DEVELOPMENTAL BIOLOGY OF STYER GROUP

POLLEN AS RELATED TO HAPLOID INDUCTION

by

TSAN-YU CHIU

BSc. (Horticultural Science), Taiwan, 2003

A THESIS SUBMITTED IN PARTIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

in The FACULTY OF GRADUATE STUDIES ( Science)

The University of British Columbia

August, 2006

© Tsan-Yu Chiu, 2006 ABSTRACT Exacum Styer Group, a group of interspecific hybrids, possesses several valuable traits for ornamental use. However, current production techniques are limited to asexual reproduction of selected genotypes. Typically, inbred line development is highly desirable for commercial introduction of Fl hybrids as well as to facilitate genetic research. Unfortunately, this group suffers from severe inbreeding depression, precluding sexually derived inbred lines. In order to avoid inbreeding depression, double haploid plant development is one alternative approach. Therefore, this research developed fundamental information related to the successful application of this technology. Initially, fourteen genotypes were evaluated for pollen viability ranging from 0% to 80.5%. Three genotypes, with different levels of pollen viability, were selected and characterized for microsporogenesis, microgametogenesis, pollen morphology, and chromosome number. Microsporogenesis was normal among all three selected genotypes. However, the low fertility genotype and sterile genotype did not complete normal microgametogenesis and were found to have abnormal exine structure, perhaps indicating a dysfunctional tapetum layer. In contrast, the fertile genotype completed microgametogenesis, produced normal exine, and produced functional pollen grains. At anthesis, pollen was shed at the bi-nucleate stage. The chromosome numbers of the three genotypes evaluated ranged from 50 (fertile genotype) to 66 (low pollen viability). The individual chromosomes ranged in size from 0.3 um (dot-shaped) to 2.65 um (rod-shaped) with all three genotypes containing each form.

Following characterization of pollen development, reprogramming treatments (e.g., temperature, mannitol, media composition and plant growth regulators) to induce haploid embryogenesis were applied at the mid uninucleate to early binucleate stage. After cold treatment (10°C) for 7 days, microspores maintained normal nuclei development without symmetrical divisions observed, suggesting that no androgenic switch occurred. Furthermore, heat treatment (35°C) for 4 days induced nuclei degradation and microspore non-viability. In addition to temperature treatments, mannitol treatments did not induce symmetrical divisions. When anthers were cultured under common temperature treatments, regardless of media composition, nuclei displayed a similar response indicating temperature was more effective in influencing nuclei development than media

ii composition. Three different auxin and cytokinin combinations were evaluated on androgenic calli/embryo induction. However, none of the combinations successfully induced calli/embryo formation.

iii TABLE OF CONTENTS

TABLE OF CONTENTS iv LIST OF TABLES vi LIST OF FIGURES viii LIST OF ACRONYMS xii ACKNOWLEDGMENTS xiii

Chapter 1. General introduction and literature review 1 1.1 General introduction 1 1.2.1 Exacum in Sri Lanka 3 1.2.2 Exacum species and horticultural potential 4 1.3 Pollen Biology 5 1.3.1 Male gametophyte development 5 1.3.2 Male sterility 6 1.3.3 Androgenesis 7 1.4 Factors related to successful androgenic induction 8 1.4.1 The natural attributes of the donor plant 8 1.4.1.1 Genotype 8 1.4.1.2 Pollen developmental stage 8 1.4.2 Appropriate microspore reprogramming treatments 9 1.4.2.1 Effect of temperature shock on androgenesis 9 1.4.2.2 Effect of mannitol stress on androgenesis 9 1.4.3 Culture environment 10 1.5 Justification and objectives 11

Chapter 2. Characteristics of pollen development in Exacum Styer Group 19 2.1 Introduction 19 2.2 Material and Methods 20 2.2.1 Plant material 20 2.2.2 In vivo pollen viability 21 2.2.3 Microsporogenesis with acetocarmine stain 21 2.2.4 Microgametogenesis with DAP I stain 22 2.2.5 Pollen morphology by scanning electron microscopy (SEM) 22 2.2.6 In vitro pollen hydration 23 2.2.7 Chromosome counts with Giemsa stain 23 2.3 Results 23 2.3.1 Pollen viability 23 2.3.2 Microsporogenesis and microgametogenesis 24 2.3.3 Pollen characteristics by scanning electron microscopy (SEM) 25 2.3.4 In vitro pollen hydration 25 2.3.5 Chromosome counts 26 2.4 Discussion 26

Chapter 3. Stressor effects on in vitro development of microspores from a fertile

iv genotype of Exacum Styer Group 44 3.1 Introduction 44 3.2 Material and Methods 47 3.2.1 Plant materials and tissue collection 47 3.2.2 Media preparation 48 3.2.3 Disinfestation and culture procedures 48 3.2.4 Reprogramming treatments 48 3.2.5 Cytological examinations 49 3.2.6 Statistics 49 3.3 Results 50 3.3.1 Effects of temperature and media composition on microspore development 50 3.3.2 Effects of temperature and plant growth regulator combinations on colli formation 50 3.4 Discussion 51

Chapter 4. Concluding remarks and future research 61

Appendix I Characteristics of pollen development and chromosome numbers from additional genotypes in Exacum Styer Group 66 AL 1 Introduction 66 Al. 2 Material and Methods 66 Al. 1.2.1 Plant materials 66 Al. 1.2.2 Microsporogenesis and microgametogenesis with acetocarmine stain 67 ALL2.3 Chromosome counts with Giemsa stain 67 Al. 3 Results and Discussion 68

Appendix II Stressor effects on in-vitro development of microspores from a sterile genotype of Exacum Styer Group 73 AIL 1 Introduction 73 AH. 2 Material and Methods 73 AIL 2.1.1 Plant materials and tissue collection 73 AH. 2.1.2 Media preparation 74 AH. 2.1.3 Disinfestation and culture procedures 74 AIL 2.1.4 Reprogramming treatments 75 All. 2.1.5 Cytological examinations 75 AIL 2.1.6 Statistics 75 AH. 3 Results and Discussion 76

References 78 LIST OF TABLES

Table 1.1 Summary of taxonomic treatments for Sri Lankan Exacum 12

Table 1.2 Plant growth regulator combinations used in anther/microspore culture induction media 13

Table 2.1 Parentage of 14 Exacum Styer Group genotypes; by conversion, pistillate parent is listed first followed by the staminate parent 31

Table 2.2 Pollen viability (percent) of 14 Exacum Styer Group genotypes based on acetocarmine stain 32

Table 2.3 Chromosome numbers of three Exacum Styer group genotypes. Chromosome number and frequency of observations (in parentheses) are shown. Most frequent counts are highlighted in bold 33

Table 3.1 Plant growth regulator treatments applied to anthers of Exacum Styer Group genotype 01-09-01 56

Table 3.2 Cold treatment and media effects on pollen development and calli formation from the Exacum Styer Group genotype 01-09-01 57

Table 3.3 Heat treatment and media effects on Exacum Styer Group genotype 01-09-01 pollen development and calli formation 58

Table 3.4 Cold treatment and plant growth regulator combination effects (2,4-D and BA or 2,4-D and NAA) on calli formation of Exacum Styer Group genotype 01-09-01 59

Table 3.5 Cold treatment and plant growth regulator combination (2,4-D and Kinetin) effects on calli formation of Exacum Styer Group genotype 01-09-01 60

vi Table Al. 1 Chromosome numbers of two additional Exacum Styer Group genotypes. Chromosome number and frequency of observations (in parentheses) are shown 69

Table AIL 1 Cold treatment and media effects on pollen development and calli formation from Exacum Styer Group genotype 01-37-37 77

Table All.2 The effect of various BA and 2,4-D combinations on calli formation of isolated anthers of Exacum Styer Group genotype 01-37-37 77

vii LIST OF FIGURES

Fig. 1.1 Map displaying the distribution of Exacum species around the Indian Ocean Basin and their proposed dispersal routes (Yuan et al, 2005) 14

Fig. 1.2 Taxonomic treatment of the Genus Exacum and of the 8 taxa native to Sri Lanka (Sumanasinghe, 1986) 15

Fig. 1.3 Simplified schematic of microgametophyte formation in 16

Fig. 1.4 Simplified schematic chart of alternative pathways of androgenesis in plants 17

Fig. 1.5 Flowchart of procedures used in this study 18

Fig. 2.1 Pollen development in Exacum Styer Group genotype 01-09-01 by acetocarmine stain. (A) Prophase I, anther length 0.5 cm; (B) Telophase I, anther length 0.5 cm; (C) Prophase II, anther length 0.5 cm; (D) Metaphase II to telophase II, anther length 0.5 cm; (E) Tetrad stage, anther length 0.6 cm; (F) Mature pollen at anthesis, anther length 1.2cm. Scale bar = 10 um 34

Fig. 2.2 Pollen development in Exacum Styer Group genotype 01-42-03 by acetocarmine stain. (A) Prophase I to metaphase I, anther length 0.4 cm; (B) Telophase I to prophase II, anther length 0.4 cm; (C) Prophase II to metaphase II, anther length 0.4 cm; (D) Tetrad stage, anther length 0.5 cm; (E) Microspores released from tetrads, anther length 0.5 cm; (F) Mature pollen at anthesis (darkly stained viable pollen and translucent non viable pollen), anther length 1.1 cm. Scale bar = 10 urn 35

viii Fig. 2.3 Pollen development in Exacum Styer Group genotype 01-37-37 by acetocarmine stain. (A) Prophase I, anther length 0.4 cm; (B) Telophase I, anther length 0.4 cm; (C) Metaphase II and Telophase II, anther length 0.4 cm; (D) Tetrads, anther length 0.5 cm; (E) Microspore released from tetrads, anther length 0.5 cm; (F) Mature but non-viable pollen, at anthesis, anther length 1.1 cm. Scale bar = 10 urn 36

Fig. 2.4 Microsporogenesis in Exacum Styer Group genotype 01-09-01 stained with DAPI. (A) Uninucleate, anther length 1 cm; (B) Prophase during mitotic division, anther length 1 cm; (C) Metaphase during mitotic division, anther length 1 cm ; (D) Telophase during mitotic division; anther length 1 cm; (E) Early binucleate, anther length 1.1-1.2 cm; (F) Late binucleate at anthesis, anther length 1.2 cm. Scale bar = 10 um...37

Fig. 2.5 Microsporogenesis in Exacum Styer Group genotype 01-42-03 stained with DAPI. (A) Uninucleate stage, anther length 0.9 cm; (B) Viable pollen at binucleate stage (arrow), and non-viable pollen without nuclei at anthesis, anther length 1.2 cm. Scale bar = 10 um 38

Fig. 2.6 Cytological stages corresponding to anther development of Exacum Styer Group. Pollen mother cells start meiotic division when anthers are 0.4/0.5 cm in length and complete microspore released when anthers are 0.6 cm in length. The first pollen mitotic division starts when anthers are 1 cm in length and results in one condensed generative cell and one diffused vegetative cell. The generative cell becomes spindle shaped when anthers reach 1.2 cm 39

ix Fig. 2.7 Scanning electron micrographs of pollen from three Exacum Styer Group genotypes varying in pollen viability. (A and D) Genotype 01-09-01 with high pollen viability; (B and E) Genotype 01-42-03 with low pollen viability; (C and F) Genotype 01-37-37 with sterile pollen 40

Fig. 2.8 In vitro pollen hydration in Exacum Styer Group genotype 01-09-01. (A) Pollen structures at anthesis (i.e., no external moisture) and (B) Hydrated pollen following 30 sec exposure to distilled water. Non-viable pollen do not change shape with hydration (Arrow). Scale bar = 20 um 41

Fig. 2.9 Photomicrographs and corresponding drawings of somatic chromosomes of three Exacum Styer Group genotypes. (A) Genotype 01-09-01, 2n=50; (B) Genotype 01-42-03, 2n=66; and (C) Genotype 01-37-37, 2n=54. Scale bar = 5 um 42

Fig. 2.10 Pedigrees of the three Exacum Styer Group genotypes. The most common chromosome numbers of the three genotypes are proposed in parentheses. The chromosome numbers of parental lines are shown 43

Fig. 3.1 Microspore nuclei development from Exacum Styer Group genotype 01-09-01 after 7 days of culture on various base media supplemented with either sucrose or mannitol cultured under either room temperature or cold (10°C). (A-D) 10°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. (E-H) 25°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. Scale bar = 10 um 61

Fig. 3.2 Microspore nuclei development from Exacum Styer Group genotype 01-09-01 after 4 days of culture on various base media supplemented with either sucrose or mannitol cultured under either room temperature or heat (35°C). (A-D) 35°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. (E-H) 25°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. Scale bar = 10 um 62

x Fig. AI.l Pollen development in Exacum Styer Group genotype 01-47-21. (A) Prophase I, anther length 0.4 cm; (B) Metaphase II and Telophase II, anther length 0.4 cm; (C) Tetrads, anther length 0.5-0.6 cm; (D) Mature pollen at anthesis, anther length 1.2 cm. Scale.bar = 10 um 70

Fig. AI.2 Pollen development in Exacum Styer Group genotype 01-83-17. (A) Prophase I, anther length 0.5 cm; (B) Tetrads, anther length 0.5-0.6 cm; (C) Pollen at late uninucleate stage, anther length 0.9cm; (D) Pollen at late uninucleate to early binucleate stage anther length 1.0cm. Scale bar = 10 um 71

Fig. AI.3 Pollen development of Exacum Styer Group genotype 01-15-01. (A) Prophase to metaphase II, anther length 0.5 cm; (B) Tetrads, anther length 0.5-0.6 cm; (C) Pollen at mid- to late uninucleate stage, anther length 0.9 cm; (D) Pollen at late uninucleate to early binucleate stage anther length 1.1 cm. Scale bar = 10 um 72

xi LIST OF ACRONYMS

2,4-D: 2,4-Dichlorophenoxyacetic acid 2iP: 6-(Y,Y-dimethylallyamino) purine 8-HQS: 8-Hydroxyquinoline sulfate BA: N6-benzyladenine BSC: Biological species concept DAPI: 4,6-Diamidino-2-phenylindole DH: Double haploid GP-1: Primary gene pool GP-2: Secondary gene pool GP-3: Tertiary gene pool IAA: Indole-3-acetic acid IBA: Indole-3-butyric acid IOB: Indian ocean basin Kin: Kinetin MS medium: Murashige and Skoog medium (1962) Mya: Million years ago NAA: a-naphthaleneacetic acid NN medium: Nitsch and Nitsch medium (1969) PAA: Phenylacetic acid PGRs: Plant growth regulators PVP: Polyvinylpyrrolidone QTL: Quantitative trait loci SEM: Scanning electron microscope TIBA: 2,3,5-Triiodobenzoic acid ZEA: Zeatin

xii ACKNOWLEDGMENTS

I truly appreciate my supervisor, Dr. Andrew Riseman, for his guidance, support and his extraordinary patience during the timely discussions throughout these years. Many thanks to my advisory committee members, Dr. Brian Ellis and Dr. Shawn Mansfield for their constructive suggestions and support. My special appreciation is extended to Peter Kalynak for his suggestions and help with plant tissue culture techniques. Also, many thanks to all the members at the Bioimaging Facility, especially to Dr. Elaine Humphrey, Derrick Home, and Kim Rensing, for their help with the microscopy. Many thanks to all my colleagues in the Center for Plant Research, especially Alain Boucher, Diane Edwards, Faride Unda, and Laura Sand, for their encouragement and support. To my friends from Immaculate Conception Parish, Vancouver, for sharing my happiness and tears throughout my studies. To all my friends in Taiwan especially, Dr. Yung-YI Lee, for the valuable suggestions and encouragement. Finally, to my parents and my younger sister, for your love and encouragement throughout the many tough challenges during these years. Without your support, I don't think I could have finished this degree.

xiii Chapter 1. General introduction and literature review

1.1 General introduction Exacum Styer Group () are interspecific hybrids derived from 5 taxa native to Sri Lanka (Riseman et ah, 2005). These hybrids display significant variation in fertility as is expected from their interspecific background (Riseman, 2006). In addition to this variation among genotypes, E. Styer Group is in general intolerant of inbreeding and suffers significant loss of vigor following two generations of self-pollination thus inhibiting inbred line development via sexual reproduction (Riseman, personal communication). However, the development of inbred lines is highly desirable for use in breeding and academic research programs. One possible approach to develop homozygous lines that does not rely on sexual inbreeding is the production of double haploid (DH) plants by reprogramming gametophyte development (Kasha and Maluszynski, 2003). Reprogramming induces mitotic divisions following meiosis leading to the development of haploid embryos. Once haploid embryos are produced, subsequent treatment with colchicine to double the genome can produce homozygous plants with the original chromosome complement. There are several advantages to using haploid embryos to develop inbred lines as compared to traditional breeding including:

(1) Haploid embryos allow for more accurate identification of genotypes without the complication of dominance; (2) Haploid embryos allow for effective in vitro selection of desirable agronomic characteristics (Guo et al., 1999); (3) Haploid embryos facilitate further genetic analysis, such as gene identification and quantitative trait loci mapping.

