, 208; ctaCgag @ 218; cgtAagg @ 312; tgaAgttaTGAaCg/aca @ 384, 389–391, 393; cagCatt[–1–]agT[–27–]tgtatt[–2–]gcGggc @ 402, 409, 447; ttgGctgcCgga @ 465,

470; gtaCTagg @ 482–483; gtaAttgatacGgtggCtgg @ 521, 529, 534; ctgTcCAaatgAcTtta @ 602, 604–605, 610, 612; ttgAgtg @ 671; taaGtga @ 1071.

66

ITS (aligned sequences with gaps, 734 bp): tgcGgaa @ 33; gatAccc/[–1–]g @ 51; tttAAgag @ 68–69; ccaAAtcttaCCcTaaacacaaGag/ct @ 106–107, 113–114, 116,

125; aaaGccCtac @ 136, 139; gaaGgagggAgcAtgc @ 169, 175, 178; tgcTcttctccGccc @ 198, 206; tacGatGaag @ 220, 223; aa/gaGaaT[–1–

]cttgtcttAggacaACaaTcat @ 244, 247, 257, 263–264, 267; gcaTcttgcgccctGAggt @

384, 395–396; ccTaag @ 405; attCTCgAtTccCa[–1–]GccttCttgtaaCgagaTggc @

438–440, 442, 444, 447, 450, 455, 462, 467; agtCggt @ 481; ctaGcagCatc @ 510,

514; ggcCtcccAAtAgCTtAcTggtcCgac @ 544, 549–550, 552, 554–555, 557, 559,

564; taaTtat @ 580; cgcAAgTacacaGtctgtgAaCggatgGtAgc @ 603–604, 606, 618,

625, 627, 633, 635; ccgAgTGCtct @ 648, 650–652; cacCttGGctttGtAAtCtgaccTcaaatCag @ 662, 665–666, 671, 673–674, 676, 682,

688.

On Poaceae: On Poaceae: Bouteloua dactyloides (Nutt.) Columbus; syn.

Buchloë dactyloides (Nutt.) Engelm.

Specimen examined: USA. OKLAHOMA: Kingfisher County. 6 JULY

1986, D. R. Huff (WSP 71313. HERBARIUM). Type on male and female sex forms of Buchloë dactyloides.

Known distribution: N. America (Kansas, Oklahoma, Texas, Nebraska,

Mexico).

67

BIBLIOGRAPHY

Clinton GP. 1902. North America Ustilagineae. J Myc 8:149–156.

Feldmann RM. 1998. Parasitic castration of the crab, Tumidocarcinus giganteus

glaessner, from the Miocene of New Zealand: Coevolution within the crustacean. J

Paleontology 72 (3):493–498.

Fischer GW and Holton CS. 1957. Biology and Control of the Smut Fungi. Ronald Press,

New York.

Henning P. 1896. Myxomycetes, Phycomycetes, Ustilagineae and Uredineae. Hedwigia

35:207–262.

Kellerman WA and Swingle WT. 1889. New species of Kansas fungi. J. of Mycology

5:11-14.

68

Chapter 4

INDUCED HERMAPHRODITISM CAUSED BY PARASITIC CASTRATION DISPLAYS AN ANCESTRAL PHENOTYPE BY REMOVING A SEX DETERMINING ONTOGENETIC LAYER FROM DIOECIOUS BUFFALOGRASS

ABSTRACT: A fascinating illustration of parasitic castration is the phenomenon known as induced hermaphroditism that occurs in a wide variety of organisms wherein a parasite induces opposite sex organs in individuals which would otherwise bear only a single sex. The only known example in the grass family Poaceae where a parasite induces hermaphroditism in both male and female sex forms occurs in buffalograss by a fungus we refer to as pistil smut. How parasites induce hermaphroditism in their hosts is unknown. Here we show that pistil smut induces hermaphroditism in male buffalograss by down-regulating a putative female-suppressor gene, BdTs2, homologous to maize

Tasselseed2 (Ts2). We found that high expression levels of BdTs2 correlate with selective abortion of pistils within hermphroditic floral precursor leading to unisexual male florets. Pistil smut infection inhibits BdTs2 expression and, in doing so, removes a sex determining ontogenetic layer permitting development of an ancestral-like hermaphroditic phenotype.

INTRODUCTION

Buffalograss [Buchloë dactyloides (Nutt.) Engelm., = Bouteloua dactyloides

(Nutt.) J. T. Columbus] is a perennial grass native to the North American

69 and historically known to be a primary food source of the region’s once immense herds of American bison (Bison bison L.). The separate male and female dioecious sex forms of buffalograss are morphologically so different that originally each was considered a separate species. However, upon receiving an uncommon monoecious sex form, George

Engelmann, the famous 19th century physian/botanist and a founding member of the

National Academy of Science, discovered that the dimorphic male and female plants were actually different sex forms of the same species; which he placed in a monotypic genus Buchloë derived from the Greek words boubalos meaning buffalo and chloë meaning grass (Engelmann 1859). Ever since then, buffalograss has served as the definition of dioecious sex expression among North American grasses (Hitchcock 1951,

Gould 1975) (FIG. 4.1, 4.2).

When we first observed the purple feathery stigmas of female sex organs protruding from flowers (termed florets in grasses) of male buffalograss (FIG. 4.3) we mistakenly thought we had discovered the hermaphroditic ancestor from which buffalograss evolved its dioecious breeding system. Our anticipation grew even more as the florets began to swell as though setting seed. However, our enthusiasm soon turned to confusion as we realized that it wasn’t seed at all, but rather swollen ovaries filled with a black powdery mass emitting a rotten fish odor. It was only later that we learned these

“seeds” were actually smut balls of tightly packed teliospores produced by a parasitic fungus originally named Tilletia buchloëana by Kellerman and Swingle in 1889

(Kellerman and Swingle 1889) and has recently been proposed to be reclassified into a new monotypic genus, Salmacisia Huff & Chandra, and named Salmacisia buchloëana

70

(Kell. & Swingl.) Huff & Chandra. Because S. buchloëana induces pistil (female sex organs) development and sporulates only within host ovaries, we refer to it as pistil smut.

Parasitically induced sex alterations of unisexual organisms is a phenomenon known as induced hermaphroditism and early taxonomists, including Linneaus, likely made similar mistakes as ours in misidentifying fungal induced hermaphroditic phenotypes of dioecious species as new hermaphroditic varieties (Fischer and Holton

1957, Hood and Antonovics 2003). To date, the most studied instance of induced hermaphroditism is caused by anther smut (Microbotryum violaceum (Pers) G. Deml &

Oberw. (=Ustilago violacea (Pers) Fuckel) which induces the development of male sex organs (stamens) in female plants of campion species (Silene spp.) and sporulates only within the anthers of infected florets (Uchida et al. 2003). Recently, we have shown that pistil smut induces hermaphroditism in not only male plants of buffalograss by inducing the development of otherwise vestigial pistils (female sex organs), but also in female plants by inducing hypertrophy of otherwise vestigial stamens. In addition to inducing hermaphroditism, pistil smut also induces the development of additional pistillate florets in both male and female plants (FIG. 4.4 vs. 4.5 and FIG. 4.6 vs. 4.7). Unlike anther smut, pistil smut is a systemic fungus that infects and grows intercellularly within vegetative meristems. When infected vegetative meristems transition into a flowering apex, fungal hyphae colonize the developing inflorescences inducing hermaphroditism. Fungal induced and naturally occurring host ovaries are supplanted with teliospores and are hence, reproductively sterile (i.e. parasitically castrated).

Parasitic castration is a class of diseases affecting animals, mollusks, insects, and plants where host gonads are sterilized by parasites preventing host sexual reproduction

71

(Hayes et al. 2002, Clay 1991, Boudoin 1975, Shukalyuk 2005). In doing so, parasites not only prevent the host from generating new resistance mechanisms but also, in many cases, diverts the unused energy and resources from host sexual reproduction towards the vegetative growth (“body”) of the host. Numerous examples of parasitic castration by

fungi have been shown to increase plant vegetative growth, vegetative biomass and

survivability which in turn enhance the parasite’s own growth and reproduction (Clay

1991, Boudoin 1975). However, we are unaware of any example of parasitic castration in plants where increased amounts of resources are allocated towards the development of

host sexual reproductive structures. Despite the interesting biology of induced hermaphroditism and parasitic castration, the molecular or biochemical mechanism underlying these phenomena are yet to be discovered.

Here we show a molecular basis for pistil smut induced hermaphroditism in male buffalograss. We will provide the first corollary evidence for the role of a putative

female-suppressor gene, homologous to Tasselseed2 (Ts2), in buffalograss unisex male floret development. We found that pistil smut infection perturbs the host’s sex determining pathway by removing this ontogenetic layer from unisexual male floret formation thereby permitting the development of a sex which the plant is genetically programmed not to exhibit.

Current theoretical models for the separation of sexes in plants accentuate that unisexual flowers have evolved, perhaps multiple times, from their hermaphroditic ancestors (Barrett 2005, Charlesworth 1978 ). The evolutionary transition from hermaphroditic precursor to unisexual florets requires the involvement of many sex- determining genes including male-suppressors and female-suppressors (Westergaard

72

1958). To date, many sex-determining genes have been theorized but only one putative

female-suppressor gene, Ts2, has ever been cloned, sequenced and characterized (Delong et al. 1993). In wild-type maize (Zea mays L.), Ts2 encodes for a short-chain dehydrogenase reductase (SDR) protein and presumably causes pistil abortion within

hermaphroditic floral precursor at the gynoecial ridge stage of floral development

resulting in maize’s unisexual male florets (Delong et al. 1993, Calderon-Urrea and

Dellaporta 1999). Mutations of Ts2, first described in 1920, allow pistils to complete development in otherwise male florets of maize (Emerson 1920). An ortholog of Ts2,

Gsf1 (gynomoecious sex form 1) in a wild relative of maize, Tripsacum dactyloides (L.)

L., exhibit expression patterns that also correlate with pistil abortion within staminate florets suggesting a common sex determination mechanism in tribe Andropogoneae

(subfamily Panicoideae) (Li et al. 1997).

Pistil smut induced hermaphroditism of male buffalograss phenotypically mimics maize Ts2 and Tripsacum Gsf1 mutants. Based on the genetic relatedness of buffalograss

(subfamily Chloridoideae) with maize and Tripsacum, all are members of the

monophyletic PACC clade (Panicoideae, Arundinoideae, Centothecoideae, and

Chloridoideae) of the grass family Poaceae (GPWG 2001), our experimental hypothesis is

that a Ts2-like gene is a key player in the separation of sexes during buffalograss unisex

floret development and that pistil smut might, either directly or indirectly, regulate Ts2-

like activity, thereby inducing hermaphroditic florets.

73

MATERIALS AND METHODS

Plant Material.— Seed of a Mexican diploid (2n=20) race of buffalograss were germinated on moistened paper disks at 30°C in the dark. Seedlings were transplanted into greenhouse potting soil (Promix®, Premier Horticulture, Inc., Quakertown , PA) and after six weeks of growth, each individual genotype was vegetatively propagated into four replicate clones, two of which were infected with pistil smut by embedding teliospores into the soil surface close to the base of the plant and two plants were left non-infected. A total of 49 genotypes (21 female and 28 male; a total of 196 clones) grown in the greenhouse (80° C day/ 70° C night) under natural day-length conditions were examined to study the effects of pistil smut infection.

Morphological stages of buffalograss inflorescence development.—

Morphological sequence of buffalograss inflorescence development from floret initiation until just prior to anthesis was categorized into six stages, namely, floral apex, pre-boot, boot, late-boot, emerging, pre-anthesis. Criteria used to characterize various stages of inflorescence development were based on the first appearance of the flag leaf with subsequent expansion and finally fully emergence of the inflorescence from within the flag leaf. Scanning electron microscopy was performed in order to capture the development of floral organs corresponding to the morphological development of buffalograss inflorescence.

Scanning Electron Microscopy.— Spikelets from non-infected and infected buffalograss of both sex forms from the six stages of inflorescence development were dissected to remove outer and inner glumes in order to expose the developing sex organs.

74

Specimens were fixed in 2.8 % gluteraldehyde and 0.02% Triton X in 0.1 M HEPES

buffer, pH 7.2, overnight at 4º C. The specimens were then post-fixed in 1% osmium

tetraoxide for 2 hrs followed by dehydration in an alcohol series (25%, 50%, 70%, 85%,

95%, and 100% ethanol for 5 min each) in preparation for critical point drying. After

critical drying, the specimens were mounted onto aluminium stubs using conductive

colloidal silver followed by sputter coating (10 nm) with gold particles in a vacuum.

Secondary electron emission images of the prepared conductive specimens were taken

using a JSM5400 scanning electron microscope with an accelerating voltage of 10 kv.

