Citation
Somatic embryogenesis, long-term regeneration and epigenetic variation in tissue culture regenerants of diploperennial teosinte (Zea diploperennis iltis, Doebley & Guzman)

Material Information

Title:
Somatic embryogenesis, long-term regeneration and epigenetic variation in tissue culture regenerants of diploperennial teosinte (Zea diploperennis iltis, Doebley & Guzman)
Creator:
Pedrosa, Luis Filipe, 1946-
Publication Date:
Language:
English
Physical Description:
vii, 334 leaves : ill., photos ; 29 cm.

Subjects

Subjects / Keywords:
Callus ( jstor )
Corn ( jstor )
Embryos ( jstor )
Internodes ( jstor )
Regeneration ( jstor )
Somatic embryogenesis ( jstor )
Specimens ( jstor )
Tassels ( jstor )
Tillers ( jstor )
Tissue culture techniques ( jstor )
Botany thesis Ph. D
Dissertations, Academic -- Botany -- UF
Somatic embryogenesis ( lcsh )
Zea diploperennis ( lcsh )
Zea diploperennis -- Cytology ( lcsh )
Zea diploperennis -- Genetics ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 306-333).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Luis Filipe Pedrosa.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
030441843 ( ALEPH )
31280148 ( OCLC )

Downloads

This item has the following downloads:


Full Text










SOMATIC EMBRYOGENESIS, LONG-TERM REGENERATION
AND EPIGENETIC VARIATION IN TISSUE CULTURE REGENERANTS OF
DIPLOPERENNIAL TEOSINTE (ZEA DIPLOPERENNIS ILTIS, DOEBLEY & GUZMAN)














By

LUIS FILIPE PEDROSA
















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA 1993


























This dissertation is dedicated to my mother, Maria Luisa, who did not live to see its completion, and to my son, David, joy of my life.













ACKNOWLEDGMENTS




I would like to express my thanks to Dr. Indra K. Vasil, for providing valuable guidance and the logistics necessary to bring this piece of research to a good end.

My appreciation also goes to Dr. H.H. Iltis (University of Wisconsin), for promptly providing the diploperennial teosinte seeds needed to produce the populations used in this study.

It is also with gratitude that I acknowledge the members of my graduate committee, Dr. Thomas J. Sheehan, Dr. Norris H. Williams, Dr. Henry C. Aldrich and Dr. William Louis Stern, as well as Dr. Vimla Vasil and Dr. Walter S. Judd for all their helpful suggestions and continued support.

Special thanks are particularly due to Dr. Flona Redway and Mr. Mark Taylor, for keeping me going in times of distress, and to Mr. Bart Schutzman, for sharing many fascinating incursions into the worlds of plant science, statistics, computers and photography.

Last, but not the least, my gratitude and thanks go to my loving son, David, for having the special endurance it takes to have a doctoral student as a single parent.















iii
















TABLE OF CONTENTS




Page

ACKNOWLEDGEMENTS ................................................................................ III

A B ST RA C T ............................................................................................................................. vi

CHAPTERS

1 INTRODUCTION .......................................................... 1

2 LITERATURE REVIEW ......................................................... 6

3 SOMATIC EMBRYOGENESIS AND LONG-TERM REGENERATION IN
CALLUS CULTURES OF DIPLOPERENNIAL TEOSINTE ......................... 29

Introduction ................................................................. ........................... 29
Materials and Methods ..................................... ..... ............... 34
Plant Material ...................................... .............................................. 34
Induction of Somatic Embryogenesis ....................................... 36
Callus Maintenance and Multiplication ....................................... 37
Data Collection and Analysis ........................................ ........ 38
Plant Regeneration .................................................................... 39
R esults ...................................................................... ............................. 40
Induction of Somatic Embryogenesis ........................................ 40
Callus Maintenance ..................................................................... 71
Plant Regeneration .................................... ................................. 86
Discussion and Conclusions ......................................... .......... 88

4 PHENOTYPIC VARIATION IN TISSUE-CULTURE-DERIVED PLANTS
OF DIPLOPERENNIAL TEOSINTE ........................................................ 101

Introduction ...................................... 101
Materials and Methods ..................................... 101
Morphometric Analyses ..................................... 101
Cytological Analyses ..................................... 106
Results ............................................................................................. 108
Univariate Comparisons ..................................... 108
Multivariate Comparisons ..................................... 127
The Effect of Gibberellic Acid ......................................................... 203
Discussion and Conclusions ..................................... 236



iv








5 ELECTROPHORETIC ANALYSIS OF TISSUE-CULTURE-DERIVED
PLANTS OF DIPLOPERENNIAL TEOSINTE .................. ...................... 257

Introduction ...................................... 257
Materials and Methods ..................................... 258
Plant Material ..................................... 258
Preparation of Starch Gels .................................. 258
Protein Extraction and Gel Loading ..................................... 259
Electrophoretic Buffers and Staining Systems .................................. 261
Electrophoresis ................................................................................. 261
Enzyme Activity Staining ..................................... 264
Analysis of Banding Patterns ..................................... 265
Results ........................................ 265
Selection of Electrophoretic Buffers and Enzyme Systems
for Studying Isozymes in Diploperennial Teosinte ......................... 265
Isozyme Analysis of Tissue Culture Regenerants .......................... 268
Discussion and Conclusions ................................................................ 268

6 CLOSING REMARKS ............................................................................. 297


REFERENCES ........................................................................................................ . 306

BIOGRAPHICAL SKETCH ....................................... 334






























v








Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy


SOMATIC EMBRYOGENESIS, LONG-TERM REGENERATION
AND EPIGENETIC VARIATION IN TISSUE CULTURE REGENERANTS OF
DIPLOPERENNIAL TEOSINTE (ZEA DIPLOPERENNIS ILTIS, DOEBLEY & GUZMAN)

By

Luis Filipe Pedrosa

December 1993

Chairman: Indra K. Vasil
Major Department: Botany


High frequency somatic embryogenesis and long-term maintenance of embryogenic calli were obtained from scutellar tissue of diploperennial teosinte immature zygotic embryos. The developmental stage of the embryo and the inclusion of a strong auxin in the initiation medium were the only critical factors for success.

Embryogenic calli were comparable to those formerly described for the majority of the Gramineae. An unusual callus type, which was never associated with soft, friable, non-embryogenic callus, was isolated and used to establish all the regenerable cultures used in this study. Manipulation of the carbon sources in the medium, as well as addition of casamino acids and L-proline, enhanced the embryogenic response and growth of the cultures that, under optimal conditions, maintained regeneration potential for over four years.

Morphologically normal plants were routinely recovered from the callus cultures for over two years. Regeneration of aberrant phenotypes, however, gradually became more frequent in older cultures. At the end of three years all plants from tissue culture were distinctly dwarf, multitillering and male sterile, displaying incomplete development and/or feminization of the tassel. All had the normal diploid chromosome complement.





vi









The fate of the induced morphological variation was studied for the next two years, using 100 plants each from the Ro and the R1. One hundred plants from seed were used as a control. Combined univariate and multivariate statistical analysis of 26 morphometric traits, covering both vegetative and floral characters from the tissue culture regenerants, their sexual progeny and control plants, revealed that all anomalous traits in the Ro were transient in nature and faded with time, although some persisted for a period of over 18 months. Gibberellin applications completely replaced the effect of time, and led to complete phenotypic recovery to normal of the Ro plants. Complete recovery also resulted from a passage through a single cycle of sexual reproduction, as there were no anomalies in the R1 plants.

Isozyme analysis did not show a link between the morphological variation and the corresponding biochemical variation that, of necessity, must precede or accompany it. Regardless of their phenotypic differences, all tissue culture regenerants showed the same isozyme banding patterns, which were the same as those from the controls.



























vii












CHAPTER 1

INTRODUCTION



The occurrence of phenotypic variation in plants regenerated from tissue cultures is a well documented, yet little understood phenomenon (Bayliss 1980; Larkin and Scowcroft 1981; D'Amato 1985; Karp and Bright 1985; L6rz and Brown 1986; L6rz et al. 1988). Phenotypic variation is frequently associated with regeneration systems involving an unorganized callus phase and it may affect quantitative as well as qualitative traits. Deviations from the expected phenotypic and genetic uniformity have been reported at the morphological, karyological, physiological, biochemical and molecular levels (Meins 1983; Orton 1984; D'Amato 1985; Scowcroft 1985; Ahloowalia 1986; Gould 1986; Semal 1986; Lrz et al. 1988).

The theoretically unexpected but apparently widespread generation of variability during tissue culture, recently termed "somaclonal variation" (Larkin and Scowcroft 1981), has been proposed as a potential source of unique and interesting new traits for use in crop improvement (Larkin and Scowcroft 1981; Larkin 1985). Since the introduction of this concept in 1981, much controversy has arisen concerning the real prospects of somaclonal variation as a contribution to ongoing or new plant breeding programs (Karp and Bright 1985; Ahloowalia 1986; L6rz and Brown 1986; Semal 1986; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K. Vasil 1987, 1988). Part of this controversy is based upon the fact that, although variation may be present in the regenerants, it is frequently transient in nature or is lost during sexual reproduction. Moreover, limited field evaluations with diverse genotypes of putative somaclones have shown that most somaclonal variation is either useless or has a limited applicability in direct varietal upgrading (Ahloowalia and Sherington 1985; Ryan et al. 1987; Baillie et al.



1






2


1992; Qureshi et al. 1992). Whenever heritable variation is present, the type and frequency of variants suggest that somaclonal variation is akin to non-directed random mutagenesis, which generates a large amount of unwanted variation often expressed in the form of traits that either are not novel or are agronomically unfavorable (Semal 1986; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K Vasil 1987, 1988). It is noteworthy that, despite the rhetoric and exaggeration surrounding the issue for almost a decade, no commercial varieties of any major crop plant have yet been produced or are grown on a large scale as a result of using tissue-culture-generated variation.

As time gradually wore most of the original glamor out of the issue of usefulness of the newly expressed traits, questions such as their persistence over time and across generations, and the quest for the genetic or physiological mechanisms underlying their origin, slowly took its place in the minds of the scientific community. There is no doubt that some of the culture-induced variation has a genetic basis and is heritable. It has often been argued, however, that much of the observed variability is only transient and epigenetic, physiological or developmental in nature and therefore nonheritable and useless in breeding (Karp and Bright 1985; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K. Vasil 1987, 1988). When this is the case the plants showing variation can no longer be thought of as potential assets to breeding; they can, however, still be usefully incorporated in programs where genetic fidelity, rather than change, is the goal, on the condition that the detected variation can be proven to be no more than a transient culture-induced "carry-over" effect.

Tissue-culture-generated variation has been correlated with a variety of factors. These include, among others, the original type and age of the explant, genotype of the donor plant, duration of culture and regeneration pathway. In Gramineae, where two possible regeneration pathways (embryogenic and organogenic) have been recognized and described in detail, most of the reported occurrences of variation in tissue-culturederived populations have been related to regeneration from organogenic callus cultures






3


(Morrish et al. 1987). Accumulating evidence seems to indicate that more stringent selection in favor of normal cells, believed to be associated with the embryogenic pathway of regeneration, tends to prevent the production of somaclonal variants (I.K. Vasil and V. Vasil 1986; Kobayashi 1987; Morrish et al. 1987; I.K. Vasil 1987, 1988; Dolezel and Binarova 1989; Gmitter et al. 1991). The selective advantage of cytologically or genetically normal cells (diplontic selection), expressed at the very onset of the embryogenic response in callus cultures, results in the production of embryos that consist wholly of normal euploid cells (e.g., Cavallini et al. 1987; Gmitter et al. 1991). Selection pressure can, however, be critical even at later stages, as evidenced by the fact that developing somatic embryos with cytological or genetic aberrations, or chimerism, often fail to differentiate into adult plants (Cavallini et a/. 1987; Cavallini and Natali 1989). The apparent intolerance, during embryogenesis, to alterations of the normal genome thus effectively prevents (or greatly limits) the transmission of anomalies to populations produced by this regeneration pathway.

Although karyological and phenotypic stability can often be correlated with the embryogenic pathway, varying frequencies of abnormal regenerants from embryogenic tissue cultures have also been reported in the literature. These may have resulted from cultures where both the embryogenic and organogenic pathways coexist, as has been described in wheat (Karp and Maddock 1984), in which case the variation reported may result from plants being regenerated from the mixed callus via the organogenic pathway (Maddock 1985). Variants may also be transient in nature (Larkin and Scowcroft 1983a,b; Irvine 1984), the result of epigenetic rather than genetic changes in the regenerated plants.

There is still much to be learned about the nature and origin of the variation found in tissue-culture-derived plants. The genetic (or epigenetic) nature of the variation has not always been properly studied or documented in the past. In many cases the regenerated plants are morphologically aberrant to the point of being unable






4


to complete a life cycle, or have their sexual expression at maturity modified or supressed, which prevents the progeny testing that constitutes the ultimate proof of genetic (heritable) variation (Meins 1983; Karp and Bright 1985). In such cases, alternative ways of distinguishing between genetic and epigenetic changes have been proposed (Meins 1983) and should be utilized whenever possible. Evidence of epigenetic rather than heritable variation in cases involving the embryogenic pathway would add additional support to the idea that somatic embryogenesis produces a lower overall level of genetic variation and that it should be the regeneration method of choice whenever maintenance of genetic fidelity, rather than variation, is the primary goal of the investigation.

Further work is needed to determine the level of stability attained through somatic embryogenesis in tissue cultures. The methods to employ in such studies have been outlined in the literature (Morrish et al. 1987). They include detailed analysis of the somatic variation present in the regenerants, as compared and contrasted to appropriate controls, as well as its heritability. In addition to the traditional chromosome cytology, such studies must focus on both quantitative and qualitative genetic traits and, whenever possible, biochemical and/or molecular analyses should be used to further assess possible alterations of the genome that may not be expressed as variant morphological traits. Regeneration must be strictly monitored, particularly in the species for which both possible regeneration pathways have been reported. Because the transient nature of some epigenetic effects may be obscured by their long-term expression, the use of a perennial (for which time is not a limiting factor) rather than an annual species is recommended and the barriers of sexual sterility or incompatibility must not be present, to permit a thorough analysis of the sexual progeny.

This dissertation reports the results of one such study in a perennial relative of maize, diploperennial teosinte (Zea diploperennis Iltis, Doebley & GuzmAn). The objectives of this research were (1) to develop a long-term efficient regeneration system






5


from cultures of diploperennial teosinte; (2) to characterize the morphology of the regeneration pathway; (3) to evaluate the genetic fidelity of the tissue-culture-derived plants through morphometric and cytological characterization of the regenerants, their progeny and control plants; (4) to evaluate biochemically the genetic fidelity of the tissue-culture-derived plants through electrophoretic characterization of the regenerants; and (5) to interpret the nature of variation in regenerants and/or their progeny, should any be detected.












CHAPTER 2

LITERATURE REVIEW



Introduction


Early in the sixties Morel first performed his classic experiments with the proliferation in vitro of orchid shoot apical meristems (Morel 1960; 1963; 1964a,b; 1965a,b). His discoveries soon led to the commercial mass production of these rare and beautiful plants and, by promoting the development of what is now known as the aseptic method of micropropagation, initiated a major revolution in the horticultural world. In the years that followed, extensive research expanded the realm of micropropagation to a wide range of taxa, now covering virtually the complete spectrum of vascular plants.

The production of tissue-culture-derived clones has been achieved in a very large number of species, including ornamentals, vegetable crops, fruit crops, trees and agronomic crops (Conger 1981; I.K. Vasil 1986). Based on the principle that regeneration from cultured somatic cells faithfully reproduces the original donor plant in potentially unlimited numbers, the system permits the large scale reproduction of elite germplasm, providing the grower (and the consumer) with superior products at a fraction of the cost and time needed by conventional means.

It was soon discovered, however, that not all cultures were suitable for cloning. When plants were regenerated from adventitious buds phenotypic and genetic changes were not uncommon, especially when a stage of unorganized cell proliferation (callus) was involved. The consistent uniformity and trueness-to-type needed in clonal propagation could still be achieved, as long as the cultures were based on induced



6






7


axillary branching from preexisting shoot tips and lateral bud explants, an approach that completely bypasses the presence of the undifferentiated callus phase. Observations made on such populations of meristem-propagated plants have shown that the frequency of variants obtained by this method is not higher than that observed when using traditional methods of clonal propagation (Conger 1981; Beauchesne 1982). Occasional sports are attributed to chance mutations, that occur at very low frequencies. This phenotypic uniformity from shoot tip cultures has been attributed to the higher level of genetic stability normally associated with shoot meristems, which act as a safe "storage" place for the normal genetic information of the plant (D'Amato 1985; I.K. Vasil 1987).

Long recognized as an unavoidable nuisance, the characteristic variation found in callus cultures and their regenerates was brought to the limelight in 1981 when Larkin and Scowcroft proposed that such variation, which they termed "somaclonal", was an inherent result of the plant cell culture itself and that it showed great promise as a significant source of novel traits potentially useful for plant improvement (Larkin and Scowcroft 1981). Under the newly coined name of "somaclonal variation" the concept was widely adopted throughout the eighties, when a large number of publications covered in varying detail a wide range of aberrations associated with tissue cultures, many of them known but never publicized before. Among others, they included numerical and gross structural chromosome changes in the cultures and their regenerates (D'Amato 1985), cryptic alterations in the nuclear, chloroplast and mitochondrial genomes (Hanson 1984; L6rz and Brown 1986), breakdown of chimeric structures (Cassells 1985; Preil 1986) and changes due to the elimination of infectious agents (Cassells 1985; Larkin and Scowcroft 1981).

The popularity of the idea is reflected in the numerous reviews that over the last several years tried to assess the nature of somaclonal variation and its potential contribution to plant breeding (Larkin and Scowcroft 1983b; Orton 1984; Mitra 1985;






8


Evans and Sharp 1986, 1988; Dewald and Moore 1987; Larkin 1987; Sala and Biasini 1987; Larkin et al. 1989). Although specific examples of applications have been reported (e.g., Evans 1989) there is, however, a growing opinion that somaclonal variation has not fulfilled its initial promise as a potential source of economically valuable germplasm. A critical analysis of the success so far achieved, now that one decade has passed, reveals that despite massive efforts by scores of dedicated researchers worldwide, no significant commercial varieties of any of the major crop species have yet been developed as a result of incorporating tissue-culture-generated genetic variation into a breeding program (I.K. Vasil 1990).

Much of the variation that has been found in cultures or their regenerates is either not novel or has doubtful usefulness in breeding (I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K. Vasil 1987, 1988). It may even be deleterious in nature. In addition, long-term epigenetic changes, which are not unusual in tissue culture regenerants, are often mistaken for true genetic variation (Karp and Bright 1985; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K. Vasil 1987, 1988; Karp 1989). Epigenetic changes are, in fact, the result of directed physiological alterations on the expression of the genome. They have been shown to be potentially reversible and, by definition, nonheritable (Meins 1983), and are therefore useless in breeding. The distinction between genetic and epigenetic variation may be even further complicated by the fact that certain changes in the genetic material, such as DNA methylation (Lrz and Brown 1986) and amplification (Cullis 1986), may be heritable under certain conditions, but may also become reversible (Phillips et al. 1990).

In view of the failure to produce the promised boon of new traits for breeding, somaclonal variation is again regarded today the way it had been from the very beginning-a seemingly inevitable nuisance to most current fields of applied research that make use of tissue cultures. Its occurrence is clearly detrimental in clonal micropropagation (Beauchesne 1982), including the use of somatic embryogenesis and






9


synthetic seed technologies as potential means of automated mass propagation (Redenbaugh et al. 1987; Ammirato 1989). It is also a particular concern in the newly emerging discipline of genetic transformation through gene transfer into cultured cells, a technique that seeks to introduce selected genes into elite lines without the cointroduction of unwanted culture-induced variability (LSrz and Brown 1986; Goodman et al. 1987; Lazzeri and L6rz 1988; de Klerk 1990).

It would be highly desirable to be able to control the genetic instability associated with tissue cultures. No overall consensus exists, however, on general applicable guidelines for such control (Karp 1989). There is a lack of information on the underlying genetic and physiological mechanisms that are the basis of variation and the interaction of the factors affecting them. The resolution of this problem, currently under intense investigation, should be a greater contribution to plant science than the proposed direct exploitation of the phenotypic variation itself (Karp and Bright 1985). This chapter briefly summarizes the nature of the phenotypic characteristics normally associated with somaclonal variation in Gramineae and presents the current understanding on how such variation relates to a range of genotypic or epigenetic changes as they affect the cultured cells in vitro.


The Nature of Tissue-culture-induced Variation

Tissue-culture-induced variation is a general phenomenon of widespread occurrence (Bayliss 1980; Larkin and Scowcroft 1981; D'Amato 1985; Karp and Bright 1985; L6rz and Brown 1986; L6rz et al. 1988). It seemingly affects most plant species, regardless of taxonomic position or ploidy level, including our most important crops. It has been described in seed crops such as tobacco, tomato, and alfalfa, in vegetatively propagated crops such as potato, in cereals and other grasses and in ornamental plants, such as the garden geranium and petunia (Larkin and Scowcroft 1981; Karp and Bright 1985; Larz and Brown 1986). As previously mentioned, tissue-






10


culture-induced variation is directly related to regeneration systems in vitro that involve the passage through an unorganized callus phase (Larkin and Scowcroft 1981; Karp and Bright 1985). It is normally absent in cultures derived from proliferation of shoot apical or axillary meristems (D'Amato 1985; Karp and Bright 1985).

Somaclonal variation can be observed at the morphological, karyological, physiological, biochemical and molecular levels (Meins 1983; Orton 1984; D'Amato 1985; Gould 1986; L6rz et al. 1988). It may affect quantitative as well as qualitative traits. The phenotypic changes so produced, sometimes at an unusually high frequency, may occur in individual plants or affect whole populations and often concern agronomically important traits. Examples of tissue-culture-induced variation in Gramineae are presented in Table 2.1. Variant traits are normally found in the heterozygous form, but may also occur as homozygous as well, an intriguing phenomenon quite characteristic of this kind of variation and for which at present there still is no plausible explanation (Karp and Bright 1985; Karp 1989).

Not all variation found in somaclonal variants is genetic and stable. In the vast literature dealing with the subject there are many examples of transient or temporary phenotypic variation. These may vary from what has been loosely termed a "carry-over" effect, like the typical shoot proliferation/root inhibition associated with the use of cytokinins in the regeneration system, to long-term effects that may be easily confused with stable genetic variation, if not for the fact that they are not always found in the following generation after selfing. This variation, termed epigenetic (Meins 1983), may be expressed at the morphological level, frequently as alterations in size or growth habit, as well as at the physiological level, for example, as habituation or vitrification. It has been correlated to the stress of growth under in vitro conditions and can often be duplicated by growing plants from seed under conditions that mimic the tissue culture environment. For example, when maize seedlings were grown from normal non-tissueculture-derived seeds maintained in closed containers with synthetic media, as would









Table 2.1 Examples of tissue-culture-induced variation in Gramineae



Species Type of References
(Common Name) Variation


Avena sativa karyological Cummings et al. 1976
(Oat) reduced plant height McCoy et al. 1982 heading date Dahleen 1989 seed protein
flag leaf area seed weight
seed number yield
bundle weight sterility

Bothriochloa spp. karyological Taliaferro et al. 1989
(Old World Bluestems) reduced plant height dwarfism
foliage colour suppression of flowering inflorescence length reduced fertility isozyme banding patterns

Cymbopogon winterianus herbage yield Mathur et al. 1988
(Citronella Grass) oil content

Festuca arundinacea isozyme banding patterns Dahleen and Eizenga 1990
(Tall Fescue)

Hordeum vulgare polyploidy Orton 1980
(Barley) karyological
plant height Ahloowalia 1986 stem thickness extra flag leaves supernumerary spikes chloroplast DNA Day and Ellis 1985 rDNA spacer fragments Breiman et al. 1987b hordein proteins hordein proteins Karp et al. 1987

Lolium multiflorum plant height Ahloowalia 1986
(Rye Grass) leaf length leaf width
spike shape
fertility






12



Table 2.1 (Continued)



Species Type of References
(Common Name) Variation


Oryza sativa reduced height Oono 1978
(Rice) reduced fertility heading date
soluble protein Sun et al. 1979 mRNA levels
plant height Schaeffer 1982 dwarfism Sun et al. 1983 tiller number
heading date
grain per pannicle
grain weight
chlorophyll mutations
reduced height Schaeffer et al. 1984 improved yield
improved protein
dwarfism Oono 1985 nuclear DNA Zheng et al. 1987

Pennisetum americanum karyological Swedlund and Vasil 1985
(Pearl Millet) reversion to fertility Smith et al. 1987

Saccharum officinarum isozyme banding patterns Heinz and Mee 1971
(Sugarcane) auricle length sheath colour
pubescence
erectness Liu et al. 1972 cane diameter Lat and Lantin 1976 stalk length
stalk weight
disease resistance Uu and Chen 1976 cane and sugar yield
disease resistance Heinz et al. 1977 disease resistance Larkin and Scowcroft 1983a disease resistance Krishnamurthi 1982

Secale cereale secalin proteins Bebeli et al. 1990
(Rye)

Triticum aestivum reduced height Ahloowalia 1982
(Bread Wheat) stem thickness Maddock et al. 1983 reduced fertility Sears et al. 1984 heading date Ahloowalia and Sherington 1985 chloroplast DNA Day and Ellis 1984






13



Table 2.1 (Continued)



Species Type of References
(Common Name) Variation


Triticum aestivum karyological Karp and Maddock 1984
(Bread Wheat) (cont.) plant height Larkin et al. 1984 tiller number
heading date
grain and glume colour
gliadin proteins
gliadin proteins Cooper et al. 1986 isozyme banding patterns Davies et al. 1986 cold tolerance Lazar et al. 1988 cold tolerance Galiba and Sutka 1989

x Triticosecale karyological Armstrong et al. 1983
(Triticale) karyological Lapitan et al. 1984
kernel protein Jordan and Larter 1985 prolamin banding pattern spike length
fertility
rDNA spacer sequences Brettell et al. 1986b

Zea mays leaf shape Green 1977
(Maize) ear position pollen sterility
karyological Edallo et al. 1981 resistance to Gengenbach et al. 1981 Helminthosporium
male sterility
mtDNA restriction pattern lysine and threonine content Hibberd and Green 1982 karyological McCoy and Phillips 1982 reduced height Beckert et al. 1983 reversion to male fertility Umbeck and Gengenbach 1983 resistance to Umbeck and Gengenbach 1983 Drechslera maydis
increased vigor Hubbard et al. 1984 reduced height
altered fertility Earle and Gracen 1985 altered maturity
altered plant and ear height altered leaf color
increased tillering
increased ear number
ADH1 gene mutation Brettel et al. 1986a reduced plant height G6bel et al. 1986






14



Table 2.1 (Continued)



Species Type of References
(Common Name) Variation


Zea mays altered leaf morphology G6bel et al. 1986
(Maize) (cont.) feminization of male flowers suppression of anthers reduced ears poor germination (R1) early seedling death (R1) silk emergence Zehr et al. 1987 reduced plant height kernel characteristics pollen shed






15


occur during tissue culture, the resulting plants showed signs of variation identical to those commonly described in somaclonal variants, including reduced fertility, dwarfing, feminizaton of male flowers, etc. (L6rz et al. 1988). The underlying reasons for this unnatural behavior of plants derived from in vitro cultures are not currently understood but, because the transient nature of some epigenetic effects may be obscured by their long-term expression, progeny tests using the appropriate crosses for the variant traits must be examined to accurately determine the extent of epigenetic variation (Meins 1983; Evans and Sharp 1988).

Unique to somaclonal variation is the peculiar combination of homozygous changes at high frequency and the presence of whole population shifts for certain traits. For example, when comparing the types of mutants present in the progeny of tomato somaclones with those occurring in the progeny from chemically mutagenized plants (Gavazzi et al. 1987), the results indicated that in vitro culture had resulted in a higher frequency of some mutations and a different spectrum of variation, some mutants being found exclusively in the somaclonal population.


Origin of Tissue-culture-induced Variation

There is no complete understanding of which factors induce or otherwise affect the nature and frequency of instability in plants regenerated from tissue culture. Among the parameters to be examined (de Klerk 1990; Orton 1983), changes in the chromosome complement have the advantage of providing a visible measure of genetic variation at the cellular level that can be correlated with the physical and chemical cultural environment and the corresponding morphological variation at the plant level. For this reason there are more extensive reports on karyological variation than on any other form of tissue-culture-induced variation.






16


Karyological Variation

Numerical and structural chromosome variation in cell and tissue cultures and their regenerates has been the subject of several excellent reviews (Bayliss 1980; Constantin 1981; Wersuhn and Dathe 1983; D'Amato 1985; Karp and Bright 1985; Lee and Phillips 1988). Cytogenetic instability is normally shown by the occurrence of chromosome fragments and laggards at anaphase, multipolar spindles and the presence of micronuclei or multinucleate cells. Numerical and structural changes are further revealed by the study of meiotic chromosome behavior in regenerated plants. The full range of karyotypic variation has been observed, including ploidy changes, aneuploidy, deletions, duplications, inversions and translocations (e.g., McCoy et al. 1982; Johnson et al. 1987; Lee and Phillips 1987; Armstrong and Phillips 1988; Benzion and Phillips 1988).


Origin of Chromosome Variability

The nature and extent of chromosome variability obviously depends on the interaction of a number of exogenous and endogenous factors. In simple terms, however, it can be stated that variation in tissue-culture-derived plants either (1) was already present in the donor plant tissues or (2) is induced by the particular conditions of the culture phase (Karp 1989).


Preexisting variability. When somatic cells cease to divide, in the majority of angiosperms, at least a part of the differentiating tissues goes through one or several additional rounds of endoreduplication of the nuclear DNA without an intervening mitosis. This leads to a geometric increase in the DNA content of the nucleus (4C,8C,16C. . .). Mature tissues, thus, generally exhibit varying levels of endopolyploidy in a varying proportion of their cells, a condition termed polysomaty (cf., D'Amato 1975, 1977, 1985). When callus is initiated from polysomatic tissues, at least some of the endopolyploid cells may be induced to divide and proliferate, and thus produce variant






17


genotypes. The degree of endoreduplication is not generally uniform. It is often dependent on the age and type of tissues and organs (cf., D'Amato 1985) and even on the mode of plant cultivation (Pijnacker et al. 1989). Therefore, choosing explants with a low degree of polysomaty is one means of reducing variability. Explants comprising meristematic or very young tissues are especially recommended for establishing cultures (Lee and Phillips 1988; Karp 1989). Differences are also found among and within species, demonstrating the genetic control of polysomaty; some species are nonpolysomatic altogether (D'Amato 1985).

In some plants the apical meristems, and consequently the mature tissues, comprise a mosaic of cells with varying proportions of different aneuploid chromosome numbers. This condition, termed aneusomaty, is transmitted to and generally enhanced in callus cultures derived from such tissues (e.g., Heinz et al. 1969). Diplontic selection may act during regeneration to give rise to normal or near normal plants, but aneusomatic plants are more commonly produced from these callus cultures (e.g., Heinz and Mee 1971). When plants are regenerated from aneusomatic tissues and their origin is multicellular (as suggested by the presence of chromosome number mosaics in mixoploid regenerants), some of the regenerants may be chimeric and bear sectors (sectorial and mericlinal chimeras) or entire histogenetic layers (periclinal chimeras) of different chromosomal constitution. Sectorial ploidy chimeras have been detected, for example, among regenerants in maize (McCoy and Phillips 1982; Benzion and Phillips 1988). Most will probably remain unnoticed, however, since karyological analyses are mostly done on root tip cells or microspore mother cells, which are derived from single histogenic layers.

As mentioned earlier, the resolution of chimeric structures may also be a source of tissue-culture-induced variation, particularly important in vegetatively propagated plants (Skeene and Barlass 1983; Preil 1986). If the component histogenic layers of a chimeral cultivar are karyotypically distinct, this can contribute to chromosome






18


variability in the corresponding callus cultures and their regenerates (e.g., Skirvin and Janick 1976). The common production of chromosomal mosaics in the regenerants is also evidence for the multicellular origin of the in vitro regenerated shoot apices in these cases (cf., D'Amato 1985).

Chromosome structural changes and spontaneous gene mutations may also occur in mature tissues. The former are generally detected by aberrations observed in the mitotic cycle. The latter, however, may escape detection, since most spontaneous mutations are recessive and rarely have an immediate effect on the phenotype of the heterozygous diploid (cf., D'Amato 1985). Both structural changes and gene mutations may add an extra component to in vitro produced variation.

The acceptance that at least part of the observed somaclonal variation originates from preexisting variation (cf., Morrish et al. 1990) has strong implications on experimental design. To account for the possibility of this origin, it is critical that all experiments involving the analysis of somaclonal variants have appropriate controls. Furthermore, careful record keeping is essential as the origin and genealogy of each regenerant must be clearly known (Karp 1989).


Tissue-culture-induced variability. In addition to the variability possibly introduced with the explant, variant karyotypes may also arise during callus induction and subsequent growth. The induction of variation during callus initiation and subculture is particularly well studied in non-polysomatic plants where cultures are initiated from uniformly diploid cell populations. In polysomatic species culture-induced variation is revealed by variability within calli derived from single cells or from protoplasts (e.g., Karp et al. 1982).

The most frequently reported change in culture is polyploidization; aneuploidy is also commonly found. Polyploidy may originate through endoreduplication followed by mitosis, by restitution nucleus formation, i.e., chromosome doubling due to spindle






19


failure or by nuclear fusion in multinucleate cells. Aneuploidy may arise through unequal chromosome segregation at mitosis or by amitotic nuclear fragmentation followed by mitosis (cf., D'Amato 1985).


Genotype and Ploidy Level. There is reasonable evidence to believe that there is a genotypic component to instability in tissue cultures. This has been demonstrated in a variety of crop species, where some cultivars consistently give rise to a wider range of somaclonal variation than others (e.g., Galiba et al. 1985; Linacero and Vazquez 1986b; Larkin 1987). The ploidy level of the starting material, for example, has been found to be a strong determinant of the frequency of ploidy changes in culture. In Gramineae, where comparisons have been made between polyploid and diploid genotypes of the same species, the former tended to show higher variation in culture. For example, 12 of 47 regenerates from two 4x ryegrass (Lolium multiflorum L.) genotypes lost up to three chromosomes, while all the regenerants of the respective 2x genotypes retained the normal diploid chromosome number (Jackson and Dale 1988). There seems to exist a positive correlation between polyploidization and the occurrence of aneuploidy, an indication that there is a higher tolerance to aneuploid variation at higher ploidy levels (cf., Larkin and Scowcroft 1981; D'Amato 1985). Further examples of genotypic influence on the frequency of chromosome doubling, aneuploidy or structural changes were noted in rice (Sun et al. 1983) and oats (McCoy et al. 1982).

Chromosome structural changes have been observed in mitotic figures in callus of varying age. They may originate from chromatid breakage and subsequent healing or fusion of broken ends, in which case they can bring about continuing variation by initiating a breakage-fusion-bridge cycle that can be self-perpetuating over a number of cell generations (cf., McClintock 1978). Structural changes may occur at the diploid level, but their incidence was also found to be higher after polyploidization (cf., D'Amato 1985).






20


Media Comoosition and Tissue Culture Procedures. The level of chromosome variability and its fate during culture is determined by the extent of preexisting variation, the rate of de novo production of variant karyotypes and their relative competitive ability. Continuing production of new karyotypes with identical chances of survival would lead to an accumulation of variants and, consequently, a steady Increase in heterogeneity of the culture (Bayliss 1980). However, since in reality not all the variants are equally good competitors, such cultures eventually reach an equilibrium where variability does not increase any further. It has been shown that differences in the competitive ability of different cell lines can be strongly influenced (to the point of becoming reversed) simply by changing the transfer frequency or the composition of the nutrient medium (Bayliss 1980). The result is that, after a period of rapid increase in variability as formerly described, one or a few cell lines with a high selective advantage gradually emerge to make up the bulk of or the whole cell population in culture. Much attention has been devoted to the effects of medium constituents, especially growth regulators, on chromosome variability, but data have been conflicting and difficult to interpret (Karp and Bright 1985; Karp 1989). It is particularly difficult to separate the roles of substances as triggers of genetic changes from agents promoting the selective growth of specific variant cell types (Bayliss 1980; Karp 1989). Because of their involvement in the control of cell division (cf., Skoog and Miller 1957; Furuya 1984; Gould 1984), auxins (in particular 2,4-D) and cytokinins have been considered as prime candidates in promoting chromosome variability through perturbations of the mitotic process (Karp and Bright 1985; Karp 1989). While some of the data in the literature seem to support this view, most do not. In the particular case of grasses and cereals, where 2,4-D has been the hormone of choice for establishing embryogenic callus cultures, there is absolutely no evidence that the auxin (or any other growth regulators) is mutagenic at all; the choice of explant and the regeneration pathway are much more important factors for consideration (Morrish et al. 1987). Reports of changes in ploidy






21


level associated with growth regulators (e.g., Ghosh and Gadgil 1979) may be only the result of the hormones promoting cell division at specific ploidy levels (Karp 1989).

The influence of time in culture on chromosome instability has long been recognized (Torrey 1967; Sacrist&n and Melchers 1969; Zagorska et al. 1974; Jones and Murashige 1974; D'Amato 1977; Ghosh and Gadgil 1979; Boucaud and Caultier 1981; Sutter and Langhans 1981; L6rz and Scowcroft 1983; Westerhof et al. 1984; Stimart 1986; Cassells and Morrish 1987; Binarova and Dolezel 1988; Lee and Phillips 1988; Karp 1989; Swartz 1991). In general, the longer explants remain in the culture phase, the greater the chromosomal instability; variant karyotypes commonly accumulate with increasing age of cultures. Because, as described earlier, some of the variation found in the callus is usually transmitted to the regenerants, the relative proportion of variant plants produced during successive passages generally also increases. In Gramineae, an increase in regenerants displaying polyploidy, aneuploidy or chromosome structural changes has been observed, for example, in maize (Lee and Phillips 1987; Armstrong and Phillips 1988; Benzion and Phillips 1988), rice (Fukui 1986), oat (McCoy et al. 1982), barley (Orton 1980) and triploid ryegrass (Ahloowalia 1983). In most of these cases, the proportion of karyotypically normal plants was 100%, or close to 100%, during the first transfers. As a consequence, restricting the number of callus passages in culture has been frequently recommended as an effective means of reducing variability, whenever genetic stability is essential to the research (Karp 1989). There is also good evidence that frequent transfers, as compared to extended subculture intervals, tend to yield more stable callus cultures (Cassells and Morrish 1987; Lee and Phillips 1988; Karp 1989).

Evolution of chromosome variability may finally be dependent on whether cells are grown as callus masses on solid medium or in liquid medium as cell suspension cultures. Where comparisons have been made, a higher level of variability was normally found in the callus cultures (e.g., Singh and Harvey 1975; Ashmore and Shapcott 1989).






22


Regeneration from protoplast-derived cultures is generally associated with unusually high levels of chromosomal variation (Karp 1989). This may be a reflection of the increased complexity and length of time in culture associated with protoplast regeneration protocols, or it may be simply a consequence of the additional physiological stress imposed by the removal of the cell wall.


Regeneration Pathway

Histological and morphological studies have provided evidence that plant regeneration from callus cultures may occur by two different distinct pathways: adventitious shoot morphogenesis (organogenesis) or somatic embryogenesis (Morrish et al. 1987; I.K. Vasil 1987). Both morphogenic processes involve a complex sequence of genetically controlled developmental steps. Since karyologically aberrant cells may be impaired in their control of one or more of these steps, morphogenesis acts as a screen that prevents a portion of the variant karyotypes from being transmitted to the regenerated plants. This would explain the common observation that the increasing chromosomal variability often associated with aging cultures is negatively correlated with their regeneration potential (e.g., Murashige and Nakano 1967; Torrey 1967; Smith and Street 1974; Zagorska et al. 1974; Orton 1985; Jha and Sen 1987). Differences in the extent to which karyotypically variant cells are excluded from plant regeneration are found among and within species, demonstrating that the process is under genetic control.

Callus cultures have been mostly described as masses of dedifferentiated cells proliferating in a random and unorganized way. Although this seems to be true for the majority of the dicots, histological investigations have revealed that in the monocots a different type of callus often occurs. This has been described as compact and organized and is commonly interpreted as a mass of proliferating partially suppressed embryos in various stages of development, rather than completely undifferentiated






23


tissue. Compact calli of this nature are predominantly produced by meristematic tissues in immature and young tissues of inflorescences, embryos and leaves (e.g., Hunault 1979; Dale 1980; V. Vasil and I.K. Vasil 1980, 1981a; Lu and I.K. Vasil 1981, 1982; Wernicke et al. 1981, 1982; Hanning and Conger 1982; Lu et al. 1982; Thomas and Scott 1985). Because organized compact calli are often associated with regeneration via the embryogenic pathway they have been termed embryogenic calli. In Gramineae, by far the most well studied group in the monocots, production of embryogenic callus has now been reported in most cereal and other grass species, suggesting that somatic embryogenesis may be the most common pathway of plant regeneration in these plants (I.K. Vasil 1982, 1983a,b, 1985, 1987; I.K. Vasil and V. Vasil 1986).

In many species that regenerate from embryogenic callus cultures a strong selective constraint seems to exist, that minimizes the amount of chromosome variation observed in the regenerants. This was observed, for example, in cultures and regenerants of several monocots such as daylilies (Krikorian et al. 1981), asparagus (Becker and Reuther 1986) and most grasses, including cereals (I.K. Vasil and V. Vasil 1986a; Morrish et al. 1987). The same apparent karyological stability seems to apply also to embryogenic cultures derived from zygotic embryos of conifers (Papes et al. 1983; Schuller et al. 1989), which are composed of masses of proliferating proembryos rather than disorganized calli (cf., von Arnold and Wallin 1988).

The relationship between the callus growth pattern and the level of chromosome variability is particularly evident in cultures where both embryogenic and nonembryogenic types coexist. In Gramineae, for example, compact organized embryogenic callus may occur together with a more dedifferentiated, friable and more or less non-morphogenic type of callus (cf., I.K. Vasil 1985, 1987; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987). Comparative cytogenetic data on the two types of calli (e.g., Scheunert et al. 1978; Swedlund and I.K. Vasil 1985; Singh 1986) show that the






24


embryogenic callus generally exhibits a lower degree of karyotypic variability than that found in other kinds of cultures (e.g., Orton 1980; Ahloowalia 1982). Since the nature of the callus can determine both its ability to regenerate and the stability of the regenerants (I.K. Vasil and V. Vasil 1986; Morrish et al. 1987), it is conceivable that the primordial nature of embryogenic calli implies an organized pattern and a rate of cell division comparable to those in meristems of intact plants (I.K. Vasil 1985, 1987). As in meristems, these kinds of cultures would then retain a high level of karyological stability. This higher chromosomal stability of embryogenic calli and the fact that somatic embryogenesis per se seems to be a highly selective process which strongly favors regeneration from normal cells, would explain why only a small portion of the karyological variation, if any, in embryogenic cultures is recovered in the regenerants (I.K. Vasil 1983a, 1987; Swedlund and I.K. Vasil 1985). Regenerants from embryogenic cultures tend to be diploid, or at least euploid, even when regenerating from calli or suspension cultures where polyploidy or aneuploidy had been formerly observed (e.g., Swedlund and I.K. Vasil 1985; Hahne and Hoffmann 1986; Sengupta et al. 1988; Taniguchi and Tanaka 1989).

Based on the high level of karyotypic stability found in both the embryogenic calli and their regenerants in several grass species (e.g., Hanna et al. 1984; Armstrong and Green 1985; Swedlund and I.K. Vasil 1985; Rajasekaran et al. 1986), I.K. Vasil (1983a, 1985, 1987) postulated a higher stability of embryogenic cultures compared with organogenic ones and recommended their use whenever genetic stability and uniformity are needed. Such generalization, however, has been the theme of debate. Many claim that, although karyological and phenotypic stability can often be correlated with the embryogenic pathway, varying frequencies of abnormal regenerants from embryogenic tissue cultures have also been reported in the literature (G6bel et al. 1986; Karp 1989). These may have resulted from cultures where both the embryogenic and organogenic pathways coexisted, as has been described in wheat (Karp and Maddock






25


1984) where the variation reported may result from plants being regenerated from the mixed callus via the organogenic pathway (Maddock 1985). In maize, where both regeneration pathways also have been reported (Green and Phillips 1975; Green 1982; Earle and Gracen 1985; Armstrong and Phillips 1988), a detailed analysis of maize plants regenerated from immature embryo-derived embryogenic callus cultures and their sexual progeny showed variation occurring at both the morphological and molecular levels (G6bel et al. 1986; Brown et al. 1991). However, the authors were the first to acknowledge that ". . .although the intention was to use somatic, embryo-derived structures for plant regeneration, it cannot be ascertained that all of the regenerants are somatic and embryo derived ... ." (p. 23 in G6bel et al. 1986). In an identical study, but where plants were regenerated from strictly controlled embryogenic callus cultures of Napier grass (Pennisetum purpureum K. Schum.), Shenoy and I.K. Vasil (1992) showed a complete absence of variation at the morphological, biochemical (14 different isozyme loci) and molecular (RFLP analysis of the mitochondrial, plastid and nuclear genomes) levels. Identical results were observed for the mitochondrial, plastid and nuclear genomes in sugarcane (Chowdhury and I.K. Vasil 1993) and meadow fescue (Vall6s et al. 1993), lending further support to the suggestion that regeneration via somatic embryogenesis largely avoids phenotypic and genetic instability. Finally, the variants that have been reported may also be transient in nature (Larkin and Scowcroft 1983b; Irvine 1984), resulting from epigenetic rather than genetic changes in the regenerated plants.

Regardless of the final verdict on the applicability of somatic embryogenesis as a way of obtaining stable regenerants from tissue cultures, it does have advantages as a mode of regeneration. It imparts genetic uniformity, a consequence of the fact that somatic embryos, like their zygotic counterparts, derive largely from single cells (cf., Morrish et al. 1987; I.K. Vasil 1987). This prevents the occurrence of mosaics and






26


chimeras at the plant level (Morrish et al. 1987; I.K. Vasil 1987; Armstrong and Phillips 1988; Karp 1989).


Possible Causes of Chromosomal Variability


By emphasizing chromosomal structural rearrangements, some of the possibilities of how culture-induced alterations of the intracellular environment could lead to chromosomal changes have been recently detailed (Lee and Phillips 1988). Consistent with various reports where rearrangements have been shown to involve breakage preferably at or close to heterochromatic regions (SacristAn 1971; McCoy et al. 1982; Lapitan et al. 1984; Murata and Orton 1984; Johnson et al. 1987; Lee and Phillips 1987; Benzion and Phillips 1988), it was speculated that late replication, characteristic of heterochromatic chromosome segments, may be further delayed under culture conditions, causing perturbations to DNA synthesis. If such delay lasts until mitosis, formation of nonreplicated heterochromatin bridges, with consequent breakage at anaphase, would be the consequence (Lee and Phillips 1988).

While breakage events within or near heterochromatin would be the primary source of chromosome rearrangements, further variation can result from breakagefusion-bridge (BFB) cycles, which tend to be perpetuated through the formation of dicentric chromosomes (Lee and Phillips 1988). Repeated breakage associated with BFB cycles may trigger various forms of genome reorganization, from gross and minor translocations to excision and splicing of transposable elements, affecting even chromosomes not involved in the BFB cycle at all (McClintock 1978). Activation of transposable elements as a result of culture-induced chromosomal changes has been confirmed in maize (Peschke et al. 1986, 1987, 1991), for example, and it has been speculated that transposon-mediated variability could be a general phenomenon in callus cultures (Larkin and Scowcroft 1981; Larkin et al. 1984; Toncelli et al. 1985; L6rz and Brown 1986).






27


Another possible component on the induction of variation, proposed from evidence obtained from other organisms, could be the nucleotide pool imbalance associated with nutrient depletion in plant tissue cultures. This, in turn, would cause disturbed DNA replication and repair, which could lead to chromosome breakage, reciprocal and nonreciprocal sister chromatid exchange and mitotic crossing-over or somatic recombination events between nonhomologous chromosomes (Lee and Phillips 1988). Accumulation of toxic and potentially 'automutagenic' metabolites in ageing cultures could also have the same effect (cf., D'Amato 1985).

Phillips et al. (1990) attempted to define a unified theory that would encompass in a single cause such seemingly unrelated events as single gene mutations, activation of transposable elements, qualitative trait variation and chromosome breakage. They postulated that the methylation status of the DNA, easily altered by the tissue culture environment (e.g., nucleotide pool imbalance, as formerly mentioned), is at the origin of DNA modifications that, in one way or another, could lead to all of the most frequent disturbances that occur in plant tissue cultures. This would imply that the mechanism of in vitro-induced variation is epigenetic rather than genetic, an interesting approach that certainly may help explain many of the unique characteristics of somaclonal variation, in particular its high frequency and widespread occurrence, the presence of seemingly stable mutational events that are lost upon crossing (e.g., Oono 1985) and the peculiar observation that some variant phenotypes are characteristic and found only in the somaclonal populations.

Finally, and in another class by itself, the presence of an altered cytoskeleton has also been related to tissue-culture-induced mitotic disturbances. Studies have shown that specific microtubular arrays associated with cell division may be absent or less distinct in callus cells than in intact meristems (cf., Seagull 1989). The chromosome variation frequently observed in vitro could thus be caused by a malfunctioning of the mitotic and cytokinetic apparatus. This would be particularly true






28


in protoplast cultures, which are known to have a disturbed microtubular organization (Hahne and Hoffmann. 1986).

Considerable progress has been made over the years in our perception of tissue-culture-induced variation. The major lesson to retain, though, is that a great deal more remains to be learned about. Gathering accurate information at the morphological, karyological, biochemical and molecular levels is critical, but solid theories are likely to be crafted only with careful integration of all this knowledge, an approach that so far has been somewhat neglected.












CHAPTER 3

SOMATIC EMBRYOGENESIS AND LONG-TERM REGENERATION
IN CALLUS CULTURES OF DIPLOPERENNIAL TEOSINTE



Introduction

Diploperennial teosinte (diploid perennial teosinte, Zea diploperennis Iitis, Doebley & Guzm~n) is a newly discovered perennial teosinte, or '"wild maize", from the highlands of Mexico (Guzmin Mejia 1978; Iltis et al. 1979) (Fig. 3.1). Taxonomically a relative of maize (Z. mays L.), Z. diploperennis was described as endemic in nature to a small localized area of the Sierra de Manantl&n, a mountain range of western Mexico, where it grows alongside small streams and, sometimes, on the edges of maize fields or in grazed pastures.

Morphologically, diploperennial teosinte differs from perennial teosinte [Z. perennis (Hitchcock) Reeves and Mangelsdorf], a rare tetraploid species endemic to the same area, by the presence of dimorphic rhizomes, a more robust growth habit and larger tassels. Specially relevant, however, is that Z. diploperennis is a diploid (2n = 2x = 20), which permits the potential transmission of its many agriculturally useful traits to maize. For this it has been called the "botanical find of the century" (Nault and Findley 1981).

Genetically, the species is a hybridizer's dream. Diploperennial teosinte carries genes for immunity or tolerance to several important pests and diseases in its germplasm (Nault 1980; Nault et al. 1980, 1982), together with a variety of other agronomically important traits. Zea diploperennis produces fertile offspring when crossed with maize and, because it retains the primitive perennial trait in a diploid condition, it is the major potential genetic source for the production of a perennial


29






























Fig. 3.1 Mature flowering plant of diploperennial teosinte, Zea diploperennis.

Fig. 3.2 Plants of diploperennial teosinte grown under greenhouse-controlled
conditions. The metal frames above the plants supported black cloth, used to simulate short day conditions needed to force flowering out of
season.





31








i






32


maize, a goal that has long eluded breeders (Iltis et al. 1979; Nault and Findley 1981; Crosswhite 1982). Since its introduction in the late seventies, the species has been successfully incorporated in a number of maize breeding programs, many of them currently undergoing evaluation and selection (Cgmara-Herndndez and Mangelsdorf 1981; Galinat 1981; Mangelsdorf et al. 1981; Nault and Findley 1981; Crosswhite 1982; Findley et al. 1982; Magoja and Pischedda 1986; Magoja and Palacios 1987; Palacios and Magoja 1987, 1988; Pischedda and Magoja 1987, 1988, 1990; Corcuera and Magoja 1988a,b,c; Carlson and Price 1989a,b).

Other than its promising potential in breeding, diploperennial teosinte makes an excellent choice for studies involving tissue cultures and tissue-culture-generated variation. It is a close relative of maize, a species that is particularly well characterized at the tissue culture level (Green and Phillips 1975; Springer et al. 1979; Green 1982; Lu et al. 1982, 1983; Chang 1983; Novak et al. 1983; V. Vasil et al. 1983, 1984, 1985; Holbrook and Molnar 1984; Armstrong and Green 1985; Duncan et al. 1985; Lowe et al. 1985; Radojevic 1985; Fahey et al. 1986; Hodges et al. 1986; Petolino and Jones 1986; Rhodes et al. 1986; Suprasanna et al. 1986; V. Vasil and I.K. Vasil 1986; Close and Ludeman 1987; Conger et al. 1987; Wang 1987; Wilkinson and Thompson 1987; Duncan and Widholm 1988; Pescitelli et al. 1989; Pareddy and Petolino 1990; Songstad et al. 1992) as well as at the karyological, biochemical and molecular levels and for which somaclonal variation has been formerly reported and extensively evaluated (e.g., Gengenbach et al. 1981; Hibberd and Green 1982; McCoy and Phillips 1982; Umbeck and Gengenbach 1983; Brettel et al. 1986a; Gbbel et al. 1986; Lee and Phillips 1987; Armstrong and Phillips 1988). In addition, being diploid, perennial and fully fertile, diploperennial teosinte should produce regenerants from tissue culture that can be analyzed over long periods of time and across generations, an absolute need when attempting to clarify the origin and nature of somaclonal variation.






33


Protocols that lead to the successful tissue culture and regeneration of diploperennial teosinte have been published (Prioli et al. 1984, 1985; Sondahl et al. 1984; Swedlund and Locy 1988). As for other members of Gramineae (V. Vasil and I.K. Vasil 1984a; I.K. Vasil 1985, 1987; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987),

(1) use of very young, immature explants containing mostly undifferentiated meristematic cells, (2) addition to the media of relatively high concentrations of synthetic auxins and (3) early recognition and selection of regenerable calli were the keys to the successful initiation and maintenance of totipotent cultures. Reports on regeneration from tissue cultures of diploperennial teosinte, however, are conflicting. Upon culturing apical and lateral meristems of Zea diploperennis, Prioli et al. (1984) reported regeneration from a semi-friable callus from which adventitious shoots arose. Although the morphology of regeneration was not investigated, it was claimed that typical somatic embryogenesis did not occur (Prioli et al. 1984, 1985). Conversely, Swedlund and Locy (1988) reported a system much closer to that commonly described for other species of Gramineae (V. Vasil and I.K. Vasil 1984a), with both nonembryogenic and embryogenic calli being formed from immature zygotic embryos and shoot apices. Regeneration occurred by somatic embryogenesis, with the embryogenic cells described as small, densely cytoplasmic and starchy. Somatic-embryo-derived plants did not differ in chromosome number or overall morphology from the original donor plants (Swedlund and Locy 1988).

This chapter reports in detail the results of studies leading to the development of a protocol for the reproducible induction of somatic embryogenesis and long-term (several years) regeneration from callus cultures of diploperennial teosinte. In particular, it describes the effect of using alternative compounds exhibiting auxin-like activity and demonstrates the potential importance of L-proline and other medium additives in optimizing the embryogenic response.






34

Materials and Methods


Plant Material

Plants of diploperennial teosinte were raised from seed obtained from Professor H.H. Iltis (Department of Botany, University of Wisconsin-Madison). The wild-collected seeds, harvested near the type location in Mexico, were accompanied by the following information:



Zea diploperennis
ex Herbarium, University of Wisconsin, Madison, Wisconsin.
Zea diploperennis Iltis, Doebley & GuzmAn.
Plants 15-25 dm tall in dense, many-stemmed (30 to 100 in each) colonies.
a) Thickets along stream bordered by Alnus trees.
b) Maize fields, and abandoned maize fields, with scattered Crataegus
mexicana.
Flat valley bottoms at Las Joyas (cf. El Chante topo-sheet), 7.8 Km W by WSW of
Rincon de Manantlan, ca. 16 Km SW of El Chante, Jalisco, Mexico, JAN1984.
19� 35' 50" N, 1040 17' W, Alt. ca. 1,900 m.
H. H. Iltis and R. Guzmcn M., No. 29115.
Field work supported by National Science Foundation Grant No. BM S74-21861;
O. N. Allen Herbarium Fund; Research Committee, Graduate School, University of
Wisconsin-Madison; and Pioneer Hi-bred International.



To these were also added other accessions, obtained as vegetative propagations from Drs. D. Pring and W. Judd (Plant Pathology and Botany Departments, University of Florida, respectively). These plants were also derived from seed donated by Professor Iltis and probably originate from the same field collection.

The fruits were planted, one caryopsis per gallon plastic container, in Metromix� Vegetable Plug Mix amended with 20% Perlite. Germination occurred in 3-5 days and a large number of plants were raised to maturity. These were grown at a density of 9 pots/m2 in a greenhouse (Fig. 3.2). All plants were fertilized once per growth cycle with Osmocote�, a slow-release fertilizer (17-6-10 plus minor elements). Mean average






35


daytime temperatures ranged from 20o-250C in winter to 300-350C in summer. The temperature was not allowed to fall below 160C during cold winter nights but it often rose well above 350C during summer.

Because teosintes are naturally short-day plants (Emerson 1924; Mangelsdorf 1974), flowering was either induced naturally, under the influence of the short days of winter and early spring, or was forced whenever needed by keeping the plants in darkness for ca. 16 hours a day (Emerson 1924; Mangelsdorf 1974). Early application of short-day treatments to specimens grown from seed, however, was best avoided, as it consistently resulted in the production of short plants that, although physiologically mature, flowered only rather sparsely.

Young elongating leaf bases, young tassels and mature and immature zygotic embryos were used as explants. For both leaf bases and young developing tassels terminal shoots were cut from the plants, washed with tap water and surface sterilized with 90% ethanol. In a sterile laminar-flow hood the green mature outer leaves were removed to expose the white to yellowish-green inner expanding leaves or the immature staminate inflorescences. Sections from the region slightly below to 2 cm above the shoot apex were sliced transversely into two millimeter thin segments that were placed in presterilized disposable plastic Petri dishes (100 mm diameter, 15 mm high) on approximately 25 ml of culture medium, one plant per dish. The normal polarity was maintained for the leaf sections in culture.

Controlled pollinations were performed by bagging both the staminate and the pistillate inflorescences before anthesis and hand-pollinating each individual ear when pollen was available. Pollinations were carried out with pollen from a different plant or from tillers of different physiological ages in the same plant because under our conditions maturation of the male inflorescences preceded that of the female by several days. Immature ears were harvested when the embryos were at the desired developmental stage, normally between the 8th and 10th day after pollination. This was






36


done except for the embryo size study, for which ears were harvested every day from day 5 to day 15. The ears were de-husked and the green caryopses were surface sterilized in 10% Clorox containing two drops of a surfactant (Tween 20) for 20 minutes, then rinsed in several changes of sterile distilled water. The fruit cases were then carefully removed in a sterile laminar-flow hood to expose the internal structures.

Immature embryos, ranging in size from less than one millimeter to five millimeters, were carefully dissected and cultured as previously described for maize (Lu et al. 1983) and diploperennial teosinte (Swedlund and Locy 1988). The embryos were usually isolated on the same day the ear was harvested, but occasionally the ears were refrigerated overnight and the embryos were dissected the following day. Embryos were placed on the surface of the culture medium, embryonic axis down, in presterilized disposable plastic Petri dishes (100 mm diameter, 15 mm high) containing approximately 25 ml of culture medium, ten embryos per dish. Whenever possible, pieces of the semi-solid endosperm (the stage of development between day 9 and day 13 after pollination) were also placed in culture following the same protocol. At very young stages of development, when the embryo was too small to be safely dissected, the whole dissected ovary was placed in culture.


Induction of Somatic Embrvogenesis

The basal nutrient medium for induction of somatic embryogenesis consisted of Murashige and Skoog's (1962) major and minor nutrients and vitamins, 100 mg/I myoinositol and 5% coconut milk (MSC basal medium). The following parameters were tested for their influence on induction of somatic embryogenesis:

1. Embryo developmental stage, expressed by the length of the embryo as

measured from the base to the tip of the scutellum.

2. Auxin type and concentration: a series of four IAA conjugates, including IAAla

(indoleacetyl-D,L-alanine), IAAsp (indoleacetyl-D,L-aspartic acid), IAGly






37


(indoleacetyl-glycine) and IAPhe (indoleacetyl-D,L-phenylalanine), NAA (naphthaleneacetic acid), 2,4-D (2,4-dichlorophenoxyacetic acid), dicamba (3,6-dichloro-2-methoxybenzoic acid) and picloram (4-amino-3,5,6-trichloropicolinic acid). All auxins were tested at the following concentrations: 2.5, 5.0, 10.0

and 25.0 pM.

3. Sucrose: 3%, 6% and 12%.

4. L-proline: 12 mM.

All media were solidified by addition of 2 g/l Gelritem and pH was adjusted to 5.8 before autoclaving at 121OC for 20 minutes. L-proline and all the auxins were filtersterilized and added after autoclaving. The Petri dishes were sealed with Parafilm and cultures were incubated at 280C in the dark. Callus Maintenance and Multiplication

Except for the factor(s) under testing, a basal nutrient medium containing Murashige and Skoog's major and minor nutrients and vitamins with 100 mg/ myoinositol, 12 mM L-proline, 500 mg/I casamino acids, 2% sucrose, 1% glucose, 10 pM picloram and 2 g/l Gelrite, pH 5.8, was used in all experiments. The following parameters were evaluated for their effect on growth rate of embryogenic calli:

1. Auxin type: 2,4-D, dicamba and picloram, at an equimolar concentration of

10pM.