Several protocols have been established to produce haploid and DH plants from various explants sources and using various reprogramming treatments. Haploid production via anther culture was first demonstrated in Datura innoxia M. following anther culture on a specific medium (Guha and Maheshwari 1964). The enclosed pollen

1 grains spontaneously developed into haploid embryos that emerged from anther after 6 to 7 weeks (Guha and Maheshwari 1964; 1966). Haploidy was confirmed by cytological examination (Guha and Maheshwari, 1966). Following the successful development of microspore-derived plants from D. innoxia in the late 1960's, more than 247 species representing 88 genera and 34 families of angiosperms have been successfully used in androgenic plant generation (Khush and Virmani, 1996). Unfortunately, no member of the Gentianaceae has been used to generate haploid plants by any process to date. However, before a reasonable attempt at haploid production can be made with this germplasm, several critical pieces of information need to be obtained. Specifically, pollen development needs to be characterized with special attention paid to the timing of the first mitotic division. Delineating this stage (i.e., late uninucleate to early binucleate) is reported to be crucial for the successful reprogramming of gamete development (Reynolds, 1997). In addition, several external factors, such as the physical stress for reprogramming and the appropriate culture medium for embryo development, must be determined. The general goals of this research are: (1) To characterize pollen development in E. Styer Group; (2) To evaluate stressor effects on in vitro development of microspores from a fertile genotype of E. Styer Group.

1.2 The Genus Exacum Exacum L. (Gentianaceae) are small annual or biennial herbaceous plants native to the continents or archipelagos in the Indian Ocean (Fig. 1-1). Currently, 64 species of Exacum are recognized occurring from western Africa through eastern Papua New Guinea (Yuan et al., 2005). The dispersal of Exacum exhibits two large and one small discontinuous distributions around the Indian Ocean Basin (IOB) with one large group concentrated in the southwest Africa-Madagascar region and the second large group concentrated in the Indo-Malay archipelagos; a third species-poor group is limited to the southern tip of the Arabian Peninsula (Klackenberg, 1985). According to Yuan et al. (2005), the current distribution of Exacum species is quite uneven over its range with only 2 species in the whole of continental Africa and only 5 species occurring in the Indo-Malay archipelagos, while on three island nations, a total of 57 species representing

2 90% of species in the genus, are found (i.e., Madagascar- 38 species, Socotra- 3 species, and Sri Lanka/southern India- 16 species). Phylogenetic and distribution patterns indicate that Exacum arose in part of Gondwanaland, and that during the mid-Cretaceous period (around 140 Mya), several vicariance events occurred resulting in the colonization of 3 small regions with high species abundance (Yuan et al., 2005). The epicenter of Exacum on Gondwanaland was probably located near Madagascar due to the abundance of more primitive Exacum species (Yuan et al., 2005). In addition, two dispersal events might have occurred relatively recently (4.7 Mya) with one migration from Madagascar to Africa, and the other from India through the South East Asia archipelago to northern Australia. These recent migrations and radiations resulted in several relatively distantly related species clusters representing the main clades we observe today (Yuan et al., 2005). The fact that several Sri Lanka Exacum taxa are problematic in terms of assignment to a robust taxonomic group (e.g., species vs. sub-species vs. varieties) supports Yuan's analyses and prediction of rapid evolution occurring within this group.

1.2.1 Exacum in Sri Lanka From Sri Lanka, Cramer (1981) recognized 5 species of Exacum, E. axillare Thwaites., E. petiolare Griseb., E. sessile L., E. walkeri Arn. and a highly variable E. trinervium (L.) Druce (= E. zeylanicum Roxb.) with four subspecies, subsp. macranthum, subsp. pallidum, subsp. ritigalensis, and subsp. trinervium (Table 1-1). The taxa within E. trinervium are collectively known as the E. trinervium complex (Sumanasinghe, 1986). Acceptance of the first four taxa as valid species was without dispute among taxonomists. However, due to significant morphological variation within the E. trinervium complex, this group remains questionable. In a more recent taxonomy, Klackenberg (1985) confirmed the four non-contentious species recognized by Cramer, re-established E. pedunculatum L. as a valid taxon, elevated subsp. macranthum and subsp. pallidum to species rank, and merged subsp. ritigalensis within E. trinervium. Part of the problem associated with species delineation is which of the definitions of a 'species' is used. One common and robust definition is the biological species concept (BSC) proposed by Mayr (1942). It defines a species as "a group of actually or potentially-inter breeding populations which are reproductively isolated from other such groups" and is reported to

3 provide an unambiguous delineation of a species (Hancock, 2004). Based on BSC, Riseman et al. (2006) concluded that subsp. trinervium and subsp. ritigalensis are conspecific, and that subsp. ritigalensis deserved a subspecies status. These conclusions were based on cytology, crossability and pollen viability studies on Exacum taxa and their hybrids.

1.2.2 Exacum species and horticultural potential L. (Persian violet) is the only Exacum species to be commercialized to date and has broad appeal as both a potted and a bedding plant. However, the genus contains other taxa that have significant horticultural appeal. Of note are the species native to Sri Lanka that display a wide range of valuable traits including large flowers ranging in color from violet to white, dark glossy green foliage, and various growth habits. Unfortunately, no one species contained sufficient trait combinations to warrant direct introduction. Therefore, a directed breeding program using five taxa was initiated with the goal to combine the most desirable traits into genotypes worthy of introduction. Though interspecific hybridization using E. pedunculatum, E. macranthum, E. pallidum, E. trinervium and E. trinervium subsp. ritigalensis, a new horticultural group was developed and named Exacum Styer Group (Riseman et al, 2005). As mentioned in Section 1.2.1, the species definition I choose is the BSC that defines species as intercrossing populations that are reproductively isolated from other such groups. However, this definition is not absolute. Just because populations are reproductively isolated in nature does not mean they are not able to intercross and produce viable and fertile offspring. Known as interspecific hybridization, this process of gene pool mixing is a fundamental process in evolution but often produces offspring that display some level of infertility. Hybridization among the Sri Lanka species resulted in viable progeny but fertility varied by parental combination and chromosome number (Riseman et al, 2005). In order to explain interspecies fertility, Harlan and de Wet (1972) suggested a 'gene pool system' that recognizes three types of genie assemblages: (1) Primary gene pool (GP-1) where hybridizations within a relatively narrow gene pool are easy with fertile hybrids generally produced (equivalent to BSC);

4 (2) Secondary gene pool (GP-2) where hybridizations within a broader gene pool are possible but difficult with hybrids produced that are week with low fertility; (3) Tertiary gene pool (GP-3) where hybridizations within this largest gene pool are impossible or produce hybrids that are completely sterile. Using this system, the Sri Lankan Exacum taxa used in hybridizations are thought to be within either GP-2 or GP-3 depending on the parental taxa used. The genetic similarity between many of the Exacum taxa allowed some degree of compatibility. However, the genetic compatibility was divergent enough to produce progeny with reduced fertility.

1.3 Pollen Biology 1.3.1 Male gametophyte development Male gametophyte development includes three main developmental stages: (1) sporogenous cell differentiation and meiosis; (2) post-meiosis development; and (3) microspore mitosis within the anthers (Chaudhury, 1993). The anther contains three primary layers: LI, L2 and L3, which will differentiate into different tissues (Goldberg et al, 1993). The LI layer gives rise to the epidermis. After primary formation, the epidermis shows limited differentiation except for forming the stomium which facilitates anther dehiscence in a mature anther (Shivanna, 2003). The L3 layer differentiates into connective tissues and the central vascular bundle. The connective tissues also show little further differentiation except for the development of the intersporangial septum (e.g., separation between the two locules of each half anther) and inner tapetum (Shivanna, 2003). The L2 layer gives rise to archesporia where cells divide periclinally and give rise to an outer primary parietal layer and an inner primary sporogenous layer in later development (Shivanna, 2003). The primary parietal cells form the endothecium, middle layers, and outer tapetum layers following differentiation (Shivanna, 2003). The tapetum is a transitory layer that surrounds the sporogenous tissues and plays several important functions in microspore development including supplying nutrients to the developing microspores and deposition of pollen wall components (Shivanna, 2003). The primary soporogenous layer gives rise to the 2n pollen mother cells which undergo meiosis giving rise to four In microspores (Shivanna, 2003). Following meiosis, these microspores are enclosed by a callose wall, forming the tetrad. When the callose wall is

5 degraded by hydrolytic enzymes synthesized by the tapetum, the four microspores are released (Shivanna, 2003). This process, from diploid mother cell to haploid microspore, is termed microsporogenesis (Shivanna, 2003). Following release from the tetrad, several additional developmental events occur. While the cytoskeleton for the pollen wall is established during meiosis, wall synthesis and exine deposition increase after release from the tetrads (McCormick, 1993). Based on morphology, chemistry, and function, mature pollen walls are composed of three main layers: intine, exine, and pollen coat (Shivanna, 2003). During pollen wall formation, the tapetum degenerates and supplies sporopollenin precursors, for exine production, pollen coat substances, and exine proteins (Shivanna, 2003). Soon after release from the tetrads, the nucleus momentarily remains in a central location with other organelles. However, a large vacuole soon develops with the nucleus migrating to the periphery. Following nuclear migration, most of the organelles (i.e., plastids and mitochondria) move to the side opposite the nucleus resulting in a polarity of organelle distribution (Shivanna, 2003). After redistribution, pollen initiate their first mitotic division resulting in two asymmetric daughter cells- a smaller generative cell and a larger vegetative cell. Following the first mitotic division, the generative cell undergoes a second mitotic division yielding the two sperm cells (Reynolds, 1997; Touraev et al., 1997) (Fig. 1.3). All plants undergo this second mitotic division. However, the timing of the second mitotic division (i.e., pre- or post pollen germination) has been used as a taxonomic character. In general, dicotyledonous plants (e.g., Exacum) shed pollen at the 2-cell (binucleate) stage (i.e., one vegetative and one generative cell) and delay the second mitotic division until after pollen germination while most monocotyledonous plants shed pollen at the 3-cell (trinucleate) stage (i.e., one vegetative and two sperm cells) (Reynolds, 1997). However, it should be noted that this characterization is not absolute and that some dicotyledonous plants shed pollen at the trinucleate stage.

1.3.2 Male sterility Male sterility in plants is defined as plants that are unable to produce functional microgametophytes (Shivanna, 2003). It is a common phenotype observed in interspecific hybrids and is commonly caused by either genie or chromosomal factors that

6 alter gene function during gametophyte development (Stebbins, 1971). Stebbins (1971) further divided the cause of infertility into two distinct groups. Genie sterility is defined when interspecific progeny have abnormal sex organs and/or abnormal meiosis caused by disharmony between parental gene combinations. Chromosomal sterility, on the other hand, is when gametes receive incomplete sets of whole chromosomes or unbalanced combinations of chromosome segments composed of partly homologous chromosome regions. Several male sterile phenotypes have been identified and varied from stamen-less to abnormal gametophyte development as described for Zea mays L. (Albertsen and Philip, 1981), Lycopersicon esculentum Mill. (Kaul, 1988), Oryza sativa L. (Hu and Rutger, 1992), and Arabidopsis thaliana L. (Chaudhury, 1993).

1.3.3 Androgenesis Androgenesis refers to the production of new plants from only the male gametophyte and eliminating fertilization from the reproduction process. Anthers are cultured in vitro, and their microspores are triggered or reprogrammed to become calli/embryos rather than microgametophytes. There are several extant pathways that explain how this process occurs (Fig. 1.4). The Al pathway, as described by Sunderland and Dunwell (1974), follows normal microgametophyte development in that the pollen cell's nucleus divides asymmetrically forming one large vegetative cell and one small generative cell. However, after the asymmetrical division, the vegetative cell continues to undergo mitotic division eventually forming an embryo without the involvement of the generative cell. This pathway was found in the dicotyledonous plants Nicotiana tabacum L. (Sunderland and Wicks, 1973), D. innoxia (Vasil, 1980), and Z. mays (Barnabas et al, 1987). An alternative to this pathway (termed A2 pathway), the asymmetric mitotic division is similar to that of the Al pathway, but the embryo is generated from the generative cell rather than from the vegetative cell. The A2 pathway has been identified in the Hyoscyamus niger L. (Raghavan, 1976; 1978). In yet another pathway described by Sunderland and Dunwell (1974), the nucleus undergoes symmetrical division during the first mitosis with two equal size nuclei produced. These two nuclei then equally contribute to embryo formation by continued mitotic divisions. This third pathway has been identified in Asparagus officinalis L. (e.g.,

7 asparagus) (Peng et al., 1997), and Brassica napus cv. Topas (Telmer et al., 1995). Once microgametophyte reprogramming has occurred, with commitment to an alternate developmental pathway, haploid plants can be generated either indirectly from calli or directly from reprogrammed microspores (Sporoy and Munshi, 1996). However, before this practice can become feasible for a specific species/genotype, three factors should be considered: (1) the natural attributes of the donor plant (i.e., pollen fertility, developmental stage, and genotype), (2) appropriate microspore reprogramming treatment, and (3) culture environment (i.e. protocols for tissue culture manipulation and microspore culture).

1.4 Factors related to successful androgenic induction 1.4.1 The natural attributes of the donor plant 1.4.1.1 Genotype Successful androgenic induction is often genotype dependent (Maheshwari et al, 1980). For example, only 5 of 12 genotypes of N. tobacum (Nitsch, 1969), 3 out of 43 Vitis vinifera lines (Gresshoff and Doy, 1974), and 3 out of 18 A. thalinana lines (Gresshoff and Doy, 1972) were responsive to androgenic induction. Furthermore, research by Jacobsen and Sopory (1978) suggests that androgenic potential is a selectable trait. They identified and accumulated genes favoring androgenic embryo formation by using a clone of Solanum tuberosum L. After breeding several generations, a specific strain was developed and selected that proved to possess better androgenic response than either parent. These results suggest that a significant genetic component is involved in the androgenic response.

1.4.1.2 Pollen developmental stage Another key factor for successful androgenic reprogramming is the developmental stage of the pollen. In all pathways listed above that lead to haploid calli/embryo formation, the literature indicates that pollen were most susceptible to reprogramming when at the late-uninucleate to binucleate stage (Reynolds, 1997). However, several N. tabacum cultivars have been shown to be susceptible to microgametophyte reprogramming as early as the tetrad stage (Nakata and Tanaka, 1968; Carlson, 1970).

8 Currently, no specific mechanism has been identified for how microspores are reprogrammed from their normal microgametophyte pathway to an androgenic pathway.

1.4.2 Appropriate microspore reprogramming treatments 1.4.2.1 Effect of temperature shock on androgenesis Temperature shock is a common treatment to reprogram microgametophyte development and trigger androgenesis. Whole flower buds or isolated anthers are usually subjected to either an elevated (> 30°C) or a lowered (4°C) temperature just about the time of the first mitotic division or while microspores are at the late uninucleate to early binucleate stage. A key event associated with successful reprogramming via a temperature treatment is the alteration of the first mitotic division from the normal asymmetric division to a symmetrical division. This change apparently increases the likelihood of a committed change toward haploid embryo formation (Nitsch and Norreel, 1973). Commonly, cold treatments induce a greater efficiency because they are thought to slow microspore development and metabolic activity, thereby allowing the 'arrested' microspores to be maintained at an optimum stage for induction over a longer time period (Duncan and Heberle-Bors, 1976; Wenzel et al, 1977; Vasil and Nitsch, 1975). In addition, cold treatments increase the time microspores can be nursed by maternal tissue (Sangwan and Sangwan-Norreel, 1987). However, heat shock is also effective in inducing haploid embryos (Lindquist, 1986). The beneficial effects of heat treatments may involve heat shock response (Lindquist, 1986), synchronization of the microspore populations, and increases in the number of competent microspores (Dunwell et al., 1983).

1.4.2.2 Effect of mannitol stress on androgenesis Anthers treated with mannitol is another common procedure to reprogram microspore development. These treatments are thought to have several beneficial effects on promoting embryo induction such as (1) increasing microspore DNA replication and chromosome doubling (Li et al, 1995a), (2) increasing endogenous hormone levels and peroxidase activity (Li et al, 1995b), and (3) inducing changes in soluble protein composition in the anthers (Li and Hu, 1995). In barley anther culture,

9 exposing anthers to 0.3 M mannitol treatment for 4 days (at 25°C) instead of to a metabolizable sugar, was found to be superior to other pretreatments requiring cold exposure (Roberts-Oehlschlager and Dunwell, 1990). However, a combination of heat shock and mannitol treatments induced a greater percentage of embryogenic microspores and a greater number of total embryos (Tourave et al., 1996a) than non-stressed cultures (Stauffer et al., 1991) in Hordeum vulgare L. anther culture. The duration of mannitol exposure is also crucial in reprogramming efficiency. The efficiency of androgenic embryo formation of Malus domestica Borkh. (e.g., apple) was reduced due to over exposure to mannitol (from 2 days to 4 days) (Hofer, 2004).

1.4.3 Culture environment Culture media provide the nutrients and plant growth regulators (PGR) required for microspores to undergo androgenesis and generally contain both inorganic and organic constitutes. Among the organic constitutes, sucrose (2-3%) is common and serves as a carbon source for growth and development. Several base formula have been successfully used for anther culture and include MS medium (Murashige and Skoog, 1962) for Paeonia hybrida Pall. (Sunderland and Dunwell, 1974), B5 medium (Gamborg et al., 1968) for Cucumis sativus L (Ashok Kumar et al., 2003), N6 medium for O. sativa (Chu et ah, 1975), and Nitsch and Nitsch medium for N. tabacum (Nitsch and Nitsch, 1969). With no particular pattern, each of these base media have been successfully tailored for haploid production in specific applications.