Isolation of full length Ts2 homolog (BdTs2) from buffalograss.— Hotstart,

touchdown PCR was performed to amplify a 497 nt region of Ts2 from buffalograss

genome to be used as a probe for gene expression studies. Amplification reaction mixture

(per 25 μl) contained: 1 X buffer, 2 mM MgCl2, 0.25 mM dNTPs, 5% DMSO, 1 M

Betaine, 2.5 U Taq-polymerase, and 200-300 ng genomic DNA (extracted by standard

CTAB procedure). Forward (5’CGCGCCGCCAAGAGCA TCCC3’) and reverse

(5’CGCCGCCGTCC ACGACCAGGTTG3’) primers were each included in the reaction mixture at a concentration of 0.15 μM. PCR cycling parameters were: initial denaturation

for 4 min @ 95º C followed by 16 cycles of 1 min @ 94º C, 1 min @ 65º C (reducing

temperature 1º C per 2 cycles), 1 min @ 72 º C followed by 19 cycles of 1 min @ 94º C,

1 min @ 57º C, 1 min @ 72 º C, with a final extension for 4 min @ 72 º C. The resulting

amplicon was gel-purified using QIAquick Gel Extraction Kit (Qiagen Science, Inc.,

Germantown, MD) and was cloned in pCR®II-TOPO® (Invitrogen, Carlsbad, CA). 3’

and 5’ Rapid Amplification of cDNA Ends (RACE) was performed in order to isolate 3’

and 5’ UTR regions of BdTs2. First strand cDNA synthesis and 3’ and 5’ RACE, based

75

on nested PCR, using adapter primers and BdTs2 specific primers was carried out

according to the guidelines of BD SMART™ RACE cDNA amplification kit (BD

Clontech, Mountain View, CA). In order to sequence full length coding region of Ts2

(1,226 nt) we designed forward (5’CGGCCCCCAATAG ATACAT3’) and reverse

(5’TCACAAGCC GATGAGGTTTC3’) primers in the 5’ and 3’ region of BdTs2 respectively. Similar PCR conditions as above were used to amplify full length BdTs2 which was then cloned in pCR®II-TOPO® (Invitrogen, Carlsbad, CA). Cloned full length BdTs2 was sequenced with the ABI Hitachi 3730XL DNA Analyzer (Applied

Biosystems, Foster City, California). BLASTn (Basic Local Alignment Search Tool) program from the national public database, NCBI (http://www.ncbi.nih.gov/) was used to

extract amino acid sequence for Ts2 from maize (AAC37345) and Ts2 homologs from

Tripsacum (AAB57738), buffalograss (ABE65370) and Bouteloua spp. (AAR16175,

AAR17511, AAR06288). Multiple alignment of amino acid sequences was conducted using CLUSTALW using default alignment parameters.

Quantitative Real-Time PCR.— Total RNA was extracted using RNAqueous®-

4 PCR Kit (Ambion, Austin, Texas). Two replicate RNA samples, pooled over genotypes, were extracted from each factor combination of sex form x infection treatment for each developmental stage. The internal oligonucleotide probe

(5’CGCAGAGGCCGTGCT CTTCCTG3’) was labeled with florescent dye 6-FAM at the 5’ end and with quencher dye BHQ-1 at the 3’ end (Biosearch Technologies Inc.,

Richmond, CA). The buffalograss Ts2 amplicon upstream PCR primer corresponded to base 399 to base 417 (5’CGCTCAGGCCCAGGGACA3’) and the downstream PCR primer corresponded to base 457 to base 478 (5’ATACATCTC CGGCCACAACCT 3’).

76

18s ribosomal RNA (rRNA) served as an endogenous control and was amplified along with the target Ts2 amplicon to standardize the amount of sample RNA added to each reaction. PCR cycling parameters used were 50º C for 3 min, 95º C for 10 min, and then

45 cycles of 95º C for 15 sec and 60º C for 1 min. Real time detection of fluorescence emissions was performed on an ABI PRISM 7700 (Applied Biosystems, Bedford, MA).

Each replicate RNA sample was measured three times. Data acquisition and analysis were carried out using Sequence Detection Software (Applied Biosystems, Bedford,

MA). Fold differences in the amount of target RNA were obtained through normalization to 18s rRNA expression levels.

RNA in situ hybridization.— Developing inflorescences from non-infected male plants were fixed in ice-cold fixative (4% (w/v) paraformaldehyde ; 4% (v/v) Triton-X in

1X PBS) under vacuum and then incubated for 12 hrs at 4º C with gentle shaking followed by dehydration in an ethanol series (30%, 40%, 50%, 60%, 70%, 85%, and

95%, 60 min each). Ethanol was removed and replaced with a series of ethanol and histoclear mixture followed by paraffin wax (Oxford Labware, St. Louis, MO) infiltration at 60ºC. Inflorescences were sectioned, 8 µm thick, with a Finnese microtome and mounted on Probe-On plus slides (Fisher Biotech, Pittsburgh, PA). DIG-labeled antisense RNA probe (497 nt) was prepared by in vitro transcription of BdTs2 using T7 polymerase and RNA in situ hybridization was performed following Jackson et al.

(1994).

77

RESULTS

We categorized buffalograss inflorescence development into six stages according to the following criteria (FIG. 4.8): Floral apex, represents the transition from vegetative

meristem to flowering apex by a slight elongation of the flowering culm; Pre-boot, when

the ligule of the flag leaf becomes visible and the inflorescence begins to enlarge within

the flag leaf sheath; Boot, when the sheath becomes noticeably swollen by enlargement

of the inflorescence; Late-boot, prior to emergence of inflorescence from sheath;

Emerging, first appearance of the inflorescence from within the sheath; Pre-anthesis, first visible sign of flower opening. Buffalograss florets typically exhibit an asynchronous pattern of growth during the early stages of inflorescence development, however, as the inflorescence matures, more and more florets synchronously reach the anthesis stage of development.

We began our investigation by examining the ontogenesis of unisexual florets in buffalograss at the microscopic level and found that florets begin development as hermaphrodites, initiating both male (androecium) and female (gynoecium) organ primordia (FIG. 4.9). As the hermaphroditic precursor reachs the gynoecial ridge stage of

development, male and female suppressors begin a process of selective abortion thereby

forming the subsequent ontogenetic layers leading to unisexual floral development. Male

suppressors result in the selective abortion of male sex organs in florets destined to

become unisexual female. Similarly, female suppressors result in the selective abortion

of female sex organs in florets destined to become unisexual male. These ontogenetic

constraints laid down by suppressor factors are characterized by vestigial organ remains

78

of the aborted sex organs (FIG. 4.9). Florets acquire their unisexual identity as the

unaborted sex organs are allowed to develop and reach full maturity (more details in

Appendix B). A similar pattern of unisexual floret development is described in maize

(Mitchell and Diggle 2005, Cheng et al. 1983, Irish and Nelson 1993) and Tripsacum (Li

et al. 1997) and may provide evidence for a common genetic mechanism.

We amplified, cloned, and sequenced the full-length (1,226 nt.) buffalograss

homolog of Ts2, BdTs2 (more details in Appendix C), and found 89% nucleotide

sequence homology with maize Ts2 (ZmTs2) and 85% amino acid sequence homology with ZmTs2 protein. Although the substrate for Ts2 is unknown, Ts2 belongs to SDR protein subfamily and presumably plays a role in steroidal-like pathway (Wu et al. 2007).

SDRs display GxxGxG and YxxxK amino acid motifs among its orthologs (Malcomber

and Kellogg 2006) and the multiple alignment of predicted amino acid sequences of

ZmTs2 and homologs from Tripsacum, buffalograss, and close relatives of buffalograss

(Bouteloua spp.) share GARGIG and YTASK amino acid motifs (yellow shaded areas in

FIG. 4.10). Amino acid sequence alignment shows high consensus except for a hyper

variable region corresponding to amino acid position 246 to 259 of buffalograss (gray

shaded area in FIG. 4.10) which interestly is shared with maize and Tripsacum but not

with Bouteloua spp.. We observed that this hyper variable region contains different

combinations of direct and inverted repeats as well as palindronic nucleotide sequences

(FIG. 4.11) and resides in the vicinity of Delong et al.’s (1993) Ac transposon insertion

site. As such, we speculate this hyper variable region might represent a regional hot-spot

for transposon activity.

79

Quantitative real-time PCR results indicated that BdTs2 expression levels were

similarly low for early stages of either male or female inflorescence development.

However, BdTs2 expression reached a considerably high level (8-fold increase) in male

inflorescences compared with female inflorescences shortly after the late boot stage of

inflorescence development. The high levels of BdTs2 expression in developing male

inflorescences correlate with the absence of pistils in male florets while the low levels of

BdTs2 expression in developing female inflorescences correlate with the presence of

pistils in female florets. These results suggest a role for BdTs2 in male unisexual floret

development (FIG. 4.12). RNA in situ hybridization demonstrated that the spatial expression of BdTs2 in immature male florets was localized in sub-epidermal layers of the vestigial gynoecium whose development had been arrested precisely at the gynoecial

ridge stage (inset FIG. 4.12 circled). Thus, the spatial and temporal correlation between

BdTs2 expression and pistil abortion suggests that BdTs2 is presumably one of the

ontogenetic constraints that prevent the complete development of the hermaphroditic

precursor as such, thereby leading to the formation of unisexual male florets. These

results are spatially and developmentally comparable to expression patterns for ZmTs2

(Delong et al. 1993) and Gsf1 (Li et al. 1997). We therefore conclude that BdTs2

exhibits a similar role as ZmTs2 and Gsf1 during the ontogenesis of unisexual male

florets from hermaphroditic floral precursors and provides evidence for the same genetic

mechanism of male sex determination (i.e. Ts2-induced pistil abortion) in buffalograss as

in maize and Tripsacum.

To test our hypothesis regarding the regulation of BdTs2 activity by pistil

smut we found that BdTs2 expression levels were substantially reduced (20-fold)

80

for infected male plants compared with the same genotypes not infected (FIG. 4.12).

Moreover, expression levels of BdTs2 in either male or female plants infected with

pistil smut were lower than in non-infected female plants. Thus, low expression

levels of BdTs2 in florets of infected male, infected female, and non-infected female

plants correspond to the presence of pistils whereas high expression levels in non-

infected males correspond to the absence of pistils (FIG. 4.12). From these results,

together with the loss of function mutations reported for ZmTs2 (Delong et al. 1993) and Gsf1 (Li et al. 1997), we conclude that pistil smut induces hermaphroditism in male buffalograss by either directly or indirectly down-regulating the expression of

BdTs2. In doing so, pistil smut infection removes an ontogenetic layer from the hermaphroditic precursor presumably by preventing BdTs2 mediated cellular arrest of pistil development and thereby allowing the unconstrained development of vestigial gynoeciums within florets that were otherwise destined to become unisexual male. A parallel molecular mechanism causing induced hermaphroditism in female campion by anther smut has yet to be discovered. In unisexual florets of infected female buffalograss, stamen growth appears unconstrained reaching a stage of being no longer vestigial but nonetheless underdeveloped (FIG. 4.12). Thus, in a

way, pistil smut infection removes ontogenetic layers from unisexual florets of both

male and female buffalograss sex forms and induces the expression of an

hermaphroditic phenotype in each. Our results offer insight into the molecular and

developmental basis for other parasitically induced hermaphroditic systems such as

male crabs infected with barnacles (Sacculina spp.) (Shukalyuk et al. 2005) and

81 perhaps chemically induced hermaphroditism such as the frog-atrazine system

(Hayes et al. 2002).

DISCUSSION

In 1889, William A. Kellerman, an instrumental early American mycologist and

Walter T. Swingle, who subsequently became a legendary agricultural botanist (Seifriz

1953) discovered an unusual smut fungus whose ‘monstrosity consists solely in its ability to produce ovaries in male plants’ (Kellerman and Swingle 1889). How parasites induce hermaphroditism in their hosts is unknown. Here we show that pistil smut infection induces ovary production in genetically male buffalograss by down-regulating a putative female suppressor gene, BdTs2, thereby allowing the development of an otherwise vestigial pistil. Whether BdTs2 is down-regulated directly from a product produced by the pistil smut fungus itself or by the host buffalograss in response to infection is not known. Furthermore, since other Tasselseed loci are known to exist in maize, including

Ts1 which is required for functional Ts2 activity (Irish et al. 1994), the exact target of pistil smut infection on sex expression of buffalograss may not directly be BdTs2 but may lie upstream of BdTs2. We also observed that BdTs2 continues to express within the arrested gynoecium throughout the remaining stages of buffalograss inflorescence development (FIG. 4.12). This result raises a particularly interesting question: If Ts2 evolved solely as a sex-determining gene in grasses, then why should BdTs2 continue to express in male florets well past the stage of sex determination? Although the evolutionary origin of Ts2 is unknown, its potential function as a mediator of cell death

82

has been postulated to have served as a tool in other developmental pathways prior to acquiring its role as a female-suppressor gene (Le Roux and Kellogg 1999, Kellogg 2000,

Kellogg 2001). Clearly, more research is needed to determine the regulatory role and

possible functions of Ts2.