2. Auxin concentration: only picloram was used, at the concentrations of 2.5, 5.0,

10.0, 20.0 and 30.0 pM.

3. Carbohydrates: glucose and sucrose, in the following combinations: glucose

3%; glucose 2% + sucrose 1%; glucose 1% + sucrose 2% and sucrose 3%.

4. Casamino acids: 500 mg/l.

5. L-proline: 12, 25, 50 and 100 mM.






38


Ten pieces of embryogenic callus were initially placed in each Petri dish. To minimize the lag phase and avoid carry-over effects from the maintenance medium all cultures under evaluation were subcultured twice, at 4 week intervals, in the experimental medium. All measurements were taken on the second subculture. Whenever a best combination of factors was determined, callus pieces from that treatment were tested for regeneration, to assess if the factor(s) that optimized growth in tissue culture had any effect on the expression of totipotency. This was done by placing pieces of embryogenic callus in a regeneration medium consisting of Murashige and Skoog's major and minor nutrients and vitamins with 100 mg/I myo-inositol, 1 pM picloram and 2.5 pM kinetin and visually evaluating the potential to produce plantlets in vitro.


Data Collection and Analysis


Induction of somatic embryogenesis. Immature embryos were checked for growth every few days and scored for the presence of embryogenic or nonembryogenic callus four weeks after culture initiation. Hard, nodular, creamy-white callus with a convoluted surface, comparable to embryogenic calli from other grass species, was classified as embryogenic (E). Soft, friable, translucent callus with a watery appearance was classified as nonembryogenic (NE). No friable embryogenic (Type-II) callus, equivalent to that formerly described for maize (Green 1982, 1983), was produced. For both kinds the frequency of callus formation was converted to a percentage (number of explants showing that callus type / total number of explants in culture x 100). All percentage data were transformed using the arc-sine transformation for proportions (Steel and Torrie 1980) before analysis . The data presented in Tables 3.1 through 3.5 were analyzed by analysis of variance using the General Linear Models procedure of the Statistical Analysis System (SAS) for the PC, Release 6.04. All plotting and curve






39


fitting in Figs. 3.16 through 3.30 was done with Microcal Origin', version 2.75, running under Microsoft" Windows" 3.1.


Callus maintenance and proliferation. Fresh weights for all treatments were measured every other day for twenty six days. Results were plotted as growth curves, showing variation of fresh weight as a function of time. The data plotted in Figs. 3.21, 3.23, 3.25, 3.27 and 3.29 are the mean values and standard errors of the mean (SE) from the five replicates. Growth rates were calculated from the growth curves during the exponential phase of growth (day 2 through day 26) by linear regression techniques. The fitting function was determined by using the General Unear Models procedure of the Statistical Analysis System (SAS) for the PC, Release 6.04 and the calculated slopes were plotted against the variable under study. The results are shown in Tables 3.6 through 3.10 and Figs. 3.22, 3.24, 3.26, 3.28 and 3.30. All plotting and curve fitting in Figs. 3.16 through 3.30 was done with Microcal Originm, version 2.75, running under Microsoft� Windows' 3.1.


Plant Regeneration

Cultural procedures. Plants were regenerated from embryogenic cultures by transferring callus pieces onto MS basal medium with 0 or 1 pM picloram, with or without 2.5 pM of a cytokinin. Cytokinins tested included 6-furfurylaminopurine (kinetin), N6-benzyladenine (BA), N6-(2-isopentenyl)adenine (2iP) and zeatin. Regenerating callus pieces were exposed to a 16-hour photoperiod under cool white fluorescent tubes at 280C.

Germinated somatic embryos were transferred to half strength MS salts and vitamins with 2% sucrose in individual test tubes and kept under the same environmental conditions for further development and rooting. After establishment of an adequate root system, the young plants were transplanted into Conetainersm and gradually hardened off under high humidity until established, normally in 2-3 weeks.






40


Acclimated plants were transferred to gallon plastic containers, in Metromix� Vegetable Plug Mix amended with 20% Perlite, fertilized with Osmocote (17-6-10 plus minor elements) and grown to maturity in a greenhouse.

Scanning electron microscopy. Regenerating embryogenic callus pieces were dissected and fixed for scanning electron microscopy following the protocol detailed in V. Vasil and I.K. Vasil (1984b). Specimens were viewed with a Hitachi S-450 scanning electron microscope and photographed using Polaroid. Type 55 positive/negative black-and-white film.


Results


Induction of Somatic Embryogenesis

Donor tissue. All tissues placed in culture responded at least partially to the conditions of the culture medium but, regardless of the cultural conditions, embryogenic callus could be isolated only from endosperm explants and immature embryos. Sections from young expanding leaves started reacting as early as 4 days after being placed in culture. Typically, there was a swelling of the cut surface in contact with the medium after 5-7 days in culture. Some explants, mostly those originating from dark colored (anthocyanin-rich) plants, produced relatively large amounts of phenolic compounds that accumulated at the cut ends and diffused into and discolored the culture medium. In such cases, to prevent the toxic effect on the tissues normally associated with the presence of such compounds (V. Vasil and I.K. Vasil 1981), the explants were transferred either to clear portions of medium in the same dish or to a new dish as required until discoloration ceased, normally after 2-3 transfers. Soft friable callus started forming at a low frequency by the end of the second week, together with root hairs and suppressed root primordia that soon degenerated into gelatinous masses that ceased proliferation. At the end of four weeks, the frequency of leaf






41


explants responding to culture was less than 10% in all treatments and there was no trace of embryogenic callus. The same overall response was found in explants originating from immature male inflorescences, except that no roots were ever formed and the frequency of callus formation was close to 0%. Whenever callus was produced by these explants, it was soft, watery and nonembryogenic.

Cell proliferation was somewhat better from immature ovary explants. In most treatments the dissected ovaries proliferated profusely at the cut bases and produced large amounts of a soft, translucent, friable and nonembryogenic callus that could not be maintained by subculture. Despite the variations introduced in the medium no embryogenic callus was ever recovered from ovary cultures as well.

Forty-seven percent of the endosperm cultures, initiated in MSC basal medium containing 3% sucrose and 10 pM 2,4-D, produced embryogenic callus. This callus, that could be maintained easily through subculture, was in all respects undistinguishable from embryogenic calli derived from immature embryos (see below). Regeneration from endosperm-derived embryogenic cultures has been formerly achieved in other families of plants (e.g., Tulecke et al. 1988). The objectives of this study, however, precluded the use of polyploidy, and there were no attempts to obtain plants from these cultures or otherwise to make any further use of them.

Embryogenic callus was readily initiated from immature zygotic embryos. Cell proliferation in culture was evident in both immature and mature embryos, but the nature and time of appearance of the emerging calli varied considerably with the chemical constitution of the medium as well as with the developmental stage of the embryo. The appropriate developmental stages were assessed by measuring the embryo length, a parameter that is more reliable than embryo age (measured as postpollination time) for the selection of competent embryos in grass tissue cultures (V. Vasil and I.K. Vasil 1982).






42


Under the cultural conditions of this study, a typical greenhouse-grown plant of diploperennial teosinte produced between 8 and 12 caryopses per ear (female inflorescence). Not all caryopses matured at the same time and embryos dissected from different fruits within one single ear were frequently at different stages of development even after controlled pollinations. This is either a result of the stigmas (silks) having different lengths, the longer ones causing a delayed fertilization of the egg, or an environmental effect, as discussed for other species of Gramineae (V. Vasil and I.K. Vasil 1982). Since time alone could not be used to predict embryo development, a large number of inflorescences were used to ensure that enough embryos could be found at all the desired stages. Fig. 3.3 illustrates the relative development of both the endosperm and embryos in the fruit, as related to postpollination time. It also summarizes the overall response in culture of the embryos, as dictated by their size.

Three kinds of calli were produced in embryo cultures of diploperennial teosinte (Fig. 3.4). As described for most other grass species (I.K. Vasil 1987), the scutellar tissue of competent embryos in culture proliferated at the periphery, near the edges, giving rise to a ridge of highly organized tissue that soon proliferated into a creamywhite, compact and highly convoluted embryogenic type of callus (Fig. 3.5). A soft, translucent, friable callus was often produced from the same explant, by proliferation of the tissues in direct contact with the medium (Figs. 3.5 and 3.6). Such callus, which could be formed even in the absence of the nodular type, did not show any potential for embryogenesis.

A second type of embryogenic callus was occasionally formed from some of the proliferating scutella. Always produced at very low frequencies, this new kind of embryogenic callus, which was structurally undistinguishable from the more common type, could be maintained easily through several cycles of subculture without ever being associated with the presence of soft, friable, nonembryogenic callus, even in















DAYS POST POLLINATION 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15



EMBRYO LENGTH (mm) <1 1-2 1-3 2-3 2-4 3-4 3-5 4-5 5



ENDOSPERM STAGE [LIQUID GET SOLID



RESPONSE IN CULTURE NONE " NE CALLUS CA GERMINATION NO YES







Fig. 3.3 Embryo length, endosperm development, callus production and germination in embryo cultures of diploperennial teosinte during the first 15 days post pollination.

















Fig. 3.4 Somatic embryogenesis in diploperennial teosinte: immature zygotic embryo at the
time of culture. The early coleoptilar stage of development (shown) was the most
responsive to culture conditions.




Fig. 3.5 Somatic embryogenesis in diploperennial teosinte: a peripheral ridge of embryogenic
callus, sometimes producing more or less well defined proembryoids, was the typical first response to the conditions in vitro. Friable unorganized callus proliferated from
the embryo axis in contact with the medium.





45





























Fig. 3.6 Somatic embryogenesis in diploperennial teosinte: callus production
from an immature zygotic embryo, with nodular, compact, organized embryogenic growth (Type A embryogenic callus, top left), as well as
friable, unorganized, nonembryogenic growth (right).

Fig. 3.7 Somatic embryogenesis in diploperennial teosinte: Type C
embryogenic callus. Note the highly organized nature of the embryogenic callus and the complete absence of friable
nonembryogenic callus.







47

















Fig. 3.8 Somatic embryogenesis in diploperennial teosinte: scanning electron micrograph of
type C embryogenic callus. The most distinctive feature is the highly organized surface, spurting ridges and cup-shaped structures, that were interpreted as
developing scutella.




Fig. 3.9 Somatic embryogenesis in diploperennial teosinte: upon transfer to regeneration
medium, the small cup-shaped structures (Fig. 3.8) develop into prominent scutella.
A shoot meristem will soon develop on the concave side.






49



























Fig. 3.10 Somatic embryogenesis in diploperennial teosinte: scanning electron
micrograph of developing somatic embryos. At the center of a lobed scutellum (top left) a shoot meristem has developed, now surrounded by a coleoptilar ring (center). A somatic proembryo can also be seen
(top right).

Fig. 3.11 Somatic embryogenesis in diploperennial teosinte: atypical somatic
embryo, at the early coleoptilar stage. Note the aberrant scutellum and the leafy appendages at the margins, covered with trichomes.
Regardless of these anomalous features, which are a result of hormonal imbalance in culture, atypical embryos develop into normal
plants.






51





























Fig. 3.12 Somatic embryogenesis in diploperennial teosinte: regenerating
embryogenic callus, with a developing somatic embryo.
Fig. 3.13 Somatic embryogenesis in diploperennial teosinte: fully formed
somatic embryo, showing the white opaque scutellum, the coleoptile
and the coleorhiza, partially embedded in the callus mass.























































cv)K
LC)


















Fig. 3.14 Somatic embryogenesis in diploperennial teosinte: germinating somatic embryos,
showing the white compact scutellum, elongating coleoptile and a short, rounded coleorhiza. Radicle protrusion is inhibited at this stage by components of the
medium.




Fig. 3.15 Somatic embryogenesis in diploperennial teosinte: germinated somatic embryos
(twin embryos) showing a shared white compact scutellum, dark colored coleoptiles
and developing first leaf. Root elongation is normal at this stage.






































LC4C LOV

i i i'

In~mm ia






56


aging cultures. This is in sharp contrast with the type more commonly described for species of Gramineae (I.K. Vasil 1987), including diploperennial teosinte (Swedlund and Locy 1988). Embryogenic cultures of most grasses are actually a mix of embryogenic and nonembryogenic cells, a consequence of a continuous conversion of the embryogenic cells into the nonembryogenic type. To maintain the embryogenic potential of the cultures and to ensure that plant regeneration is from somatic embryos only, the embryogenic calli must be carefully selected and transferred every subculture (1.K. Vasil 1987), an added burden to the already fastidious task of initiating and maintaining such cultures.

The occurrence of more than one kind of embryogenic callus in diploperennial teosinte embryo cultures is not unique among Gramineae. Early recognition and isolation of appropriate callus types are becoming increasingly important in the establishment of successful long-term regeneration protocols. In maize, for example, regeneration from long-term embryogenic callus and suspension cultures is dependent on the production of a soft, friable, embryogenic (Type-II) callus. Recognition and isolation of this particular kind of callus, first described by Green (1982, 1983), was paramount in the production of transgenic maize plants (Fromm et al. 1990; GordonKamm et al. 1990). Likewise, the identification of a callus type for long-term maintenance and regeneration in wheat (Type C, Redway et al. 1990a) was essential to the establishment of regenerable cell suspension cultures (Redway et al. 1990b) and totipotent protoplasts (V. Vasil et al. 1990). Using Type C embryogenic callus as the regeneration system was the key to the successful production of transgenic plants in wheat (V. Vasil et a/. 1992).

By analogy to the embryogenic callus types formerly described in wheat, the more common compact callus found in diploperennial teosinte embryogenic cultures was termed Type A. Uke its wheat namesake, teosinte Type A embryogenic callus occurred at a relatively high frequency, and always in conjunction with nonembryogenic






57


growth (Fig. 3.6). The tendency to degenerate easily into nonembryogenic callus made strict selection at every subculture essential to a problematic long-term maintenance. Following the same terminology, the second kind of embryogenic callus found in embryogenic cultures of diploperennial teosinte was designated Type C (Fig. 3.7). Usually found at very low frequencies, teosinte Type C embryogenic callus was easy to maintain in long-term cultures and regenerated freely even after several years in vitro, characteristics that it shared with both maize Type-II and wheat Type C embryogenic calli. Unlike any of these, however, it was not derived from any preexisting or aged callus; it was isolated directly from the proliferating immature embryos. Structurally comparable to wheat, teosinte Type C callus was compact, nodular and scutellar in nature (rather than soft and friable, like maize Type-II). The most striking feature of teosinte Type C callus, however, was the absence of nonembryogenic proliferation, even in aging cultures, a trait that is not shared by either maize or wheat embryogenic calli (Fig. 3.7).

Scanning electron microscopy revealed the compact convoluted callus to be nothing but a large mass of continuously proliferating scutella. Provided with a well differentiated epidermal layer, this callus spurted numerous cup-shaped structures at the surface (Fig. 3.8), comparable to those found in other embryogenic grass cultures, where they have been described as similar to the scutellum in both morphology and structure (V. Vasil and I.K. Vasil 1981; Lu and I.K. Vasil 1982). Type C embryogenic callus could be maintained indefinitely by subculture, without any further organization taking place. When transferred to a regeneration medium, however, the cup-shaped structures became more prominent (Fig. 3.9), and a well defined shoot apex, surrounded by a coleoptilar ring, formed at their center (Fig. 3.10). The ring further developed into a more or less well defined coleoptile, from which the first leaves emerged as the newly formed embryo precociously germinated. Structures showing aberrant morphology, an obvious consequence of the hormonal imbalance in culture,






58


were not uncommon (Fig. 3.11). Scutella were often lobed, or formed small leafy appendages at the margins. Sometimes the whole scutellum became a leafy structure with characteristic trichomes on the surface, or several scutella fused laterally to form one single ridge of tissue associated with multiple embryonic apices.

Individual small segments of embryogenic callus could be separated and transferred independently to an appropriate regeneration medium. When kept under light they greatly enlarged, the whole fragment becoming a prominent white scutellum with a deep red coleoptile (Figs. 3.12 and 3.13) from which the first green leaves emerged (Figs 3.14 and 3.15). A developing primary root, protruding from a well defined coleorhiza, was perfectly visible in these germinating somatic embryos (Fig.

3.15).


Embryo developmental stage. Embryogenic callus cultures in Gramineae are often initiated from immature embryos, normally at a specific early stage of development (Lu et al. 1983). Cultures have also been obtained, however, from fully mature embryos and whole caryopses (e.g., McDaniel et al. 1982; Mehta et al. 1982; Botti and I.K. Vasil 1983; Rao et al. 1985). Since it is not predictable how any individual species will perform in culture, an experiment was conducted to find out the particular sequential pattern, if any, in the capacity to produce embryogenic callus from embryos of diploperennial teosinte.

The influence of embryo length on the formation of embryogenic callus is illustrated in Fig. 3.16. The relative frequencies of both embryogenic and nonembryogenic calli formed, as influenced by embryo size, are shown in Table 3.1. All experiments were done in MSC medium with 3% sucrose and 10 pM 2,4-D.

Cell proliferation resulting in the formation of calli occurred at all size ranges. Frequencies varied from a low of 33% for the smaller embryos (0-1 mm range) to a high of up to 90% in the upper size ranges (size ranges 2-3, 3-4 and 4-5 mm were not






59


100
STotal Embryogenic Callus
- = Type C Embryogenic Callus
Z
O 80


O 60

. J

o 40



o 20




0-1 1-2 2-3 3-4 4-5 EMBRYO LENGTH (mm)

Fig. 3.16 Influence of embryo length on embryogenic callus formation from immature embryo cultures of diploperennial teosinte.


Table 3.1 Influence of embryo length on callus formation from immature embryo cultures of diploperennial teosinte.



Embryo length NE callus Type A Type C Total Total
(mm) only* E callus* E callus* E callus* callus*

0-1 33.3a 0.0a 0.0a 0.0a 33.3a 1-2 0.0b 66.7b 1 0.0b 76.7b 76.7b 2-3 70.00 16.7a 0.0a 16.7a 86.7bc 3-4 80.0c 10.0a 0.0a 10.0a 90.00
4-5 76.7c 0.0a 0.08 0.0a 76.7bc


Percentage of immature embryos that formed calli. Each value is the mean of three replication (ten embryos per treatment for each replicate). Values within each column marked with the same letter
are not significantly different (P < 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980).
Each medium contained:
MS basal salts and vitamins
5% coconut milk
3% sucrose 10pM 2,4-D.
Embryo length as tested.






60


significantly different at the 95% confidence level). Production of embryogenic calli, however, was restricted to a much narrower developmental range. Embryogenic callus was produced at a high frequency (77%) by embryos in the 1-2 mm length range only. The same type of callus was also observed in embryos 2-3 and 3-4 mm long; their frequency of formation, however, was not significantly different from zero (95% confidence level). No embryogenic callus was ever produced by the smallest (s 1 mm) or the largest (> 4 mm, mature) embryos.

Type C embryogenic callus was produced at a relatively low frequency (10%) and only by embryos in the 1-2 mm size range (Fig. 3.16, Table 3.1). The occurrence of nonembryogenic callus alone was frequent in the larger size cultures (> 2 mm) and null in the 1-2 mm embryos. Germination was common in all embryos larger than 2 mm and occasional in the 1-2 mm range. Since embryos smaller than 1 mm and larger than 2mm gave mostly unsatisfactory results, they were not used in any of the subsequent experiments.


Auxin type and concentration. Problems associated with the use of 2,4-D in tissue culture investigations (e.g., Saunders and Bingham 1975; Collins et al. 1978; Hangarter et al. 1980) led some authors to look for possible substitutes that would be more effective and less toxic in vitro. Having in view the successful use of some alternative compounds (Vian 1976; Collins et al. 1978; Hangarter et al. 1980; Hanning and Conger 1982; Conger et al. 1983), experiments were made to compare the effect of 2,4-D with two other herbicides with strong auxin-like properties, dicamba and picloram. Naphthaleneacetic acid, which has been used alone or in combination with 2,4-D in plant tissue cultures, was also studied as a possible substitute. Four commercially available indoleacetylamino acid conjugates, considered to be more stable sources of auxin than the labile free IAA and less toxic than the persistent synthetic auxin analogs (Hangarter et al. 1980) were also tested. The results are summarized in Table 3.2 and






61


Fig. 3.17. All experiments were done in MSC medium with 3% sucrose, with 1-2 mm embryos.
The frequency of embryogenic response varied considerably with both the type and the concentration of the compound being used. No response was ever obtained in the absence of an auxin in the medium, neither was it in the presence of the IAA conjugates. Relatively high concentrations (25 pM) of naphthaleneacetic acid induced only moderate proliferation (35%) in the explants, and the callus formed in such conditions showed no embryogenic potential. This callus was also formed at 10 pM, but at a very low frequency (15%, statistically not different from zero at the 95% confidence level). No response of any kind was obtained at any of the lower concentrations (Table 3.2).

All three auxin herbicides (2,4-D, dicamba and picloram) promoted callus formation at all concentrations tested, and both embryogenic and nonembryogenic calli were produced under their influence. The results of Duncan's Multiple Range test using the arc-sine transformed data show that the frequencies of production of nonembryogenic callus alone were lowest at a concentration of 10 pM, and were significantly lower (10% and 3%, as compared to 20% for 2,4-D) in media containing dicamba and picloram, respectively (Table 3.2). The same concentration of 10 pM yielded the highest frequencies of embryogenic callus production, this time significantly higher (80% and 87%, as compared to 70% for 2,4-D) in media containing dicamba and picloram, respectively (Table 3.2, Fig. 3.17). The capacity to minimize nonembryogenic and maximize embryogenic callus production clearly makes this the most effective concentration to use when initiating embryogenic callus cultures in diploperennial teosinte. The shape of the fitting curves further demonstrates that 2,4-D is relatively less effective at all concentrations, when compared to either dicamba and picloram (Fig. 3.17). The frequency of response is significantly higher for both dicamba and picloram (95% confidence level) at optimal levels, and at least comparable to 2,4-D at






62

Table 3.2 Influence of auxin type and concentration on callus formation from
immature embryo cultures of diploperennial teosinte.



Auxin NE callus Type A Type C Total Total (pM) only* E callus* E callus* E callus* callus*


0 0.0h 0.0d 0.0C 0.0h 0.00 2.5 0.0h 0.0d 0.OC O.Oh 0.0e NAA 5 0.0h 0.0d O.OC O.Oh 0.0*
10 15.0gh 0.0d 0.0C 0.0h 15.0de 25 35.0bod 0.0d 0.0 . 0.0h 35.0cd

0 0.0h 0.0d 0.00 0.0h 0.0
2.5 20.0defg 0.0d 0.00 0.0h 20.0de 2,4-D 5 30.0cdef 36.7b 0.0c 36.7de 66.7b
10 20.0defg 66.7a 3.3bc 70.0b 90.0a 25 66.7a 0.0d 0.0c 0.0h 66.7b

0 0.0h 0.0d 0.0C 0.0h 0.0 2.5 16.7efgh 23.3bc 0.00 23.3'fg 40.00 Dicamba 5 40.0bc 23.3bc 20.0a 43.3d 83.3a
10 1 0.0gh 66.7a 13.3ab 80.0a 90.0a 25 50.0b 10.0cd 0.00 10.0gh 60.0b

0 0.0h 0.0d 0.00 0.0h 0.0* 2.5 6.7gh 23.3bc 3.3bc 26.7ef 33.3cd Picloram 5 33.3cde 36.7b 20.0a 56.7c 90.0a
10 3.3gh 76.7a 1 0.0abc 86.7a 90.0a 25 1 0.gh 1 6.7cd 0.0 16.7fgh 26.7cd


Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter
are not significantly different (P s 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980).
Each medium contained:
MS basal salts and vitamins
5% coconut milk
3% sucrose
Auxin (and concentration) as tested.
All embryos in the 1-2 mm range.































Fig. 3.17 Influence of auxin type and concentration on embryogenic callus
formation from immature embryo cultures of diploperennial teosinte.

Light gray: Total embryogenic callus
Dark gray: Type C embryogenic callus







64



100
2,4-D

80



60



40



20


o 100 . ......
DICAMBA
O

S80o U






o
Lu
O 20


100- -*
PICLORAM

80



60



40



20
0 201 10 AUXIN CONCENTRATION (pM) (log scale)
.. . .. . .

60......

....O...








AUXIN CONCETRATION ( .M ............






65


sub- and supra- optimal concentrations, implying a higher overall effectiveness of both dicamba and picloram at all concentrations (and, in particular, at the lower ones). It also suggests a lower toxicity of both dicamba and picloram at the high end. A slightly, but significant, better overall response to picloram, as compared to dicamba (Table 3.2) would suggest that this is the auxin of choice for the initiation of embryogenic callus cultures in diploperennial teosinte.

Type C embryogenic callus was induced at the optimal concentration (10 pM) by all three auxins. It was also produced at suboptimal concentrations in the presence of either dicamba or picloram. This would also indicate, and add further support to the idea, that these regulators are more effective and represent an improvement over the use of 2,4-D in the induction of somatic embryogenesis.

Osmoticum. High levels of exogenous sucrose have been found to increase the frequency of somatic embryogenesis in tissue cultures of maize (Lu et al. 1983; Tomes 1986; Close and Ludeman 1987). Since there are indications that the same may apply to diploperennial teosinte (Swedlund and Locy 1988), an experiment was conducted to compare the effect of various levels of sucrose in the medium on the capacity to produce embryogenic callus from immature zygotic embryos of diploperennial teosinte. The results of these experiments are summarized in Table 3.3 and Fig. 3.18. All experiments were done in MSC medium containing 10 pM 2,4-D. All embryos were in the 1-2 mm range.

The concentration of sucrose in the medium had only a moderate effect on the nature and efficiency of callus formation, although it had a marked repercussion on both the growth rate of the newly formed calli and the germination of the embryos. The production of nonembryogenic callus alone was significantly decreased with increasing sucrose concentrations, to the point of being virtually nonexistent at the higher concentrations (Table 3.3). The same general trend happened to the total callus







66


1000
STotal Embryogenic Callus
=Type C Embryogenic Callus

z
O 80T T......





0 40



O 20
U







2 6 12

SUCROSE CONCENTRATION (%)


Fig. 3.18 Influence of sucrose concentration on embryogenic callus formation
from immature embryo cultures of diploperennial teosinte.


Table 3.3 Influence of sucrose concentration on callus formation from immature
embryo cultures of diploperennial teosinte.



Sucrose NE callus Type A Type C Total Total
(%) only* E callus* E callus' E callus* callus*

3 16.7a 63.3a 6.7a 70.0a 86.7

6 6.7ab 80.0a 3.3* 83.3a 90.0a 12 3.3b 66.7a 0.0a 66.7* 70.0b


Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter
are not significantly different (P s; 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980).
Each medium contained:
MS basal salts and vitamins
5% coconut milk
10pJM 2,4-D
Sucrose concentration as tested.
All embryos in the 1-2 mm range.
.. . .... .. .
.. . .. . .
.. . ..-. . . . . . .. . .
.. . . . . . . .. . .. . . . .
......40.
.. . .. . . .. ,. . .
.. . . . . . . . .
.. . . . . . . .. .. . . . .
o 20..
.. .... .
40 -......
2. : ... ..

Fig 3.18 Inleneo. scoe.ocntain.nemr..nccllsfrmto fromimmtureembyo ultues f diloprenial eosnte

Tal . nlec fscoecnetato nclu omto rmimtr embryo ulture of diloperen.al tosinte

Sucos N calu Tpe TpeC Tta Tta
(%)~~~~...... ony' Ecals.Ecals' Ecals' clls
3 1 ....... 63..a 6 a7.0..7
6~. ..a .00 .3 .......9008.
1~203b6a00 6..7a.

Pecetae f mmtue mbyo ta frmd ali.Eah ale s.h.ma.oftheereliaton.(e
embrys pe tretmen foreachreplcate. Vaues ithi eac colmn .mr.edwiththe.a.e.ette
are~~~~~~......... noXinfcnl ifeet( .5Xb ucn utpe ag et(te adTre,18)
Each.mediu.contained
MS basalsalts.an.vitamin






67


production, which was significantly lower at 12% than at 3% or 6%. The level of sucrose in the medium, however, did not have as much influence on the quality and quantity of embryogenic callus produced (Table 3.3). No significant differences (95% confidence level) were found in the induction of embryogenic callus at the different sucrose levels, although a trend seems to exist favoring the intermediate level of 6% (Table 3.3, Fig. 3.18).

Embryogenic calli formed at higher sucrose concentrations were more compact and opaque and less convoluted than those formed at the lower concentration of 3%. They also developed at a considerably slower rate (as judged by qualitative visual observation) than that observed at the lower concentration level. In addition, cultures induced in high sucrose did not survive unless later subcultured to a low osmoticum medium (3% sucrose). Embryo germination, which was normally accompanied by nonembryogenic callus proliferation from the nodal region, was greatly reduced at the concentration of 6% and totally suppressed at 12%.

L-proline. The reported stimulation of somatic embryogenesis in maize by the addition of L-proline (Armstrong and Green 1982, 1985; Green et al. 1983) prompted an experiment to evaluate the effect of L-proline on the induction of somatic embryogenesis from immature zygotic embryos in diploperennial teosinte. The results of this experiment are summarized in Table 3.4 and Fig. 3.19. All experiments were done in MSC medium containing 3% sucrose and 10 pM 2,4-D. All embryos were in the 1-2 mm range.

At the single concentration (12 mM) tested no significant (95% confidence level) improvement was found in the embryogenic response as a result of the addition of Lproline. A trend seemed to exist, though, indicating that higher levels of L-proline might have helped prevent nonembryogenic callus formation, and promote embryogenesis (Table 3.4). That this might be the case is partially indicated by results obtained when testing the influence of L-proline in long-term callus maintenance (see below).






68


100
M Total Embryogenic Callus q , I Type C Embryogenic Callus
z



O 60



o 40


0 20




0 12 PROLINE CONCENTRATION (mM)

Fig. 3.19 Influence of proline on embryogenic callus formation from immature
embryo cultures of diploperennial teosinte.


Table 3.4 Influence of proline on callus formation from immature embryo
cultures of diploperennial teosinte.


Proline NE callus Type A Type C Total Total
(mM) only" E callus" E callus* E callus* callus*

0 1 0.0a 73.3a 3.3a 76.7a 86.7a 12 3.3a 56.7a 30.0a 86.7a 90.0a


Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter
are not significantly different (P s 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980).
Each medium contained:
MS basal salts and vitamins
5% coconut milk
3% sucrose 10pM 2,4-D
Proline as tested.
All embryos in the 1-2 mm range.






69


Combined effects. After making these refinements in the culture conditions, a final experiment was conducted to reexamine the effect of all factors that had been tested, when combined at their optimum levels, in a single medium and to compare their integrated performance to their individual effect. The results of this experiment are summarized in Table 3.5 and Fig. 3.20. The composite medium, hereafter referred to as PIC1, was MSC containing 6% sucrose, 10 pM picloram and 12 mM L-proline. All embryos were in the 1-2 mm range.

Combining all optimized factors in a single medium did not improve the overall performance of the system, as compared to each optimized factor taken individually. Total callus, Type A embryogenic callus and total embryogenic callus production showed no significant differences (95% confidence level) in media containing the optimized factors individually or in combination (Table 3.5).

Type C embryogenic callus was formed at a low frequency and only in the L-proline-containing medium and in the composite medium (Table 3.5). None was produced in the media containing sucrose or picloram at the optimal levels. This is in apparent contradiction with what was formerly observed when testing the effect of both sucrose and picloram (Tables 3.3 and 3.2, respectively). This unique type of callus, however, always tended to occur at very low frequencies. This was particularly true in the experiments that tested the individual effect of both those factors, where its frequency was often not significantly different from zero (Tables 3.2 and 3.3). Its production, however, might be promoted by the presence of L-proline (Tables 3.4 and

3.5).

Despite the relatively low occurrence of Type C embryogenic callus, its nature and characteristics suggested that it would be the ideal kind for long-term culture experiments, since it completely avoided the tedious, albeit essential, selection associated with every subculture in the more common Type A callus (I.K. Vasil 1987).







70



100
ETotal Embryogenic Callus
Type C Embryogenic Callus









J..........
o... 20 --iiiiiiiiiii i .iiiiiiii.i .X










SUCROSE L-PROLINE PICLORAM COMPOSITE
(6%) (12 mM) (10 pM)





Picoram, 10M 6.7a a 0.0a 83.3a 90.0a.. Composite 6.7a 80.0a 6.7b 60.7a a
























Percentage of immature embryos that formed calli. Each value is the mean of three replication (ten








10 pM pictoram 12 mM proline
All embros in the 1-2 mm ranae.
U- . .. .............. .X
U ) .. ........ 11 -1.1 ...........,..... .....
............. ...........
-.........
-XX . ............ ..
.... 40..........
0.................
................z. .....



















MS~~.... baa sa. an v.min
5% coconut.milk
6% sucrose 10MM ..pic .....l...r.am...
12..... mM.. .pro. ..ine...
Allemro i th 12 m rng ......






71


Therefore, all experiments reported hereafter were performed with Type C embryogenic callus only.


Callus Maintenance

The optimized medium developed for culture initiation from immature embryos (PIC1) also proved adequate for callus growth. In addition to examining the effect of this medium on long-term maintenance, however, the use of other compounds known to influence the growth rate and viability of cultures was also investigated. These studies were prompted by the report of Duncan et al. (1985), which indicated that an increase in the levels of reduced nitrogen, together with the addition of glucose to the medium (medium "D", Duncan et al. 1985), were particularly beneficial in the initiation and growth of maize tissue cultures.

Preliminary experiments indicated that, as with maize, such combination could also be helpful in the long-term maintenance of regenerable callus from diploperennial teosinte. As a result of these preliminary trials a new medium evolved that promoted superior growth of the embryogenic callus cultures. Much like PIC1, this medium (termed PIC2) was based on Murashige and Skoog's basal salts and vitamins and contained both L-proline and picloram, at the same levels as before (12 mM and 10 pM, respectively). Coconut milk, however, was replaced by the addition of 500 mg/I casamino acids to the medium and sucrose alone (at 6%) was substituted by a combination of both sucrose (2%) and glucose (1%). Once the qualities of this medium were recognized, the importance of the various components was then evaluated, to determine their individual effect on the long-term growth of the embryogenic callus cultures.


Auxin type and concentration. Since both dicamba and picloram had proven superior to 2,4-D at inducing embryogenic callus cultures from immature embryos, an experiment was performed to determine the relative efficiency of the three auxins at






72


maintaining embryogenic callus growth. The results are summarized in Table 3.6 and Figs. 3.21 and 3.22. All experiments were done using PIC2 medium, containing 10 pM of either 2,4-D, dicamba or picloram as the auxin source.

Regression analysis of the growth curves demonstrated a significant linear relationship between fresh weight and time for the three compounds under study (Fig. 3.21, Table 3.6). All three auxins (2,4-D, dicamba and picloram) promoted callus growth at the concentration tested (10 pM). The slopes of the linear regression lines, however, indicate that, once again, 2,4-D was relatively ineffective at sustaining growth when compared to both dicamba and picloram (Table 3.6, Fig. 3.22). At this concentration, callus growth rates were significantly higher for both dicamba and picloram (95% confidence level). A significantly better response to picloram, as compared to dicamba, suggested that this again was the best auxin source for the regular maintenance of embryogenic callus cultures in diploperennial teosinte. In addition to the effect on callus growth rate, supplementing the medium with either dicamba or picloram also resulted in suppression of both differentiation and callus browning that tended to occur towards the end of the subculture period (20-30 days) when 2,4-D was used as the auxin source.

A second experiment was conducted to determine whether the auxin concentration that best promoted callus initiation was also the best to support callus growth. The results of this experiment are summarized in Table 3.7 and Figs. 3.23 and 3.24. All experiments were done using PIC2 medium, with varying picloram concentrations.

Regression analysis of the growth curves again demonstrated a significant linear relationship between fresh weight and time for all concentrations (Fig. 3.23, Table 3.7). Picloram promoted callus growth at the whole range of concentrations tested. The Gaussian fit indicated that the optimal concentration for sustained proliferation was 10 pM (Table 3.7, Fig. 3.24). At this concentration embryogenic calli maintained the







73



5000
AUXIN

o 2,4-D
o DICAMBA
4000 A PICLORAM E .

1

C 3000


LU

C 2000 UL




1000

0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days)


Fig. 3.21 Embryogenic callus growth curves (fresh weight) with various auxins.



Table 3.6 Regression analysis for growth with different auxins.




Auxin (10 pM) Linear Regression* R2



2,4-D 895.1 + 18.1xa 0.993 Dicamba 899.0 + 69.9xb 0.999 Picloram 847.4 + 155.7xc 0.999



Based on data points from five replications (ten callus pieces per treatment for each replicate).
Slopes marked with the same letter are not significantly different (P s 0.05).
Each medium contained:
MS basal salts and vitamins
12 mM L-proline
500 mg/I casamino acids
20 g/I sucrose 10 g/I glucose
10 pM 2,4-D, dicamba or picloram, as tested
2 g/ Gelrite
pH 5.8






74



160

140

12010080
0 iiiiiiiiiiiiiiiiiiiiiiii

40


0 22,4-D DICAMBA PICLORAM AUXIN TYPE

Fig. 3.22 Effect of varying auxin type on embryogenic callus growth in
diploperennial teosinte (see also Table 3.7).

Each medium contained:
MS basal salts and vitamins
12 mM L-proline
500 mg/I casamino acids
20 g/l sucrose 10 g/I glucose
10 pM 2,4-D, dicamba or picloram as tested
2 g/I Gelrite
pH 5.8

Data points represent the mean (and SE) of five replications (ten callus
pieces per treatment in each replicate).







75



5000- PICLORAM


o 2.5 pM A 5 pM
4000- o 10 pM E v 20 pM
o 30 pM

S3000




L. 2000




1000 1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days)


Fig. 3.23 Embryogenic callus growth curves (fresh weight) with various picloram concentrations.


Table 3.7 Regression analysis for growth with various picloram concentrations.



Picloram (uM) Linear Regression* R2


2.5 1185.9 + 15.1a 0.942 5 1190.8 + 66.6b 0.998 10 887.9 + 156.6c 0.999 20 1023.8 + 60.1d 0.998 30 1148.0 + 33.6e 0.996


Based on data points from five replications (ten callus pieces per treatment for each replicate).
Slopes marked with the same letter are not significantly different (P s 0.05).
Each medium contained:
MS basal salts and vitamins
12 mM L-proline
500 mg/I casamino acids
20 g/l sucrose 10 g/l glucose
Picloram concentrations as tested
2 g/l Gelrite
pH 5.8






76


160

c 140- 120

100

80
UJ
60

F- 400
CC 202.5 5 10 20 40 PICLORAM CONCENTRATION (pM) (log scale)

Fig. 3.24 Effect of varying picloram concentration on embryogenic callus growth
in diploperennial teosinte (see also Table 3.8).

Each medium contained:
MS basal salts and vitamins
12 mM L-proline
500 mg/I casamino acids
20 g/l sucrose 10 g/I glucose
Picloram concentrations as tested
2 g/I Gelrite
pH 5.8

Data points represent the mean (and SE) of five replications (ten callus
pieces per treatment in each replicate).






77


characteristic creamy white color and convoluted surface and could be easily subcultured for indefinite periods of time, without ever forming nonembryogenic friable callus.

Both sub- and supra- optimal concentrations led to reduced callus growth. At 2.5 pM the proliferating scutella increased in size, and differentiation of leaf-like appendages, often covered with trichomes, occurred at a low frequency. Red pigmentation, due to anthocyanin biosynthesis, was also occasionally observed. Both events indicated that at this concentration picloram was no longer capable of completely supressing differentiation. At the highest concentrations growth was also reduced. At 30 pM callus pieces gradually changed color to a deeper shade and eventually died.

No friable callus was ever formed at any of the concentrations tested, which shows the peculiarity of embryogenic callus Type C. Callus proliferation was never observed in the complete absence of auxin.


Carbohydrates. Claims that glucose, in addition to sucrose, contributed to increased proliferation in maize embryogenic callus cultures (Duncan et al. 1985) led to the design of an experiment to determine the effect of different combinations of glucose and sucrose on diploperennial teosinte callus growth. The results of this experiment are summarized in Table 3.8 and Figs. 3.25 and 3.26. All experiments were done using PIC2 medium, with varying sugar combinations.

As in the former experiments, regression analysis of the growth curves confirmed a significant linear relationship between fresh weight and time for all sugar combinations (Fig. 3.25, Table 3.8). Glucose alone (3%) was unable to support adequate growth. In addition to the poor proliferation rate, callus pieces changed to a darker hue and became soft, in evident contrast with the harder, healthier calli formed in all sucrose containing media. Of these, the results of linear regression clearly indicated







78


5000
CARBOHYDRATES

a GLUCOSE (3%)
A GLUCOSE (2%)+ SUCROSE (1%)
4000- o GLUCOSE (1%)+ SUCROSE (2%) o SUCROSE (3%) g

2 3000
uJ



Cr
Lu. 2000




1000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days)


Fig. 3.25 Embryogenic callus growth curves (fresh weight) with various carbohydrate combinations.


Table 3.8 Regression analysis for growth with different carbohydrate combinations.



Carbohydrate(s) Unear Regression* R2


Glucose (3%) 1119.3 + 15.8xa 0.988 Glucose (2%) + Sucrose (1%) 998.4 + 60.2xb 0.977 Glucose (1%) + Sucrose (2%) 853.6 + 156.0xc 0.998 Sucrose (3%) 1038.1 + 86.8xd 0.997


Based on data points from five replications (ten callus pieces per treatment for each replicate).
Slopes marked with the same letter are not significantly different (P s 0.05).
Each medium contained:
MS basal salts and vitamins
12 mM L-proline
500 mg/l casamino acids
Carbohydrate combinations as tested
10pM picloram
2 g/l Gelrite
pH 5.8






79



160

140

_ 120S100

so -i
0



E 40
o
0 20


GLUCOSE (3%) GLUCOSE (2%) GLUCOSE (1%) SUCROSE (3%) SUCROSE (1%) SUCROSE (2%) CARBOHYDRATES

Fig. 3.26 Effect of varying carbohydrate combinations on embryogenic callus
growth in diploperennial teosinte (see also Table 3.9).

Each medium contained:
MS basal salts and vitamins
12 mM L-proline
500 mg/I casamino acids
Carbohydrate combinations as tested
10 pM picloram
2 g/I Gelrite
pH 5.8

Data points represent the mean (and SE) of five replications (ten callus
pieces per treatment in each replicate).






80


that the best combination for optimal growth was obtained with the addition of glucose

(1%) to sucrose (2%) (Table 3.8, Fig. 3.26). This sugar combination, the same used in medium "D" of Duncan et al. (1985), almost doubled the culture growth rate, when compared to the more commonly used plain sucrose (3%) in the medium.

Other than the variation in growth rate, no other noticeable differences were seen on the physical appearance of embryogenic calli grown in different sucrose containing combinations. No plausible reason is currently known for the apparent beneficial effect of the glucose-sucrose sugar combination.


Casamino acids. Substantial growth increases were reported for maize tissue cultures when casein hydrolysate levels were increased in the media (Duncan et al. 1985). The same promotional effect was also observed in diploperennial teosinte callus cultures when casein hydrolysate was used as a replacement for coconut milk (Swedlund and Locy 1988). An experiment was thus conducted to determine whether adding or omitting casamino acids (a vitamin-free product of casein hydrolysate) from the medium had any effect on embryogenic callus cultures of diploperennial teosinte. The results of this experiment are summarized in Table 3.9 and Figs. 3.27 and 3.28. All experiments were done using PIC2 medium, with or without casamino acids.

As usual, fresh weight varied linearly with time (Fig. 3.27, Table 3.9). Near regression of the data for the two growth curves confirmed a significant role for the presence of casamino acids in the medium: the addition of 500 mg/ casamino acids alone accounted for an almost doubling in the growth rate of the cultures (Table 3.9, Fig. 3.28). The convoluted surface of the embryogenic calli was also more prominent in the presence of casamino acids, which suggests that in addition to the quantitative effect on proliferation and growth, the product might also have an influence on the embryogenic process per se.







81



5000
CASAMINO ACIDS

o 0 mg/I
o 500 mg/I
4000



I

2 3000 LU




U 2000




1000 1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days)


Fig. 3.27 Embryogenic callus growth curves (fresh weight) with and without casamino acids.


Table 3.9 Regression analysis for growth with casamino acids.




Casamino Acids (mg/1) Linear Regression* R2



0 1126.3 + 87.0xa 0.999
500 832.3 + 159.0xb 0.998



Based on data points from five replications (ten callus pieces per treatment for each replicate).
Slopes marked with the same letter are not significantly different (P s 0.05).
Each medium contained:
MS basal salts and vitamins
12mM L-proline
0 or 500 mg/I casamino acids, as tested
20 g/I sucrose 10 g/I glucose
10,pM picloram
2 g/I Gelrite
pH 5.8






82



160

-o 140

O~ 120S8060
E ,.





CC 20
40- ....


0 500 CASAMINO ACIDS (mg/I)

Fig. 3.28 Effect of casamino acids on embryogenic callus growth in
diploperennial teosinte (see also Table 3.10).

Each medium contained:
MS basal salts and vitamins
12mM L-proline
0 or 500 mg/I casamino acids as tested
20 g/I sucrose 10 g/I glucose
10 pM picloram
2 g/I Gelrite
pH 5.8

Data points represent the mean (and SE) of five replications (ten callus
pieces per treatment in each replicate).






83


L-proline. Although not statistically supported, the greatest mean frequency of Type C embryogenic callus production occurred when 12 mM L-proline was added to the medium (Tables 3.4 and 3.5). Since observations suggested that had higher concentrations been used better results might have been obtained, an experiment was conducted to determine the effect of increasing levels of L-proline on the embryogenic callus cultures. The results of this experiment are summarized in Table 3.10 and Figs. 3.29 and 3.30. All experiments were done using PIC2 medium, with varying L-proline concentrations.

Again, regression analysis showed a significant linear relation between fresh weight and time for all concentrations (Fig. 3.29, Table 3.10). L-proline was not essential for callus growth, as moderate proliferation was obtained even when it was completely absent from the medium. Addition of the amino acid, however, produced a dramatic increase in the growth rate with increasing concentrations, with rates up to four times those measured for the control (Table 3.10, Fig. 3.30). Regression analysis of the slopes demonstrated a nonlinear increase in growth rate with increasing L-proline concentrations. The fitting curve was suggestive of a Michaelis-Menten equation, indicating that a factor (probably an enzyme or group of enzymes involved in membrane transport or cellular metabolism of L-proline) became limiting at higher concentrations, no further appreciable gain being then obtained by increasing the amino acid level in the medium (growth rates at 50 mM and 100 mM were not significantly different at the 95% confidence level). Comparable responses were reported in maize for the initiation of friable embryogenic callus (Armstrong and Green 1985) as well as fresh and dry weight increases and frequency of somatic embryo formation (V. Vasil and I.K. Vasil 1986) in the presence of proline.

Noteworthy was the fact that, in this experiment, faster growth was not correlated with a concurrent increase in embryogenic performance. Improved embryogenesis, subjectively judged by the level of convolution (proliferating scutella) on







84



6000
PROLINE

o 0 mM
5000- o 12 mM
A 25 mM
v 50 mM
E
4000- 0 100 mM I- 4000 I


I 3000


LL
2000



1000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days)


Fig. 3.29 Embryogenic callus growth curves (fresh weight) with various proline concentrations.


Table 3.10 Regression analysis for growth with proline.




Proline (mM) Linear Regression* R2



0 1147.4 + 54.0xa 0.997 12 845.2 + 147.9xb 0.998 25 923.7 + 158.9xc 0.997 50 750.7 + 186.3xd 0.997 100 640.5 + 194.0xd 0.989



Based on data points from five replications (ten callus pieces per treatment for each replicate).
Slopes marked with the same letter are not significantly different (P s 0.05).
Each medium contained:
MS basal salts and vitamins
500 mg/l casamino acids
20 g/l sucrose 10 g/I glucose
10 pM picloram
2 g/l Gelrite
Proline concentrations as tested
pH 5.8






85


200






C,
"O




: 50
I




0 20 40 60 80 100 PROLINE CONCENTRATION (mM)

Fig. 3.30 Effect of varying L-proline concentration on embryogenic callus growth
in diploperennial teosinte (see also Table 3.11).

Each medium contained:
MS basal salts and vitamins
500 mg/I casamino acids
20 g/l sucrose 10 g/I glucose
10 2M picloram
2 g/l Gelrite
Proline concentrations as tested
pH 5.8

Data points represent the mean (and SE) of five replications (ten callus
pieces per treatment in each replicate).






86


the callus surface and the visual appearance of the calli, both conditions that cannot be translated in numbers, was observed with increasing concentrations of L-proline up to 25 mM, at which the best compromise between growth and embryogenesis was attained. At 50 mM the callus turned whiter and became smoother and at 100 mM, regardless of the superior growth rate, all calli were no more than big growing white compact balls of tissue, without any surface features that would suggest somatic embryogenesis. When challenged for regeneration, these otherwise healthy calli completely failed to produce plants. The fact that callus growth and embryogenesis were greatest at different L-proline levels indicates that the two phenomena are under independent control, and suggests that conditions conducive to fast proliferation and growth are not necessarily good for the attainment of an optimal embryogenic response.

From this experiment a new medium (PIC3) arose, identical in composition to PIC2 but containing 25 mM L-proline. This medium promoted both excellent growth and embryogenesis in callus cultures of diploperennial teosinte. All subsequent work involving the maintenance of these cultures was thus performed with PIC3.


Plant Regeneration

Lowering the auxin concentration or simply ommiting it from the medium is the only basic requirement to induce regeneration from grass embryogenic callus cultures (I.K. Vasil 1987). As formerly indicated this was done upon completion of every experiment. Callus pieces were taken from the medium containing the best combination of factors and tested for regeneration, to assess if the component(s) that had optimized growth also had any effect on the expression of totipotency. This was done by placing pieces of embryogenic callus in a regeneration medium consisting of Murashige and Skoog's major and minor nutrients and vitamins with 100 mg/I myo-






87


inositol, 1 pM picloram and 2.5 pM kinetin, and visually evaluating the potential to produce plantlets in vitro.

This simple medium produced generally good results. However, because preliminary observations suggested that the presence of either auxins or cytokinins in the regeneration medium might not be needed at all, an additional experiment was conducted to determine the effect, if any, of these two kinds of substances.

Plants were regenerated with comparable ease from all auxin/cytokinin combinations, even their total absence. Over a period of 3-4 weeks the cup-shaped structures on the callus surface increased in size and turned bright white, due to starch accumulation. Small (3-5 mm) callus pieces often turned into single deformed scutella from which a coleoptile soon developed. At this stage normal germination occurred, with the first green leaves growing through the coleoptile tip and the radicle protruding from a well developed coleorhiza. Various numbers (generally 1-10) of somatic embryos were produced from larger callus pieces.

Presence of auxin was not needed for regeneration, and it did not influence the average number of plantlets (normally 5-15) that could be obtained per dish (5 callus pieces/dish). It did, however, delay somatic embryo germination by approximately a week. Plantlets formed in the presence of auxin usually had fewer deformities (e.g., leaf curling) than plantlets regenerated without it, a fact of no consequence as all plants soon grew out of epigenetic anomalies.

The average number of plantlets formed per dish was also not affected by the type of cytokinin employed. In the presence of BA (N6-benzyladenine), however, the white scutella tended to become green and resembled leafy structures, often with trichomes. This greening of the scutellum, which was particularly evident in the absence of auxin, had no obvious effect on the morphology of the plantlets so derived. Anthocyanin pigmentation was present in all treatments as scattered dots on the scutellar structures or as a deep red color on the mature coleoptiles.






88


Embryogenic callus cultures kept in optimized media were still regenerable after a period of four years in vitro and, although not challenged for regeneration at that time, were still growing vigorously six years after being placed in culture. As cultures aged, the requirement for auxin did not change, but the need for a cytokinin in the medium became more pronounced. BA was the most effective for that purpose. At levels that varied from 2.5 pM (initial concentration for all cytokinins tested) up to 5.0 pM (needed in 4 year old cultures), addition of N6-benzyladenine to the regeneration medium consistently promoted formation of plantlets from callus cultures, even when other cytokinins no longer had an effect at comparable concentrations.

Discussion and Conclusions



Donor tissue. One of the most important success factors in the initiation of regenerable cultures in Gramineae is the appropriate choice of the original explant. The use of immature tissues which still maintain meristematic activity and competence has been essential to the development of cell and tissue culture systems in this family of plants.

Immature embryos and young inflorescences and leaves are the most commonly used sources of explants for initiation of embryogenic cultures in the grass family (I.K. Vasil 1987). Under the cultural conditions of this study, however, immature leaves and inflorescences of diploperennial teosinte, although at the deemed appropriate stages of development, failed to initiate any embryogenic callus growth, indeed almost any kind of callus growth at all.

Production of embryogenic callus seems to be confined to particular stages of development, even when meristematic tissues are concerned. The existence of these temporal and developmental "competence windows" has been demonstrated in many cases and is particularly well documented in Gramineae. For example, immature leaf






89


explants containing or being close to the apical meristem, as well as older, fully differentiated tissues, are generally unable to proliferate and give rise to embryogenic calli. This gradual decline in embryogenic competence, that is not necessarily related to the loss of mitotic activity (Joarder et al. 1986; Taylor and I.K. Vasil 1987), has been demonstrated in a number of cereals and other grasses (Wernicke and Brettell 1980; Haydu and I.K. Vasil 1981; Lu and I.K. Vasil 1981; Wernicke et al. 1981; Hanning and Conger 1982; Alfinetta et al. 1983; Ho and I.K. Vasil 1983; Wernicke and Milkovits 1984; Joarder et al. 1986; Linacero and Vazquez 1986a; Wenzler and Meins 1986; Rajasekaran et al. 1987b; Pareddy and Petolino 1990; Songstad et al. 1992), including diploperennial teosinte (Swedlund and Locy 1988).

From these and other studies it became clear that older tissues lose their capacity to respond in culture as they differentiate and lose their meristematic potential. However, sections containing or being close to the apical meristem will also not proliferate, a condition somewhat more difficult to explain. As in competent explants embryogenic calli are normally formed in association with developing vascular tissues (I.K. Vasil 1985), it has been suggested that the transport of growth regulators through the vascular system could be instrumental in achieving the delicate hormonal balance that controls tissue competence and, therefore, embryogenesis. Since apical areas lack a differentiated vascular system, endogenous regulators are transported mostly by cell to cell diffusion and are probably partially metabolized and never reach the adequate balance needed for the development of competence in culture (I.K. Vasil 1987). In the present study, in the few cases where any proliferation was obtained in immature inflorescences and young leaves, that proliferation was indeed topologically related to the location of the sectioned parallel veins. No major growth of any kind, however, was obtained in any of the treatments, regardless of the relative position of the explant in relation to the shoot apical meristem.






90


Accumulating evidence indicates that, assuming the developmental status of the explant is adequate, it is possible to obtain callus and regeneration from any plants deemed recalcitrant by basically varying the chemical and physical environment of the explant until the right combination of factors is achieved (Morrish et al. 1987). Relatively simple changes in media nutrient composition and hormone levels, for example, have led to spectacular increases in response in some otherwise recalcitrant genotypes (Lu et al. 1983; Duncan et al. 1985; Hanzel et al. 1985). The environmental conditions under which donor plants are grown, however, are also critical in determining the level of competence of the explant, no matter how adequate it might intrinsically be (V. Vasil and I.K. Vasil 1982; Morrish et al. 1987). This is particularly true for greenhouse grown heliophytes, or field grown crops that are subject to some kind of cultural stress during the growing season.

The natural conditions during growth (e.g., temperature, rainfall and photoperiod), as well as the cultural conditions (e.g., irrigation or the application of pesticides or fertilizers), are crucial in determining the physiological condition of the donor plant and, therefore, the response of the prospective explants in culture (Morrish et al. 1987). Lu et al. (1983), for example, showed that embryos obtained from the same cultivar but grown under different environmental conditions produced different responses in culture. They reported that ". . .there was more variability in the response of embryos from the same cultivar grown at different times than amongst embryos obtained from different cultivars of a single planting ... ." (Lu et al. 1983). Such variation may be extreme to the point of the explants not responding at all.

Failure to meet in full some physiological requirement of the explants, or of the original donor plants, would explain the setback in initiating embryogenic cultures from immature inflorescences and young leaf explants in this study, even when preexisting successful protocols developed for the same species (Swedlund and Locy 1988) were






91


followed. That further work is needed, however, is also indicated by the fact that these authors, too, reported only a meager (5%) percentage of success from such explants.

To minimize the unavoidable variation in cultural conditions to which field grown plants are exposed, the use of the controlled environment of the greenhouse has been recommended (Lu et al. 1983; Armstrong and Green 1985). It has also been noted, however, that greenhouse conditions may be less than ideal (Lu et al. 1983). This was obvious in the present study, where immature embryos from plants grown in the open yielded a much better response in culture (as judged by qualitative visual observation) than embryos from greenhouse grown plants.


Embryo developmental stage. Many factors affect the frequency of formation of embryogenic calli from immature embryos of diploperennial teosinte. The developmental stage of the embryo and the presence of a strong auxin in the medium were the most critical for success.

Diploid perennial teosinte embryos at different stages of development had a distinct sequential pattern in the capacity to produce embryogenic callus. Very young embryos (s 1 mm) showed no response when cultured, whereas embryos with a differentiated scutellum that was already opaque and starting to accumulate starch (1-2 mm) produced the best embryogenic response. This gradually tapered off in larger embryos (2-4 mm), and completely disappeared in fully mature ones (2 4 mm).

The differences found in the embryogenic response of immature embryos of diploperennial teosinte were identical to those recorded for other members of Gramineae (Green and Phillips 1975; Beckert 1982; V. Vasil and I.K. Vasil 1982; Lu et al. 1983; Maddock et al. 1983; Armstrong and Green 1985; Morrish et al. 1987). Again they indicate and confirm the existence of a competence window in developing embryos, comparable in all respects to that formerly discussed for young leaf explants. How the






92


expression of competence is regulated is not known, but it certainly must involve steps of gene expression that are under developmental control.


Auxin type and concentration. As more is learned about the factors affecting somatic embryogenesis, it becomes apparent that the type, concentration and time of application of plant growth regulators in the medium are all particularly critical to longterm success. In dicots, where exogenous auxin levels readily reach toxicity, determination of the lowest effective concentrations and shortest exposure times to the regulator has been a continuous concern (Ammirato 1989).

Indoleacetic acid, the endogenous auxin, is mostly inefficient in the initiation of embryogenic callus cultures, probably because it is rapidly metabolized or destroyed by enzymatic oxidation in vivo (Good et al. 1982). Amide and ester conjugates of the native auxin have been identified in a variety of plant tissues. They seem to be involved in the protection of IAA from destruction by peroxidases (Andreae and Good 1955; Cohen and Bandurski 1978) and the detoxification of tissues in the presence of excessive IAA (Andreae and Good 1955; Good et al. 1956). The fact that both kinds of conjugates can be easily hydrolyzed under physiological conditions to release free IAA in vivo (Hangarter and Good 1981) suggested their potential use as slow-release auxin sources for tissue cultures in vitro (Feung et al. 1977; Hangarter et al. 1980; Good et al. 1982; Pence and Caruso 1984). Prospective advantages of using such compounds included a continuous supply of near physiological levels of auxin, which would promote cell division (callus formation) without impairing future regeneration capabilities (Good et al. 1982), a problem often found with the more stable and more powerful synthetic analogs.

IAA alone as an auxin source, however, does not elicit embryogenic response in most monocots, particularly in Gramineae (Rajasekaran et al. 1987a). This would explain why explants were not responsive to the IAA-aminoacid conjugates, even when






93


considering that they are more stable sources of auxin than the labile free IAA or the persistent synthetic IAA analogs (Hangarter et al. 1980). The same lack of responsiveness to IAA-conjugates was found by Rajasekaran et al. (1987a) in cultured leaf explants of Napier grass.

Tissue cultures of Gramineae are unique in that a strong auxin is often the only growth regulator needed to elicit embryogenic response (I.K. Vasil 1987). Other than the native auxin, indoleacetic acid, most of the weaker synthetic analogs such as NAA, that yield better results than 2,4-D in dicot tissue cultures (e.g., Lazzeri et al. 1987), are also ineffective when used alone to initiate an embryogenic response in members of the grass family. In the present experiments, only nonembryogenic callus was produced by competent immature embryos in the presence of NAA and only when concentrations were high. This agrees with the observation that whenever NAA is used in tissue cultures of Gramineae, it most commonly is in combination with some of the stronger auxins, such as 2,4-D (e.g., Haydu and I.K. Vasil 1981).

The most commonly used compound for triggering embryogenesis in the majority of plants, and in particular Gramineae, is 2,4-D. A powerful selective herbicide with strong auxin activity, well tolerated by most monocots, 2,4-D has been associated, however, with toxicity and mutagenicity (Karp and Bright 1985; Karp 1989) and it is thought to be associated with the loss of regeneration capacity in long-term cultures (e.g., Smith and Street 1974; Saunders and Bingham 1975). Although these claims are probably unfounded, as attested by the wide variety of Gramineae where it has been, and it still is, successfully used to initiate and maintain stable regenerable tissue cultures (I.K. Vasil 1987), 2,4-D has been advocated as a possible generator of instability, and it has been indicated as good practice to limit its use in plant tissue cultures (Karp 1989) whenever alternatives can be found.