In over 80% of anther culture research, either auxins or cytokinins are required in the medium for androgenesis (Maheshwari et al., 1980, 1982). A combination of 2,4-Dichlorophenoxyacetic acid (2,4-D) (0.5-2 mg l"1) and Kinetin (Kin) (0.5-1 mg l"1) was commonly used in cereal anther culture for xTriticosecale Wittmack (Wang and Hu, 1984), Triticum aestivum L. (Pauk et al., 2003), Z. mays (Barnabas, 2003), Secale cereale L. (Immonen and Tenhola-Rioininen, 2003), and Phleum pretense L. (Puli and Guo, 2003). In dicotyledon species, a combination of N6-benzyladenine (BA) (1 mg 1"') anda-naphthaleneacetic acid (NAA) (2 mg I"1) was successfully used on A. officinalis (Peng et al., 1997) and Helianthus annuus L. (Saji and Sujatha, 1998). However, other auxin and cytokinin combinations, such as Indole-3-butyric acid (IBA) (0.2 mg 1"') and

10 Kin (0.2 mg l"1), have been used in apple (M domestica) anther culture (Hofer, 2004). The most common PGR combinations are summarized (Table 1.2). Regardless of specific combinations used, the effect of growth regulators was cultivar and concentration dependent (Arnison et al., 1990).

1.5 Justification and objectives

Currently, there are no reports detailing pollen developmental biology or androgenic induction for Exacum. Therefore, in this study, I detailed these biological phenomena in the E. Styer Group, and carried out preliminary reprogramming treatments as guided by the following set of objectives: (1) To evaluate pollen viability from fourteen genotypes of E. Styer Group; (2) To describe characteristics of microsporogenesis and microgametogenesis in selected E. Styer Group genotypes; (3) To characterize pollen morphology of the fertile and sterile genotypes of E. Styer Group genotypes; (4) To evaluate the pollen hydration ability of the fertile, reduced viability and sterile genotypes; (5) To determine chromosome numbers of several E. Styer Group genotypes; (6) To associate objectives 2-4 with pollen fertility; (7) To characterize the cytology of microspore development under various stress treatments using a fertile genotype; (8) To evaluate combinations of plant growth regulators on calli/embryo induction from microspores of a fertile genotype.

11 Table 1.1 Summary of taxonomic treatments for Sri Lankan Exacum

Cramer (1981) Klackenberg(1983) Sumanasinghe (1986) E. axillare E. axillare E. axillare E. petiolare E. petiolare E. petiolare E. sessile E. sessile E. sessile E. walkeri E. walkeri E. walkeri E. trinervium E. trinervium E. trinervium subspecies: subspecies: ritigalensis ritigalensis trinervium macrathum pallidum E. macrathum E. macrathum E. pallidum E. pallidum E. pedunculatum (Valid E. pedunculatum species reestablished)

12 Table 1.2 Plant growth regulator combinations in anther/ microspore culture induction

1 1 Reference Crop (explant) Basal Medium Auxin (mgl ) Cytokinin (mgl" ) Zeatin NAA IAA 2,4-D ffiA PAA 2,3,5- BA Kineti TTBA n 1 Jacquard etal., 2003 Barley (anther) C3 2 1 Barley FHG Cistueefa/., 2003 (m if*rr»Qr*f\rP i 0.5 Barnabas et al., 2003 Wheat (anther) W4 2 0.1 Barnabas et al., 2003 Maize (anther) YP 0.5 Triticale (anther) 190-2 1.5 Tuvesson et al., 2003 induction Zapata-Arias et al., 2003 Rice(anther) MS orN6 0.5-2 0.5 Pulli and Guo, 2003 Rye (microspore) PG96-M 1.5 Brassica species B5 or NLN Ferrie, 2003 (microspore) Tai and Xiong, 2003 Potato (anther) Uhr 85 0.1 2.5 Tourave and Heberle-Bors, Tobacco AT3 2003 (microspore) Asparagus Asparagus (anther) Wolyn and Nicols, 2003 anther culture 2 1 LTsay (1996) Hofer, 2003 Apple (anther) modified MS 0.2 0.2 Germana, 2003 Citrus (anther) N6 0.02 0.02 0.5 1 0.5 Ashok Kumar et al., 2003 Cucumis (anther) B5 0.5 0.3 Saji and Sujatha, 1998 Sunflower (anther) MS 2 1 —T—r

30 H y\\ —v 20 -\ \ v Arabian Peninsula

10 -\ // Socotra CD // (3spp) T3 Africa Sri Lanka .-i OH / (16spp) ' Gondwanaland / / / components -10 H / A

-20 H f n Madagascar/' (38spp)^/ 1> v -30 H Indian Ocean o I i I i I i I i I i I i I—i—I—I I I 30 40 50 60 70 80 90 100 110 120 Longitude (E)

Fig. 1.2 Map displaying the distribution of Exacum species around the Indian Ocean Basin and their proposed dispersal routes (Yuan et al., 2005).

14 Kingdom: Plantae Phylum: Magnoliophyta Class: Magnoliopsida Order: Family: Gentianaceae (gentians) Tribe: Exaceae Genus: Exacum Species: E. axillare E. macranthum* E. pallidum * E. pedunculatum * E. petiolare E. sessile E. trinervium* E. trinervium subsp. ritigalensis E. walkeri

Fig 1.2 Taxonomic treatment of the Genus Exacum and of the 8 taxa native to Sri Lanka (Sumanasinghe, 1986). *species used to form E. Styer Group

15 Pollen (1) Microspore (2) Vegetative cell (3) Vegetative cell (n) Mother cell (n) (2n) (n) • Sperm (n) Generative cell (n) Sperm (n)

(1) Microsporogenesis-meiosis (2) Microgametogenesis- asymmetrical division of mitosis I (3) Microgametogenesis- symmetrical division of mitosis I

Fig. 1.3 Simplified schematic of microgametophyte formation in plants.

16 (3) Microspore (1) Vegetative cell (n) Calli or embryo (n) (n) Generative cell (n)

Symmetrical cell (n) (4) Calli or embryo (n)

Symmetrical cell (n)

(1) Micrgametogenesis- asymmetrical division of mitosis I (2) Androgenesis- symmetrical division of mitosis I (3) Androgenesis- either vegetative or generative cell continues symmetrical mitotic division (4) Androgenesis- both cells continue symmetrical mitotic division

Fig. 1.4 Simplified schematic chart of alternative pathways of androgenesis in plants.

17 Chapter 2. Characteristics of pollen development in Exacum Styer Group

Fertile genotype

01-09-01

First Mitotic division commenced approximately when anthers 1-1.1cm in length

Chapter 3. Stressor effects on in-vitro development of microspores from a fertile genotype of Exacum Styer Group

Late uninucleate to early bi-nucleate anthers ( 0.8 -1.1 cm in length) used as explants

10 C or 25 "C on MS medium and 35 °C or 25 C on MS medium N&N medium supplemented with and N&N medium supplemented 30 gl"1 sucrose or 63.8 g I"1 with 30 gl'1 sucrose or 63.8 g I'1 mannitol for 7 days mannitol for 7 days 1 Following 7 days 10 °C treatment, pollen maintained normal development regardless media

Cultured on MS media supplemented with 30gl~1 sucrose for 7 days at 10 °C was chosen for PGR experiments I With Different BA With different BA With different Kin and 2,4-D and NAA and 2,4-D combinations combinations combinations I J 1 With all PGR combinations, calli production was observed from the filaments. No calli were emerged from anthers

Fig. 1.5 Flowchart of procedures used in this study.

18 Chapter 2. Characteristics of pollen development in Exacum Styer Group

2.1 Introduction The accepted taxonomic treatment for Exacum supports 64 valid species with 16 species native to Sri Lanka and the southern tip of India (Yuan et al., 2005). Of these taxa, five species were collected, evaluated, and bred for horticultural merit (Riseman and Craig, 1995). After twelve generations of controlled hybridization and selection for improved horticultural traits (i.e., flower forms and colors, production characteristics), a unified population evolved and was subsequently name Exacum Styer Group (Riseman et al, 2005). For a variety of reasons, we are interested in developing haploid plants through anther/microspore culture. Once haploid plants are obtained, they can then be chemically treated with colchicine to reestablish the 2n complement and produce double haploid (DH) plants. The advantages of DH plants are based on their 100% homozygosity and include more efficient early generation selection, quantitative trait loci evaluation, and Fl hybrid production (Kasha and Maluszynski, 2003). However, before haploid plants can be developed via anther/pollen culture, the basic biology of pollen development needs to be characterized to ensure the tissue used is at an appropriate developmental stage. In addition, this characterization will serve as the 'control' developmental sequence to identify any alterations in development following haploid induction treatments. Only limited information is available about pollen viability or pollen development in E. Styer Group. In general, the pollen viability (i.e., in vitro germination test) of the parental taxa used in E. Styer Group's development was as follows: E. macranthum- 0.21%, E. trinervium- 24.6%, and E. trinervium subsp. ritigalensis- 30.2-42.8%. In addition, the pollen viability of the primary interspecific hybrids was highly variable ranging from 71.6% (E. trinervium x E. trinervium subsp. ritigalensis hybrids) to 0.4% (E. trinervium subsp. ritigalensis x E. macranthum hybrids) (Riseman, 1990). This reduction in fertility or the development of sterility in interspecific hybrids is a common

19 phenomenon. It can be caused by either genie or chromosomal factors that alter gene function for any process during normal male gametophyte development (Stebbins, 1971; Horner and Palmer, 1995). For example, in safflower {Carthamus tinctorius L.) hybrids, the sterile plants possessed pollen with varying exine thickness and a lost of finely articulate symmetrical peaks on the surface layer. The germination pores were generally obscured and appeared to be closed with whole grains appearing deeply shrunken (Carapetian, 1994). Unfortunately, after twelve generations of hybridization among E. Styer Group accessions, we know little about the basis for the observed reduced fertility and sterility expressed in some genotypes. In order to better understand the basis for this phenomenon and to gain insights in to the developmental biology of pollen for future application in haploid production, the specific objectives of this research were: (1) To evaluate pollen viability from fourteen genotypes of E. Styer Group; (2) To describe characteristics of microsporogenesis and microgametogenesis in selected E. Styer Group genotypes; (3) To characterize pollen morphology of fertile and sterile genotypes of E. Styer Group; (4) To evaluate the pollen hydration ability of the fertile, reduced viability and sterile genotypes (5) To determine chromosome numbers of several E. Styer Group genotypes; (6) To associate any of the above with pollen fertility.

2.2 Material and Methods 2.2.1 Plant material Exacum Styer Group was bred from several species and taxa native to Sri Lanka (i.e., E. pedunculatum, E. macranthum, E. pallidum, E. trinervium, and E. trinervium subsp. ritigalensis) (Riseman et al., 2005). Genotypes included were selected from the E. Styer Group breeding program at the University of British Columbia, Vancouver, BC. In total, 14 genotypes (01-09-01, 01-37-37, 01-37-50, 01-37-69, 01-42-03, 01-47-21, 01-47-49, 01-50-46, 01-83-17, 04-06-01, 04-06-05, 04-06-06, 04-06-09, and 04-15-01) were used for this part of the research. All experimental plants were derived from seed or micropropagated cuttings. The

20 cuttings or seedlings were transferred to the Horticulture Greenhouse and transplanted into 10 cm pots filled with a commercial potting mix (West Creek Farms, BC, Canada; 75 % peat and 25% perlite plus a starter charge of NPK; pH adjusted by dolomite and limestone to between 5.5 and 6.5). The temperature of the greenhouse was maintained at 22-24°C day and 20-24°C night. From June through September plants were grown under natural light, while during the rest of the year, ambient light was supplemented with 200 umoles supplied by high-pressure sodium lamps. For supplemental lighting, an 18 hour photoperiod was used (6:00-22:00). Humidity varied from 50-80% depending on time of year and weather with the most humid conditions typically during summer. Plants were sub-irrigated as needed with fertilizer 15-5-15 Cal-Mag (The Scotts Company, Marysville, Ohio). Zinc-EDTA (Plant Products Co. Ltd., Brampton, Ontario) at 10-15 g l"1 was applied every second week as a prophylactic against zinc deficiency. Sulfur (Safer Ltd. Scarboruogh, Ontario) at 12 g l"1 was applied every week to control powdery mildew.

2.2.2 In vivo pollen viability Pollen viability was assessed by acetocarmine stain and observed under a light microscope (Motic B5 Professional Series, Richmond, Canada) at 40X mag. Individual pollen grains were scored as viable when shaped round and stained red, and non-viable when they were oblong shaped and remained unstained. For each genotype, flowers at anthesis were collected, with one anther randomly selected for observation. Anthers were macerated in a drop of stain and a cover slip applied with gentle pressure. For each slide, three random fields were observed and total viable and non-viable pollen grains counted.

2.2.3 Microsporogenesis with acetocarmine stain Three genotypes (01-09-01, 01-37-37, and 01-42-03) were used to characterize microsporogenesis and were selected based on their varying pollen fertility. Flower buds that contained anthers between 0.4-0.8 cm in length were collected between 9:00-13:00. Anthers were excised and fixed for 24 h at 4°C in Carnoy's solution (3 parts 95 % alcohol: 1 part acetic acid) supplemented with ferric chloride (4 g ml"1 in 49 ml

21 Carnoy's solution). After fixation, anthers were washed three times in distilled water and stored in 70% ETOH at 4°C until examination. To determine the duration of meiosis, anthers were fresh collected at 9:00, 10:00, 11:00, 12:00 and 13:00 or collected at 10:00 for a five consecutive days. A minimum of ten anthers were examined for each genotype and collection time to ensure a sufficient sample size. Microsporogenetic events were observed under a light microscope (Motic B5 Professional Series, Richmond, Canada) at 40X mag, with digital photomicrographs captured with a Nikon Coolpix 4500 camera (Nikon Corp., Japan) attached to the microscope.

2.2.4 Microgametogenesis with DAP I stain Two genotypes selected based on their varying pollen fertility (01-09-01, 01-42-03) were used to characterize microgametogenesis. Anthers were collected between pre-anthesis (approximately 0.8 cm in length) through anthesis (approximately 1.2 cm in length) at 0.1 cm intervals. The anthers were fixed as described above. However, for visualization, 4,6-diamidino-2-phenylindole (DAPI) (5 ug ml"1) was used instead of acetocarmine and the preparations were kept in the dark for 15 min following maceration. The stage of microgametogenesis was determined by visual observation with an Axioplan fluorescent microscope (Zess Axioplan 2 Imaging, Germany) at 100X mag. Digital photomicrographs were captured with an attached Zess Axioplan 2 Imaging camera at the Bioimaging Center, UBC.

2.2.5 Pollen morphology by scanning electron microscopy (SEM) Three genotypes selected based on their varying pollen fertility. (01-09-01, 01-37-37, and 01-42-03) were used to characterize pollen morphology via SEM. Anthers at anthesis were manually manipulated to release mature pollen onto a SEM aluminum stub. Since mature pollen was dehydrated at maturity, critical point drying was deemed unnecessary. Once mounted, samples were directly sputter coated with gold for 3.25 mins. The outer structure of Exacum pollen was imaged under a scanning electron microscope (Hitachi S 2600N, Japan) at the Bioimaging Center, UBC. The voltage was set at 5 KV with a working distances of approximately 10 mm.

22 2.2.6 In vitro pollen hydration Three genotypes (01-09-01, 01-37-37, and 01-42-03) of E. Styer Group were used to evaluate the ability of pollen to hydrate via either direct exposure to distilled water or by exposure to actocarmine stain solution. Mature anthers were manually manipulated to release mature pollen onto a dry slide. Pollen was observed under a light microscope (Motic B5 Professional Series, Richmond, Canada) at 40X mag, and digital photographs taken with an attached Nikon Coolpix 4500 camera. Next, a drop of distilled water was placed on the slide and a cover slip applied. Following 30 sec, observations were repeated as described above.

2.2.7 Chromosome counts with Giemsa stain In the three genotypes (01-09-01, 01-37-37, and 01-42-03) of E. Styer Group, root-tips were harvested from only actively growing healthy plants in the UBC Horticulture Greenhouse, Vancouver, BC. Root rips were collected between 12:30-13:30 under sunny conditions. Once collected, roots were pretreated for five hours (shaken) with 0.002 M 8-Hydroxyquinoline sulfate (8-HQS). After pretreatment, the roots were washed with distilled water three times and fixed in Carnoy's solution (3 parts 95% alcohol: 1 part acetic acid) for 24 h at 4°C. For storage until use, roots were washed with distilled water three times and store in 70% ETOH at 4°C. For observation, roots were washed with distilled water three times and digested for 2 h at 37°C in a 8.25 U u.1" pectinase plus 6.55x 10" U ul" cellulase solution. After digestion, roots were again washed with distilled water three times. Tips were then macerated on slides in a drop of methanol and acetic acid mix (3 parts 100% methanol: 1 part acetic acid). The solution was ignited to burn off the solution and to fix the cells on the slide. Once dried,

the tissue was stained with Giemsa solution (3 ml Giemsa stain, 30 ml 0.067 M Na2HP04, 30 ml 0.067 M KH2PO4) for five minutes, washed under tap water to remove excess dye, and allowed to air-dry. Observations and chromosome counts were made with a light microscope (Zess Axioplan 2 Imaging, Germany) at 100X mag.