We also show that pistil smut infection induces hermaphroditism in female

buffalograss by allowing the development of otherwise vestigial stamens. Within

grasses, development of pistils in unisex female florets and the development of stamens

in unisex male florets are believed to be under separate genetic control as the

developmental arrest of opposite sexes does not covary in evolutionary time (Malcomber

and Kellogg 2006). Nonetheless, pistil smut has the ability to alter these separate genetic

sex-determining mechanisms at the same time. We speculate that pistil smut infection is

either directly targeting male and female-suppressor loci independently or is acting early

in the developmental sequence by regulating meristem determinacy genes (Bortiri and

Hake 2007) and in turn indirectly regulating downstream evolutionary layers including

sex-determining loci. In either scenario, pistil smut infection is phenotypically altering

the dimorphic male and female buffalograss florets to resemble a morphologically

indistinguishable hermaphroditic phenotype.

We show that pistil smut infection allows the development of a sex which male

plants are genetically programmed not to exhibit or, in other words, infection prevents the

sex of the hermaphroditic precursor from being changed into unisexual male. In doing

so, the fungus permits host expression of an ancestral hermaphroditic phenotype that existed back in time and is not the result of any permanent genetic change within the host

as is the case for an atavism or a reversion. And so, in a way, we did discover the

83 hermaphroditic ancestor of buffalograss, in the form of a retrospective phenotype, or retrophenotype. These fungal induced retrophenotypes raise interesting coevolutionary biology questions, for example: do such patterns of parasitism happen by chance or have parasites, like pistil smut, played a role in the evolution of separate sexes in their hosts?

(Wilson 1983).

84

BIBLIOGRAPHY

Barrett SCH. 2002. The evolution of plant sexual diversity. Nature Reviews Genetics 3,

274-284.

Bensen R.J, G.S Johal, V.C Crane, J.T Tossberg, P.S Schnable, R.B Meeley, and S.P

Briggs. 1995. Cloning and characterization of the maize An1 gene. Plant cell 7(1): 75–

84.

Bortiri E and Hake S. 2007. Flowering and determinacy in maize. J Expt. Botany

8(5):909-916.

Boudoin, M. 1975. Host castration as a parasitic strategy. Evolution 29 (2): 335-352.

Calderon-Urrea A and Dellaporta SL. 1999. Cell death and cell protection genes

determine the fate of pistils in maize. Development. 126: 435-441.

Charlesworth, B. and Charlesworth, D. 1978. A model for the evolution of dioecy and

gynodioecy. Am. Nat. 112: 975-997.

Cheng PC, Greyson RI and Walden DB. 1983. Organ initiation and the development of

unisexual flowers in the tassel and ear of Zea mays. Am. J. Bot. 70(3): 450-462.

Clay K. 1991. Parasitic castration of plants by fungi. Trends Ecol. Evol. 6: 162-166.

Delong A, Calderon-Urrea A and Dellaporta SL. 1993. Sex determination gene

TASSELSEED2 of maize encodes a short-chain alcohol dehydrogenase required for

stage-specific floral organ abortion. Cell 74: 757-768.

85

Delong A, Calderon-Urrea A and Dellaporta, SL. 1993. Sex determination gene

TASSELSEED2 of maize encodes a short-chain alcohol dehydrogenase required for

stage-specific floral organ abortion. Cell 74: 757-768.

Engelmann G. 1959. Two new dioecious grasses of the United States. Trans. St. Louis

Acad. Sci. 1: 431-443.

Fischer GW and Holton,CS. 1957. Biology and Control of the Smut Fungi. Ronald Press,

New York.

Gould FW. 1975. The Grasses of Texas. Texas A&M University Press, College Station.

Grass Phylogeny Working Group. 2001. Phylogeny and subfamilial classification of the

grasses (Poaceae). Annals of the Missouri Botanical Garden. 88:373-457.

Hayes TB, Collins A, Lee M, Mendoza M, Noriega N, Stuart AA, Vonk A. 2002.

Hermaphroditic, demasculinized frogs after exposure to the herbicide atrazine at low

ecologically relevant doses. Proc Natl Acad Sci USA 99:5476–5480.

Hitchcock AS. 1951. Manual of the grasses of the United States. Misc. Publ. No. 200.

Washington, DC: U.S. Department of Agriculture, Agricultural Research

Administration. 1051 p.

Hood ME, Antonovics J. 2003. Plant species descriptions show signs of disease. Proc.

Royal Soc. London Series B (Suppl.) 270:S156-S158.

Irish EE and Nelson LJ. 1994. Interaction between tasselseed genes and other sex

determining genes in maize. Developmenatl Genetics 15:155-171.

Irish EE and Nelson TM. 1993. Development of TASSELSEED2 inflorescence in maize.

Am. J. Bot. 80(3): 292-299.

86

Jackson D, Veit B and Hake S. 1994. Expression of maize KNOTTED1 related

homeobox genes in the shoot apical meristem predicts patterns of morphogenesis in the

vegetative shoot. Development 120: 405-413.

Kellerman WA and Swingle WT. 1889. New species of Kansas fungi. J. of Mycology

5:11-14.

Kellogg EA. 2000. The grasses: A case study in macroevolution. Annu. Rev. Ecol Syst.

31: 217-238.

Kellogg EA. 2001. Evolutionary history of the grasses. Plant Physiology 125(3): 1198-

1205.

Le Roux LG and Kellogg EA. 1999. Floral development and the formation of unisexual

spikelets in the Andropogoneae (Poaceae). Am. J. Bot. 86: 354-368.

Li D, Blakey CA, Dewald C and Dellaporta, SL. 1997. Evidence for a common sex

determining mechanism for pistil abortion in maize and its wild relative Tripsacum.

Pro.c Natl. Acad. Sci. 94: 4217-4222.

Malcomber ST and Kellogg EA. 2006. Evolution of unisexual flowers in grasses

(Poaceae) and the putative sex-determining gene, TASSELSEED2 (TS2). 170:885-

899.

Mitchell CH and Diggle PK. 2005. The evolution of unisexual flower: Morphological and

functional convergence results from diverse developmental transition. Am. J. Bot.

92(7): 1068–1076.

Seifriz W. 1953. Walter T. Swingle. 1871-1952. Science 118(3063):288–9.

87

Shukalyuk AI, Isaeva VV, Pushchin II and Dolganov SM. 2005. Effects of the

Briarosaccus callosus Infestation on the Commercial Golden King Crab Lithodes

aequispina. J. Parasitology 91 (2): 1502-1504.

Westergaard M. 1958. The mechanism of sex determination in dioecious flowering

plants. Adv. Genet. 9:217-281.

Wilson MF. 1983. Plant reproductive Ecology. John Wiley & Sons, New York.

88

4.11 4.2 4.3

4.4 Florets 4.5 Florets

Stigma

Style

Anthers Ovary filled Reduced with anthers teliospores Florets 4.6 4.7

Stigma Stigma

Pistil Pistil Style

Style Ovary Ovary Hyper trophied anthers

Figure 4.1 to Figure 4.7: Pistil smut induced hermaphroditism and development of additional florets in buffalograss.

89

4.1, Male inflorescences extend above the plant foliage, typically bearing 2-3 spikes, with

functional, pollen producing anthers protruding out from within the florets. 4.2, Female

inflorescences remain low within the foliage, typically bearing 3-4 spikes (seed dispersal

capsules), with purple feathery stigmas protruding out from within the florets. 4.3,

Inflorescence of infected male plant containing induced female sex organs (pistils) as

evidenced by the presence of purple, feathery stigmas. 4.4, Spikelet of non-infected male

plant containing two florets each with three developed anthers. 4.5, Spikelet of infected

male plant containing three florets each with three reduced anthers and a fully developed

pistil supplanted with teliospores (smut balls). 4.6, Spikelet of non-infected female plant containing a single fully developed pistil. 4.7, Spikelet of infected female plant with two florets each with a fully developed pistil and hypertrophied anthers.

90

Male 4.8

Floral apex Pre-boot Boot Late-boot Emerging Pre-anthesis

Female

Floral apex Pre-boot Boot Late-boot Emerging Pre-anthesis

Figure 4.8: Stages of male and female buffalograss inflorescence development.

91

4.9

50 μm GY

AN Floral primordia

50 μm

Hermaphroditic GY Gynoecial ridge stage precursor AN

Male Female suppressor suppressor

AN

GY Sex determination vGY X

100 µm vAN

Pistil

vAN Maturation AN

50 μm Female unisex floret Male unisex floret

92

Figure 4.9: Scanning electron micrographs depicting the ontogenetic events of

buffalograss unisexual floret development.

Florets, containing male (androecium, (AN)) and female (gynoecium, (GY)) organs, initiate development as hermaphrodites in floral precursor. Suppressor factors signal the

transition from a hermaphroditic precursor to unisexual florets at the gynoecial ridge stage of floret development. Female suppressors allow continued development of the

androecium into a whorl of three stamens but leave behind a vestigial remnant of the

gynoecium (vGY) (note: one of the three stamens removed (X) to make gynoecium

visible). Male suppressors allow continued development of gynoecium into a pistil but

leave behind vestigial remnants of the androecium (vAN). As these florets mature, they

acquire an identity of either unisexual male or unisexual female.

93

4.10

BdTs2 ------MQAAAMPALDPLPEKSHAHQ----TPHHGWESNGGAAAVVAPTPAPRKLDGKV 49 OsTs2 -----ASYAAAAMPALDVLPEK------PHHGWESNGGAA--VAPTPAPRRLDGKV 43 BhTs2 -----ASYAAAAMPALDVLPEK------PHHGWESNGAAA--VAPTPAPRRLDGKV 43 BtTs2 ----- ASYAAAAMPALDLLSEK------SHHGWDGTGAAA--VAPTPAPRRLDGKV 43 ZmTs2 MHASLASYAAAAMPALDLRPEIAHAHQPVMSPSHHGWDGNGATA---VPTPMPKRLDGKV 57 Gsf1 M HASLASYAAAAMPALDLRPEIAHAHQPVMSPSHHGWDGNGAAA---VPTPMPKRLDGKV 57 ******** * **** * * *** * *****

BdTs2 AIVTGGARGIGEAIVRLFAKHGARVVIADIDAAAGDALAAALGPQVSCVRCDVSVEDDVG 109 OsTs2 AIVTGGARGIGEAIVRLFVKHGARVVIADIDAAVGDALAAALGPQVSCVRCDVSVEDDVK 103 BhTs2 AIV TGGARGIGEAIVRLFARHGARVVIADVDAAAGDALAAALGPQVSCVRCDVSVEDDVR 103 BtTs2 AIVTGGARGIGEAIVRLFVKHGARVVIADIDDAAGDALAAALGPQVSCVRCDVSVEEDVK 103 ZmTs2 AIV TGGARGIGEAIVRLFAKHGARVVIADIDDAAGEALASALGPQVSFVRCDVSVEDDVR 117 Gsf1 AIVTGGARGIGEAIVRLFAKHGARVVIADIDDAAGEALAAALGPQVSFVRCDVSVEEDVR 117 ****************** ********* * * * *** ******* ******** **

BdTs2 RAVEWAVARHG-RLDVLCNNAGVLGRQTRAAKSILSFDAAEFDAVLRVNALGAALGMKHA 168 OsTs2 R AVEWAVARHG-RLDVLCNNAGVLGRQTRAAKSILSFDAGEFDRVLRVNALGAALGMKHA 162 BhTs2 RAVEWAVARHG-RLDVLCNNAGVLGRQTRAAKSILSFDAGEFDRVLRVNALGAALGMKHA 162 BtTs2 RAVEWAVARHG-RLDVLCNNAGVLGRQTRAAKSILSFDAGEFDRVLRVNALGTALGMKHA 162 ZmTs2 R AVDWALSRHGGRLDVYCNNAGVLGRQTRAARSILSFDAAEFDRVLRVNALGAALGMKHA 177 Gsf1 RAVDWALSRHGGRLDVYCNNAGVLGRQTRAAKSILSFDAGEFDRVLRVNALGAALGMKHA 177 *** ** *** **** ************** ******* *** ******** *******

BdTs2 ALAMAP RRAGSIVSVSSVAGVLGGLGPHAYTASKHAIVGLTKNAACELGAHGIRVNCVSP 228 OsTs2 ALAMAPRRAGSIVSVASVAGVLGGLGPHAYTASKHAIVGLTKNAACELGAHGIRVNCVSP 222 BhTs2 ARAMAPRRAGSIVSVASVAGVLGGLGPHAYTASKHAIVGLTKNAACELGAHGIRVNCVSP 222 BtTs2 ALAMAP RRAGSIVSVASVAGVLGGLGPHAYTASKHAIVGLTKNAACELGAHGIRVNCVSP 222 ZmTs2 ARAMAPRRAGSIVSVASVAAVLGGLGPHAYTASKHAIVGLTKNAACELRAHGVRVNCVSP 237 Gsf1 A RAMAPRRAGSIVSVASVAGVLGGLGPHAYTASKHAIVGLTKNAACELGAHGVRVNCVSP 237 * ************* *** **************************** *** *******

BdTs2 F GVATNMLINAWRQGHADGGGGDD---DVDIDIAVPSDEEVEKMEEVVRGFATLKGPTLR 285 OsTs2 FGVATPMLVNAWRQGHAA------AVPSDEEVEKMEEVVRGFATLKGPTLR 267 BhTs2 F GVATPMLINAWRQGHAD------AVPSDEEVEKMEEVVRGLATLKGPTLR 267 BtTs2 FGVATPMLINAWRQGHAD------AVPSDEEVEKMEEVVRGFATLKGPTLR 267 ZmTs2 FGVATPMLINAWRQGHDDATADADRDLDLDLDVTVPSDQEVEKMEEVVRGLATLKGPTLR 297 Gsf1 F GVATPMLINAWRQGHDGAAD-----AELDLDINVPSDQEVEKMDAGA--VVDINVPFKF 290 ***** ** ******* **** ***** *

BdTs2 PRDIAEAVLFLASDESRYVSGHNLVVDGGVTTSRNLIGL 324 OsTs2 P------268 BhTs2 P ------268 BtTs2 P------268 ZmTs2 P RDIAEAVLFLASDEARYISGHNLVVDGGVTTSRNLIGL 336 Gsf1 LFSFTDGRYIFGFSKANDQASHRMLDTFSKKKIMKFLII 329

Figure 4.10: Multiple sequence alignment of predicted amino acids for Ts2 gene.