Reports have shown that other herbicides with auxin-like properties, such as dicamba or picloram, can be used as successful alternatives to 2,4-D in establishing




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EW9BUUIJA_INP02G INGEST_TIME 2015-04-13T19:34:22Z PACKAGE AA00030033_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

SOMATIC EMBRYOGENESIS, LONG-TERM REGENERATION AND EPIGENETIC VARIATION IN TISSUE CULTURE REGENERANTS OF DIPLOPERENNIAL TEOSINTE (ZEA DIPLOPERENNIS ILTIS, DOEBLEY & GUZMAN) By LUIS FILIPE PEDROSA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1993

PAGE 2

This dissertation is dedicated to my mother, Maria Luisa, who did not live to see its completion, and to my son, David, joy of my life.

PAGE 3

ACKNOWLEDGMENTS I would like to express my thanks to Dr. Indra K. Vasil, for providing valuable guidance and the logistics necessary to bring this piece of research to a good end. My appreciation also goes to Dr. H.H. litis (University of Wisconsin), for promptly providing the diploperennial teosinte seeds needed to produce the populations used in this study. It is also with gratitude that I acknowledge the members of my graduate committee, Dr. Thomas J. Sheehan, Dr. Norris H. Williams, Dr. Henry C. Aldrich and Dr. William Louis Stern, as well as Dr. Vimla Vasil and Dr. Walter S. Judd for all their helpful suggestions and continued support. Special thanks are particularly due to Dr. Fiona Redway and Mr. Mark Taylor, for keeping me going in times of distress, and to Mr. Bart Schutzman, for sharing many fascinating incursions into the worlds of plant science, statistics, computers and photography. Last, but not the least, my gratitude and thanks go to my loving son, David, for having the special endurance it takes to have a doctoral student as a single parent. iii

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS Hi ABSTRACT vi CHAPTERS 1 INTRODUCTION 1 2 LITERATURE REVIEW 6 3 SOMATIC EMBRYOGENESIS AND LONG-TERM REGENERATION IN CALLUS CULTURES OF DIPLOPERENNIAL TEOSINTE 29 Introduction 29 Materials and Methods 34 Plant Material 34 Induction of Somatic Embryogenesis 36 Callus Maintenance and Multiplication 37 Data Collection and Analysis 38 Plant Regeneration 39 Results 40 Induction of Somatic Embryogenesis 40 Callus Maintenance 71 Plant Regeneration 86 Discussion and Conclusions 88 4 PHENOTYPIC VARIATION IN TISSUE-CULTURE-DERIVED PLANTS OF DIPLOPERENNIAL TEOSINTE 101 Introduction 101 Materials and Methods 1 01 Morphometric Analyses 1 01 Cytological Analyses 1 06 Results 108 Univariate Comparisons 1 08 Multivariate Comparisons 1 27 The Effect of Gibberellic Acid 203 Discussion and Conclusions 236 iv

PAGE 5

5 ELECTROPHORETIC ANALYSIS OF TISSUE-CULTURE-DERIVED PLANTS OF DIPLOPERENNIAL TEOSINTE 257 Introduction 257 Materials and Methods 258 Plant Material 258 Preparation of Starch Gels 258 Protein Extraction and Gel Loading 259 Electrophoretic Buffers and Staining Systems 261 Electrophoresis 261 Enzyme Activity Staining 264 Analysis of Banding Patterns 265 Results 265 Selection of Electrophoretic Buffers and Enzyme Systems for Studying Isozymes in Diploperennial Teosinte 265 Isozyme Analysis of Tissue Culture Regenerants 268 Discussion and Conclusions 268 6 CLOSING REMARKS 297 REFERENCES 306 BIOGRAPHICAL SKETCH 334 v

PAGE 6

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SOMATIC EMBRYOGENESIS, LONG-TERM REGENERATION AND EPIGENETIC VARIATION IN TISSUE CULTURE REGENERANTS OF DIPLOPERENNIAL TEOSINTE (ZEA DIPLOPERENNIS ILTIS, DOEBLEY & GUZMAN) By Luis Filipe Pedrosa December 1 993 Chairman: Indra K. Vasil Major Department: Botany High frequency somatic embryogenesis and long-term maintenance of embryogenic calli were obtained from scutellar tissue of diploperennial teosinte immature zygotic embryos. The developmental stage of the embryo and the inclusion of a strong auxin in the initiation medium were the only critical factors for success. Embryogenic calli were comparable to those formerly described for the majority of the Gramineae. An unusual callus type, which was never associated with soft, friable, non-embryogenic callus, was isolated and used to establish all the regenerable cultures used in this study. Manipulation of the carbon sources in the medium, as well as addition of casamino acids and L-proline, enhanced the embryogenic response and growth of the cultures that, under optimal conditions, maintained regeneration potential for over four years. Morphologically normal plants were routinely recovered from the callus cultures for over two years. Regeneration of aberrant phenotypes, however, gradually became more frequent in older cultures. At the end of three years all plants from tissue culture were distinctly dwarf, multitillering and male sterile, displaying incomplete development and/or feminization of the tassel. All had the normal diploid chromosome complement. vi

PAGE 7

The fate of the induced morphological variation was studied for the next two years, using 100 plants each from the R 0 and the R v One hundred plants from seed were used as a control. Combined univariate and multivariate statistical analysis of 26 morphometric traits, covering both vegetative and floral characters from the tissue culture regenerants, their sexual progeny and control plants, revealed that all anomalous traits in the R 0 were transient in nature and faded with time, although some persisted for a period of over 1 8 months. Gibberellin applications completely replaced the effect of time, and led to complete phenotypic recovery to normal of the Ro plants. Complete recovery also resulted from a passage through a single cycle of sexual reproduction, as there were no anomalies in the Ri plants. Isozyme analysis did not show a link between the morphological variation and the corresponding biochemical variation that, of necessity, must precede or accompany it. Regardless of their phenotypic differences, all tissue culture regenerants showed the same isozyme banding patterns, which were the same as those from the controls. vii

PAGE 8

CHAPTER 1 INTRODUCTION The occurrence of phenotypic variation in plants regenerated from tissue cultures is a well documented, yet little understood phenomenon (Bayliss 1 980; Larkin and Scowcroft 1981; D'Amato 1985; Karp and Bright 1985; Lorz and Brown 1986; Lorz era/. 1988). Phenotypic variation is frequently associated with regeneration systems involving an unorganized callus phase and it may affect quantitative as well as qualitative traits. Deviations from the expected phenotypic and genetic uniformity have been reported at the morphological, karyological, physiological, biochemical and molecular levels (Meins 1983; Orton 1984; D'Amato 1985; Scowcroft 1985; Ahloowalia 1986; Gould 1986; Semal 1986; Lorz era/. 1988). The theoretically unexpected but apparently widespread generation of variability during tissue culture, recently termed "somaclonal variation" (Larkin and Scowcroft 1981), has been proposed as a potential source of unique and interesting new traits for use in crop improvement (Larkin and Scowcroft 1981; Larkin 1985). Since the introduction of this concept in 1981, much controversy has arisen concerning the real prospects of somaclonal variation as a contribution to ongoing or new plant breeding programs (Karp and Bright 1985; Ahloowalia 1986; Lorz and Brown 1986; Semal 1986; I.K. Vasil and V. Vasil 1986; Morrish er al. 1987; I.K. Vasil 1987, 1988). Part of this controversy is based upon the fact that, although variation may be present in the regenerants, it is frequently transient in nature or is lost during sexual reproduction. Moreover, limited field evaluations with diverse genotypes of putative somaclones have shown that most somaclonal variation is either useless or has a limited applicability in direct varietal upgrading (Ahloowalia and Sherington 1985; Ryan era/. 1987; Baillie etal. 1

PAGE 9

2 1992; Qureshi et al. 1992). Whenever heritable variation is present, the type and frequency of variants suggest that somaclonal variation is akin to non-directed random mutagenesis, which generates a large amount of unwanted variation often expressed in the form of traits that either are not novel or are agronomically unfavorable (Semal 1 986; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K Vasil 1987, 1988). It is noteworthy that, despite the rhetoric and exaggeration surrounding the issue for almost a decade, no commercial varieties of any major crop plant have yet been produced or are grown on a large scale as a result of using tissue-culture-generated variation. As time gradually wore most of the original glamor out of the issue of usefulness of the newly expressed traits, questions such as their persistence over time and across generations, and the quest for the genetic or physiological mechanisms underlying their origin, slowly took its place in the minds of the scientific community. There is no doubt that some of the culture-induced variation has a genetic basis and is heritable. It has often been argued, however, that much of the observed variability is only transient and epigenetic, physiological or developmental in nature and therefore nonheritable and useless in breeding (Karp and Bright 1985; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K. Vasil 1987, 1988). When this is the case the plants showing variation can no longer be thought of as potential assets to breeding; they can, however, still be usefully incorporated in programs where genetic fidelity, rather than change, is the goal, on the condition that the detected variation can be proven to be no more than a transient culture-induced "carry-over" effect. Tissue-culture-generated variation has been correlated with a variety of factors. These include, among others, the original type and age of the explant, genotype of the donor plant, duration of culture and regeneration pathway. In Gramineae, where two possible regeneration pathways (embryogenic and organogenic) have been recognized and described in detail, most of the reported occurrences of variation in tissue-culturederived populations have been related to regeneration from organogenic callus cultures

PAGE 10

3 (Morrish et al. 1987). Accumulating evidence seems to indicate that more stringent selection in favor of normal cells, believed to be associated with the embryogenic pathway of regeneration, tends to prevent the production of somaclonal variants (I.K. Vasil and V. Vasil 1986; Kobayashi 1987; Morrish et al. 1987; I.K. Vasil 1987, 1988; Dolezel and Binarova 1989; Gmitter et al. 1991). The selective advantage of cytologically or genetically normal cells (diplontic selection), expressed at the very onset of the embryogenic response in callus cultures, results in the production of embryos that consist wholly of normal euploid cells (e.g., Cavallini et al. 1 987; Gmitter ef al. 1991). Selection pressure can, however, be critical even at later stages, as evidenced by the fact that developing somatic embryos with cytological or genetic aberrations, or chimerism, often fail to differentiate into adult plants (Cavallini ef al. 1 987; Cavallini and Natali 1989). The apparent intolerance, during embryogenesis, to alterations of the normal genome thus effectively prevents (or greatly limits) the transmission of anomalies to populations produced by this regeneration pathway. Although karyological and phenotypic stability can often be correlated with the embryogenic pathway, varying frequencies of abnormal regenerants from embryogenic tissue cultures have also been reported in the literature. These may have resulted from cultures where both the embryogenic and organogenic pathways coexist, as has been described in wheat (Karp and Maddock 1984), in which case the variation reported may result from plants being regenerated from the mixed callus via the organogenic pathway (Maddock 1985). Variants may also be transient in nature (Larkin and Scowcroft 1983a,b; Irvine 1984), the result of epigenetic rather than genetic changes in the regenerated plants. There is still much to be learned about the nature and origin of the variation found in tissue-culture-derived plants. The genetic (or epigenetic) nature of the variation has not always been properly studied or documented in the past. In many cases the regenerated plants are morphologically aberrant to the point of being unable

PAGE 11

to complete a life cycle, or have their sexual expression at maturity modified or supressed, which prevents the progeny testing that constitutes the ultimate proof of genetic (heritable) variation (Meins 1983; Karp and Bright 1985). In such cases, alternative ways of distinguishing between genetic and epigenetic changes have been proposed (Meins 1983) and should be utilized whenever possible. Evidence of epigenetic rather than heritable variation in cases involving the embryogenic pathway would add additional support to the idea that somatic embryogenesis produces a lower overall level of genetic variation and that it should be the regeneration method of choice whenever maintenance of genetic fidelity, rather than variation, is the primary goal of the investigation. Further work is needed to determine the level of stability attained through somatic embryogenesis in tissue cultures. The methods to employ in such studies have been outlined in the literature (Morrish etal. 1987). They include detailed analysis of the somatic variation present in the regenerants, as compared and contrasted to appropriate controls, as well as its heritability. In addition to the traditional chromosome cytology, such studies must focus on both quantitative and qualitative genetic traits and, whenever possible, biochemical and/or molecular analyses should be used to further assess possible alterations of the genome that may not be expressed as variant morphological traits. Regeneration must be strictly monitored, particularly in the species for which both possible regeneration pathways have been reported. Because the transient nature of some epigenetic effects may be obscured by their long-term expression, the use of a perennial (for which time is not a limiting factor) rather than an annual species is recommended and the barriers of sexual sterility or incompatibility must not be present, to permit a thorough analysis of the sexual progeny. This dissertation reports the results of one such study in a perennial relative of maize, diploperennial teosinte (Zea diploperennis litis, Doebley & Guzman). The objectives of this research were (1) to develop a long-term efficient regeneration system

PAGE 12

from cultures of diploperennial teosinte; (2) to characterize the morphology of the regeneration pathway; (3) to evaluate the genetic fidelity of the tissue-culture-derived plants through morphometric and cytological characterization of the regenerants, their progeny and control plants; (4) to evaluate biochemically the genetic fidelity of the tissue-culture-derived plants through electrophoretic characterization of the regenerants; and (5) to interpret the nature of variation in regenerants and/or their progeny, should any be detected.

PAGE 13

* CHAPTER 2 LITERATURE REVIEW Introduction Early in the sixties Morel first performed his classic experiments with the proliferation in vitro of orchid shoot apical meristems (Morel 1960; 1963; 1964a,b; 1965a,b). His discoveries soon led to the commercial mass production of these rare and beautiful plants and, by promoting the development of what is now known as the aseptic method of micropropagation, initiated a major revolution in the horticultural world. In the years that followed, extensive research expanded the realm of micropropagation to a wide range of taxa, now covering virtually the complete spectrum of vascular plants. The production of tissue-culture-derived clones has been achieved in a very large number of species, including ornamentals, vegetable crops, fruit crops, trees and agronomic crops (Conger 1981; I.K. Vasil 1986). Based on the principle that regeneration from cultured somatic cells faithfully reproduces the original donor plant in potentially unlimited numbers, the system permits the large scale reproduction of elite germplasm, providing the grower (and the consumer) with superior products at a fraction of the cost and time needed by conventional means. It was soon discovered, however, that not all cultures were suitable for cloning. When plants were regenerated from adventitious buds phenotypic and genetic changes were not uncommon, especially when a stage of unorganized cell proliferation (callus) was involved. The consistent uniformity and trueness-to-type needed in clonal propagation could still be achieved, as long as the cultures were based on induced 6

PAGE 14

7 axillary branching from preexisting shoot tips and lateral bud explants, an approach that completely bypasses the presence of the undifferentiated callus phase. Observations made on such populations of meristem-propagated plants have shown that the frequency of variants obtained by this method is not higher than that observed when using traditional methods of clonal propagation (Conger 1981; Beauchesne 1982). Occasional sports are attributed to chance mutations, that occur at very low frequencies. This phenotypic uniformity from shoot tip cultures has been attributed to the higher level of genetic stability normally associated with shoot meristems, which act as a safe "storage" place for the normal genetic information of the plant (D'Amato 1985; I.K. Vasil 1987). Long recognized as an unavoidable nuisance, the characteristic variation found in callus cultures and their regenerates was brought to the limelight in 1 981 when Larkin and Scowcroft proposed that such variation, which they termed "somaclonal", was an inherent result of the plant cell culture itself and that it showed great promise as a significant source of novel traits potentially useful for plant improvement (Larkin and Scowcroft 1981). Under the newly coined name of "somaclonal variation" the concept was widely adopted throughout the eighties, when a large number of publications covered in varying detail a wide range of aberrations associated with tissue cultures, many of them known but never publicized before. Among others, they included numerical and gross structural chromosome changes in the cultures and their regenerates (D'Amato 1985), cryptic alterations in the nuclear, chloroplast and mitochondrial genomes (Hanson 1984; Lorz and Brown 1986), breakdown of chimeric structures (Cassells 1985; Preil 1986) and changes due to the elimination of infectious agents (Cassells 1985; Larkin and Scowcroft 1981). The popularity of the idea is reflected in the numerous reviews that over the last several years tried to assess the nature of somaclonal variation and its potential contribution to plant breeding (Larkin and Scowcroft 1983b; Orton 1984; Mitra 1985;

PAGE 15

8 Evans and Sharp 1 986, 1 988; Dewald and Moore 1 987; Larkin 1 987; Sala and Biasini 1987; Larkin et al. 1989). Although specific examples of applications have been reported (e.g., Evans 1989) there is, however, a growing opinion that somaclonal variation has not fulfilled its initial promise as a potential source of economically valuable germplasm. A critical analysis of the success so far achieved, now that one decade has passed, reveals that despite massive efforts by scores of dedicated researchers worldwide, no significant commercial varieties of any of the major crop species have yet been developed as a result of incorporating tissue-culture-generated genetic variation into a breeding program (I.K. Vasil 1990). Much of the variation that has been found in cultures or their regenerates is either not novel or has doubtful usefulness in breeding (I.K. Vasil and V. Vasil 1986; Morrish er al. 1987; I.K. Vasil 1987, 1988). It may even be deleterious in nature. In addition, long-term epigenetic changes, which are not unusual in tissue culture regenerants, are often mistaken for true genetic variation (Karp and Bright 1985; I.K. Vasil and V. Vasil 1986; Morrish er al. 1987; I.K. Vasil 1987, 1988; Karp 1989). Epigenetic changes are, in fact, the result of directed physiological alterations on the expression of the genome. They have been shown to be potentially reversible and, by definition, nonheritable (Meins 1983), and are therefore useless in breeding. The distinction between genetic and epigenetic variation may be even further complicated by the fact that certain changes in the genetic material, such as DNA methylation (L6rz and Brown 1986) and amplification (Cullis 1986), may be heritable under certain conditions, but may also become reversible (Phillips etal. 1990). In view of the failure to produce the promised boon of new traits for breeding, somaclonal variation is again regarded today the way it had been from the very beginning-a seemingly inevitable nuisance to most current fields of applied research that make use of tissue cultures. Its occurrence is clearly detrimental in clonal micropropagation (Beauchesne 1982), including the use of somatic embryogenesis and

PAGE 16

9 synthetic seed technologies as potential means of automated mass propagation (Redenbaugh er a/. 1987; Ammirato 1989). It is also a particular concern in the newly emerging discipline of genetic transformation through gene transfer into cultured cells, a technique that seeks to introduce selected genes into elite lines without the cointroduction of unwanted culture-induced variability (Lorz and Brown 1986; Goodman etal. 1987; Lazzeri and Lorz 1988; de Klerk 1990). It would be highly desirable to be able to control the genetic instability associated with tissue cultures. No overall consensus exists, however, on general applicable guidelines for such control (Karp 1989). There is a lack of information on the underlying genetic and physiological mechanisms that are the basis of variation and the interaction of the factors affecting them. The resolution of this problem, currently under intense investigation, should be a greater contribution to plant science than the proposed direct exploitation of the phenotypic variation itself (Karp and Bright 1985). This chapter briefly summarizes the nature of the phenotypic characteristics normally associated with somaclonal variation in Gramineae and presents the current understanding on how such variation relates to a range of genotypic or epigenetic changes as they affect the cultured cells in vitro. The Nature of Tissue-culture-induced Variation Tissue-culture-induced variation is a general phenomenon of widespread occurrence (Bayliss 1980; Larkin and Scowcroft 1981; D'Amato 1985; Karp and Bright 1985; Lorz and Brown 1986; Lorz etal. 1988). It seemingly affects most plant species, regardless of taxonomic position or ploidy level, including our most important crops. It has been described in seed crops such as tobacco, tomato, and alfalfa, in vegetatively propagated crops such as potato, in cereals and other grasses and in ornamental plants, such as the garden geranium and petunia (Larkin and Scowcroft 1981; Karp and Bright 1985; Lorz and Brown 1986). As previously mentioned, tissue-

PAGE 17

10 culture-induced variation is directly related to regeneration systems in vitro that involve the passage through an unorganized callus phase (Larkin and Scowcroft 1 981 ; Karp and Bright 1985). It is normally absent in cultures derived from proliferation of shoot apical or axillary meristems (D'Amato 1985; Karp and Bright 1985). Somaclonal variation can be observed at the morphological, karyological, physiological, biochemical and molecular levels (Meins 1983; Orton 1984; D'Amato 1985; Gould 1986; Ldrz et al. 1988). It may affect quantitative as well as qualitative traits. The phenotypic changes so produced, sometimes at an unusually high frequency, may occur in individual plants or affect whole populations and often concern agronomically important traits. Examples of tissue-culture-induced variation in Gramineae are presented in Table 2.1. Variant traits are normally found in the heterozygous form, but may also occur as homozygous as well, an intriguing phenomenon quite characteristic of this kind of variation and for which at present there still is no plausible explanation (Karp and Bright 1985; Karp 1989). Not all variation found in somaclonal variants is genetic and stable. In the vast literature dealing with the subject there are many examples of transient or temporary phenotypic variation. These may vary from what has been loosely termed a "carry-over" effect, like the typical shoot proliferation/root inhibition associated with the use of cytokinins in the regeneration system, to long-term effects that may be easily confused with stable genetic variation, if not for the fact that they are not always found in the following generation after selfing. This variation, termed epigenetic (Meins 1983), may be expressed at the morphological level, frequently as alterations in size or growth habit, as well as at the physiological level, for example, as habituation or vitrification. It has been correlated to the stress of growth under in vitro conditions and can often be duplicated by growing plants from seed under conditions that mimic the tissue culture environment. For example, when maize seedlings were grown from normal non-tissueculture-derived seeds maintained in closed containers with synthetic media, as would

PAGE 18

11 Table 2.1 Examples of tissue-culture-induced variation in Gramineae Species (Common Name) Type of Variation References Avena sativa (Oat) Bothriochloa spp. (Old World Bluestems) Cymbopogon winterianus (Citronella Grass) Festuca arundinacea (Tall Fescue) Hordeum vulgare (Barley) Lolium multiflorum (Rye Grass) karyological reduced plant height heading date seed protein flag leaf area seed weight seed number yield bundle weight sterility karyological reduced plant height dwarfism foliage colour suppression of flowering inflorescence length reduced fertility isozyme banding patterns herbage yield oil content Cummings et al. 1 976 McCoy era/. 1982 Dahleen 1 989 Taliaferro er al. 1 989 Mathuref al. 1988 isozyme banding patterns Dahleen and Eizenga 1 990 polyploidy karyological plant height stem thickness extra flag leaves supernumerary spikes chloroplast DNA rDNA spacer fragments hordein proteins hordein proteins plant height leaf length leaf width spike shape fertility Orton 1 980 Ahloowalia 1 986 Day and Ellis 1 985 Breiman et al. 1 987b Karp era/. 1987 Ahloowalia 1 986

PAGE 19

12 Table 2.1 (Continued) Species (Common Name) Type of Variation References Oryza sativa (Rice) Pennisetum americanum (Pearl Millet) Saccharum officinarum (Sugarcane) Secale cereale (Rye) Triticum aestivum (Bread Wheat) reduced height reduced fertility heading date soluble protein mRNA levels plant height dwarfism tiller number heading date grain per pannicle grain weight chlorophyll mutations reduced height improved yield improved protein dwarfism nuclear DNA karyological reversion to fertility isozyme banding patterns auricle length sheath colour pubescence erectness cane diameter stalk length stalk weight disease resistance cane and sugar yield disease resistance disease resistance disease resistance secalin proteins reduced height stem thickness reduced fertility heading date chloroplast DNA Oono 1978 Sun etal. 1979 Schaeffer 1982 Sun etal. 1983 Schaeffer ef al. 1 984 Oono 1 985 Zheng ef al. 1 987 Swedlund and Vasil 1 985 Smith ef al. 1987 Heinz and Mee 1 971 Liu era/. 1972 Lat and Lantin 1 976 Liu and Chen 1 976 Heinz et al. 1977 Larkin and Scowcroft 1 983a Krishnamurthi 1 982 Bebeli ef al. 1 990 Ahloowalia 1 982 Maddock ef al. 1 983 Sears ef al. 1 984 Ahloowalia and Sherington 1 985 Day and Ellis 1984

PAGE 20

13 Table 2.1 (Continued) Species (Common Name) Type of Variation References Triticum aestivum (Bread Wheat) (cont.) x Triticosecale (Triticale) Zea mays (Maize) karyological plant height tiller number heading date grain and glume colour gliadin proteins gliadin proteins isozyme banding patterns cold tolerance cold tolerance karyological karyological kernel protein prolamin banding pattern spike length fertility rDNA spacer sequences leaf shape ear position pollen sterility karyological resistance to Helminthosporium male sterility mtDNA restriction pattern lysine and threonine content karyological reduced height reversion to male fertility resistance to Drechslera maydis increased vigor reduced height altered fertility altered maturity altered plant and ear height altered leaf color increased tillering increased ear number ADH1 gene mutation reduced plant height Karp and Maddock 1 984 Larkin et al. 1 984 Cooper et al. 1 986 Davies et al. 1 986 Lazar et al. 1 988 Galiba and Sutka 1989 Armstrong er al. 1 983 Lapitan ef al. 1 984 Jordan and Larter 1 985 Brettell etal. 1986b Green 1 977 Edallo era/. 1981 Gengenbach et al. 1 981 Hibberd and Green 1 982 McCoy and Phillips 1982 Beckert ef al. 1 983 Umbeck and Gengenbach 1 983 Umbeck and Gengenbach 1983 Hubbard ef al. 1 984 Earle and Gracen 1 985 Brettelef al. 1986a Gobel etal. 1986

PAGE 21

14 Table 2.1 (Continued) Species Type of References (Common Name) Variation Zea mays altered leaf morphology Gobel et al. 1 986 (Maize) (cont.) feminization of male flowers suppression of anthers reduced ears poor germination (Ri) early seedling death (flj) silk emergence Zehr et al. 1 987 reduced plant height kernel characteristics pollen shed

PAGE 22

15 occur during tissue culture, the resulting plants showed signs of variation identical to those commonly described in somaclonal variants, including reduced fertility, dwarfing, feminizaton of male flowers, etc. (Lorz er a/. 1988). The underlying reasons for this unnatural behavior of plants derived from in vitro cultures are not currently understood but, because the transient nature of some epigenetic effects may be obscured by their long-term expression, progeny tests using the appropriate crosses for the variant traits must be examined to accurately determine the extent of epigenetic variation (Meins 1983; Evans and Sharp 1988). Unique to somaclonal variation is the peculiar combination of homozygous changes at high frequency and the presence of whole population shifts for certain traits. For example, when comparing the types of mutants present in the progeny of tomato somaclones with those occurring in the progeny from chemically mutagenized plants (Gavazzi et al. 1987), the results indicated that in vitro culture had resulted in a higher frequency of some mutations and a different spectrum of variation, some mutants being found exclusively in the somaclonal population. Origin of Tissue-culture-induced Variation There is no complete understanding of which factors induce or otherwise affect the nature and frequency of instability in plants regenerated from tissue culture. Among the parameters to be examined (de Klerk 1990; Orton 1983), changes in the chromosome complement have the advantage of providing a visible measure of genetic variation at the cellular level that can be correlated with the physical and chemical cultural environment and the corresponding morphological variation at the plant level. For this reason there are more extensive reports on karyological variation than on any other form of tissue-culture-induced variation.

PAGE 23

16 Karyoloqical Variation Numerical and structural chromosome variation in cell and tissue cultures and their regenerates has been the subject of several excellent reviews (Bayliss 1980; Constantin 1981; Wersuhn and Dathe 1983; D'Amato 1985; Karp and Bright 1985; Lee and Phillips 1988). Cytogenetic instability is normally shown by the occurrence of chromosome fragments and laggards at anaphase, multipolar spindles and the presence of micronuclei or multinucleate cells. Numerical and structural changes are further revealed by the study of meiotic chromosome behavior in regenerated plants. The full range of karyotypic variation has been observed, including ploidy changes, aneuploidy, deletions, duplications, inversions and translocations (e.g., McCoy ef al. 1 982; Johnson ef al. 1 987; Lee and Phillips 1 987; Armstrong and Phillips 1 988; Benzion and Phillips 1988). Origin of Chromosome Variability The nature and extent of chromosome variability obviously depends on the interaction of a number of exogenous and endogenous factors. In simple terms, however, it can be stated that variation in tissue-culture-derived plants either (1) was already present in the donor plant tissues or (2) is induced by the particular conditions of the culture phase (Karp 1989). Preexisting variability . When somatic cells cease to divide, in the majority of angiosperms, at least a part of the differentiating tissues goes through one or several additional rounds of endoreduplication of the nuclear DNA without an intervening mitosis. This leads to a geometric increase in the DNA content of the nucleus (4C,8C,16C. . .). Mature tissues, thus, generally exhibit varying levels of endopolyploidy in a varying proportion of their cells, a condition termed polysomaty (cf., D'Amato 1975, 1977, 1985). When callus is initiated from polysomatic tissues, at least some of the endopolyploid cells may be induced to divide and proliferate, and thus produce variant

PAGE 24

17 genotypes. The degree of endoreduplication is not generally uniform. It is often dependent on the age and type of tissues and organs (cf., D'Amato 1985) and even on the mode of plant cultivation (Pijnacker et al. 1 989). Therefore, choosing explants with a low degree of polysomaty is one means of reducing variability. Explants comprising meristematic or very young tissues are especially recommended for establishing cultures (Lee and Phillips 1988; Karp 1989). Differences are also found among and within species, demonstrating the genetic control of polysomaty; some species are nonpolysomatic altogether (D'Amato 1985). In some plants the apical meristems, and consequently the mature tissues, comprise a mosaic of cells with varying proportions of different aneuploid chromosome numbers. This condition, termed aneusomaty, is transmitted to and generally enhanced in callus cultures derived from such tissues (e.g., Heinz et al. 1969). Diplontic selection may act during regeneration to give rise to normal or near normal plants, but aneusomatic plants are more commonly produced from these callus cultures (e.g., Heinz and Mee 1971). When plants are regenerated from aneusomatic tissues and their origin is multicellular (as suggested by the presence of chromosome number mosaics in mixoploid regenerants), some of the regenerants may be chimeric and bear sectors (sectorial and mericlinal chimeras) or entire histogenetic layers (periclinal chimeras) of different chromosomal constitution. Sectorial ploidy chimeras have been detected, for example, among regenerants in maize (McCoy and Phillips 1 982; Benzion and Phillips 1988). Most will probably remain unnoticed, however, since karyological analyses are mostly done on root tip cells or microspore mother cells, which are derived from single histogenic layers. As mentioned earlier, the resolution of chimeric structures may also be a source of tissue-culture-induced variation, particularly important in vegetatively propagated plants (Skeene and Barlass 1983; Preil 1986). If the component histogenic layers of a chimeral cultivar are karyotypically distinct, this can contribute to chromosome

PAGE 25

18 variability in the corresponding callus cultures and their regenerates (e.g., Skirvin and Janick 1976). The common production of chromosomal mosaics in the regenerants is also evidence for the multicellular origin of the in vitro regenerated shoot apices in these cases (cf., D'Amato 1985). Chromosome structural changes and spontaneous gene mutations may also occur in mature tissues. The former are generally detected by aberrations observed in the mitotic cycle. The latter, however, may escape detection, since most spontaneous mutations are recessive and rarely have an immediate effect on the phenotype of the heterozygous diploid (cf., D'Amato 1985). Both structural changes and gene mutations may add an extra component to in vitro produced variation. The acceptance that at least part of the observed somaclonal variation originates from preexisting variation (cf., Morrish et al. 1990) has strong implications on experimental design. To account for the possibility of this origin, it is critical that all experiments involving the analysis of somaclonal variants have appropriate controls. Furthermore, careful record keeping is essential as the origin and genealogy of each regenerant must be clearly known (Karp 1989). Tissue-culture-induced variability . In addition to the variability possibly introduced with the explant, variant karyotypes may also arise during callus induction and subsequent growth. The induction of variation during callus initiation and subculture is particularly well studied in non-polysomatic plants where cultures are initiated from uniformly diploid cell populations. In polysomatic species culture-induced variation is revealed by variability within calli derived from single cells or from protoplasts (e.g., Karp era/. 1982). The most frequently reported change in culture is polyploidization; aneuploidy is also commonly found. Polyploidy may originate through endoreduplication followed by mitosis, by restitution nucleus formation, i.e., chromosome doubling due to spindle

PAGE 26

19 failure or by nuclear fusion in multinucleate cells. Aneuploidy may arise through unequal chromosome segregation at mitosis or by amitotic nuclear fragmentation followed by mitosis (cf., D'Amato 1985). Genotype and Ploidy Level . There is reasonable evidence to believe that there is a genotypic component to instability in tissue cultures. This has been demonstrated in a variety of crop species, where some cultivars consistently give rise to a wider range of somaclonal variation than others (e.g., Galiba et al. 1985; Linacero and Vazquez 1986b; Larkin 1987). The ploidy level of the starting material, for example, has been found to be a strong determinant of the frequency of ploidy changes in culture. In Gramineae, where comparisons have been made between polyploid and diploid genotypes of the same species, the former tended to show higher variation in culture. For example, 12 of 47 regenerates from two 4x ryegrass (Lolium multiflorum L.) genotypes lost up to three chromosomes, while all the regenerants of the respective 2x genotypes retained the normal diploid chromosome number (Jackson and Dale 1988). There seems to exist a positive correlation between polyploidization and the occurrence of aneuploidy, an indication that there is a higher tolerance to aneuploid variation at higher ploidy levels (cf., Larkin and Scowcroft 1981 ; D'Amato 1985). Further examples of genotypic influence on the frequency of chromosome doubling, aneuploidy or structural changes were noted in rice (Sun etal. 1983) and oats (McCoy era/. 1982). Chromosome structural changes have been observed in mitotic figures in callus of varying age. They may originate from chromatid breakage and subsequent healing or fusion of broken ends, in which case they can bring about continuing variation by initiating a breakage-fusion-bridge cycle that can be self-perpetuating over a number of cell generations (cf., McClintock 1978). Structural changes may occur at the diploid level, but their incidence was also found to be higher after polyploidization (cf., D'Amato 1985).

PAGE 27

20 Media Composition and Tissue Culture Procedures . The level of chromosome variability and its fate during culture is determined by the extent of preexisting variation, the rate of de novo production of variant karyotypes and their relative competitive ability. Continuing production of new karyotypes with identical chances of survival would lead to an accumulation of variants and, consequently, a steady increase in heterogeneity of the culture (Bayliss 1980). However, since in reality not all the variants are equally good competitors, such cultures eventually reach an equilibrium where variability does not increase any further. It has been shown that differences in the competitive ability of different cell lines can be strongly influenced (to the point of becoming reversed) simply by changing the transfer frequency or the composition of the nutrient medium (Bayliss 1980). The result is that, after a period of rapid increase in variability as formerly described, one or a few cell lines with a high selective advantage gradually emerge to make up the bulk of or the whole cell population in culture. Much attention has been devoted to the effects of medium constituents, especially growth regulators, on chromosome variability, but data have been conflicting and difficult to interpret (Karp and Bright 1985; Karp 1989). It is particularly difficult to separate the roles of substances as triggers of genetic changes from agents promoting the selective growth of specific variant cell types (Bayliss 1980; Karp 1989). Because of their involvement in the control of cell division (cf., Skoog and Miller 1957; Furuya 1984; Gould 1984), auxins (in particular 2,4-D) and cytokinins have been considered as prime candidates in promoting chromosome variability through perturbations of the mitotic process (Karp and Bright 1985; Karp 1989). While some of the data in the literature seem to support this view, most do not. In the particular case of grasses and cereals, where 2,4-D has been the hormone of choice for establishing embryogenic callus cultures, there is absolutely no evidence that the auxin (or any other growth regulators) is mutagenic at all; the choice of explant and the regeneration pathway are much more important factors for consideration (Morrish et al. 1987). Reports of changes in ploidy

PAGE 28

21 level associated with growth regulators (e.g., Ghosh and Gadgil 1979) may be only the result of the hormones promoting cell division at specific ploidy levels (Karp 1989). The influence of time in culture on chromosome instability has long been recognized (Torrey 1 967; Sacristan and Melchers 1 969; Zagorska et a/. 1 974; Jones and Murashige 1974; D'Amato 1977; Ghosh and Gadgil 1979; Boucaud and Caultier 1 981 ; Sutter and Langhans 1 981 ; Lorz and Scowcroft 1 983; Westerhof ef a/. 1 984; Stimart 1 986; Cassells and Morrish 1 987; Binarova and Dolezel 1 988; Lee and Phillips 1988; Karp 1989; Swartz 1991). In general, the longer explants remain in the culture phase, the greater the chromosomal instability; variant karyotypes commonly accumulate with increasing age of cultures. Because, as described earlier, some of the variation found in the callus is usually transmitted to the regenerants, the relative proportion of variant plants produced during successive passages generally also increases. In Gramineae, an increase in regenerants displaying polyploidy, aneuploidy or chromosome structural changes has been observed, for example, in maize (Lee and Phillips 1987; Armstrong and Phillips 1988; Benzion and Phillips 1988), rice (Fukui 1986), oat (McCoy er a/. 1982), barley (Orton 1980) and triploid ryegrass (Ahloowalia 1983). In most of these cases, the proportion of karyotypically normal plants was 100%, or close to 100%, during the first transfers. As a consequence, restricting the number of callus passages in culture has been frequently recommended as an effective means of reducing variability, whenever genetic stability is essential to the research (Karp 1989). There is also good evidence that frequent transfers, as compared to extended subculture intervals, tend to yield more stable callus cultures (Cassells and Morrish 1987; Lee and Phillips 1988; Karp 1989). Evolution of chromosome variability may finally be dependent on whether cells are grown as callus masses on solid medium or in liquid medium as cell suspension cultures. Where comparisons have been made, a higher level of variability was normally found in the callus cultures (e.g., Singh and Harvey 1975; Ashmore and Shapcott 1989).

PAGE 29

22 Regeneration from protoplast-derived cultures is generally associated with unusually high levels of chromosomal variation (Karp 1989). This may be a reflection of the increased complexity and length of time in culture associated with protoplast regeneration protocols, or it may be simply a consequence of the additional physiological stress imposed by the removal of the cell wall. Regeneration Pathway Histological and morphological studies have provided evidence that plant regeneration from callus cultures may occur by two different distinct pathways: adventitious shoot morphogenesis (organogenesis) or somatic embryogenesis (Morrish era/. 1987; I.K. Vasil 1987). Both morphogenic processes involve a complex sequence of genetically controlled developmental steps. Since karyologically aberrant cells may be impaired in their control of one or more of these steps, morphogenesis acts as a screen that prevents a portion of the variant karyotypes from being transmitted to the regenerated plants. This would explain the common observation that the increasing chromosomal variability often associated with aging cultures is negatively correlated with their regeneration potential (e.g., Murashige and Nakano 1967; Torrey 1967; Smith and Street 1974; Zagorska era/. 1974; Orton 1985; Jha and Sen 1987). Differences in the extent to which karyotypically variant cells are excluded from plant regeneration are found among and within species, demonstrating that the process is under genetic control. Callus cultures have been mostly described as masses of dedifferentiated cells proliferating in a random and unorganized way. Although this seems to be true for the majority of the dicots, histological investigations have revealed that in the monocots a different type of callus often occurs. This has been described as compact and organized and is commonly interpreted as a mass of proliferating partially suppressed embryos in various stages of development, rather than completely undifferentiated

PAGE 30

23 tissue. Compact calli of this nature are predominantly produced by meristematic tissues in immature and young tissues of inflorescences, embryos and leaves (e.g., Hunault 1979; Dale 1980; V. Vasil and I.K. Vasil 1980, 1981a; Lu and I.K. Vasil 1981, 1 982; Wernicke er a/. 1 981 , 1 982; Hanning and Conger 1 982; Lu et al. 1 982; Thomas and Scott 1985). Because organized compact calli are often associated with regeneration via the embryogenic pathway they have been termed embryogenic calli. In Gramineae, by far the most well studied group in the monocots, production of embryogenic callus has now been reported in most cereal and other grass species, suggesting that somatic embryogenesis may be the most common pathway of plant regeneration in these plants (I.K. Vasil 1982, 1983a,b, 1985, 1987; I.K. Vasil and V. Vasil 1986). In many species that regenerate from embryogenic callus cultures a strong selective constraint seems to exist, that minimizes the amount of chromosome variation observed in the regenerants. This was observed, for example, in cultures and regenerants of several monocots such as daylilies (Krikorian et al. 1981), asparagus (Becker and Reuther 1986) and most grasses, including cereals (I.K. Vasil and V. Vasil 1986a; Morrish et al. 1987). The same apparent karyological stability seems to apply also to embryogenic cultures derived from zygotic embryos of conifers (Papes et al. 1983; Schuller era/. 1989), which are composed of masses of proliferating proembryos rather than disorganized calli (cf., von Arnold and Wallin 1988). The relationship between the callus growth pattern and the level of chromosome variability is particularly evident in cultures where both embryogenic and nonembryogenic types coexist. In Gramineae, for example, compact organized embryogenic callus may occur together with a more dedifferentiated, friable and more or less non-morphogenic type of callus (cf., I.K. Vasil 1985, 1987; I.K. Vasil and V. Vasil 1 986; Morrish et al. 1 987). Comparative cytogenetic data on the two types of calli (e.g., Scheunert et al. 1978; Swedlund and I.K. Vasil 1985; Singh 1986) show that the

PAGE 31

24 embryogenic callus generally exhibits a lower degree of karyotypic variability than that found in other kinds of cultures (e.g., Orton 1980; Ahloowalia 1982). Since the nature of the callus can determine both its ability to regenerate and the stability of the regenerants (I.K. Vasil and V. Vasil 1986; Morrish et al. 1987), it is conceivable that the primordial nature of embryogenic calli implies an organized pattern and a rate of cell division comparable to those in meristems of intact plants (I.K. Vasil 1985, 1987). As in meristems, these kinds of cultures would then retain a high level of karyological stability. This higher chromosomal stability of embryogenic calli and the fact that somatic embryogenesis per se seems to be a highly selective process which strongly favors regeneration from normal cells, would explain why only a small portion of the karyological variation, if any, in embryogenic cultures is recovered in the regenerants (I.K. Vasil 1983a, 1987; Swedlund and I.K. Vasil 1985). Regenerants from embryogenic cultures tend to be diploid, or at least euploid, even when regenerating from calli or suspension cultures where polyploidy or aneuploidy had been formerly observed (e.g., Swedlund and I.K. Vasil 1985; Hahne and Hoffmann 1986; Sengupta et al. 1988; Taniguchi and Tanaka 1989). Based on the high level of karyotypic stability found in both the embryogenic calli and their regenerants in several grass species (e.g., Hanna et al. 1984; Armstrong and Green 1985; Swedlund and I.K. Vasil 1985; Rajasekaran et al. 1986), I.K. Vasil (1983a, 1985, 1987) postulated a higher stability of embryogenic cultures compared with organogenic ones and recommended their use whenever genetic stability and uniformity are needed. Such generalization, however, has been the theme of debate. Many claim that, although karyological and phenotypic stability can often be correlated with the embryogenic pathway, varying frequencies of abnormal regenerants from embryogenic tissue cultures have also been reported in the literature (Gobel etal. 1986; Karp 1989). These may have resulted from cultures where both the embryogenic and organogenic pathways coexisted, as has been described in wheat (Karp and Maddock

PAGE 32

25 1 984) where the variation reported may result from plants being regenerated from the mixed callus via the organogenic pathway (Maddock 1985). In maize, where both regeneration pathways also have been reported (Green and Phillips 1 975; Green 1 982; Earle and Gracen 1985; Armstrong and Phillips 1988), a detailed analysis of maize plants regenerated from immature embryo-derived embryogenic callus cultures and their sexual progeny showed variation occurring at both the morphological and molecular levels (Gobel et a/. 1986; Brown et al. 1991). However, the authors were the first to acknowledge that ". . .although the intention was to use somatic, embryo-derived structures for plant regeneration, it cannot be ascertained that all of the regenerants are somatic and embryo derived " (p. 23 in Gobel ef al. 1986). In an identical study, but where plants were regenerated from strictly controlled embryogenic callus cultures of Napier grass (Pennisetum purpureum K. Schum.), Shenoy and IX Vasil (1992) showed a complete absence of variation at the morphological, biochemical (14 different isozyme loci) and molecular (RFLP analysis of the mitochondrial, plastid and nuclear genomes) levels. Identical results were observed for the mitochondrial, plastid and nuclear genomes in sugarcane (Chowdhury and IX Vasil 1993) and meadow fescue (Valles et al. 1993), lending further support to the suggestion that regeneration via somatic embryogenesis largely avoids phenotypic and genetic instability. Finally, the variants that have been reported may also be transient in nature (Larkin and Scowcroft 1983b; Irvine 1984), resulting from epigenetic rather than genetic changes in the regenerated plants. Regardless of the final verdict on the applicability of somatic embryogenesis as a way of obtaining stable regenerants from tissue cultures, it does have advantages as a mode of regeneration. It imparts genetic uniformity, a consequence of the fact that somatic embryos, like their zygotic counterparts, derive largely from single cells (cf., Morrish et al. 1987; IX Vasil 1987). This prevents the occurrence of mosaics and

PAGE 33

26 chimeras at the plant level (Morrish etal. 1987; I.K. Vasil 1987; Armstrong and Phillips 1988; Karp 1989). Possible Causes of Chromosomal Variability By emphasizing chromosomal structural rearrangements, some of the possibilities of how culture-induced alterations of the intracellular environment could lead to chromosomal changes have been recently detailed (Lee and Phillips 1988). Consistent with various reports where rearrangements have been shown to involve breakage preferably at or close to heterochromatic regions (Sacristan 1 971 ; McCoy et al. 1 982; Lapitan et al. 1 984; Murata and Orton 1 984; Johnson er al. 1 987; Lee and Phillips 1987; Benzion and Phillips 1988), it was speculated that late replication, characteristic of heterochromatic chromosome segments, may be further delayed under culture conditions, causing perturbations to DNA synthesis. If such delay lasts until mitosis, formation of nonreplicated heterochromatin bridges, with consequent breakage at anaphase, would be the consequence (Lee and Phillips 1988). While breakage events within or near heterochromatin would be the primary source of chromosome rearrangements, further variation can result from breakagefusion-bridge (BFB) cycles, which tend to be perpetuated through the formation of dicentric chromosomes (Lee and Phillips 1988). Repeated breakage associated with BFB cycles may trigger various forms of genome reorganization, from gross and minor translocations to excision and splicing of transposable elements, affecting even chromosomes not involved in the BFB cycle at all (McClintock 1978). Activation of transposable elements as a result of culture-induced chromosomal changes has been confirmed in maize (Peschke er al. 1986, 1987, 1991), for example, and it has been speculated that transposon-mediated variability could be a general phenomenon in callus cultures (Larkin and Scowcroft 1 981 ; Larkin et al. 1 984; Toncelli et al. 1 985; Lorz and Brown 1986).

PAGE 34

27 Another possible component on the induction of variation, proposed from evidence obtained from other organisms, could be the nucleotide pool imbalance associated with nutrient depletion in plant tissue cultures. This, in turn, would cause disturbed DNA replication and repair, which could lead to chromosome breakage, reciprocal and nonreciprocal sister chromatid exchange and mitotic crossing-over or somatic recombination events between nonhomologous chromosomes (Lee and Phillips 1988). Accumulation of toxic and potentially 'automutagenic' metabolites in ageing cultures could also have the same effect (cf., D'Amato 1985). Phillips ef al. (1 990) attempted to define a unified theory that would encompass in a single cause such seemingly unrelated events as single gene mutations, activation of transposable elements, qualitative trait variation and chromosome breakage. They postulated that the methylation status of the DNA, easily altered by the tissue culture environment (e.g., nucleotide pool imbalance, as formerly mentioned), is at the origin of DNA modifications that, in one way or another, could lead to all of the most frequent disturbances that occur in plant tissue cultures. This would imply that the mechanism of in vitro-induced variation is epigenetic rather than genetic, an interesting approach that certainly may help explain many of the unique characteristics of somaclonal variation, in particular its high frequency and widespread occurrence, the presence of seemingly stable mutational events that are lost upon crossing (e.g., Oono 1985) and the peculiar observation that some variant phenotypes are characteristic and found only in the somaclonal populations. Finally, and in another class by itself, the presence of an altered cytoskeleton has also been related to tissue-culture-induced mitotic disturbances. Studies have shown that specific microtubular arrays associated with cell division may be absent or less distinct in callus cells than in intact meristems (cf., Seagull 1989). The chromosome variation frequently observed in vitro could thus be caused by a malfunctioning of the mitotic and cytokinetic apparatus. This would be particularly true

PAGE 35

28 in protoplast cultures, which are known to have a disturbed microtubular organization (Hahne and Hoffmann. 1986). Considerable progress has been made over the years in our perception of tissue-culture-induced variation. The major lesson to retain, though, is that a great deal more remains to be learned about. Gathering accurate information at the morphological, karyological, biochemical and molecular levels is critical, but solid theories are likely to be crafted only with careful integration of all this knowledge, an approach that so far has been somewhat neglected.

PAGE 36

CHAPTER 3 SOMATIC EMBRYOGENESIS AND LONG-TERM REGENERATION IN CALLUS CULTURES OF DIPLOPERENNIAL TEOSINTE Introduction Diploperennial teosinte (diploid perennial teosinte, Zea diploperennis litis, Doebley & Guzman) is a newly discovered perennial teosinte, or "wild maize", from the highlands of Mexico (Guzman Mejia 1978; litis ef a/. 1979) (Fig. 3.1). Taxonomically a relative of maize (Z. mays L), Z. diploperennis was described as endemic in nature to a small localized area of the Sierra de Manantlan, a mountain range of western Mexico, where it grows alongside small streams and, sometimes, on the edges of maize fields or in grazed pastures. Morphologically, diploperennial teosinte differs from perennial teosinte [Z. perennis (Hitchcock) Reeves and Mangelsdorf], a rare tetraploid species endemic to the same area, by the presence of dimorphic rhizomes, a more robust growth habit and larger tassels. Specially relevant, however, is that Z. diploperennis is a diploid (2n = 2x = 20), which permits the potential transmission of its many agriculturally useful traits to maize. For this it has been called the "botanical find of the century" (Nault and Findley 1981). Genetically, the species is a hybridizer's dream. Diploperennial teosinte carries genes for immunity or tolerance to several important pests and diseases in its germplasm (Nault 1980; Nault ef a/. 1980, 1982), together with a variety of other agronomically important traits. Zea diploperennis produces fertile offspring when crossed with maize and, because it retains the primitive perennial trait in a diploid condition, it is the major potential genetic source for the production of a perennial 29

PAGE 37

Fig. 3.1 Mature flowering plant of diploperennial teosinte, Zea diploperennis. Fig. 3.2 Plants of diploperennial teosinte grown under greenhouse-controlled conditions. The metal frames above the plants supported black cloth, used to simulate short day conditions needed to force flowering out of season.

PAGE 39

32 maize, a goal that has long eluded breeders (litis et al. 1 979; Nault and Findley 1 981 ; Crosswhite 1982). Since its introduction in the late seventies, the species has been successfully incorporated in a number of maize breeding programs, many of them currently undergoing evaluation and selection (Camara-Hernandez and Mangelsdorf 1981; Galinat 1981; Mangelsdorf er al. 1981; Nault and Findley 1981; Crosswhite 1982; Findley etal. 1982; Magoja and Pischedda 1986; Magoja and Palacios 1987; Palacios and Magoja 1987, 1988; Pischedda and Magoja 1987, 1988, 1990; Corcuera and Magoja 1988a,b,c; Carlson and Price 1989a,b). Other than its promising potential in breeding, diploperennial teosinte makes an excellent choice for studies involving tissue cultures and tissue-culture-generated variation. It is a close relative of maize, a species that is particularly well characterized at the tissue culture level (Green and Phillips 1975; Springer etal. 1979; Green 1982; Lu er al. 1982, 1983; Chang 1983; Novak er al. 1983; V. Vasil er al. 1983, 1984, 1985; Holbrook and Molnar 1 984; Armstrong and Green 1 985; Duncan et al. 1 985; Lowe ef al. 1985; Radojevic 1985; Fahey etal. 1986; Hodges etal. 1986; Petolino and Jones 1986; Rhodes et al. 1986; Suprasanna et al. 1986; V. Vasil and I.K. Vasil 1986; Close and Ludeman 1987; Conger et al. 1987; Wang 1987; Wilkinson and Thompson 1987; Duncan and Widholm 1988; Pescitelli etal. 1989; Pareddy and Petolino 1990; Songstad et al. 1992) as well as at the karyological, biochemical and molecular levels and for which somaclonal variation has been formerly reported and extensively evaluated (e.g., Gengenbach etal. 1981; Hibberd and Green 1982; McCoy and Phillips 1982; Umbeck and Gengenbach 1983; Brettel er al. 1986a; Gobel er al. 1986; Lee and Phillips 1987; Armstrong and Phillips 1988). In addition, being diploid, perennial and fully fertile, diploperennial teosinte should produce regenerants from tissue culture that can be analyzed over long periods of time and across generations, an absolute need when attempting to clarify the origin and nature of somaclonal variation.

PAGE 40

33 Protocols that lead to the successful tissue culture and regeneration of diploperennial teosinte have been published (Prioli et al. 1984, 1985; Sondahl et al. 1984; Swedlund and Locy 1988). As for other members of Gramineae (V. Vasil and I.K. Vasil 1984a; I.K. Vasil 1985, 1987; I.K. Vasil and V. Vasil 1986; Morrish ef al. 1987), (1) use of very young, immature explants containing mostly undifferentiated meristematic cells, (2) addition to the media of relatively high concentrations of synthetic auxins and (3) early recognition and selection of regenerable calli were the keys to the successful initiation and maintenance of totipotent cultures. Reports on regeneration from tissue cultures of diploperennial teosinte, however, are conflicting. Upon curturing apical and lateral meristems of Zea diploperennis, Prioli ef al. (1984) reported regeneration from a semi-friable callus from which adventitious shoots arose. Although the morphology of regeneration was not investigated, it was claimed that typical somatic embryogenesis did not occur (Prioli et al. 1984, 1985). Conversely, Swedlund and Locy (1 988) reported a system much closer to that commonly described for other species of Gramineae (V. Vasil and I.K. Vasil 1984a), with both nonembryogenic and embryogenic calli being formed from immature zygotic embryos and shoot apices. Regeneration occurred by somatic embryogenesis, with the embryogenic cells described as small, densely cytoplasmic and starchy. Somatic-embryo-derived plants did not differ in chromosome number or overall morphology from the original donor plants (Swedlund and Locy 1988). This chapter reports in detail the results of studies leading to the development of a protocol for the reproducible induction of somatic embryogenesis and long-term (several years) regeneration from callus cultures of diploperennial teosinte. In particular, it describes the effect of using alternative compounds exhibiting auxin-like activity and demonstrates the potential importance of L-proline and other medium additives in optimizing the embryogenic response.

PAGE 41

34 Materials and Methods Plant Material Plants of diploperennial teosinte were raised from seed obtained from Professor H.H. litis (Department of Botany, University of Wisconsin-Madison). The wild-collected seeds, harvested near the type location in Mexico, were accompanied by the following information: Zea diploperennis ex Herbarium, University of Wisconsin, Madison, Wisconsin. Zea diploperennis litis, Doebley & Guzman. Plants 1 5-25 dm tall in dense, many-stemmed (30 to 1 00 in each) colonies. a) Thickets along stream bordered by Alnus trees. b) Maize fields, and abandoned maize fields, with scattered Crataegus mexicana. Flat valley bottoms at Las Joyas (cf. El Chante topo-sheet), 7.8 Km W by WSW of Rincon de Manantlan, ca. 16 Km SW of El Chante, Jalisco, Mexico, JAN1984. 1 9° 35' 50" N, 1 04° 1 7' W, Alt. ca. 1 ,900 m. H. H. litis and R. Guzman M., No. 291 15. Field work supported by National Science Foundation Grant No. BM S74-21 861 ; O. N. Allen Herbarium Fund; Research Committee, Graduate School, University of Wisconsin-Madison; and Pioneer Hi-bred International. To these were also added other accessions, obtained as vegetative propagations from Drs. D. Pring and W. Judd (Plant Pathology and Botany Departments, University of Florida, respectively). These plants were also derived from seed donated by Professor litis and probably originate from the same field collection. The fruits were planted, one caryopsis per gallon plastic container, in Metromix® Vegetable Plug Mix amended with 20% Perlite. Germination occurred in 3-5 days and a large number of plants were raised to maturity. These were grown at a density of 9 pots/m 2 in a greenhouse (Fig. 3.2). All plants were fertilized once per growth cycle with Osmocote®, a slow-release fertilizer (17-6-10 plus minor elements). Mean average

PAGE 42

35 daytime temperatures ranged from 20°-25°C in winter to 30°-35°C in summer. The temperature was not allowed to fall below 16°C during cold winter nights but it often rose well above 35°C during summer. Because teosintes are naturally short-day plants (Emerson 1924; Mangelsdorf 1974), flowering was either induced naturally, under the influence of the short days of winter and early spring, or was forced whenever needed by keeping the plants in darkness for ca. 16 hours a day (Emerson 1924; Mangelsdorf 1974). Early application of short-day treatments to specimens grown from seed, however, was best avoided, as it consistently resulted in the production of short plants that, although physiologically mature, flowered only rather sparsely. Young elongating leaf bases, young tassels and mature and immature zygotic embryos were used as explants. For both leaf bases and young developing tassels terminal shoots were cut from the plants, washed with tap water and surface sterilized with 90% ethanol. In a sterile laminar-flow hood the green mature outer leaves were removed to expose the white to yellowish-green inner expanding leaves or the immature staminate inflorescences. Sections from the region slightly below to 2 cm above the shoot apex were sliced transversely into two millimeter thin segments that were placed in presterilized disposable plastic Petri dishes (100 mm diameter, 15 mm high) on approximately 25 ml of culture medium, one plant per dish. The normal polarity was maintained for the leaf sections in culture. Controlled pollinations were performed by bagging both the staminate and the pistillate inflorescences before anthesis and hand-pollinating each individual ear when pollen was available. Pollinations were carried out with pollen from a different plant or from tillers of different physiological ages in the same plant because under our conditions maturation of the male inflorescences preceded that of the female by several days. Immature ears were harvested when the embryos were at the desired developmental stage, normally between the 8th and 10th day after pollination. This was

PAGE 43

36 done except for the embryo size study, for which ears were harvested every day from day 5 to day 15. The ears were de-husked and the green caryopses were surface sterilized in 1 0% Clorox containing two drops of a surfactant (Tween 20) for 20 minutes, then rinsed in several changes of sterile distilled water. The fruit cases were then carefully removed in a sterile laminar-flow hood to expose the internal structures. Immature embryos, ranging in size from less than one millimeter to five millimeters, were carefully dissected and cultured as previously described for maize (Lu era/. 1983) and diploperennial teosinte (Swedlund and Locy 1988). The embryos were usually isolated on the same day the ear was harvested, but occasionally the ears were refrigerated overnight and the embryos were dissected the following day. Embryos were placed on the surface of the culture medium, embryonic axis down, in presterilized disposable plastic Petri dishes (100 mm diameter, 15 mm high) containing approximately 25 ml of culture medium, ten embryos per dish. Whenever possible, pieces of the semi-solid endosperm (the stage of development between day 9 and day 1 3 after pollination) were also placed in culture following the same protocol. At very young stages of development, when the embryo was too small to be safely dissected, the whole dissected ovary was placed in culture. Induction of Somatic Embrvoqenesis The basal nutrient medium for induction of somatic embryogenesis consisted of Murashige and Skoog's (1962) major and minor nutrients and vitamins, 100 mg/l myoinositol and 5% coconut milk (MSC basal medium). The following parameters were tested for their influence on induction of somatic embryogenesis: 1. Embryo developmental stage, expressed by the length of the embryo as measured from the base to the tip of the scutellum. 2. Auxin type and concentration: a series of four IAA conjugates, including lAAIa (indoleacetyl-D.L-alanine), lAAsp (indoleacetyl-D,L-aspartic acid), lAGly

PAGE 44

37 (indoleacetyl-glycine) and lAPhe (indoleacetyl-D,L-phenylalanine), NAA (naphthaleneacetic acid), 2,4-D (2,4-dichlorophenoxyacetic acid), dicamba (3,6-dichloro-2-methoxybenzoic acid) and picloram (4-amino-3,5,6-trichloropicolinic acid). All auxins were tested at the following concentrations: 2.5, 5.0, 10.0 and 25.0 /iM. 3. Sucrose: 3%, 6% and 1 2%. 4. L-proline: 12 mM. All media were solidified by addition of 2 g/l Gelrite™ and pH was adjusted to 5.8 before autoclaving at 121°C for 20 minutes. L-proline and all the auxins were filtersterilized and added after autoclaving. The Petri dishes were sealed with Parafilm® and cultures were incubated at 28°C in the dark. Callus Maintenance and Multiplication Except for the factor(s) under testing, a basal nutrient medium containing Murashige and Skoog's major and minor nutrients and vitamins with 100 mg/l myoinositol, 12 mM L-proline, 500 mg/l casamino acids, 2% sucrose, 1% glucose, 10 /l/M picloram and 2 g/l Gelrite, pH 5.8, was used in all experiments. The following parameters were evaluated for their effect on growth rate of embryogenic calli: 1. Auxin type: 2,4-D, dicamba and picloram, at an equimolar concentration of 10pM. 2. Auxin concentration: only picloram was used, at the concentrations of 2.5, 5.0, 10.0,20.0 and 30.0 pM. 3. Carbohydrates: glucose and sucrose, in the following combinations: glucose 3%; glucose 2% + sucrose 1 %; glucose 1 % + sucrose 2% and sucrose 3%. 4. Casamino acids: 500 mg/l. 5. L-proline: 1 2, 25, 50 and 1 00 mM.