2.3 Results 2.3.1 Pollen viability

23 Pollen viability was determined for 14 genotypes, with only four genotypes producing >20% viable pollen (Table 2.2). The remaining ten genotypes were either sterile or with reduced pollen viability (2-6%). Of the siblings evaluated, one pair (01-47-21 and 01-47-49) ranged in pollen viability between 0% and 5.6%, while the other group of siblings (01-37-37, 01-37-50, 01-37-69) equally produced non-viable pollen.

2.3.2 Microsporogenesis and microgametogenesis Microsporogenesis and microgametogenesis were characterized for E. Styer Group: genotypes 01-09-01 (high pollen viability), 01-42-03 (low pollen viability) and 01-37-37 (sterile). In general, pollen development across genotypes followed a typical pattern for angiosperm species, and as an example, displayed for genotype 01-09-01 (Fig. 2.1). Meiosis was synchronous without irregularities or bridges observed. When anthers grew to approximately 0.5 cm in length, pollen mother cells started meiosis (Fig. 2.1 A-D) and formed tetrahedral tetrads by 0.6 cm in length (Fig. 2.1 E). Mature pollen was found in anthers 1.2 cm in length (Fig. 2.1 F). Cytokinesis in these genotypes belongs to the simultaneous type (e.g., cell wall forms when the meiosis is completed). The entire meiotic process occurred over 120 hours. A comparable pattern of pollen development was also found in the low pollen viability genotype (Fig. 2.2 A-E). However, at anthesis, a mixed population of viable and non-viable microspores was observed (Fig. 2.2 F). Microsporogenesis in the sterile genotype was comparable to the other genotypes (Fig. 2.3). However, meiosis began earlier than that of the high fertility genotype, typically at 0.4 cm anther length (Fig. 2.3 A-C). This sterile genotype appeared to progress through meiosis, produce tetrads, and release microspores, normally (Fig. 2.3 D&E). However, all released microspores were non-viable (Fig. 2.3 F). Normal microgametogenesis was observed from the high pollen viability genotype. The first mitotic division commenced at anther lengths between 1.0-1.1 cm (Fig. 2.4 A-D). Asymmetrical cytokinesis in the microspore produced two unequal sized cells; one large vegetative cell with a diffusive nucleus and a smaller generative cell with a condensed nucleus (Fig. 2.4 E). The generative cell was spherical in shape at first, but became

24 spindle-shaped when anthers reached 1.2 cm in length (e.g., when petals initiated pigment production) (Fig. 2.4 F). Pollen was determined to be bi-nucleate at anthesis. In the low pollen viability genotype, throughout the process (e.g., released from tetrad (anther lengths 0.6 cm) to late uninucleate (anther length 0.6 to 0.9 cm)), microspores maintained normal nuclei structure (Fig. 2.5 A). Microspores destined to be fertile continued their normal development and formed generative and vegetative cells (Fig. 2.5 B). However, those microspores with an abnormal mitotic division I matured into pollen without detectable nuclei. As an overview of E. Styer Group, microsporogenesis and microgametogenesis, a summary figure of the developmental stages and corresponding anther lengths is shown (Fig. 2.6).

2.3.3 Pollen characteristics by scanning electron microscopy (SEM) Pollen exine and morphology were examined for a high pollen viability genotype, a low pollen viability genotype, and a sterile genotype. The high pollen viability genotype produced normal oblong shaped pollen (30um x 16um) with uniform cell borders (Fig. 2.7 A). However, in the low pollen viability and sterile genotypes, pollen with an irregular appearance was present in approximately the same percentage as overall pollen viability (Fig. 2.7 B-C). Higher resolution images revealed that viable pollen, regardless of genotype, exhibited a smoother exine sculpture with three well-formed germination pores as compared to non-viable pollen which lacked these features (Fig. 2.7 D-E). In addition, pollen from the sterile genotype was uniformly misshapen with a shrunken morphology (Fig. 2.7 F).

2.3.4 In vitro pollen hydration Pollen from the fertile, partially fertile and sterile genotypes were evaluated for morphological changes associated with hydration. Prior to hydration, mature pollen were oblong shaped as previous described, regardless of genotype (Fig. 2.8 A). Thirty seconds following exposure to water, hydration was complete with fertile pollen swelling into spheres while non-viable pollen failed to hydrate and remained oblong shaped. The percentage of non-hydrated pollen was positively associated with both the percent pollen

25 with shrunken morphology and viability stain results (Fig. 2.8 B).

2.3.5 Chromosome counts Suitable tissue and slide preparation protocols proved to be difficult and resulted in the production of only a few mounts with cells appropriately conditioned for mitotic observation. Also, due to the relatively high number of small chromosomes, identical counts were not obtained from all cells. Therefore, the data presented include both the range in chromosome numbers found for an individual, their frequency, and identification of the most confident count (Table 2.3). As expected, chromosome morphology and number varied across interspecific accessions. Among the three genotypes evaluated, individual chromosomes ranged in size from 0.3 um (dot-shaped) to 2.65 um (rod-shaped) (Fig. 2.9) and total chromosome counts ranged 50 (01-09-01) to 66 (01-42-03) (Table 2.3). This range is deemed very broad and extended beyond the highest and lowest chromosome numbers found in the parental taxa (Fig. 2.10).

2.4 Discussion This research details pollen development (i.e., microsporogenesis and microgametogenesis) of E. Styer Group and provides morphological and cytological comparisons between genotypes varying in fertility. The observation of simultaneous cytokinesis in E. Styer Group confirms previous research on other Exacum species (Rao and Chinnappa, 1983; Lakshminarayana and Maheswari Devi, 1985). In this study, the tetrads were tetrahedral and were similar to those observed in E. bicolor Roxb., E. petiolare Griseb., E. pumilum Griseb. (Maheswari Devi, 1962) and E. pedunculatum (Rao and Chinnappa, 1983), while both tetrahedral and isobilateral tetrads were found in E. saulieri L., but with the tetrahedral form dominant (Lakshminaray and Maheswari Devi, 1985). Worth mentioning is that E. pedunculatum is one of the parents of the E. Styer Group and may have contributed to this phenotype. At anthesis, mature pollen is binucleate (i.e., two-cell stage). However, reports indicate that this condition is not uniform within Exacum with E. pedunculatum (Rao and Chinnappa, 1983), E. Saulieri L. (Lakshminaray and Maheswari Devi, 1985), and E. bicolor (Maheswari Devi, 1962) have been shown to be trinucleate at anthesis.

26 In general, the duration of meiosis in pollen mother cells is highly variables across plant species and is influenced by four main factors: (1) ploidy level of the organism; (2) nuclear DNA content; (3) genotype; and (4) environmental factors (e.g., temperature). In this study, Exacum pollen mother cells required approximately 120 h to complete meiosis. This is the 1st known report on the duration of meiosis for any Exacum species or hybrid. Chromosome counts indicate a wide range in number from 50 to 66 (Table 2.3). This range is deemed broad and may indicate genomic instability, genomic incompatibility, or polyploidy incompatibility. The ploidy level of the parental taxa was proposed as follows: E. macranthum- hexaploid with a base number of 9 (2n=6x=54); E. pallidum- hexaploid with base number of 9 plus aneuploid reduction (2n=6x-2=52) and E. trinervium - hexaploid with a base number of 10 (2n=6x=60) (Sumanasighe, 1986). Compared to known reports in other polyploid species, meiotic duration ranged from 20 h in Veronica chamaedrys L. (4x=28) (Bennett, 1973) to as long as 144 h (4x=24) in Tradescantia reflexa Raf. (Sax & Edmonds, 1933; Bennett, 1971). Although T. reflexa had a similar chromosome number as V. chamaedrys, the 3C nuclear DNA content of 7! reflexa is 52 times greater than V. chamaedrys (i.e., 144.9 pg v.s. 2.8 pg) (Bennett et al., 1977). In other words, the plants with greater nuclear DNA content require more time to complete meiosis than those plants with less DNA. Comparing the meiotic duration among known polyploids, E. Styer Group appears to take a long time to complete meiosis. However, no information on total nuclear DNA content is available. Therefore, these two issues can not be resolved at this time for this group. In general, three developmental stages are recognized for successful development of viable pollen: (1) differentiation of sporogenous cells and meiosis, (2) post-meiosis development, and (3) microspore mitosis. Aberration during any of these stages results in the production of abnormal and non-functional male gametophytes (Chaudhury, 1993). In all three genotypes evaluated (i.e., fertile, partially fertile and sterile), the differentiation of sporogenous cells was normal and meiosis was synchronous with no irregularities or bridges observed. Therefore, the 1st required stage during microspore development is normal and appears not to influence final pollen fertility of a genotype. Post-meiosis development includes the degradation of the callose wall surrounding the tetrads thereby releasing the microspores into the locule and intine/exine development

27 (Chaudhury, 1993). Most microspores from both the fertile and sterile genotypes appear to progress through post-meiosis development normally through the release of the immature pollen grains from the tetrads. However, only microspores from the fertile genotype continued normally with post-meiosis development with normal exine deposition (Fig. 2.7 A&D). Relatively few microspores from the partially fertility genotype (Fig. 2.7 B&E) and no microspores from the sterile genotype, proceeded through normal exine deposition (Fig. 2.7 C&F). Observations of pollen viability and morphology indicate that a positive association was present between non-viable pollen (i.e., unstained) and oblong shaped pollen (as opposed to fully round pollen). However, from non-hydrated SEM observations, viable pollen could not be discerned from non-viable pollen by the oblong shape alone (Fig. 2.7). Based on results from the in vitro pollen hydration evaluations, both viable and non-viable pollen of Exacum are oblong shaped at anthesis (i.e., dehydrated) and only viable pollen are able to hydrate normally (Fig. 2.8). The shrunken pollen morphology associated with non-viable pollen may have been caused by several factors including abnormal pollen wall formation and cytoplasmic leakage or degeneration (Taylor et al, 1998). Shrunken pollen from Arabidopsis thaliana has been associated with degenerated cytoplasm that fails to maintain osmotic potential and turgidity (Taylor et al, 1998). For example, the male sterile mutant msll produces collapsed microspores at dehiscence and was associated with failed pollen wall structure. Detailed transmission electron microscope (TEM) observations of the exine structures showed that the exine was irregular and collapsed. /?-galactosidase activity was detected in the locules but not in the microspores suggesting that the pollen grain's cytoplasmic contents leaked into the surrounding area. In addition, in the mutant ms8, empty flattened shells of pollen grains were found in the locules. Observations showed that pollen exine was continuous around the surface of the pollen but that the germination pores were absent or reduced. The protoplasmic contents were also found in the locules suggesting that the protoplast had exuded through the incomplete germination pore zone (Taylor et al, 1998). In Exacum Styer Group, SEM observations showed that pollen from the sterile genotype has abnormal germination pores (i.e., without or reduced) on the exine (Fig. 2.7 F). With abnormal exine structure, the protoplasmic contents may

28 not be maintained in the pollen and may lead to shrunken pollen morphology. However, other issues may have caused this morphology. Most abnormal pollen wall phenotypes described have been associated with abnormal tapetum function (Taylor et al., 1998). In Arabidopsis male sterile mutant, ms 9, irregular pollen without exine was caused by abnormal tapetum function (Taylor et al., 1998). Unfortunately, information on the tapetum's role in viable and non-viable pollen development of E. Styer Group is lacking. Therefore, future research should investigate both the developmental biology of the tapetum and its role in non-viable pollen production. Abnormal microspore mitosis is another factor that may lead to pollen sterility (Chaudhury, 1993). In the low pollen viability genotype (01-42-03), all pollen contained nuclei that appeared normal when released from the tetrads to just before first mitotic division. However, only a small percentage of the pollen proceeded with normal development progressing through the first mitotic division yielding one vegetative and one generative cell, while the majority of pollen failed to progress in development and contained degraded nuclei. In Petunia PRK1 transgenic plants, pollen that did not have normal microspore mitosis were arrested at the uninucleate stage, and soon after lost their nuclei (Lee et al., 1996). This result suggests that the non-viable pollen produced by some E. Styer Group genotypes may involve an abnormal mitotic process first appearing during the first mitotic division in microgametogenesis. Genome structure has also been implicated in variable fertility of interspecific hybrids and is often due to incompatibilities between the parental taxa leading to euploidy and/or aneuploidy. As mentioned, the chromosome numbers of the three genotypes evaluated ranged from 2n=66 (01-42-03) to 2n=50 (01-09-01) while those of the parental taxa ranged from 2n=52 (E. pallidum) to 2n=60 (E. trinervium) (Sumanashinghe, 1986). Of note is the accession with the lowest chromosome number was also the accession with the highest pollen viability. Therefore, the 50 chromosomes within the fertile genotype carry sufficient genetic information to produced functional gametes. However, in accessions with higher chromosome numbers, pollen fertility was reduced or was sterile. In addition, the genetic factors mentioned may also influence female gametophyte development. In the accessions evaluated, both male and female gamete viability were

29 positively associated based on breeding performance (e.g., accessions with high pollen viability also produced high numbers of viable seedlings) (data not shown). In reviewing the pedigrees of these accessions, I found that a common cross combination (i.e., Fl back crossed to E. trinervium subsp. ritigalensis) might indicate the source of the sterility (Fig. 2.10). In general, higher chromosome numbers often contain redundant copies of genes to prevent the loss of critical genetic material. However, when higher chromosome numbers are associated with reduced fertility, it may indicate disharmonious interactions between gene copies (Stebbins, 1971). In this study, the characteristics of microsporogenesis and microgametogenesis of E. Styer Group were described. In addition, the comparative morphology of fertile and sterile E. Styer Group genotypes was completed. The development of pollen from the sterile genotype was characterized as having an abnormal first mitotic division and aberrant exine surface. All alterations leading to reduced fertility occurred after meiosis.

30 Table 2.1 Parentage of 14 Exacum Styer Group genotypes; by convention, pistillate parent is listed first followed by the staminate parent. Genotype* Parental Hybridization

01-09-01 86-15-02x86-35-01 01-37-37 89-10-09x86-35-01 01-37-50 89-10-09x86-35-01 01-37-69 89-10-09x86-35-01

01-42-03 89-73-06x86-35-01 01-47-21 91-68-01 x 91-45-18 01-47-49 91-68-01 x 91-45-18

01-50-46 92-13-30x92-54-25

01-83-17 95-23-27x91-66-26

04-06-01 02-171-03 x03-11-16

04-06-05 02-171-03 x 03-11-16 04-06-06 02-171-03 x03-11-16 04-06-09 02-171-03 x 03-11-16 04-15-01 03-36-01 x 03-06-28

* first number denotes the year that the seed were sown, the second number denotes the parental combination and the third number denotes the individual sibling.

31 Table 2.2 Pollen viability (percent) of 14 Exacum Styer Group genotypes based on acetocarmine stain. Code of Viability

genotypes Mean ± SD(%) Range (%)

01-09-01 54.5 ± 4.85 35.6- 80.5 01-42-03 2.05 ± 0.8 3.05-7.21 01-47-21 5.6 ± 2.3 0-3.0 01-83-17 48.2 ± 3.06 37.0-61.5 04-15-01 27.8 ± 5.49 31.8-40.5 04-06-01 20.9 ± 6.45 3.3- 42.6 01-37-37 0 0 01-37-50 0 0 01-37-69 0 0

01-47-49 0 0 01-50-46 0 0 04-06-05 0 0

04-06-06 0 0

04-06-09 0 0

32 Table 2.3 Chromosome numbers of three Exacum Styer Group genotypes. Chromosome number and frequency of observations (in parentheses) are shown. Most frequent counts are highlighted in bold.

Genotype Range in chromosome number 01-09-01 48(1) 50(5) 54(2) 56(1) 64(1) 01-37-37 54(4) 56(1) 58(3) 60(1) 66(1) 01-42-03 52(1) 54(1) 56(1) 60(1) 62(1) 64(1) 66(3) 68(1)

33 B ^

Fig. 2.1 Pollen development in Exacum Styer Group genotype 01-09-01 by acetocarmine stain. (A) Prophase I, anther length 0.5 cm; (B) Telophase I, anther length 0.5 cm; (C) Prophase II, anther length 0.5 cm; (D) Metaphase II to telophase II, anther length 0.5 cm; (E) Tetrad stage, anther length 0.6 cm; (F) Mature pollen at anthesis, anther length 1.2cm. Scale bar = 10 um.

34 Fig. 2.2 Pollen development in Exacum Styer Group genotype 01-42-03 by acetocarmine stain. (A) Prophase I to metaphase I, anther length 0.4 cm; (B) Telophase I to prophase II, anther length 0.4 cm; (C) Prophase II to metaphase II, anther length 0.4 cm; (D) Tetrad stage, anther length 0.5 cm; (E) Microspores released from tetrads, anther length 0.5 cm; (F) Mature pollen at anthesis (darkly stained viable pollen and translucent non viable pollen), anther length 1.1 cm. Scale bar = 10 urn.

35 Fig. 2.3 Pollen development in Exacum Styer Group genotype 01-37-37 by acetocarmine stain. (A) Prophase I, anther length 0.4 cm; (B) Telophase I, anther length 0.4 cm; (C) Metaphase II and Telophase II, anther length 0.4 cm; (D) Tetrads, anther length 0.5 cm; (E) Microspore released from tetrads, anther length 0.5 cm; (F) Mature but non-viable pollen, at anthesis, anther length 1.1 cm. Scale bar = 10 um

36 Fig. 2.4 Microsporogenesis in Exacum Styer Group genotype 01-09-01 stained with DAPI. (A) Uninucleate, anther length 1 cm; (B) Prophase during mitotic division, anther length 1 cm; (C) Metaphase during mitotic division, anther length 1 cm ; (D) Telophase during mitotic division; anther length 1 cm; (E) Early binucleate, anther length 1.1-1.2 cm; (F) Late binucleate at anthesis, anther length 1.2 cm. Scale bar = 10 nm.