94

Ts2 amino acid sequences obtained from buffalograss (BdTs2), Opizia stonolifera (=

Bouteloua dimorpha) (OsTs2), Bouteloua hirsuta (BhTs2) and Bouteloua trifida (BtTs2),

maize (ZmTs2) and Tripsacum (Gsf1) using Clustal W program. Colored alphabets indicate amino-acids according to the standard genetic code. * indicate sites with

conserved amino acid. Yellow shaded areas indicate GARGIG and YTASK amino acid

motifs characteristic of SDR protein subfamily. Gray shaded area indicates a hyper

variable region present in maize, Tripsacum and buffalograss.

95

>Buchloë dactyloides (Buffalograss) 4.11 TCTTGACCGGCCGAAGTTCGAGCCGTGCTCCGCGTCAACGCGCTGGGCGCCGCGCTCGGGATG AAGCACGCCGCGCTCGCCATGGCGCCGCGCCGCGCGGGCAGCATCGTNTCCGTCTCCAGCGTC GCCGGCGTGCTCGGCGGGCTGGGCCCGCACGCGTACACCGCCTCCAAGCACGCCATCGTCGG GCTCACCAAGAACGCCGCCTGCGGGCTCGGCGCGCACGGCATCCGCGTCAACTGCGTCTCGCC 1 CTTCGGCGTCGCCACGAACATGCTCATCAACGCGTGGCGCCAGGGCCACGCCGACG[GCGG 36 CGGCGGCGACGACGACGTCGACATCGACATCG]CCGTGCCCAGCGACGAGGA 77 GGTGGAGAAGATGGAGGAGGTGGTCAGGGGGTTCGCCACGCTCAAAGGACCCACGCTCAGGC CCAGGGACATCGCAGAGGCCGTGCTCTTCCTGGCCAGCGACGAGTCCAGATACATCTCCGGCC ACAACCTGGTCGTGGAACGGCGGCGA

[GCGGCGGCGGCG ACGACGACG TCGACATCGACA TCG] 1 36 (GCG)4 (ACG)3 (TCGACA)2

DIRECT REPEATS:

1 6 7 12 GCGGCG GCGGCG

13 24 25 36 ACGACGACGTCG ACATCGACATCG

52 59 60 69 70 77 AGGAGGTG GAGAAGATGG AGGAGGTG

8 27 44 63 CGGCGACGACGACGTCGACATCGACATCGCCGTGCCCAGCGACGAGGAGGTGGAGA

DIRECT REPEAT MISMATCH FOOTPRINT DELONG’S (1993) AC INSERTION SITE

INVERT REPEATS: 2 20 21 39 CGGCGGCGGCGACGACGAC GTCGACATCGACATCGCCG

3 23 24 44 GGCGGCGGCGACGACGACGTC GACATCGACATCGCCGTGCCC

2 25 28 51 CGGCGGCGGCGACGACGACGTCGA TCGACATCGCCGTGCCCAGCGACG

Figure 4.11: Hypervariable region (36 nt.), comprising of a combination of direct, inverted, repeats present in buffalograss, maize and Tripsacum but absent in Bouteloua

species. 96

Anthers 4.12 450

400

350

300

250

200

150

Relative Ts2 mRNA expression leve 100

50

0 Boot Boot Boot Boot Pre-Boot Pre-Boot Pre-Boot Pre-Boot Emerging Emerging Emerging Emerging Late-Boot Late-Boot Late-Boot Late-Boot Floral apex Floral apex Floral apex Floral apex Pre-Anthesis Pre-Anthesis Pre-Anthesis Pre-Anthesis Male Female Infected male Infected female

97

Figure 4.12: Role of the female suppressor gene, BdTs2, in buffalograss unisexual floret development and its regulation by pistil smut.

Quantitative real time PCR analysis of BdTs2 expression for different stages of inflorescence development in male and female genotypes infected and not infected by pistil smut. Pictures of mature florets correspond to their respective sex forms of either non-infected or infected plants. (inset) RNA in-situ hybridization of digoxyginin labeled

BdTs2 probe to subepidermal tissue of gynoecium (circled) at the gynoecial ridge stage of development in a unisexual male floret.

98

Chapter 5 TEN-FOLD INCREASE IN POTENTIAL SEED YIELD DUE TO FUNGAL INFECTION OF BUFFALOGRASS

ABSTRACT: Sex expression of dioecious buffalograss is environmentally stable with 1:1 sex ratios. The present study shows that the feminizing effects of pistil smut fungus infection on dioecious buffalograss sex expression causes a shift in the phenotypic sex ratios from approximately 1:1 to almost 100% hermaphroditic sex form in the population. Detailed morphological analysis supplemented with scanning electron microscopy revealed other secondary effects of fungal infection like induction of additional pistillate floret per spikelet in each sex form as well as early and increased flowering leading to ten-fold increase in ovary production. Parasitic castration may have the potential of altering resource allocation towards vegetative growth of the host.

Although all ovaries of infected plants are sterile, infection by pistil smut reallocates plants resources from vegetative growth towards ovary production thereby enhancing the potential for overall seed yield. This study provides the first evidence for parasitically enhanced sexual reproductive allocation (SRA).

INTRODUCTION

Plants differentially allocate energy and resources towards either vegetative growth, the process of sexual reproduction, or maintenance, depending on their life- history characteristics (Wilson 1983). Sexual reproductive allocation (SRA) is

99 essentially the amount of energy and resources that an organism devotes or allocates towards the process of sexual reproduction (Charnov 1982, Wilson 1983), resulting in genetically variable progenies. Similarly, vegetative reproductive allocation (VRA) is the amount of energy and resources that plants allocate towards its survival through vegetative propagation (Charnov 1982, Wilson 1983), resulting in genetically uniform clones of the parent plant.

Evolution molds SRA and VRA differently in different organisms depending on environmental adaptations and life history strategies (Charnov 1982).

Researchers often find that a tradeoff (negative correlation) exists between SRA and VRA. Perennial plants exhibit higher VRA as they survive and proliferate vegetatively by accumulating greater proportion of photosynthates in leaves than in seeds, for example most turfgrass species including buffalograss. A reduced amount of resource allocation towards sexual reproduction (SRA) in perennial grasses is one of the major reasons for their reduced seed yields and harvest index.

Annual plants, on the other hand, allocate all resources towards SRA after a brief

VRA phase. For example, wheat, rice, maize, and other important agronomic crops that have been bred for thousands of years for increased seed yields; a classical example is the introduction of dwarfing genes in wheat and rice leading to green revolution in 1960s (Borlaug et al. 1969).

In agriculture, a measure of SRA is commonly called seed yield which is capable of being described by a multiplicative formula of components, for example:

Seed Yield = Number of Inflorescences/Plant x Number of

Seeds/Inflorescence x Weight/Seed.

100

The ideal plant form that plant breeders attempt to achieve through their genetic

manipulation typically have improved seed yield as a main goal. However, achieving gains in seed yield is commonly observed to be complicated by the polygenic nature of the trait and by negative correlations among the individual seed yield components. Furthermore, breeding for high yielding varieties and is a long-term commitment. Moreover, various biotic and abiotic stresses pose serious threats and difficult challenges along the way. In all environments, possibly the greatest threat to productivity is diseases, especially those caused by fungal pathogens (Reynolds and Borlaug 2006), for example smut diseases. Grain or seed smuts not only fill up host ovaries with fungal spores rendering them functionally and economically worthless, but they also limit the ability of the plant to sexually reproduce, i.e. parasitically castrated (Clay 1991). In many cases, smut infection takes the unused energy and resources that were going to be utilized for the sexual reproduction of the plant and redirects them towards the vegetative growth of the plant (Clay 1991, Kover 2000). In other words, parasitic castration, in general, is known to decrease SRA and enhance VRA of its host, causing a detrimental effect on the potential seed yield and harvest index of the plant. However, no example of any example of parasitic castration seems to exist in the literature increase in the potential SRA of its host.

In the present study I will document various secondary effects of pistil smut

(Tilletia buchloëana Kell & Swing.; recently proposed as Salmacisia buchloëana

(Kell & Swing.) Huff & Chandra) infection on both male and female plants of dioecious buffalograss (Buchloë dactyloides (Nutt.) Engelm.). Pistil smut fungus

101

has been known to induce the development of female sex organs (pistils) in florets

of male buffalograss (Kellerman and Swingle 1889). Here I will show how the parasitic pistil smut fungus demonstrates an ability to increase the overall SRA of buffalograss by systematically enhancing various seed yield components. What’s even more amazing is that the fungus performs this task in such a coordinated fashion that negative tradeoffs among the yield components are essentially lacking.

Even so, natural and induced pistils within florets of male and female buffalograss are supplanted with fungal teliospores and are hence reproductively sterile.

However, if these secondary effects of pistil smut infection can be realized without the presence of the fungus, then these induced seed-bearing structures provide potential for dramatically enhancing the harvest index of buffalograss and, in general, other perennial grasses.

MATERIALS AND METHODS

Isolation of pistil smut.— The pistil smut fungus used in this study was a single isolate originally collected from an individual male buffalograss plant growing in a short grass prairie located in Kingfisher County, Oklahoma in 1984 (Huff et al. 1987).

Plant Material.— Seeds of sixty genotypes of Mexican diploid (2n=20) (Huff et al. 1992) buffalograss were germinated on moistened paper disc and, in November of

2004, were transplanted into plastic pots (15.25 cm diameter) containing potting soil

(Promix®, Premier Horticulture, Inc., Quakertown, Pennsylvania) in the greenhouse (26

102

C day/21° C night) under natural day length conditions. Each genotype was vegetatively

propagated, at the 4 to 6 tiller stage, into four clonal replicate plants.

Host infection by pistil smut.— Two clonal replicate plants of each genotype

were left uninfected and two of which were inoculated by embedding teliospores into the

soil surface close to the base of vegetative shoots. Inoculated plants were saturated with

water and kept sealed in clear plastic bags in order to maintain high humidity conditions,

for approximately 6 weeks to allow teliospores to germinate and the fungus to enter the

plant. A total of 49 genotypes (21 female and 28 male; a total of 196 clones) grown in

the greenhouse (80° C day/ 70° C night) under natural day-length conditions were

examined to study the effects of pistil smut infection.

Resource partitioning analysis.— In June 2005 and 2006, all plants were

harvested 3 cms above the potting soil surface and separated into sexual reproductive

biomass (all flowering shoots and stems) and vegetative biomass (all vegetative shoots

and stems). These plant material were placed in brown paper bags, air-dried in the

greenhouse for two weeks and then received a final drying for 24 hrs @ 55° C and

weighed. Ability to vegetatively propagate buffalograss allowed me to use Student’s

paired t-test, which was conducted using Minitab (Minitab® release 14.2, Minitab Inc,

2005) to analyze the differences in biomass accumulation between paired samples of non-

infected and infected genotypic clones. Two-factor analysis of variance (Two-way

ANOVA) was used to test difference between two years of data collection.