PAGE 45

38 Ten pieces of embryogenic callus were initially placed in each Petri dish. To minimize the lag phase and avoid carry-over effects from the maintenance medium all cultures under evaluation were subcultured twice, at 4 week intervals, in the experimental medium. All measurements were taken on the second subculture. Whenever a best combination of factors was determined, callus pieces from that treatment were tested for regeneration, to assess if the factor(s) that optimized growth in tissue culture had any effect on the expression of totipotency. This was done by placing pieces of embryogenic callus in a regeneration medium consisting of Murashige and Skoog's major and minor nutrients and vitamins with 100 mg/l myo-inositol, 1 fjM picloram and 2.5 /jM kinetin and visually evaluating the potential to produce plantlets in vitro. Data Collection and Analysis Induction of somatic embrvoqenesis . Immature embryos were checked for growth every few days and scored for the presence of embryogenic or nonembryogenic callus four weeks after culture initiation. Hard, nodular, creamy-white callus with a convoluted surface, comparable to embryogenic calli from other grass species, was classified as embryogenic (E). Soft, friable, translucent callus with a watery appearance was classified as nonembryogenic (NE). No friable embryogenic (Type-ll) callus, equivalent to that formerly described for maize (Green 1982, 1983), was produced. For both kinds the frequency of callus formation was converted to a percentage (number of explants showing that callus type / total number of explants in culture x 100). All percentage data were transformed using the arc-sine transformation for proportions (Steel and Torrie 1980) before analysis . The data presented in Tables 3.1 through 3.5 were analyzed by analysis of variance using the General Linear Models procedure of the Statistical Analysis System (SAS) for the PC, Release 6.04. All plotting and curve

PAGE 46

39 fitting in Figs. 3.16 through 3.30 was done with Microcal Origin™, version 2.75, running under Microsoft® Windows™ 3.1 . Callus maintenance and proliferation . Fresh weights for all treatments were measured every other day for twenty six days. Results were plotted as growth curves, showing variation of fresh weight as a function of time. The data plotted in Figs. 3.21 , 3.23, 3.25, 3.27 and 3.29 are the mean values and standard errors of the mean (SE) from the five replicates. Growth rates were calculated from the growth curves during the exponential phase of growth (day 2 through day 26) by linear regression techniques. The fitting function was determined by using the General Linear Models procedure of the Statistical Analysis System (SAS) for the PC, Release 6.04 and the calculated slopes were plotted against the variable under study. The results are shown in Tables 3.6 through 3.10 and Figs. 3.22, 3.24, 3.26, 3.28 and 3.30. All plotting and curve fitting in Figs. 3.16 through 3.30 was done with Microcal Origin™, version 2.75, running under Microsoft® Windows™ 3.1 . Plant Regeneration Cultural procedures . Plants were regenerated from embryogenic cultures by transferring callus pieces onto MS basal medium with 0 or 1 fjM picloram, with or without 2.5 pM of a cytokinin. Cytokinins tested included 6-furfurylaminopurine (kinetin), N 6 -benzyladenine (BA), N 6 -(2-isopentenyl)adenine (2iP) and zeatin. Regenerating callus pieces were exposed to a 1 6-hour photoperiod under cool white fluorescent tubes at 28°C. Germinated somatic embryos were transferred to half strength MS salts and vitamins with 2% sucrose in individual test tubes and kept under the same environmental conditions for further development and rooting. After establishment of an adequate root system, the young plants were transplanted into Conetainers™ and gradually hardened off under high humidity until established, normally in 2-3 weeks.

PAGE 47

40 Acclimated plants were transferred to gallon plastic containers, in Metromix Vegetable Plug Mix amended with 20% Perlite, fertilized with Osmocote® (17-6-10 plus minor elements) and grown to maturity in a greenhouse. Scanning electron microscopy . Regenerating embryogenic callus pieces were dissected and fixed for scanning electron microscopy following the protocol detailed in V. Vasil and I.K. Vasil (1984b). Specimens were viewed with a Hitachi S-450 scanning electron microscope and photographed using Polaroid™ Type 55 positive/negative black-and-white film. Results Induction of Somatic Embrvoqenesis Donor tissue . All tissues placed in culture responded at least partially to the conditions of the culture medium but, regardless of the cultural conditions, embryogenic callus could be isolated only from endosperm explants and immature embryos. Sections from young expanding leaves started reacting as early as 4 days after being placed in culture. Typically, there was a swelling of the cut surface in contact with the medium after 5-7 days in culture. Some explants, mostly those originating from dark colored (anthocyanin-rich) plants, produced relatively large amounts of phenolic compounds that accumulated at the cut ends and diffused into and discolored the culture medium. In such cases, to prevent the toxic effect on the tissues normally associated with the presence of such compounds (V. Vasil and I.K. Vasil 1981), the explants were transferred either to clear portions of medium in the same dish or to a new dish as required until discoloration ceased, normally after 2-3 transfers. Soft friable callus started forming at a low frequency by the end of the second week, together with root hairs and suppressed root primordia that soon degenerated into gelatinous masses that ceased proliferation. At the end of four weeks, the frequency of leaf

PAGE 48

41 explants responding to culture was less than 1 0% in all treatments and there was no trace of embryogenic callus. The same overall response was found in explants originating from immature male inflorescences, except that no roots were ever formed and the frequency of callus formation was close to 0%. Whenever callus was produced by these explants, it was soft, watery and nonembryogenic. Cell proliferation was somewhat better from immature ovary explants. In most treatments the dissected ovaries proliferated profusely at the cut bases and produced large amounts of a soft, translucent, friable and nonembryogenic callus that could not be maintained by subculture. Despite the variations introduced in the medium no embryogenic callus was ever recovered from ovary cultures as well. Forty-seven percent of the endosperm cultures, initiated in MSC basal medium containing 3% sucrose and 1 0 2,4-D, produced embryogenic callus. This callus, that could be maintained easily through subculture, was in all respects undistinguishable from embryogenic calli derived from immature embryos (see below). Regeneration from endosperm-derived embryogenic cultures has been formerly achieved in other families of plants (e.g., Tulecke et al. 1988). The objectives of this study, however, precluded the use of polyploidy, and there were no attempts to obtain plants from these cultures or otherwise to make any further use of them. Embryogenic callus was readily initiated from immature zygotic embryos. Cell proliferation in culture was evident in both immature and mature embryos, but the nature and time of appearance of the emerging calli varied considerably with the chemical constitution of the medium as well as with the developmental stage of the embryo. The appropriate developmental stages were assessed by measuring the embryo length, a parameter that is more reliable than embryo age (measured as postpollination time) for the selection of competent embryos in grass tissue cultures (V. Vasil and IX Vasil 1982).

PAGE 49

42 Under the cultural conditions of this study, a typical greenhouse-grown plant of diploperennial teosinte produced between 8 and 12 caryopses per ear (female inflorescence). Not all caryopses matured at the same time and embryos dissected from different fruits within one single ear were frequently at different stages of development even after controlled pollinations. This is either a result of the stigmas (silks) having different lengths, the longer ones causing a delayed fertilization of the egg, or an environmental effect, as discussed for other species of Gramineae (V. Vasil and I.K. Vasil 1982). Since time alone could not be used to predict embryo development, a large number of inflorescences were used to ensure that enough embryos could be found at all the desired stages. Fig. 3.3 illustrates the relative development of both the endosperm and embryos in the fruit, as related to postpollination time. It also summarizes the overall response in culture of the embryos, as dictated by their size. Three kinds of calli were produced in embryo cultures of diploperennial teosinte (Fig. 3.4). As described for most other grass species (I.K. Vasil 1987), the scutellar tissue of competent embryos in culture proliferated at the periphery, near the edges, giving rise to a ridge of highly organized tissue that soon proliferated into a creamywhite, compact and highly convoluted embryogenic type of callus (Fig. 3.5). A soft, translucent, friable callus was often produced from the same explant, by proliferation of the tissues in direct contact with the medium (Figs. 3.5 and 3.6). Such callus, which could be formed even in the absence of the nodular type, did not show any potential for embryogenesis. A second type of embryogenic callus was occasionally formed from some of the proliferating scutella. Always produced at very low frequencies, this new kind of embryogenic callus, which was structurally undistinguishable from the more common type, could be maintained easily through several cycles of subculture without ever being associated with the presence of soft, friable, nonembryogenic callus, even in

PAGE 50

43 CO CVJ 10 10 4 in c CD CO TJ O E ® CD "fO 1 i C CD CD QO £ Q. c " © o o to E I UJ o z O Si z —J I E £ x lV z UJ _) O > cr m UJ UJ O < H CO DC UJ CL CO 8 z UJ UJ cr H _l =) O z UJ CO z 2 CO UJ cr z O h< z DC UJ C5 CO CO

PAGE 51

Is II > .2 F II Rrt 5 cu « E E o. E -2 T.% |* 'go n J2 c w C i_ If . .9-1 2 -*— » .9 .52 « o o 0 X! E CD CJ 3 i: a > o 2 2 ~ « E o (73 CD CL P 8 .E cd « E f 2 -° £ £5 : to a> 0 w 0)TJ « E-o cd a> Q. O N cn CD co -Q CD CD w hi Q-TJ O) _, CD CD c « = O CD — w o c a> x> a w £ i •55 w . s T3 O 0 E -2 ? c2 -«— « £ E w -c T3 O>.0 OTJ o w 3 c co CD 2 O c q. a> o CD TT-C r'1 E c c o • o o ^ o8 2 2 2 g CD ,913 fi 2
PAGE 52

45

PAGE 53

Somatic embryogenesis in diploperennial teosinte: callus production from an immature zygotic embryo, with nodular, compact, organized embryogenic growth (Type A embryogenic callus, top left), as well as friable, unorganized, nonembryogenic growth (right). Somatic embryogenesis in diploperennial teosinte: Type C embryogenic callus. Note the highly organized nature of the embryogenic callus and the complete absence of friable nonembryogenic callus.

PAGE 54

47

PAGE 55

u a> ro -C N Q. C "O CO CO 0) 2 o | Erf c ® © a; £ 5 • ' CO C CD -C |3 toI • • CD "« 0) > " CO C CO O = CD CO "D ** tj a> « 9c w 5 2 E Q. 2 CD ^ Q_ TJ * CO S C 3 = CO co CO • .£ |o|| 11= £ CO CO TJ ft ~ CD o C CO © *O) c 2 o °E .» 2 0) Q_ is ** Q. c o o a> Q. > a) 3 CD TJ CO .£ So 9> .52 co ® C ££ 1u o 8.2 o. .9-D > TJ CD CD rQ"" 0 .E co <.<2 ? o CO Q. CO CD 3 = C O CD > o if w w E^'S 0 E O r£ ^ C o co .2 o E TJ ^ O CD CO co E < CO CO a> co d> Ll

PAGE 56

49

PAGE 57

Fig. 3.10 Somatic embryogenesis in diploperennial teosinte: scanning electron micrograph of developing somatic embryos. At the center of a lobed scutellum (top left) a shoot meristem has developed, now surrounded by a coleoptilar ring (center). A somatic proembryo can also be seen (top right). Fig. 3.11 Somatic embryogenesis in diploperennial teosinte: atypical somatic embryo, at the early coleoptilar stage. Note the aberrant scutellum and the leafy appendages at the margins, covered with trichomes. Regardless of these anomalous features, which are a result of hormonal imbalance in culture, atypical embryos develop into normal plants.

PAGE 58

51

PAGE 59

Fig. 3.12 Somatic embryogenesis in diploperennial teosinte: regenerating embryogenic callus, with a developing somatic embryo. Fig. 3.13 Somatic embryogenesis in diploperennial teosinte: fully formed somatic embryo, showing the white opaque scutellum, the coleoptile and the coleorhiza, partially embedded in the callus mass.

PAGE 60

53

PAGE 61

co""° • o cd .c .a c E 0 3 O 2 £ 0 c o Q. E o o ~ o (0 -C E W s « — £ a) PQ-oi t o n) S 9 ..O(0 S ™"F £ c O 03 W (!) O) *C T3 _ o a> f «S 2 3 o.® .52 o CO CO *; E a. o o © cd .y to El CD ±= cO u o»| E g a) CO co o E Q. o , T1 W •— co CO CD C CD E 3 co co o a> ^1 P 9 M o O fl o © CO O o "D v. CD -C CD «3c? E c w C = CO CD .2 "F en© ±= -•— ' *— < flt 3 * CD q C W CO E co § CD © a) t O) CL > £ On o g. CD 0) CO 3 or CD .b ffi 0)_c h O CO *_Q CO c E | "5. ® o o E > 5 CD CD CO T3 if? (0 co 6) T — CO 6)

PAGE 62

55

PAGE 63

56 aging cultures. This is in sharp contrast with the type more commonly described for species of Gramineae (I.K. Vasil 1987), including diploperennial teosinte (Swedlund and Locy 1988). Embryogenic cultures of most grasses are actually a mix of embryogenic and nonembryogenic cells, a consequence of a continuous conversion of the embryogenic cells into the nonembryogenic type. To maintain the embryogenic potential of the cultures and to ensure that plant regeneration is from somatic embryos only, the embryogenic calli must be carefully selected and transferred every subculture (I.K. Vasil 1987), an added burden to the already fastidious task of initiating and maintaining such cultures. The occurrence of more than one kind of embryogenic callus in diploperennial teosinte embryo cultures is not unique among Gramineae. Early recognition and isolation of appropriate callus types are becoming increasingly important in the establishment of successful long-term regeneration protocols. In maize, for example, regeneration from long-term embryogenic callus and suspension cultures is dependent on the production of a soft, friable, embryogenic (Type-ll) callus. Recognition and isolation of this particular kind of callus, first described by Green (1982, 1983), was paramount in the production of transgenic maize plants (Fromm et al. 1 990; GordonKamm er al. 1990). Likewise, the identification of a callus type for long-term maintenance and regeneration in wheat (Type C, Redway er al. 1 990a) was essential to the establishment of regenerable cell suspension cultures (Redway et al. 1 990b) and totipotent protoplasts (V. Vasil et al. 1990). Using Type C embryogenic callus as the regeneration system was the key to the successful production of transgenic plants in wheat (V. Vasil era/. 1992). By analogy to the embryogenic callus types formerly described in wheat, the more common compact callus found in diploperennial teosinte embryogenic cultures was termed Type A. Like its wheat namesake, teosinte Type A embryogenic callus occurred at a relatively high frequency, and always in conjunction with nonembryogenic

PAGE 64

57 growth (Fig. 3.6). The tendency to degenerate easily into nonembryogenic callus made strict selection at every subculture essential to a problematic long-term maintenance. Following the same terminology, the second kind of embryogenic callus found in embryogenic cultures of diploperennial teosinte was designated Type C (Fig. 3.7). Usually found at very low frequencies, teosinte Type C embryogenic callus was easy to maintain in long-term cultures and regenerated freely even after several years in vitro, characteristics that it shared with both maize Type-ll and wheat Type C embryogenic calli. Unlike any of these, however, it was not derived from any preexisting or aged callus; it was isolated directly from the proliferating immature embryos. Structurally comparable to wheat, teosinte Type C callus was compact, nodular and scutellar in nature (rather than soft and friable, like maize Type-ll). The most striking feature of teosinte Type C callus, however, was the absence of nonembryogenic proliferation, even in aging cultures, a trait that is not shared by either maize or wheat embryogenic calli (Fig. 3.7). Scanning electron microscopy revealed the compact convoluted callus to be nothing but a large mass of continuously proliferating scutella. Provided with a well differentiated epidermal layer, this callus spurted numerous cup-shaped structures at the surface (Fig. 3.8), comparable to those found in other embryogenic grass cultures, where they have been described as similar to the scutellum in both morphology and structure (V. Vasil and I.K. Vasil 1981; Lu and I.K. Vasil 1982). Type C embryogenic callus could be maintained indefinitely by subculture, without any further organization taking place. When transferred to a regeneration medium, however, the cup-shaped structures became more prominent (Fig. 3.9), and a well defined shoot apex, surrounded by a coleoptilar ring, formed at their center (Fig. 3.10). The ring further developed into a more or less well defined coleoptile, from which the first leaves emerged as the newly formed embryo precociously germinated. Structures showing aberrant morphology, an obvious consequence of the hormonal imbalance in culture,

PAGE 65

58 were not uncommon (Fig. 3.11). Scutella were often lobed, or formed small leafy appendages at the margins. Sometimes the whole scutellum became a leafy structure with characteristic trichomes on the surface, or several scutella fused laterally to form one single ridge of tissue associated with multiple embryonic apices. Individual small segments of embryogenic callus could be separated and transferred independently to an appropriate regeneration medium. When kept under light they greatly enlarged, the whole fragment becoming a prominent white scutellum with a deep red coleoptile (Figs. 3.12 and 3.13) from which the first green leaves emerged (Figs 3.14 and 3.15). A developing primary root, protruding from a well defined coleorhiza, was perfectly visible in these germinating somatic embryos (Fig. 3.15). Embryo developmental stage . Embryogenic callus cultures in Gramineae are often initiated from immature embryos, normally at a specific early stage of development (Lu et al. 1983). Cultures have also been obtained, however, from fully mature embryos and whole caryopses (e.g., McDaniel et al. 1982; Mehta era/. 1982; Botti and I.K. Vasil 1983; Rao et al. 1985). Since it is not predictable how any individual species will perform in culture, an experiment was conducted to find out the particular sequential pattern, if any, in the capacity to produce embryogenic callus from embryos of diploperennial teosinte. The influence of embryo length on the formation of embryogenic callus is illustrated in Fig. 3.16. The relative frequencies of both embryogenic and nonembryogenic calli formed, as influenced by embryo size, are shown in Table 3.1 . All experiments were done in MSC medium with 3% sucrose and 1 0 2,4-D. Cell proliferation resulting in the formation of calli occurred at all size ranges. Frequencies varied from a low of 33% for the smaller embryos (0-1 mm range) to a high of up to 90% in the upper size ranges (size ranges 2-3, 3-4 and 4-5 mm were not

PAGE 66

59 100 o 80 -I cc O 60-| CO => < o o LU CD o > cc CD LU 40 20 1 I Total Embryogenic Callus 8888881 Type C Embryogenic Callus m T T 0-1 1-2 2-3 3-4 EMBRYO LENGTH (mm) 4-5 Fig. 3.16 Influence of embryo length on embryogenic callus formation from immature embryo cultures of diploperennial teosinte. Table 3.1 Influence of embryo length on callus formation from immature embryo cultures of diploperennial teosinte. Embryo length (mm) NE callus only* Type A E callus* Type C E callus* Total E callus* Total callus* 0-1 33.3 a 0.0 a 0.0 a 0.0° 33.3 a 1-2 0.0 b 66.7 b 10.0 b 76.7 b 76.7 b 2-3 70.0° 16.7 a 0.0 a 16.7 a 86.7 bc 3-4 80.0° 10.0 a 0.0 a 10.0* 90.0 C 4-5 76.7 C 0.0 a 0.0 a 0.0° 76.7 bc Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter are not significantly different (P < 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980). Each medium contained: MS basal salts and vitamins 5% coconut milk 3% sucrose 10 /jM 2,4-D. Embryo length as tested.

PAGE 67

60 significantly different at the 95% confidence level). Production of embryogenic calli, however, was restricted to a much narrower developmental range. Embryogenic callus was produced at a high frequency (77%) by embryos in the 1 -2 mm length range only. The same type of callus was also observed in embryos 2-3 and 3-4 mm long; their frequency of formation, however, was not significantly different from zero (95% confidence level). No embryogenic callus was ever produced by the smallest (< 1 mm) or the largest (> 4 mm, mature) embryos. Type C embryogenic callus was produced at a relatively low frequency (10%) and only by embryos in the 1-2 mm size range (Fig. 3.16, Table 3.1). The occurrence of nonembryogenic callus alone was frequent in the larger size cultures (> 2 mm) and null in the 1 -2 mm embryos. Germination was common in all embryos larger than 2 mm and occasional in the 1-2 mm range. Since embryos smaller than 1 mm and larger than 2mm gave mostly unsatisfactory results, they were not used in any of the subsequent experiments. Auxin type and concentration . Problems associated with the use of 2,4-D in tissue culture investigations (e.g., Saunders and Bingham 1975; Collins et al. 1978; Hangarter et al. 1 980) led some authors to look for possible substitutes that would be more effective and less toxic in vitro. Having in view the successful use of some alternative compounds (Vian 1976; Collins et al. 1978; Hangarter et al. 1980; Hanning and Conger 1982; Conger er al. 1983), experiments were made to compare the effect of 2,4-D with two other herbicides with strong auxin-like properties, dicamba and picloram. Naphthaleneacetic acid, which has been used alone or in combination with 2,4-D in plant tissue cultures, was also studied as a possible substitute. Four commercially available indoleacetylamino acid conjugates, considered to be more stable sources of auxin than the labile free IAA and less toxic than the persistent synthetic auxin analogs (Hangarter er al. 1 980) were also tested. The results are summarized in Table 3.2 and

PAGE 68

61 Fig. 3.17. All experiments were done in MSC medium with 3% sucrose, with 1-2 mm embryos. The frequency of embryogenic response varied considerably with both the type and the concentration of the compound being used. No response was ever obtained in the absence of an auxin in the medium, neither was it in the presence of the IAA conjugates. Relatively high concentrations (25 /l/M) of naphthaleneacetic acid induced only moderate proliferation (35%) in the explants, and the callus formed in such conditions showed no embryogenic potential. This callus was also formed at 10 /jM, but at a very low frequency (15%, statistically not different from zero at the 95% confidence level). No response of any kind was obtained at any of the lower concentrations (Table 3.2). All three auxin herbicides (2,4-D, dicamba and picloram) promoted callus formation at all concentrations tested, and both embryogenic and nonembryogenic calli were produced under their influence. The results of Duncan's Multiple Range test using the arc-sine transformed data show that the frequencies of production of nonembryogenic callus alone were lowest at a concentration of 10 fiM, and were significantly lower (10% and 3%, as compared to 20% for 2,4-D) in media containing dicamba and picloram, respectively (Table 3.2). The same concentration of 10 iiM yielded the highest frequencies of embryogenic callus production, this time significantly higher (80% and 87%, as compared to 70% for 2,4-D) in media containing dicamba and picloram, respectively (Table 3.2, Fig. 3.17). The capacity to minimize nonembryogenic and maximize embryogenic callus production clearly makes this the most effective concentration to use when initiating embryogenic callus cultures in diploperennial teosinte. The shape of the fitting curves further demonstrates that 2,4-D is relatively less effective at all concentrations, when compared to either dicamba and picloram (Fig. 3.17). The frequency of response is significantly higher for both dicamba and picloram (95% confidence level) at optimal levels, and at least comparable to 2,4-D at

PAGE 69

62 Table 3.2 Influence of auxin type and concentration on callus formation from immature embryo cultures of diploperennial teosinte. Auxin NE callus Type A Type C Total Total only* E callus* E callus* E callus* callus* 0 0.0 h 0.0 d 0.0 C 0.0 h 0.0 e 2.5 0.0 h 0.0 d 0.0 C 0.0 h 0.0 8 h 1 A A NAA 5 0.0 h o.o d 0.0° 0.0 h o.o e 10 15.0 f 9h o.o d o.o c 0.0 h 15.0^ 25 35.0 bcd o.o d o.o c 0.0 h 35.0 cd 0 0.0 h o.o d o.o c 0.0 h 0.0 e 2.5 20.0 def 9 o.o d o.o c o.o h 20.0° de 2,4-D 5 30.0 cdef 36.7 b 0.0° 36.7 de 66.7 b 10 20.0 de, 9 66.7 a 3.3 bc 70.0 b 90.0 a 25 66.7 a o.o d 0.0° 0.0 h 66.7 b 0 0.0 h o.o d 0.0° 0.0 h 0.0« 2.5 1 6 jefgh 23.3 bc 0.0 C 23.3 e, 9 40.0° Dicamba 5 40.0 bc 23.3 bc 20.0° 43.3 d 83.3 a 10 10.09 h 66.7 a 13.3 ab 80.0° 90.0 a 25 50.0 b 10.0 cd 0.0° 10.09 h 60.0 b 0 0.0 h 0.0 d 0.0 C 0.0 h 0.0* 2.5 6.79" 23.3 bc 3.3 bc 26.7* 33.3 cd Picloram 5 33 3 cde 36.7 b 20.0 a 56.7 C 90.0 a 10 3.39 h 76.7 a 1 0.0 abc 86.7 a 90.0 a 25 10.09 h 16.7 cd 0.0° 26.7 cd Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter are not significantly different (P < 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980). Each medium contained: MS basal salts and vitamins 5% coconut milk 3% sucrose Auxin (and concentration) as tested. All embryos in the 1-2 mm range.

PAGE 70

Fig. 3.17 Influence of auxin type and concentration on embryogenic callus formation from immature embryo cultures of diploperennial teosinte. Light gray: Total embryogenic callus Dark gray: Type C embryogenic callus

PAGE 71

AUXIN CONCENTRATION (\\M) (log scale)

PAGE 72

65 suband supraoptimal concentrations, implying a higher overall effectiveness of both dicamba and picloram at all concentrations (and, in particular, at the lower ones). It also suggests a lower toxicity of both dicamba and picloram at the high end. A slightly, but significant, better overall response to picloram, as compared to dicamba (Table 3.2) would suggest that this is the auxin of choice for the initiation of embryogenic callus cultures in diploperennial teosinte. Type C embryogenic callus was induced at the optimal concentration (10 yM) by all three auxins. It was also produced at suboptimal concentrations in the presence of either dicamba or picloram. This would also indicate, and add further support to the idea, that these regulators are more effective and represent an improvement over the use of 2,4-D in the induction of somatic embryogenesis. Osmoticum . High levels of exogenous sucrose have been found to increase the frequency of somatic embryogenesis in tissue cultures of maize (Lu et a/. 1 983; Tomes 1986; Close and Ludeman 1987). Since there are indications that the same may apply to diploperennial teosinte (Swedlund and Locy 1988), an experiment was conducted to compare the effect of various levels of sucrose in the medium on the capacity to produce embryogenic callus from immature zygotic embryos of diploperennial teosinte. The results of these experiments are summarized in Table 3.3 and Fig. 3.18. All experiments were done in MSC medium containing 1 0 /jM 2,4-D. All embryos were in the 1 -2 mm range. The concentration of sucrose in the medium had only a moderate effect on the nature and efficiency of callus formation, although it had a marked repercussion on both the growth rate of the newly formed calli and the germination of the embryos. The production of nonembryogenic callus alone was significantly decreased with increasing sucrose concentrations, to the point of being virtually nonexistent at the higher concentrations (Table 3.3). The same general trend happened to the total callus

PAGE 73

66 100 ] Total Embryogenic Callus | Type C Embryogenic Callus 80 O 6040 20 X T 2 6 SUCROSE CONCENTRATION (%) 12 Fig. 3.18 Influence of sucrose concentration on embryogenic callus formation from immature embryo cultures of diploperennial teosinte. Table 3.3 Influence of sucrose concentration on callus formation from immature embryo cultures of diploperennial teosinte. Sucrose NE callus Type A Type C Total Total (%) only* E callus* E callus* E callus* callus* 3 16.7 a 63.3 a 6.7 a 70.0 a 86.7 a 6 6.7 ab 80.0 a 3.3 a 83.3 a 90.0 a 12 3.3 b 66.7 a 0.0 a 66.7 a 70.0 b Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter are not significantly different (P < 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980). Each medium contained: MS basal salts and vitamins 5% coconut milk 10/jM2,4-D Sucrose concentration as tested. All embryos in the 1-2 mm range.

PAGE 74

67 production, which was significantly lower at 12% than at 3% or 6%. The level of sucrose in the medium, however, did not have as much influence on the quality and quantity of embryogenic callus produced (Table 3.3). No significant differences (95% confidence level) were found in the induction of embryogenic callus at the different sucrose levels, although a trend seems to exist favoring the intermediate level of 6% (Table 3.3, Fig. 3.18). Embryogenic calli formed at higher sucrose concentrations were more compact and opaque and less convoluted than those formed at the lower concentration of 3%. They also developed at a considerably slower rate (as judged by qualitative visual observation) than that observed at the lower concentration level. In addition, cultures induced in high sucrose did not survive unless later subcultured to a low osmoticum medium (3% sucrose). Embryo germination, which was normally accompanied by nonembryogenic callus proliferation from the nodal region, was greatly reduced at the concentration of 6% and totally suppressed at 1 2%. L-proline . The reported stimulation of somatic embryogenesis in maize by the addition of L-proline (Armstrong and Green 1982, 1985; Green etal. 1983) prompted an experiment to evaluate the effect of L-proline on the induction of somatic embryogenesis from immature zygotic embryos in diploperennial teosinte. The results of this experiment are summarized in Table 3.4 and Fig. 3.19. All experiments were done in MSC medium containing 3% sucrose and 10 /iM 2,4-D. All embryos were in the 1 -2 mm range. At the single concentration (12 mM) tested no significant (95% confidence level) improvement was found in the embryogenic response as a result of the addition of Lproline. A trend seemed to exist, though, indicating that higher levels of L-proline might have helped prevent nonembryogenic callus formation, and promote embryogenesis (Table 3.4). That this might be the case is partially indicated by results obtained when testing the influence of L-proline in long-term callus maintenance (see below).

PAGE 75

68 100 80 < cc O 60 -| CO _l _l O 40 g z LU CD O > DC 00 LU 20 Total Embryogenic Callus Type C Embryogenic Callus T 0 12 PROLINE CONCENTRATION (mM) Fig. 3.19 Influence of proline on embryogenic callus formation from immature embryo cultures of diploperennial teosinte. Table 3.4 Influence of proline on callus formation cultures of diploperennial teosinte. from immature embryo Proline (mM) NE callus only* Type A E callus* Type C E callus* Total E callus* Total callus* 0 12 10.0 a 3.3 a 73.3 a 56.7* 3.3 a 30.0 a 76.7 a 86.7 a 86.7° 90.0 a Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter are not significantly different (P < 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980). Each medium contained: MS basal salts and vitamins 5% coconut milk 3% sucrose 10/JM2.4-D Proline as tested. All embryos in the 1-2 mm range.

PAGE 76

69 Combined effects . After making these refinements in the culture conditions, a final experiment was conducted to reexamine the effect of ail factors that had been tested, when combined at their optimum levels, in a single medium and to compare their integrated performance to their individual effect. The results of this experiment are summarized in Table 3.5 and Fig. 3.20. The composite medium, hereafter referred to as PIC1, was MSC containing 6% sucrose, 10 /l/M picloram and 12 mM L-proline. All embryos were in the 1-2 mm range. Combining all optimized factors in a single medium did not improve the overall performance of the system, as compared to each optimized factor taken individually. Total callus, Type A embryogenic callus and total embryogenic callus production showed no significant differences (95% confidence level) in media containing the optimized factors individually or in combination (Table 3.5). Type C embryogenic callus was formed at a low frequency and only in the L-proline-containing medium and in the composite medium (Table 3.5). None was produced in the media containing sucrose or picloram at the optimal levels. This is in apparent contradiction with what was formerly observed when testing the effect of both sucrose and picloram (Tables 3.3 and 3.2, respectively). This unique type of callus, however, always tended to occur at very low frequencies. This was particularly true in the experiments that tested the individual effect of both those factors, where its frequency was often not significantly different from zero (Tables 3.2 and 3.3). Its production, however, might be promoted by the presence of L-proline (Tables 3.4 and 3.5). Despite the relatively low occurrence of Type C embryogenic callus, its nature and characteristics suggested that it would be the ideal kind for long-term culture experiments, since it completely avoided the tedious, albeit essential, selection associated with every subculture in the more common Type A callus (I.K. Vasil 1987).

PAGE 77

70 100 80 Total Embryogenic Callus Type C Embryogenic Callus £ 60 CO => 40 20 — T— SUCROSE (6%) X L-PROLINE (12 mM) —J— PICLORAM (10 uM) COMPOSITE Fig. 3.20 Influence of the optimized factors, used singly or in combination, on embryogenic callus formation from immature embryo cultures of diploperennial teosinte. Table 3.5 Influence of the optimized factors, used singly or in combination, on callus formation from immature embryo cultures of diploperennial teosinte. Treatment NE callus only* Type A E callus* Type C E callus* Total E callus* Total callus* Picloram, 10/l/M 6.7 a 83.3 a 0.0 a 83.3 a 90.0 a Sucrose, 6% 6.7 a 76.7 a 0.0 a 76.7 a 83.3 a Proline, 12 mM 6.7 a 70.0 a 10.0 b 80.0 a 86.7 a Composite 6.7 a 80.0° 6.7 b 86.7 a 93.3 a Percentage of immature embryos that formed calli. Each value is the mean of three replications (ten embryos per treatment for each replicate). Values within each column marked with the same letter are not significantly different (P <. 0.05) by Duncan's Multiple Range test (Steel and Torrey, 1980). The composite medium contained: MS basal salts and vitamins 5% coconut milk 6% sucrose 10 fjM picloram 12 mM proline All embryos in the 1-2 mm range.

PAGE 78

71 Therefore, all experiments reported hereafter were performed with Type C embryogenic callus only. Callus Maintenance The optimized medium developed for culture initiation from immature embryos (PIC1) also proved adequate for callus growth. In addition to examining the effect of this medium on long-term maintenance, however, the use of other compounds known to influence the growth rate and viability of cultures was also investigated. These studies were prompted by the report of Duncan er a/. (1985), which indicated that an increase in the levels of reduced nitrogen, together with the addition of glucose to the medium (medium "D", Duncan er a/. 1985), were particularly beneficial in the initiation and growth of maize tissue cultures. Preliminary experiments indicated that, as with maize, such combination could also be helpful in the long-term maintenance of regenerable callus from diploperennial teosinte. As a result of these preliminary trials a new medium evolved that promoted superior growth of the embryogenic callus cultures. Much like PIC1, this medium (termed PIC2) was based on Murashige and Skoog's basal salts and vitamins and contained both L-proline and picloram, at the same levels as before (12 mM and 10 fjM, respectively). Coconut milk, however, was replaced by the addition of 500 mg/l casamino acids to the medium and sucrose alone (at 6%) was substituted by a combination of both sucrose (2%) and glucose (1%). Once the qualities of this medium were recognized, the importance of the various components was then evaluated, to determine their individual effect on the long-term growth of the embryogenic callus cultures. Auxin type and concentration . Since both dicamba and picloram had proven superior to 2,4-D at inducing embryogenic callus cultures from immature embryos, an experiment was performed to determine the relative efficiency of the three auxins at

PAGE 79

72 maintaining embryogenic callus growth. The results are summarized in Table 3.6 and Figs. 3.21 and 3.22. All experiments were done using PIC2 medium, containing 1 0 ijM of either 2,4-D, dicamba or picloram as the auxin source. Regression analysis of the growth curves demonstrated a significant linear relationship between fresh weight and time for the three compounds under study (Fig. 3.21, Table 3.6). All three auxins (2,4-D, dicamba and picloram) promoted callus growth at the concentration tested (10 jl/M). The slopes of the linear regression lines, however, indicate that, once again, 2,4-D was relatively ineffective at sustaining growth when compared to both dicamba and picloram (Table 3.6, Fig. 3.22). At this concentration, callus growth rates were significantly higher for both dicamba and picloram (95% confidence level). A significantly better response to picloram, as compared to dicamba, suggested that this again was the best auxin source for the regular maintenance of embryogenic callus cultures in diploperennial teosinte. In addition to the effect on callus growth rate, supplementing the medium with either dicamba or picloram also resulted in suppression of both differentiation and callus browning that tended to occur towards the end of the subculture period (20-30 days) when 2,4-D was used as the auxin source. A second experiment was conducted to determine whether the auxin concentration that best promoted callus initiation was also the best to support callus growth. The results of this experiment are summarized in Table 3.7 and Figs. 3.23 and 3.24. All experiments were done using PIC2 medium, with varying picloram concentrations. Regression analysis of the growth curves again demonstrated a significant linear relationship between fresh weight and time for all concentrations (Fig. 3.23, Table 3.7). Picloram promoted callus growth at the whole range of concentrations tested. The Gaussian fit indicated that the optimal concentration for sustained proliferation was 10 /7M (Table 3.7, Fig. 3.24). At this concentration embryogenic calli maintained the

PAGE 80

73 TIME (days) Fig. 3.21 Embryogenic callus growth curves (fresh weight) with various auxins. Table 3.6 Regression analysis for growth with different auxins. Auxin (1 0 yM) Linear Regression* R 2 2,4-D 895.1 + 18.1x a 0.993 Dicamba 899.0 + 69.9x b 0.999 Picloram 847.4 + 1 55.7x c 0.999 Based on data points from five replications (ten callus pieces per treatment for each replicate). Slopes marked with the same letter are not significantly different (P < 0.05). Each medium contained: MS basal salts and vitamins 12 mM L-proline 500 mg/l casamino acids 20 g/l sucrose 10 g/l glucose 10 /jM 2,4-D, dicamba or picloram, as tested 2 g/l Gelrite pH 5.8

PAGE 81

74 |j 140 2,4-D DICAMBA AUXIN TYPE PICLORAM Fig. 3.22 Effect of varying auxin type on embryogenic callus growth in diploperennial teosinte (see also Table 3.7). Each medium contained: MS basal salts and vitamins 12 mM L-proline 500 mg/l casamino acids 20 g/l sucrose 1 0 g/l glucose 10pM 2,4-D, dicamba or picloram as tested 2 g/l Gelrite pH 5.8 Data points represent the mean (and SE) of five replications (ten callus pieces per treatment in each replicate).

PAGE 82

75 TIME (days) Fig. 3.23 Embryogenic callus growth curves (fresh weight) with various picloram concentrations. Table 3.7 Regression analysis for growth with various picloram concentrations. Picloram (/l/M) Linear Regression* R2 2.5 1185.9 + 15.1 a 0.942 5 1190.8 + 66.6 b 0.998 10 887.9 + 156.6° 0.999 20 1023.8 + 60. 1 d 0.998 30 1148.0 + 33.6° 0.996 Based on data points from five replications (ten callus pieces per treatment for each replicate). Slopes marked with the same letter are not significantly different (P s 0.05). Each medium contained: MS basal salts and vitamins 12 mM L-proline 500 mg/l casamino acids 20 g/l sucrose 10 g/l glucose Picloram concentrations as tested 2 g/l Gelrite pH 5.8

PAGE 83

76 160 2.5 5 10 20 40 PICLORAM CONCENTRATION (mM) (log scale) Fig. 3.24 Effect of varying picloram concentration on embryogenic callus growth in diploperennial teosinte (see also Table 3.8). Each medium contained: MS basal salts and vitamins 12 mM L-proline 500 mg/l casamino acids 20 g/l sucrose 1 0 g/l glucose Picloram concentrations as tested 2 g/l Gelrite pH 5.8 Data points represent the mean (and SE) of five replications (ten callus pieces per treatment in each replicate).

PAGE 84

77 characteristic creamy white color and convoluted surface and could be easily subcultured for indefinite periods of time, without ever forming nonembryogenic friable callus. Both suband supraoptimal concentrations led to reduced callus growth. At 2.5 11M the proliferating scutella increased in size, and differentiation of leaf-like appendages, often covered with trichomes, occurred at a low frequency. Red pigmentation, due to anthocyanin biosynthesis, was also occasionally observed. Both events indicated that at this concentration picloram was no longer capable of completely supressing differentiation. At the highest concentrations growth was also reduced. At 30 /jM callus pieces gradually changed color to a deeper shade and eventually died. No friable callus was ever formed at any of the concentrations tested, which shows the peculiarity of embryogenic callus Type C. Callus proliferation was never observed in the complete absence of auxin. Carbohydrates . Claims that glucose, in addition to sucrose, contributed to increased proliferation in maize embryogenic callus cultures (Duncan et al. 1 985) led to the design of an experiment to determine the effect of different combinations of glucose and sucrose on diploperennial teosinte callus growth. The results of this experiment are summarized in Table 3.8 and Figs. 3.25 and 3.26. All experiments were done using PIC2 medium, with varying sugar combinations. As in the former experiments, regression analysis of the growth curves confirmed a significant linear relationship between fresh weight and time for all sugar combinations (Fig. 3.25, Table 3.8). Glucose alone (3%) was unable to support adequate growth. In addition to the poor proliferation rate, callus pieces changed to a darker hue and became soft, in evident contrast with the harder, healthier calli formed in all sucrose containing media. Of these, the results of linear regression clearly indicated

PAGE 85

78 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days) Fig. 3.25 Embryogenic callus growth curves (fresh weight) with various carbohydrate combinations. Table 3.8 Regression analysis for growth with different carbohydrate combinations. Carbohydrate(s) Linear Regression* R2 Glucose (3%) 1119.3 + 15.8x a 0.988 Glucose (2%) + Sucrose (1 %) 998.4 + 60.2x b 0.977 Glucose (1 %) + Sucrose (2%) 853.6 + 156.0x c 0.998 Sucrose (3%) 1038.1 + 86.8x d 0.997 Based on data points from five replications (ten callus pieces per treatment for each replicate). Slopes marked with the same letter are not significantly different (P < 0.05). Each medium contained: MS basal salts and vitamins 12 mM L-proline 500 mg/l casamino acids Carbohydrate combinations as tested 10 ijM picloram 2 g/l Gelrite pH 5.8

PAGE 86

79 160 .g 140 "§) 120 100 80 60 40 20 1 GLUCOSE (1%) SUCROSE (2%) 1 GLUCOSE (3%) 1 GLUCOSE (2%) SUCROSE (1%) SUCROSE (3%) CARBOHYDRATES Fig. 3.26 Effect of varying carbohydrate combinations on embryogenic callus growth in diploperennial teosinte (see also Table 3.9). Each medium contained: MS basal salts and vitamins 12 mM L-proline 500 mg/l casamino acids Carbohydrate combinations as tested 10 /iM picloram 2 g/l Gelrite pH 5.8 Data points represent the mean (and SE) of five replications (ten callus pieces per treatment in each replicate).

PAGE 87

80 that the best combination for optimal growth was obtained with the addition of glucose (1%) to sucrose (2%) (Table 3.8, Fig. 3.26). This sugar combination, the same used in medium "D" of Duncan et a/. (1985), almost doubled the culture growth rate, when compared to the more commonly used plain sucrose (3%) in the medium. Other than the variation in growth rate, no other noticeable differences were seen on the physical appearance of embryogenic calli grown in different sucrose containing combinations. No plausible reason is currently known for the apparent beneficial effect of the glucose-sucrose sugar combination. Casamino acids . Substantial growth increases were reported for maize tissue cultures when casein hydrolysate levels were increased in the media (Duncan er a/. 1985). The same promotional effect was also observed in diploperennial teosinte callus cultures when casein hydrolysate was used as a replacement for coconut milk (Swedlund and Locy 1988). An experiment was thus conducted to determine whether adding or omitting casamino acids (a vitamin-free product of casein hydrolysate) from the medium had any effect on embryogenic callus cultures of diploperennial teosinte. The results of this experiment are summarized in Table 3.9 and Figs. 3.27 and 3.28. All experiments were done using PIC2 medium, with or without casamino acids. As usual, fresh weight varied linearly with time (Fig. 3.27, Table 3.9). Linear regression of the data for the two growth curves confirmed a significant role for the presence of casamino acids in the medium: the addition of 500 mg/l casamino acids alone accounted for an almost doubling in the growth rate of the cultures (Table 3.9, Fig. 3.28). The convoluted surface of the embryogenic calli was also more prominent in the presence of casamino acids, which suggests that in addition to the quantitative effect on proliferation and growth, the product might also have an influence on the embryogenic process per se.

PAGE 88

81 0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days) Fig. 3.27 Embryogenic callus growth curves (fresh weight) with and without casamino acids. Table 3.9 Regression analysis for growth with casamino acids. Casamino Acids (mg/l) Linear Regression* R2 0 1126.3 + 87.0x a 0.999 500 832.3 + 159.0x b 0.998 * Based on data points from five replications (ten callus pieces per treatment for each replicate). Slopes marked with the same letter are not significantly different (P s 0.05). Each medium contained: MS basal salts and vitamins 12mM L-proline 0 or 500 mg/l casamino acids, as tested 20 g/l sucrose 10 g/l glucose ^0^lM picloram 2 g/l Gelrite pH 5.8

PAGE 89

82 CASAMINO ACIDS (mg/l) Fig. 3.28 Effect of casamino acids on embryogenic callus growth in diploperennial teosinte (see also Table 3.10). Each medium contained: MS basal salts and vitamins 12mM L-proline 0 or 500 mg/l casamino acids as tested 20 g/l sucrose 1 0 g/l glucose 1 0 /l/M picloram 2 g/l Gelrite pH 5.8 Data points represent the mean (and SE) of five replications (ten callus pieces per treatment in each replicate).

PAGE 90

83 L-proline . Although not statistically supported, the greatest mean frequency of Type C embryogenic callus production occurred when 12 mM L-proline was added to the medium (Tables 3.4 and 3.5). Since observations suggested that had higher concentrations been used better results might have been obtained, an experiment was conducted to determine the effect of increasing levels of L-proline on the embryogenic callus cultures. The results of this experiment are summarized in Table 3.10 and Figs. 3.29 and 3.30. All experiments were done using PIC2 medium, with varying L-proline concentrations. Again, regression analysis showed a significant linear relation between fresh weight and time for all concentrations (Fig. 3.29, Table 3.10). L-proline was not essential for callus growth, as moderate proliferation was obtained even when it was completely absent from the medium. Addition of the amino acid, however, produced a dramatic increase in the growth rate with increasing concentrations, with rates up to four times those measured for the control (Table 3.10, Fig. 3.30). Regression analysis of the slopes demonstrated a nonlinear increase in growth rate with increasing L-proline concentrations. The fitting curve was suggestive of a Michaelis-Menten equation, indicating that a factor (probably an enzyme or group of enzymes involved in membrane transport or cellular metabolism of L-proline) became limiting at higher concentrations, no further appreciable gain being then obtained by increasing the amino acid level in the medium (growth rates at 50 mM and 100 mM were not significantly different at the 95% confidence level). Comparable responses were reported in maize for the initiation of friable embryogenic callus (Armstrong and Green 1985) as well as fresh and dry weight increases and frequency of somatic embryo formation (V. Vasil and I.K. Vasil 1986) in the presence of proline. Noteworthy was the fact that, in this experiment, faster growth was not correlated with a concurrent increase in embryogenic performance. Improved embryogenesis, subjectively judged by the level of convolution (proliferating scutella) on

PAGE 91

84 6000 1000 -L, , , 1 1 1 1 1 1 1 1 , 1 ^ 0 2 4 6 8 10 12 14 16 18 20 22 24 26 TIME (days) Fig. 3.29 Embryogenic callus growth curves (fresh weight) with various proline concentrations. Table 3.10 Regression analysis for growth with proline. Proline (mM) Linear Regression* R 2 0 1147.4 + 54.0x a 0.997 12 845.2 + 147.9x b 0.998 25 923.7 + 158.9x9 0.997 50 750.7 + 186.3x d 0.997 100 640.5 + 194.0x d 0.989 Based on data points from five replications (ten callus pieces per treatment for each replicate). Slopes marked with the same letter are not significantly different (P < 0.05). Each medium contained: MS basal salts and vitamins 500 mg/l casamino acids 20 g/l sucrose 10 g/l glucose 10 /jM picloram 2 g/l Gelrite Proline concentrations as tested pH 5.8

PAGE 92

85 o H 1 1 1 1 • 1 1 • r 0 20 40 60 80 100 PROLINE CONCENTRATION (mM) Fig. 3.30 Effect of varying L-proline concentration on embryogenic callus growth in diploperennial teosinte (see also Table 3.1 1). Each medium contained: MS basal salts and vitamins 500 mg/l casamino acids 20 g/l sucrose 1 0 g/l glucose 1 0 /jM picloram 2 g/l Gelrite Proline concentrations as tested pH 5.8 Data points represent the mean (and SE) of five replications (ten callus pieces per treatment in each replicate).

PAGE 93

86 the callus surface and the visual appearance of the calli, both conditions that cannot be translated in numbers, was observed with increasing concentrations of L-proline up to 25 mM, at which the best compromise between growth and embryogenesis was attained. At 50 mM the callus turned whiter and became smoother and at 100 mM, regardless of the superior growth rate, all calli were no more than big growing white compact balls of tissue, without any surface features that would suggest somatic embryogenesis. When challenged for regeneration, these otherwise healthy calli completely failed to produce plants. The fact that callus growth and embryogenesis were greatest at different L-proline levels indicates that the two phenomena are under independent control, and suggests that conditions conducive to fast proliferation and growth are not necessarily good for the attainment of an optimal embryogenic response. From this experiment a new medium (PIC3) arose, identical in composition to PIC2 but containing 25 mM L-proline. This medium promoted both excellent growth and embryogenesis in callus cultures of diploperennial teosinte. All subsequent work involving the maintenance of these cultures was thus performed with PIC3. Plant Regeneration Lowering the auxin concentration or simply ommiting it from the medium is the only basic requirement to induce regeneration from grass embryogenic callus cultures (I.K. Vasil 1987). As formerly indicated this was done upon completion of every experiment. Callus pieces were taken from the medium containing the best combination of factors and tested for regeneration, to assess if the component(s) that had optimized growth also had any effect on the expression of totipotency. This was done by placing pieces of embryogenic callus in a regeneration medium consisting of Murashige and Skoog's major and minor nutrients and vitamins with 100 mg/l myo-

PAGE 94

87 inositol, 1 ijM picloram and 2.5 /iM kinetin, and visually evaluating the potential to produce plantlets in vitro. This simple medium produced generally good results. However, because preliminary observations suggested that the presence of either auxins or cytokinins in the regeneration medium might not be needed at all, an additional experiment was conducted to determine the effect, if any, of these two kinds of substances. Plants were regenerated with comparable ease from all auxin/cytokinin combinations, even their total absence. Over a period of 3-4 weeks the cup-shaped structures on the callus surface increased in size and turned bright white, due to starch accumulation. Small (3-5 mm) callus pieces often turned into single deformed scutella from which a coleoptile soon developed. At this stage normal germination occurred, with the first green leaves growing through the coleoptile tip and the radicle protruding from a well developed coleorhiza. Various numbers (generally 1-10) of somatic embryos were produced from larger callus pieces. Presence of auxin was not needed for regeneration, and it did not influence the average number of plantlets (normally 5-1 5) that could be obtained per dish (5 callus pieces/dish). It did, however, delay somatic embryo germination by approximately a week. Plantlets formed in the presence of auxin usually had fewer deformities (e.g., leaf curling) than plantlets regenerated without it, a fact of no consequence as all plants soon grew out of epigenetic anomalies. The average number of plantlets formed per dish was also not affected by the type of cytokinin employed. In the presence of BA (N 6 -benzyladenine), however, the white scutella tended to become green and resembled leafy structures, often with trichomes. This greening of the scutellum, which was particularly evident in the absence of auxin, had no obvious effect on the morphology of the plantlets so derived. Anthocyanin pigmentation was present in all treatments as scattered dots on the scutellar structures or as a deep red color on the mature coleoptiles.

PAGE 95

88 Embryogenic callus cultures kept in optimized media were still regenerable after a period of four years in vitro and, although not challenged for regeneration at that time, were still growing vigorously six years after being placed in culture. As cultures aged, the requirement for auxin did not change, but the need for a cytokinin in the medium became more pronounced. BA was the most effective for that purpose. At levels that varied from 2.5 /l/M (initial concentration for all cytokinins tested) up to 5.0 /vM (needed in 4 year old cultures), addition of N 6 -benzyladenine to the regeneration medium consistently promoted formation of plantlets from callus cultures, even when other cytokinins no longer had an effect at comparable concentrations. Discussion and Conclusions Donor tissue . One of the most important success factors in the initiation of regenerable cultures in Gramineae is the appropriate choice of the original explant. The use of immature tissues which still maintain meristematic activity and competence has been essential to the development of cell and tissue culture systems in this family of plants. Immature embryos and young inflorescences and leaves are the most commonly used sources of explants for initiation of embryogenic cultures in the grass family (I.K. Vasil 1987). Under the cultural conditions of this study, however, immature leaves and inflorescences of diploperennial teosinte, although at the deemed appropriate stages of development, failed to initiate any embryogenic callus growth, indeed almost any kind of callus growth at all. Production of embryogenic callus seems to be confined to particular stages of development, even when meristematic tissues are concerned. The existence of these temporal and developmental "competence windows" has been demonstrated in many cases and is particularly well documented in Gramineae. For example, immature leaf

PAGE 96

89 explants containing or being close to the apical meristem, as well as older, fully differentiated tissues, are generally unable to proliferate and give rise to embryogenic calli. This gradual decline in embryogenic competence, that is not necessarily related to the loss of mitotic activity (Joarder er a/. 1986; Taylor and IX Vasil 1987), has been demonstrated in a number of cereals and other grasses (Wernicke and Brettell 1980; Haydu and IX Vasil 1981; Lu and IX Vasil 1981; Wernicke ef a/. 1981; Hanning and Conger 1982; Alfinetta er a/. 1983; Ho and IX Vasil 1983; Wernicke and Milkovits 1984; Joarder ef a/. 1986; Linacero and Vazquez 1986a; Wenzler and Meins 1986; Rajasekaran ef a/. 1987b; Pareddy and Petolino 1990; Songstad et al. 1992), including diploperennial teosinte (Swedlund and Locy 1988). From these and other studies it became clear that older tissues lose their capacity to respond in culture as they differentiate and lose their meristematic potential. However, sections containing or being close to the apical meristem will also not proliferate, a condition somewhat more difficult to explain. As in competent explants embryogenic calli are normally formed in association with developing vascular tissues (I.K. Vasil 1985), it has been suggested that the transport of growth regulators through the vascular system could be instrumental in achieving the delicate hormonal balance that controls tissue competence and, therefore, embryogenesis. Since apical areas lack a differentiated vascular system, endogenous regulators are transported mostly by cell to cell diffusion and are probably partially metabolized and never reach the adequate balance needed for the development of competence in culture (I.K. Vasil 1987). In the present study, in the few cases where any proliferation was obtained in immature inflorescences and young leaves, that proliferation was indeed topological^ related to the location of the sectioned parallel veins. No major growth of any kind, however, was obtained in any of the treatments, regardless of the relative position of the explant in relation to the shoot apical meristem.