37 Fig. 2.5 Microsporogenesis in Exacum Styer Group genotype 01-42-03 stained with DAPI. (A) Uninucleate stage, anther length 0.9 cm; (B) Viable pollen at binucleate stage (arrow), and non-viable pollen without nuclei at anthesis, anther length 1.2 cm. Scale bar = 10 um.

38 Microsporogenesis

Fig. 2.6 Cytological stages corresponding to anther development of Exacum Styer Group. Pollen mother cells start meiotic division when anthers are 0.4/0.5 cm in length and complete microspore released when anthers are 0.6 cm in length. The fist pollen mitotic division starts when anthers are 1 cm in length and results in one condensed generative cell and one diffused vegetative cell. The generative cell becomes spindle shaped when anthers reach 1.2 cm. Fig. 2.7 Scanning electron micrographs of pollen from three Exacum Styer Group genotypes varying in pollen viability. (A and D) Genotype 01-09-01 with high pollen viability; (B and E) Genotype 01-42-03 with low pollen viability; (C and F) Genotype 01-37-37 with sterile pollen.

40 Fig. 2.8 In vitro pollen hydration in Exacum Styer Group genotype 01-09-01. (A) Pollen structures at anthesis (i.e., no external moisture) and (B) Hydrated pollen following 30 sec exposure to distilled water. Non-viable pollen do not change shape with hydration (Arrow). Scale bar = 20 um.

41 D % ft % • • • • « •

% • ft 4 • m\\r * % ft t • • • • • * . s * ^ • ft % • • 7

B E

%.> V • • c

7

Fig. 2.9 Photomicrographs of somatic chromosomes of three Exacum Styer Group genotypes. (A) Genotype 01-09-01, 2n=50; (B) Genotype 01-42-03, 2n=66; and (C) Genotype 01-37-37, 2n=54. Scale bar = 5 urn.

42 85-117-02 E. ritigalensis (2n=60) 86-15-10 85-102-12 E. macranthum (2n=54) 89-10-09

01-37-37 (2n=54) 85-117-12|£ ritigalensis (2n=60)

85-123-23\E. trivervium (2n=60) 86-35-01 85-113-01 \E. pallidum (2n=52)

85-117-02 E. ritigalensis (2n=60) 86-15-17 89-73-06 85-102-01 E. macranthum (2n=54)

85-117-12|£ ritigalensis (2n=60) 01-42-03 (2n=66)

85-123-231^ trivervium (2n-60) 86-35-01 86-113-01 \E. pallidum (2n=52)

85-117-02|£ ritigalensis (2n=60) 86-15-02 85-102-01 \E. macranthum (2n=54) 01-09-01 (2n=50) 85-123-23|£ trivervium (2n=60) 86-35-01 Wi-mE pallidum (2n=52)

Fig. 2.10 Pedigrees of the three Exacum Styer Group genotypes. The most common chromosome numbers of the three genotypes are proposed in parentheses. The chromosome numbers of parental taxa are shown.

43 Chapter 3. Stressor effects on in vitro development of microspores from a fertile genotype of Exacum Styer Group

3.1 Introduction In normal sexual reproduction of angiosperm species, the 2n chromosome number is reduced via meiosis to yield In microspores and megaspores. Following meiosis, each microspore undergoes an asymmetrical mitotic cellular division that results in one larger vegetative cell and one smaller generative cell within the mature microspore. One last mitotic division of the generative cell leads to the formation of the two sperm cells that take part in double fertilization. (Chapterl; Fig. 1.3). When a sperm cell fertilizes an egg cell, a diploid zygote (eventual embryo) is formed and the 2n chromosome number is reestablished. However, this normal process is susceptible to purposeful manipulation whereby embryo formation can be separated from the fertilization process. In vitro production of embryos from microspores (i.e., haploid embryos) without fertilization was first reported from cultured anthers of Datura innoxia Mill. (Guha and Maheshwari 1964; 1966) and is termed androgenesis (Reynolds, 1997). Since this first observation, many refinements to the technique have been achieved and include a better understanding of the reprogramming requirements, tissue manipulation, and culture conditions (Sopory and Munshi, 1996). This research has resulted in the generation of haploid plants of many plant species (e.g., tobacco, barley, asparagus etc.) that can either be used directly or used to form double haploid (DH) individuals. DH plants, due to their homozygosity, open valuable opportunities in plant breeding/genetic research and developmental biology (Kasha and Maluzynski, 2003). However, most plants are recalcitrant to androgenetic induction. Apparently, lack of understanding surrounding the mechanism of androgenesis hinders further application of this technology. Specifically, Maraschin et al. (2005b) identified three critical phases for successful DH induction: (1) acquisition of embryonic potential, (2) initiation of cell division, and (3) embryo-like pattern formation. Reprogramming gametophyte development is the first key step in inducing haploid plant production. Gametogenesis is susceptible to artificial stress and has been successfully altered to induce haploid embryo formation. For example, temperature

44 shock, mannitol treatment, and plant growth regulators, etc. can induce microspores to switch from a microgametophytic pathway to an embryonic pathway (Touraev et al, 1997). Developmental stage is also crucial for successfully inducing androgenesis (Maheshwari et al., 1980). In order to gain embryonic potential, the reprogramming treatment should be applied at a specific time during microspores' developmental cycle; commonly near the time of the first pollen mitotic division (Chapter 1; Fig. 1.3; Fig 1.4, Steps 1&2) (Raghavan, 1976). However, the timing for haploid induction is often species and genotype dependent (Sopory and Munshi, 1996). For example, optimal timing in Lycopersicon esculentum Mill, appears to be early in microspore development (i.e., between meiotic prophase I to telophase II; Chapter 1; Fig 1.3 Step 1) (Shtereva et al, 1998), while in Nicotinana tabacum L. cultivars, it is as late as the tetrad stage (Nakata and Tanaka, 1968). Regardless of individual species variation, reprogramming treatments are most often successful when applied during the early stages of microspore development while when applied during the later stages (i.e., when the vegetative cell starts to accumulate starch), androgenesis can no longer to be induced (Binarova et al, 1997; Touraev et al, 1997). The most common reprogramming treatments to induce androgenesis include temperature shock and mannitol treatment (Sangwan and Sangwan-Norreel, 1987; Dunwell et al, 1983). It is thought that these stressors disrupt normal metabolism and alter cytoskeleton development in the cytoplasm (Maraschin et al, 2005b). Morphologically, induced microspores display star-like vacuoles with subsequent symmetrical cell division during the first mitotic division, rather than the normal asymmetrical division (Maraschin et al, 2005b). The star-like structure is formed by a central nucleus with fragmented vacuoles. This phenomenon has been observed in several species including Hordeum vulgare L. (Maraschin et al, 2005a), Triticum aestivum L. (Indrianto et al, 2001), Brassica napus L. cvs Topas and Optima (Zaki and Dickinson, 1991), and ./V. tabacum (Touraev et al, 1996a and b). Treating cells with low or high temperatures can induce acceptable levels of androgenesis, but efficacy is species specific (Sopory and Munshi, 1996). In addition to temperature manipulation, mannitol is commonly used to reprogram microspore development. Exposing barley anther cultures to 0.3 M mannitol for 4 days (at 25°C), instead of a metabolizable sugar, was

45 found to be superior to other pretreatments requiring cold exposure (Roberts-Oehlschlager and Dunwell, 1990). However, a combination of heat shock and mannitol treatments to H. vulgare anthers induced a greater percentage of embryogenic microspores and a greater number of total embryos (Tourave et al, 1996a) than non-stressed cultures (Stauffer et al., 1991). Effective haploid production protocols are presently available for commonly grown plants, including Zea mays L. (Barnabas, 2003), Asparagus officinalis L. (Wolyn and Nichols, 2003) and several Brassica species (Ferrie, 2003). However, there are no haploid embryo production protocols for any Exacum species or hybrid.

Once successful reprogramming of microspore development has occurred, proper and effective culture conditions (i.e., media composition, environmental conditions) for the explants need to be developed (Sopory and Munshi, 1996). Three commonly used basal media formulas that have been shown suitable for anther culture include: Murashige and Skoog (1962), B5 (Gamborg et al, 1968) and Nitsch & Nitsch (Nitsch and Nitsch, 1969). However, each of these media generally requires slight modifications to the levels of specific components (i.e., plant growth regulators, minerals, organic compounds) (Sopory and Munishi, 1996). In most cases, 2-3% sucrose is incorporated as the sole carbon source for tissue use (Beyl, 2000).

Plant growth regulators (PGRs) are either natural or artificial chemicals that regulate physiological processes within a plant. They are commonly incorporated into plant tissue culture media to regulate explants growth and organ development (Beyl, 2000). In anther/microspore culture, various auxins, cytokinins and combinations are used to induce calli/embryo formation (Maheshwari et al, 1982). For example, in cereal anther/microspore culture, 2,4-Dichlorophenoxyacetic acid (2,4-D) (0.5-2 mg l"1) and kinetin (0.5-1 mg l"1) were employed (Wang and Hu, 1984), while in the dicotyledon species, A. officinalis (Peng et al, 1997) and Helianthus annuus L. (Saji and Sujatha, 1998), the combination of BA (1 mg I"1) and NAA (2 mg 1"') was effective. Also, BA (0.3 mg I"1) combined with 2,4-D (0.5 mg l"1) was successfully used in Cucumis sativus L. (Ashok Kumar et al, 2003) (Table 1.2). Based on these reports, optimal PGR combination to induce embryogenesis appears to be species dependent and must be determined empirically for each new plant evaluated.

46 In this chapter, I (1) characterize the cytology of microspore development under various stress treatments and (2) evaluate several combinations of plant growth regulators (i.e., BA and 2,4-D combinations; BA and NAA combinations and Kin and NAA combinations) on calli/ embryo induction from microspores.

3.2 Material and Methods 3.2.1 Plant materials and tissue collection The fertile genotype 01-09-01 was micropropagated in the plant tissue culture laboratory of UBC's Botanical Garden and Center for Plant Research. The micropropagation protocol used was as follows: explants approximately 1-1.5 cm long with two axillary meristems were excised from 6-8 week-old mother plants and all leaves trimmed to 0.7-1 cm in length. The explants were cultured on multiplication medium which contained Anderson macro and micronutrients (Anderson, 1984) and supplemented with 20 g l"1 sucrose, 0.5 mg l"1 MES, 0.5 mg 1"' polyvinylpyrrolidone (PVP), 0.6 g l"1 activated charcoal, 3 g l"1 agar and 1.5 g l"1 phytagel. The pH was adjusted to 5.8 with 1 N sodium hydroxide and/or 1 N hydrochloric acid. Plant growth regulators incorporated for multiplication were 2iP (1-2 mg l"1) and NAA (0.01 mg l"1). Shoots 10 cm in length were harvested, basally treated with 10 mM NAA solution for 10 minutes, stuck into cell-packs filled with media (1 peat: 1 perlite), and placed into a mist chamber for 3 weeks. Rooted cuttings were transferred to the Horticulture Greenhouse and transplanted into 10 cm pots filled with a commercial potting mix (West Creek Farms, BC, Canada; 75% peat and 25% perlite plus a starter charge of NPK; pH adjusted by dolomite and limestone to between 5.5 and 6.5). The temperature of the greenhouse was maintained at 22-24°C day and 20-24°C night. From June through September, plants were grown under natural light, and while during the rest of the year, an 18 hr (6:00-22:00) ambient light was supplemented with 200 umoles m"2 sec"1 supplied by high-pressure sodium lamps. Humidity varied from 50-80% depending on time of year and weather with the most humid conditions typically during summer. Plants were sub-irrigated as needed with fertilizer 15-5-15 Cal-Mag fertilizer (The Scotts Company, Marysville, Ohio). Zinc-EDTA (Plant Products Co. Ltd., Brampton, Ontario) at 10-15 g l"1 was applied every second week as a prophylactic against zinc deficiency. Sulfur

47 (Safer Ltd. Scarboruogh, Ontario) at 12 g l"1 was applied every week to control powdery mildew. Visible buds were observed at week 12. These buds were then collected and used for subsequent experiments. Previous studies characterized the in vivo development of the male gametophyte for this genotype and concluded that anthers 1-1.1 cm in length contained microspores at the late uninucleate to early binucleate stage. Therefore, buds with anthers 0.8-1.2 cm in length were collected for treatment. Anthers were divided into two groups based on anticipated developmental stage: 0.8-0.9 cm in length (i.e., mid to late uninucleate) and 1-1. lcm in length (i.e., late uninucleate to early binucleate). All bud collections were made between 9:00-13:00.

3.2.2 Media preparation Two basal media, MS salts (Murashige and Skoog, 1962) and N&N salts (Nitsch and Nitsch, 1969) without plant growth regulators were used for cytological investigations. Both media were supplemented with 30 g l"1 sucrose, except for the mannitol stress. All media were solidified with 8 g l"1 agar (Sigma, Canada). The pH of the medium was adjusted to 5.8 with 1 N sodium hydroxide and/or 1 N hydrochloric acid prior to autoclaving at 121 °C for 20 min. In addition, any plant growth regulators were added before autoclaving.

3.2.3 Disinfestation and culture procedures Flower buds were first disinfested in 40% bleach (2.4% sodium hypochlorite) for 15 min followed by three rinses with sterile distilled water. Anthers were aseptically excised and cultured in 60 x 15 mm Petri dishes (Fisher Scientific In., Canada), each containing 10 ml media. Filaments were removed from the anther as best as possible. All cultures were maintained at 25°C in the dark.

3.2.4 Reprogramming treatments Temperature Temperature treatments were applied to excised cultured anthers as either cold (10°C for 7 days) or heat (35°C for 4 days) exposure. All temperature treatments were applied

48 in the dark. For the control treatments, cultures were maintained at 25°C for the same duration as the treated tissue. Mannitol stress For the mannitol stress, the MS and N&N media described above were adjusted to contain 63.8 g l"1 (0.7 M) mannitol instead of sucrose. Tissue was maintained on these media for either 4 or 7 days depending on temperature treatments before the cytological observations were made. All other culture conditions were as described above.

Plant growth regulators MS base medium supplemented with one of three combinations of auxin and cytokinin were evaluated for calli/embryo production (Table 3.1). All other culture conditions were as described above.

3.2.5 Cytological examinations For each experimental unit (i.e., one bud with five anthers), one anther was randomly selected at predetermined intervals (i.e., 0 and day 4 or 7 days depending on reprogramming treatment) and fixed in Carnoy's solution (3 parts 95% ETOH: 1 part acetic acid) for 24 h at 4°C. After fixation, anthers were washed three times with distilled water and kept in 70% ETOH at 4°C until examination. For staining, a whole anther was macerated in a drop of 4'-6-diamidino-2-phenylindole (DAPI, 5 ug ml"1), the coverslip applied with gentle pressure, and allowed to stand for five min in the dark. DAPI stained tissue was observed with a Zeiss Axioplan 2 epiflorescent microscope (Carl Zeiss Inc., North America) at 60X and 100X mag. Digital micrographs were recorded at the beginning of the experiment and at the conclusion of each treatment period to assess any change in microspore development.

3.2.6 Statistics Temperature and mannitol stress Treatments to experimental units (i.e., individual buds containing five anthers) were applied in a split-plot experimental design. The main plots were media with 4 levels (MS, N&N, sucrose, mannitol), and subplots were temperatures with 2 levels of either

49 cold (10°C and 25°C) or heat (25°C and 35°C). Each treatment had five replications. The whole experiment was repeated three times.

Plant growth regulators Treatments to experimental units (i.e., individual buds containing five anthers) were applied in a split-plot experimental design. The main plots were PGR combinations (e.g., combinations I and II were with 5 levels and combination III was with 12 levels) (Table 3.1), and subplots were temperatures with 2 levels (10°C and 25°C). Each treatment had eight replications. The whole experiment was repeated three times. The observations were made after 6 weeks of culture.

3.3 Results 3.3.1 Effects of temperature and media composition on microspore development DAPI stain showed that anthers 0.8-0.9 cm in length contained microspores at the mid to late uninucleate stage, while anthers 1-1.1 cm in length contained microspores at the late uninucleate to binucleate stage. Following seven days of cold treatment, microspores were found to have undergone normal asymmetrical mitotic division with no deviations observed (Fig. 3.1; A-D). In contrast, all microspores from the control cultures kept at constant room temperature (25°C) contained degraded nuclei (Fig. 3.1; E-H). Following four days of heat treatment, all heat-treated microspores contained degraded nuclei (Fig 3.2; A-D). Similar to the cold treatment control cultures, the heat treatment control cultures also contained microspores with mostly degraded nuclei with less than 1% appearing normal (Fig 3.2; E-H). There were no significant effects of initial developmental stage or media on microspore development, regardless of temperature treatment or carbon source (Table 3.2; 3.3).