Analysis of components of sexual reproductive allocation.— Sexual

reproductive allocation is sub-divided into various seed yield components, for example, number of inflorescence per plant, number of spikes per inflorescence, number of

103

spikelets per spike, number of florets per spikelet, and seed weight or weight of smut balls for infected plants. Data on various seed yield components was collected, in May

2006, for female and male genotypes infected and non-infected with pistil smut.

Student’s paired t-test was conducted using Minitab (Minitab® release 14.2, Minitab Inc,

2005) to analyze the differences in various yield components between paired samples of non-infected and infected genotypic clones.

RESULTS

Pistil smut was found to substantially alter host resource partitioning between

vegetative and sexual reproductive biomass in both 2005 and 2006 (TABLE 1). Although

significant differences were found between years for biomass accumulation in both male

and female genotypes upon infection, each year showed similar trends in growth pattern

and resource partitioning. Total aboveground plant biomass was substantially reduced

for both female and male plants upon infection (FIG. 5.1 vs. FIG. 5.2 and TABLE 5.1, 5.2)

suggesting that pistil smut is a virulent parasite. Infection reduced vegetative plant

biomass to an even greater extent than total aboveground plant biomass, however, sexual

reproductive biomass was greatly increased in both male and female sex forms infected

with pistil smut. Taken together, these results have an additive effect on increasing

sexual reproductive allocation (SRA) and decreasing vegetative allocation in infected

plants (TABLE 1, 2). Overall, SRA was increased 36.6-fold in 2005 and 1.69-fold in 2006

for infected female plants, and 7.93-fold in 2005 and 2.15-fold in 2006 for infected male

104

plants compared to their non-infected clones. Significant increase in sexual reproductive biomass upon pistil smut infection resulted, primarily, from increased number of flowering as well as earlier onset of flowering in both male and female plants as compared to non-infected plants. Differences observed in two year of data collection are

probably because at the time of first data collection in June 2005, the plants were too

immature (only 7-months young) to flower as compared to June 2006 data collection.

In June 2006, data was collected on estimated components of SRA (TABLE 3).

Number of inflorescences per plant, spikes per inflorescence and florets per spikelet were significantly increased in both male and female genotypes upon pistil smut infection.

Number of spikelets per spike was significantly increased in females upon infection, however, they were significantly decreased at 20% level of probability in males, probably due to a negative trade-off. Overall, this remarkable increase in SRA is primarily due to the ability of pistil smut infection to regulate meristem determinacy within infected plants

(by increasing the frequency of flowering shoots; FIG. 5.3) and their floral meristems

(producing extra florets) (FIG. 5.4). This latter observation confirms Engelmann’s

(Engelmann 1959) that male florets may contain a third floret per spikelet and explains

the line drawing by Kellerman and Swingle (Kellerman and Swingle 1989) (1889, Fig.

11 therein) depicting an infected male spikelet with three florets.

In female plants, fungal infection induces an overall 12.6-fold increase in

ovary production. In male plants, the fungus induces a 2.3-fold increase in floret

number and over 95% of these florets contain a well developed fungal induced pistil, while the remaining 5% escape infection and appear to function as normal pollen-producing male florets (TABLE 3). One caveat that deserves mentioning at

105

this point is that mature ovaries (induced or not) of all infected florets (male or

female) are filled with the teliospores of the pistil smut and are thus, reproductively

sterile. For this reason I refer to the fungal-induced increase in overall buffalograss seed yield as a “potential” seed yield increase. In addition, the weight per seed

component of seed yield can only be calculated for non-infected female plants whereas for infected plants of both sexes this component is represented as weight

per smut ball (teliospore filled ovary) (TABLE 3).

I found that seed yield per non-infected female plant is 1/10 that of the smut

ball yield for infected plants. Assuming that the energetic costs to the host-plant in

filling its ovaries with fungal teliospores is roughly equivalent, to that of filling

uninfected ovaries with starchy endosperm during seed formation, then the potential

for seed yield increase in buffalograss would be substantial. The total weight of

smut balls produced from infected males is almost twice to that of infected females

(1.09 gm for females and 2.70 gm for males) but the number of induced male pistils

is approx. four times greater than that of infected female plants. Thus, either

buffalograss physiology limits the amount of ovary filling that may occur or, that

there are structural differences between male and female ovaries. Finally, if the

effects of fungal infection were applied across all sexual reproductive components,

including the equivalency of seed weight, and was indexed based on total

aboveground plant dry weight, it would represent an estimated 61-fold-increase in

Harvest Index (0.19 % for a male–female pair of uninfected plants vs. 11.72% for a

pair of infected plants). If such an increase in SRA were capable of being realized

without the fungus, it would certainly seem to have application for improving

106 perennial grain crops whose low seed yield has been problematic for commercialization.

DISCUSSION

Smuts are economically important due to their devastating impact on food grain crops worldwide. As such, nearly all efforts are focused on the prevention, control, and embargo identification of smuts in world agriculture. However, buffalograss-pistil smut provides a non-threatening model system in which to exploit the ability of smut infection to regulate its host sex expression and meristem determinacy at the molecular and biochemical levels. If pistil smut’s ability to manipulate host sexual reproductive allocation were better understood and capable of being realized without the fungus, it may actually benefit agriculture by improving the harvest index of grain crops in general and specifically in perennial grain crops whose low seed yields have been problematic for commercialization (Cox et al. 2006).

Describing in more general terms the immense importance of breeding for high seed yielding cultivars for the world population, I would like to quote Jonathan Swift in

"Gullivars Travels" from 1726:

"And he gave it for his opinion, that whoever could make two ears of corn or two blades of grass to grow upon a spot of ground where only one grew before, would deserve better for mankind, and do more essential service to his country, than the whole range of people that just only is speaking about it".

107

BIBLIOGRAPHY

Borlaug NE, Narvaez I, Aresvik O, Anderson RG. 1969. Green revolution yields a golden

harvest. Columbia J World Business 4(5):9-19.

Charnov EL. 1982. The Theory of Sex Allocation. Princeton University press, New

Jersey.

Clay K. 1991. Parasitic castration of plants by fungi. Trends Ecol. Evol. 6: 162-166.

Cox TS, Glover JD, Van-Tassel, Cox CM, DeHaan LR. 2006. Bioscience 56:649-659.

Huff DR, Zagory D and Wu L. 1987. Report of Buffalograss Bunt (Tilletia buchloëana)

in Oklahoma. Plant Dis. 71:651.

Huff DR and Wu L. 1992. Distribution and inheritance of inconstant sex forms in natural

populations of dioecious buffalograss (Buchloë dactyloides). Am. J. Bot. 79(2): 207-

215.

Kellerman WA and Swingle WT. 1889. New species of Kansas fungi. J. of Mycology

5:11-14.

Kover PX. 2000. Effects of parasitic castration on plant resource allocation. Oecologia

123:48-56.

Reynold MP and Borlaug NE. 2006. Impacts of breeding on international collaborative

whaet improvement. J Agricultural Science 144:3-17.

Wilson MF. 1983. Plant reproductive Ecology. John Wiley & Sons, New York.

108

Non-infected buffalograss 5.1

Decreased levels of SRA Increased levels of VRA

Infected buffalograss

5.2

Increased levels of SRA Decreased levels of VRA

5.3

Non-infected male genotypes

Same male genotypes infected by pistil smut

Figure 5.1 to Figure 5.3: Effect of pistil smut infection on buffalograss resource partitioning.

109

5.1, Non-infected buffalograss allocates substantial amounts of resources towards

vegetative growth of the plant, and therefore exhibits relatively high levels of VRA and

low levels of SRA. 5.2, Buffalograss infected with pistil smut allocates substantial amounts of resources towards sexual reproductive growth of the plant, and therefore exhibits relatively high levels of SRA and low levels of VRA. 5.3, Genotypes of male buffalograss infected with pistil smut exhibits earlier onset of flowering as compared to the same genotypes not infected.

110

Non-infected male Infected male 5.4

Tertiary

Secondary

Primary

Figure 5.4: Developmental changes in inflorescence architecture of male buffalograss induced by pistil smut.

111

Table 5.1: Mean resource partitioning for identical genotypes (clones) of either female or male sex forms of buffalograss that were infected or non-infected with the pistil smut fungus. Means are the averages of genotypes for two replications from 2005 summer harvest of greenhouse grown plants. Mean differences between non-infected and infected clones were tested for significance using Student’s paired t-test. Standard error = (SE).

Plant biomass per plant Allocation per plant Genotype sample Vege- Sexual Vege- Sexual size1 Total2 tative3 Reprod.4 tative 5 Reprod.6 Sex form (N) (gm) (gm) (gm) (%) (%)

Female non-infected 23 80.5 80.1 0.37 99.5 0.5 infected 23 47.8 39.4 8.97 81.2 18.8 mean difference 32.7*** 40.7*** -8.60*** 18.2*** -18.2*** (SE) (3.62) (3.85) (1.41) (0.03) (0.03) fold difference7 -0.41 -0.51 +23.2 -0.18 +36.6

Male non-infected 31 78.3 77.4 0.94 98.5 1.5 infected 31 46.5 40.9 5.53 86.6 13.4 mean difference 31.8*** 36.4*** -4.60*** 11.9*** -11.9*** (SE) (3.00) (3.19) (0.78) (0.02) (0.02) fold difference7 -0.41 -0.47 +4.88 -0.12 +7.93

112

Table 5.2: Mean resource partitioning for identical genotypes (clones) of either female or male sex forms of buffalograss that were infected or non-infected with the pistil smut fungus. Means are the averages of genotypes for two replications from 2006 summer harvest of greenhouse grown plants. Mean differences between non-infected and infected clones were tested for significance using Student’s paired t-test. Standard error = (SE).

Plant biomass per plant Allocation per plant Genotype sample Vege- Sexual Vege- Sexual size1 Total2 tative3 Reprod.4 tative 5 Reprod.6 Sex form (N) (gm) (gm) (gm) (%) (%)

Female non-infected 23 51.6 48.3 4.26 92.2 8.43 infected 23 35.8 27.8 8.00 77.3 22.7 mean difference 13.6*** 20.5*** -3.74** 14.9*** -14.9*** (SE) (4.82) (4.45) (1.38) (0.03) (0.03) fold difference7 -0.31 -0.42 +0.88 -0.16 +1.69

Male non-infected 31 40.7 37.7 2.93 92.4 7.59 infected 31 31.2 24.4 6.76 76.0 23.9 mean difference 9.46** 13.3*** -3.83*** 16.4*** -16.4*** (SE) (3.60) (3.53) (0.80) (0.03) (0.03) fold difference7 -0.23 -0.35 +1.31 -0.18 +2.15

113

Table 5.3: Mean estimated components of sexual reproductive allocation for identical genotypes (clones) of either female or male sex forms of buffalograss that were infected or non-infected with the pistil smut fungus Means are the averages of genotypes for two replications from 2006 summer harvest of greenhouse grown plants. Mean differences between non-infected and infected clones were tested for significance using Two- tail Student’s paired t-test. Standard error = (SE).

Estimated components of SRA

Genotype Inflore- Genotype Weight of 100 Seed Estimated sample scences sample Spikes Spikelets Florets Florets Ovary seeds or smut or sb Harvest size /plant size /inflor. /spike /spikelets /plant8 /plant balls (sb) yield9 index10 Sex form (N) (#) (N) (#) (#) (#) (#) (#) (mg) (gm) (%)

Female non-infected 23 16.4 21 2.91 2.71 1.00 129 129 78.3 (seed) 0.10 0.19 infected 23 49.5 21 3.43 4.95 2.02 1697 1697 64.4 (sb) 1.09 3.05 mean difference -33.1*** -0.52*** -2.24*** -1.02*** (SE) (5.85) (0.17) (0.68) (0.09) fold difference7 +2.01 +0.18 + 0.83 +1.02 +12.6 +12.6 -1.78 9.9 +15.1

Male non-infected 31 31.1 28 2.82 12.39 2.00 2173 0 0 0 ---- infected 31 64.7 28 3.29 11.61 2.88 7117 6761 40.0 (sb) 2.70 8.67 mean difference -33.7*** -0.46*** 0.79† -0.88*** (SE) (5.19) (0.12) (0.55) (0.06) fold difference7 +1.08 +0.15 - 0.08 +0.44 +2.3 ------

†, ** and *** significant at the 20%, 2% and 1% level of probability, respectively. 1 - Each genotype was vegetatively propagated into four replicate clonal plants, two replicate clones were infected with pistil smut and two remained non-infected. 2 - Total plant biomass = Vegetative biomass + Sexual reproductive biomass. 3 - Vegetative biomass = dry weight of non-flowering vegetative shoots and stems. 4 - Sexual reproductive biomass = dry weight of flowering shoots and stems. 5 - Vegetative Allocation = Vegetative biomass / Total plant biomass. 6 - SRA, Sexual Reproductive Allocation = Sexual reproductive biomass / Total plant biomass. 7 - Gain (+) or loss (-) for infected clones compared to non-infected clones measured as fold difference. 8 - # floret/plant = # inflorescence/plant x # spikes/inflorescence x # spikelet/spike. 9 - Seed or smut ball yield = # ovaries/plant x weight of a seed or smut ball. 10 - Estimated harvest index = seed or smut ball yield 9/ total plant biomass 2.