PAGE 97

90 Accumulating evidence indicates that, assuming the developmental status of the explant is adequate, it is possible to obtain callus and regeneration from any plants deemed recalcitrant by basically varying the chemical and physical environment of the explant until the right combination of factors is achieved (Morrish et al. 1 987). Relatively simple changes in media nutrient composition and hormone levels, for example, have led to spectacular increases in response in some otherwise recalcitrant genotypes (Lu era/. 1983; Duncan era/. 1985; Hanzel ef al. 1985). The environmental conditions under which donor plants are grown, however, are also critical in determining the level of competence of the explant, no matter how adequate it might intrinsically be (V. Vasil and I.K. Vasil 1982; Morrish et al. 1987). This is particularly true for greenhouse grown heliophytes, or field grown crops that are subject to some kind of cultural stress during the growing season. The natural conditions during growth (e.g., temperature, rainfall and photoperiod), as well as the cultural conditions (e.g., irrigation or the application of pesticides or fertilizers), are crucial in determining the physiological condition of the donor plant and, therefore, the response of the prospective explants in culture (Morrish ef al. 1 987). Lu ef al. (1 983), for example, showed that embryos obtained from the same cultivar but grown under different environmental conditions produced different responses in culture. They reported that ". . .there was more variability in the response of embryos from the same cultivar grown at different times than amongst embryos obtained from different cultivars of a single planting. . . ." (Lu etal. 1983). Such variation may be extreme to the point of the explants not responding at all. Failure to meet in full some physiological requirement of the explants, or of the original donor plants, would explain the setback in initiating embryogenic cultures from immature inflorescences and young leaf explants in this study, even when preexisting successful protocols developed for the same species (Swedlund and Locy 1 988) were

PAGE 98

91 followed. That further work is needed, however, is also indicated by the fact that these authors, too, reported only a meager (5%) percentage of success from such explants. To minimize the unavoidable variation in cultural conditions to which field grown plants are exposed, the use of the controlled environment of the greenhouse has been recommended (Lu et al. 1983; Armstrong and Green 1985). It has also been noted, however, that greenhouse conditions may be less than ideal (Lu et al. 1983). This was obvious in the present study, where immature embryos from plants grown in the open yielded a much better response in culture (as judged by qualitative visual observation) than embryos from greenhouse grown plants. Embryo developmental stage . Many factors affect the frequency of formation of embryogenic calli from immature embryos of diploperennial teosinte. The developmental stage of the embryo and the presence of a strong auxin in the medium were the most critical for success. Diploid perennial teosinte embryos at different stages of development had a distinct sequential pattern in the capacity to produce embryogenic callus. Very young embryos (< 1 mm) showed no response when cultured, whereas embryos with a differentiated scutellum that was already opaque and starting to accumulate starch (1-2 mm) produced the best embryogenic response. This gradually tapered off in larger embryos (2-4 mm), and completely disappeared in fully mature ones (> 4 mm). The differences found in the embryogenic response of immature embryos of diploperennial teosinte were identical to those recorded for other members of Gramineae (Green and Phillips 1975; Beckert 1982; V. Vasil and IX Vasil 1982; Lu era/. 1983; Maddock etal. 1983; Armstrong and Green 1985; Morrish etal. 1987). Again they indicate and confirm the existence of a competence window in developing embryos, comparable in all respects to that formerly discussed for young leaf explants. How the

PAGE 99

92 expression of competence is regulated is not known, but it certainly must involve steps of gene expression that are under developmental control. Auxin type and concentration . As more is learned about the factors affecting somatic embryogenesis, it becomes apparent that the type, concentration and time of application of plant growth regulators in the medium are all particularly critical to longterm success. In dicots, where exogenous auxin levels readily reach toxicity, determination of the lowest effective concentrations and shortest exposure times to the regulator has been a continuous concern (Ammirato 1989). Indoleacetic acid, the endogenous auxin, is mostly inefficient in the initiation of embryogenic callus cultures, probably because it is rapidly metabolized or destroyed by enzymatic oxidation in vivo (Good et al. 1982). Amide and ester conjugates of the native auxin have been identified in a variety of plant tissues. They seem to be involved in the protection of IAA from destruction by peroxidases (Andreae and Good 1955; Cohen and Bandurski 1978) and the detoxification of tissues in the presence of excessive IAA (Andreae and Good 1955; Good et al. 1956). The fact that both kinds of conjugates can be easily hydrolyzed under physiological conditions to release free IAA in vivo (Hangarter and Good 1 981 ) suggested their potential use as slow-release auxin sources for tissue cultures in vitro (Feung et al. 1 977; Hangarter et al. 1 980; Good et al. 1982; Pence and Caruso 1984). Prospective advantages of using such compounds included a continuous supply of near physiological levels of auxin, which would promote cell division (callus formation) without impairing future regeneration capabilities (Good et al. 1982), a problem often found with the more stable and more powerful synthetic analogs. IAA alone as an auxin source, however, does not elicit embryogenic response in most monocots, particularly in Gramineae (Rajasekaran ef al. 1987a). This would explain why explants were not responsive to the lAA-aminoacid conjugates, even when

PAGE 100

93 considering that they are more stable sources of auxin than the labile free IAA or the persistent synthetic IAA analogs (Hangarter ef a/. 1980). The same lack of responsiveness to lAA-conjugates was found by Rajasekaran ef al. (1 987a) in cultured leaf explants of Napier grass. Tissue cultures of Gramineae are unique in that a strong auxin is often the only growth regulator needed to elicit embryogenic response (I.K. Vasil 1987). Other than the native auxin, indoleacetic acid, most of the weaker synthetic analogs such as NAA, that yield better results than 2,4-D in dicot tissue cultures (e.g., Lazzeri ef al. 1987), are also ineffective when used alone to initiate an embryogenic response in members of the grass family. In the present experiments, only nonembryogenic callus was produced by competent immature embryos in the presence of NAA and only when concentrations were high. This agrees with the observation that whenever NAA is used in tissue cultures of Gramineae, it most commonly is in combination with some of the stronger auxins, such as 2,4-D (e.g., Haydu and I.K. Vasil 1981). The most commonly used compound for triggering embryogenesis in the majority of plants, and in particular Gramineae, is 2,4-D. A powerful selective herbicide with strong auxin activity, well tolerated by most monocots, 2,4-D has been associated, however, with toxicity and mutagenicity (Karp and Bright 1985; Karp 1989) and it is thought to be associated with the loss of regeneration capacity in long-term cultures (e.g., Smith and Street 1974; Saunders and Bingham 1975). Although these claims are probably unfounded, as attested by the wide variety of Gramineae where it has been, and it still is, successfully used to initiate and maintain stable regenerable tissue cultures (I.K. Vasil 1987), 2,4-D has been advocated as a possible generator of instability, and it has been indicated as good practice to limit its use in plant tissue cultures (Karp 1989) whenever alternatives can be found. Reports have shown that other herbicides with auxin-like properties, such as dicamba or picloram, can be used as successful alternatives to 2,4-D in establishing

PAGE 101

94 totipotent callus lines in grasses (Dudits et al. 1975; Vian 1976; Collins ef al. 1978; Conger era/. 1982, 1983; Hanning and Conger 1982; Irvine etal. 1983; Gray ef al. 1984; McDonnell and Conger 1 984; Duncan ef al. 1 985; Gray and Conger 1 985; Close and Ludeman 1 987; Hunsinger and Schauz 1 987; Papenfuss and Carman 1 987; Carman et al. 1988; Fitch and Moore 1990) and other monocots (Beyl and Sharma 1983; Phillips and Luteyn 1983; Zimmermann and Read 1986; Phillips and Hubstenberger 1987; Lu ef al. 1989). In some of these studies, where comparisons were made, both compounds had a beneficial effect over 2,4-D in the frequency of somatic embryos as well as embryogenic callus formation over a wide range of concentrations (e.g., Close and Ludeman 1 987; Fitch and Moore 1 990) and/or the longevity and regenerability of the cultures (e.g., Fitch and Moore 1990). The results of the present study indicate that both dicamba and picloram are more effective than 2,4-D in initiating and maintaining embryogenic callus cultures from zygotic embryos of diploperennial teosinte. Although the response curve is basically the same for all three compounds, both dicamba and picloram outperformed 2,4-D in initiating embryogenic cultures at suboptimal concentrations and were better tolerated at supraoptimal levels. At all concentrations, in particular at the optimal or near optimal concentration of 10 ^M, dicamba and picloram both promoted embryogenic callus formation and inhibited nonembryogenic growth significantly better than 2,4-D. In addition, both were more adequate than 2,4-D to the maintenance of continued longterm growth. Although basically equivalent to dicamba in the capacity to initiate embryogenesis, picloram was significantly better at suboptimal concentrations. Picloram also significantly outperformed all other auxins tested in maintaining long-term embryogenic cultures of diploperennial teosinte. It should thus be the auxin of choice whenever embryogenesis is to be studied in this grass species. Embryogenic callus cultures kept in optimized picloram-containing media were still regenerable after a

PAGE 102

95 period of four years in vitro and are still growing vigorously six years after being placed in culture. Osmoticum . Elevated osmolality can affect the induction of embryogenesis, as well as the maturation of the somatic embryos so derived. High osmotic conditions, created by elevated levels of sucrose or addition of hexitols, such as sorbitol, to the medium promoted both direct embryogenesis from immature zygotic embryos of sunflower (Finer 1 987) and secondary embryogenesis from the apical region of somatic embryos in soybean (Finer 1988). Increased sucrose concentrations promoted somatic embryogenesis in papaya (Litz and Conover 1983), Norway spruce (von Arnold 1987), sunflower (McCann etal. 1988), carrot (Kamada et al. 1988) and cucumber (Chee and Tricoli 1988). High sucrose also promoted microspore embryogenesis in canola (Dunwell and Thurling 1985) and, in combination with sorbitol and mannitol in the medium, increased embryogenic efficiency in long-term cultures of Vigna aconitifolia (Kumar er al. 1988). Elevated sugars also promoted more normal development of tissue-culture-derived carrot somatic embryos (Ammirato 1983) and helped prepare them for survival to dessication in synthetic seed production protocols (Kitto and Janick 1985). These results, and more, clearly indicate that the extracellular osmotic environment is critical in the induction and expression of somatic embryogenesis in plant tissue cultures. In Gramineae, increased osmolality improved the efficiency of somatic embryogenesis in maize embryogenic tissue cultures (Lu er al. 1983; Rapela 1984, 1985; Duncan etal. 1985; Tomes 1986; Close and Ludeman 1987). Swedlund and Locy (1988) used an intermediate sucrose concentration of 6% in the isolation of embryogenic callus cultures of diploperennial teosinte, which indicated that the same promotional effect of high sucrose in the medium might apply to this species as well. The results of the present study, however, don't support this view. Addition of up to

PAGE 103

96 1 2% sucrose to the medium did not significantly alter the frequency of 2,4-D induced embryogenesis or the site and mode of formation of embryogenic calli. It did affect, however, the nature of the embryogenic callus formed and its ability to grow and regenerate. This, however, was obviously no more than the expression of a physiological response to the osmotic stress imposed by the high sugar concentrations. Transferring callus pieces induced at high osmotica to a medium low in sucrose soon reestablished normal growth. A possible interpretation for the apparent lack of effect of high sucrose levels found in the present work is that, although the osmotic potential of the medium is an important component in the induction of the embryogenic response, it becomes critical only when other factors that affect the formation of embryogenic callus, such as the developmental and physiological state of the explants at the time of excision and culture, and the type and concentration of auxin in the medium, are not optimized. L-proline . Nitsch (1 977) and Sozinov er al. (1 981 ) first recorded the effect of L-proline as a specific stimulator of embryo formation, during androgenesis. Other reports have since confirmed a promotional role for L-proline on the induction or maintenance of embryogenesis in a number of plants (e.g., Nuti-Ronchi et al. 1984; Stuart and Strickland 1984a,b; Meijer and Brown 1987; Chibbar et al. 1987; Lu et al. 1989), including species of Gramineae (Armstrong and Green 1982; Sheridan 1982; Green er al. 1 983; Rapela 1 984, 1 985; Armstrong and Green 1 985; Earle and Gracen 1985; Fahey er al. 1986; V. Vasil and I.K. Vasil 1986; Trigiano and Conger 1987; Armstrong and Phillips 1988; Pareddy and Petolino 1990; Songstad era/. 1992). The physiological role of L-proline in plant tissue cultures is not fully understood. L-proline (and other small organic molecules), are often present in high concentrations in cells under certain situations of environmental stress, in particular those involving osmoregulation (Jefferies 1980; Le Rudulier er al. 1984). In halophytes, it was

PAGE 104

97 suggested that these compounds may serve as cytoplasmic osmotica, to balance the high vacuolar ion concentrations (Jefferies 1980). They might also help control the solubility of proteins and other biopolymers under the conditions of low water availability (Schobert 1977). Other possible roles suggested for L-proline are as a storage compound for reduced nitrogen (Britikov et al. 1970), or as a source of energy and reducing power (Stewart ef al. 1966). The fact that developing zygotic maize embryos contain high levels of free proline, as compared to other amino acids (Sheridan 1982), also led to the suggestion that the stimulatory influence of proline on somatic embryogenesis in maize could be related to the creation in the medium of a nitrogen metabolism environment similar to that found during zygotic embryogenesis in that species (Armstrong and Green 1985). Nuti-Ronchi et al. (1 984) reported that L-proline, when added in a wide range of concentrations to carrot hypocotyl cultures during early callus growth in the presence of 2,4-D, extended the duration and number of mitotic divisions and significantly stimulated the number of embryoids regenerated upon transfer to a hormone-free regeneration medium. They concluded that the increment in mitotic index observed in the presence of L-proline (and other additives) was connected with the formation of cells competent for embryogenesis; proembryonic masses, rather than plain callus, were formed as a result of mitotic divisions in the presence of the amino acid. V. Vasil and I.K. Vasil (1986) also showed, at concentrations that ranged from 5 to 160 mM, a concurrent beneficial effect of L-proline on growth (fresh weight, dry weight) and frequency of somatic embryo formation in maize friable (Type-ll) embryogenic callus cultures. As in the former reports, the results of the present study also indicated that both callus growth and embryogenesis were stimulated by additions of L-proline to the medium. The two processes, however, seemed to be under independent control; the levels of L-proline that optimized embryogenesis in the callus cultures were considerably lower than those that promoted proliferation and growth. A comparable

PAGE 105

98 effect of L-proline concentration was reported in maize for the frequency of both Type-ll callus and somatic embryo formation (Armstrong and Green 1985), as optimal levels of L-proline were of the same order of magnitude as those reported in the present study. In order for cells to differentiate, changes in gene expression, and hence cell phenotype, must occur during the cell cycle, so that new cells generated by subsequent mitotic divisions become different from the original mother cell, rather than mere copies of it. It seems logical to think that such a quantal cell cycle (Holtzer and Rubenstein 1977), involving the reactivation, as well as the shutting down, of specific genes whose activities dictate the final fate of the prospective new cell, must also occur during the onset of somatic embryogenesis. If this is so, it is conceivable that L-proline (as well as other compounds) at appropriate concentrations might be involved in the direct or indirect regulation of genes that contribute to the control of the embryogenic response. The suggestion of an indirect involvement of L-proline in gene regulation is not new (Nuti-Ronchi et al. 1984). The reason for the promotional effect on cell proliferation, however, could be quite different, a more aspecific nutritional effect being more difficult to rule out in this case. It is evident that the multiplicity of hypotheses regarding the mode of action of L-proline on tissue cultures is only a measure of our current ignorance of the system. Further research on the mechanism through which L-proline exerts its beneficial effect on somatic embryogenesis is obviously much needed. Casamino acids. The presence of casein hydrolysate has long been associated with embryogenesis, as early studies showed a strong promotional effect on growth and embryogenic response in carrot callus and suspension cultures as a result of its addition to the media (Syono 1965; Wetherell and Dougall 1976). The results of the present study fully agree with those early observations, as well as more recent ones that indicate that the same stimulatory effects also occur in grass tissue cultures (Dale et al.

PAGE 106

99 1981; Genovesi and Collins 1982; Green etal. 1983; Ozias-Akins and I.K. Vasil 1983a,b; Gray etal. 1984; Duncan etal. 1985; Gray and Conger 1985). The exact role of casein hydrolysate on embryogenesis is not known, although it probably acts primarily as a major source of reduced nitrogen. Eighteen different amino acids account for up to 73% of the total available nitrogen in casein hydrolysate (Sigma Chemical Co. analytical data). It is no surprise, then, that several investigators have tried to replace this preparation with some of its components, taken singly or in combination. No systematic studies, however, have been conducted to determine which factor, or combination of factors, would produce the same promotional effects. Contradictory reports exist in the literature, and amino acids found in casein hydrolysate have been variously described as promoting (e.g., L-alanine, L-glutamine, L-proline), inhibiting (e.g., L-glutamine, histidine, leucine, methionine) or having no apparent effect on growth and embryogenesis in a variety of plant species (Gamborg 1 970; Bailey et a/. 1972; Behrend and Mateles 1975; Wetherell and Dougall 1976; Stuart and Strickland 1984a). This suggests that different species may be responsive to different amino acids or combinations of such and clearly indicates again that more work is needed in this area. Conclusions . A successful protocol was developed that allowed high frequency induction of somatic embryogenesis and long-term maintenance of embryogenic calli from scutellar tissue of diploperennial teosinte immature zygotic embryos. The only critical factors for success were the developmental stage of the embryo (embryos 1 -2 mm long yielded the best results) and the inclusion of a strong auxin (at the appropriate concentration) in the initiation medium. For this purpose 10 /;M picloram was significantly better than all other auxin combinations tried. At the levels tested, increased sucrose concentration or inclusion of L-proline in the medium had little or no effect on the initiation of the embryogenic cultures, although observations indicate that

PAGE 107

100 levels of L-proline in excess of 12 mM might have a promotional effect on the induction of embryogenesis. Morphologically, embryogenic calli of diploperennial teosinte are comparable to those formerly described for the majority of Gramineae (I.K. Vasil 1987). An unusual type of embryogenic callus was isolated, though, which could be easily maintained through several cycles of subculture without ever being associated with the presence of soft, friable, nonembryogenic callus, even in aging cultures. These calli (termed Type C, by analogy with wheat), which were otherwise indistinguishable from the more common type (termed Type A, following the same terminology), were used to establish all the regenerable cultures used in this study. Although by no means essential, manipulation of the carbon sources in the medium, as well as addition of casamino acids and L-proline, greatly enhanced the embryogenic response and growth of the cultures that, under optimal conditions, readily maintained regeneration potential for a period of over 4 years.

PAGE 108

CHAPTER 4 PHENOTYPIC VARIATION IN TISSUE-CULTURE-DERIVED PLANTS OF DIPLOPERENNIAL TEOSINTE Introduction Once long-term regenerating embryogenic cultures of diploperennial teosinte became a practical reality, the purpose of the investigation turned to the evaluation of the phenotypic expression of regenerants from such cultures. This chapter reports the results of a detailed morphometric study that reflects the morphologies of living specimens of diploperennial teosinte, regenerated from tissue cultures through somatic embryogenesis, as they changed over time and across the barrier of sexual reproduction. Using routine univariate and multivariate statistical analyses, morphometric characters of the regenerants were compared and contrasted to those recorded from control plants (raised from wild collected seed) grown under the same cultural conditions over the same period of time. Detailed karyological observations in the tissue-culture-derived plants were used as well, to complement at the cytological level the results obtained from the morphometric studies. Materials and Methods Morphometric Analyses Populations of 20 to 30 individuals were regenerated from somatic embryos (see Chapter 3 for details) about every six months, for a period of three years. Only plants that could be unquestionably identified as originating from somatic embryogenesis were used. These were raised to maturity and visually assessed for variation. 101

PAGE 109

102 All plants derived from tissue culture during the first two years looked phenotypically normal. At about two and a half years, the presence of dwarf multitillering plants became increasingly frequent among the regenerants. At the end of three years, all plants regenerated from tissue culture were grossly abnormal. At that time, and for the next two years, the fate of morphometric variation in the regenerants was followed in three populations: (1) a greenhouse-grown population of 100 specimens, all derived from the three-year-old cultures; (2) a control population of 100 plants derived from wild collected seed [seeds courtesy of Prof. H.H. litis (H.H. litis and R. Guzman Accession No. 29115)]; and (3) a population of 100 plants, raised from selfing the regenerants. Observations and Variables The present study is based upon examination of 1 00 mature specimens from each of the three above mentioned populations. The plants were grown at a density of 9 pots/m 2 in a greenhouse of the Department of Horticultural Sciences, University of Florida, in Gainesville, Florida, USA, between November 1988 and November 1990. All plants were potted in gallon plastic containers, in Metromix Vegetable Plug Mix amended with 20% Perlite and fertilized once per growth cycle with Osmocote® (17-6-10 plus minor elements, 25 g per pot). Mean average daytime temperatures ranged from 20-25°C in winter to 30-35°C in summer. The temperature was not allowed to fall below 1 6°C during cold winter nights but it often went well above 35°C during hot dry spells in the summer. All measurements were made at complete maturity of the plants, just before the articulated ears started to break apart to disperse the caryopses. All specimens harvested for evaluation were cut at soil level and individually scored for a set of morphometric characters that included 21 quantitative and 5 qualitative (subjectively scored) traits, selected to provide a representative characterization of the external

PAGE 110

103 morphology of the plants (Table 4.1). They included, among others, many of the characters used by previous workers in ecological, evolutionary and taxonomic studies of this group of plants (Wellhausen et al. 1952; Wilkes 1967; Smith era/. 1981) as well as traits that are often reported to be affected by in vitro culture (e.g., male fertility). Size characters measured in centimeters were recorded to the nearest centimeter using a measuring tape. Those measured in millimeters were recorded to the nearest 0.1 mm using a dial caliper. Male fertility was assessed by collecting pollen directly from shedding tassels, on the first or second day after anthesis, between 09:00 and 10:00. The pollen grains were stained either without fixation, using 1% 2,3,5triphenyltetrazolium chloride (TTC) in phosphate buffer, pH 7.2 (Diakonu 1 961 ; Miller 1 982) or placed in 70% ethanol for later use. In the latter case staining for viability was done with either an iodine-potassium iodide (l 2 -KI) solution (Burnham 1982) or with aceto-carmine. All three staining methods worked equally well. A small tag with the plant identification number was always included in each pollen-containing vial. In all cases, to determine the degree of visibly aborted pollen, a drop of stain containing pollen grains in suspension was mounted on a microscope slide and covered with a cover slip. Viable pollen was opaque and stained dark, whereas nonviable pollen looked empty and did not stain. Since aborted grains tend to float to the outside, the entire slide was counted (Pittenger anf Frolik 1950). For every plant the results were processed to give the percentage of viable grains (number of stained pollen grains / total number of pollen grains x 100). All populations, with the exception of that obtained by selfing or sib-crossing the regenerants, were evaluated three times. Measurements were made in the spring of 1989 and again in the spring of 1990, when the plants flowered naturally under the influence of the short winter days. The last set of measurements was made in the fall of 1990, after the plants were induced to flower by being kept under artificial short-day

PAGE 111

104 Table 4.1 Characters used in the morphometric analysis of Zea diploperennis (diploperennial teosinte) populations. 1 . Plant Height, Main Stem (cm) 2. Plant Biomass (g/plant, aerial part) 3. Number of Tillers (total tillers per plant) 4. Average Tiller Size (percent of main culm) (0 = no tillers; 1 = 1-25%; 2 = 26-50%; 3 = 51-75%; 4 = 76-100%) 5. Number of Flowered Tillers 6. Percentage of Flowered Tillers 7. Leaf Sheath Color (0 = green; 1 = light pink; 2 = medium pink; 3 = red) 8. Leaf Sheath Length (third leaf from apex, main stem, mm) 9. Leaf Blade Length (third leaf from apex, main stem, mm) 1 0. Leaf Blade Width (third leaf from apex, main stem, mm) 1 1 . Leaf Area (0.75 x leaf length x leaf width, cm 2 ) 12. Leaf Ratio (leaf length/leaf width, main stem) 13. Leaf Blade/Sheath Ratio (leaf blade length/leaf sheath length, main stem) 1 4. Number of Internodes (main stem) 1 5. Internode Length (third internode below apical inflorescence, main stem, mm) 16. Internode Width (third internode below apical inflorescence, main stem, mm) 17. Internode Ratio (internode length/internode width, main stem) 18. Internode/Sheath Ratio (internode length/leaf sheath length, main stem) 19. Flowering Time (1 = early; 2 = mid-season; 3 = late) 20. Number of Primary Tassel Branches 21. Attitude of Tassel Branches (0 = no branches; 1 = oblique; 2 = horizontal; 3 = drooping) 22. Sex of Tassel Flowers (0 = female; 1 = both; 2 = male) 23. Length of Central Tassel Spike (mm) 24. Length of Longest Primary Tassel Branch (mm) 25. Tassel Ratio (length of longest primary tassel branch/length of central spike) 26. Male Fertility (pollen stainability, %)

PAGE 112

105 conditions (Emerson 1924; Mangelsdorf 1974). The population raised from selfed regenerants was evaluated only once, in the fall of 1 990. To determine if the observed variations were permanent or transient, gradually dissipating over time, a number of cycles of accelerated growth was induced by severely cutting back the plants several times between evaluations, to force the repeated production of new shoots from the thin, wiry, perennial underground rhizomes. Terminology In the statistical analyses that follow, the measurements made on the population raised from wild collected seed are collectively termed W (for wild). Following the recommendations by Chaleff (1981), measurements taken from the tissue-culture derived plants are collectively designated Ro (for Ro generation). Measurements made on progeny obtained by self-fertilization of the regenerated plants are termed Ri (for Ri generation). Statistical Analysis For every collection simple descriptive statistics, including means, standard deviations and coefficients of variation (CV) were calculated for each character in all of the groups studied (IV, R 0 , Ri) (Sokal and Rohlf 1969). The average coefficient of variation for each group was also computed, to permit comparison of univariate variability among groups. In addition, for every data set an analysis of variance was employed to test the significance of the differences between character means. Principal component analyses of all data sets were used to scrutinize variation among the morphometric characters and to provide a multivariate summary of morphological relationships among the groups. Canonical discriminant analysis was used to recognize the linear combination of the original variables that maximized the distances between the centroids relative to the variance within the groups, i.e., that linear combination of characters that provided maximum separation between the a priori

PAGE 113

106 groups (e.g., IV, R 0 , Ri). A stepwise discriminant analysis permitted to appraise the relative importance of the different characters in effecting that separation. As for principal components analysis, bivariate plots and/or histograms of the individuals' scores on the one or two axes permitted the visual assessment of the position of any specimen in this study. Computations were made to include in the output the average squared canonical correlations (ASCC, a measure of the discriminatory power of the function), the eigenvalues (representing the ratio of between-class variation to withinclass variation) and the Mahalanobis D 2 (or generalized distance between pairs of groups, a measure of the morphometric dissimilarity between the groups). Application of this extremely powerful statistical tool reduces the non-discriminating variability within groups while maximizing the discriminating variability among them. This absolutely requires prior identification of the specimens belonging to each group, and combined use of principal components analysis to define the a priori groups was occasionally necessary. All data processing was done using PC-SAS, version 6.04 (SAS Institute, Inc.). Input matrices for SAS and output tables and graphs for publication were created with Microsoft® Excel 4.0 (Microsoft Corporation) and MicroCal Origin™ 2.8 (MicroCal Inc.), respectively, running under Microsoft® Windows™ 3.1. Cytological Analyses Root tip squashes were used to determine the chromosome number in 1 0 wild type plants, the original mother plant, and all plants from the R 0 and Ri generations. Thirty five randomly selected R 0 and Rt plants were additionally scrutinized for major aberrations in chromosome morphology. Several standard protocols for the pretreatment, fixation and staining of grass chromosomes (Bhaduri and Ghosh 1954; Brown 1967; Chen 1968, 1969) were tested unsuccessfully. Acceptable results were finally obtained with an adaptation of Sallee's

PAGE 114

107 (1982) original protocol for maize chromosomes, as follows: apical root meristems were collected in March-April, between 10:00 and 1 1 :00, from greenhouse-grown plants of all populations. Vigorously growing plants were carefully unpotted and three to four centimeter long healthy roots were harvested, 4-5/plant, from the outside surface of the root ball. Selection of the proper collection time was critical for success. The root tips were immediately placed in a prefixation solution (consisting of the supernatant of an emulsion containing five drops of mono-bromonaphthalene, four drops of dimethylsulphoxide (DMSO) and 100 ml tap water) for one hour at room temperature. The sequence of addition of the chemicals during the preparation of the solution is critical (Sallee 1982). This solution is saturated in mono-bromonaphthalene, since this compound is basically insoluble in water. Selection of monobromonaphthalene and addition of DMSO were both important for the adequate spreading of the metaphase plates. Chromosomes failed to spread adequately when pretreated with other chemicals commonly used for the same purpose (colchicine, 8hydroxyquinoline) or when treated with mono-bromonaphthalene in the absence of DMSO. After pretreatment, the root tips were transferred into glacial acetic acid for overnight fixation at 4°C. Use of FAA or other common fixatives containing ethanol, or long-term storage of the root tips in 70% ethanol as recommended by Sallee (1982), inhibited later staining of the chromosomes. Roots could be stored for unlimited periods of time in glacial acetic acid without impairment of staining. Upon fixation the root tips were rinsed in tap water for 1 0 minutes, hydrolyzed for 1 4 minutes in 1 N HCI at 60°C and then placed into Feulgen stain at 4°C for several hours, until an intense purple coloration developed in the meristem. For successful staining hydrolysis time was crucial, staining time was not. Slides were made from root tip squashes dissociated in lacto-propiono-orcein and covered with a cover slip. Adequate metaphase plates were examined with a Zeiss

PAGE 115

108 Photomicroscope II equipped with 1 6x and 40x Neofluar objectives and photographed with a 100x planapochromatic objective in combination with a 1.25x Optovar. Three root tips were examined in each specimen, until a total of 1 0 chromosome counts per plant were obtained. Results Univariate Comparisons Means, standard deviations and coefficients of variation for both the experimental and control populations, and for the four harvesting times that constitute the basis for the current analytical study (Spring 1 989, Spring 1 990, Fall 1 990 and Ri generation) are presented in Tables 4.2, 4.4, 4.6 and 4.8, respectively. Likewise, the percent difference between means, as well as the F values (used to test the significance of the differences between character means) are shown in Tables 4.3, 4.5, 4.7 and 4.9, respectively. Spring 1989 . Highly significant differences in both vegetative and floral traits were encountered between the R 0 and the control (W) populations during the first flowering season following regeneration from tissue culture (Spring 1989, Tables 4.2 and 4.3). The means of R 0 and W differed highly significantly in 16 of the 18 vegetative characters and in all eight floral traits considered in this study. For five of the 16 vegetative characters, the means for the regenerants were significantly larger than for the control, with percent increases ranging from 1 1 .4 for average tiller size to almost 300 for the total number of tillers per plant. With the exception of leaf sheath color, which in all likelihood represents a genetically controlled trait that happened to have a value in the clonal population different from the mean for the complex control population, all traits showing increases (number of tillers per plant, number of flowered tillers, plant biomass and relative tiller size) were highly correlated, changes in one affecting, or

PAGE 116

109 Table 4.2 Statistics of location and dispersion for two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture (Ro, 100 specimens). Values for the Spring 1 989. Control (W) Regenerants {R 0 ) Character Mean 0 CV Mean a CV 1 178.12 31.84 17.87 64.83 9.49 14.64 2 167.26 30.85 18.44 240.59 46.54 19.35 3 8.42 2.88 34.19 33.34 7.17 21.51 4 3.59 0.55 15.38 4.00 0.00 0.00 5 7.03 2.42 34.46 18.50 5.32 28.77 6 84.87 15.23 17.95 56.34 12.97 23.02 7 1.04 0.58 56.22 2.60 0.80 30.92 8 92.05 15.83 17.20 73.22 2.44 3.33 9 310.10 61.60 19.86 186.09 35.59 19.12 10 22.30 3.59 16.08 22.25 4.53 20.38 11 51.65 12.59 24.38 31.12 9.08 29.17 12 14.36 4.17 29.05 8.72 2.45 28.13 1 T I o O.HO n 7R OO 1 ft 0 n 4Q 1Q 70 14 8.49 1.13 13.34 6.95 0.87 12.50 15 267.02 49.66 18.60 48.41 6.55 13.53 16 2.70 0.62 23.03 2.52 0.61 24.09 17 103.84 31.01 29.87 20.25 5.19 25.64 18 2.96 0.63 21 .26 0.66 0.09 13.33 19 1.84 0.85 46.17 1.00 0.00 0.00 20 1.03 0.92 88.85 0.00 0.00 21 0.92 0.80 86.96 0.00 0.00 22 1.42 0.50 34.93 0.84 0.37 43^86 23 114.33 20.24 17.71 46.55 20.63 44.32 24 66.61 50.85 76.34 0.00 0.00 25 0.56 0.42 74.90 0.00 0.00 26 93.22 4.83 5.18 0.00 0.00 Average CV 33.09 20.71

PAGE 117

110 Table 4.3 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture (R 0 , 100 specimens). Values for the Spring 1989, sorted by percent difference between means. Character Mean (W) Mean (Ro) % F 3 8.42 33.34 296.0 1040.12*" 5 7.03 18.50 163.2 384.67*" 7 1.04 2.60 150.0 246.25*** 2 167.26 240.59 43.8 172.46*** 4 3.59 4.00 11.4 55.12*** 10 22.30 22.25 -0.2 0.01 16 2.70 2.52 -6.7 4.34* 14 8.49 6.95 -18.1 116.38*" 8 92.05 73.22 -20.5 138.23*** 13 3.43 2.54 -25.9 96.37*** 6 84.87 56.34 -33.6 203.34*** 12 14.36 8.72 -39.3 135.96*** 11 51.65 31.12 -39.7 174.96*** 9 310.10 186.09 -40.0 303.87*** 1 178.12 64.83 -63.6 1162.83"* 18 2.96 0.66 -77.7 1309.33*" 17 103.84 20.25 -80.5 706.68*** 15 267.02 48.41 -81.9 1904.93*** 22 1.42 0.84 -40.8 88.10*** 19 1.84 1.00 -45.7 97.78*** 23 114.33 46.55 -59.3 549.92*** 20 1.03 0.00 -100.0 126.68*** 21 0.92 0.00 -100.0 132.25*** 24 66.61 0.00 -100.0 171 .59*** 25 0.56 0.00 -100.0 178.27*** 26 93.22 0.00 -100.0 37320.92*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; *** P < 0.001).

PAGE 118

111 being affected by, changes in any other. Particularly impressive is the influence of the strong multitillering habit on the total plant biomass, that more than compensates for the negative effect of the reduced plant height found in this population (see below). For the remaining 1 1 characters, the means for the regenerants were significantly smaller than those for the control, with the percent decrease ranging from 18.1 , for the number of internodes, to 81 .9, for internode length. Here, the characters showing the greatest shrinkage (internode ratio, internode/sheath ratio, plant height) are all related to the internode length. Of a lesser magnitude, but also highly significant are the leaf blade length and associated traits (leaf area, leaf ratio, leaf blade/sheath ratio), the leaf sheath length and the number of internodes. The percentage of flowered tillers was also significantly lower than in the control as was, to a much lesser extent, the internode width. In addition to the differences uncovered at the vegetative level, highly significant differences between character means were also evident in the floral expression of the regenerants (which, as formerly pointed out, was assessed exclusively by characteristics of the tassel). These ranged from the complete suppression of lateral branching (and, of course, all related variables, such as the attitude of tassel branches, length of longest tassel branch and tassel ratio) and pollen production to a reduction in the central spike length, accompanied by an intriguing shift toward feminization of the tassel flowers and a tendency for early flowering. In sum, all specimens originating from tissue culture, when compared to adequate controls of the same physiological age and grown under the same cultural conditions, could at this time be accurately described as dwarf multitillering plants with strongly reduced or totally suppressed male sexual expression. The reduced stature of the plants (63.6% less than the controls) was mostly due to shorter internodes (-81 .9%) rather than fewer internodes (-18.1%) and was associated with a corresponding reduction in leaf length. This was considerably more pronounced on the blade (-40.0%)

PAGE 119

112 than on the sheath (-20.5%). Since neither internode nor leaf widths were significantly affected in the regenerants, both stems and leaves were disproportionately short and broad when compared to the controls (internode ratio: -80.5; leaf ratio: -39.3%), which greatly accentuated the dwarf appearance of the plants. Female flowering, not statistically assessed in this study, did not seem impaired in any way. All Ro plants, as well as the respective controls, had a normal diploid chromosome complement of 20, and no obvious aberrations were found (Figs. 4.1 through 4.3). Spring 1990 . After roughly a year of forced accelerated regrowth (the plants were severely cut back four times) the same character set was measured again, in the same two original populations. Highly significant differences in both vegetative and floral traits were again apparent, although somewhat less pronounced, between the Ro and the control (W) populations during this season (Spring 1990, Tables 4.4 and 4.5). The means of Ro and W differed highly significantly in 1 5 of the 1 8 vegetative characters and in six of the eight floral traits under assessment. As in Spring 1 989, average leaf sheath color was significantly deeper in the regenerants than in the control population. This is not entirely unexpected since, as discussed before, this is conceivably a genetic trait characteristic of the regenerants that is here being contrasted to the much wider variability for the same character found in a natural population. Puzzling, however, was the fact that (as in Spring 1 989) there was no complete uniformity for this trait in the plants regenerated from tissue culture, as would certainly be expected in a clonal population for a character that is under strict genetic control. For three of the remaining 14 vegetative characters, the means for the regenerants were larger than for the control, at a highly significant level. AS in the previous harvest, the total number of tillers and the number of flowered tillers were both significantly higher in the regenerants, although to a much lesser degree than seen before (39.4% and 31 .3% more, respectively, as compared to 296.0% and 1 63.2% in

PAGE 120

c TO Q. h— O c o T5 C D) o co c 2 < £ a) CL 0) O
PAGE 121

1 14

PAGE 122

115 Table 4.4 Statistics of location and dispersion for two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture (Ro, 100 specimens). Values for the Spring 1990. Control (W) Regenerants (R 0 ) Character Mean a CV Mean a CV 1 145.75 26.05 17.87 77.73 4.04 5.20 2 271 .54 40.57 14.94 196.82 37.89 19.25 3 14.77 2.98 20.16 20.59 5.14 24.98 4 3.56 0.50 14.01 3.44 0.50 14.50 5 11.61 2.87 24.76 15.24 4.76 31 .23 6 78.78 12.41 15.76 72.90 6.62 9.08 7 1.08 0.61 56.88 2.44 0.50 20.45 8 78.62 16.37 20.83 72.10 4.27 5.93 9 260.18 56.74 21 .81 201.54 40.57 20.13 10 23.11 4.59 19.86 23.01 3.98 17.29 11 44.88 12.78 28.47 34.66 8.85 25.53 12 11.80 3.94 33.39 9.07 2.69 29.61 1 o O A-i d.4 1 0.91 ""70 26.73 2.80 0.54 19.46 14 8.51 1.49 17.56 8.97 1.34 14.99 15 221 .54 40.57 18.31 82.10 4.27 5.21 16 2.70 0.62 23.03 2.98 0.30 10.00 17 86.23 25.60 29.69 27.83 2.92 10.50 18 2.92 0.70 23.91 1.14 0.08 6.94 19 1.98 0.83 41.85 2.00 0.00 0.00 20 1.35 1.09 80.44 0.00 0.00 21 1.37 1.09 79.46 0.00 0.00 22 1.63 0.49 29.77 1.44 0.50 34.65 23 116.24 8.85 7.62 99.59 5.14 5.16 24 70.11 43.99 62.75 0.00 0.00 25 0.60 0.38 63.19 0.00 0.00 26 94.39 3.51 3.72 24.67 13.98 56.67 Average CV 30.65 17.58

PAGE 123

116 Table 4.5 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture (R 0 , 100 specimens). Values for the Spring 1990, sorted by percent difference between means. Character Mean (W) Mean (Ro) % F 7 1.08 2.44 125.9 295.34"* 3 14.77 20.59 39.4 95.92"* 5 11.61 15.24 31.3 42.63*** 16 2.70 2.98 10.4 16.02*** 14 8.51 8.97 5.4 5.24* 10 23.11 23.01 -0.4 0.03 4 3.56 3.44 -3.4 2.89 6 78.78 72.90 -7.5 17.48*** 8 78.62 72.10 -8.3 14.84*" 13 3.41 2.80 -17.9 33.88*** 9 260.18 201 .54 -22.5 70.68*** 11 44.88 34.66 -22.8 43.23"* 12 11.80 9.07 -23.1 32.79*** 2 271 .54 196.82 -27.5 181.18*" 1 145.75 77.73 -46.7 665.80*** 18 2.92 1.14 -61 .0 640.28*** 15 221.54 82.10 -62.9 1168.55"* 17 86.23 27.83 -67.7 513.70*** 19 1.98 2.00 1.0 0.06 22 1.63 1.44 -11.7 7.45** 23 116.24 99.59 -14.3 264.44*** 26 94.39 24.67 -73.9 2339.02*** 20 1.35 0.00 -100.0 154.54"* 21 1.37 0.00 -100.0 158.39*** 24 70.11 0.00 -100.0 253.97*** 25 0.60 0.00 -100.0 250.46*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *" P < 0.001).

PAGE 124

117 Spring 1989). Plant biomass, however, was no longer compensated by the larger number of tillers, neither was it by the slight increase in plant height observed in the regenerants, being now significantly lower (-27.5%) than in the controls (see below). The relative uniformity in tiller size found in the Ro population in Spring 1 989 was lost, as differences for that trait were no longer significant between the two populations, and internode width, marginally smaller for the regenerants in the former harvest, was now significantly larger than the controls. For the remaining 11 characters, the means for the regenerants were significantly smaller than those for the control, with the percent decrease ranging from 7.5 for the percentage of flowered tillers to 67.7 for the internode ratio. The characters showing the greatest reduction (internode ratio, internode/sheath ratio, plant height) were all related to the internode length once again. This was still considerably short, although to a lesser extent than before (62.9% shorter than the control, as compared to 81.9% in the previous spring). The same trend towards reduced dimensions was also found for the leaf blade length and associated traits (leaf ratio, leaf area, leaf blade/sheath ratio), the leaf sheath length and the percentage of flowered tillers. All these characters showed mean values significantly lower than in the control, but the differences between the two populations were all considerably less pronounced than in the previous measurements. The number of internodes was marginally higher in the regenerants. At the floral level, highly significant differences between character means were evident once again. As in the previous harvesting time, lateral tassel branches were completely absent in the R 0 population, which naturally affected the mean values for all associated traits (tassel ratio, length of longest tassel branch and attitude of tassel branches). Male fertility, now restored (which permitted the production of the flj), was greatly reduced (73.9% less than the control) as was the central spike length. This reduction in dimensions, however, was less pronounced than in the previous spring

PAGE 125

118 (-14.3%, as compared to -59.3%), which indicates that the overall morphology of the floral parts was following the same trend found in the vegetative body. The previously observed shift toward feminization of the tassel flowers was still apparent, but considerably less pronounced in this season, and the tendency for earliness was no longer present. In brief, all R 0 specimens measured in Spring 1990 showed the same characteristics described in Spring 1989, but in a more attenuated form. Plants were still multitillering and of a reduced stature and showed an aberrant floral morphology, albeit with a trend towards normality. As observed before, height reduction (plants were now 46.7% shorter than the control, as compared to 63.6% in Spring 1989) was again associated with a decrease in internode length (now 62.9% less in the Ro than in the control, as compared to 81.9% in the previous harvesting time) rather than fewer internodes; the number of internodes was now marginally higher in the R 0 than in the controls. As before, too, reductions in length were found in both the leaf blade and the leaf sheath in the R 0 \ those on the blade (-22.5%) were more pronounced than in the sheath (-8.3%). Both differences, however, were considerably less than those found the previous year (-40.0% and -20.5%, respectively). In addition, both internodes and leaves showed less distortion (internode ratio: -67.7%; leaf ratio: -23.1%, as compared to -80.5% and -39.3% in Spring 1 989) and male fertility, completely suppressed before, was now partially restored. The shift towards feminization of tassel flowers in the tissue culture regenerants, although still significant, was less pronounced than in Spring 1989, and the entire population did flower uniformly during the mid-season, rather than early in the season the year before. Fall 1990 . During the late spring and summer of 1 990 it became evident that a considerable recovery to normality was taking place in the tissue-culture-derived plants. Under the influence of the seasonable long days and good cultural practice, strong

PAGE 126

119 healthy growth was produced by all plants, that were additionally cut back twice before being artificially induced to flower in the fall. Then, roughly a year and a half after regeneration from tissue culture, the same two original populations were appraised one last time for the same character set. Not surprisingly, the dissimilarities between the R 0 and the control (W) population had decreased considerably over the summer months. Many of the vegetative traits (including number of internodes, plant biomass, leaf blade/sheath ratio, plant height, relative tiller size and number of tillers) or floral traits (such as length of the longest primary tassel branch and tassel ratio) that had shown significant differences to the controls in previous harvesting times, were now only marginally significant or not significant at all (Tables 4.6 and 4.7, Fall 1990). For these characters, tissue-cultureinduced change apparently ceased, or became imperceptibly small after a period of 18 months in cultivation. Highly significant differences between the means of Ro and W were encountered in 1 1 of the 1 8 vegetative characters and in five of the eight floral traits. Among the differences in vegetative traits, the most dramatic was leaf sheath color. For the first time ever, every single specimen in the whole population of tissue-culture-derived plants showed an absolutely uniform solid purple leaf sheath color, even on the plants that had not shown such coloration in previous harvesting times. This would further suggest that purple is the natural color for that organ in the clonal R 0 population. The full expression of the controlling genes probably was partially suppressed in some of the plants during the previous seasons. Four of the remaining 1 0 vegetative characters showed means that were larger in the regenerants than in the control at a highly significant level. Unlike in the two previous seasons, however, these were no longer related to the number of tillers or associated traits, such as the number of flowered tillers, which were now significantly smaller in the R 0 than in the controls (see below). Means significantly higher than those

PAGE 127

120 Table 4.6 Statistics of location and dispersion for two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture (Ro, 100 specimens). Values for the Fall 1990. Control (W) Regenerants (R 0 ) Character Mean o CV Mean a CV 1 1 49.56 30.21 20.20 150.81 7.28 4.83 2 289.43 30.41 10.51 297.71 27.22 9.14 3 16.29 2.71 16.65 15.22 3.30 21.67 4 3.50 0.50 14.36 3.44 0.50 14.50 5 13.44 2.64 19.62 A r\ r\r\ 10.80 3.35 30.98 6 82.74 10.33 12.49 69.66 8.29 11.90 7 1 .08 0.61 56.88 3.00 0.00 0.00 o 8 o r\ A A 80.1 1 -4 ~~r f~~\ o 17.23 21 .50 89.28 8.29 9.28 9 250.1 8 56.74 22.68 288.62 16.37 5.67 10 27.29 2.44 8.95 25.39 1.37 5.41 11 51.11 11.99 23.47 54.95 4.06 7.38 12 9.27 2.39 25.83 11.40 0.95 8.37 13 3.23 0.90 27.82 3.26 0.31 9.45 14 8.92 1.37 15.34 9.35 1.09 11.61 15 216.82 37.89 17.48 174.97 18.29 10.46 16 2.91 0.49 16.67 3.19 0.14 4.29 17 76.23 17.47 22.92 54.87 6.01 10.96 18 2.81 0.68 24.06 1.97 0.24 12.28 19 1.97 0.81 41.10 2.00 0.00 0.00 20 1.53 0.92 59.81 2.44 0.50 20.45 21 1.67 0.91 54.53 1.00 0.00 0.00 22 1.44 0.50 34.65 2.00 0.00 0.00 23 116.69 9.69 8.31 122.38 6.77 5.54 24 84.33 28.34 33.61 92.03 5.35 5.81 25 0.73 0.25 34.82 0.75 0.05 6.75 26 95.12 3.63 3.81 92.40 4.66 5.05 Average CV 24.93 8.91

PAGE 128

121 Table 4.7 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture {R 0 , 100 specimens). Values for the Fall 1990, sorted by percent difference between means. Character Mean (W) Mean (Ro) % F 7 1.08 3.00 177.8 976.86"* 12 9.27 11.40 23.0 68.61*** 9 250.18 288.62 15.4 42.37*** 8 80.11 89.28 11.4 23.01*** 16 2.91 3.19 9.6 30.35*** 11 51.11 54.95 7.5 9.20** 14 8.92 9.35 4.8 6.06* 2 289.43 297.71 2.9 4.12* 13 3.23 3.26 0.9 0.08 1 149.56 1 50.81 0.8 0.16 4 3.50 3.44 -1.7 0.72 3 16.29 15.22 -6.6 6.28* 10 27.29 25.39 -7.0 46.25*** 6 82.74 69.66 -15.8 97.50*** 15 216.82 174.97 -19.3 98.92*** 5 13.44 10.80 -19.6 38.40*** 17 76.23 54.87 -28.0 133.61*** 18 2.81 1.97 -29.9 136.53*** 20 1.53 2.44 59.5 76.23*** 22 1.44 2.00 38.9 126.00*** 24 84.33 92.03 9.1 7.13" 23 116.69 122.38 4.9 23.15"* 25 0.73 0.75 2.7 0.84 19 1.97 2.00 1.5 0.14 26 95.12 92.40 -2.9 21.19*** 21 1.67 1.00 -40.1 54.12*" Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 129

122 in the controls were found for leaf blade length (1 5.4% longer than the control) and the associated leaf ratio (+23.0%) and leaf area (+7.5%), as well as for leaf sheath length (+11.4%). Mean internode width was also significantly larger in the regenerants, in good accordance with the observations from the previous spring. Means for the regenerants were significantly smaller than those for the control for the remaining six characters, with the percent decrease ranging from 7.0 for the leaf blade width to 29.9 for the internode/sheath ratio. The characters showing the greatest reduction, internode/sheath ratio (-29.9%) and internode ratio(-28.0%), were again associated with the internode length (that was 19.3% shorter than the control). Differences for the number of flowered tillers and leaf blade width were somewhat less pronounced (albeit significant), but contrasted sharply with the values from the previous seasons when no reductions were found. The number of internodes and plant biomass were marginally larger and the number of tillers marginally smaller in the tissue-culturederived plants than in the control. The most dramatic differences, however, were found in the tassel. Highly significant differences were found for five of the eigth character means being compared. Of these, the means for the number and length of the primary tassel branches (produced for the first time in this season) as well as the means of the length of the primary tassel branch were significantly higher than the means for the control, in very sharp contrast with the values for all the previous harvesting times. In addition, all tassels produced exclusively male flowers, all traces of feminization having completely disappeared (female florets could still be found in some of the control specimens). Of the remaining traits, the attitude of the tassel branches was significantly different from the control, probably reflecting again a character under strict genetic control (there was absolute uniformity for the expression of this trait in the entire R 0 population). Male fertility was significantly lower than the control (but only by 2.9%) and

PAGE 130

123 the tassel ratio showed no significant differences. All Ro plants flowered uniformly during mid-season. In summary, the trend towards normality apparently suggested by the values measured during Spring 1990 was basically fully realized during Fall 1990, after a summer of exceptionally good growth. Ro specimens measured in Fall 1990 were normal or nearly normal (not significantly different from the control) for most vegetative and floral traits. For the first time since regeneration from tissue cultures, this now highly uniform population (average CV = 8.91, as compared to 20.71 and 17.58 for the previous two harvest seasons) could no longer be described as composed of dwarf, multitillering plants with an altered sexual male expression, but rather as a perfectly normal population whose main difference from the seed raised control resided in its greater uniformity in size, color and sexual expression. With plant height, leaf morphology and normal sexuality basically restored, differences, if any, pointed to a superior performance from the clone. Internode length, and its associated internode ratio and internode/sheath ratio, however, were still moderatly (but significantly) trailing behind the control. The Ri Population . The production of viable pollen by the R 0 plants during Spring 1 990 permitted for the first time the development of an R 1 (selfed regenerants) population. This new group of 100 plants was grown alongside with the original populations and was evaluated, in Fall 1990, for all the traits in the usual character set. A new control population of 1 00 individuals raised from wild collected seed was also grown, to ensure that the R 1 plants would be compared to controls of the same physiological age. A number of highly significant differences were still found between the R 1 and the control populations. Again, both vegetative and floral traits were affected ( Tables 4.8 and 4.9). The means of the R 1 and the control differed highly significantly in eight of

PAGE 131

124 Table 4.8 Statistics of location and dispersion for two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (IV, 100 specimens) and seed from selfed tissue culture regenerants (flj, 100 specimens). Control (W) Selfed Regenerants (fir) Character Mean o CV Mean a CV 1 187.51 41.75 22.27 188.77 42.32 22.42 2 244.38 54.27 22.21 251 .36 60.82 24.20 3 10.71 2.94 27.47 9.18 2.19 23.91 4 2.98 0.83 27.80 3.44 0.50 14.50 5 8.79 2.61 29.65 8.31 2.00 24.11 6 82.43 12.00 14.55 90.93 8.20 9.02 7 1.35 1.09 80.44 1.87 0.79 42.08 8 93.78 20.96 22.35 90.83 17.41 19.17 9 340.35 78.74 23.14 314.62 60.04 19.08 10 26.70 4.09 15.32 30.01 5.08 16.93 11 67.85 18.08 26.64 70.54 17.17 24.34 12 13.12 4.00 30.52 10.84 3.07 28.30 13 3.76 1.08 28.75 3.56 0.84 23.71 14 8.92 1.37 15.34 9.39 1.09 11.62 15 279.84 60.47 21.61 248.09 34.80 14.03 16 2.98 0.86 28.74 3.13 0.76 24.42 17 101.97 37.22 36.50 83.90 23.13 27.57 18 3.10 0.83 26.66 2.82 0.58 20.65 19 1.87 0.81 43.43 1.44 0.50 34.65 20 1.35 1.05 77.64 1.55 1.17 75.27 21 1.35 1.05 77.64 1.13 0.79 69.64 22 1.56 0.50 31.98 1.45 0.50 34.48 23 118.14 24.76 20.95 120.33 12.49 10.38 24 66.23 40.32 60.88 68.82 41.21 59.88 25 0.54 0.34 62.70 0.56 0.34 60.05 26 93.22 4.83 5.18 94.02 3.41 3.62 Average CV 33.86 28.39

PAGE 132

125 Table 4.9 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 1 00 specimens) and seed from selfed tissue culture regenerants (Ri, 100 specimens). Values sorted by percent difference between means. Character Mean (W) Mean (Ri) % F 7 1.35 1.87 38.5 15.03*" 4 2.98 3.44 15.4 22.62*** 10 26.70 30.01 12.4 25.77*** 6 82.43 90.93 10.3 34.22*** 14 8.92 9.39 5.3 7.21" 16 2.98 3.13 5.0 1.80 11 67.85 70.54 4.0 1.16 2 244.38 251 .36 2.9 0.73 1 187.51 188.77 0.7 0.04 8 93.78 90.83 -3.1 1.17 13 3.76 3.56 -5.3 2.29 5 8.79 8.31 -5.5 2.13 9 340.35 314.62 -7.6 6.75* 18 3.10 2.82 -9.0 8.03** 15 279.84 248.09 -11.3 20.71*** 3 10.71 9.18 -14.3 17.38"* 12 13.12 10.84 -17.4 20.43*** 17 101.97 83.90 -17.7 17.00*** 20 1.35 1.55 14.8 1.63 24 66.23 68.82 3.9 0.20 25 0.54 0.56 3.7 0.30 23 118.14 120.33 1.9 0.62 26 93.22 94.02 0.9 1.83 22 1.56 1.45 -7.1 2.43 21 1.35 1.13 -16.3 2.82 19 1.87 1.44 -23.0 20.35*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *" P < 0.001).

PAGE 133

126 the 18 vegetative characters, and only in one of the eight floral traits. The selfed regenerants, now highly heterogeneous as a result of genetic recombination, showed an average deeper color than the wild types. Not completely unexpected, this is probably a consequence of genetic drift, since the Ro mother plants from which the Ri originated were very dark colored. Of the remaining seven vegetative characters showing highly significant differences, three (relative tiller size, leaf blade width and percentage of flowered tillers) had means that were larger in the R 1 than in the control, with percent increases of 1 5.4%, 1 2.4% and 1 0.3%, respectively. The number of internodes was also marginally larger in the Ri than in the controls (+5.3%). Characters showing significantly lower means could be grouped in three classes: (1) one involving internode related dimensions, such as internode ratio (-17.7%), internode length (-11.3%) and internode/sheath ratio (-9.0%); (2) one involving leaf related dimensions, such as leaf ratio (-17.4%) and leaf blade length (-7.6%); and (3) the number of tillers (-14.3%). Of the floral traits only one, flowering time, was significantly different between the populations: specimens of the Ri tended to flower somewhat earlier than the controls. This is probably another consequence of genetic drift. Comparison of the average coefficients of variation for each group (Tables 4.2, 4.4, 4.6 and 4.8) also showed that the variability in the R 1 was within the limits found in seed-derived controls (average CV = 28.39, Table 4.8), an expected consequence of the population no longer being a clone. Hence, the R 1 can be described as a fundamentally normal population; the means for most vegetative characters and all floral morphological traits previously shown to diverge significantly between the R 0 and the control no longer differed significantly in the selfed regenerants. Just as for the R 0 , all plants of the ftj, as well as the respective controls, had the normal diploid chromosome complement for the species (2n = 2x = 20, litis ef a/. 1979), without any visible aberrations (Fig. 4.4).

PAGE 134

127 The results from previous harvests showed that, in general, the effect of time was apparently to reduce the magnitude of the differences between the tissue-culturederived plants and the corresponding controls. In more than one instance such differences were completely obliterated as a result of time alone. For those characters that time did not normalize in the R 0 , tissue-culture-induced change ceased in this generation (Ri). This further indicates that the variation previously found in the Ro is transient in nature and fails to be transmitted sexually. Differences affecting both internode and leaf length, however, seem to have somehow prevailed through the cycle of meiosis and syngamy. Although the differences found for these traits were now greatly attenuated, their apparent persistence in the Ri may indicate the existence of a condition that can be transmitted to the sexual progeny. Multivariate Comparisons Principal Components Analysis In using the exploratory capacities of principal components analysis, three approaches are of interest. First, it can be asked if there are any detectable differences between the component scores of R 0 (or fy) and their respective controls (W) for the four harvesting times, when the corresponding data sets are pooled and analyzed as a single group. Second, it can be asked to what extent the passage of time (or a sexual cycle) induces changes in the components themselves, by analyzing the four harvests separately and comparing the corresponding eigenvectors (factor loadings) and eigenvalues (expressed as percents of trace). Last, and because results from the univariate analysis apparently suggest a certain level of heterogeneity in the clonal population from tissue culture, it can be asked to what extent, if any, such heterogeneity does indeed exist, and what factor(s) are most involved in causing it.

PAGE 135

128 Analysis as a single group . Ordination on the first two principal components of individuals of Ro (or and W, when all harvesting times are analyzed as a single group, is shown in Fig. 4.5. The associated factor loadings and percents of trace are given in Table 4.10. Table 4.11 details both the vegetative and floral traits that load most significantly (P < 0.005) on the first two components, in decreasing order of magnitude. Mean scores for the first two principal components of individuals of R 0 (or Ri) and W are compared in Table 4.12, covering all harvesting times. Sixteen of the 26 variables are significantly loaded on PC I, which subsumes 39.2% of the variation in the data (Table 4.10). Among the most significant, and of like sign, are male fertility, internode length and plant height, denoting a positive correlation between male fertility and vegetative size, under the conditions of this analysis (Table 4.11). All other vegetative characters of like sign that load significantly on PC I are size factors, associated with either internode length (internode/sheath ratio and internode ratio) or leaf blade length (leaf blade length, leaf area and leaf blade/sheath ratio). Tassel-related traits, including the length of the central tassel spike, length of the longest primary tassel branch, tassel ratio and the number and attitude of the primary tassel branches also load significantly on PC I and are of like sign, indicating a positive correlation between vegetative size and degree of male expression (Table 4.11). The number of tillers and associated traits (number of flowered tillers), also significantly represented, have loadings with a negative sign implying a negative correlation between the number of tillers and both the size and fertility of the plants (Table 4.11). PC I, therefore, ordinates individuals on the basis of plant size, tassel differentiation and male fertility. Leaf sheath color, internode width and plant biomass load moderately heavily on the second principal component, as do the number of primary tassel branches, sex of tassel flowers, length of longest primary tassel branch and tassel ratio (Table 4.11).

PAGE 136

~ o c J= .£ Q) i2f • 4 "IS i§«" 0) a> c CO to c o 5 .a -6 -o -§ .2 w a> ob 3= 2 a> ej cd $ o co o c a> cd c tS E° E J! o 5 eg co 3_JL CO co C » co 03 = o o _ o = E * *o A X t5 "~ iro w c-g o E a5 re ^ c -2 = .2 3 "2 to £5 co -*— ' c CL CO TJ £ 2 w "g o ~ co f» CD C x> o o t* cd CD T3 il CO CD « s P o 0 ^ Qo O CO -2>

PAGE 137

130 •* CO
PAGE 138

131 Table 4.10 Factor loadings and percents of trace from principal components analysis of Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 1 00 specimens) and derived from tissue culture (Ro or Ru 100 specimens each). Data sets from all harvesting seasons, analyzed as a single group. Principal Component Character I II III 1 . Plant Height 0.271 -0.045 0.040 2. Plant Biomass 0.039 0.243 0.266 3. k 1 I _ _ r -r~" 1 1 _ Number of Tillers -0.262 0.130 0.038 4. Relative Tiller Size -0.109 -0.018 0.030 5. Number of Flowered Tillers -0.216** 0.063 0.124 6. Percentage of Flowered Tillers 0.173* -0.224 0.113 7. Leaf Sheath Color -0.125 0.305 -0.059 8. Leaf Sheath Length 0.147 0.206** -0.169* 9. Leaf Blade Length 0.236** 0.007 -0.396*** \ u. Leal Diaae wiain U. 1 1 o 0.145 11. Leaf Area 0.233** 0.086 -0.190* 12. Leaf Ratio 0.142 -0.089 -0.473*** 13. Leaf Blade/Sheath Ratio 0.159* -0.171 -0.283*** 14. Number of Internodes 0.112 -0.004 0.250*** 15. Internode Length 0.278*** -0.172 0.020 16. Internode Width 0.093 0.290*** -0.328*** 17. Internode Ratio 0.214** -0.330 0.171* 18. Internode/Sheath Ratio 0.244** -0.284 0.118 19. Flowering Time 0.078 -0.128 0.159* 20. Number of Primary Tassel Branches 0.205** 0.344*** 0.062 21. Attitude of Tassel Branches 0.199** 0.177* 0.118 22. Sex of Tassel Flowers 0.143 0.264*** 0.007 23. Length of Central Tassel Spike 0.243** 0.090 0.138 24. Length of Longest Primary Tassel Branch 0.237** 0.253*** 0.117 25. Tassel Ratio 0.231" 0.234** 0.125 26. Male Fertility 0.285*** -0.041 0.116 Percent of trace 39.2 10.0 8.5 Marked entries show variables that are significantly represented in each principal component (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 139

132 Table 4.1 1 Characters most significantly represented (P < 0.005) in the first two axes of a principal components analysis of Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 100 specimens) and derived from tissue culture (Ro or ftj, 100 specimens each). Data sets from all harvesting seasons, analyzed as a single group. Component Character Loading PC I (39.2%) Internode Length Plant Height Number of Tillers Internode/Sheath Ratio Leaf Blade Length Leaf Area Number of Flowered Tillers Internode Ratio 0.278 0.271 -0.262 0.244 0.236 0.233 -0.216 0.214 Male Fertility Length of Central Tassel Spike Length of Longest Primary Tassel Branch Tassel Ratio 0 285 0.243 0.237 0.231 PC II (10.0%) Leaf Sheath Color Internode Width Plant Biomass 0.305 0.290 0.243 Number of Primary Tassel Branches Sex of Tassel Flowers Length of Longest Primary Tassel Branch Tassel Ratio 0.344 0.264 0.253 0.234

PAGE 140

133 Table 4.12 Statistics of location and dispersion and results of analysis of variance comparing PC I and PC II scores in Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (IV, 100 specimens) and derived from tissue culture (Ro or Ri, 100 specimens each). Data sets from all harvesting seasons, analyzed as a single group. Season Control Regenerants Mean o Mean a Spring PCI 1.89 1.44 -6.20 0.55 2753.05*** 1989 PC II -1.65 1.63 0.43 0.65 140.12*** Spring PC I 0.92 1.32 -3.74 0.70 968.15*** 1990 PC II -0.49 1.58 -0.11 0.48 5.18* Fall PCI 1.08 1.04 1.03 0.40 0.18 1990 PC II 0.25 1.28 2.16 0.33 208.64*** Selfed PC I 2.68 1.55 2.34 1.28 2.84 Regenerants PC II -0.62 1.75 0.02 1.45 7.91" Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *" P < 0.001).