3.3.2 Effects of temperature and plant growth regulator combinations on calli formation Neither calli nor embryos emerged from anthers cultured with any of the PGR combinations evaluated (Table 3.4 and 5). However, after two weeks of culture on all PGR supplemented media, calli appeared from the attachment site of the filament, perhaps from remnant filament tissue.

50 3.4 Discussion Both the timing of and type of reprogramming stress are critical factors in disrupting normal microspore development and inducing embryogenesis for haploid embryo production. In an attempt to separate these two issues, I chose to concentrate on specific reprogramming treatments at a given developmental stage. All stress treatments were applied to anthers containing microspores between the mid-uninucleate and early binucleate stages. Unfortunately, no androgenic calli or embryos were induced from this developmental period, regardless of reprogramming treatment. In general, crop breeders agree that the most effective time during microspore development for reprogramming is between the late uninucleate to the early binucleate stage (Reynolds, 1997); however, significant species variation exists. For example, N. tabacum cultivars, microspores as early as the tetrad stage still maintained embryogenic potential (Nakata and Tanaka, 1968). In addition, androgenic studies in L. esculentum showed that sterile mutants at prophase- metaphase I were the most responsive for DH induction (Shtereva et al., 1998). However, my observations from an Exacum sterile genotype following cold treatments during meiosis failed to identify any androgenic response (data not shown). Therefore, the possibility of using sterile genotypes is preserved but that they remain a less likely resource for success.

In addition, a symmetrical division after the 1st mitosis is the most common key indicator of microspore embryogenesis induction (Peng et al., 1997; Hamaoka et al., 1991). In my anther culture experiments, I did not observed any symmetrical embryogenic divisions but instead, only normal asymmetrical gametophytic divisions. However, reports indicate that haploid embryos can still be induced after this normal asymmetrical division from either of the resulting generative (e.g., Hyoscyamus niger L.; Raghavan, 1976; 1978) or vegetative cells (e.g., Z. mays; Barnabas et al., 1987). I would therefore propose that anthers of E. Styer Group may still have embryogenic potential and future research using a broader range of microspore developmental stages should be evaluated. However, many factors other than timing could have also influenced my results. These factors include genotypes specificity for embryogenic potential and reprogramming treatments.

51 The genotype of the donor plant plays an important role in the androgenic induction (Maheshwari et al, 1980). In this research, I only evaluated one genotype (01-09-01) of E. Styer Group. However, research on N. tabacum anther cultures showed that only 5 of 12 genotypes were responsive (Nitsch, 1969). In addition, only 3 out of 43 lines in Vitis vinifera L. (Gresshoff and Doy, 1974) and 3 out of 43 lines in Arabidopsis thalinana (Gresshoff and Doy, 1972) were responsive to the anther culture. In addition, significant genotypic variation exists in E. Styer Group for general responses to plant tissue culture (Riseman and Chennareddy, 2004). Therefore, I suggest that evaluation of additional genotypes of this group may lead to the identification of responsive genotypes useful for androgenic induction.

Reprogramming microspore development with temperature treatments of various magnitudes and durations have been an effective stress for androgenesis induction. Unfortunately, observations from my heat treatments (i.e., 35°C for 4 d) indicate the stress intensity (i.e., combination of magnitude and duration) used was deleterious for this germplasm and resulted in complete nuclei degradation. In general, a relatively narrow range of temperatures i.e., between 30-35°C, have been effective for reprogramming when applied to anther cultures (Sopory and Munshi, 1996). However, a much broader range of exposure durations have been effective in inducing androgenesis and range from 12 h for Brassica oleracea var. italica (i.e., broccoli) (Arnison et al., 1990) to 6 d in T. aestivum (i.e., spring wheat) (Li et al., 1988). The tolerance of microspores to temperature treatments is species and often genotype dependent (Sopory and Munshi, 1996). Due to complete nuclei degradation following my heat treatment, I conclude that the temperature intensity selected was inappropriate for this species (or these genotypes) and future work should concentrate on modifying treatment intensity by both modulating the magnitude and the duration of the treatment. Cold exposure is an alternative temperature stress that has been effective for androgenesis induction. In my research, microspores exposed to the cold treatment (10°C for 7 d) remained healthy with normal shaped nuclei throughout the experiment. This suggests that the intensity of cold stress used was less deleterious than the heat stress; however, it did not induce symmetrical division. The first successful production of haploid plants following cold treatment was in D. innoxia (Nitsch and Norrel, 1973).

52 Since then, the cold treatments have been widely used and recommended for many recalcitrant species (Maheshwari et al, 1980; 1982). Therefore, I suggest that cold treatments may still provide an opportunity for microspore reprogramming. However, the intensity of the cold treatment did not appear effective and future studies should evaluate expanding the range of both the magnitude and duration of the treatment. In anther culture, medium composition can have a significant effect on androgenesis induction with both carbon source and PGRs implicated. Replacing sucrose with mannitol has been successfully used to induce androgenesis (Tourave et al., 1997). Unfortunately, Exacum microspores did not respond to the mannitol treatment. In mannitol-based media, microspores are exposed to two stressors simultaneously, carbohydrate restriction and osmotic pressure. This combined effect has successfully switched the normal gametogenesis into androgenesis in H. vulgare (e.g., barley) anther culture (Wojnarowiez et al, 2004). In addition, mannitol treatments are reported to promote embryo induction by (1) increasing microspore DNA replication and chromosome doubling (Li et al, 1995a), (2) increasing endogenous hormone levels and peroxidase activity (Li et ah, 1995b), and (3) inducing changes in soluble protein composition in the anthers (Li and Hu, 1995). As with temperature, the duration of mannitol exposure is crucial in reprogramming efficiency. For example, in Malus domestica (e.g., apple) microspore culture, extending mannitol exposure from 2 d to 4 d decreased the efficiency of embryo formation due to reduced pollen viability during the treatment process (Hofer, 2004). From my observations, regardless of media, there were significant differences in nuclei development when anthers were cultured under various temperature conditions. These results suggest that temperature treatments play a more important role in influencing Exacum microspore development than the mannitol treatments. Therefore, future research on evaluating stress treatments should focus on temperature modification and not on carbon source.

Choosing appropriate combinations and concentrations of plant growth regulators is crucial in callus or embryo formation in anther culture (Sopory and Munshi, 1996). Based on established protocols on cereals (Wang and Hu, 1984) and dicots anther/microspore culture protocols (i.e., A. officinalis (Peng et al., 1997) and Hellanthus annuus (Saji and Sujatha, 1998)), three auxin and cytokinin combinations were evaluated

53 (Table 3.1). Unfortunately, the three PGR combinations selected did not successfully induce embryos or calli from microspores. In addition to the three combinations I selected, it was reported that the combination of IAA (0.1 mg l"1) and BA (2.5 mg l"1) was successful in potato anther culture (Tai and Xiong, 2003), while IBA (0.2 mg l"1) and Kinetin (0.2 mg l'1) was successful in apple haploid plant production (Hofer, 2003). In addition to single auxin and cytokinin combinations, multiples of each have been effective. For example, two auxins (0.02 mg l"1 NAA and 0.02 mg l"1 2,4-D) combined with three cytokinins (0.5 mg l"1 BA, 1 mg 1-1 Kinetin, and 0.5 mg l"1 Zeatin) were successful for embryo induction in citrus anther culture (Germana, 2003). Therefore, additional evaluations using a broader range of PGRs as well as more complex combinations should be tested for the ability to induce androgenesis in Exacum. Regardless of the PGRs used to promote calli production or organogenesis, the effected progenitor tissue is critical in producing the desired products. In order to obtain haploid calli or embryos, microgametes must be the progenitor tissue; all other anther tissue will yield 2n material. Based on the position of the anther on the media and the region from which calli arose, I conclude that the calli formed in all cultures was from filament remnants and not from microgametes. Filament removal is a common procedure in anther culture (Keller, 1984) often producing four to five times more embryos than when stamens are left intact (Arnison et al., 1990). However, embryo formation of high androgenic capacity species or genotypes will not be influenced by this procedure. A possible negative role of filaments in anther culture is as a nutrient sink depriving the microspores of nutrients and growth regulators required for embryogenesis (Arnison et al., 1990). In order to avoid this potential problem, future research should evaluate microspore culture as an alternative to anther culture.

This report is the first evaluation of E. Styer Group microspore development and the effects of reprogramming treatments on androgenesis. Although the treatments evaluated failed to successfully reprogram normal gametophytic development, descriptions of basic biological development were developed and indicate several areas for future research. The most promising area to focus future investigations is on the magnitude and duration of the cold treatments as these were viewed as the least lethal treatments. However, additional research should evaluate a broader range of microspore

54 developmental stages, additional auxin and cytokinin combinations, and additional genotypes.

55 Table 3.1 Plant growth regulator treatments applied to anthers of Exacum Styer Group genotype 01-09-01. Plant Growth Concentration (mg l"1) Regulators Combination BA 0 0.1 0.5 0.5 1 I 2,4-D 0 1 1 2 2 Combination BA 0 0.1 0.5 0.5 1 II NAA 0 1 1 2 2

Combination Kin 0 1 2 III 2,4-D 0 0.5 1 2 0 0.5 1 2 0 0.5 1 2

56 Table 3.2 Cold treatment and media effects on pollen development and calli formation from the Exacum Styer Group genotype 01-09-01.

Media2 Temperature Duration Anther Microspores Microspores Microspores (°C) (days) length with one with two with degraded (cm) nucleus or symmetrical nuclei (%)y two nuclei (%)y asymmetrical nuclei (%)y MSS 25 7 0.8-0.9 0 0 100 25 7 1-1.1 0 0 100 NNS 25 7 0.8-0.9 0 0 100 25 7 1-1.1 0 0 100 MSS 10 7 0.8-0.9 100 0 0 10 7 1-1.1 100 0 0 NNS 10 7 0.8-0.9 100 0 0 10 7 1-1.1 100 0 0

MSM 25 7 0.8-0.9 0 0 100 25 7 1-1.1 0 0 100 NNM 25 7 0.8-0.9 0 0 100 25 7 1-1.1 0 0 100 MSM 10 7 0.8-0.9 100 0 0 10 7 1-1.1 100 0 0 NNM 10 7 0.8-0.9 100 0 0 10 7 1-1.1 100 0 0 ZMSS: Basal MS media supplemented with 30 gl"1 sucrose; NNS: Basal N&N media supplemented with 30 gl"1 sucrose; MSM: Basal MS media supplemented with 63.8 g 1 mannitol; NNM: Basal N&N media supplemented with 63.8 g l"1 mannitol. y Overall pollen viability was 54.5 (± 4.85 %) with only viable microspores containing nuclei. If over 50 % microspores contained normal nuclei, it was designated as 100 %.

57 Table 3.3 Heat treatment and media effects on Exacum Styer Group genotype 01-09-01 pollen development and calli formation.

Media2 Temperature Duration Anther Microspores Microspores Microspores (°Q (days) length with one with two with degraded (cm) nucleus or symmetrical nuclei (%)y two nuclei (%)y asymmetrical nuclei (%)y MSS 25 4 0.8-0.9 1 0 99 25 4 1-1.1 1 0 99 NNS 25 4 0.8-0.9 1 0 99 25 4 1-1.1 1 0 99 MSS 35 4 0.8-0.9 0 0 100 35 4 1-1.1 0 0 100 NNS 35 4 0.8-0.9 0 0 100 35 4 1-1.1 0 0 100

MSM 25 4 0.8-0.9 1 0 99 25 4 1-1.1 1 0 99 NNM 25 4 0.8-0.9 1 0 99 25 4 1-1.1 1 0 99 MSM 35 4 0.8-0.9 0 0 100 35 4 1-1.1 0 0 100 NNM 35 4 0.8-0.9 0 0 100 35 4 1-1.1 0 0 100 MSS: Basal MS media supplemented with 30 gl"1 sucrose; NNS: Basal N&N media supplemented with 30 gl"1 sucrose; MSM: Basal MS media supplemented with 63.8 g 1" mannitol; NNM: Basal N&N media supplemented with 63.8 g l"1 mannitol. y Overall pollen viability was 54.5 (± 4.85 %) with only viable microspores containing nuclei. If over 50 % microspores contained normal nuclei, it was designated as 100 %. 1% represents only a few nuclei intact.

58 Table 3.4 Cold treatment and plant growth regulator combination effects (2,4-D and BA or 2,4-D and NAA) on calli formation of Exacum Styer Group genotype 01-09-01. Duration of 10°C Auxin Cytokinin No. of anthers Calli formation treatment (days) (mgf1) (mgl"') cultured (%)

2,4-D BA 0 0 0 120 0 1 0.1 120 0 1 0.5 120 0 2 0.5 120 0 2 1 120 0

7 0 0 120 0 1 0.1 120 0 1 0.5 120 0 2 0.5 120 0 2 1 120 0

NAA BA 0 0 0 120 0 1 0.1 120 0 1 0.5 120 0 2 0.5 120 0 2 1 120 0

7 0 0 120 0 1 0.1 120 0 1 0.5 120 0 2 0.5 120 0 2 1 120 0

59 Table 3.5. Cold treatment and plant growth regulator combination (2,4-D and Kinetin) effects on calli formation of Exacum Styer Group genotype 01-09-01. Duration of cold Auxin Cytokinin No. of anthers Calli formation treatment (days) (mgr1) (mgl"1) cultured (%)

2,4-D Kinetin 0 0 0 120 0 0.5 0 120 0 1 0 120 0 2 0 120 0 0 0.5 120 0 0.5 0.5 120 0 1 0.5 120 0 2 0.5 120 0 0 1 120 0 0.5 1 120 0 1 1 120 0 2 1 120 0

7 0 0 120 0 0.5 0 120 0 1 0 120 0 2 0 120 0 0 0.5 120 0 0.5 0.5 120 0 1 0.5 120 0 2 0.5 120 0 0 1 120 0 0.5 1 120 0 1 1 120 0 2 1 120 0

60 D A c • m • *• * %

i * H E F G

i r «i * .* •>••*, f ^ r mmmS*

Fig. 3.1. Microspore nuclei development from Exacum Styer Group genotype 01-09-01 after 7 days of culture on various base m supplemented with either sucrose or mannitol cultured under either room temperature or cold (10°C). (A-D) 10°C treatments w different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. (E-H) 25°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. Scale 10 um. A B t D

1p- mam w \ \ ^ •m

* •

E F G H

« * • -t m > •

Fig. 3.2 Microspore nuclei development from Exacum Styer Group genotype 01-09-01 after 4 days of culture on various base media supplemented with either sucrose or mannitol cultured under either room temperature or heat (35°C). (A-D) 35°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. (E-H) 25°C treatments with different media: MS with sucrose; N&N with sucrose; MS with mannitol; and N&N with mannitol, respectively. Scale 10 um. Chapter 4. Concluding remarks and future research

The main objectives of this research were to characterize microspore development in Exacum Styer Group and identify key issues related to haploid induction. Many factors affect haploid induction and vary depending on species or genotype, timing and magnitude of reprogramming treatments, and culture conditions. Therefore, the first part of my thesis addressed pollen developmental biology of several genotypes of E. Styer Group ranging in pollen viability. The second part of my thesis evaluated how developmental stage, media, and stressors affect microspore development. Chapter 2 details pollen developmental biology of E. Styer Group. Among genotypes, pollen viability was highly variable ranging from sterile to over 80 % viability. Although pollen viability was variable, anther lengths correlated well with microspore developmental stage. The three selected genotypes exhibited normal meiosis regardless the fertility. In addition to detailing microsporogenesis and microgametogenesis, pollen morphology and nuclei development among the genotypes was also evaluated. In general, non-viable microspores exhibited a shrunken (empty appearance), misshapen exine morphology and nuclei that degraded before anthesis. In contrast, viable pollen exhibited a non-shrunken (full appearance), smoother exine sculpture with several well-formed germination pores. All pollen produced normal nuclei following meiosis but sometime between the completion of meiosis and the start of mitosis I, reduced fertility genotypes deviated from the normal developmental process to yield non-viable pollen grains. I hypothesize that in reduced fertility genotypes, a tapetum layer with altered function is negatively affecting the production of viable pollen. This hypothesis may be tested by observations with a transmission electron microscope (TEM) of ultra thin sections of anthers appropriately stained for key components or processes. In addition, TEM may be useful in identifying aberrant cytological changes that may occur during androgenesis.

This is the first report of chromosome numbers for E. Styer Group genotypes. The chromosomes were mostly dot-shaped except for two pairs of medium-long chromosomes. In order to develop a stronger link between chromosome numbers and the duration of meiosis, genomic DNA content should be determined by flow cytometry. Furthermore, once the haploid calli/embryos are induced from anther culture, the ploidy

63 level will need to be determined through similar methods. Therefore, nuclei isolation protocols for E. Styer Group need to be established in the near future. Chapter 3 characterizes pollen development under in vitro conditions following stress treatments. In general, the most effective androgenic induction time is between the late uninucleate to the early binucleate stage (Reynolds, 1997). In this study, the initial developmental stages for reprogramming treatments were fixed during mid-uninucleate stage to early binucleate stage. Unfortunately, microspores during this developmental period did not show any androgenic development. However, in N. tabacum cultivars, microspores as early as the tetrad stage still maintained embryogenic potential (Nakata and Tanaka, 1968). Therefore, the possibility of successfully using earlier developmental stages is preserved. Despite this report, I feel earlier developmental stages are not a strong source tissue for androgenic reprogramming. I suggest that microspores at the uninucleate to early binucleate stage in E. Styer Group should be the primary developmental stage tested for future reprogramming treatments. In the androgenic induction, genotype fertility is a key factor to be considered. In this research, I only evaluated one fertile genotype of E. Styer Group. This genotype did not respond to the reprogramming treatments. However, in L. esculentum, male sterile genotypes were successfully used to produce haploid embryos when reprogramming treatments were applied to microspores during meiosis. However, this reprogramming process did not occur in E. Styer Group (data not shown) and suggests that sterile genotypes are not suitable material for androgenic induction. I suggest that only pollen fertile genotypes are used for future experiments.