114

Chapter 6

RATIONALE FOR FUTURE STUDIES ON BUFFALOGRASS-PISTIL SMUT INTERACTION

The buffalograss-pistil smut system has several distinct advantages as a model system: 1) both pistil smut fungus and host buffalograss can be grown in large numbers,

2) an ability to vegetatively propagate buffalograss enables us to investigate and analyze paired samples of identical replicates of buffalograss genotypes, 3) unlike systems studying individual mutants, buffalograss-pistil smut interaction represents a system that allows us to examine the overall coordination of genes within complex polygenic traits, for example, SRA and various seed yield components, and 4) since buffalograss is a member of the grass family Poacea, results from buffalograss-pistil smut system are anticipated to have wide application to the world’s important agricultural grain and cereal crops

The present study provides potential molecular evidence for Ts2 involvement in the separation of sexes resulting in the dioecious sex expression of buffalograss. Tasselseed2

(Ts2) encodes for a short chain dehydorgenase reductase (SDR) protein with high sequence similarity to hydroxysteroid dehydrogenase (Delong et al. 1993). Ts2 and Ts2- like SDR proteins have also been identified from other grass species and even from some dicots like Silene latilofia, and Arabidopsis thaliana. Although DNA and amino acid sequences of these Ts2-like SDR proteins show high homology with maize Ts2, their expression patterns are different from that of maize Ts2 (Malcomber and Kellogg 2006).

Unlike maize Ts2, which expresses in the developing gynoecium of the florets destined to

115

male, Sta1 and ATA1, Ts2-like genes from S. latifolia and A. thaliana, respectively, expresses in the tapetal cells of the stamen (Lebel-Hardenack et al. 1997). This indicates that gynocial expression of Ts2 is novel to Panicoides and Chloridoides (FIG. 6.1), and

represents ectopic expression relative to ancestral tapetal expression (Malcomber and

Kellogg 2006). In addition to the reports of tapetal expression, Ts2 has also been shown

to express in plant parts other than floral organs, for example, culm, leaves, roots of

maize (Malcomber and Kellogg 2006) and leaves and nodes of buffalograss (Appendix

D). Ectopic expression of Ts2 (heterotopy) suggests that Ts2 has more general developmental role and is not just a sex-determining gene but rather it has been co-opted for sex determination function during the course of evolution. However, DeLong et al.,

(1993) showed that Activator (Ac)-tagged mutation of Ts2 locus in maize causes feminization of tassels, and Ac excision event resulted in the production of somatic revertant sectors, which is correlated with the reactivation of male developmental program. This suggests that Ts2 plays a pivotal role in determining the sexual fate of floral meristem.

In the present study, I show that pistil smut has an ability to down-regulate the expression of BdTs2 thereby inducing hermaphroditism in otherwise unisexual florets of male buffalograss. Pistil smut also, typically, induces various other secondary changes in nearly 100% of all infected male and female plants. However, the exact mechanistic pathway controlling these secondary changes is unknown. At first thought, hormones must be involved; either the fungus produces a hormone recognized by the plant or the fungus elicits a host response in the form of a phytohormone, which may either

116

negatively or positively influence the cascading sequence of events involved in floral and

inflorescence development (more details on Appendix E).

Fungi are known to produce hormones such as gibberellic acids (GAs) (Rojas et

al. 2001) and local tissue levels of GA in maize are known to determine whether or not

stamens are suppressed (Irish 1997). However, the application of near-lethal doses of GA

and GA-inhibitors has not been observed to induce hermaphroditism in buffalograss (Yin

and Quinn 1995). Nevertheless, the role of GA can not be ruled out because some GA

mutants of maize are also unresponsive to applied GA (Ross 1997). Applications of

Ethephon®, an ethylene hormone, has also not been observed to induce hermaphroditism

in buffalograss (Huff, unpublished). Basse (2002) recently suggested that tumor

formation in maize as well as increased flowering in A. thaliana is caused by a gene

product of Ustilago maydis fungus that shows similarity to a 5α-steroid reductase which

interferes with the brassinosteroid pathway in plants (Basse et al. 2002). Brassinosteroids

are a class of potent phytohormones that control plant growth and development, male

fertility, apical dominance, senescence and flowering (Altman 1998). Fungal

interference with the brassinosteroid pathway would therefore, seem to be a potential

candidate for, not only inducing pistils in male buffalograss florets, but also may explain

some of the additional secondary effects observed resulting from fungal infection.

The range of potential substrates for Ts2 encoded SDR enzyme is very broad, for example, sugars, alcohol, aromatic compounds, and gibberellins or steroid-like compound (brassinolide). Ts2 expression has been shown to be transcriptionally regulated by various hormone levels (Malcomber and Kellogg 2006). Rice Ts2 has been shown to be up-regulated in callus tissue exposed to GA and down-regulated when

117

exposed to ABA (a stress hormone). Based on these findings, another proposition can be

that pistil smut infection either directly or indirectly (as a result of stress associated with

infection) induces ABA signaling pathway in infected buffalograss, thereby, altering GA

levels and ultimately leading to the down-regulation of Ts2.

Various sterilizing pathogens, ranging from endophytes to smuts, are known to cause parasitic castration of their respective hosts where the parasite specifically targets and destructs the host gonads (gamete-producing organs) (Clay 1991). Resistance- virulence coevolution governs the rate at which hosts and parasites interact and their iterative process always begins in the host. However, it is difficult to imagine a coevolutionary process where the host is reproductively sterile and incapable of producing genetically variable progeny as a result of parasitic castration. Thus, the coevolutionary defense for a host against virulence by a parasitic castrator would ultimately depend on somatic mutation. Induced hermaphroditism in buffalograss is a unique form of parasitic castration in that both floral sex organs are sterile but for different reasons; ovaries (induced or not) are sterilized as a result of being supplanted by teliospore production while the stamens are spore free but sterile as a result of being underdeveloped. Based on floral organ source-sink relationships (Charnov 1982), I propose that stamens are underdeveloped in hermaphroditic florets because pistil smut sporulation creates such a strong sink within induced ovaries that insufficient resources

remain for normal stamen development. This line of reasoning is in agreement with

parasitic castration theory (Boudoin 1975, Clay 1991). Therefore, it seems reasonable

that if only one sex is parasitized, then any somatic host mutation that reallocates

118

resources to the opposite sex function (i.e. sex allocation) would have a selective

advantage.

Fossil evidence for parasitically induced sex change in hosts suggests a long term relationship between host and the castrator (Feldmann 1998). Assuming pistil smut was a

disease causing agent of the hermaphroditic ancestor of buffalograss, I reason that a mutation based selective advantage led to Ts2 evolving its function as a female suppressor gene by arresting pistil development within the hermaphroditic ancestor, thereby preventing fungal sporulation, and consequently reallocating resources towards normal stamen development, resulting in the male sex form. Thus, the pistil organ death programming of Ts2 would seem to benefit the ancestral host as a type of defense mechanism by eliminating pistil smut’s venue for horizontal transmission. Consequently, pistil smut would become entombed within these male plants, capable of persisting only through intercellular mycelial propagation (vertical transmission). Thus, host defense

provides selection pressure for pistil smut to coevolve an over-riding virulence

mechanism to down-regulate Ts2 thereby inducing female organs in male sex forms and hence restoring its capability for horizontal transmission (more detail in Appendix F).

During the above coevolutionary process, the selective advantage for Ts2 as a host defense mechanism would reduce the evolutionary cost (Pannell 2002,

Charlesworth 1984) for its female-suppressor function thereby facilitating unisexual

flower evolution. This reasoning lends support to the notion that parasitic castration

might have played an important role in the evolution of unisexual flowers in

buffalograss (Wilson 1983). This scenario is certainly debatable and alternative

explanations exist for the evolutionary origin of pistil smut induced

119

hermaphroditism in male buffalograss. However, I think we can test certain aspects

of this argument by creating an artificial androdioecious population by introducing

a pistil specific promoter-Ts2 transgene into an hermaphroditic grass-systemic bunt

(Tilletia sp.) system to determine a) if Ts2 prevents fungal sporulation, by removing

the pistil and, b) if male function is retained. Regardless of the evolutionary origin,

I demonstrate that pistil smut inhibits Ts2 pistil cell death programming and removes a sex determining ontogenetic layer from the evolutionary hermaphroditic precursor in unisexual male florets.

Broader implications of this study include: 1) Results showing the pathological regulation of a cell death gene causing anatomical changes in buffalograss may provide a hint for developmental mechanisms underlying sex altering effects of parasitic castration in other plants as well as animals species. 2) This study suggests that secondary effects of parasitic castration including behavioural, anatomical and functional modifications are constrained by evolutionary history of the host and that the removal of various evolutionary layers by the pathogen leads to the reversion of host to a more primitive form. 3) Finally this study suggests that pathological influence can potentially serve as a strong selective force in the host for the evolution of separate sexes as a means to escape parasitic virulence, rather than solely as a means to increase diversity to optimize resource allocation per se.

120

BIBLIOGRAPHY

Charnov EL. 1982. The Theory of Sex Allocation. Princeton University press, New

Jersey.

Altman T. 1998. A tale of dwarfs and drugs: brassinosteroids to the rescue. Trends Genet.

14: 490-495.

Basse CW, Kerschbamer C, Brustmann M, Altmann T and Kahmann R. 2002. Evidence

for a Ustilago maydis steroid 5α-reductase by functional expression in Arabidopsis

det2-1 mutants. Plant Physiol 129: 717-732.

Boudoin M. 1975. Host castration as a parasitic strategy. Evolution 29 (2): 335-352.

Charlesworth D. 1984. Androdioecy and the evolution of dioecy. Biol. J. Linn. Soc. 23:

333-348.

Clay K. 1991. Parasitic castration of plants by fungi. Trends Ecol. Evol. 6: 162-166.

Delong A, Calderon-Urrea A and Dellaporta SL. 1993. Sex determination gene

TASSELSEED2 of maize encodes a short-chain alcohol dehydrogenase required for

stage-specific floral organ abortion. Cell 74: 757-768.

Feldmann RM. 1998. Parasitic castration of the crab, Tumidocarcinus giganteus

glaessner, from the Miocene of New Zealand: Coevolution within the crustacean. J

Paleontology 72 (3): 493-498.

Irish EE. 1997. Class II tassel seed mutations provide evidence for multiple types of

inflorescence meristems in maize (Poaceae). Am J Bot 84 (11): 1502-1515.

121

Lebel-Hardenack S, Ye D, Kountnikova H, Saedler H, Grant SR. 1997. Conserved

expression of a TASSELSEED2 homolog in the tapetum of the dioecious Silene

latifolia and Arabidopsis thaliana. Plant Journal 12:515-526.

Malcomber ST and Kellogg EA. 2006. Evolution of unisexual in grasses (Poacea) and the

putative sex determining gene TASSELSEED2 (Ts2) New Phytologist 170:885-899.

Pannell JR. 2002. The evolution and maintenance of androdioecy. Annu. Rev. Ecol. Syst.

33: 397–425.

Rojas MC, Hedden P, Gaskin P and Tudzynski B. 2001. The P450–1 gene of Gibberella

fujikuroi encodes a multifunctional enzyme in gibberellin biosynthesis. Proc Natl Acad

Sci U. S. A. 98: 5838–5843.

Ross JJ, Murfet IC and Reid JB. 1997. Gibberellin mutants. Plant Physiol 100: 550-560.

Wilson MF. 1983. Plant reproductive Ecology. John Wiley & Sons, New York.

Yin T and Quinn J. 1995. Tests of a mechanistic model of one hormone regulating both

sexes in Buchloe dactyloides (Poaceae). Am J Bot 82:745-751.