PAGE 141

134 PC II, that subsumes 1 0.0% of the variation in the data, thus ordinates individuals on the basis of those traits. A number of observations of biological relevance emerge from the analysis of both the tabulated data and the graphic modeling of the different plant populations. First, there is a distinct difference in statistical dispersion between the Ro population and the corresponding wild-type controls at all three harvesting times, no doubt a consequence of the clonal nature of the Ro, that becomes more pronounced as the population ages (Table 4.12, Fig. 4.5; cf., average coefficients of variation in Tables 4.2, 4.4 and 4.6). Confirming this interpretation is the fact that variability is fully restored to normal values in the R 1 (Table 4.12, Fig. 4.5; cf., average coefficients of variation in Table 4.8). Noteworthy, however, is the fact that a number of Ro plants stray from the pack during the first cultural season (Spring 1989). These plants diverge from the main R 0 group along both principal components axes, indicating that they would be different in size, color and/or male sex expression. This assumption, already hinted by observations of the univariate analysis (cf., Table 4.2), was fully confirmed by multivariate analysis of the R 0 plants alone (see below), which also permitted the physical identification of the variant plants. A second observation of biological pertinence is the fact that in all populations (including the controls) mean values for the morphometric characters drifted with time (Fig. 4.5). Significant differences were found between harvesting times in the mean PC I and PC II scores for the seed-derived population (W) (ANOVA: 13.06 < F < 73.16, P < 0.001 for all; mean PC I scores for Spring 1990 and Fall 1990 were not significantly different). This is not completely unexpected, and it represents the variation of a natural population as it gradually changes with both age and season. Similar significant seasonal differences were also found in the mean PC I and PC II scores for the tissueculture-derived populations (R 0 , Ri). These differences, however, were considerably more pronounced than those found in the controls (ANOVA: 44.80 < F < 3514.16,

PAGE 142

135 P < 0.001 for all). Tissue-culture-derived plants, thus, changed considerably more over time and across generations than would be expected on the basis of the performance of the controls (Fig. 4.5). The most striking observations of this study, however, become apparent when mean PC I and PC II scores for the tissue-culture-derived plants and the respective controls are compared pairwise for every harvest time (Table 4.12, Fig. 4.5). Differences in mean PC I scores between tissue culture regenerants and control plants were highly significant on the first two harvests, the magnitude of the difference being much greater for Spring 1989 than for Spring 1990 (Table 4.12). At both times mean PC I scores were much lower for the tissue-culture-derived plants than for the seed-derived controls. Such prominent differences, however, could no longer be identified either on the last harvest that involved the R 0 generation (Fall 1990) or on the R 1 (Table 4.12), a conclusion confirmed by inspection of Fig. 4.5. This, once again, distinguishes the original R 0 individuals as anomalous multitillering plants of reduced stature and fertility, with unusually short and broad stems and leaves, and modified tassel morphology. It also shows, beyond any doubt, that such anomalies become increasingly attenuated with time, completely disappearing in the sexual progeny. Evidently, one effect of time (or a sexual cycle) is to reduce the magnitude of the morphological dimorphism detected on PC I, a multivariate recapitulation of the previous univariate findings. Significant differences were also found, at all harvesting times, between regenerants and control plants with respect to their PC II scores. These differences were particularly pronounced in the third harvest season (Table 4.12, Fig. 4.5). The higher values of the mean PC II scores normally associated with the R 0 certainly do reflect the influence of leaf sheath color on the second principal component. This variable becomes increasingly important as differences in the quantitative morphometric characters gradually dissipate over time. Its weight becomes maximal during Fall 1 990

PAGE 143

136 when R 0 plants, then basically morphologically normal, turned out for the first time uniformly dark colored (highest score for the variable). Analysis as separate groups . Ordination on the first two principal components of individuals of Ro (or Ri) and W, analyzed as separate groups, is shown for all the harvesting times in Fig. 4.6. Comparison of factor loadings and percents of trace associated with the separate principal components analyses for each harvest time are presented in Table 4.13. Table 4.14 details both the vegetative and floral traits that load most significantly (P < 0.005) on the first two components, in decreasing order of magnitude. Mean scores for the first two principal components of individuals of Ro (or Ri) and W are compared in Table 4.15, covering all harvest times. For each principal component, considerable differences exist in the values of both character loadings and percents of trace for the four harvest times (Table 4.13). Seasonal variation in the most significant PC I and PC II eigenvectors (Table 4.14) reveals that the variables used to describe variation change with the passage of time, an indication that different characters change at different rates, over time and across generations. Size, fertility and tassel differentiation, which dominate the descriptive ability of PC I in the first two harvest seasons, become gradually less important, while other characters, such as color, come to dominate later in time (Table 4.14). Decreasing PC I eigenvalues, paralleled by increasing values for PC II and PC III indicate a progressive decline in total variation with the passage of time (or a sexual cycle). This, and the decreasing differences in mean PC I scores between tissue culture regenerants and control plants (Table 4.15, Fig. 4.6), clearly reflect an extensive convergence of traits between the R 0 and the seed-derived controls. Significantly different mean PC II scores between the R 1 and the controls reveal, however, that a slight (PC II accounts for only 12.1% of the total variation), but detectable, variation still remains in the selfed regenerants (Table 4.15). In good agreement with previous

PAGE 144

C-C (I) o ID C t ffi C 2 a O "D Si 0 •c _ 0) CO "O C CD c (0 0) ^ CL o Q. TJ go g) o "6 ~o CD CD (C tO «— » O E=i o o to * !| •cd to — 'to o »to" c c 2 o to to .£ "5 P Q. 6a c E • o © S3 E a 3 o F2 o 8 o> OS CD 5 OT £ c t: -ff 9o_ c a> 2 o m « o o> CD C .E E -2 "to f 1 i QS. CO CO £.C o o = o ffl o E tX O >to to ° E "S « CO o a> _a> cu « -Q 3 2 i co -c 5 to x to *= 0) > CO -»— ' c el's.!?. £| a> c tj o 2 8 i= TJ ° § 0) TJ to 1 2 Qo O co CO CD CD -* Ll

PAGE 145

138

PAGE 146

139 Table 4.13 Comparison of factor loadings and percents of trace from principal components analyses of Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 1 00 specimens) and derived from tissue culture (Ro or R 1t 100 specimens each). Data sets from all harvesting seasons, analyzed as separate groups. PC I S89 S90 F90 Ri 1. Plant Height 0.244" 0.277*** 0.032 0.126 2. Plant Biomass -0.186* 0.204** 0.030 -0.068 3. Number of Tillers -0.242" -0.202** -0.131 -0.189* 4. Relative Tiller Size -0.134 0.023 -0.027 -0.069 5. Number of Flowered Tillers -0.215" -0.158* -0.222** -0.213" 6. Percentage of Flowered Tillers 0.184* 0.078 -0.255*** -0.040 7. Leaf Sheath Color -0.198" -0.229** 0.304*** 0.059 8. Leaf Sheath Length 0.178* 0.099 0.213" 0.215" 9. Leaf Blade Length 0.226** 0.204** 0.279*** 0.252*** 10. Leaf Blade Width 0.007 0.024 -0.135 0.100 11. Leaf Area 0.202** 0.179* 0.216" 0.274*** i_t?di naiiu (J. loo 0.141 0.279 0.136 13. Leaf Blade/Sheath Ratio 0.170* 0.150* 0.040 0.033 14. Number of Internodes 0.159* -0.056 0.017 -0.040 15. Internode Length 0.249*** 0.285*** -0.204** 0.010 16. Internode Width 0.070 -0.044 0.266*** 0.314*** 17. Internode Ratio 0.222** 0.248*** -0.306*** -0.251*** 18. Internode/Sheath Ratio 0.239** 0.262*** -0.293*** -0.165* 19. Flowering Time 0.146 -0.011 -0.035 -0.092 20. Number of Primary Tassel Branches 0.197" 0.243** 0.259*** 0.333*** 21. Attitude of Tassel Branches 0.191* 0.240" -0.100 0.262*** 22. Sex of Tassel Flowers 0.157* 0.101 0.289*** 0.193* 23. Length of Central Tassel Spike 0.224** 0.253*** 0.108 0.271*" 24. Length of Longest Primary Tassel Branch 0.211" 0.267*** 0.138 0.326*** 25. Tassel Ratio 0.211" 0.264*** 0.093 0.282*** 26. Male Fertility 0.252*** 0.287*** -0.110 0.007 Percent of trace 55.5 39.8 25.1 17.3 (continued)

PAGE 147

Table 4.13 (Continued) 140 PC II Character S89 S90 F90 1. Plant Height 0.047 -0.053 0.267*** 0.015 2. Plant Biomass 0.255*" -0.148 0.006 0.042 3. Number of Tillers 0.177* -0.131 -0.109 0.237** 4. Relative Tiller Size 0.018 -0.145 0.034 -0.164* 5. Number of Flowered Tillers 0.151* -0.229** -0.029 0.201" 6. Percentage of Flowered Tillers -0.116 -0.296*** 0.125 -0.063 7. Leaf Sheath Color 0.284*** 0.153* -0.180* -0.059 8. Leaf Sheath Length 0.089 0.270*** 0.095 -0.066 9. Leaf Blade Length -0.104 0.280*** 0.234** 0.373*** 10. Leaf Blade Width 0.194* 0.174* 0.189* -0.339*** 11. Leaf Area -0.002 0.337*** 0.323*** 0.068 12. Leaf Ratio -0.161* 0.109 0.125 0.469*** 13. Leaf Blade/Sheath Ratio -0.206** 0.077 0.149* 0.381*** 14. Number of Internodes 0.015 -0.132 -0.095 -0.105 15. Internode Length -0.038 -0.105 0.369*** 0.259*** l b. internode wiatn -0.196 0.462 0.155* 0.028 17. Internode Ratio -0.054 -0.280*** 0.185* 0.150* 18. Internode/Sheath Ratio -0.079 -0.223** 0.186* 0.275*** 19. Flowering Time -0.182* -0.177* 0.019 0.192* 20. Number of Primary Tassel Branches 0.324*** 0.079 0.083 0.016 21. Attitude of Tassel Branches 0.271*** 0.034 0.320*** 0.071 22. Sex of Tassel Flowers 0.414*** 0.201** 0.010 0.003 23. Length of Central Tassel Spike 0.203** 0.016 -0.231** -0.112 24. Length of Longest Primary Tassel Branch 0.311*** 0.030 0.299*** 0.035 25. Tassel Ratio 0.293*** 0.018 0.360*** 0.061 26. Male Fertility -0.050 -0.097 0.070 -0.038 Percent of trace 8.2 10.5 10.9 12.1 (continued)

PAGE 148

141 Table 4. 1 3 (Continued) PC III Character S90 F90 fl, 1. Plant Height 0.160* -0.064 0.011 0.190* 2. Plant Biomass 0.129 -0.077 -0.086 0.036 3. Number of Tillers 0.005 0.247*** 0.042 -0.278*** 4. Relative Tiller Size -0.058 -0.106 0.038 0.057 5. Number of Flowered Tillers 0.065 0.280*** 0.061 -0.231** 6. Percentage of Flowered Tillers 0.191* 0.148 0.057 0.100 7. Leaf Sheath Color 0.084 0.021 0.026 0.053 8. Leaf Sheath Length -0.097 -0.080 -0.200** 0.043 9. Leaf Blade Length -0.176* 0.321 *** 0.324*** 0.237** 10. Leaf Blade Width 0.230** -0.493*** -0.139 0.246** 11. Leaf Area -0.039 -0.067 0.266*** 0.374*** 12. Leaf Ratio -0.267** 0 530*** 0 31 7*** 0 021 13. Leaf Blade/Sheath Ratio -0.133 0.370*** 0.479*** 0.175* 14. Number of Internodes 0.290*** -0.125 -0.052 -0.019 15. Internode Length 0.129 -0.004 0.099 0.173* 16. Internode Width -0.493*** 0.096 0.046 0.199** 17. Internode Ratio 0.282*** -0.034 0.047 -0.084 18. Internode/Sheath Ratio 0.187* 0.037 0.222** 0.097 19. Flowering Time 0.304*** 0.044 0.067 0.030 20. Number of Primary Tassel Branches -0.202** 0.006 -0.178* -0.263*** 21. Attitude of Tassel Branches -0.182* -0.033 -0.234** -0.319*** 22. Sex of Tassel Flowers 0.031 0.022 -0.035 0.109 23. Length of Central Tassel Spike 0.142 -0.095 -0.006 -0.098 24. Length of Longest Primary Tassel Branch -0.184* -0.009 -0.366*** -0.344*** 25. Tassel Ratio -0.176* -0.002 -0.351*** -0.359*** 26. Male Fertility 0.112 -0.044 -0.039 -0.003 Percent of trace 7.5 9.0 10.0 10.7 Marked entries show variables that are significantly represented in each principal component (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 149

142 Table 4.14 Characters most significantly represented (P < 0.005) in the first two axes of a principal components analysis of Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 100 specimens) and derived from tissue culture (Ro or Ru 100 specimens). nnmnonent Character Loading PC 1* (55.5%) Internode Length Plant Height Number of Tillers Internode/Sheath Ratio Leaf Blade Length Internode Ratio Number of Flowered Tillers 0.249 0.244 -0.242 0.239 0.226 0.222 -0.215 Male Fertility Length of Central Tassel Spike 0.252 0.224 PC ll a (8.2%) Leaf Sheath Color Plant Biomass 0.284 0.255 Sex of Tassel Flowers Number of Primary Tassel Branches Length of Longest Primary Tassel Branch Tassel Ratio Attitude of Tassel Branches 0.414 0.324 0.311 0.293 0.271 (continued)

PAGE 150

143 Table 4.14 (Continued) Component Character Loading PC l b Internode Length 0.285 Plant Height 0.277 (39.8%) Internode/Sheath Ratio 0.262 Internode Ratio 0.248 Leaf Sheath Color -0.229 Male Fertility 0.287 Length of Longest Primary Tassel Branch 0.267 Tassel Ratio 0.264 Length of Central Tassel Spike 0.253 Number of Primary Tassel Branches 0.243 Attitude of Tassel Branches 0.240 PC ll b Internode Width 0.462 Leaf Area 0.337 (1 0.5%) Percentage of Flowered Tillers -0.296 Leaf Blade Length 0.280 Internode Ratio -0.280 Leaf Sheath Length 0.270 Number of Flowered Tillers -0.229 Internode/Sheath Ratio -0.223 (continued)

PAGE 151

144 Table 4.14 (Continued) Component Character Loading PC l c (25.1%) Internode Ratio Leaf Sheath Color Internode/Sheath Ratio Leaf Blade Length Leaf Ratio Internode Width Percentage of Flowered Tillers Number of Flowered Tillers Leaf Area Leaf Sheath Length -0.306 0.304 -0.293 0.279 0.279 0.266 -0.255 -0.222 0.216 0.213 Sex of Tassel Flowers Number of Primary Tassel Branches 0.289 0.259 PC ll c (10.9%) Internode Length Leaf Area Plant Height Leaf Blade Length 0.369 0.323 0.267 0.234 Tassel Ratio Attitude of Tassel Branches Length of Longest Primary Tassel Branch Length of Central Tassel Spike 0.360 0.320 0.299 -0.231 (continued)

PAGE 152

145 Table 4.14 (Continued) Component onaracter LUdUII iy PC l d (17.3%) Internode Width Leaf Area Leaf Blade Length Internode Ratio Leaf Sheath Length Number of Flowered Tillers 0.314 0.274 0.252 -0.251 0.215 -0.213 Number of Primary Tassel Branches Tassel Ratio Length of Central Tassel Spike Attitude of Tassel Branches 0.333 0.282 0.271 0.262 PC ll d (12.1%) Leaf Ratio Leaf Blade/Sheath Ratio Leaf Blade Length Leaf Blade Width Internode/Sheath Ratio Internode Length Number of Tillers 0.469 0.381 0.373 -0.339 0.275 0.259 0.237 a Data sets from Spring 1 989 {W,R 0 ) b Data sets from Spring 1 990 (W, R 0 ) c Data sets from Fall 1 990 (W,R 0 ) d Data sets from Selfed Regenerants (W,Ri)

PAGE 153

146 Table 4.15 Statistics of location and dispersion and results of analysis of variance comparing PC I and PC II scores in Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture {Ro or R 1t 100 specimens each). Data sets from all harvesting seasons, analyzed as separate groups. Season Control Regenerants Mean o Mean o Spring PC I 3.64 1.39 -3.64 0.58 2321 .76 1989 PC II -0.10 1.62 0.10 1.27 0.98 Spring PC I 3.01 1.42 -3.01 0.71 1423.96 1990 PC II -0.33 1.96 0.33 1.20 8.22 Fall PC I -2.03 2.05 2.03 0.77 345.14 1990 PC II 0.54 2.18 -0.54 0.58 22.85 Selfed PC I -0.16 2.29 0.16 1.94 1.20 Regenerants PC II 0.80 1.77 -0.80 1.39 50.48 Marked entries indicate significant differences between means (** 0.01 > P > 0.001 ; "* P < 0.001).

PAGE 154

147 univariate findings, the differences found in the Ri plants involve leaf related characters (leaf length and width, the associated leaf ratio, and leaf blade/sheath ratio), internode length related traits (internode length, internode/sheath ratio) and the number of tillers. The plants were otherwise virtually undistinguishable from the controls (Table 4.14). Analysis of the R o. A number of characters, both vegetative and floral, in the R 0 showed higher coefficients of dispersion than would be expected from a population of strictly clonal origin. This effect, particularly noticeable in the first harvest season (cf., Table 4.2), suggested that a certain level of heterogeneity might exist among the regenerants, indicating that more than one variant phenotype could be present in the tissue-culture-derived plants. Application of principal components analysis to the data sets from the R 0 confirmed the univariate assumptions, and permitted the identification and characterization of the variants. Ordination on the first two principal components of the individuals of Ro, with all harvest times analyzed as a single group, is shown in Fig. 4.7. The associated factor loadings and percents of trace are given in Table 4.16. The presence, in Spring 1989, of two independent groups of tissue-culture-derived plants is undeniable. These segregate on the second component axis (Figs. 4.7, 4.8), revealing the existence of differences in plant biomass, leaf sheath color and internode width (Table 4.16). The two groups converge in a single one during the Spring and Fall of 1 990, as they both evolve along the first component, increasing in size and fertility with the passage of time (Fig. 4.7, Table 4.16). Figure 4.9 illustrates the ordination on the first two principal components of the individuals of Ro, when the four harvest times are analyzed as separate groups. Comparison of factor loadings and percents of trace associated with the individual analyses for each harvest time are presented in Table 4.17. Analyzing the populations for the different harvest seasons as separate groups makes direct comparison of the

PAGE 155

" o Q. W O) 8 1 f l CD Q_ © rT 5 » _ § rt C ** CD CD ~ 5 >^ 111! CO e co lip m It C , gw = cd S i N^oE 2x2 to c (1) co 2 CD 03 « co p co JO « nQ CO ** ,J= =}£_,
PAGE 156

149 — i 1 ! 1 1 ° 1990 o -C5 z _ SPRI i i i t r CD 00 LCD cr o. n °° ° CO CO z 111 z o Q. o o — I < CL o z cr Q(%6 00 II INdNOdWOO IVdlONIdd

PAGE 157

150 Table 4.16 Factor loadings and percents of trace from principal components analysis of a Zea diploperennis (diploperennial teosinte) population derived from tissue culture (R 0 , 100 specimens). Data sets from all harvesting seasons, analyzed as a single group. Principal Component Character I II III 1 I . Plant I— lairtht ridiii ntJiyru n 1 1 q U. I I » 0 CtOA o tL. Plant DiArrtQee U. I *to U.OC7C7 0 007 r> o. iNUiiiuei ui i nicio \J, 1 uu o on A -n 1 no -0 1 ftO O. -o 1 ri U. I O 1 U. IU I 0 OOfi U.UOH -O 970* 7 Leaf Shpath Color 0 090 1 paf Shpflth 1 pnnth l—^Cll Ul l^Oll 1 L_d lull 1 0 901 0 147 -O 10Q q 1 paf Rlfldp 1 pnnth LCGI LCI IUU 1 U.UUU -O 971 * 10. Leaf Blade Width 0.101 0.099 0.177 11. Leaf Area 0.221 0.017 -0.130 12. Leaf Ratio 0.133 -0.125 -0.379*** 13. Leaf Blade/Sheath Ratio 0.158 -0.176 -0.302** 14. Number of Internodes 0.147 -0.156 0.318** 15. Internode Length 0.252* 0.035 0.028 16. Internode Width 0.139 -0.373*** -0.197 17. Internode Ratio 0.242* 0.110 0.078 18. Internode/Sheath Ratio 0.245* -0.014 0.079 19. Flowering Time 0.183 -0.291* 0.290* 20. Number of Primary Tassel Branches 0.237* 0.129 -0.083 21. Attitude of Tassel Branches 0.243* 0.128 -0.085 22. Sex of Tassel Flowers 0.205 0.026 0.218 23. Length of Central Tassel Spike 0.214 -0.073 0.303** 24. Length of Longest Primary Tassel Branch 0.243* 0.128 -0.083 25. Tassel Ratio 0.242* 0.127 -0.084 26. Male Fertility 0.250* 0.017 0.020 Percent of trace 58.3 10.9 7.7 Marked entries show variables that are significantly represented in each principal component (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 158

Fig. 4.8 Histograms of the scores of specimens of a Zea diploperennis (diploperennial teosinte) population derived from tissue culture (Ro, 1 00 specimens) on the first two principal components extracted from a matrix of correlation coefficients between twenty six morphometric variables. Data sets from all harvesting seasons, analyzed as a single group. All score frequencies follow a normal (Gaussian) distribution except for the PC II scores in the Spring 1989, which show a bimodal frequency distribution, and reveal the presence of two discrete groups of plants in the Ro, at that harvest time.

PAGE 159

152

PAGE 160

C CD N "3 o B _ §i2 CO c x: « coo i9 co" c o co V% Q> Q. CO O C 6 w O O) xi .E CO co CD CO x: "cO E p o I If. CO 8 Q) O Set 5 £ N s>!§ « 3-d 3 CD CO E ° o ® 2 2 fi-S a ** J5 "2 TJ 4-> (0 CD CO > o c C CD P E in v> c 2 CD c CD CD CD CO •*-» c CO CD c CD cn CD CD OC cn CD co CL E J E 5 3 ^ = 8 ^ 2 O * Q O) "° re E cd | § * 1 1 *= *= C m CO CO CO -C >, CL C "5 CL "p CD TJ CL o CD W ir O > > co O Q.5 5 CO DC £ m ® -» CO <» £ to 2X5 CD — U O CD O w (J) TJ fD CD c co -5 — CD *L O QO WOO E XI E 3 c Id 3 CD C x> c o CL CO 2 L_ o o CD x: >^ X) T3 0 "cO o tj c 2 CO CO c g CO co CD o o CO CO 3 •D > TJ C 03 -* oS

PAGE 161

154

PAGE 162

155 Table 4.17 Comparison of factor loadings and percents of trace from principal components analyses of Zea diploperennis (diploperennial teosinte) populations derived from tissue culture (Ro or flj, 100 specimens each). Data sets from all harvesting seasons, analyzed as separate groups. PC Character S89 S90 F90 1. Plant Height 0.333" 0.173 0.343** 0.110 2. Plant Biomass 0.278* -0.029 0.160 -0.096 3. Number of Tillers 0.314" -0.251* -0.347** -0.210 4. Relative Tiller Size 0.000 -0.061 -0.026 -0.109 5. Number of Flowered Tillers 0.229* -0.271 * -0.360" -0.214 6. Percentage of Flowered Tillers -0.033 -0.265* -0.321 ** 0.002 7. Leaf Sheath Color 0.327** -0.007 0.000 0.000 8 l paf Sheath Lencrth 0.030 0.220 0.145 0.200 9. Leaf Blade Length -0.138 0.344** 0.154 0.246* 10. Leaf Blade Width 0.070 0.137 0.308** 0.107 11. Leaf Area -0.046 0.363** 0.342** 0.264* 12. Leaf Ratio -0.134 0.163 -0.092 0.116 13. Leaf Blade/Sheath Ratio -0.143 0.290* -0.049 0.037 14. Number of Internodes 0.076 -0.041 0.112 -0.072 15. Internode Length 0.226 0.076 0.256* 0.012 16. Internode Width -0.306** 0.350** 0.187 0.300** 17. Internode Ratio 0.314" -0.292* 0.173 -0.267* 18. Internode/Sheath Ratio 0.223 -0.128 0.107 -0.169 19. Flowering Time 0.000 0.000 0.000 -0.067 20. Number of Primary Tassel Branches 0.000 0.000 0.010 0.339** 21. Attitude of Tassel Branches 0.000 0.000 0.000 0.273* 22. Sex of Tassel Flowers 0.311" 0.158 0.000 0.149 23. Length of Central Tassel Spike 0.304** 0.197 0.218 0.231* 24. Length of Longest Primary Tassel Branch 0.000 0.000 0.151 0.342** 25. Tassel Ratio 0.000 0.000 -0.044 0.313" 26. Male Fertility 0.000 0.188 -0.094 -0.005 Percent of trace 43.5 22.1 17.4 17.9 (continued)

PAGE 163

156 Table 4.17 (Continued) PC II Character S89 S90 F90 Ri 1. Plant Height 0.040 -0.175 0.245* -0.186 2. Plant Biomass 0.041 -0.256* 0.241 * 0.063 3. Number of Tillers 0.011 0.278* 0.220 0.400*** 4. Relative Tiller Size 0.000 -0.106 -0.034 -0.035 5. Number of Flowered Tillers 0.070 0.273* 0.222 0.380*** 6. Percentage of Flowered Tillers 0.089 0.192 0.186 -0.044 7. Leaf Sheath Color 0.023 0.070 0.000 0.015 8. Leaf Sheath Length 0.089 0.002 -0.120 0.025 9. Leaf Blade Length 0.526*" 0.323** -0.114 0.025 10. Leaf Blade Width 0.194 -0.369** -0.084 -0.384*** 11. Leaf Area 0.484"* -0.011 -0.157 -0.262* 12. Leaf Ratio 0.221 0.459*** -0.019 0.269* 13. Leaf Blade/Sheath Ratio 0.512"* 0.339** 0.047 -0.003 14. Number of Intemodes -0.101 -0.171 0.148 -0.122 15. Internode Length 0.218 -0.001 0.402*** -0.112 lb. Internode Width u.uoy 17. Internode Ratio 0.063 -0.076 0.465*** -0.022 18. Internode/Sheath Ratio 0.198 0.003 0.427*** -0.094 19. Flowering Time 0.000 0.000 0.000 -0.050 20. Number of Primary Tassel Branches 0.000 0.000 0.198 0.222 21. Attitude of Tassel Branches 0.000 0.000 0.000 0.228* 22. Sex of Tassel Flowers 0.072 0.037 0.000 -0.339** 23. Length of Central Tassel Spike 0.076 -0.212 -0.061 -0.025 24. Length of Longest Primary Tassel Branch 0.000 0.000 0.091 0.220 25. Tassel Ratio 0.000 0.000 0.128 0.243* 26. Male Fertility 0.000 -0.212 -0.016 -0.025 Percent of trace 14.7 15.6 14.3 12.0 (continued)

PAGE 164

157 Table 4.17 (Continued) PC III onaracter S89 S90 F90 Ri 1. Plant Height -0.029 0.450*** 0.060 0.052 2. Plant Biomass -0.021 0.289* -0.180 0.067 3. Number of Tillers 0.055 0.172 0.136 0.056 4. Relative Tiller Size 0.000 -0.058 -0.173 -0.091 5. Number of Flowered Tillers 0.177 0.208 0.133 0.074 6. Percentage of Flowered Tillers 0.163 0.265* 0.114 0.060 7. Leaf Sheath Color 0.003 0.086 0.000 0.032 8. Leaf Sheath Length -0.179 0.009 0.164 -0.061 9. Leaf Blade Length 0.087 0.090 0.532 _ . mm*** 0.466 10. Leaf Blade Width -0.612"* 0.001 -0.250* -0.173 11. Leaf Area -0.387*** 0.061 0.222 0.235* 12. Leaf Ratio 0.532*** 0.084 0.531*** 0.431 *** 13. Leaf Blade/Sheath Ratio 0.119 0.096 0.161 0.449*** 14. Number of Internodes 0.002 0.182 -0.157 -0.128 15. Internode Length 0.152 0.488*** 0.189 0.210 16. Internode Width ill l \s ill vy \s w v M %i -0.008 -0.001 0.283* 0.188 17. Internode Ratio 0.110 0.241* 0.071 -0.054 18. Internode/Sheath Ratio 0.194 0.355** 0.044 0.206 19. Flowering Time 0.000 0.000 0.000 0.109 20. Number of Primary Tassel Branches 0.000 0.000 -0.004 -0.131 21. Attitude of Tassel Branches 0.000 0.000 0.000 -0.145 22. Sex of Tassel Flowers -0.027 0.146 0.000 0.000 23. Length of Central Tassel Spike -0.046 0.174 -0.026 -0.154 24. Length of Longest Primary Tassel Branch 0.000 0.000 -0.073 -0.180 25. Tassel Ratio 0.000 0.000 -0.040 -0.172 26. Male Fertility 0.000 0.158 -0.025 -0.035 Percent of trace 10.9 12.5 12.6 11.0 Marked entries show variables that are significantly represented in each principal component (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *" P < 0.001).

PAGE 165

158 seasonal differences more difficult, as both eigenvalues and eigenvectors for the individual principal components change with the passage of time (Table 4.17). Doing it, however, increases analytical resolution for each particular season, as fewer data sets are evaluated and the principal components are scaled to best fit the factors prevailing in each specific season. The enhanced analytical ability of the procedure, in turn, makes it simpler to identify critically the individual plants belonging to each group and, hence, to compare them. The results from these analyses fully confirm the presence of two discrete groups of plants (hereafter designated Type A and B Regenerants) in Spring 1 989 (Figs. 4.10 and 4.11). The two groups, whose elements can be individually located by the graphic model (Fig. 4.9), have highly significant differences in mean PC I scores, a pattern that is not, however, repeated in subsequent harvest seasons (Fig. 4.12, Table 4.18). Type A Regenerants comprise a group of 80 specimens that basically conform to the morphological archetype formerly established for Spring 1989 Ro plants. In brief, they can be described as multitillering plants of reduced stature and fertility, with atypical proportions and poorly developed (short and unbranched) tassels. Type B Regenerants, in turn, are a smaller group of 20 plants that, although still fitting the same overall definition, differ significantly from the former group in having fewer tillers (and, consequently, fewer flowered tillers), an even shorter plant height, wider internodes (and a correspondingly smaller internode ratio) and a lighter color (Table 4.17). In addition, the relatively stouter stems only partially compensate for the combined reduction in both number of tillers and plant height, which results in a reduced plant biomass. Differences are also significant at the floral level, where an otherwise even shorter tassel spike is, in many plants, completely replaced by an ear (all plants where this was observed fall in this group). This is an extreme expression of the feminization of the male inflorescence already mentioned previously.

PAGE 166

Fig. 4.10 Fig. 4.11 Vegetative and floral morphology in diploperennial teosinte. Center: seed-derived control Right: tissue-culture-derived plant (type A) Left: tissue-culture-derived plant (type B) Vegetative and floral morphology in diploperennial teosinte. Left: tissue-culture-derived plant (type A) Right: tissue-culture-derived plant (type B) The plant in the center is morphologically intermediate between the two extremes.

PAGE 167

160

PAGE 168

Fig. 4.12 Histograms of the scores of specimens of a Zea diploperennis (diploperennial teosinte) population derived from tissue culture (Ro or Ri, 100 specimens each) on the first two principal components extracted from a matrix of correlation coefficients between twenty six morphometric variables. Data sets from all harvesting seasons, analyzed as separate groups. All score frequencies follow a normal (Gaussian) distribution except for the PC I scores in the Spring 1989, which show a bimodal frequency distribution, and reveal the presence of two discrete groups of plants in the R 0 , at that harvest time.

PAGE 169

PC I SCORES PC II SCORES

PAGE 170

163 Table 4.18 Statistics of location and dispersion and results of analysis of variance comparing PC I and PC II scores for the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (flo) derived from tissue culture. Data sets from all harvesting seasons, analyzed as separate groups. Type A regenerants Type B regenerants Season F Mean o Mean a Spring 1989 PCI PC II 1.34 0.03 0.89 1.66 -5.37 -0.13 1.34 1.78 737.45* 0.15 Spring 1990 PCI PC II -0.18 -0.02 2.18 1.82 0.72 0.06 1.90 1.81 2.87 0.03 Fall 1990 PC I PC II -0.05 0.12 1.93 1.85 0.20 -0.50 2.12 1.36 0.25 2.00 Marked entries indicate significant differences between means (***P < 0.001).

PAGE 171

164 The exact identification by the graphic model of all individual Type B plants (Fig. 4.9) permitted the application of the methods of univariate analysis to the two Ro subpopulations. The results from these fully support all the previous conclusions, and permit their quantification (Tables 4.19 through 4.24). They also add to those conclusions by revealing, for example, that the reduced height found in Type B plants (31 .9% less than Type A) is due to a highly significant shorter internode length (1 6.6% shorter than Type A) rather than fewer internodes, which also results in a highly significant smaller internode/sheath ratio (16.2% less than Type A). Type B plants also have significantly longer leaf blades (+14.2%) as well as a correspondingly greater leaf ratio (+19.8%) and leaf blade/sheath ratio (+15.0%). The exact identification of the individual Type B specimens also revealed that these plants did not randomly originate in culture, as all 20 plants were derived from one callus line. The facts that (1 ) both groups evolve concurrently in the same direction [they become virtually identical by Spring 1990 (Figs. 4.7 and 4.9, Table 4.18)], and that (2) there are no aberrant traits in the Ri (Fig. 4.12) indicate that the differences found between the two morphological types are, once again, transient in nature and not sexually transmitted. Genetically homologous, the two plant groups gradually develop to become indistinguishable, before growing fully normal later in time. Hence, Type B represents nothing but an extreme in the expression of the tissue-culture-induced epigenetic variation manifested by all the plants that were the object of this study. Canonical Discriminant Analysis Because univariate analyses demonstrated significant differences for many morphometric characters between tissue-culture-derived plants and the seed-derived controls, as well as within the tissue culture regenerants themselves, calculation of a discriminant function designed to maximize group differences should reveal substantial

PAGE 172

165 Table 4.19 Statistics of location and dispersion for the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population {R 0 ) derived from tissue culture. Values for the Spring 1 989. Type A Regenerants Type B Regenerants Character — Mean o CV Mean o CV 1 69.25 1.52 2.20 47.15 6.96 14.76 2 258.09 26.49 10.26 1 70.60 A O O ~f 43.87 AC 7^ 25.71 3 36.63 2.41 6.58 20.20 4.07 20.1 b 4 4.00 0.00 0.00 4.00 0.00 0.00 5 20.04 4.56 22.74 12.35 0.0» df . 4b 6 55.00 12.74 23.16 51 .70 i O Qi On T7 7 3.00 0.00 0.00 h r\r\ 1 .00 0.00 U.UU 8 73.35 2.40 3.28 d.DO 9 1 80.95 34.07 18.83 one cc 20b. bo o4.oo 1 Dot) 10 22.59 4.69 20.74 20.90 3.66 17.52 11 30.69 8.86 28.88 32.85 9.94 30.27 12 8.39 2.51 29.90 10.05 1.70 16.89 13 2.47 0.46 18.58 2.84 0.50 17.70 14 7.03 0.84 11.98 6.65 0.93 14.04 15 50.08 6.21 12.40 41 .75 2.27 5.43 16 2.26 0.18 8.09 3.57 0.58 16.11 17 22.33 3.31 14.83 11.95 2.06 17.27 18 0.68 0.08 12.25 0.57 0.04 6.30 19 1.00 0.00 0.00 1.00 0.00 0.00 20 0.00 0.00 0.00 0.00 21 0.00 0.00 0.00 0.00 22 1.00 0.00 0.00 0.20 0.41 205^20 23 55.19 3.13 5.67 12.00 24.63 205.21 24 0.00 0.00 0.00 0.00 25 0.00 0.00 0.00 0.00 26 0.00 0.00 0.00 0.00

PAGE 173

166 Table 4.20 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (A?o) derived from tissue culture. Values for the Spring 1989, sorted by percent difference between means. Character Mean (A) Mean (B) % F 16 2.26 3.57 58.0 302.90*** 12 8.39 10.05 19.8 7.86** 13 2.47 2.84 15.0 10.18" 9 1 80.95 206.65 14.2 9.02" 6 55.00 61.70 12.2 4.42* 1 1 30.69 32.85 7.0 0.91 4 4.00 4.00 0.0 8 73.35 72.70 -0.9 1.14 14 7.03 6.65 -5.4 3.04 10 22.59 20.90 -7.5 2.26 18 0.68 0.57 -16.2 32.42*** 15 50.08 41.75 -16.6 34.59*** 1 69.25 47.15 -31.9 693.97"* 2 258.09 170.60 -33.9 130.45*** 5 20.04 12.35 -38.4 49.84*** 3 36.63 20.20 -44.9 546.57"* 17 22.33 11.95 -46.5 178.32*** 7 3.00 1.00 -66.7 00 *** 20 0.00 0.00 • 21 0.00 0.00 24 0.00 0.00 25 0.00 0.00 26 0.00 0.00 19 1.00 1.00 0.0 23 55.19 12.00 -78.3 237.88*** 22 1.00 0.20 -80.0 313.60*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P £ 0.001).

PAGE 174

167 Table 4.21 Statistics of location and dispersion for the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (R 0 ) derived from tissue culture. Values for the Spring 1 990. Type A Regenerants Type B Regenerants Mean a CV Mean a CV 1 77.68 4.09 5.26 77.95 3.95 5.07 2 199.43 38.54 19.32 186.40 34.13 18.31 3 20.50 5.39 26.30 20.95 4.10 19.56 4 3.44 0.50 14.52 3.45 0.51 14.79 5 15.21 5.02 32.99 15.35 3.65 23.75 6 72.93 6.79 9.31 72.80 6.03 8.29 7 2.39 0.49 20.53 2.65 0.49 18.47 8 72.14 4.20 5.82 71 .95 4.67 6.50 9 199.63 41 .65 20.87 209.20 35.82 17.12 10 22.91 3.94 17.21 23.44 4.20 17.91 11 34.23 9.07 26.51 36.40 7.86 21 .58 12 9.02 2.67 29.57 9.31 2.83 30.37 13 2.77 0.56 20.29 2.91 0.46 15.90 14 9.11 1.36 14.92 8.40 1.14 13.60 15 82.21 4.31 5.24 81.65 4.22 5.17 16 2.97 0.31 10.34 3.02 0.26 8.69 17 27.95 3.11 11.14 27.35 1.98 7.24 18 1.14 0.08 7.07 1.14 0.07 6.54 19 2.00 0.00 0.00 2.00 0.00 0.00 20 0.00 0.00 0.00 0.00 21 0.00 0.00 0.00 0.00 22 1.34 0.48 35.58 1.85 0.37 19.80 23 98.66 5.08 5.15 103.30 3.54 3.43 24 0.00 0.00 0.00 0.00 25 0.00 0.00 0.00 0.00 26 22.23 13.97 62.87 34.45 8.99 26.09

PAGE 175

168 Table 4.22 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (Ro) derived from tissue culture. Values for the Spring 1990, sorted by percent difference between means. Character Mean (A) Mean (B) % F 7 2.39 2.65 10.9 4.59* 11 34.23 36.40 6.3 0.97 13 2.77 2.91 5.1 1.05 9 199.63 209.20 4.8 0.89 12 9.02 9.31 3.2 0.18 10 22.91 23.44 2.3 0.29 3 20.50 20.95 2.2 0.12 16 2.97 3.02 1.7 0.40 5 15.21 15.35 0.9 0.01 1 77.68 77.95 0.3 0.07 4 3.44 3.45 0.3 0.01 18 1.14 1.14 0.0 0.08 6 72.93 72.80 -0.2 0.01 8 72.14 71.95 -0.3 0.03 15 82.21 81.65 -0.7 0.28 17 27.95 27.35 -2.1 0.67 2 199.43 186.40 -6.5 1.91 14 9.11 8.40 -7.8 4.66* 26 22.23 34.45 55.0 13.82*** 22 1.34 1.85 38.1 20.15*** 23 98.66 103.30 4.7 14.83*** 19 2.00 2.00 0.0 20 0.00 0.00 21 0.00 0.00 24 0.00 0.00 • 25 0.00 0.00 Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; *** P < 0.001).

PAGE 176

169 Table 4.23 Statistics of location and dispersion for the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (R 0 ) derived from tissue culture. Values for the Fall 1990. Type A Regenerants Type B Regenerants Mean a CV Mean o CV 1 1 50.73 7.39 4.91 151.15 6.97 4.61 2 299.58 27.67 9.24 290.25 24.57 8.46 3 15.18 3.45 22.73 15.40 2.68 17.42 4 3.44 0.50 14.52 3.45 0.51 14.79 5 10.83 3.50 32.30 10.70 2.74 25.57 6 69.96 8.45 12.07 68.45 7.70 11.25 7 3.00 0.00 0.00 3.00 0.00 0.00 8 89.35 8.09 9.05 89.00 9.26 10.40 9 287.84 16.81 5.84 291 .75 14.48 4.96 10 25.35 1.36 5.38 25.54 1.44 5.64 11 54.74 4.15 7.58 55.80 3.62 6.49 12 11.39 0.95 8.34 11.46 1.00 8.69 13 3.24 0.31 9.42 3.31 0.32 9.67 14 9.45 1.10 11.65 8.95 0.94 10.55 15 175.48 18.41 10.49 172.95 18.14 10.49 16 3.19 0.14 4.39 3.20 0.13 3.93 17 55.09 6.26 11.36 54.00 4.95 9.17 18 1.98 0.25 12.55 1.96 0.22 11.37 19 2.00 0.00 0.00 2.00 0.00 0.00 20 2.46 0.50 20.38 2.35 0.49 20.82 21 1.00 0.00 0.00 1.00 0.00 0.00 22 2.00 0.00 0.00 2.00 0.00 0.00 23 121.18 6.71 5.54 127.20 4.63 3.64 24 92.08 5.33 5.79 91.85 5.57 6.06 25 0.76 0.05 6.84 0.72 0.03 4.18 26 92.21 4.70 5.09 93.15 4.57 4.90

PAGE 177

170 Table 4.24 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (Ro) derived from tissue culture. Values for the Fall 1990, sorted by percent difference between means. Character Mean (A) Mean (B) % F 13 3.24 3.31 2.2 0.66 1 1 54.74 55.80 1.9 1.10 3 15.18 15.40 1.4 0.07 9 287.84 291 .75 1.4 0.91 10 25.35 25.54 0.7 0.29 1 ? 11.39 11.46 0.6 0.09 1 150.73 151.15 0.3 0.05 4 3.44 3.45 0.3 0.01 16 3.19 3.20 0.3 0.08 7 3.00 3.00 0.0 8 89.35 89.00 -0.4 0.03 18 1.98 1.96 -1.0 0.11 5 10.83 10.70 -1 .2 0.02 15 175.48 172.95 -1 .4 0.30 17 55.09 54.00 -2.0 0.52 6 69.96 68.45 -2.2 0.53 2 299.58 290.25 -3.1 1.90 14 9.45 8.95 -5.3 3.48 23 121.18 127.20 5.0 14.36"* 26 92.21 93.15 1.0 0.64 19 2.00 2.00 0.0 21 1.00 1.00 0.0 22 2.00 2.00 0.0 24 92.08 91.85 -0.2 0.03 20 2.46 2.35 -4.5 0.81 25 0.76 0.72 -5.3 10.38" Marked entries indicate significant differences between means (" 0.01 > P > 0.001 ; *" P < 0.001).

PAGE 178

171 separation between these a priori groups. The following results show that this expectation is indeed realized. Analysis of all data sets as a single group . Projection on the first two canonical variables of individuals of R 0 (or fl r ) and W, when all harvesting times are analyzed as a single group, is shown in Fig. 4.13. Table 4.25 displays the associated standardized canonical coefficients and shows the cumulative average squared canonical correlations (ASCC) from the step-wise discriminant analysis, and details in descending order the unique contribution of the various morphometric traits to the discriminatory capability of the model. Generalized distances between group centroids are given for all groups in Table 4.26. Mean scores for the two canonical variables of individuals of Ro (or Ri) and W are compared in Table 4.27, covering all harvesting times. In the step-wise discriminant analysis male fertility is entered first, followed by plant biomass, internode length and leaf sheath color, in basic agreement with results from previous analyses. In all, 25 of the 26 variables are significantly involved in the definition of the discriminant function, which shows the difficulty of the model to discriminate concurrently all the populations from all harvest seasons (total ASCC = 0.559). With these 25 variables included, there is significant heterogeneity between the group centroids of the tissue-culture-derived population and the seedderived control for all but the last harvest season (Tables 4.26 and 4.27). The model, then, successfully discriminates the Ro from the wild types during Spring 1989, Spring 1 990 and Fall 1 990. It also clearly renders the attenuation, with time, of the differences among those populations, to the point of finally being unable to discriminate the R 1 and the corresponding control (Fig. 4.13, Table 4.27). Analysis of all data sets as separate groups . Frequency distributions of the discriminant scores on the first canonical variable of individuals of Ro (or fl?) and W, when all harvesting times are analyzed as separate groups, is shown in Fig. 4.14.

PAGE 179

fill a> d) CO ±= ** > > _ "D o ™ 0) I? (0 Q. 3 O c c 2 £ 5 £ /n TO c o) co » w co c
PAGE 180

173 1 1 * •% 1 1 1 I ! SPRING 1990 \ r ; i i ! i ' W " -10 -5 ( : : : : : — — : : : o j i*h o ° ° i i LO FALL 19S ^ o up m o 10 II 319VIHVA 1V0IN0NV0 LD _l m < o z o z < o

PAGE 181

174 Table 4.25 Standardized canonical coefficients and cumulative average squared canonical correlations (ASCC) from a step-wise discriminant analysis calculated to discriminate two Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 100 specimens) and derived from tissue culture {R 0 or Ru 100 specimens each). Data sets from all harvesting seasons, analyzed as a single group. Canonical Variable Step Character ASCC 8 *** *** *** 1 Male Fertility 4.652 1.946 0.139 2 Plant Biomass 0.008 0.418 0.205 3 Internode Length -0.663 -1.526 0.249 4 Leaf Sheath Color -0.116 -0.100 0.299 5 Number of Tillers -1.474 2.597 0.333 6 Length of Central Tassel Spike 0.801 -2.528 0.366 7 Flowering Time 0.016 -0.417 0.390 8 Leaf Blade Width -0.222 0.067 0.415*" 9 Number of Flowered Tillers 0.578 -1.325 0.433 10 Percentage of Flowered Tillers -0.147 0.468 0.454 1 1 Length of Longest Primary Tassel Branch -1 .682 5.21 3 0.464 12 Tassel Ratio 1.600 -4.451 0.481 13 Attitude of Tassel Branches 0.114 0.081 0.493 14 Relative Tiller Size -0.008 0.164 0.504 15 Internode/Sheath Ratio 2.041 1.796 0.510 16 Internode Width -0.405 -0.463 0.515 17 Leaf Area -0.052 0.291 0.524"* 18 Sex of Tassel Flowers -0.108 -0.202 0.530*** 19 Plant Height -0.058 0.651 0.535*** 20 Internode Ratio -0.480 -0.420 0.538*** 21 Leaf Sheath Length 0.803 0.421 0.539*** 22 Number of Primary Tassel Branches 0.039 0.006 0.542*** 23 Leaf Ratio -0.260 0.021 0.544*** 24 Leaf Blade Length 0.682 0.489 0.556 25 Number of Internodes 0.067 -0.183 0.558 26 Leaf Blade/Sheath Ratio -0.293 -0.506 0.559 *** *** *** *** *** *** *** *** *** The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. Marked entries show variables that significantly contribute to the discriminatory capability of the model (** 0.01 >P>0.001; *** P<0.001).

PAGE 182

175 Table 4.26 Mahalanobis' distances 8 between group centroids from a canonical discriminant analysis calculated to discriminate two lea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 100 specimens) and derived from tissue culture (Ro or Ru 100 specimens each). Comparative results for all harvesting periods, when all data sets are pooled together and analyzed as a single group. To group From group W (S89) fl 0 (S89) W(S90) fl 0 (S90) W(F90) fl 0 (F90) IV(fl f ) fl*(fl,) W(S89) 0.0 402.4 16.3 199.2 25.4 42.7 10.5 16.9 flo(S89) 0.0 369.8 99.9 369.4 337.8 395.0 386.5 W(S90) 0.0 180.0 3.7 20.8 9.8 14.2 flo(S90) 0.0 185.1 160.1 196.6 193.8 W(F90) 0.0 21.6 12.8 14.4 fl 0 (F90) 0.0 29.4 26.5 W(Ri) 0.0 7.6 Ri(Ri) 0.0 8 The Mahalanobis' D 2 (or generalized distance between two groups) is a measure of the morphometric dissimilarity between the groups. The entries in bold show distances between the centroids of Ro (or Ri) and the corresponding controls for the four harvest seasons.

PAGE 183

176 Table 4.27 Statistics of location and dispersion and results of analysis of variance comparing CAN I and CAN II scores in Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 100 specimens) and derived from tissue culture (Ro or Ri, 100 specimens each). Data sets from all harvesting seasons, analyzed as a single group. Season Control Regenerants Mean o Mean o Spring 1989 CAN I CAN II 4.51 -0.33 0.84 1.00 -15.08 2.96 0.63 1.50 35008.07*" 333.32*** Spring 1990 CAN I CAN II 3.86 0.35 0.69 0.87 -8.52 -4.54 2.00 0.65 3410.12"* 2015.63*** Fall 1990 CAN I CAN II 3.84 0.81 0.74 0.88 2.59 0.06 0.68 0.54 154.80*** 52.50*** Selfed Regenerants CAN I CAN II 4.53 0.22 0.99 1.27 4.28 0.45 0.65 0.95 4.46 2.08 Marked entries indicate significant differences between means (*** P < 0.001).

PAGE 184

177 Table 4.28 displays, for every harvest season, the standardized canonical coefficients associated with each separate analysis and shows the cumulative average squared canonical correlations (ASCC) from the step-wise discriminant analyses detailing, in descending order, the morphometric traits that, in each season, significantly contributed to the discriminatory capability of the model. Average squared canonical correlations, eigenvalues and generalized distances between group centroids are given for all harvests in Table 4.29. Mean scores for the first canonical variable of individuals of Ro (or fir) and W are compared in Table 4.30, covering all harvesting times. As was the case with principal components analysis, the increased analytical power attained by processing separately the four harvest seasons improved the resolution of the models; tissue-culture-derived populations and respective controls were now discriminated at all harvest seasons, including the Rl Doing so, however, considerably narrowed the descriptive value of the method as, by their very own nature and unlike principal components analyses, discriminant analyses tend to overemphasize the discriminant value of single (or just a few) characters, in detriment to other equally or even more important (in biological terms) discriminant traits. Discrimination of the Ro from the respective control, in Spring 1 989, is based almost entirely on male fertility (ASCC = 0.995, Table 4.28). Other entries that significantly contribute to the discriminatory power of the model include plant height, number of tillers, length of central tassel spike and, to a lesser extent, internode ratio. The model is not further improved with the incorporation of any additional variables (Table 4.28). With these five variables included, there is a highly significant heterogeneity among the group centroids. The model clearly discriminates the R 0 plants from the seed-derived controls (Fig. 4.14, Tables 4.29 and 4.30). Discrimination is still strongly based on male fertility (ASCC = 0.922, Table 4.28) in Spring 1990. Other contributing variables include internode length, leaf sheath color, tassel ratio and sex of tassel flowers and, to a lesser extent, internode width, internode

PAGE 185

^ co c re gQ. © o •= w -Qo e> § rt lis 1 * C tO Q) CD N >> re 2 re o CO #r w wj ^ o C 15 "8 S 8 g $ 1 S .55 2 i^j w © « » 0 "O E D) (0 c .ill* "g I s g A) (J ° w d) 52 X « 're 3 xi 2 8 co • *s 3 x: re « *-> — CO °&E E"2 re O re 5 re aT_ m Q. o> c < >. re g (D CD cn re ~ Q

PAGE 186

179 i 1 1 1 1 1 1 — i 1 1 r 3 3 8 8 2° § 8 8 ° A0N3n03Uzl

PAGE 187

180 Table 4.28 Standardized coefficients from a canonical discriminant analysis and cumulative average squared canonical correlations (ASCC) from a step-wise discriminant analysis calculated to discriminate two Zea diploperennis (diploperennial teosinte) populations, raised from wildcollected seed (W, 1 00 specimens) and derived from tissue culture (Ro or ft j, 100 specimens). Step Character CAN l a ASCC i i . Malp Fprtilitv 12.939 0.995"* 2 Plant Heiaht 0.383 0.995" o. Mi imhpr nf Tillprc -1 .367 0.995" 4 lci i^u i vji vwi luai i aooci v_/^ji r\C/ 1.062 0.995" R o. IntprnnHp Ratio -0.035 0.996* 6 Leaf Ratio -0.175 0.996 7 Intprnodp Width II 1 lv 1 1 1 w V V IVl 11 1 -0.304 0.996 8. Lenath of Lonaest Primarv Tassel Branch -1 .820 0.996 9. Attitude of Tassel Branches 0.400 0.996 10. Number of Primary Tassel Branches 0.386 0.996 11. Tassel Ratio 0.961 0.996 12. Average Tiller Size 0.105 0.996 13. Internode/Sheath Ratio 2.947 0.996 14. Percentage of Flowered Tillers -0.045 0.996 15. Leaf Sheath Length 0.422 0.996 16. Internode Length -2.356 0.996 17. Leaf Blade/Sheath Ratio -0.851 0.996 18. Leaf Blade Length 1.555 0.996 19. Leaf Blade Width -0.106 0.996 20. Sex of Tassel Flowers -0.112 0.996 21. Plant Biomass -0.071 0.996 22. Flowering Time -0.053 0.996 23. Leaf Sheath Color -0.061 0.996 24. Number of Internodes 0.032 0.996 25. Leaf Area -0.201 0.996 26. Number of Flowered Tillers -0.016 0.996 (continued)

PAGE 188

181 Table 4.28 (Continued) Step Character CAN l b ASCC 1 . Male Fertilitv 4.610 0.922*** 2 Internode Lenath -0.216 0.947*** 3. Leaf Sheath Color -0.381 0.952*** 4 Taccol Ratio -9.444 0.955*** 5 Sex of Tassel Flowers -0.212 0.958*** 6. Internode Width -0.636 0.959** 7. Internode Ratio -1.008 0.961** 8. Plant Heiaht 0.629 0.962* 9. Lenath of Central Tassel Soike -2.034 0.963* 10. Lenath of Lonaest Primarv Tassel Branch 10.292 0.969* 11. Internode/Sheath Ratio 2.254 0.970* 12. Leaf Sheath Length 0.754 0.971* 13. Number of Internodes -0.061 0.971 14. Leaf Blade Width -0.565 0.971 15. Number of Tillers -1 .219 0.971 16. Plant Biomass 0.108 0.971 17. Leaf Area -0.042 0.972 18. Leaf Ratio -0.897 0.972 19. Average Tiller Size 0.053 0.972 20. Number of Primary Tassel Branches 0.155 0.972 21. Leaf Blade Length 0.873 0.972 22. Number of Flowered Tillers 1.225 0.972 23. Percentage of Flowered Tillers -0.419 0.973 24. Flowering Time -0.025 0.973 25. Leaf Blade/Sheath Ratio -0.098 0.973 26. Attitude of Tassel Branches -0.007 0.973 (continued)

PAGE 189

182 Table 4.28 (Continued) Cton Olfc!p f > harar*tpr WllCUCtWlHI CAN l c ASCC 1 1 . I paf Sheath Color -2.21 7 0.831*" O IntprnnHp 1 pnnth II IICI 1 HJLJt! LCI IUU 1 0.575 0.853*** 3 w\ 1 paf Ratio 2.246 0.861*** 4 Plant Rinmac;<5 -0.21 5 0.869*** 5 w\ 1 pnnth of I onoest Primarv Tassel Branch Lvl lull 1 vl Lvl IU vvl 1 1 11 1 IUI y 1 -w> \_> \_/ l ui Ml IWI I 1.149 0.874* Attiti iHp nf Taccpl RrannhPQ rAllllUUC VJI 1 uOOCl Ul Ol Iwl ICO 0.502 0.889* 7 Pprppntanp of Flowprpri Tillprs 0.172 0.893* 8 1 paf Area 2.690 0.895 9 w . Avpranp Tillpr Si7P nVCI uUw 1 lllwl UILC 0.143 0.896 I u. Mi i mhcr of Tillprc INUIIIUcl Ul IIIIUlo -0 066 w. \S\J\J 0 898 I I . -0 094 w. ww^T 0 899 W . Www 19 1 paf RlaHp 1 pnnth Leal Diauc i_ci lyu I -5 282 0 900 1 ^ I o. Plant Wpinht i lain nciyi u -0 223 0 900 1 A Coy of Taccpl Flnwprc Oca kji I aojci rivjwciw 0 194 w. 1 w"T 0 901 W . 1 1 fi Mnmhor of Priman/ Taccpl RrannhPQ INUI 1 lUcl \J\ i 1 II l Id! y I aoocl Dl al iwl ico -0 131 w. 1 \J 1 0 901 w. ww 1 Malp Fprtilitw iviaic i ciLiiiiy 0 070 w \J 1 w 0 902 17 1 # • 1 pnnth of Central Ta*^pl Snikp LCI IMU 1 vl vvl HI dl I aooci UUIrXv -0.588 0.902 18. Tassel Ratio -1 .691 0.904 19. Leaf Sheath Length 0.692 0.904 20. Leaf Blade/Sheath Ratio 0.612 0.905 21. Number of Internodes 0.052 0.905 22. Internode Width 0.009 0.905 23. Number of Flowered Tillers 0.210 0.905 24. Leaf Blade Width -0.109 0.906 25. Internode/Sheath Ratio 0.102 0.906 26. Internode Ratio -0.087 0.906 (continued)

PAGE 190

183 Table 4.28 (Continued) oiep ============ ^========== v^rididoici CAN l d ASCC i . r oTOciudyc ui nuwcicu iiiicio 0 614 0.147*** o c. IntornnHo I onnth -0.237 0.273*** 3 I paf Sheath Color 0.216 0.345*** 4 1 paf Rlarip Width 0.793 0.404*** c Flnworinn Tlmp nuwciiiiy ihmc -0.564 0.443*** e o. Plant RinmaGG l ten ii uiuiiiaoo 0.242 0.458* 7 Mi imhpr nf IntPrnnHPG 0.176 0.471 * Q Av/pranp Tillpr Si7P nvci aye i mci wiz.c 0.210 0.480 q £3. INUIIIUCI w 1 IIIICIO 0.020 0.487 TpcqpI Ratio 1 .827 0.490 I I . 1 annth r\f 1 onnoet Primarw Taccol Rranoh LUIiylll Ul LUliyWoL r 1 III Idly 1 doom Did) loll -2 050 0 499 1 9 Mumhpr of Primarv Ta^cpl RranohpG INLil 1 IUCI \Jl l 1 II 1 IOI y 1 QOOvl LJ 1 C*l Iwl IvQ 0.404 0.507 I O. Oca Ul 1 doocl rlUWcio U, 1 uu 0 51 1 A+titi iHo nf Taccol Rranph^Q /AllllUUC Ul l aoocl Dl CSI IUI ICO -0 173 0 514 HIlClllUUC VVIVJIM 0 365 0 517 16 1 paf Shpath 1 pnnth -0.785 0.518 17 IntprnorJp/Shpath Ratio -0.757 0.519 18. Plant Height o!l93 0^522 19. Internode Ratio 0.319 0.523 20. Length of Central Tassel Spike 0.157 0.524 21. Leaf Ratio 1.945 0.525 22. Leaf Blade Length -2.185 0.535 23. Leaf Area 0.961 0.536 24. Male Fertility 0.047 0.537 25. Number of Flowered Tillers -0.281 0.537 26. Leaf Blade/Sheath Ratio -0.012 0.537 a Data sets from Spring 1989 (W,R 0 ) b Data sets from Spring 1 990 (W, R 0 ) c Data sets from Fall 1 990 (W,R 0 ) d Data sets from Selfed Regenerants (\N,Ri) The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. Marked entries show variables that significantly contribute to the discriminatory capability of the model (* 0.05>P>0.01 ; ** 0.01 >P>0.001 ; ***P<0.001).

PAGE 191

184 Table 4.29 Comparison of average squared canonical correlations, eigenvalues for the first canonical variable and Mahalanobis' generalized (D 2 ) distances between groups from canonical discriminant analyses calculated to discriminate two lea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 100 specimens) and derived from tissue culture (Ro or Ri, 100 specimens each). Data sets from all harvesting seasons, analyzed as separate groups. Data Set ASCO Eigenvalue 6 D 2 * Spring 1 989 (W,R 0 ) Spring 1990 (W,R 0 ) Fall 1990 (W,R 0 ) Selfed Regenerants (W,fl T ) 8 The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. b The eigenvalue represents the ratio of between-class variation to withinclass variation for the canonical variable. 0 The Mahalanobis' D 2 (or generalized distance between two groups) is a measure of the morphometric dissimilarity between the groups. 0.996 272.4 1079.0 0.973 35.4 140.3 0.906 9.6 38.0 0.537 1 .2 4.6

PAGE 192

185 Table 4.30 Statistics of location and dispersion and results of analysis of variance comparing CAN I scores in two Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed {W, 100 specimens) and derived from tissue culture {Ro or Ri, 100 specimens each). Data sets from all harvesting seasons, analyzed as separate groups. Season Control Regenerants Mean a Mean a Spring 1 989 16.42 1.36 -16.42 0.39 53932.08"* Spring 1 990 5.92 1.06 -5.92 0.94 7013.83*" Fall 1990 3.08 1.34 -3.08 0.46 1897.76*** Selfed Regenerants -1.07 1.10 1.07 0.89 229.70*** Marked entries indicate significant differences between means (*** P < 0.001).

PAGE 193

186 ratio, plant height, length of central tassel spike, length of longest primary tassel branch, internode/sheath ratio and leaf sheath length (Table 4.28). With these 12 variables combined, the model again shows a highly significant separation between groups, while also revealing a much shorter distance between them (Fig. 4.14, Tables 4.29 and 4.30). For the first time in Fall 1990 male fertility was no longer an important component of the discriminating function. This showed that the R 0 was now normal (or nearly so) for that trait. Group separation was now strongly influenced by differences in leaf sheath color (ASCC = 0.831 , Table 4.28), together with internode length, leaf ratio, plant biomass and, to a much lesser extent, the length of the longest primary tassel branch, the attitude of tassel branches and the percentage of flowered tillers (Table 4.28). With these seven variables alone segregation between the centroids of the two groups was again highly significant. The discriminating model also revealed, however, an additional reduction in the gap separating them (Fig. 4.14, Tables 4.29 and 4.30). In sharp contrast with the previous harvest seasons, where the discriminating model was strongly dominated by a single trait, definition of the discriminant function for the selfed regenerants and the respective controls was based on a number of weakly represented variables, including the percentage of flowered tillers, internode length, leaf sheath color, leaf blade width, flowering time, plant biomass and number of internodes (Table 4.28). The resulting model was correspondingly much weaker than any of those from previous seasons (total ASCC = 0.537, Table 4.29). While still able to significantly discriminate between groups, the discriminant function revealed that the variability between groups was now of the same order of magnitude as the variability within the groups (eigenvalue = 1 .2, Table 4.29). There is ample overlapping between the individuals of the two groups (Fig. 4.14, Tables 4.29 and 4.30). In sum, the use of multiple discriminant analysis, while of a considerably less descriptive value than the previous methods, leads by and large to the same overall conclusions. Differences between the Ro and seed-derived control plants are

PAGE 194

187 substantial, and based mostly on male fertility and plant size. They are also greatly attenuated with the passage of time. Differences between the R 1 and the respective controls are much less pronounced and of a more diffuse nature, dominated by a shorter stature of the tissue-culture-derived plants and by a more intense average coloration and an earlier flowering time (both traits that, as previously discussed, could be the result of genetic drift in the Ri). Analysis of the Ro . The results of univariate analysis suggested the presence of more than one distinct group in the R 0 population (in particular in Spring 1989), an assumption that was fully confirmed by the subsequent principal components analysis of the Ro data sets. The techniques of multiple discriminant analysis were also applied to the same data sets (Type A and Type B regenerants being the newly defined a priori groups) to further confirm (or otherwise refute) the reality of the existence of such groups. Projection on the first three canonical variables of the individuals of R 0 , with the data sets from all harvest times analyzed as a single group and Type A and Type B regenerants defined as the a priori groups, is shown in Figs. 4.15 and 4.16. Table 4.31 displays the standardized canonical coefficients and the cumulative average squared canonical correlations (ASCC) from the step-wise discriminant analysis showing, in descending order, the morphometric traits that most significantly contributed to the discrimination of the two regenerant types. The generalized distances between group centroids are given for all groups in Table 4.32, and the mean scores of Type A and Type B individuals are compared, for the three canonical variables, in Table 4.33 covering all harvesting times. The magnitude of the differences found in the evolving R 0 population, as it changed over time, was much greater than that found between the two regenerant types within each harvest season, as shown in Figs. 4.15 and 4.16. The three harvest

PAGE 195

Fig. 4.15 Relative similarity of specimens from the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (Ro) derived from tissue culture, as shown by their projection on the first three canonical variables from a multiple discriminant function analysis calculated to discriminate the two regenerant types. Data sets from all harvesting seasons, analyzed as a single group. The specimens from each harvest season are discriminated on the plane of the first two canonical variables. The two regenerant types are resolved, in each season, on the third canonical variable (see Fig. 4.16). Solid circles: Type A Regenerants Open circles: Type B Regenerants

PAGE 196

189

PAGE 197

Fig. 4.16 Histograms of the scores of specimens of the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (flo) derived from tissue culture, on the third canonical variable from a multiple discriminant function analysis calculated to discriminate the two regenerant types. Data sets from all harvesting seasons, analyzed as a single group. Light Gray: Type A Regenerants Dark Gray: Type B Regenerants

PAGE 198

191

PAGE 199

192 Table 4.31 Standardized canonical coefficients and cumulative average squared canonical correlations (ASCC) from a step-wise discriminant analysis calculated to discriminate the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (R 0 ) derived from tissue culture. Data sets from all harvesting seasons, analyzed as a single group. Canonical Variable Step Character ASCC» I II III 1 Flowering Time -0.51 4 1.034 -0.857 0.200 2 Attitude of Tassel Branches 1.152 -0.072 -2.246 0.400 3 Leaf Sheath Color 0.000 0.000 1.159 n CA-r*** 0.547 4 Length of Central Tassel Spike 0.000 0.000 2.893 0.568 5 Internode Width 0.000 0.000 -1 .524 r\ CQO*** 6 Sex of Tassel Flowers 0.000 0.000 -0.232 O.bOo 7 Male Fertility 0.000 0.000 -0.207 0.D21 8 Tassel Ratio 0.000 0.000 5.097 0.637"* 9 Length of Longest Primary Tassel Branch n nnn u.uuu n nnn u.uuu -o.o 1 D 10 Plant Height 0.000 0.000 2.134 0.652"* 11 Number of Tillers 0.000 0.000 0.784 0.654*** 12 Percentage of Flowered Tillers 0.000 0.000 -0.353 0.656*** 13 Number of Internodes 0.000 0.000 -0.195 0.660* 14 Internode Ratio 0.000 0.000 -2.481 0.661 15 Plant Biomass 0.000 0.000 0.123 0.663 16 Leaf Sheath Length 0.000 0.000 0.845 0.666 17 Internode/Sheath Ratio 0.000 0.000 2.886 0.667 18 Internode Length 0.000 0.000 -0.835 0.668 19 Number of Primary Tassel Branches 0.000 0.000 -0.143 0.669 20 Leaf Ratio 0.000 0.000 0.529 0.669 21 Leaf Blade Width 0.000 0.000 0.068 0.670 22 Number of Flowered Tillers 0.000 0.000 0.056 0.670 23 Leaf Blade Length 0.000 0.000 -0.662 0.670 24 Leaf Area 0.000 0.000 0.577 0.670 25 Leaf Blade/Sheath Ratio 0.000 0.000 -0.129 0.670 26 Relative Tiller Size 0.000 0.000 -0.028 0.670 a The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. Marked entries show variables that significantly contribute to the discriminatory capability of the model (* 0.05>P>0.01 ; *** P<0.001).