Neither temperature treatments tested successfully reprogrammed microgametogenesis. Moreover, the heat treatments had substantial negative effects on nuclei development. In addition, evaluations of media composition, i.e., carbon source, PGR, indicate that neither component had a significant impact on androgenic reprogramming but rather, indicated that temperature may have a more important role in influencing nuclei development. In particular, when anthers were cultured under room temperature, all nuclei gradually degraded, regardless of media composition. Therefore, in order to maintain normal nuclei development under control conditions, suitable culture conditions (e.g., temperature, light or dark, basal salts composition, plant growth

64 regulators combinations, charcoal supplement etc.) need to be developed. Among the auxin and cytokinin combinations evaluated, none successfully induced androgenic calli/embryos. The concentrations of auxins and cytokinins I evaluated ranged from 0 to 2 mg l"1. I suggest future research should evaluate higher concentrations of PGRs (i.e., 5 or 10 mg l"1). Furthermore, other auxin and cytokinin types, and their combinations, should be valuated for androgenic induction. In addition to the three main factors evaluated, genotype of the donor plant should be considered. Unfortunately, 1 only evaluated one genotype that was not responsive to the induction treatments. Therefore, future research should evaluate a wide range of genotypes for androgenic potential. Under PGR conditions, calli were observed. These calli appeared to be derived from remnant filament tissue and not microgametes. Unfortunately, calli/embryos formed from 2n anther tissues are not useful for haploid production. In order to address this issue, microspore isolation and culture protocols should be established. Successful haploid induction involves the interaction of many factors. Besides the factors I detailed above, other factors are reported to influence the production of haploid calli or embryos. These factors include the physiological role of the donor plants conditions (e.g., the age of the donor plants, the growing conditions), other reprogramming treatments (e.g. carbon dioxide treatment (Raina, 1989); gamma-radiation (Laurain et al, 1993)). By characterizing pollen developmental biology of E. Styer Group and identifying key issues related to haploid induction, my research provides useful information that promotes the successful development of haploid plants for this group.

65 Appendix I Characteristics of pollen development and chromosome numbers from additional genotypes in Exacum Styer Group

Al. 1 Introduction This experiment was intended to provide additional information on pollen development of Exacum Styer Group using additional genotypes. Chapter 2 detailed that in the genotypes evaluated, anther length were positive associated to meiosis stage and that meiosis was initiated when anthers reached 0.4-0.5 cm in length with microspores released from tetrads when anthers reached 0.5-0.6 cm in length. At anthesis, viable pollen were stained by acetocarmine and hydrated properly.

Al. 2 Material and Methods AL 1.2.1 Plant materials E. Styer Group was bred from several species and taxa native to Sri Lanka (i.e., E. pedunculatum, E. macranthum, E. pallidum, E. trinervium, and E. trinervium subsp. ritigalensis) (Riseman et al., 2005). Genotypes included were selected from the E. Styer Group breeding program at the University of British Columbia, Vancouver, BC. In total, 3 genotypes (01-47-21, 01-83-17, and 04-15-01) were used for this part of the research. All experimental plants were derived from seed or micropropagated cuttings. The cuttings or seedlings were transferred to the Horticulture Greenhouse and transplanted into 10 cm pots filled with a commercial potting mix (West Creek Farms, BC, Canada; 75 % peat and 25% perlite plus a starter charge of NPK; pH adjusted by dolomite and limestone to between 5.5 and 6.5). The temperature of the greenhouse was maintained at 22-24°C day and 20-24°C night. From June through September, plants were grown under natural light, and while during the rest of the year, ambient light was supplemented with 200 umoles supplied by high-pressure sodium lamps. For supplemental lighting, an 18 hr (6:00-22:00) photoperiod was used. Humidity varied from 50-80% depending on time of year and weather with the most humid conditions typically during summer. Plants were sub-irrigated as needed with fertilizer 15-5-15 Cal-Mag (The Scotts Company, Marysville, Ohio). Zinc-EDTA (Plant Products Co. Ltd., Brampton, Ontario) at 10-15 g

66 r was applied every second week as a prophylactic against zinc deficiency. Sulfur (Safer Ltd. Scarboruogh, Ontario) at 12 g f1 was applied every week to control powdery mildew.

ALL2.2 Microsporogenesis and microgametogenesis with acetocarmine stain Three genotypes (01-47-21, 01-83-17, and 04-15-01) were used to characterize microsporogenesis. Flower buds that contained anthers between 0.4-0.8 cm in length were collected between 9:00-13:00. Anthers were excised and fixed for 24 h at 4°C in Carnoy's solution (3 parts 95 % alcohol: 1 part acetic acid) supplemented with ferric chloride (4 g ml"1 in 49 ml Carnoy's solution). After fixation, anthers were washed three times in distilled water and stored in 70% ETOH at 4°C until examination. A minimum often anthers were examined for each genotype and collection time to ensure a sufficient sample size. Microsporogenetic events were observed under a light microscope (Motic B5 Professional Series, Richmond, Canada) at 40X mag, with digital photomicrographs captured with a Nikon Coolpix 4500 camera (Nikon Corp., Japan) attached to the microscope.

AL 1.2.3 Chromosome counts with Giemsa stain Two genotypes (01-37-08 and 01-47-21) were used to evaluate the chromosome numbers. Root tips were used for all chromosome counts. E. Styer Group root-tips were harvested from only actively growing healthy plants in the UBC Horticulture Greenhouse, Vancouver, BC. Root rips were collected between 12:30-13:30 under sunny conditions. Once collected, roots were pretreated for five hours (shaken) with 0.002 M 8-Hydroxyquinoline sulfate (8-HQS). After pretreatment, the roots were washed with distilled water three times and fixed in Carnoy's solution (3 parts 95% alcohol: 1 part acetic acid) for 24 h at 4°C. For storage until use, roots were washed with distilled water three times and store in 70% ETOH at 4°C. For observation, roots were washed with distilled water three times and digested for 2 h at 37°C in a 8.25 U ul"1 pectinase plus 6.55x 10"3 U ui"1 cellulase solution.. After digestion, roots were again washed with distilled water three times. Tips were then macerated on slides in a drop of methanol and acetic acid mix (3 parts 100% methanol: 1 part acetic acid). The solution

67 was ignited to burn off the solution and to fix the cells on the slide. Once dried, the

tissue was stained with Giemsa solution (3 ml Giemsa stain, 30 ml 0.067 M Na2HP04, 30 ml 0.067 M KH2PO4) for five minutes, washed under tap water to remove excess dye, and allowed to air-dry. Observations and chromosome counts were made with a light microscope (Zess Axioplan 2 Imaging, Germany) at 100X mag.

Al. 3 Results and Discussion The results from microsporogenesis and microgametogenesis evaluations of genotypes 01-47-21, 01-83-17 and 01-15-01 confirmed that meiosis initiated around anther lengths of 0.4-0.5 cm . The remainder of microspore development followed the pattern I detailed in the main thesis (Fig. Al. 1-3). In addition, the chromosome numbers of two additional genotypes from E. Styer Group were evaluated. However, due to the same difficulties in making appropriately conditioned samples, identical counts were not obtained from all cells (Table Al. 1). Therefore, total chromosome counts were very broad within one genotype, ranging from 50-64 for genotype 01-37-08 and from 48-71 for genotype 01-47-21.

68 Table Al. 1 Chromosome numbers of two additional Exacum Styer Group genotypes. Chromosome number and frequency of observations (in parentheses) are shown.

Genotype Range in chromosome number 01-37-08 50(1) 52(1) 64(1) 01-47-21 48(1) 49(1) 56(1) 71(1)

69 Fig. AL 1 Pollen development in Exacum Styer Group genotype 01-47-21. (A) Prophase I, anther length 0.4 cm; (B) Metaphase II and Telophase II, anther length 0.4 cm; (C) Tetrads, anther length 0.5-0.6 cm; (D) Mature pollen at anthesis, anther length 1.2 cm. Scale bar = 10 um.

70 4* ^ m

Fig AL 2 Pollen development in Exacum Styer Group genotype 01-83-17. (A) Prophase I, anther length 0.5 cm; (B) Tetrads, anther length 0.5-0.6 cm; (C) Pollen at late uninucleate stage, anther length 0.9 cm; (D) Pollen at late uninucleate to early binucleate stage anther length 1.0 cm. Scale bar = 10 um.

71 Fig Al. 3 Pollen development in Exacum Styer Group genotype 01-15-01. (A) Prophase to metaphase II, anther length 0.5 cm; (B) Tetrads, anther length 0.5-0.6 cm; (C) Pollen at mid- to late uninucleate stage, anther length 0.9 cm; (D) Pollen at late uninucleate to early binucleate stage anther length 1.1 cm. Scale bar =10 um.

72 Appendix II Stressor effects on in-vitro development of microspores from a sterile genotype of Exacum Styer Group

AH. 1 Introduction In order to gain embryonic potential, the reprogramming treatment is normally applied at the time near the first pollen mitotic division (Chapter 1; Fig. 1.3; Fig 1.4, Steps 1&2) using pollen fertile genotypes (Raghavan, 1986). However, the timing for haploid induction and successful use of reduced fertility or sterile genotypes are often species and genotype dependent (Sopory and Munshi, 1996). For example, in a Lycopersicon esculentum Mill, male sterile genotype, the androgenic response was induced earlier in microspore development (i.e., between meiotic prophase I to telophase II; Chapter 1; Fig 1.3 Step 1) (Shtereva et al., 1998) than normal. This research attempted to reprogram pollen development of a sterile genotype before the abortive development occurred.

AIL 2 Material and Methods All. 2.1.1 Plant materials and tissue collection The sterile genotype 01-37-37 was micropropagated in the plant tissue culture laboratory of UBC's Botanical Garden and Center for Plant Research. The micropropagation protocol used was as follows: explants approximately 1-1.5 cm long with two axillary meristems were excised from 6-8 week-old mother plants and all leaves trimmed to 0.7-1 cm in length. The explants were cultured on multiplication medium which contained Anderson macro and micronutrients (Anderson, 1984) and supplemented with 20 g f'sucrose, 0.5 mg l"1 MES, 0.5 mg l'1 polyvinylpyrrolidone (PVP), 0.6 g l"1 activated charcoal, 3 g l"1 agar and 1.5 g l"1 phytagel. The pH was adjusted to 5.8 with 1 N sodium hydroxide and/or 1 N hydrochloric acid. Plant growth regulators incorporated for multiplication were 2iP (1-2 mg l"1) and NAA (0.01 mg l"1). Shoots 10 cm in length were harvested, basally treated with 10 mM NAA solution for 10 minutes, stuck into cell-packs filled with media (1 peat: 1 perlite), and placed into a fog chamber for 3 weeks. Rooted cuttings were transferred to the Horticulture

73 Greenhouse and transplanted into 10 cm pots filled with a commercial potting mix (West Creek Farms, BC, Canada; 75% peat and 25% perlite plus a starter charge of NPK; pH adjusted by dolomite and limestone to between 5.5 and 6.5). The temperature of the greenhouse was maintained at 22-24°C day and 20-24°C night. From June through September, plants were grown under natural light, and while during the rest of the year, an 18 hr (6:00-22:00) ambient light was supplemented with 200 umoles m"2 sec"1 supplied by high-pressure sodium lamps. Humidity varied from 50-80% depending on time of year and weather with the most humid conditions typically during summer. Plants were sub-irrigated as needed with fertilizer 15-5-15 Cal-Mag fertilizer (The Scotts Company, Marysville, Ohio). Zinc-EDTA (Plant Products Co. Ltd., Brampton, Ontario) at 10-15 g l"1 was applied every second week as a prophylactic against zinc deficiency. Sulfur (Safer Ltd. Scarboruogh, Ontario) at 12 g l"1 was applied every week to control powdery mildew. Visible buds were observed at week 12. These buds were then collected and used for subsequent experiments. The buds with anthers 0.4-0.9 cm in length were collected for treatment. All bud collections were made between 9:00-13:00.

AH 2.1.2 Media preparation Basal media, MS salts (Murashige and Skoog, 1962) with plant growth regulators 1.0 mg l"1 BA and 3.0 mg l"1 2,4-D were used and supplemented with 30 g l"1 sucrose. Media were solidified with 8 g 1"' agar (Sigma, Canada). The pH of the medium was adjusted to 5.8 with 1 N sodium hydroxide and/or 1 N hydrochloric acid prior to autoclaving at 121 °C for 20 min. In addition, any plant growth regulators were added before autoclaving.

All. 2.1.3 Disinfestation and culture procedures Flower buds were first disinfested in 40% bleach (2.4% sodium hypochlorite) for 15 min followed by three rinses with sterile distilled water. Anthers were aseptically excised and cultured in 60 x 15 mm Petri dishes (Fisher Scientific In., Canada), each containing 10 ml media. Filaments were removed from the anther as best as possible. All cultures were maintained at 25°C in the dark.

74 AH 2.1.4 Reprogramming treatments Temperature Cold treatment (10°C or 5°C) for 0, 3, or 7 days exposure were applied to either excised anthers or whole buds. All temperature treatments were applied in the dark. For the control treatments (0 days for either cold treatment), cultures were maintained at 25°C. All other conditions were as previously described.

Plant growth regulators Culture media were composed of MS base medium supplemented with 2,4-D (0.01, 0.1, or 1 mg l"1) and BA (1, 2, or 3 mg l"1) and were evaluated for ability to induce calli/embryo production

AIL 2.1.5 Cytological examinations For each experimental unit (i.e., one bud with five anthers), one anther was randomly selected at the beginning of the experiment to served as the control developmental stage with another anther collected after 2 weeks culture. Anthers were fixed in Carnoy's solution (3 parts 95 % alcohol: 1 part acetic acid) supplemented with ferric chloride (4 g ml"1 in 49 ml Carnoy's solution) for 24 h at 4°C. After fixation, anthers were washed three times with distilled water and kept in 70% ETOH at 4°C until examination. For staining, a whole anther was macerated in a drop of acetocarmine solution and observed under a light microscope (Motic B5 Professional Series, Richmond, Canada) at 40X mag. Digital photomicrographs were captured using a Nikon Coolpix 4500 camera (Nikon Corp., Japan) attached to the microscope.

AIL 2.1.6 Statistics Temperature Treatments to experimental units (i.e., individual buds containing five anthers) were applied in a split-plot experimental design. The main plots were temperatures with 2 levels of either cold (5°C or 10°C) and subplots were durations with 3 levels (0, 3, or 7 days). Each treatment had three replications. The whole experiment was repeated

75 three times.

Plant growth regulators Treatments to experimental units (i.e., individual buds containing five anthers) were applied in a split-plot experimental design. The main plots were PGR combinations with 9 levels (2,4-D 0.01, 0.1 or 1 mgl"1 and BA 1, 2 or 3 mgl"1) and subplots were cold treatments (5°C) durations with 2 levels (0 or 7 days). Each treatment had three replications. The whole experiment was repeated three times. Observations were made after 6 weeks of culture.

AIL 3 Results and Discussion From earlier observations of this genotype, anthers 0.4 - 0.9 cm in length were found to contain pollen mother cells near the end of meiosis, or pollen grains at the tetrad stage. After 2 weeks' culture, those microspores exposed to low temperatures/plant growth regulator combinations developed into abnormal oblong shape pollen grains that were morphologically similar in appearance to pollen produced under greenhouse conditions. No anthers produced embryos or calli in this experiment. Apparently, treatments applied to anthers or intact buds had no effect on pollen development, neither reprogramming them into an androgenic pathway nor restoring fertility. Androgenic studies in L. esculentum showed that pollen from sterile mutants between prophase- metaphase I were the most responsive for DH induction (Shtereva et al., 1998). However, my observations from an Exacum sterile genotype failed to identify any androgenic response. Therefore, I conclude that using sterile genotypes for androgenic induction is less likely to be successful than using fertile genotypes

76 Table AII.l Cold treatment and media effects on pollen development and calli formation from Exacum Styer Group genotype 01-37-37. Temperature Duration of Explants for Pollen development Calli emerged from (°C) treatment temperature anthers (%) (days) treatment 5 0 Anther Oblong shape 0 Bud Oblong shape 0 3 Anther Oblong shape 0 Bud Oblong shape 0 7 Anther Oblong shape 0 Bud Oblong shape 0 n U 10 0 Anther Oblong shape 0 Bud Oblong shape 0 3 Anther Oblong shape 0 Bud Oblong shape 0 7 Anther Oblong shape 0 Bud Oblong shape 0

Table AII.2 The effect of various BA and 2,4-D combinations on calli formation of isolated anthers of Exacum Styer Group genotype 01-37-37. Auxin Cytokinin No. of anthers Calli formation 2,4-D (mgr1) BA (mgr1) culture

0.01 1 45 N 2 45 N 3 45 N

0.1 1 45 N 2 45 N 3 45 N

1 1 45 N 2 45 N 3 45 N

77 References 1. Albertsen MC & Phillips RL (1981) Developmental cytology of 13 genetic male-sterile loci in maize. Can J. Genet. Cytol. 23: 195-208

2. Anderson WC (1984) A revised tissue culture medium for shoot multiplication of rhododendron. J. Amer. Soc. Hort. Sci. 109: 343-347

3. Arnison PG, Donaldson A, Jackson A, Ho LCC & Keller WA (1990) The influence of various physical parameters on anthers culture of broccoli (Brassica oleracea var. italica). Plant Cell Tiss. Org. Cult. 20: 147-155

4. Ashok Kumar HG, Murthy HN & Paek KY (2003) Embryogenesis and plant regeneration from anther cultures of Cucumis sativus. L. Sci. Hort. 98: 213-222

5. Barnabas B, Fransz PF, Schel JHN (1987) Ultrastructural studies on pollen embryogenesis in maize (Zea mays L.). Plant Cell Rep. 6: 212-215

6. Barnabas B (2003) Anther culture of maize (Zea mays L.) In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

7. Bennett MD (1971) The duration of meiosis. Proc. Roy. Soc. Lond. B. 178: 277-299

8. Bennett MD (1973) The duration of meiosis. In: Balls M. & Billet FS (eds.) The cell cycle in development and differentiation, ppl 11-131. Cambridge University Press, UK

9. Bennett MD, Lewis KR, & Harberd DJ (1977) The time and duration of meiosis (and discussion). Phil. Trans. R. Soc. Lond. B. 277:201-226

78 10. Beyl CA (2000) Getting started with tissue culture-media preparation, sterile technique, and laboratory equipment. In: Trigiano RN, Gray DJ (eds.) Plant tissue culture concepts and laboratory exercises. Second edition (pp.21-44). CRC Press, U.S.A.