122

Cucumis sativus_AF286651 6.1 Arabidopsis thaliana_AY082345 Leersia virginica_DQ384230 Muhlenbergia sobolifera_DQ384248 Bouteloua hirsuta_AY434538 Bouteloua hirsuta_AY434527 Bouteloua hirsuta_AY434525 Buchloe dactyloides_DQ457002 Bouteloua dimorpha_AY426297 Chloridoideae Bouteloua dimorpha_AY426296 Bouteloua dimorpha_AY426295 Bouteloua trifida_AY426327 Bouteloua hirsuta_AY434539 Eragrostis tef_DQ384249 Spartina pectinata_DQ384250 Ixophorus unisetus_DQ384251 Ixophorus unisetus_DQ384252 italica_DQ384253 Sorghum bicolor_DQ384254 Panicoideae Tripsacum dactyloides_U89270 Zea mays_L20621 Orthoclada laxa_DQ384255 Zizania aquatica_DQ384231 Zizania aquatica_DQ384232 Avena sativa_DQ384235 Brachypodium distachyon_DQ384236 Danthonia spicata_DQ384261 Oryza sativa_AK070356 Hordeum vulgare_DQ384262 Monocots Orthoclada laxa_DQ384263 Brachypodium distachyon_DQ384237 Psathyrostachys juncea_DQ384238 Psathyrostachys juncea_DQ384239 Melica altissima_DQ384240 Melica altissima_DQ384241 Lithachne humilis_DQ384242 Lithachne humilis_DQ384243 Olyra latifolia_DQ384244 Bambusoideae Pariana radiciflora_DQ384245 Pariana radiciflora_DQ384246 Otatea acuminata_DQ384247 Ehrharta erecta_DQ384233 Ehrharta erecta_DQ384234 Pharus lappulaceus_DQ384258 Pharoideae Pharus lappulaceus_DQ384259 Anomochloa marantoidea_DQ384256 Anomochloa marantoidea_DQ384257 Anomochloa marantoidea_DQ384260 Eragrostis tef_DQ384228 Flagellaria indica_DQ384229 Pennisetum glaucum_DQ384225 Pharus lappulaceus_DQ384226 Otatea acuminata_DQ384227 Orthoclada laxa_DQ384223 Chasmanthium latifolium_DQ384222 Setaria italica_DQ384224 Citrullus lanatus_AB018559 Silene latifolia_U53828

50 substitution per site

Figure 6.1: One of the most parsimonious tree displaying relationship of 58 Tasselseed2

(Ts2) and related short-chain dehydrogenase/reductase (SDR) genes.

123

MP tree is generated from PAUP version 4.0 using General Time Reversible (GTR) model, invariant sites and gamma distribution (GTR+I+G). (●) represent taxa for which the role of Ts2 as a sex determining locus has been described.

124

Chapter 7

SUMMARY AND CONCLUSION

The present study is the most comprehensive study conducted to date providing molecular, evolutionary and biochemical understanding of induced hermaphroditism caused in dioecious buffalograss as a result of parasitic castration by pistil smut. I show that the pistil smut fungus induces hermaphroditism in not only male sex forms of buffalograss by inducing the development of otherwise vestigial pistils, but also in female sex forms by inducing hypertrophy of otherwise vestigial stamens (male sex organs).

Both male and female sex organs within these induced hermaphroditic florets are parasitically castrated but apparently for different sex-specific reasons. In addition to inducing hermaphroditism, the fungus also induces the development of additional pistillate florets in both infected male and female plants. Phylogenetic analyses based on maximum parsimony, maximum likelihood and genetic distance of rDNA sequences show that pistil smut exhibits an accelerated rate of nucleotide substitution, which places pistil smut outside the current taxonomic circumscription of genus Tilletia suggesting that pistil smut is unlike any other species of Tilletia. Because insufficient evidence was displayed to accommodate pistil smut within the genus Tilletia, we proposed a new Latin binomial combination Salmacisia buchloëana to describe pistil smut, the only species known to induce hermaphroditism in order Tilletiales.

A mechanism describing how parasites induce hermaphroditism in their hosts is unknown. In order to understand how pistil smut induces hermaphroditism in

125

buffalograss, I first demonstrate through scanning electron micrographic images that florets begin development as bisexual hermaphrodites, initiating both male and female

organ precursor. Unisexual florets are then formed by the selective abortion of opposite

sex organs within hermaphroditic florets. In general, this pattern of unisexual floret development is known as a Type 1 form of sexual dimorphism and is characterized by the vestigial organ remains of the opposite sex within unisexual florets (Mitchell and Diggle

2005). Moreover, I provide the first corollary evidence for the role of a putative female- suppressor gene, homologous to Tasselseed2 (Ts2), in buffalograss unisex male floret development. Temporal and spatial expression studies of the cloned full length BdTs2 showed high expression levels of BdTs2 RNA in the sub-epidermal layer of gynoecium within would-be staminate florets, and is correlated with the degeneration of the gynoecium. This suggests that, like in maize, the development of staminate florets in buffalograss is under genetic control of Ts2 as a regulator of programmed cell death causing vacuolization and the loss of cellular integrity of the sub-epidermal cells of gynoecium (Cheng et al. 1983). Sex expression of buffalograss is environmentally stable and no sex chromosomes have been reported in buffalograss. However, my results provide the first report suggesting a very strong correlation between the expression patterns of BdTs2 with the male sex developmental program in buffalograss. A common sex-determination mechanism for pistil abortion has been suggested for species within the maize tribe Andropogoneae and possibly even the maize subfamily Panicoideae (Le

Roux and Kellogg 1999). This shared trait adds to other sex determination studies which

have important evolutionary and mechanistic implications for understanding the forces of

natural selection that resulted in unisexual flower evolution.

126

Here I show that pistil smut induces hermaphroditism in male buffalograss by

down-regulating a putative female-suppressor gene, BdTs2, homologous to maize

Taseelseed2 (Ts2). I found that pistil smut infection perturbs the host’s sex determining

pathway by removing this ontogenetic layer from unisexual male floret formation thereby

permitting the development of a sex which the plant is genetically programmed not to

exhibit. In other words, down-regulation of BdTs2, presumably, prevents BdTs2

mediated cellular arrest of pistil development and allows the unconstrained development

of vestigial gynoeciums within florets that were otherwise destined to become unisexual

male. I also show that the infection by the fungus reallocates plants resources from being

utilized for vegetative growth towards a dramatic increase in ovary production thereby enhancing the potential for overall seed yield. This study provides the first evidence for

parasitically induced hermaphroditism and for parasitically enhanced sexual reproductive

allocation (SRA).

Present research has shown that parasitic castration of buffalograss by the pistil

smut fungus in the form of induced hermaphroditism not only results in alteration of sex

expression but also manipulates resource allocation and life history traits of the host. As

such, pistil smut is permitting expression of an ancestral-like buffalograss;

hermaphroditic and annual-type phenotype that presumably existed back in time in the

form of retrospective phenotype or retrophenotype.

127

BIBLIOGRAPHY

Cheng PC, Greyson RI and Walden DB. 1983. Organ initiation and the development of

unisexual flowers in the tassel and ear of Zea mays. Am. J. Bot. 70(3): 450-462.

Le Roux LG and Kellogg EA. 1999. Floral development and the formation of unisexual

spikelets in the Andropogoneae (Poaceae). Am. J. Bot. 86: 354-368.

Mitchell CH and Diggle PK. 2005. The evolution of unisexual flower: Morphological and

functional convergence results from diverse developmental transition. Am. J. Bot.

92(7): 1068–1076.

128

Appendix A

PISTIL SMUT INFECTION OF BUFFALOGRASS RELATIVES BELONGING TO PACC CLADE UNDER CONTROLLED CONDITIONS OF GREENHOUSE

An attempt to infect six members of PACC clade including, Zea mays, Sorghum

bicolor, Tripsacum dactyloides, Bouteloua hirsuta, B. gracilis, and B. curtipendula under

controlled conditions like high humidity and high temperature in greenhouse. Pistil smut

teliospores were embedded in the centre of potting soil (Promix®, Premier Horticulture,

Inc., Quakertown, PA) contained within clay pots (15.25 cm diameter) in the greenhouse

(26º C day, 21º C night) under natural day length conditions. Well-watered pots were

then kept sealed in opaque white plastic bags in order to maintain high humidity and high temperature conditions, for about 1-week to allow the teliospores to absorb water and

initiate germination. Seeds of all six grasses were then sowed into the soil separately

along with embedding more teliospores into the soil surface close to the seeds. Pots were

again re-sealed in the bags, for about 6-weeks to allow the fungus enter the plants.

An attempt to infect maize, sorghum, Tripsacum, and Bouteloua spp. under

controlled conditions of greenhouse failed. No secondary effects of pistil smut infection

were observed on these close relatives of buffalograss suggesting that pistil smut

infection is host specific with a narrow host range.

129

Appendix B

SIZE MEASUREMENT OF SEX ORGAN WITHIN UNISEXUAL MALE AND FEMALE BUFFALOGRASS FLORETS

Dioecious buffalograss has male and female florets on separate individuals. Each

buffalograss floret, both male and female, initiates as a perfect floret with both male and female sex organs developing in the same flower structure. Unisexual florets are formed by the selective abortion of opposite sex organs as the developing hermaphroditic floret reaches the gynoecial ridge stage of development. Selective abortion is characterized by the vestigial organ remains of the opposite sex while the attending sex organ is allowed to develop to full mature size within unisexual florets (TABLE B.1).

130

TABLE B.1: Means, standard deviations (std), and fold differences of morphological characteristics between pistil smut infected and non-infected genotypes of buffalograss.

Mean ± std

Sample Pistil Ovary Anther Filament size length† width length length Sex form (N) (mm) (mm) (mm) (mm)

Male non-infected 11 vestigial vestigial 1.79±0.36 0.73±0.05 infected 11 3.32±0.37 0.75±0.18 1.26±0.38 0.53±0.07 fold diff. - - -0.30*** -0.27***

Female non-infected 4 5.96±0.48 0.61±0.13 vestigial vestigial infected 4 6.09±0.67 0.95±0.17 0.80±0.42 0.34±0.05 fold diff. ns‡ +0.55*** - -

† Pistil length represents the summation of ovary, style and stigma length. ‡ ns, non-significant and *** significant at the 1% level of probability.

131

Appendix C

SOUTHERN BLOTTING TO DETERMINE TS2 COPY NUMBER IN DIPLOID BUFFALOGRASS

Copy number of Ts2 within diploid (2n=20) race of buffalograss was estimated

using PCR and preliminary confirmation was done by Southern blot analysis.

Approximately 15 µg of total genomic DNA from six buffalograss male genotypes was

extracted using CTAB and then digested with BamHI, and EcoRI at 37º C for overnight.

Digested DNA was separated on 0.8% agarose gel, blotted onto Hybond N+ nylon membrane (Amersham) under alkaline conditions. Ts2 probes were 32P-dCTP labeled

using random priming method using Prime-It II Random Primer kit, (Stratagene, La Jolla,

CA) according to the manufacture’s directions, and were hybridized onto the nylon membrane carrying digested buffalograss DNA for 16 h at 65º C. After hybridization, blots were washed at 65º C twice in 2 X saline sodium citrate (SSC) and 0.5% sodium dodecyl sulfate (SDS) for 20 min each, and twice with 0.1 X SSC/0.1% SDS.

Figure C.1 with six lanes; first three lanes are three male genotypes digested with

BamHI, and next three lanes are same three male genotypes digested with EcoRI,

According to NEB-cutter, these two restriction enzymes do not have restriction site within the coding region of BdTs2. Blot shows that there are two bands in each lane, presumably because BdTs2 is heterozygous at this locus. This is reasonable because according to Westergaard (1958), for 1:1 dioecious sex expression, one sex form either has to be heterogametic (XY chromosomes) or heterozygous at the sex determining loci.

Further study needs to be done with male as well as female genotypes digested with

132 restriction enzymes flanking BdTs2 and with the ones that cut the gene within the coding region. Higher ploidy levels (tetraploid and hexaploid) should also be investigated to confirm the copy number of the gene.

BamH1 EcoR1 C.1

M1 M2 M3 M1 M2 M3

Figure C.1: Southern blot of diploid buffalograss genome with BdTs2.

First three lanes are three male genotypes, namely, M1, M2 and M3, digested with

BamHI, and next three lanes are same three male genotypes digested with EcoRI.

133

Appendix D

BDTS2 EXPRESSION IN NODES AND LEAVES OF BUFFALOGRASS

Quantitative real-time PCR was performed on various different stages of

inflorescence development as well as on leaves and nodes of both male and female sex

forms of buffalograss, infected and non-infected with pistil smut. Figure D.1 shows the

BdTs2 expression levels in leaves and nodes of buffalograss suggesting that Ts2 might

have a more general role of in development than specifically sex determination.

2000 D.1 1800

1600

1400

1200

1000

800

600

Relative Ts2 mRNA expression levels expression Ts2 mRNA Relative 400

200

0 Nodes Leaves Nodes Leaves Nodes Leaves Nodes Leaves Plant tissue type Male Female Infected male Infected female

Figure D.1: BdTs2 expression in leaves and nodes of male and female buffalograss infected with pistil smut, as determined by quantitative real-time PCR.

134

Appendix E

PRELIMINARY METABOLOMIC ANALYSIS AND BIOASSAY TO ISOLATE THE TRIGGER COMPOUND FROM BUFFALOGRASS UPON PISTIL SMUT INFECTION

Metabolomic analysis.— Samples of non-infected and fungal infected

buffalograss inflorescences, along with a fungal culture frowing on potato dextrose media

were homogenized in liquid nitrogen in order to isolate compounds produced by the

fungus upon infection or compounds produced by the plant in response to fungal

infection. Homogenized samples were extracted using polar (water) and non-polar

solvents (methanol, acetone) for 10-12 hrs using a Soxhlet extraction apparatus.