PAGE 200

193 Table 4.32 Mahalanobis' distances 8 between group centroids from a canonical discriminant analysis calculated to discriminate the two morphological types (A, 80 specimens; B, 20 specimens) found in a lea diploperennis (diploperennial teosinte) population (Ro) derived from tissue culture. Comparative results for all harvesting periods, when all data sets are pooled together and analyzed as a single group. To group From group A(S89) B(S89) A(S90 ) B (S90) A(F90) B(F90) A(S89) 0.0 173.9 1.5 X10 9 1.5 x10 s 2.9 x10 s 2.9 x10 s B(S89) 0.0 1.5x10 s 1.5x10 s 2.9x10 s 2.9x10 s A(S90) 0.0 7.2 1.5 x10 s 1.5 x10 s B(S90) 0.0 1.5 x10 s 1.5 x10 s A(F90) 0.0 4.1 B(F90) 0.0 a The Mahalanobis' D 2 (or generalized distance between two groups) is a measure of the morphometric dissimilarity between the groups. The entries in bold show distances between the centroids of Type A and Type B regenerants for the three harvest seasons.

PAGE 201

194 Table 4.33 Statistics of location and dispersion and results of analysis of variance comparing CAN I, CAN II and CAN III scores for the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (Ro) derived from tissue culture. Data sets from all harvesting seasons, analyzed as a single group. Season Type A regenerants Type B regenerants Mean a Mean a Spring CAN I -0.09 0.00 -0.09 0.00 1989 CAN II -1.41 0.00 -1.41 0.00 CAN III 2.64 0.54 -10.55 2.63 1772.38"* Spring CAN I -1.18 0.00 -1.18 0.00 1990 CAN II 0.78 0.00 0.78 0.00 CAN III -0.14 1.15 0.55 1.10 5!88* Fall CAN I 1.26 0.00 1.26 0.00 1990 CAN II 0.63 0.00 0.63 0.00 CAN III -0.03 0.37 0.13 0.31 3^26 Marked entries indicate significant differences between means C 0.05 > P > 0.01 ; *** P < 0.001).

PAGE 202

195 times for the R 0 are separated on the plane of the first two canonical variables, based exclusively on the unique contributions of two discriminant variables (Table 4.31), flowering time (the plants flowering earlier in Spring 1989 and Spring 1990 than in Fall 1 990) and the attitude of the tassel branches (there is a progressive branching of the tassels with the passage of time). The two regenerant populations (Type A and Type B) are resolved, in each season, on the third canonical variable (Fig. 4.16). Widely different in Spring 1989, the two morphological types gradually merge together over the next two harvest seasons, and are virtually indistinguishable in Fall 1990 (Fig. 4.16, Tables 4.32 and 4.33). The morphometric traits that are significantly involved in the discrimination of Type A and Type B regenerants include primarily leaf sheath color (more intense in Type A), plant height and number of tillers (greater in Type A, in Spring 1989), percentage of flowered tillers and internode width (greater in Type B, in the same harvest season) and, to a much lesser extent, the number of internodes. Significant characteristics of the tassel in the discriminant model include sex of tassel flowers, male fertility, length of the central tassel spike and of the longest primary tassel branch-as well as the corresponding tassel ratio, which reflects the pronounced reduction in development of the tassel and the sex reversal found in Type B regenerants (in Spring 1989). The results of the discriminant analysis, thus, reinforce the validity of the previous assumptions and fully confirm the findings of the univariate and principal components analyses. Application of the discriminant function to the R 0 data sets, analyzed as separate groups, leads basically to the same conclusions. Separation of the two regenerant types is complete (ASCC = 1) in Spring 1989 (Fig. 4.17, Tables 4.35 and 4.36), based solely on differences in leaf sheath color (Table 4.34). Segregation of the two populations is considerably less efficient in subsequent seasons (Fig. 4.17, Tables 4.35 and 4.36). The eigenvalues reveal that in both the Spring and Fall 1990 the differences between the two groups are of the same or a lesser order of magnitude than those

PAGE 203

. c < co co £ _ Q_ CO _ c co © Hi O) Q_ O o ° lots CO c c CD (1) O) ID 2 CO CL O 3 O E o g> £.Q. £ S £ N • m n) x: 3 TS c 0) o 15 c E c o CO "O TJ © 5 co 3 e (0 CO « E o CO CD . CO C CD E o 'o O CD CO CD CO co ~ CM O CD co oo E c CO CD o>E O o II X co c o *= j= o _ §« c c o 0» mi: 73 .S co of IS 2 _g> co -= 5 *= c CO c CD CD C C CD 0J CD 3 tree < CO 0) 0> OCL .>»>« CO co yco — E Q) g" 3 CL cl E ^ _i CD CO ~ Q N «* co

PAGE 205

198 Table 4.34 Standardized coefficients from a canonical discriminant analysis and cumulative average squared canonical correlations (ASCC) from a step-wise discriminant analysis calculated to discriminate the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (R 0 ) derived from tissue culture. Step Character CAN l a ASCC 1 . Leaf Sheath Color 1.000 1 .000"* 2. Plant Height 0.000 1.000 3. Plant Biomass 0.000 1.000 4. Number of Tillers 0.000 1.000 5. Relative Tiller Size 0.000 1.000 6. Number of Flowered Tillers 0.000 1.000 7. Percentage of Flowered Tillers 0.000 1.000 8. Leaf Sheath Length 0.000 1.000 9. Leaf Blade Length 0.000 1.000 10. Leaf Blade Width 0.000 1.000 11. Leaf Area 0.000 1.000 12. Leaf Ratio 0.000 1.000 13. Leaf Blade/Sheath Ratio 0.000 1.000 14. Number of Internodes 0.000 1.000 15. Internode Length 0.000 1.000 16. Internode Width 0.000 1.000 17. Internode Ratio 0.000 1.000 18. Internode/Sheath Ratio 0.000 1.000 19. Flowering Time 0.000 1.000 20. Number of Primary Tassel Branches 0.000 1.000 21. Attitude of Tassel Branches 0.000 1.000 22. Sex of Tassel Flowers 0.000 1.000 23. Length of Central Tassel Spike 0.000 1.000 24. Length of Longest Primary Tassel Branch 0.000 1.000 25. Tassel Ratio 0.000 1.000 26. Male Fertility 0.000 1.000 (continued)

PAGE 206

199 Table 4.34 (Continued) Step Character CAN l b ASCC 1 . Sex of Tassel Flowers 0.475 0.171"* 2. Male Fertility 0.289 0.225* 3, Number of Internodes -0.534 0.262* 4. Leaf Sheath Color 0.357 0.295* 5. Leaf Sheath Lenath -2.644 0.317 6. Internode/Sheath Ratio -2.741 0.323 7. Leaf Area -1 .299 0.326 8. Number of Tillers 5.667 0.334 9. Internode Ratio II IIVI 1 IV\J V 1 IV* H 1.080 0.336 10. Number of Flowered Tillers -6.909 0.339 11. Percentage of Flowered Tillers 1.868 0.367 12. Internode Width 0.942 0.372 13. Plant Height 0.158 0.375 14. Internode Length 1.239 0.377 15. Plant Biomass -0.027 0.377 16. Leaf Blade/Sheath Ratio -0.164 0.378 17. Leaf Blade Width 0.814 0.381 18. Leaf Blade Length 1.623 0.382 19. Length of Central Tassel Spike 0.324 0.382 20. Leaf Ratio -0.371 0.382 21. Average Tiller Size 0.022 0.382 22. Flowering Time 0.000 0.225 23. Number of Primary Tassel Branches 0.000 0.225 24. Attitude of Tassel Branches 0.000 0.225 25. Length of Longest Primary Tassel Branch 0.000 0.225 26. Tassel Ratio 0.000 0.225 (continued)

PAGE 207

200 Table 4.34 (Continued) Step Character CAN l c ASCC 1. Length of Central Tassel Spike -4.786 0.128"* 2. Number of Internodes 0.432 0.178* 3. Leaf Blade/Sheath Ratio -2.386 0.202 4. Number of Tillers -1.075 0.212 5. Number of Flowered Tillers 0.693 0.225 6. Length of Longest Primary Tassel Branch 4.233 0.235 7. Tassel Ratio -4.547 0.248 8. Plant Biomass 0.159 0.254 9. Leaf Blade Width -1 .491 0.258 10. Internode Width 1.625 0.263 11. Internode/Sheath Ratio 1.312 0.268 12. Plant Height -0.254 0.274 13. Leaf Area 1.514 0.276 14. Internode Length -4.362 0.283 15. Internode Ratio 3.692 0.292 16. Leaf Sheath Length -1.114 0.294 17. Number of Primary Tassel Branches 0.077 0.294 18. Male Fertility -0.053 0.295 1 9. Average Tiller Size 0.019 0.295 20. Percentage of Flowered Tillers 0.106 0.295 21. Leaf Ratio -0.416 0.295 22. Leaf Blade Length 0.336 0.295 23. Leaf Sheath Color 0.000 0.295 24. Flowering Time 0.000 0.295 25. Attitude of Tassel Branches 0.000 0.295 26. Sex of Tassel Flowers 0.000 0.295 a Data sets from Spring 1 989 {W,R 0 ) b Data sets from Spring 1 990 {W, R 0 ) c Data sets from Fall 1 990 (W,R 0 ) The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. Marked entries show variables that significantly contribute to the discriminatory capability of the model (* 0.05>P>0.01 ; *** P<0.001).

PAGE 208

201 Table 4.35 Comparison of average squared canonical correlations, eigenvalues for the first canonical variable and Mahalanobis' generalized (D 2 ) distances between groups from canonical discriminant analyses calculated to discriminate the two morphological types (A, 80 specimens; B, 20 specimens) found in a lea diploperennis (diploperennial teosinte) population (R 0 ) derived from tissue culture. Data sets from all harvesting seasons, analyzed as separate groups. Data Set ASCC a Eigenvalue 6 [)2c Spring 1 989 1.000 9.4 x10 s Spring 1 990 0.382 0.6 3.8 Fall 1990 0.295 0.4 2.6 a The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. b The eigenvalue represents the ratio of between-class variation to withinclass variation for the canonical variable. 0 The Mahalanobis' D 2 (or generalized distance between two groups) is a measure of the morphometric dissimilarity between the groups.

PAGE 209

202 Table 4.36 Statistics of location and dispersion and results of analysis of variance comparing CAN I scores for the two morphological types (A, 80 specimens; B, 20 specimens) found in a Zea diploperennis (diploperennial teosinte) population (Ro) derived from tissue culture. Data sets from all harvesting seasons, analyzed as separate groups. Season Type A regenerants Type B regenerants Mean o Mean o Spring 1 989 0.50 0.00 -1.99 0.00 00 "* Spring 1990 -0.39 1.00 1.56 1.01 60.70*" Fall 1990 0.32 1.04 -1 .28 0.80 41 .00"* Marked entries indicate significant differences between means (*** P < 0.001).

PAGE 210

203 found within a single group (Table 4.35). The morphometric traits used to discriminate the two morphological types are the sex of the tassel flowers (and, to a much lesser extent, male fertility, number of internodes and leaf sheath color) in Spring 1 990 (Table 4.34), and the length of the central tassel spike (and the number of internodes) in Fall 1990 (Table 4.34). The Effect of Gibberellic Acid Mutants are known in maize that produce a reduced stature in the plant. In many such mutants, this characteristic may also be associated with morphological changes in the tassel (Coe and Poethig 1982). The fact that some of these aberrant plants can be normalized by applications of gibberellic acid, an endogenous plant growth regulator that is known to influence both cell elongation and sex expression in maize, prompted an experiment to evaluate the effect of gibberellic acid applications on the expression of the abnormal traits in tissue-culture-derived plants. A group of ten plants was selected among both the Ro and the control (W) populations, and the individual clumps were split and potted into separate gallon containers. This provided for an experimental and control subsample of the original populations, constituted by individuals of exactly the same genotype. Type B specimens were selected in the R 0 , since they represented the most extreme expression of the aberrant phenotype. The plants were grown for a month, cut back and allowed to sprout again under standard cultural conditions. The two groups of plants (10 specimens each) of both the R 0 and W were then sprayed with either a 1 0" 4 M solution of GA 3 or plain deionized water, as a control. Plants were sprayed to run off, and no attempt was made to quantify the exact amount of chemical absorbed. A surfactant was added to the solutions, in order to facilitate absorption of the growth regulator by the leaves. Spray applications were done twice a week for four weeks. The plants were then allowed to grow to maturity, when they were assessed for the usual set of 26

PAGE 211

204 morphometry traits. The experiment was repeated and similar results were obtained both times. Univariate comparisons . Means, standard deviations and coefficients of variation for the control and treated populations (both IV and R 0 ), are presented in Tables 4.37 and 4.38, respectively. The percent difference between means, as well as the F values (used to test the significance of the differences between character means) are shown, for all paired populations, in Tables 4.39 through 4.43. Considerable differences are found in most vegetative and floral (tassel) morphometric traits between the untreated tissue-culture-derived plants and the seedderived controls (Table 4.39). Not unexpectedly, these basically follow, with minor variations, the pattern previously described for those two populations in Spring 1989 (Table 4.3). Significant differences are also found between the untreated and the GA3treated Ro plants (Table 4.40), the hormonal treatment affecting, one way or another, most of the vegetative and floral traits under assessment. What was unanticipated, however, was that these differences were comparable, for most morphometric traits, to those found between the R 0 and the control (cf., Tables 4.39 and 4.40). This indicates that the dwarf sterile regenerants tend to become virtually identical to the seed-derived controls as a result of the GA3 treatment, a conclusion that is fully substantiated by pairwise comparisons of the treated R 0 plants and the untreated and treated controls (Tables 4.41 and 4.42, respectively). The most significant differences found between the GA 3 -treated R 0 and the seed-derived plants involve leaf sheath color, average tiller size and the attitude of the tassel branches (all characters likely to be genetically different in the two populations) as well as the tassel ratio, most other traits being statistically identical in the two populations. The univariate analyses indicate little effect from the GA3 application on the seed-derived controls (Table 4.43). Normalization at both the vegetative and floral levels is thus almost complete as a result of gibberellin

PAGE 212

205 Table 4.37 Statistics of location and dispersion for two Zea diploperennis (diploperennial teosinte) control populations (20 specimens each), raised from wild collected seed, that were either left untreated or were treated with a 10" 4 M solution of gibberellic acid (GA3) as a foliar spray. Untreated Controls GA3-Treated Controls Character Mean 0 CV Mean 0 CV 1 1 88.40 39.56 21.00 198.35 56.03 28.25 2 200.55 54.84 27.35 233.75 58.30 24.94 3 13.05 3.75 28.72 10.70 3.16 29.57 4 3.70 0.47 12.71 3.30 0.47 14.25 5 9.90 3.23 32.59 9.30 3.13 33.66 6 76.10 12.39 16.28 86.75 12.20 14.06 7 1.60 1.05 65.39 1.60 1.05 65.39 8 101.15 23.64 23.37 103.80 28.12 27.09 9 344.45 93.88 27.25 355.70 98.52 27.70 10 23.21 4.71 20.28 23.99 4.86 20.25 11 60.35 21 .93 36.34 64.20 23.42 36.47 1 iL i o.4y 0.0b o~7 on 0/.0U i o.4y R OK o.yo 00. 4o 13 3.49 0.96 27.50 3.55 0.98 27.68 14 8.50 1.15 13.50 8.45 1.05 12.43 15 283.40 70.52 24.88 31 1 .45 83.59 26.84 16 3.29 1.05 31.89 2.81 1.01 36.14 17 96.00 42.67 44.45 127.85 64.72 50.62 18 2.91 0.86 29.55 3.17 1.08 34.04 19 2.05 0.76 37.03 2.10 0.72 34.20 20 2.05 0.76 37.03 2.20 0.77 34.90 21 2.05 0.76 37.03 2.05 0.76 37.03 22 2.00 0.00 0.00 2.00 0.00 0.00 23 131.75 35.50 26.94 139.15 36.80 26.45 24 91.15 23.64 25.94 105.90 20.82 19.66 25 0.73 0.24 32.47 0.80 0.24 29.36 26 92.75 4.66 5.02 91.00 4.98 5.48 Average CV 27.78 28.27

PAGE 213

206 Table 4.38 Statistics of location and dispersion for two Zea diploperennis (diploperennial teosinte) populations derived from tissue culture (Ro, 20 specimens each) that were either left untreated or were treated with a 1 0 4 M solution of gibberellic acid (GA3) as a foliar spray. Untreated Regenerants (R 0 ) G/VrTreated Regenerants Character Mean 0 cv Mean O CV 1 57.65 11.44 19.85 179.05 10.76 6.01 2 176.85 38.32 21.67 184.15 36.80 19.99 3 18.25 4.33 23.71 9.60 2.74 28.56 4 4.00 0.00 0.00 3.00 0.00 0.00 5 14.55 6.17 42.40 7.60 2.80 36.82 6 77.95 22.05 28.29 79.55 17.17 21.59 7 1.00 0.00 0.00 3.00 0.00 0.00 8 72.65 4.82 6.63 100.75 14.93 14.82 9 192.05 41 .27 21 .49 351 .45 83.59 23.78 10 21.71 4.42 20.37 21 .71 4.13 19.01 11 32.05 11.69 36.48 58.60 21 .86 37.31 12 8.98 1.73 19.31 16.33 3.29 20.17 13 2.66 0.64 24.00 3.60 1.14 31.81 14 7.20 0.95 13.21 7.45 0.94 12.68 15 55.15 6.12 11.10 244.05 27.21 11.15 16 2.94 0.67 22.98 2.88 0.73 25.45 17 19.85 5.82 29.30 91.00 30.66 33.69 18 0.76 0.09 12.09 2.46 0.39 15.68 19 2.00 0.00 0.00 2.00 0.00 0.00 20 0.00 0.00 2.45 0.51 20.83 21 0.00 0.00 1.00 0.00 0.00 22 1.00 0.00 o!oo 2.00 0.00 0.00 23 47.85 10.47 21.89 131 .40 29.60 22.53 24 0.00 0.00 110.80 17.91 16.17 25 0.00 0.00 0.26 0.41 157.47 26 0.00 0.00 91.70 5.19 5.66 Average CV 17.85 22.35

PAGE 214

207 Table 4.39 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, one raised from wild collected seed (W, 20 specimens) and one derived from tissue culture (Ro, 20 specimens). Values sorted by percent difference between means. Character Mean (W) Mean (Ro) % F 5 9.90 14.55 47.0 8.92** 3 13.05 18.25 39.8 16.50*** 4 3.70 4.00 8.1 8.14" 6 76.10 77.95 2.4 0.11 10 23.21 21 .71 -6.5 1.08 16 3.29 2.94 -10.6 1.58 2 200.55 176.85 -11.8 2.51 14 8.50 7.20 -15.3 1 5.22*** 13 3.49 2.66 -23.8 10.32" 8 101.15 72.65 -28.2 27.91*** 7 1.60 1.00 -37.5 6.58* 12 15.49 8.98 -42.0 22.71*** 9 344.45 1 92.05 -44.2 44.1 r** 11 60.35 32.05 -46.9 25.93*** 1 188.40 57.65 -69.4 201 .64*** 18 2.91 0.76 -73.9 123.28*** 17 96.00 19.85 -79.3 62.54*** 15 283.40 55.15 -80.5 207.93*** 19 2.05 2.00 -2.4 0.09 22 2.00 1.00 -50.0 00 *" 23 131.75 47.85 -63.7 102.78"* 20 2.05 0.00 -100.0 145.84*" 21 2.05 0.00 -100.0 145.84*" 25 0.73 0.00 -100.0 189.67*" 24 91.15 0.00 -100.0 297.33*** 26 92.75 0.00 -100.0 7939.22*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; "* P < 0.001).

PAGE 215

208 Table 4.40 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations derived from tissue culture (Ro, 20 specimens each) that were either left untreated or were treated with a 1Q 4 M solution of gibberellic acid (GA3) as a foliar spray. Character Mean (Ro/Gfi*) Mean (Ro) % F 5 7.60 14.55 91.4 21 .06"* 3 9.60 18.25 90.1 57.03"* 4 3.00 4.00 33.3 *** 00 16 2.88 2.94 2.1 0.05 10 21 .71 21.71 0.0 0.00 6 79.55 77.95 -2.0 0.07 14 7.45 7.20 -3.4 0.70* 2 184.15 176.85 -4.0 0.38 13 3.60 2.66 -26.1 10.12" 8 100.75 72.65 -27.9 64.14*** 12 16.33 8.98 -45.0 77.93*** 11 58.60 32.05 -45.3 22.94*** 9 351 .45 192.05 -45.4 58.48*** 7 3.00 1.00 -66.7 00 "* 1 179.05 57.65 -67.8 1194.76*** 18 2.46 0.76 -69.1 367.12*** 15 244.05 55.15 -77.4 917.34*** 17 91.00 19.85 -78.2 103.99*** 19 2.00 2.00 0.0 22 2.00 1.00 -50.0 00 "* 23 131.40 47.85 -63.6 141.61*** 25 0.26 0.00 -100.0 8.07** 20 2.45 0.00 -100.0 460.80*** 24 110.80 0.00 -100.0 765.13*** 26 91.70 0.00 -100.0 6238.54*** 21 1.00 0.00 -100.0 00 — Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 216

209 Table 4.41 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, one raised from wild collected seed (IV, 20 specimens) and one derived from tissue culture and treated with a 1 0" 4 M solution of gibberellic acid as a foliar spray (Ro/Gf^). Values sorted by percent difference between means. Character Mean (W) Mean {Ro/Gfi*) % F 7 1.60 3.00 87.5 35.81*" 12 15.49 16.33 5.4 0.31 6 76.10 79.55 4.5 0.53 13 3.49 3.60 3.2 0.10 9 344.45 351 .45 2.0 0.06 8 101.15 100.75 -0.4 0.00 11 60.35 58.60 -2.9 0.06 1 188.40 179.05 -5.0 1.04 17 96.00 91.00 -5.2 0.18 10 23.21 21.71 -6.5 1.16 2 200.55 184.15 -8.2 1.23 14 8.50 7.45 -12.4 9.99" 16 3.29 2.88 -12.5 1.96 15 283.40 244.05 -13.9 5.42* 18 2.91 2.46 -15.5 4.46* 4 3.70 3.00 -18.9 44.33*** 5 9.90 7.60 -23.2 5.80* 3 13.05 9.60 -26.4 11.04" 24 91.15 110.80 21.6 8.78 20 2.05 2.45 19.5 3.82 22 2.00 2.00 0.0 23 131.75 131.40 -0.3 0.00 26 92.75 91.70 -1.1 0.45 19 2.05 2.00 -2.4 0.09 21 2.05 1.00 -51 .2 38.26*** 25 0.73 0.26 -64.4 1 9.89*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 217

210 Table 4.42 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) populations, one raised from wild collected seed (W/GA3) and one derived from tissue culture (R0/GA3), treated with a 10" 4 M solution of gibberellic acid (GA3) as a foliar spray. Values sorted by percent difference between means. Character Mean (W/GA3) Mean (R0/GA3) % F 7 1.60 3.00 87.5 35.81"* 12 15.49 16.33 5.4 0.31 16 2.81 2.88 2.8 0.08 13 3.55 3.60 1.4 0.02 9 355.70 351 .45 -1 .2 0.02 8 103.80 100.75 -2.9 0.18 6 86.75 79.55 -8.3 2.34 11 64.20 58.60 -8.7 0.61 4 3.30 3.00 -9.1 8.14" 10 23.99 21.71 -9.5 2.56 1 198.35 179.05 -9.7 2.29 3 10.70 9.60 -10.3 1.38 14 8.45 7.45 -11.8 10.03" 5 9.30 7.60 -18.3 3.28 2 233.75 184.15 -21 .2 10.35" 15 31 1 .45 244.05 -21.6 1 1 .76" 18 3.17 2.46 -22.3 7.60" 17 127.85 91 .00 -28.8 5.30* 20 2.20 2.45 11.4 1.47 24 105.90 110.80 4.6 0.64 26 91.00 91.70 0.8 0.19 22 2.00 2.00 0.0 19 2.10 2.00 -4.8 0.39 23 139.15 131.40 -5.6 0.54 21 2.05 1.00 -51 .2 38.26*** 25 0.80 0.26 -67.8 26.88*** Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001 ; *** P < 0.001).

PAGE 218

211 Table 4.43 Means, percent difference between means, and results of analysis of variance comparing vegetative (top) and floral (bottom) traits in two Zea diploperennis (diploperennial teosinte) control populations raised from wild collected seed, that were either left untreated (IV, 20 specimens) or were treated with a 1 0 4 M solution of gibberellic acid (GA3) as a foliar spray (W/GA3, 20 specimens). Values sorted by percent difference between means. Character Mean (W) Mean (W/GA3) % F 17 96.00 127.85 33.2 3.38 2 200.55 233.75 16.6 3.44 6 76.10 86.75 14.0 7.51** 15 283.40 31 1 .45 9.9 1.32 18 2.91 3.17 8.9 0.72 11 60.35 64.20 6.4 0.29 1 188.40 198.35 5.3 0.42 10 23.21 23.99 3.4 0.26 9 344.45 355.70 3.3 0.14 8 101.15 103.80 2.6 0.10 13 3.49 3.55 1.7 0.03 7 1.60 1.60 0.0 0.00 12 15.49 15.49 0.0 0.00 14 8.50 8.45 -0.6 0.02 5 9.90 9.30 -6.1 0.36 4 3.70 3.30 -10.8 7.24 16 3.29 2.81 -14.6 2.17 3 13.05 10.70 -18.0 4.59* 24 91.15 105.90 16.2 4.38 25 0.73 0.80 9.6 1.04 20 2.05 2.20 7.3 0.39 23 131 .75 139.15 5.6 0.42 19 2.05 2.10 2.4 0.05 21 2.05 2.05 0.0 0.00 22 2.00 2.00 0.0 26 92.75 91.00 -1.9 1.32 Marked entries indicate significant differences between means (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001).

PAGE 219

B c I c c I Q. x> o 8 c o g o ! Q. I CO < CD o 0 3= 0 c CD E cl o CD > 0 Q S3 CO o 2 8 £ "5 O Q) -a g co c 0 ^ 0^ _J cc 0 -*— ' c co o 0 c c 0 CL o Q. XI O W c J5 CL C o w c o "co o Q. CL CO CO < CD co -»— < o 0 3= 0 "co • c 0 E CL o 0 > 0 Q co 0 c £ co o J co 3 o o 'c 0 o cn c 0 o ° c f 0 2 e z o c >iz jo CL QC | p g c 3 CO 0 o CO -O "D 0 C 0 > S> CO co c JO CL -a 0 > c 0 "a 3 o 0 3 <0 • .52 c s E C 0 ii o o co E 0 o II I? 0 E < ° CO XT CO CT)

PAGE 220

213

PAGE 221

3 c C 2 o 9O ! a. g o I CL CL ro CO < 03 » C 0 E CL o CD > CD Q E I J2 I c 3 i I ~o c _co Qc -t (0 C CD O i= <-> 5 CD CD _ x: "a ^ a) o "co CO 0) 3 Sit E rajf 2 2 E LL |— CD CD -*— ' C CO o 'c c 9! CD CL O Cl O to c Cl C o to c o ' to o "5. cl CO CO < o § CD 3= CD "cO c CD E Cl O CD > CD Q T3 C CO CD N to i— ' O O ._: o c to -b CD CO CO CD « 0) " § E -c 2 CD TJ £ li j5 ~ (0 3 i= 8 W fD 3 -C £ li * o x tr * a ri> 2 Li. Cl o I '(0 « 0) CD XJ CJ) J. CD < > o C\J LL cb

PAGE 222

215

PAGE 223

216 application, demonstrating that the previously found effects of time (and passage through a sexual cycle) can be replaced simply by a hormonal treatment of the anomalous plants (Figs. 4.1 8 through 4.21). The effect of GA3 treatment seems to be permanent on the tissue-culture-derived plants. Repeatedly cutting these back after the original treatment consistently yielded normal plants at every new cycle of regrowth. Principal components analysis . Ordination on the first two principal components of individuals of both treated and untreated R 0 and controls is shown in Fig. 4.22. The associated factor loadings and percents of trace are given in Table 4.44, and Table 4.45 details both the vegetative and floral traits that load most significantly (P < 0.005) on the first two components, in decreasing order of magnitude. Mean scores for the first two principal components of individuals of all populations are compared in Table 4.46. The results of the univariate analysis are fully confirmed by the principal components analysis, whose graphical output clearly illustrates the fading of the differences between the tissue-culture-derived plants and the controls as a result of hormonal treatment (Fig. 4.22). There are no significant differences between the treated and untreated controls, as well as between both these and the treated regenerants, with respect to their PC I scores (Table 4.46). Neither are there significant differences in mean PC II scores among any of the populations under study (Table 4.46). The untreated regenerants, however, differ from all the other populations in their mean PC I scores at a highly significant level (Table 4.46). PC I, therefore, is the discriminating component in the analysis, revealing a common multivariate identity among the treated R 0 and the treated and untreated controls. A graphical depiction of the fate of the individual quantitative traits that load most significantly on PC I (those that contribute the most to the anomalous phenotype of the R 0 , Table 4.45) is shown, as they change in response to GA3 applications, in Figs. 4.23 through 4.31 . The trend is obvious.

PAGE 224

Fig. 4.22 Ordination of individuals from Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (IV, 20 specimens) and derived from tissue culture {Ro, 20 specimens), before and after application of gibberellin (GA 3 ) as a foliar spray, on the first two principal components extracted from a matrix of correlation coefficients between twenty six morphometric variables. Data sets from all populations, analyzed as a single group. Dark Gray: Untreated controls Light Gray: GA3-treated controls Black: Untreated regenerants White: GA3-treated regenerants

PAGE 225

218 «t 2 d LU O | 0 o o _l < o z £2 1 — 1 1 1 : • *a 1 1 © • © ® © • •v •• oo ° r o «°* _ w o •• • • ® G i i i 0 • o® o 0 o • o ® ® ® • B t -4-2 0 2 PRINCIPAL COMPONENT I (44.9%)

PAGE 226

219 Table 4.44 Factor loadings and percent of trace from principal components analysis of Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed {W, 20 specimens) and regenerated from tissue culture (Ro, 20 specimens), that were either left untreated or were treated with a 10 4 M solution of gibberellic acid (GA3) as a foliar spray. Principal Component Character I II III 1 . Plant Height 0.266 0.038 0.091 2. Plant Biomass 0.088 0.280 0.212 3. Number of Tillers -0.177 0.267 0.228 4. Average Tiller Size -0.173 0.1 47 U.UlD 5. Number of Flowered Tillers -0.1 OO c\ 000** U.ooU 6. Percentage of Flowered Tillers 0.039 -0.01 0 0.353 7. Leaf Sheath Color 0.149* 0.033 -0.228** 8. Leaf Sheath Length 0.167* 0.317** 0.199* 9. Leat Blade Lengtn 0.<24l f\ 007** -0. 1 Oo 10. Leaf Blade Width 0.080 -0.173* -0.195* 11. Leaf Area 0.217** 0.086 -0.181* 12. Leaf Ratio 0.183* 0.353** 0.022 13. Leaf Blade/Sheath Ratio 0.163* -0.056 -0.269** 14. Number of Internodes 0.129 -0.145 -0.001 15. Internode Length 0.265** 0.052 0.127 16. Internode Width 0.028 0.441** -0.257** 17. Internode Ratio 0.193* -0.263** 0.282** 18. Internode/Sheath Ratio 0.235** -0.192* 0.053 19. Flowering Time -0.014** -0.344** 0.272" 20. Number of Primary Tassel Branches 0.257** 0.013 -0.010 21. Attitude of Tassel Branches 0.224** -0.008 0.128 22. Sex of Tassel Flowers 0.275** -0.053 0.061 23. Length of Central Tassel Spike 0.252** -0.029 -0.125 24. Length of Longest Primary Tassel Branch 0.274** 0.017 -0.057* 25. Tassel Ratio 0.176* 0.033 0.338" 26. Male Fertility 0.277" -0.045 0.044 Percent of trace 44.8 10.4 7.8 Marked entries show variables that are significantly represented in each principal component (* 0.05 > P > 0.01 ; ** 0.01 > P > 0.001).

PAGE 227

220 Table 4.45 Characters most significantly represented (P < 0.005) in the first two axes of a principal components analysis of Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 20 specimens) and regenerated from tissue culture (Ro, 20 specimens), that were either left untreated or were treated with a 1 0 4 M solution of gibberellic acid (GA3) as a foliar spray. Component Character Loading PC I Plant Height Internode Length Leaf Blade Length Internode/Sheath Ratio Leaf Area 0.266 0.265 0.241 0.235 0.217 (44.8%) Male Fertility Sex of Tassel Flowers 0.277 0.275 0.274 0.257 0.252 PC II Internode Width Leaf Ratio Flowering Time Leaf Sheath Length Plant Biomass Number of Tillers Internode Ratio Leaf Blade Length Number of Flowered Tillers 0.441 0.353 -0.344 0.317 0.280 0.267 -0.263 0.237 0.233 (10.4%)

PAGE 228

221 Table 4.46 Statistics of location and dispersion and results of analysis of variance comparing PC I and PC II scores in Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 20 specimens) and derived from tissue culture (Ro, 20 specimens), that were either left untreated or were treated with a 1 0 4 M solution of gibberellic acid (GA3) as a foliar spray. Ro W W/GA3 fl 0 /GA3 Mean 0 Mean a F Mean a F Mean a F PCI PC II -5.52 0.25 0.42 0.70 1.55 0.22 1.33 2.12 513.92*" 0.00 2.39 1.45 549.53"* -0.27 2.29 0.92 1.58 1.11 719.08*** -0.20 0.91 3.03 fl 0 /GA3 W W/GA3 Mean a Mean 0 F Mean 0 F PCI PC II 1.58 1.11 -0.20 0.91 1.55 0.22 1.33 2.12 0.01 0.66 2.39 1.45 -0.27 2.29 3.95 0.01 W I/V/GA3 Mean 0 Mean a F PCI 1.55 1.33 2.39 1.45 3.65 PC II 0.22 2.12 -0.27 2.29 0.49 Marked entries indicate significant differences between means (*** P < 0.001).

PAGE 229

Fig. 4.23 Effect of GA3 applications on male fertility of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (Ro). Fig. 4.24 Effect of GA3 applications on the length of the central tassel spike of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture {Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Seed raised controls (W)

PAGE 230

223 MALE FERTILITY CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.23 CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.24

PAGE 231

Fig. 4.25 Effect of GA3 applications on the number of primary tassel branches of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (Ro). Fig. 4.26 Effect of GA3 applications on the length of the longest primary tassel branch of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Seed raised controls (W)

PAGE 232

225 NUMBER OF PRIMARY TASSEL BRANCHES CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.25 LENGTH OF LONGEST PRIMARY TASSEL BRANCH CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.26

PAGE 233

Fig. 4.27 Effect of GA3 applications on the plant height of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (R 0 ). Fig. 4.28 Effect of GA3 applications on the internode length of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture {Ro). Fig. 4.29 Effect of GA3 applications on the internode / sheath ratio of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Seed raised controls (W)

PAGE 234

227 CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.27 INTERNODE LENGTH CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.28 INTERNODE / SHEATH RATIO T T CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.29

PAGE 235

Fig. 4.30 Effect of GA3 applications on the leaf blade length of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (Ro). Fig. 4.31 Effect of GA3 applications on the leaf area of plants of Zea diploperennis (diploperennial teosinte), raised from seed (W) or derived from tissue culture (Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Seed raised controls (W)

PAGE 236

229 CONTROL + GA3 REGENERANTS REGENERANTS + GA3 POPULATION / TREATMENT Fig. 4.30 Fig. 4.31

PAGE 237

230 Canonical discriminant analysis . Projection on the first three canonical variables of individuals from all treated and untreated populations is shown in Fig. 4.32. Table 4.47 displays the associated standardized canonical coefficients, and shows the cumulative average squared canonical correlations from the step-wise discriminant analysis, detailing the contribution of the various morphometric traits to the discriminatory capability of the model. Generalized distances between group centroids are given for all groups in Table 4.48. Mean scores for the three canonical variables of individuals from all populations are compared in Table 4.49. The magnitude of the differences between the untreated regenerants and all the other populations leads to their complete separation on the first canonical variable (Fig. 4.32, Tables 4.48 and 4.49), based solely on the number of flowered tillers (Table 4.47). The remaining three populations are projected (and resolved) on the plane of the second and third canonical variables (Fig. 4.32, Tables 4.48 and 4.49). The discrimination is based on differences in leaf blade width, male fertility, plant biomass, length of central tassel spike, internode ratio and, to a lesser extent, length of the longest primary tassel branch, attitude of the tassel branches and tassel ratio (Table 4.47). Differences for the mean CAN II and CAN III scores are significant for the three groups (Table 4.49). The generalized distances among the three groups, however, are several orders of magnitude smaller than those between any of the individual groups and the untreated R 0 population (Table 4.48 and 4.49). Caution must be taken when interpreting the results from discriminant analyses. As previously noted, the discriminant function is often weighed to such a heavy dependence on just a few variables, in complete detriment of other more relevant (in biological terms) characters, that it ends up contributing considerably less to the analysis than do principal components. This is because the method of principal components, unlike discriminant analyses, is more sensitive to the covariation in large numbers of characters, and relatively insensitive to unique extremes in individual

PAGE 238

Fig. 4.32 Relative similarity of specimens from lea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 20 specimens) and derived from tissue culture (Ro, 20 specimens), before and after application of gibberellin (GA3) as a foliar spray, as shown by their projection on the first three canonical variables from a multiple discriminant function analysis calculated to discriminate the four groups. Data sets from all populations, analyzed as a single group. The dwarf, sterile R 0 plants are discriminated from normal fertile plants on the first canonical variable. All other populations are projected (and resolved) on a single separate plane, defined by the second and third canonical variables. Dark Gray: Untreated controls Light Gray: GA3-treated controls Black: Untreated regenerants White: GA3-treated regenerants

PAGE 239

232

PAGE 240

233 Table 4.47 Standardized canonical coefficients and cumulative average squared canonical correlations (ASCC) from a step-wise discriminant analysis calculated to discriminate Zea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (IV, 20 specimens) and regenerated from tissue culture (R 0 , 20 specimens), that were either left untreated or were treated with a 1 CH M solution of gibberellic acid (GA3) as a foliar spray. Canonical Variable Step Character ASCC a I II III 1 Number of Flowered Tillers 1.000 1 .507 0.144 0.333 2 Leaf Blade Width 0.000 3.596 -0.741 0.467 3 Male Fertility 0.000 A on-i -1 .301 f\ f\A O 0.018 0.549 4 Plant Biomass 0.000 A CA & 1.516 r\ HOC -0.985 0.601 5 Length of Central Tassel Spike 0.000 0.307 1 .040 0.670 6 Internode Ratio (J.UUU n a no U.4UO n con'** U.boU 7 Leaf Area 0.000 -1 .932 -1 .880 0.706" 8 Length of Longest Primary Tassel Branch 0.000 2.295 2.630 0.736* 9 Attitude of Tassel Branches 0.000 0.734 -1.153 0.755* 10 Tassel Ratio 0.000 -3.240 0.337 0.779* 11 Leaf Sheath Color 0.000 0.235 1.777 0.789 12 Internode Width 0.000 -0.164 -1.161 0.797 13 Number of Internodes 0.000 -2.097 0.334 0.799 14 Internode/Sheath Ratio 0.000 1.808 -0.760 0.805 15 Leaf Sheath Length 0.000 1.087 -0.125 0.808 16 Leaf Ratio 0.000 -1 .360 -0.425 0.810 17 Leaf Blade Length 0.000 1.478 0.108 0.813 18 Leaf Blade/Sheath Ratio 0.000 0.700 0.276 0.816 19 Percentage of Flowered Tillers 0.000 1.301 -2.059 0.823 20 Plant Height 0.000 1.272 -0.889 0.831 21 Relative Tiller Size 0.000 -2.628 4.094 0.836 22 Flowering Time 0.000 -0.436 1.135 0.842 23 Internode Length 0.000 -0.956 -1.074 0.843 24 Number of Primary Tassel Branches 0.000 -0.038 0.237 0.845 25 Number of Tillers 0.000 -1 .664 -0.300 0.845 26 Sex of Tassel Flowers 0.000 0.185 0.296 0.846 a The Average Squared Canonical Correlation (ASCC) can vary between 0 and 1. It is closer to 1 when the discriminant model shows good separation for the groups. Marked entries show variables that significantly contribute to the discriminatory capability of the model (* 0.05>P>0.01 ; *** P<0.001).

PAGE 241

234 Table 4.48 Mahalanobis' distances 8 between group centroids from a canonical discriminant analysis calculated to discriminate lea diploperennis (diploperennial teosinte) populations, raised from wild-collected seed (W, 20 specimens) and regenerated from tissue culture (R 0 , 20 specimens), that were either left untreated or were treated with a 10" 4 M solution of gibberellic acid (GA3) as a foliar spray. To group From group W/GA3 Ro R0/QA3 W/GA3 Ro R0/GA3 0.0 11.3 0.0 9.1 x 10 8 9.1 x 10 8 91.5 94.1 0.0 9.1 X10 8 0.0 8 The Mahalanobis' D 2 (or generalized distance between two groups) is a measure of the morphometric dissimilarity between the groups.

PAGE 242

235 Table 4.49 Statistics of location and dispersion and results of analysis of variance comparing CAN I and CAN II scores in Zea diploperennis (diploperennial teosinte) populations, raised from wild collected seed (W, 20 specimens) and derived from tissue culture (Ro, 20 specimens), that were either left untreated or were treated with a 1 0" 4 M solution of gibberellic acid (GA3) as a foliar spray. Ro W IV/GA3 R0/GA3 Mean a Mean o F Mean o F Mean o F CAN I -1.72 0.00 0.57 0.00 00 *** 0.57 0.00 » *** 0.57 0.00 00 CAN II 0.00 0.77 3.08 1.13 101.69*" 3.24 1.05 124.50*" -6.32 1.02 493.04 CAN III 0.00 0.88 1.69 1.32 22.85*" -1.66 1.06 29.08*" -0.03 0.60 0.01 fl 0 /GA3 W IV/GA3 Mean o Mean o F Mean o F CAN I 0.57 0.00 0.57 0.00 0.57 0.00 CAN II -6.32 1.02 3.08 1.13 765.04*" 3.24 1.05 859.10 CAN III -0.03 0.60 1.69 1.32 28.16*" -1.66 1.06 35.76 W W/GA3 Mean o Mean o F CAN I 0.57 0.00 0.57 0.00 CAN II 3.08 1.13 3.24 1.05 0.20 CAN III 1.69 1.32 -1.66 1.06 78.53' Marked entries indicate significant differences between means (*** P < 0.001).

PAGE 243

236 characters, the ones on which discriminant analysis depends. Even so, although the coefficients and loadings cannot be compared in detail, because of the differences in scaling between the two methods, the overall conclusions of the multiple discriminant analysis again closely parallelled (and, in a way, complemented) those from principal components-in short, that a simple application of GA3 to the dwarf, sterile Ro plants of diploperennial teosinte leads to complete recovery from the traumatic effects of the passage through tissue culture, and entirely replaces the healing effects of time or a sexual cycle. Discussion and Conclusions The Statistical Methodology Large data matrices, such as those analyzed in this chapter, often contain so many interrelationships that it is almost impossible to interpret them accurately at first sight. When this is the case, researchers in a rapidly expanding number of scientific fields have been resorting to multivariate descriptive statistical analysis, a designation that collectively embraces a number of methods excelling in their ability to analyze and describe large undifferentiated data sets. Statisticians often refer to these methods as exploratory, in that their goal is to obtain a parsimonious description of a large amount of data. The techniques, however, can also be used to analyze, verify and test certain hypotheses (Lebart et al. 1984). In the current study, two of these methods were employed, in addition to the relatively simple techniques of univariate statistical analysis. Principal components analysis was used to explore variation among the highly correlated morphometric characters and to outline the structure of the data, and provided a parsimonious multivariate summary of the morphological relationships within and among the groups. Canonical discriminant analysis (and a corresponding step-wise discriminant analysis)

PAGE 244

237 was used to identify the linear combination of characters that maximized differences between a priori groups and to reveal the relative importance of characters in effecting that separation. The a priori groups were the tissue-culture-derived plants (Ro or Ri) and the seed-derived controls (W), analyzed separately at each harvest. The uses and statistical foundations of principal components analysis (as well as multiple discriminant analysis, see below) are discussed in most classic multivariate statistics or numerical taxonomy texts (e.g., Seal 1964; Morrison 1967; Blackith and Reyment 1971 ; Wiley 1981 ; Lebart er a/. 1984). In brief, and taking the present study as an example, if a sample of 200 specimens, belonging to two discrete groups, is measured for 26 morphometric traits, a graph whose coordinates are the values for any two discriminant characters will display 200 points in two clusters (for a specific example of this methodology see Rajasekaran etal. 1986). Statistically, it is possible to extend the representation of the 200 points from the two dimensional space of this graph into a 26-dimensional character space that uses all the measurements in the 200 x 26 matrix. When, as in the case of the present study, the heterogeneity of the variables in each data set is high, the analyses are performed on the character correlation matrices instead (Gnanadesikan 1977). The longest axis through the cloud of points in this hyperspace, which is positioned to minimize the squared distances of all 200 points from it, is the first principal component. The second principal component is the axis, orthogonal to the first, that traverses the second longest dimension in the cloud of points. If these first two principal components are taken as the coordinates of a new graph, and the 200 points are plotted according to their projections on these axes, the pattern will reflect the two major directions of maximum variation in the 26 characters. This graph has the advantage of representing, in two dimensions, two linear combinations of all the characters used, thus reducing the influence of subjective decisions and individual characters. Effects due to differences in character scale or variance are reduced by standardizing (automatically) the character vectors to zero

PAGE 245

238 means and unit standard deviations. Once the scores of each individual on the first two principal components are plotted in the bivariate scattergrams, visual assessment of both the position of each of the known groups and the variation within each group in the directions of those components becomes possible, and statistics can be calculated for their distributions within that subset of the principal components hyperspace. Good discrimination among groups is normally obtained this way, even in cases where none of the individual morphometric characters will discriminate alone or in pairs. It might be argued that, for the data in this study, this kind of analysis may be of questionable applicability, in particular since the assumptions of multivariate normality and homogeneity of variance were not verified. The use of principal components analysis in a purely descriptive manner, however, does not require any assumptions whatever about multivariate normality, the variance-covariance matrices, or any structure of the data (Harris 1975), which makes it suitable for use even on heterogeneous populations such as those that are the object of this study. Likewise, discriminant function analyses can also be used in a strictly descriptive way to produce the two kinds of information used in this study-the relative importance of characters as discriminators between a priori groups and the relative position of the centroids of those groups (Harris 1975). All statistical techniques used in this chapter employed methods that are commonly applied to the analysis of morphological data in biological systems, in particular in the field of numerical taxonomy, which deals with problems that are similar to those discussed here. For those that, after these notes, still might insist in criticizing the methodology, it must be stated that the intent of this study is not so much to evaluate the suitability of the methods on statistical grounds, but rather to draw attention to the outcome of their application to biological data sets that differ only in the effect of a passage through tissue culture, as the populations change over time or through a sexual cycle. In every case, all assumptions derived from the statistical

PAGE 246

239 analyses were fully confirmed by careful observation of live specimens maintained in cultivation. The Biological System The apparently higher stability of embryogenic cultures, indicated by the lower overall level of genetic variability found in both the embryogenic calli and their regenerants in several grass species (e.g., Hanna et al. 1984; Armstrong and Green 1985; Swedlund and I.K. Vasil 1985; Rajasekaran et al. 1986), suggests that the embryogenic pathway of regeneration is less likely to result in the production of somaclonal variants, and should be the method of choice whenever genetic stability and uniformity are critical to the nature of the investigation (I.K. Vasil 1983a, 1985, 1987). The validity of such broad generalization, however, has been disputed, on the basis that although karyological and phenotypic stability can often be correlated with the embryogenic pathway, varying frequencies of abnormal regenerants from embryogenic tissue cultures have also been reported in the literature (e.g., Cavallini et al. 1986, 1987; Gobel et al. 1986; Linacero and Vazquez 1986b; Orton 1987; Armstrong and Phillips 1988; Karp 1989; Bebeli er al. 1990). This has led some to conclude that any correlation between somatic embryogenesis and stability is at best not exacting, and that other factors may be somewhat overriding (e.g., Karp 1989). From this perspective, it is in a way puzzling the number of publications where variation has been described in good detail, but care was not taken to monitor strictly the regeneration pathway. Since abnormal plants may result from cultures where both the embryogenic and organogenic pathways coexist, the variation reported may well result from plants being regenerated from the mixed callus by organogenesis, a regeneration pathway somewhat more tolerant of the expression of variation. An example of such a system was described in wheat (Karp and Maddock 1 984; Maddock

PAGE 247

240 1985) . In maize, where both regeneration pathways have also been described (Green and Phillips 1975; Green 1982; Armstrong and Phillips 1988), Earle and Gracen (1985) reported extensive morphological variation in the progeny of plants regenerated from callus cultures, but also recognized that they were ". . .probably dealing with "mixed" cultures in which both organogenesis and somatic embryogenesis [did] occur. . . ." (p. 141 in Earle and Gracen 1985). Likewise, variation occurring at both the morphological and molecular levels in regenerants from maize immature embryoderived embryogenic callus cultures and their sexual progeny (Gobel et al. 1 986; Brown ef al. 1 991 ) may have resulted from the fact that ". . . although the intention was to use somatic, embryo-derived structures for plant regeneration, it cannot be ascertained that all of the regenerants [were] somatic and embryo derived " (p. 23 in G6bel et al. 1986) . Regardless of the intrinsic value of these publications in reporting new expressions of tissue-culture-induced variation, they are nonetheless inconclusive in what concerns any possible correlation between such variation and the regeneration pathway. It is somewhat revealing that, when identical studies were performed with plants regenerated from rigidly controlled embryogenic callus cultures of Napiergrass (Pennisetum purpureum K. Schum.), a complete absence of variation was shown at the morphological, biochemical (14 different isozyme systems) and molecular (RFLP analysis of the mitochondrial, plastid and nuclear genomes) levels (Shenoy and I.K. Vasil 1992). Comparable results, observed for the mitochondrial, plastid and nuclear genomes in sugarcane (Chowdhury and I.K. Vasil 1993), wheat (Chowdhury etal. 1994) and meadow fescue (Valles era/. 1993), lend additional support to the proposition that regeneration via somatic embryogenesis largely avoids phenotypic and genetic instability in the regenerants. In the present study, the last of a sequence of experiments dedicated to clarifying the role of somatic embryogenesis in controlling culture-induced variation

PAGE 248

241 (Shenoy and I.K. Vasil 1992; Chowdhury and I.K. Vasil 1993; Chowdhury er al. 1994), individuals of diploperennial teosinte, produced from long-term regenerable callus cultures by strictly monitored somatic embryogenesis, preserved their phenotypic fidelity for a period of more than two years. Plants regenerated after two and a half years, however, became increasingly abnormal, to the point that after three years only morphologically aberrant plants could be produced at all. This is in good agreement with the well known effect of ageing on instability in callus cultures (Torrey 1 967; Fukui 1 983; Lorz and Scowcroft 1 983; Benzion 1984; Chandler and I.K. Vasil 1984; Lee and Phillips 1984, 1987, 1988; Armstrong 1986; Cassells and Morrish 1 987; Zehr et a/. 1 987; Armstrong and Phillips 1 988; Benzion and Phillips 1988; Lee et al. 1988; Karp 1989; Morrish ef a/. 1990), but in sharp contrast to the postulated stability associated with the embryogenic pathway. The detailed statistical analyses of 26 morphometric traits in diploperennial teosinte reveal that the aberrant features in the regenerants from three year old embryogenic callus cultures can be broadly classified into three major groups: (1) reduction in size of vegetative and floral parts, (2) altered morphology of the vegetative body (culms, leaves) and the male inflorescence (developmental arrest, feminization), and (3) reduced male fertility. These closely conform to comparable patterns formerly described as somaclonal variation in other grass species (Table 2.1). Perception of the nature and origin of variation has been very much a function of personal interpretation and, as such, has changed considerably over the years. This is particularly true for variant plants showing anomalies in size, morphology, sex expresssion and fertility, that have been described in different species and at different times as heritable genetic mutations (e.g., Cummings etal. 1976; Fukui 1983; Sun era/. 1983; Zehr er al. 1987) or as mere epigenetic effects, transient in nature and not transmissible to the progeny (e.g., Brettell and Thomas 1980; Larkin and Scowcroft

PAGE 249

242 1983b; Irvine 1984; Earle and Gracen 1985; Rajasekaran et a/. 1986; Morrish et al. 1990). The problem of correctly identifying the exact nature (genetic vs. epigenetic) of tissue-culture-induced variation is not simple. Epigenetic changes are the result of directed physiological alterations on gene expression. As such, they are potentially reversible and, by definition, nonheritable (Meins 1983). Long-term epigenetic changes, however, are not unusual in tissue culture regenerants, and they can and have been often mistaken for true genetic variation (Karp and Bright 1985; I.K. Vasil and V. Vasil 1986; Morrish et al. 1987; I.K. Vasil 1987, 1988; Karp 1989). The distinction between genetic and epigenetic variation is further complicated by the fact that certain changes in the genetic material, in particular DNA methylation (Lorz and Brown 1986) and amplification (Cullis 1986), may be heritable under certain conditions, but may also become reversible (Phillips et al. 1990). Reversible methylation, for example, may be at the origin of the dwarf rice mutant isolated by Oono (1 985) from rice tissue cultures, which was stable for eight generations of selfing but could no longer be recovered, after crossing with control plants, among the segregating progenies. Unlike many previously described cases, where the variation observed among regenerated plants often correlates well with karyological variation in the callus cultures (e.g., Edallo er al. 1981; Sears et al. 1982; Jordan and Larter 1985; Taliaferro et al. 1989), the phenotypic anomalies in the regenerants of diploperennial teosinte could not be traced to an underlying cytological variation, as none was observed in any of the variant plants or their progenies. This is by no means unique, as numerous other cases have also been reported where novel variants were found in callus-derived plants that had an otherwise normal chromosome complement (e.g., Liu and Chen 1976; Gamborg etal. 1977; Jordan and Larter 1985; Dahleen 1989; Taliaferro et al. 1989; Dahleen and Eizenga 1990).

PAGE 250

243 The fine descriptive detail provided in the present study by the concurrent analysis of 26 morphometric traits, over time and across generations, sheds some light on the nature of the extensive variation found in the tissue-culture-derived plants. All aberrant characters observed in the R 0 plants during the first season after regeneration change sequentially in time, as morphological and functional recovery takes place and the plants gradually revert to complete normality, over a period of almost two years. This observation, and the fact that completely normal plants were also obtained after a passage through one single cycle of sexual reproduction or as a result of a hormonal treatment, clearly identifies the variation in tissue culture regenerants of diploperennial teosinte as being epigenetic, transient and non-heritable. The tendency for multiple tillering in the R 0 plants, also reported in other species of Gramineae (e.g., Brettell and Thomas 1980; Larkin et al. 1984; Jordan and Larter 1985; Earle and Gracen 1985; Rajasekaran er al. 1986; Morrish et al. 1990) and sometimes sought as a useful trait, is probably no more than a carryover effect from tissue culture, a variation of the typical micro-tillering associated with the use of cytokinins in the regeneration system (Chu and Kurtz 1990). Although in the present study this condition was among the first to become completely normal (Figs. 4.33 and 4.34), there are cases where it can be rather persistent in time, and even used as a marketable trait (Chu and Kurtz 1990). Dwarf or stunted vegetative growth in plants regenerated from tissue culture has been reported multiple times in the literature (e.g., Brettell and Thomas 1980; Beckert et al. 1983; Fukui 1983; Sun etal. 1983; Larkin etal. 1984; Earle and Gracen 1985; Jordan and Larter 1985; Zehr etal. 1987; Taliaferro etal. 1989; Hashim etal. 1990; Morrish ef al. 1990). It frequently occurs associated with reduction or suppression of the inflorescences, partial or complete male sterility and a tendency for feminization of the male flowers (Table 2.1). A strong correlation was equally found, in the present study, between these groups of characters. Plants showing the most extreme reduction in

PAGE 251

244 height (Type B regenerants) also had extremely reduced (and sterile) or completely suppressed male inflorescences (in which case they became morphologically and functionally female at the apex) during the first season after regeneration from tissue culture. Plants with less extreme height reduction (Type A regenerants) still had very short, unbranched, non-functional male inflorescences, during the same season. As internode length (and the related internode sheath ratio and plant height) increased (Figs. 4.35 through 4.37), and leaf dimensions and proportions became more normal (Fig. 4.38), the same happened with the aberrant features of the tassel. These followed a well defined pattern of sequential differentiation, starting with the elongation of the tassel central spike to a certain critical length, at which time partial male fertility was restored (Figs. 4.39 and 4.40). Both length and fertility continued to increase with time, as did branch differentiation and all conditions associated with it: number of branches, length of longest primary tassel branch and tassel ratio (Figs 4.41 through 4.43). In addition, although branching was not critical to the restoration of fertility, maximal values for fertility were attained only when branching was fully restored. A special word is due regarding the evolution over time of plant biomass. Being a combination of the number of tillers and their size, biomass decreased significantly during the first year in cultivation (a result of the sharp decrease in the number of tillers during that time), but later increased (a result of the increase in tiller size) to values comparable to the control (Fig. 4.44). Complete recovery of a normal phenotype was fully achieved for all 26 morphometric traits, after nearly two years in cultivation. With time (or after a sexual cycle), the once abnormal plants from tissue culture became virtually undistinguishable from any of the seed-derived controls.