11. Binarova P, Hause G, Cenklova V, Cordewener JHG & Campagne MMV (1997) A short severe heat shock is required to induce embryogenesis in late bicellular pollen of Brassica napus L. Sex. Plant Rep. 10 (4) 200-208

12. Carapetian J (1994) Effects of safflower sterility genes on the inflorescence and pollen grains. Aust. J. Bot. 42: 325-327

13. Carlson PS (1970) Induction and isolation of auxotrophic mutants in somatic cell cultures of Nicotiana tabacum. Science 168: 487-489

14. Chaudhury AM (1993) Nuclear genes controlling male fertility. Plant Cell 5: 1277-1283

15. Chu CC, Wang CC, Sun CS, Hsu C, Yin KC, Chu CY & Pi FY (1975) Establishment of an efficient medium for anther culture of rice through comparative experiments on the nitrogen sources. Sic. Sin. 18: 659-668

16. Cramer LH (1981) Gentianaceae, pages 55-66. In: MD Dassanayake, ed., A revised handbook to the flora of Ceylon. Amerind Publishing Co., New Delhi

17. Ducan EJ & Heberle-Bors E (1976) Effect of temperature shock on nuclear phenomena in microspores of Nicotiana tabacum and consequently on plantlet production. Protoplasma 90: 173-177

18. Dunwell JM, Cornish M & De Courcel AGL (1983) Induction and growth of 'microspore dervied' embryo of Brassica napus ssp. oleifera. J. Exp. Bot. 34:

79 1768-1778

19. Ferrie A (2003) Microspore culture of Brassica species. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

20. Gamborg OL, Miller RA & Ojima K (1968) Nutrient requirements of suspension cultures of soybean root cells. Exp. Cell. Res. 50:151-158

21. Germana MA & Chiancone B (2003) Improvment of Citrus Clementina Hort. ex Tan. microspore-derived embryoid induction and regeneration. Plant Cell Rep. 22:181-187

22. Gresshoff PM & Doy CH (1972) Haploid Arabidopsis thaliana callus and plants from anther culture. Austr. J. Biol. Sci. 25: 259-264

23. Gresshoff PM & Doy CH (1974) Derivation of a haploid cell line from Vitis vinifera and the importance of the stage of meiotic development of anthers for haploid culture of this and other genera. Z. Pflanzenphysiol 73: 132-141

24. Goldberg RB, Beals TP & Sanders PM (1993) Anther development: Basic principles and potential applications. Plant Cell 5:1217-1299

25. Guha S & Maheshwari SC (1964) In vitro production of embryos from anthers of Datura. Nature 204: 497

26. Guha S & Maheshwari SC (1966) Cell division and differentiation of embryo in the pollen grains of Datura in vitro. Nature 212: 97-98

27. GuoY-D, Sewon P & Pulli S (1999) Improved embryogenesis from anther culture and plant regeneration in timothy. Plant Cell Tiss. Org. Cult. 57: 85-93

80 28. Hamaoka Y, Fujita Y & Iwai Sumio (1991) Effects of temperature on the mode of pollen development in anther culture of Brassica campestris. Physiologia plantarum 82:67-72

29. Hancock JF (2004) Plant evolution and the origin of crop species. CABI Publishing, Wallingford

30. Harlan JR & deWet JMJ (1972) A simplified classification of cultivated sorghum. Crop Science 12:172-176

31. Hofer M (2003) In vitro androgenesis in apple. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp.287-292). Kluwer Academic Publisher, Netherlands

32. Hofer M (2004) In vitro androgenesis in apple-improvement of the induction phase. Plant Cell Rep. 22:365-370

33. Horner HT & Palmer RG (1995) Mechanism of genie male sterility. Crop Sci. 35:1527-1535

34. Hu J & Rutger JN (1992) Pollen characteristics and genetics of induced and spontaneous genetic male sterile mutants in rice. Plant Breed. 19: 97-107

35. Immonen S & Tenhola-Roininen T (2003) Protocol for rye anther culture. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp.103-108). Kluwer Academic Publisher, Netherlands

36. Indrianto A, Barinova I, Touraev A, Heberle-Bors E (2001) Tracking individual wheat microspores in vitro: identification of embryogenic microspores and body

81 axis formation in the embryo. Planta 212:163-174

37. Jacobsen E & Sopory SK (1978) The influence and possible recombination of genotype on the production of microspore embryoids in anther cultures of Solarium tuberosum and hybrids. Theor. Appl. Genet. 52:119-123

38. Jacquard C, Wojnarowiez G & Clement C (2003) Anther culture in barley. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

39. Kasha KJ & Maluzynski M (2003) Production of double haploids in crop plants. An introduction. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

40. Kaul MLH (1988) Male sterility in higher plants. Monographs on theoretical and applied genetics 10. Springer-Verlag, New York

41. Keller WA (1984) Anther culture of Brassica. In: Vasil IK (Ed) Cell Culture and Somatic Cell Genetics of Plants. Vol 1 (pp302-310) Academic Press, Orlando

42. Khush GS & Virmani SS (1996) Haploid in plant breeding. In: Jain SM, Sopory SK, & Veilleux RE (eds) In vitro haploid production in higher plants, vol. 1. (pp. 11-33). Kluwer Academic Publisher, Boston

43. Klackenberg J (1985) The genus Exacum (Gentianaceae). Opera botanica, p. 1-144. AiO Print Ltd. Copenhagen

44. Lakshminarayana K & Maheswari Devi H (1985) Embryological studies in Gentianaceae. Proc. Indian Acad. Sci. (Plant Sci). 95(3): 213-219

82 45. Laurain DJ, Trimouillaux-Guiller & Chenieux JC (1993) Embryogenesis from microspores of Ginkgo biloba L. a medicinal woody species. Plant Cell Physiol. 31:1249-1251

46. Lee H-S, Karunanandaa B, McCubbin A, Gilroy S & Kao T-h (1996) PRK1, a receptor-like kinase of petunia inflate, is essential for postmeiotic development of pollen. Plant J. 9(5): 613-624

47. Li H, Javed AQ & Kutty KK (1988) The influence of different temperature treatments on anther culture response of spring wheat (Triticum aestivum). Plant Sci. 57: 55-61

48. Li WZ & Hu H (1995) Changes of soluble-protein in barley anther during the pretreatment period. J. Laiyang Agric. College 12: 6-10

49. Li WZ, Song ZH, Jing JK & Hu H (1995a) Effects of mannitol pretreatment on androgenesis of barley (Hordeum vulgare L.). Acta Bot. Sinica 37: 552-557

50. Li WZ, Song ZH, Jing JK & Hu H (1995b) Effects of different pretreatment on three endogenous hormones contents and peroxidase activity in barley anther. Acta. Bot. Sinica 37:731-737

51. Lindquist S (1986) The heat shock response. Ann. Rev. Biochem. 55:1151-1191

52. MacCormick S (1993) Male gametophyte development. Plant Cell 5:1265-1275

53. Maheshwari Devi H (1962) Embryological studies in Gentianaceae: Gentianoideae and Menyanthoideae. Proc. Indian Acad. Sci B, 56:195-216

54. Maheshwari SC, Tyagi AK, Malhotra K & Sopory SK (1980) Induction of haploid

83 from pollen grains in angiosperms, the current status. Theor. Appl. Genet. 58:193-206

55. Maheshwari SC, Rashid A & Tyagi AK (1982) Haploids from pollen grains-Retrospect and Prospect. Am. J. Bot. 69: 865-879

56. Maraschin SF, Vennik M, Lamers GEM, Spaink HP, Wang M (2005a) Time-lapse tracking of barley androgenesis reveals position-determined cell death within pro-embryos. Planta. 220:531-540

57. Maraschin SF, de Priester W, Spaink HP, Wang M (2005b) Androgenic switch: an example of plant embryogenesis from the male gametophyte perspective. Journal of Experimental Botany. 56:1711-1726

58. Mayr E (1942) Systematics and the origin of species. Columbia University Press, U.S.A.

59. Murashige T & Skoog F (1962) A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant. 15: 473-479

60. Nakata K & Tanaka M (1968) Differentiation of embryoids from developing germ cell in anther culture of tobacco. Japanese J. Genet. 43: 67-71

61. Nitsch JP & Nitsch C (1969) Haploid plants from pollen grains. Science 163: 85-87

62. Nitsch JP & Norreel B (1973) Effect d'un choc thermique sur le pouvoir embryogene du pollen de Datura culture dans l'anthere ou isole de 1' anthere. C.R. Acad Sci Paris 276:303-306

63. Nitsch JP (1969) Experimental androgenesis in Nicotiana. Phytomorphology 19: 389-404

84 64. Pauk J, Mihaly R & Puolimatka M (2003) Protocol for wheat (Triticum aestivum L.) anther culture. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

65. Peng M, Ziauddin A & Wolyn DJ (1997) Development of asparagus microspores in vivo and in vitro is influenced by gametogenic stage and cold treatment. In Vitro Cell Dev. Biol. PLANT 33:263-268

66. Pulli S & Guo Y-D (2003) Microspore culture of rye. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 151-154). Kluwer Academic Publisher, Netherlands

67. Raghavan V (1976) Role of the generative cell in androgenesis in henbane. Science 191:388-389

68. Raghavan V (1978) Origin and development of pollen embryoids and pollen calluses in cultured anther segments of Hyoscyamus niger (henbane) Am. J. Bot. 65:984-1002

69. Raina SK (1989) Tissue culture in rice-improvement: Status and potential. Adv. Agron. 42: 339-397

70. Rao KS and Chinnappa CC (1983) Studies in Gentianaceae. Microsporangium and pollen. Can. J. Bot. 61:324-336

71. Reynolds TL (1997) Pollen embryogenesis. Plant Molecular Biology 33:1-10

72. Riseman A (1990) Examination of the morphology and reproductive biology of interspecific hybrids of Exacum. MSc. thesis. The Pennsylvania State University.

85 The Pennsylvania State University. University Park, Pennsylvania. U.S.A.

73. Riseman A & Craig R (1995) Interspecific Exacum hybrids- novel germplasm for the production of a new floricultural crop XVIIIth EUCARPIA Symposium Section Ornamentals, Tel Aviv, Israel. 5-9 Mar. 1995. Acta Hort. 420: 132-134

74. Riseman A & Chennareddy S (2004) Genetypic variation in the micropropagation of Sri Lankan Exacum hybrids. J. Amer. Soc. Hort. Sci. 129: 698-703

75. Riseman A, Sumanasinghe VA, Justice D, & Craig R (2005) New name 'Styer Group' proposed for interspecific hybrids of Exacum species native to Sri Lanka. Hortscience 40:1580-1583

76. Riseman A, Sumanasinghe VA & Craig R (2006) Cytology, crossability, and pollen fertility of Sri Lanka Exacum (Gentianaceae) and their hybrids. Int. J. Plant Sci. 167(2): 191-199

77. Roberts-Oehlschlager SL & Dunwell JM (1990) Barley anther culture: Pretreatment on mannitol stimulates production of microspore-derived embryos. Plant Cell Tiss. Org. Cult. 20:235-240

78. Saji KV & Sujatha M (1998) Embryogenesis and plant regeneration in anther culture of sunflower (Helianthus annuus L.). Euphytica 103: 1-7

79. Sangwan RS & Sangwan-Norreel BS (1987) Biochemical cytology of pollen embryogenesis. Intl. Rev Cytol. 107:221-272

80. Shivanna KR (2003) Pollen biology and biotechnology. Science Publishers,U.S.A.

81. Shtereva LA, Zagorska NA, Dimitrov BD, Kruleva MM & Oanh HK (1998) Induced androgenesis in tomato (Lycopersicon esculentum Mill). II. Factors

86 affecting induction of androgenesis. Plant Cell Rep. 18: 312-317

82. Sopory SK & Munshi M (1996) Anther culture. In: Jain SM, Sopory SK, & Veilleux RE (eds) In vitro haploid production in higher plants, vol. 1. (pp. 145-176). Kluwer Academic Publisher, Boston

83. Stauffer C, Benito Moreno RM & Heberie-Bors E (1991) Seed set after pollination with in-vitro-matured, isolated pollen of Triticum aestivum. Theoretical and Applied Genetics 81:576-580

84. Stebbins GL (1971) Chromosomal evolution in higher plants. Edward Arnold Publishers, London

85. Sumanasinghe VAD (1986) Electrophoretic, cytogenetic, crossability, and morphological studies of Exacum (Gentianaceae). Ph. D. Thesis. The Pennsylvania State University. University Park, Pennsylvania. U.S.A.

86. Sunderland N & Wicks F (1973) Embryoid formation in pollen grains of Nicotiana tabacum. J. Exp. Bot. 22:213-226

87. Sunderland N & Dunwell JM (1974) Pathways in pollen embryogenesis. In: International congress of plant tissue and cell culture 3rd (p.141-167). Academic Press, New York

88. Tai GCC & Xiong XY (2003) Haploid production of potatoes by anther culture. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

89. Taylor PE, Glover JA, Lavithis M, Craig S, Jnox RB, Dennis ES, Chaudhury AM (1998) Genetic control of male fertility in Arabidopsis thaliana: structural analyses

87 of postmeiotic developmental mutants. Planta 205:492-505

90. Telmer CA, Newcomb W & Simmonds DH (1995) Cellular changes during heat shock induction and embryo development of cultured microspores of Brassica napus cv. Topas. Protoplasma 185:106-112

91. Touraev A, Indrianto A, Vicente O, Wratschko O, Heberie-Bors E (1996a) Efficient microspore embryogenesis in wheat (Triticum aestivum L.) induced by starvation at high temperature. Sexual Plant Repro. 9:209-215

92. Touraev A, Pfosser M, Vicente O, Heberle-Bors E. (1996b) Stress as the major signal controlling the developmental fate of tabacco microspores: towards a unified model of induction of microspore/pollen embryogenesis. Planta 200:144-152

93. Touraev A, Vicente O & Heberle-Bors E (1997) Initiation of microspore embryogenesis by stress. Trends Plant Sci. 2(8):297-302

94. Vasil IK and Nitsch C (1975) Experimental production of pollen haploids and their uses. Z. Pflanzenphysiol 76:191-212

95. Vasil IK (1980) Androgenic haploids. Int. Rev. Cytol. Supp. IIA: 195-223

96. Wang XZ & Hu H (1984) The effect of potato II medium for triticale anther culture. Plant Sci. Lett. 36: 237-239

97. Wenzel GF, Hoffmann F, Potrykus I and Thomas E (1977) Increased induction and chromosome doubling of androgenetic haploid rye. Theor. Appl. Genet. 51:81-86

98. Wojnarowiez G, Caredda S, Devaux P, Sangwan R & Clement C (2004) Barley anther culture: assessment of carbohydrate effects on embryo yield, green plant production and differential plastid development in relation with albinism. Journal of

88 Plant Physiology 161: 747-755

99. Wolyn DJ & Nichols B (2003) Asparagus microspore and anther culture. In: Maluszynski M, Kasha KJ, Forster BP & Szarejko I (eds.) Double haploid production in crop plants. A manual (pp. 103-108). Kluwer Academic Publisher, Netherlands

100. Yuan YM, Wohlhauser S, Moller M, Klackenberg J, Callmander MW & Kiipfer P (2005)Phylogeny and Biogeography of Exacum (Gentianaceae): A disjunction in the Indian Ocean Basin resulted from long distance dispersal and extensive radiation. Syst.Biol. 54(1): 21-34

101. Zaki MAM & Dickinson HG (1991) Microspore-derived embryos in Brassica: the significance of division symmetry in pollen mitosis I to embryogenic development. Sexual Plant Repro. 4:48-55

89