Resulting extracts were analyzed using high performance liquid chromatography-mass

spectrometry (HPLC-MS). The extracts were separated using reverse-phase

chromatography on an octadecylsilyl stationary phase and gradient elution with a

water/acetonitrile/ formic acid mobile phase. Elute was ionized via electrospray interface

and the analytes mass was determined with a time-of-flight mass spectrometer.

The complexity of the total ion chromatogram did not allow for the confident

determination of composition differences in the samples, thus narrower mass ranges were

viewed. FIGURE E.1 shows chromatograms with a mass range of 775-800amu for each of

the three samples. As can be seen in the figure, a significant peak at 14.66 minutes (*) is

present in the sample of the infected buffalograss, but not present in the non-infected

buffalograss or fungal culture. The mass spectra of the peak were extracted from the chromatograms (FIG. E.2). Although, only one observed chemical difference has been

135

provided in this example, at least two dozen differences in the composition were observed from this preliminary work.

Preliminary bioassay.— Infected plant tissue was homogenized in order to isolate compounds produced by the fungus upon infection along with compounds produced by the plant in response to fungal infection. Pistil smut growing on the artificial media was also homogenized in liquid nitrogen and extracted using polar (water) and non-polar solvents (methanol, acetone) for 10-12 hrs using a Soxhlet extraction apparatus. Before bioassay, the extract was evaporated and the dried extract was re-suspended in 4 times volume of distilled water along with a penetrating adjuvant/surfactant (ex. Liberate®) to provide uniform coverage of spray solution. Ammonium sulfate was also added at low rates to improve penetrance. Extracts were applied onto uninfected plants (5 plants per

extract), as sprays or drenches near the base of the plant, to stimulate secondary effects

such as induced hermaphroditism, additional florets, increased flowering, etc., in the

absence of the fungus.

Results of bioassay were inconclusive. No secondary effects were observed. The

original proposal was if secondary effects are observed then these crude extracts would

then be fractionated into subsets based on their biochemical properties. For example,

crude extracts may initially be fractionated based on carbon content with the high carbon

fractions representing lipids, low carbon fractions representing sugars, and the

intermediate carbon fractions representing secondary metabolites. Resulting fractions

would again be assayed for biological activity and subjected to further fractionation. This

iterative process would have continued until an active compound, or compounds, have

136 been identified. If no active products were identified, then an attempt would have been conducted to isolate and extract fungal membrane tissue of infected plants in an attempt to capture membrane-bound, heat sensitive proteins. Mass spectrometry fractionation would have been performed at the PSU Huck Institute Mass Spectrometry Facility which is capable of 200-fold fractionation.

137

E.1 Non-infected

Infected

Fungal culture

Figure E.1: Chromatographs with a mass range of 775-800 amu extracted from acetone extracts of non-infected (top) and infected infected males, (middle) male buffalograss plants and pure cultures of pistil smut fungus (bottom). A significant peak at 14.66 minutes (*) is unique to infected plants.

138

E.2 Non-infected

Infected

Fungal culture

Figure E.2: Mass spectra of the significant peak 14.66 from the three spectra: non- infected males, infected males and pure fungal cultures.

139

Appendix F

PARASITIC CASTRATION AND THE EVOLUTION OF DIOECY IN BUFFALOGRASS

We theorize that the induced hermaphroditism generated through buffalograss- pistil smut interaction is a reversion back to a more primitive morphological state (a so- called “retrophenotype”). The following is our logic to support this theory. Let us assume that the pistil smut fungus was also a disease causing agent of the hermaphroditic

ancestor of buffalograss. The effects of pistil smut disease in ancient hermaphrodites

would presumably be similar to those we have observed in induced hermaphrodites, that

is, a parasitize pistil and reduced stamens. An analogous situation can be observed in

smut (Tilletia indica Mitra) infected hermaphroditic wheat (Triticum aestivum L.) (Clay

1999). The process of sporulation within infected ovaries of either ancient/induced hermaphrodites likely creates such a strong sink that essentially all available resources

within the floret are allocated to the female sex organ with little to none remaining for the

development of male sex organs. Thus, the pistil smut regulatory control over sex

allocation within hermaphroditic florets reduces male sex organ size, function, and hence

reproductive fitness i.e. a total parasitic castration of both male and female reproduction

not unlike that of the white campion-anther smut system but for the opposite reasons.

Interestingly, at a population level, anther smut infected campions exhibit androdioecy

(hermaphrodite and male plants) whereas pistil smut infected buffalograss represents hermaphroditic population; however, in both systems, infected plants are reproductively sterile.

140

Therefore, it seems reasonable that if only one sex is parasitized in an hermaphrodite then any host mutation that provides allocation to the opposite sex would have a selective advantage. As such, any mutation that would abort pistils from ancestral hermaphroditic florets of buffalograss, like Ts2, would presumably have a selective advantage by enabling the male sex to survive and function normally even though the plant was infected with pistil smut. Such suppressor mutations are also a necessary starting point for the evolution of separate sexes. Thus, a newly arisen Ts2 female- suppressor mutation would yield an all male plant that would be resistant to pistil smut castration and, when crossed to a susceptible hermaphrodite, would sire 1:1, male

(resistant) to hermaphrodite (susceptible), segregant progeny. Assuming a continued selective advantage of such resistant Ts2 male plants, perhaps including a fortuitous increase in sexual reproductive allocations due to pistil smut infection, further incorporation of Ts2 into the population would eventually develop an androdioecious sexual system (male and hermaphroditic plants), a system whose evolutionary origins currently remain obscure.

To date, our observations and results suggest that buffalograss has a similar genetic mechanism of unisexual floret development as that of maize and eastern gamagrass. Therefore, in order to progress from this point forward to the evolution of dioecy in buffalograss, we will apply the maize model of unisexual flower development

(Delong et al. 1993) by hypothetically including two additional genes, neither of which has been sequenced; namely, Silkless (Sk1) and Dwarf1 (D1). In maize, Ts2 is epistatic to the Sk1 locus, a gene which protects a proportion of maize pistils from Ts2’s cell death programming resulting in unisexual female florets. Given its spatially-expressed,

141

epistatic control over Ts2 in maize, we reason that the original Sk1 gene would had to have established within a Ts2 genetic background. Thus, incorporation of a Sk1-like gene into a Ts2 resistant male buffalograss would give rise to individuals bearing both male and hermaphroditic florets (andromonoecy) capable of producing seed. In maize, the dwarf1,2,3,5,8 and anther-ear genes are involved in the gibberellin (GA) biosynthesis pathway and each eliminates male stamens from early developing hermaphroditic florets resulting in unisexual female florets (Irish et al. 1994). Similarly, GA mutants have been hypothesized in buffalograss to have interacted with the grazing pressure of herbivores, like the American bison, giving rise to the short-stature height of female inflorescences

(Quinn et al. 1994) and the edible female seed-capsule. Thus, we further speculate that a

GA biosynthesis gene, similar to D1 in maize, eliminated male function of hermaphrodite florets creating unisexual female florets which then became highly modified through coevolution with the herbivore, American bison, to produce the monoecious individuals of buffalograss we observe today. Due to its subdioecious tendencies (Huff and Wu

1992), we believe buffalograss evolved dioecy directly through a monoecious pathway identical to that described by Charlesworth and Charlesworth (1978), involving disruptive selection of at least one major gene followed by that of several additional genes of minor effect, that partition unisexual florets onto different individual plants in order to avoid inbreeding depression. Throughout this period of dioecy evolution, the fungus would have lost approximately 50% of its host range because unisexual male florets would be basally resistant to fungal sporulation. Even andromonoecious and monoecious plants would have exhibited partial basal resistance to pistil smut virulence.

Thus, host resistance provides a selection pressure on the fungus to evolve an over-riding

142

virulence mechanism by down-regulating Ts2 cell death programming within resistant male florets making them once again susceptible to fungal sporulation. Therefore, we believe that sometime before the spread of hexaploid dioecious buffalograss across the

North American Great Plains, that the pistil smut fungus had already regained its teliospore producing ability by inducing pistil development in unisexual male florets. In doing so, the fungus reverts male florets back to a more primitive morphological state resurrecting the hermaphroditic ancestor from which buffalograss evolved dioecy. We propose using the term “retrophenotype” to describe the reversion of host morphology back to a more primitive state by a parasite (FIG. F.1).

143

BIBLIOGRAPHY

Charlesworth B and Charlesworth D. 1978. A model for the evolution of dioecy and

gynodioecy. Am. Nat. 112:975-997.

Clay K. 1991. Parasitic castration of plants by fungi. Trends Ecol. Evol. 6:162-166... 140

Delong A., Calderon-Urrea A. and Dellaporta SL. 1993. Sex determination gene

TASSELSEED2 of maize encodes a short-chain alcohol dehydrogenase required for

stage-specific floral organ abortion. Cell 74:757-768.

Huff DR and Wu L. 1992. Distribution and inheritance of inconstant sex forms in natural

populations of dioecious buffalograss (Buchloë dactyloides). Am. J. Bot. 79(2): 207-

215.

Irish EE, Langdale JA and Nelson TM. 1994. Interaction between tasselseed genes and

other sex determining genes in maize. Developmental Genetics. 15: 155-171. 2

Quinn JA, Mowrey DP, Emanuele SM and Whalley RDB. 1994. The “Foligae is the

fruit” Hypothesis: Buchloe dactyloides (Poaceae) and the of North

America. Am J Bot. 81 (12): 1545-1554.

144

F.1 KEY Vestigial pistil Anther Hermaphrodite (bisexual) Male = stamens Filament (3) Stigma Female = pistil Style Pistil smut infection Ovary Vestigial stamen

Hermaphrodite = Bisexual Parasitic castration of female sex and reduced male function Teliospore filled Fungal smut ball = ovary and reduced stamens Host resistance – Ts2

Androdioecy – male and hermaphrodite plants

Resistant males Susceptible hermaphrodites

Silkless (Sk1) evolves spatial, epistatic control over Ts2 protecting female pistils (Calderon-Urrea A and Dellaporta SL 1999) in a proportion of florets of resistant male plants.

Andromonoecy – male and hermaphrodite florets

GA biosynthesis genes eliminate male function (Irish EE 1997) in hermaphrodite florets creating unisexual female florets which become highly modified through coevolution with the herbivore, American bison (Quinn JA et al. 1994)

Monoecy – dimorphic unisexual florets

Disruptive and density-dependent selection partitions unisexual florets onto different plants to avoid inbreeding depression (Charlesworth B and Charlesworth D 1978)

Dioecy – dimorphic unisexual individuals

Pistil smut infection overrides host resistance by down-regulating Ts2

“Retrophenotype” of ancestral hermaphrodite

Figure F.1: Model for parasitic castration and the evolution of dioecy in buffalograss.

145

VITA

Ambika Chandra

Personal Information Name: Ambika Chandra Present position: Assistant Professor, Turfgrass Breeding and Molecular Genetics Address: Texas A&M Research and Extension Center, Dallas 17360 Coit Road, Dallas, TX 75252 Contact: Email – [email protected] Phone – 972-231-5362 Fax – 972-952-9216

Education Ph.D. The Pennsylvania State University, USA – Agronomy, 2007 M.Sc. Punjab Agricultural University, India – Plant Breeding and Genetics, 2003 B.Sc. Punjab Agricultural University, India – Agriculture (Hons.), 2001

Awards and distinctions: o Recipient of university merit scholarship during undergraduate program o Recipient of George Hamilton Fellowship Award ($5000), for the year 2004-2005 o Poster entitled “Fungal Induced Sex Change in Male Buffalograss Mimics Tasselseed-2 Mutants” awarded 1st position at the annual meeting of ASA-CSSA- SSSA, Seattle, 2004 o Poster entitled “Fungal Induced Sex Change in Male Buffalograss Mimics Tasselseed-2 Mutants” awarded 2nd position at CAS graduate student poster competition, The Penn State University, 2005 o Poster entitled “Fungal Induced Sex Change in Male Buffalograss Mimics Tasselseed-2 Mutants” awarded 4th position at Gamma Sigma Delta sponsored poster competition, The Penn State University, 2005, and o Awarded $2000 from the CAS Graduate Student Competitive Grant Program on a research proposal titled “RNA Silencing of a Sex Determining gene, Tasselseed2, in Dioecious Buffalograss” , 2006. Research publications: 9 Chandra, A., Gupta, M.L., Ahuja, I., Kaur, G., Banga, S.S. (2004) Intergeneric hybridization between Erucastrum cardaminoides and two diploid crop Brassica species. Theor. Appl. Genet. 108:1620–1626. 9 Chandra A, Gupta M.L, Banga S.S, Banga S.K (2004) Production of an interspecific hybrid between Brassica fruticulosa and B-rapa. Plant Breeding. 123 (5): 497-498.