PAGE 252

Fig. 4.33 Effect of time (or a sexual cycle) on the number of tillers in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Fig. 4.34 Effect of time (or a sexual cycle) on the number of flowered tillers in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (R 0 ). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Selfed regenerants (Ri)

PAGE 253

246 Fig. 4.33 Fig. 4.34

PAGE 254

Fig. 4.35 Effect of time (or a sexual cycle) on internode length in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Fig. 4.36 Effect of time (or a sexual cycle) on internode/sheath ratio in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Selfed regenerants (Ri)

PAGE 255

248 Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.35 Fig. 4.36

PAGE 256

Fig. 4.37 Effect of time (or a sexual cycle) on plant height in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Fig. 4.38 Effect of time (or a sexual cycle) on leaf blade length in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture Light Gray: Tissue culture regenerants (Ro) Dark Gray: Selfed regenerants (Ri)

PAGE 257

250 Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.37 Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.38

PAGE 258

Fig. 4.39 Effect of time (or a sexual cycle) on the length of the central tassel spike in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Fig. 4.40 Effect of time (or a sexual cycle) on male fertility in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture Light Gray: Tissue culture regenerants (Ro) Dark Gray: Selfed regenerants (ft 7 )

PAGE 259

252 Fig. 4.39 Fig. 4.40

PAGE 260

Fig. 4.41 Effect of time (or a sexual cycle) on the number of primary tassel branches in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Fig. 4.42 Effect of time (or a sexual cycle) on the length of the primary tassel branch in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture {Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Selfed regenerants (Ri)

PAGE 261

254 180 150' 120 90' 60' 30' Number of Primary Tassel Branches Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.41 120 100 Length of Longest Primary Tassel Branch Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.42

PAGE 262

Fig. 4.43 Effect of time (or a sexual cycle) on the tassel ratio in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Fig. 4.44 Effect of time (or a sexual cycle) on plant biomass in plants of Zea diploperennis (diploperennial teosinte) regenerated from tissue culture (Ro). Light Gray: Tissue culture regenerants (Ro) Dark Gray: Selfed regenerants (Ri)

PAGE 263

256 120 Tassel Ratio 100 80 60 < 40 20 Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.43 180 Plant Biomass 150 120 90 60 30 Spring'89 Spring'90 Fall'90 Selfed Regenerants Population / Season Fig. 4.44

PAGE 264

CHAPTER 5 ELECTROPHORETIC ANALYSIS OF TISSUE-CULTURE-DERIVED PLANTS OF DIPLOPERENNIAL TEOSINTE Introduction Any morphological variation found in tissue-culture-derived plants must be, of necessity, preceded or accompanied by changes at the physiological level. These changes might involve, among other factors, differential gene activation or deactivation, which is potentially detectable on isozyme banding profiles. If that is possible, the establishment of temporal and functional correlations between changes at the biochemical level and those at the morphological level would be extremely useful in elucidating the underlying causes of the induced epigenetic variation. The problem is complicated, however, by the fact that genetic variation, previously documented and discussed at the morphological level (Chapter 2 and Chapter 4), has also been reported at the biochemical level (e.g., Larkin et al. 1984; Maddock et al. 1985; Cooper et al. 1986; Davies etal. 1986; Allicchio era/. 1987; Breiman etal. 1987a; Eizenga 1987, 1989; Karp et al. 1987; Ryan and Scowcroft 1987; Dahleen 1989; Taliaferro et al. 1989; Dahleen and Eizenga 1990). When this happens, it may be difficult to determine which variation is due to true mutational events and which results from differential gene expression, particularly if it is kept in mind, as discussed earlier, that certain changes in the patterns of gene expression may be long lived or even heritable (see Chapter 2). This chapter reports the results of a biochemical study of variation existing at the isozyme level between diploperennial teosinte abnormal regenerants from tissue culture and control plants obtained by division of the original mother plant. 257

PAGE 265

258 Materials and Methods Plant Material To biochemically characterize the tissue-culture-derived plants, all specimens from the Ro population (as defined in Chapter 4) were analyzed electrophoretically, to assess their patterns of enzyme polymorphism and contrast them to those of the original donor plant. This plant population is the same that was used for the morphometric analysis (Chapter 4) and does not require any further characterization. Sample tissues for analysis were collected during the first season after regeneration from tissue culture (Spring 1989). Preparation of Starch Gels Thirteen percent (w/v) starch gels, prepared from Sigma Potato Starch partially hydrolyzed for electrophoresis (Sigma Cat. No. S-4501), were used throughout this study. All starch used had the same lot number. Gels were prepared by mixing the partially hydrolyzed starch with a preheated buffered solution (gel buffer) and then allowing them to cool, as described below. Gel buffers were prepared from dilutions of the corresponding electrode buffers. After preparation, one third of the buffer required for one gel was added to a side-arm flask containing the preweighed starch and sucrose, and the flask was immediately swirled vigorously by hand to assure an even suspension and to avoid the formation of lumps in the starch. The remainder of the buffer was placed into a 500 ml volumetric flask. The flask containing only the buffer was heated until boiling. When the buffer began to boil, the flask was first swirled to settle the boiling action and prevent splattering when dispensing. The flask with the starch suspension was then swirled, to ensure that the starch was fully suspended, and the boiling buffer was rapidly poured in with simultaneous swirling of both flasks. At this moment the starch mixture was a thick

PAGE 266

259 paste. Additional heating with continuous agitation on a hot plate was needed to reach the proper viscosity for degassing and pouring. The mixture was heated until large bubbles appeared on the surface, at which stage it turned translucent as it reached maximum viscosity. The mixture was degassed immediately by vacuum before it started to cool, to avoid air bubbles forming in the gel that would later interfere with protein migration. The solution boiled vigorously under vacuum and was considered degassed and ready to pour when small bubbles could no longer be seen in the starch solution or on the side of the flask; only the large bubbles from boiling remained. Molds (inside dimensions of 184 mm x 158 mm x 6 mm), were made from 6 mm thick plexiglas strips sealed onto common glass plates with methylene chloride; each held ca. 300 ml of starch solution. Pouring was done fairly rapidly to reduce the possibility of creating air bubbles. Any small solid particles or lumps present in the suspension were removed at this time with a spatula, before the gel started to harden. The molds containing the cooling solution were then covered with another glass plate to prevent dehydration and ensure a smooth upper surface to the gel. Gels were best prepared in the evening and left overnight to cool and harden properly. The following morning they were further cooled by refrigeration (4-7°C) for at least an extra hour. For faster preparation the freshly poured gels could be cooled at room temperature for approximately four hours and then placed in the refrigerator for an additional hour prior to loading the samples. Protein Extraction and Gel Loading Terminal shoots about 1 0 cm long were cut from actively growing plants in the greenhouse, and the outer green leaves were removed to expose cylinders of the inner, furled, expanding young leaves. These were labeled and quick-frozen on dry ice in a styrofoam box. This step proved to be one of the major factors for success. In fact, any

PAGE 267

260 material not frozen directly in the greenhouse produced only smears or completely negative results, an indication of the need for this additional protection of the labile molecules. Once in the lab, the frozen material was cut into smaller 1 cm long segments, wrapped in aluminum foil, labelled and deep-frozen by immersion in liquid nitrogen. These were kept in a -80°C deep-freezer until further use. There seemed to be no limit to how long these segments could be kept with no apparent loss of enzyme activity, when preserved this way. For protein extraction, frozen pieces were weighed and extracted by grinding with an ice-cold (or colder) 200 mM Tris/HCI buffer, pH 7.8, with 60% glycerol and 0.2% j8-mercaptoethanol, added just before use. The buffer was added to the plant tissue at a ratio of 2:1 (tissue weight, mg : buffer volume, fj\). Keeping the samples very cold during the whole extraction procedure was crucial. Any shoot sections containing a piece of stem tissue were better discarded, as they consistently turned brown during or after extraction and lost enzyme activity. Paper wicks (approximately 1 0 mm x 1 .5 mm) were cut from thick gel blot paper (Schleicher & Schuell Cat. No. GB003). The wicks were imbibed to saturation in the leaf extracts, transferred to labelled multi-well boxes, and either used immediately or preserved at -80°C indefinitely. A slit was cut in the chilled gels, approximately 3 cm from the cathodal edge (8 cm for cathodal migration), to serve as the origin. Excess liquid from the saturated sample wicks was removed by touching them to a plain paper towel. The gel slit was then gently opened and the wicks inserted in the slit. The wicks adhered to the cut gel surface, which aided in their application. It was important to make sure that each wick was vertical, that it touched the glass plate, and that adjacent wicks did not touch each other. Twenty five sample wicks, spaced approximately 2 mm apart, were used per gel. In addition, control wicks (saturated with extracts from the original donor plant) were

PAGE 268

261 added at the first and last lanes, and spacers (wicks soaked in extraction buffer only) were placed outside these to help minimize distortion in the band migration near the edges of the gel. For the same reason, wicks were not placed in the side border zones (approximately 1 cm wide). Electrophoretic Buffers and Staining Systems Six buffer systems, developed or optimized for the starch gel electrophoresis of isozymes from maize (Stuber et al. 1988) were tested for resolving isozymes from diploperennial teosinte. The original terminology used for these systems will be followed here. The formulations for both the electrode buffers and the corresponding gel buffers are listed in Table 5.1 . Gel recipes are listed in Table 5.2. The formulations used for the enzyme staining solutions were adapted from Vallejos (1983) and Stuber era/. (1988). Electrophoresis Once all samples had been inserted, 300 ml of electrode buffer were added to each electrode buffer tank to cover the electrodes. The gels were placed into the electrode tanks at 4°C in a refrigerator, with the glass base of the gel molds sitting on an absorbent sponge cloth saturated with ice-cold water. The bridges between the electrode buffer and the gel consisted of pieces of absorbent sponge cloth, color coded for easy recognition of the polarity. The cloths were soaked in electrode buffer in the buffer tank and carefully applied to the end surfaces of the gel. A Biorad 500/200 DC Power Supply was used as the power source for electrophoresis. This was carried out under constant current in a refrigerator at 4°C. Power settings and running time varied depending upon the buffer and had to be determined experimentally (Table 5.3). An electric pulse was applied for 15 minutes with the wicks in place to drive the proteins into the gels. After this time the power was turned off, the wicks were removed

PAGE 269

262 Table 5.1 Formulas for electrode and gel buffers System Electrode Buffer Gel Buffer A 50 mM L-Histidine (7.75 g/l) pH 5.0 24 mM Citric Acid.H 2 0 pH adjusted with Citric Acid.H 2 0 B 65 mM L-Histidine (1 0.088 g/l) pH 5.7 20 mM Citric Acid.H 2 0 pH adjusted with Citric Acid.H 2 0 C 190 mM Boric Acid (11.875 g/l) pH 8.3 40 mM Lithium Hydroxide pH adjusted with LiOH CT 68 mM N-(3-Aminopropyi) pH 6.1 Morpholine (9.8 g/l) 40 mM Citric Acid.H 2 0 (8.41 g/l) pH adjusted with Citric Acid.H 2 0 D 65 mM L-Histidine (1 0.088 g/l) pH 6.5 7 mM Citric Acid.H 2 0 pH adjusted with Citric Acid.H 2 0 F 1 35 mM Trizma Base (1 6.35 g/l) pH 7.0 40 mM Citric Acid.H 2 0 pH adjusted with Citric Acid.H 2 0 4 mM L-Histidine 2mM Citric Acid. H^ (1:13 dilution of Electrode Buffer) (23 ml Buffer + 276 ml H^ = 299 ml) 9 mM L-Histidine 3mM Citric Acid. H^ (1 :7 dilution of Electrode Buffer) (40 ml Buffer + 240 ml Hj>0 = 280 ml) 9 parts Tris/Citric Acid Buffer [50 mM Trizma Base (6.2 g/l) + 7 mM Citric Acid.H^, pH 8.3] : 1 part Electrode Buffer C (270 ml Tris/Citrate Buffer C + 30 ml Electrode Buffer C = 300 ml) 3.4 mM N-(3-Aminopropyl) Morpholine 2 mM Citric Acid.H 2 0 (1 :20 dilution of Electrode Buffer) (15 ml Buffer + 285 ml H2O = 300 ml) 16 mM L-Histidine 2 mM Citric Acid.H 2 0 (1 :4 dilution of Electrode Buffer) (70 ml Buffer + 210 ml H^ = 280 ml) 9 mM Trizma Base 3 mM Citric Acid.F^O (1 :1 5 dilution of Electrode Buffer) (20 ml Buffer + 280 ml Hj>0 = 300 ml) Source: Modified from Stuber et al. (1 988)

PAGE 270

263 Table 5.2 Gel Recipes System Starch (13%) Gel Buffer Volume A 38.9 g Potato Starch 299 ml D B 36.4 g 8.4 g Potato Starch Sucrose 280 ml <-> 39.0 g 13.0g Potato starch Sucrose 300 ml CT 39.0 g 13.0g Potato starch Sucrose 300 ml D 39.0 g 9.0 g Potato starch Sucrose 300 ml F 39.0 g 9.0 g Potato starch Sucrose 300 ml Source: Modified from Stuber etal. (1988) Table 5.3 Constant Current, Initial Voltage and Running Time for Electrophoresis. System Constant Current Initial Voltage Running Time (mA) (V) (hours) A 25 310 4 B 35 300 4 C 35 200 4 CT 35 330 4 D 25 360 4 F 35 210 4

PAGE 271

264 and discarded and the slit closed. It was important not to trap air between the two sides of the gel at the origin. Solid glass rods were placed at the top and bottom of the gel to ensure that the slit was held tightly closed during the run. The gels were covered with polyethylene film to prevent desiccation, and glass plates were placed on top of the slabs to ensure good contact between the gel and the electrode bridges. Frozen ice packs were placed on the glass plates to minimize enzyme degradation. Careful maintenance of very low temperatures throughout the whole process was probably the most important factor for success. Enzyme Activity Staining Following completion of electrophoresis, the glass plate and the polyethylene wrap were removed and a diagonal slash was cut in the anodal left corner of the gel to identify the position of the first lane. Since no enzymes migrate the last 3 cm, the terminal 3 cm section of the gel was also cut off the entire length of the anodal edge to facilitate slicing and transfer to the staining boxes. Gels were sliced using a home-made gel slicer made of a piece of very thin steel wire maintained under strong tension by the bow of a hacksaw. Two plastic slicing strips of the desired thickness were placed on the bottom glass plate along both long sides of the gel. The gel was then sliced horizontally in two 3-mm thick slices by pulling the wire through, using the slicing strips to adjust the thickness of the slice through the slab. Single slices were separated with a spatula and transferred to staining boxes, where zones of enzyme activity were revealed by immersion into stain assay mixes. The top slice always had a rubbery upper surface due to its exposure to the air, and produced poorer resolution than the lower slice. This was counteracted by staining it inverted to expose the freshly cut lower surface for staining. Each slice was individually stained for a different enzyme with 50 ml of a freshly prepared staining mix. A gentle

PAGE 272

265 swirling motion was used to eliminate air bubbles trapped under the slice, as they affect the even distribution of the solution. Analysis of Banding Patterns Gels were kept in the dark and checked for staining every 15 to 30 minutes. When optimum resolution was obtained they were rinsed with tap water, visually scored, labelled, and photographed with Kodak™ Technical Pan™ 35 mm black and white film. This was then developed with full strength Kodak™ Developer D-1 9™ for four minutes to improve contrast. Zymograms were visually compared for differences in migration patterns. No attempt was made to identify single bands by using Rf (migration of the band relative to the front) or R b (migration of the band relative to a unique control band) values and no genetic analysis was performed to identify the monomeric or polymeric nature of the enzymes. Results Selection of Electrophoretic Buffers and Enzyme Systems for Studying Isozymes in Diploperennial Teosinte Seventy four enzyme staining systems were tested, in combination with each of the six buffer systems described above. The results of this survey are presented in Table 5.4. The colloquial names and enzyme commission (E.C.) numbers listed below follow the guidelines and recommendations of the nomenclature committee of the International Union of Biochemists (I.U.B. 1984). Well resolved bands were obtained with 15 enzyme activity stains-aconitase (E.C. 4.2.1.3), alcohol dehydrogenase (E.C. 1.1.1.2), arginine aminopeptidase (E.C. 3.4.11.6), aryl esterase (tested with orand /3-naphthyl substrates) (E.C. 3.1.1.2), diaphorase (E.C. 1.6.99.1), endopeptidase (E.C. 3.4.-.-), L-glutamic dehydrogenase

PAGE 273

266 Table 5.4 Enzyme activity stains and buffer systems tested for isozyme detection in Zea diploperennis. Buffer System* Enzyme Activity Stain A B C D F CT o © © © o o o © © © o o © o © © o o o © o o © o o © © 1 . Acid Phosphatase (a-naphthyl substrate) 2. Acid Phosphatase (p-naphthyl substrate) 3. Aconitase 4. Adenosine Deaminase 5. Adenylate Kinase 6. Alanine Aminotransferase 7. Alcohol Dehydrogenase 8. Aldehyde Oxidase 9. Aldolase 10. Alkaline Phosphatase (a-naphthyl substrate) 1 1 . Alkaline Phosphatase (p-naphthyl substrate) 12. Amylase (a-) 13. Arginine Aminopeptidase 14. Aryl Esterase (a-naphthyl substrate) © 15. Aryl Esterase (p-naphthyl substrate) O 16. Aryl Sulfatase .... 1 7. Ascorbate Oxidase .... 18. Carbonic Anhydrase .... 19. Catalase .... 20. Chalcone Isomerase .... 21 . Creatine Kinase .... 22. Diaphorase © © 23. Endopeptidase © 24. Formate Dehydrogenase .... 25. Fumarase .... 26. Galactose Dehydrogenase .... 27. Galactosidase (a-D-) .... 28. Galactosidase (p-D-) .... 29. Glucose Oxidase .... 30. Glucose-6-Phosphate Dehydrogenase O O O 31 . Glucosidase (p-D-) O 32. Glucuronidase (p-) .... 33. Glutamic Dehydrogenase (L-) ... © 34. Glutamate-Oxaloacetate Transaminase O © © O 35. Glutamine Synthetase .... 36. Glutathione Reductase .... 37. Glyceraldehyde-3-Phosphate Dehydrogenase .... 38. Glycerol-3-Phosphate Dehydrogenase .... 39. Guanine Deaminase .... 40. Hexokinase O O © (continued)

PAGE 274

267 Table 5.4 (Continued) Buffer System* Enzyme Activity Stain A B C D F CT 41. Hydroxyacylglutathione Hydrolase 42. Iditol Dehydrogenase (L-) 43. Isocitric Dehydrogenase © 44. Laccase 45. Lactate Dehydrogenase 46. Leucine Aminopeptidase 47. Lipoxigenase 48. Malic Dehydrogenase © o © 49. Malic Enzyme O © © 50. Mannose-6-Phosphate Isomerase 51. Nitrate Reductase 52. Nitrite Reductase 53. Nucleoside Triphosphate Pyrophosphatase 54. PEP Carboxylase 55. Peptidase 56. Peroxidase _ 57. Phosphoglucomutase e © 58. Phosphogluconate Dehydrogenase (6-) © o © 59. Phosphohexose Isomerase © o © 60. Phosphorylase — 61. Purine-Nucleoside Phosphorylase Pyrophosphatase 63. Pyruvate Kinase 64. Ribonuclease 65. Shikimic Acid Dehydrogenase © © © e o o 66. Succinic Dehydrogenase o 0 67. Superoxide Dismutase 68. Thioglucosidase 69. Triose Phosphate Isomerase o 70. Tyrosine Aminotransferase 71. Urease 72. Uroporphyrinogen I Synthase 73. Xanthine Dehydrogenase 74. Xylosidase "= no activity © = activity detected, poor migration or resolution © = fair migration or resolution © = well resolved bands

PAGE 275

268 (E.C. 1.4.1.2), hexokinase (E.C. 2.7.1.1), isocitric dehydrogenase (E.C. 1.1.1.42), malic dehydrogenase (E.C. 1.1.1.37), malic enzyme (E.C. 1.1.1.40), phosphoglucomutase (E.C. 5.4.2.2), 6-phosphogluconate dehydrogenase (E.C. 1.1.1.44) and phosphohexose isomerase (E.C. 5.3.1.9). Eight additional enzymes-acid phosphatase (E.C. 3.1.3.2) (tested with aand /3-naphthyl substrates), adenylate kinase (E.C. 2.7.4.3), aldolase (E.C. 4.1.2.13), alkaline phosphatase (E.C. 3.1.3.1) (tested with aand /3-naphthyl substrates), glutamate-oxaloacetate transaminase (E.C. 2.6.1.1) and shikimate dehydrogenase (E.C. 1.1.1.25)--showed fair migration or resolution. A number of changes in all important electrophoretic parameters were made in an attempt to improve resolution or stainability, but all other enzyme systems that were investigated showed either limited migration or resolution or had no staining activity at all. This completely prevented their further use in any subsequent work. To assess for cultureinduced variation, all 100 plants of the R 0 population were individually tested for each of the 23 enzyme activity stains that yielded well resolved or fairly well resolved banding patterns. Isozyme Analysis of Tissue Culture Reqenerants Zymograms covering 50 R 0 plants, developed for each of 12 enzyme activity stains, are shown in Figs. 5.1 through 5.12. The most obvious feature in all (including those not illustrated) was the absolute uniformity of the banding patterns from all tissueculture-derived plants. No differences could be found between Type A and Type B regenerants, neither were there any bands in the R 0 plants that could not be referred to corresponding bands in the control donor plant. Discussion and Conclusions The occurrence of phenotypic variation in plants regenerated from tissue cultures, frequently associated with regeneration systems involving an unorganized

PAGE 276

Figure 5.1 Zymogram (anodal migration) showing isozyme phenotypes for the acid phosphatase alleles (specific for cr-naphthyls), in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 277

270

PAGE 278

Figure 5.2 Zymogram (anodal migration) showing isozyme phenotypes for the acid phosphatase alleles (specific for /3-naphthyls), in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 280

Figure 5.3 Zymogram (anodal migration) showing isozyme phenotypes for the aconitase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 281

274

PAGE 282

Figure 5.4 Zymogram (anodal migration) showing isozyme phenotypes for the aryl esterase alleles (specific for a-naphthyls), in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 283

276

PAGE 284

Figure 5.5 Zymogram (anodal migration) showing isozyme phenotypes for the aryl esterase alleles (specific for /3-naphthyls), in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 285

278

PAGE 286

Figure 5.6 Zymogram (anodal migration) showing isozyme phenotypes for the diaphorase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 287

280 ttHtttt tttt««ttt> iifffttffftttii

PAGE 288

Figure 5.7 Zymogram (anodal migration) showing isozyme phenotypes for the endopeptidase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 289

282

PAGE 290

Figure 5.8 Zymogram (anodal migration) showing isozyme phenotypes for the L-glutamic dehydrogenase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 291

284 itfMtvvtt'ttttttttttiitin

PAGE 292

Figure 5.9 Zymogram (anodal migration) showing isozyme phenotypes for the malic dehydrogenase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 294

Figure 5.10 Zymogram (anodal migration) showing isozyme phenotypes for the malic enzyme alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 295

288 lillllMMlllllllllllllll

PAGE 296

Figure 5.11 Zymogram (anodal migration) showing isozyme phenotypes for the phosphoglucomutase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 297

290 MXmiMiimiiinnXMi ^ittitTtilHiiiii^iii i

PAGE 298

Figure 5.12 Zymogram (anodal migration) showing isozyme phenotypes for the 6-phosphogluconate dehydrogenase alleles, in diploperennial teosinte. Top: Regenerant lines 1 through 25 Bottom: Regenerant lines 26 through 50 Sample origin is at the top. The two outermost lanes are controls (from the original donor plant).

PAGE 299

292 iiiiliiiiilUtiliilli iUS} ftiiiiiHift Hilil}||i|||>

PAGE 300

293 callus phase, has been found in many crop species. Deviations from the expected phenotypic and genetic uniformity have been reported at the morphological, karyological, physiological, biochemical and molecular levels (Meins 1983; Orton 1984; D'Amato 1985; Scowcroft 1985; Ahloowalia 1986; Gould 1986; Semal 1986; L6rz ef a/. 1988). Variation in isozyme or seed protein banding patterns, as a result of a passage through tissue culture, has been reported repeatedly in the literature (e.g., Larkin ef a/. 1 984; Maddock er a/. 1 985; Cooper ef a/. 1 986; Davies ef a/. 1 986; Breiman ef a/. 1 987a; Eizenga 1 987, 1 989; Karp ef al. 1 987; Dahleen 1 989; Taliaferro ef al. 1 989; Dahleen and Eizenga 1990). A general interpretation of these results, however, has not been easy. In many cases the presence of biochemical variants seems to correlate well with the occurrence of karyological aberrations. This implies that these could be at the origin of the newly induced variation. Such was the case with alcohol dehydrogenase variants identified in wheat (Triticum aestivum L.) regenerants from tissue culture that were found to be associated with aneuploidy, translocations and other chromosomal aberrations (Davies ef al. 1986). In barley (Hordeum vulgare L), where regenerants derived from immature embryos were found to be relatively stable and free from variation for the banding profiles of esterase and glutamate-oxaloacetate transaminase, a single hordein variant was found, associated with meiotic aberrations (Karp ef al. 1987). And in tall fescue (Festuca arundinacea Schreb.), where isozyme changes were found in regenerants from mature embryo cultures (Eizenga 1987), the loss of isozyme bands also correlated well with the presence of karyological aberrations, of which chromosome loss was the most apparent change (Eizenga 1 989; Dahleen and Eizenga 1990). The accumulating evidence thus would seem to indicate that, as previously observed at the morphological level (see Chapter 2), variation at the biochemical level can also be associated with numerical and gross structural alterations of the

PAGE 301

294 chromosomes. In a study of wild barley (H. spontaneum Koch) regenerants, however, a variant hordein pattern was found in a plant that had an otherwise normal karyotype (Breiman etal. 1987a). Likewise, Dahleen (1989) also found extensive biochemical (and morphological) variation in oat (AVena sativa L) lines derived from selected tissue culture regenerants with no obvious karyotypic aberrations. This clearly indicates that although the role of karyological alterations cannot be ignored in the generation of such variation, there are cases where other factors might be involved as well. Clarification of the matter is further obscured by the fact that some reports may have a dual interpretation of the results. Several studies, for example, have reported variation in the gliadin banding patters of wheat (Larkin ef al. 1 984; Maddock ef al. 1 985; Cooper ef al. 1986). It is not clear, however, whether these are indeed culture-induced variants or the normal products of segregation and outcrossing (Metakovsky et al. 1987). Another possible case is the variation in esterase and peroxidase isozyme patterns reported in Bothriochloa (Taliaferro er al. 1 989) which may be epigenetic, rather than genetic, in origin. Both esterase and peroxidase are key enzymes in native auxin metabolism (Tang and Bonner 1 947; Wagenkencht and Burris 1 950; Andreae and Good 1955; Good era/. 1956; Ray 1958; Hinman and Lang 1965; Sequira and Mineo 1966; Endo 1968; Haard 1977; Huang and Haard 1977; Cohen and Bandurski 1978; Thomas and Jen 1980; Thomas ef al. 1980; Hangarter and Good 1981), and are therefore among the most likely to be influenced by the hormonal imbalance in vitro. Recent observations showed that 2,4-D may also be directly involved in the inhibition of peroxidase activity (Lee ef al. 1982; Grambow and Langenbeck-Schwich 1983). In Bothriochloa all selfed regenerants (/? 7 ) had peroxidase and esterase bands identical to their respective progenitors (R 0 ) (Taliaferro ef al. 1989). This could be the result of certain long lived physiological changes that can be transferred through one or more sexual generations (e.g., Morrish etal. 1990).

PAGE 302

295 It is noteworthy that in sharp contrast with the forementioned variation found in the regenerants from mature embryo cultures of tall fescue (Eizenga 1987, 1989; Dahleen and Eizenga 1990), plants derived from cultured immature inflorescences of the same species were fertile, had a normal chromosome complement and completely lacked any isozyme variation (Eizenga and Dahleen 1990). This striking lack of variation (as compared to the previous reports) was attributed both to a shorter period in culture and to regeneration by somatic embryogenesis, which resulted ". . . in fewer abnormalities than organogenesis, which occurred in the mature embryo culture. . . ." (Eizenga and Dahleen 1990). This would also probably explain the biochemical (and morphological) variation found in the oat lines studied by Dahleen (1989), as the original R 0 plants were selected from a karyologically heterogeneous population produced by shoot morphogenesis (McCoy era/. 1982). Additional support of the idea that somatic embryogenesis may contribute to the apparent stability and uniformity at the biochemical level of tissue culture regenerants is provided by Shenoy and I.K. Vasil (1992), who found no detectable qualitative variation in a population of Napier grass regenerated by strictly monitored somatic embryogenesis and analyzed for the activity of 14 enzyme systems. The results reported in this chapter for diploperennial teosinte also support this hypothesis. The uniformity of protein banding patterns per se is by no means definitive evidence of absence of any aberrations at the loci tested (Brown and Weir 1983). Only about one quarter of base substitutions in the DNA result in amino acid replacements detectable by routine electrophoresis, and substitutions in sequences which are not translated, including the intervening sequences or introns, are beyond the reach of electrophoretic surveys. Sufficient evidence is already available, however, to indicate that the results of these biochemical analyses are closely paralleled by results of molecular analyses at the nuclear and cytoplasmic levels (Shenoy and I.K. Vasil 1992; Chowdhury and I.K. Vasil 1993; Valles era/. 1993; Chowdhury era/. 1994).

PAGE 303

296 The apparent regulatory effect that can be achieved by (1 ) the careful choice of competent explants and by (2) strict selection of the embryogenic pathway of regeneration has been well documented at the karyological and morphological levels (Scheunert ef al. 1978; I.K. Vasil 1983a, 1985, 1987; Hanna era/. 1984; Armstrong and Green 1985; Swedlund and I.K. Vasil 1985; Hahne and Hoffmann 1986; Rajasekaran ef al. 1986; Singh 1986; Sengupta ef al. 1988; Taniguchi and Tanaka 1989). This can be extended to the biochemical and molecular levels and supports the suggestion that the combined use of these two key factors may largely avoid the undesirable phenotypic and genetic instability so often associated with systems that involve a passage through an undifferentiated callus phase in tissue culture (I.K. Vasil 1983a, 1985, 1987).

PAGE 304

CHAPTER 6 CLOSING REMARKS As stated earlier, it was the objective of the research reported in this dissertation (1 ) to develop a long-term efficient regeneration system from cultures of diploperennial teosinte; (2) to morphologically characterize the pathway of regeneration; (3) to morphometrically and cytologically characterize the regenerants, their progeny and control plants to evaluate the genetic fidelity of the tissue-culture-derived plants; (4) to electrophoretically characterize the regenerants to further evaluate at the biochemical level the genetic fidelity of the tissue-culture-derived plants; and (5) to interpret the nature of variation, should any be detected in the regenerants and/or their progeny. In the previous chapters it was shown that once the appropriate requirements are met for the choice of explants and their cultural conditions, it is possible to initiate regenerable embryogenic calli of diploperennial teosinte that can be maintained for extended periods of time (over 4 years) without a serious loss in their capacity to regenerate plants. It was also shown that although the regeneration capacity of such cultures is not greatly impaired by time (within the stated limits), the same is not true for the fidelity of their genetic expression. Extensive variation was generated in the tissue culture process, to the point that after three years every single regenerant was grossly abnormal, and in more than one way. Previous publications reported an increase in phenotypic aberrations associated with long-term cultures (e.g., Torrey 1967; Fukui 1983; Lorz and Scowcroft 1983; Benzion 1984; Chandler and I.K. Vasil 1984; Lee and Phillips 1984, 1987, 1988; Armstrong 1986; Cassells and Morrish 1987; Zehr era/. 1987; Armstrong and Phillips 1988; Benzion and Phillips 1988; Lee et a/. 1988; Karp 1989; Morrish et al. 1990). The present work also showed that a large number of qualitative 297

PAGE 305

298 and quantitative morphological characters, as well as some physiological traits, were affected by the long-term tissue culture process. The most noticeable were the overall size and proportions of the plants and their sex expression and fertility at maturity. The regenerants were distinctly dwarf multitillering male sterile plants, with an incomplete development and/or feminization of the tassel. No cytological anomalies were found in conjunction with the morphological aberrations. All plants had the normal diploid chromosome complement for the species. No variation was detected at the biochemical level (for the 23 isozyme systems tested) among the regenerants or between these and the original donor plant. The most relevant findings of this study, however, permitted by the perennial nature of this fully self-fertile species, were the clear demonstration that (1 ) the simple passage of time (a factor not available when working with annual plants), (2) a sexual cycle (not possible when sterility is involved) or (3) a hormonal treatment reverts most, if not all, of the induced aberrations back to a completely normal phenotype. All variation found in the tissue culture regenerants was physiological and non-heritable, and merely a consequence of the environmental stress represented by the tissue culture per se. The total absence of stable, transmissible variation offers full support to the proposition that the strict impositions of embryogenesis virtually eliminate any genetic variation that might arise during tissue culture (I.K. Vasil 1983a, 1985, 1987). The underlying reasons for the unnatural behavior of plants derived from in vitro cultures are not currently understood (Meins 1983; Gould 1986; Evans and Sharp 1988). Some anomalies, such as the multitillering habit found in the R 0 during the first growing season after regeneration, may be no more than a reflection of a "carry-over" effect commonly associated with the use of cytokinins in the regeneration protocol. Some long-term effects that could be easily confused with stable genetic variation were not found in the following generation after selfing. This kind of variation, mostly expressed at the morphological level by alterations in size, growth habit and sex

PAGE 306

299 expression, is termed epigenetic variation (Meins 1983). It can be duplicated by growing plants from seed under conditions that mimic the tissue culture environment (e.g., Lorz er a/. 1988), and there is a suggestion that it might be correlated with the stress of growth under in vitro conditions (e.g., Lorz ef a/. 1988; Karp 1989). The exact nature of this "stress in vitro" remains largely unknown. Developmental aberrations comparable to those observed in the course of this study, however, can also be generated in systems that do not involve in vitro conditions. In maize, for example, exposure to short days and cool nights, or auxin applications, bring about in normal plants exactly the same type of variation in fertility and development of the tassel found in long-term regenerants of diploperennial teosinte (Figs. 6.1 through 6.4) (Schaffner 1927; Richey and Sprague 1932; Choudri and Krishan 1946; Moss and Heslop-Harrison 1968). Because the environmental effects can be completely substituted by auxin treatment of the plants, it is natural to assume that they are effective by altering the native auxin metabolism (Heslop-Harrison 1957a, 1957b, 1959, 1961). The only common denominator between the situation described above and the anomalies induced by tissue culture is the presence of an auxin (required to elicit embryogenic response in tissue cultures of Gramineae), which leads to the inference that the presence of the auxin in the medium may well also be at the origin, directly or indirectly (e.g., by influencing native auxin metabolism), of the anomalies found in the tissue-culture-derived plants (Fig. 6.5). Recent observations show that synthetic auxins may indeed affect native hormone levels. Compounds like 2,4-D (as well as a number of other disubstituted phenolics) can inhibit peroxidase (including IAA oxidase) activity, which leads to an accumulation of endogenous auxin in the tissues (Lee ef al. 1 982; Grambow and Langenbeck-Schwich 1983). From this perspective, then, the nature and intensity of the effects observed in specific organs, and at different times, in the regenerants of diploperennial teosinte, is most probably correlated with the

PAGE 307

Male fertility, tassel development and sex expression in tissue-culturederived plants of diploperennial teosinte. Spring 1989 (R 0 ): The apical inflorescence is morphologically and functionally female (type B regenerant). Male fertility, tassel development and sex expression in tissue-culturederived plants of diploperennial teosinte. Spring 1990 {R 0 ): The apical inflorescence is morphologically male, but only partially developed. Branches are absent and the anthers do not protrude, which makes most plants functionally male sterile. Feminization of the basal flowers is not uncommon. Male fertility, tassel development and sex expression in tissue-culturederived plants of diploperennial teosinte. Spring 1990 {R 0 ): In some plants, transformation of tassel structures into vegetative organs can occur. In such plants, a variable number of floral appendages become transformed into leaflets. In extreme cases, the whole apical inflorescence may turn vegetative, with small plantlets replacing the spikelets. This condition, which can be easily mistaken for a genetic mutation, is only a transient physiological disorder under photoperiodic control (Coe and Poethig 1982). Male fertility, tassel development and sex expression in tissue-culturederived plants of diploperennial teosinte. Fall 1990 (R 0 , left) and selfed regenerants (flj, right): The apical inflorescence is morphologically and functionally male. Individual spikelets are fully fertile.

PAGE 309

302 5— CD J! ^ o ® Qto (0 CL 0) ° 99> co m O -D T3 o c CO cl CL 00 co o tr a. «> o E a JS & co -o CD co a) _c u c CO 0 CD E ~ « .«s eo £ |i CO -c O Q. ft 55 "E o E $ .£2 O £ iT C CD (0 CD £ o a S co CD O CO 0 T3 CD "co £ c c ? 2 CD CD > CO -»— • o> CD > CD .C -— < C o O c CO — o a) " -C F 1 c o >. CD f= TJ C C CO E = ,. o o . 3 P *" -Q "ti >> (0 .52 O CD <0 CO ~ X *5 > c.2 i c 3 Q. CO _ 2 c •c I o CD c CO CD sis* b CO w ^ O E r CO co 2-o 2 co _c -•— < CO CL "CO c CD E CL O | CD "D CD > -«— CO E CD CO c CD E CL O £• rO C © _ o Q> c 5 *~ *: w t3 9o o co A co -9 a) o O) CO E "N »-° -o _ -Q .E to g> ^ J3> > • o ai t co 55 0) 05 r 3 c 5 c — E '5 * — PI 8 5 E-S l| g E _ o .2 X >. ^ co co N *^ CO CD o3= 9-Q-c ** o C (D £ O tTJH CO TJ 3 CO w 5t CD CO T ^ CD CO CD CD — -Q CO cb£ to E in CD en

PAGE 310

303 auxin levels present in the plants at that particular time, as different developmental steps are differentially sensitive to the level of auxin concentration. Although variation in endogenous auxin levels might be invoked to explain the anomalous floral development and altered sex expression in the tissue culture regenerants, the mediation of ethylene, however, better explains the alterations found in the vegetative morphology. Both auxins and ethylene can cause a number of similar responses at the morphological and physiological levels. Endogenous IAA is capable of controlling ethylene production in vivo (Maloney et al. 1983), which may subsequently affect other metabolic processes, including the biosynthesis of other hormones. Many responses once attributed to auxin, including the forementioned alterations in floral development and sex expression, have now been traced to ethylene produced in response to the auxin treatment instead (Abeles 1973; Yang and Pratt 1978). In addition, there is also considerable evidence that 2,4-D, and probably all other synthetic auxins used in the induction of somatic embryogenesis, act on a range of developmental processes by directly increasing the levels of ethylene production in plant tissues (e.g., Holm and Abeles 1968; Goeschel and Kays 1975; Rappaport 1980, Pinfield er al. 1984). Assessing a role for ethylene in the morphological determination of abnormal traits in plants from tissue culture is not an easy task. Little attention has been paid to the possibility that ethylene may be involved in such processes and direct experimental evidence is not available upon which to draw any valid conclusions. Indirect evidence, however, indicates that factors known to alter ethylene production or which promote changes in ethylene concentration in normal plant tissues often lead to phenotypic responses that closely mimic the morphology and behavior of variant plants regenerated from tissue cultures. At the vegetative level, a typical example is in the antilodging properties of the ethylene releasing compound, Ethrel®. Ethephon (2-chloroethylphosphonic acid), the active ingredient in Ethrel®, is a growth regulator

PAGE 311

304 that has been used to reduce plant height and lodging and thereby increase harvestable yield in a variety of cereals and small grains (Dahnous et al. 1 982; Nafziger era/. 1986; Wiersma etal. 1986; Gaska and Oplinger 1988; Norberg et al. 1988, 1989; Konsler and Grabau 1989). In these plants lodging is diminished by a reduction of plant height (reduced internode length), an increase in stem diameter and a stimulation of brace root development. The application of ethephon to grass species thus results in large changes in the stalk morphology of the treated plants, many of which similar to those shown by tissue culture variants. In both cases a marked decrease in height results, due to a reduction in internode length without a change in internode number. This is often combined with a thickening of the stem which, in the case of the tissue culture variants, is particularly noticeable in the extreme case of Type B regenerants. Unquestionably, both conditions impose dwarfism on the plants, one of the most commonly reported anomalies resulting from tissue culture. The assumption of an involvement of ethylene receives additional support from the results of the GA3 applications to tissue culture regenerants reported in this dissertation. The alleviatory effects of GA3 on the aberrant traits found in the R 0 plants suggests the possibility that the growth abnormalities result from the suppression, by the conditions in vitro, of endogenous gibberellin biosynthesis, which results in an inadequate supply of hormone during the developmental period. Pinfield et al. (1 984) showed that 2,4-D (or the ethylene it produces) inhibits endogenous gibberellin production in marrow seedlings treated with various combinations of both gibberellin and 2,4-D. The same authors and others (e.g., Vyas er al. 1975) showed that gibberellins reversed the effects of 2,4-D in vivo. Gibberellins are also well known antagonists of ethylene action (e.g., Jackson and Campbell 1979; Pinfield etal. 1984). It is thus conceivable that tissue cultures might induce dwarfism, as well as alterations in fertility and floral structure and development, through the mediation of

PAGE 312

305 ethylene. Many of the variant phenotypes commonly described in the literature, when shown to be of non-genetic origin, could then be nothing but a reflection, at the morphological level, of an altered physiology produced by higher levels of ethylene induced in vitro either by the mechanical excision of the explant, by the presence of a strong synthetic auxin in the medium, or by the overall stressing conditions of the tissue culture environment per se. Obviously, additional work needs to be done along these lines to test this hypothesis.

PAGE 313

REFERENCES Abeles F.B. 1973. Ethylene in Plant Biology. Academic Press, New York. Ahloowalia B.S. 1982. Plant regeneration from callus culture in wheat. Crop Sci. 22:405-410. Ahloowalia B.S. 1 983. Spectrum of variation in somaclones of triploid ryegrass. Crop Sci. 23:1141-1147. Ahloowalia B.S. 1986. Limitations to the use of somaclonal variation in crop improvement. In J. Semal (ed.): "Somaclonal Variations and Crop Improvement", pp. 1 4-27. Martinus Nijhoff, Dordrecht, The Netherlands. Ahloowalia B.S. and J. Sherington. 1985. Transmission of somaclonal variation in wheat. Euphytica 34:525-537. Alfinetta B., Z.A. Zamora and K.J. Scott. 1983. Callus formation and plant regeneration from wheat leaves. Plant Sci. Lett. 29:183-189. Allicchio R., C. Antonioli, L. Graziani, R. Roncarati and C. Vannini. 1987. Isozyme variation in leaf-callus regenerated plants of Solanum tuberosum. Plant Sci. 53:81-86. Ammirato P.V. 1983. Embryogenesis. In D.A. Evans, W.R. Sharp, P.V. Ammirato and Y. Yamada (eds.): "Handbook of Plant Cell Culture. Vol. 1", pp. 82-123. Macmillan, New York. Ammirato P.V. 1 989. Recent progress in somatic embryogenesis. Intern. Assoc. Plant Tiss. Cult., Newslett. 57:2-16. Andreae W.A. and N.E. Good. 1955. The formation of indoleaceylaspartic acid in pea seedlings. Plant Physiol. 30:380-382. Armstrong C.L and C.E. Green. 1982. Initiation of friable, embryogenic maize callus: the role of L-proline. Agron. Abstr. 74:89. Armstrong C.L. and C.E. Green. 1985. Establishment and maintenance of friable, embryogenic maize callus and the involvement of L-proline. Planta 164:207-214. Armstrong C.L and R.L. Phillips. 1988. Genetic and cytogenetic variation in plants regenerated from organogenic and friable, embryogenic tissue cultures of maize. Crop Sci. 28:363-369. Armstrong K.C., C. Nakamura and W.A. Keller. 1983. Karyotype instability in tissue culture regenerants of triticale (x Triticosecale Wittmack) cv. Welsh from 6-month old callus cultures. Z. Pflanzenzucht. 91 :233-245. 306

PAGE 314

307 Ashmore S.E. and A.S. Shapcott. 1989. Cytogenetic studies of Haplopappus gracilis in both callus and suspension cell cultures. Theor. Appl. Genet. 78:249-259. Bailey J.M., J. King and O.L Gamborg. 1972. The ability of amino compounds and conditioned medium to alleviate the reduced nitrogen requirement of soybean cells grown in suspension cultures. Planta 1 05:25-32. Bayliss M.W. 1980. Chromosomal variation in plant tissues in culture. Intern. Rev. Cytol., Suppl. 11A:1 13-144. Beauchesne G. 1982. Appearance of plants not true to type during in vitro plant propagation. In E.D. Earle and Y. Demarly (eds.): "Variability in Plants Regenerated from Tissue Culture", pp. 268-272. Praeger Publishers, New York. Bebeli P.J., A. Karp and P.J. Kaltsikes. 1990. Somaclonal variation from cultured embryos of sister lines of rye differing in heterochromatic content. Genome 33:177-183. Becker U. and G. Reuther. 1986. Cytogenetic studies in callus cultures of Asparagus officinalis. In W. Horn, C.J. Jensen, W. Odenbach and O. Schieder (eds.): "Genetic Manipulation in Plant Breeding", pp. 425-428. Walter de Gruyter, Berlin. Beckert M. 1982. Role du scutellum dans I'obtention de plantes neoformees in vitro chez le mai's. Agronomie 2:61 1 -61 5. Beckert M., M. Pollacsek and M. Caenen. 1983. Etude de la variability genetique obtenue chez le mai's apres callogenese et regeneration de plantes in vitro. Agronomie 3:9-17. Behrend J. and R.I. Mateles. 1975. Nitrogen metabolism in plant cell suspension cultures. I. Effect of amino acids on growth. Plant Physiol. 56:584-589. Benzion G. and R.L Phillips. 1988. Cytogenetic stability of maize tissue cultures: a cell line pedigree analysis. Genome 30:318-325. Beyl C.A. and G.C. Sharma. 1983. Picloram induced somatic embryogenesis in Gasteria and Haworthia. Plant Cell Tissue Organ Cult. 2:123-132. Bhaduri P.N. and P.N Ghosh. 1954. Chromosome squashes in cereals. Stain Technol. 29:269-276. Binarova P. and J. Dolezel. 1988. Alfalfa embryogenic cell suspension culture: growth and ploidy level stability. Plant Physiol. 133:561-566. Blackith R.E. and R.A. Reyment. 1971. Multivariate Morphometries. Academic Press, New York. Botti C. and I.K. Vasil. 1983. Plant regeneration by somatic embryogenesis from parts of cultured mature embryos of Pennisetum americanum (L.) K. Schum. Z. Pflanzenphysiol. 1 1 1 :31 9-325.

PAGE 315

308 Boucaud M.T. and J.M. Caultier. 1981. Cytophotometric study and statistical analysis of ploidy evolution in cultured tissues of Nicotiana tabacum L. Physiol. Plant. 51:207-214. Breiman A., T. Felsenburg and E. Galun. 1987b. Nor loci analysis in progenies of plants regenerated from the scutellar callus of bread wheat. Theor. Appl. Genet. 73:827-831 . Breiman A., D. Rotem-Abarbanell, A. Karp and H. Shaskin. 1987a. Heritable somaclonal variation in wild barley {Hordeum spontaneum). Theor. Appl. Genet. 74:104-112. Brettell R.I.S., E.S. Dennis, W.R. Scowcroft and W.J. Peacock. 1986a. Molecular analysis of a somaclonal mutant of maize alcohol dehydrogenase. Mol. Gen. Genet. 202:235-239. Brettell R.I.S., M.A. Pallotta, J. P. Gustafson and R. Appels. 1986b. Variation at the Nor loci in triticale derived from tissue culture. Theor. Appl. Genet. 71 :637-643. Brettell R.I.S. and E. Thomas. 1980. Reversion of Texas male-sterile cytoplasm maize in culture to give fertile, T-toxin resistant plants. Theor. Appl. Genet. 58:55-58. Britikov E.A., J. Schrauwen and H.F. Linskens. 1970. Proline as a source of nitrogen in plant metabolism. Acta Bot. Neerl. 19:515-520. Brown A.H.D. and B.S. Weir. 1983. Measuring genetic variability in plant populations. In S.D. Tanksley and T.J. Orton (eds.): "Isozymes in Plant Genetics and Breeding, Part A", pp. 219-239. Elsevier, Amsterdam. Brown P.T.H., E. Gobel and H. Lorz. 1991. RFLP analysis of Zea mays callus cultures and their regenerated plants. Theor. Appl. Genet. 81 :227-232. Brown R.M. 1967. Refined smear technique for obtaining large numbers of metaphases in corn root-tips. Maize Genet. Coop. News Lett. 41 :1 93-1 94. Burnham C.R. 1982. Details of the smear technique for studying chromosomes in maize. In W.F. Sheridan (ed.): "Maize for Biological Research", pp. 107-118. University Press, University of North Dakota, Grand Forks, North Dakota. Camara-Hernandez J. and P.C. Mangelsdorf. 1981 . Crosses of Zea diploperennis with corn. Maize Genet. Coop. News Lett. 55:15-17. Carlson L.A. and S.C. Price. 1 989a. Zea diploperennis B73 adapted to U.S. corn belt. Maize Genet. Coop. News Lett. 63:106-107. Carlson L.A. and S.C. Price. 1989b. Zea diploperennis-maize hybrid adapted to the U.S. corn belt. Maize Genet. Coop. News Lett. 63:107. Carman J.G., N.E. Jefferson and W.F. Campbell. 1988. Induction of embryogenic Triticum aestivum L. calli. 1. Quantification of cultivar and culture medium effects. Plant Cell Tissue Organ Cult. 12:83-95.

PAGE 316

309 Cassells A.C. 1985. Genetic, epigenetic and non-genetic variation in tissue culture derived plants. In Schafer-Menuhr (ed.): "In Vitro Techniques. Propagation and Long Term Storage", pp. 1 1 1-120. Martinus Nijhoff, Dordrecht, The Netherlands. Cassells A.C. and F.M. Morrish. 1987. Variation in adventitious regenerates of Begonia rex Putz. 'Lucille Closon' as a consequence of cell ontogeny, callus ageing and frequency of callus subculture. Sci. Hortic. 32:135-143. Cavallini A., R. Cremonini, M.C. Lupi and A. Bennici. 1986. In vitro culture of Bellevalia romana (L.) Rchb. II. Cytological study of callus and regenerated plantlets. Protoplasma 132: 58-63. Cavallini A., M.C. Lupi, R. Cremonini and A. Bennici. 1987. In vitro culture of Bellevalia romana (L.) Rchb. III. Cytological study of somatic embryos. Protoplasma 139:66-70. Cavallini A. and L. Natali. 1 989. Cytological analyses of in vitro somatic embryogenesis in Brimeura amethystina Salisb. (Liliaceae). Plant Sci. 62:255-261 . Chaleff R.S. 1981. Genetics of Higher Plants. Applications of Cell Culture. Cambridge University Press, London. Chang Y.F. 1983. Plant regeneration in vitro from leaf tissues derived from cultured immature embryos of Zea mays L. Plant Cell Rep. 2:183-185. Chee P.P. and D.M. Tricoli. 1988. Somatic embryogenesis and plant regeneration from cell suspension cultures of Cucumis sativus L. Plant Cell Rep. 7:274-277. Chen C. 1968. Root-tip squash technique. Maize Genet. Coop. News Lett. 42:174. Chen C. 1 969. The somatic chromosomes of maize. Can. J. Genet. Cytol. 1 1 :752-754. Chibbar R.N., J. Shyluk, F. Georges and F. Constabel. 1987. Role of proline in somatic embryogenesis in cultured carrot cells. Plant Physiol. 83 (Suppl.):76. Choudri R.S. and R. Krishan. 1946. Sex differentiation in Zea mays. Sci. & Cult. 1 1 :472476. Chowdhury M.K.U. and I.K. Vasil. 1993. Molecular analysis of plants regenerated from embryogenic cultures of hybrid sugarcane cultivars (Saccharum spp.). Theor. Appl. Genet. 86:181-188. Chowdhury M.K.U., V. Vasil and I.K. Vasil. 1994. Molecular analysis of plants regenerated from embryogenic cultures of wheat {Triticum aestivum L). Theor. Appl. Genet, (in press). Chu I.Y.E and S.L Kurtz. 1990. Commercialization of plant micropropagation. In P.M. Ammirato, D.A. Evans, W.R. Sharp and Y.P.S. Bajaj (eds.): "Handbook of Plant Cell Culture, Volume 5, Ornamental Species", pp. 126-164. McGraw-Hill, New York. Close K.R. and LA. Ludeman. 1987. The effect of auxin-like plant growth regulators and osmotic regulation on induction of somatic embryogenesis from elite maize inbreds. Plant Sci. 52:81-89.

PAGE 317

310 Coe E.H. and R.S. Poethig. 1982. Genetic factors affecting plant development. In W.F. Sheridan (ed.): "Maize for Biological Research", pp. 295-300. University Press, University of North Dakota, Grand Forks, North Dakota. Cohen J.D. and R.S. Bandurski. 1978. The bound auxins: protection of indole-3-acetic acid from peroxidase-catalyzed oxidation. Planta 1 39:203-208. Collins G.B., W.E. Vian and G.C. Phillips. 1978. Use of 4-amino-3,5,6-trichloropicolinic acid as an auxin source in plant tissue cultures. Crop Sci. 18:286-288. Conger B.V (ed.). 1981. Cloning Agricultural Plants via In Vitro Techniques. CRC Press, Boca Raton, Florida. Conger B.V., G.E. Hanning, D.J. Gray and J.K. McDaniel. 1983. Direct embryogenesis from mesophyll cells of orchard grass. Science 221 :850-851 . Conger B.V., L.L. Hilenski, K.W. Lowe and J.V. Carabia. 1982. Influence of different auxins at varying concentrations on callus induction and growth from embryo and leaf-tip explants in Gramineae. Environ. Exp. Bot. 22:39-48. Conger B.V., F.J. Novak, R. Afza and K. Erdelsky. 1987. Somatic embryogenesis from cultured leaf segments of Zea mays. Plant Cell Rep. 6:345-347. Constantin M.J. 1981. Chromosome instability in cell and tissue cultures and regenerated plants. Environ. Exp. Bot. 21 :359-368. Cooper D.B., R.G. Sears, G.L Lockhart and B.L Jones. 1986. Heritable somaclonal variation in gliadin proteins of wheat plants derived from immature embryo callus culture. Theor. Appl. Genet. 71 :784-790. Corcuera V.R. and J.L. Magoja. 1988a. Diploperennial teosinte-maize hybrids: inheritance of evolutive cycle. Maize Genet. Coop. News Lett. 62:77. Corcuera V.R. and J.L. Magoja. 1988b. Diploperennial teosinte-maize hybrids: inheritance of prolificity. Maize Genet. Coop. News Lett. 62:77-78. Corcuera V.R. and J.L. Magoja. 1988c. Diploperennial teosinte-maize hybrids: inheritance of tassel traits. Maize Genet. Coop. News Lett. 62:78-79. Crosswhite F.S. 1982. Corn (Zea mays) in relation to its wild relatives. Desert Plants 3:193-202. Cullis C.A. 1986. Unstable genes in plants. Symp. Soc. Exp. Biol. 40:77-84. Cummings D.P., C.E. Green and D.D. Stuthman. 1976. Callus induction and regeneration in oats. Crop Sci. 16:465-470. D'Amato F. 1975. The problem of genetic stability in plant tissue and cell cultures. In O.H. Frankel and J.G. Hawkes (eds.): "Crop Genetic Resources for Today and Tomorrow", pp. 333-348. Cambridge University Press, New York. D'Amato F. 1977. Cytogenetics of differentiation in tissue and cell cultures. In J. Reinert and V.P.S. Bajaj (eds.): "Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture", pp. 343-464. Springer Verlag, Berlin.

PAGE 318

311 D'Amato F. 1 985. Cytogenetics of plant cell and tissue cultures and their regenerates. CRC Crit. Rev. Plant Sci. 3:73-1 12. Dahleen L.S. 1989. Somaclonal variation in oat (Avena sativa L.) lines derived from tissue culture. Ph.D. thesis. Univ. of Minnesota, St. Paul/Minneapolis. Dahleen L.S. and G.C. Eizenga. 1 990. Meiotic and isozymic characterization of plants regenerated from euploid and selfed monosomic tall fescue embryos. Theor. Appl. Genet. 79:39-44. Dahnous K., G.T. Vigue, A.G. Law, C.F. Konzak and D.G. Miller. 1982. Height and yield response of selected wheat, barley, and triticale cultivars to ethephon. Agron. J. 74:580-582. Dale P.J. 1980. Embryoids from cultured immature embryos of Lolium multiflorum. Z. Pflanzenphysiol. 100:73-77. Dale P.J., E. Thomas, R. Brettell and W. Wernicke. 1981 . Embryogenesis from cultured immature inflorescences and nodes of Lolium multiflorum. Plant Cell Tissue Organ Cult. 1 :47-55. Davies P.A., M.A. Pallatto, S.A. Ryan, W.R. Scowcroft and P.J. Larkin. 1986. Somaclonal variation in wheat: genetic and cytogenetic characterization of alcohol dehydrogenase 1 mutants. Theor. Appl. Genet. 72:644-653. Day A. and T.H.N. Ellis. 1984. Chloroplast DNA deletions associated with wheat plants regenerated from pollen: possible basis for maternal inheritance of chloroplasts. Cell 39:359-368. Day A. and T.H.N. Ellis. 1 985. Deleted forms of plastid DNA in albino plants from cereal anther culture. Curr. Genet. 9:671-678. de Klerk G.-J. 1990. How to measure somaclonal variation: a review. Acta Bot. Neerl. 39:129-144. Dewald S.G. and G.A. Moore. 1987. Somaclonal variation as a tool for the improvement of perennial fruit crops. Fruit Var. J. 41 :54-57. Diakonu P. 1961. A new method for determining the viability of corn pollen. Agrobiologiya 2: 1 93-1 98. Dolezel J. and P. Binarova. 1989. The effects of colchicine on ploidy level, morphology and embryogenic capacity of alfalfa suspension cultures. Plant Sci. 64:213-219. Dudits D., G. Nemet and Z. Haydu. 1975. Study of callus growth and organ formation in wheat (Triticum aestivum) tissue cultures. Can. J. Bot. 53:957-967. Duncan D.R. and J.M. Widholm. 1988. Improved plant regeneration from maize callus cultures using 6-benzylaminopurine. Plant Cell Rep. 7:452-455. Duncan D.R., M.E. Williams, B.E. Zehr and J.M. Widholm. 1985. The production of callus capable of plant regeneration from immature embryos of numerous Zea mays genotypes. Planta 165:322-332.

PAGE 319

312 Dunwell J.M. and N. Thurling. 1985. Role of sucrose in microspore embryo production in Brassica napus ssp. oleifera. J. Exp. Bot. 36:1 478-1 491 . Earle E.B. and V.E. Gracen. 1985. Somaclonal variation in progeny of plants from corn tissue cultures. In R.R. Henke, K.W. Hughes, M.J. Constantin and A. Hollaender (eds.): 'Tissue Culture in Forestry and Agriculture", pp. 139-152. Plenum Press, New York. Edallo S., C. Zucchinali, M. Perenzin and F. Salamini. 1981. Chromosomal variation and frequency of spontaneous mutation associated with in vitro culture and plant regeneration in maize. Maydica 26:39-56. Eizenga G.C. 1987. Cytogenetic and isozymic characterization of anther-panicle culture derived tall fescue aneuploids. Euphytica 36:1 75-1 79. Eizenga G.C. 1989. Meiotic analysis of tall fescue somaclones. Genome 32:373-379. Eizenga G.C. and L.S. Dahleen. 1990. Callus production, regeneration and evaluation of plants from cultured inflorescences of tall fescue (Festuca arundinacea Schreb.). Plant Cell Tissue Organ Cult. 22:7-15. Emerson R.A. 1 924. Control of flowering in teosinte. Short-day treatment brings early flowers. J. Hered. 15:41-48. Endo T. 1968. Indoleacetic acid oxidase activities of horseradish and other plant peroxidase isozymes. Plant Cell Physiol. 9:333-341. Evans D.A. 1989. Somaclonal variation Genetic basis and breeding applications. Trends Genet. 5:46-50. Evans D.A. and W.R. Sharp. 1986. Applications of somaclonal variation. Bio/Technology 4:528-532. Evans D.A. and W.R. Sharp. 1988. Somaclonal variation and its application in plant breeding. Intern. Assoc. Plant Tiss. Cult., Newslett. 54:2-10. Fahey J.W., J.N. Reed, T.L Readdy and G.M. Pace. 1986. Somatic embryogenesis from three commercially important inbreds of Zea mays. Plant Cell Rep. 5:35-38. Feung C.S., R.H. Hamilton and R.O. Mumma. 1977. Metabolism of indole-3-acetic acid. IV. Biological properties of amino acid conjugates. Plant Physiol. 59:91-93. Findley W.R., L.R. Nault, W.E. Styer and DT. Gordon. 1982. Inheritance of maize chlorotic dwarf virus resistance in maize x Zea diploperennis backcrosses. Maize Genet. Coop. News Lett. 56:165-166. Finer J.J. 1987. Direct somatic embryuogenesis and plant regeneration from immature embryos of hybrid sunflower (Helianthus annuus L.) on a high sucrosecontaining medium. Plant Cell Rep. 6:372-374. Finer J.J. 1 988. Apical proliferation of embryogenic tissue of soybean [Glycine max (L.) Merrill]. Plant Cell Rep. 7:238-241.

PAGE 320

313 Fitch M.M.M. and P.H. Moore. 1990. Comparison of 2,4-D and picloram for selection of long-term totipotent green callus cultures of sugarcane. Plant Cell Tissue Organ Cult. 20:157-163. Fromm M.E., F. Morrish, C. Armstrong, R. Williams, J. Thomas and T.M. Klein. 1990. Inheritance and expression of chimeric genes in the progeny of transgenic maize plants. Bio/Technology 8:833-839. Fukui K. 1986. Case histories of genetic variability in vitro: rice. In I.K. Vasil (ed.): "Cell Culture and Somatic Cell Genetics of Plants, Volume 3, Plant Regeneration and Genetic Variability", pp. 385-398. Academic Press, New York. Furuya M. 1984. Cell division patterns in multicellular plants. Ann. Rev. Plant Physiol. 35:349-373. Galiba G. and J. Sutka. 1989. Frost resistance of somaclones derived from Triticum aestivum L. winter wheat calli. Plant Breeding 1 02:1 01 -1 04. Galiba G., Z. Kertesz, J. Sulka and L. Sagi. 1985. Differences in somaclonal variation in three winter wheat Triticum aestivum varieties. Cereal Res. Commun. 13:343350. Galinat W.C. 1981. The inheritance and linkage of perennialism derived from diploperennis. Maize Genet. Coop. News Lett. 55:107. Gamborg O.L. 1970. The effects of amino acids and ammonium on the growth of plant cells in suspension culture. Plant Physiol. 45:372-375. Gamborg O.L, J. P. Shyluk, D.S. Brar and F. Constabel. 1977. Morphogenesis and plant regeneration from callus of immature embryos of sorghum. Plant Sci. Lett. 10:67-74. Gaska J.M. and E.S. Oplinger. 1988. Yield, lodging, and growth characteristics in sweet corn as influenced by ethephon timing and rate. Agron. J. 80:722-726. Gavazzi G., C. Tonelli, G. Todesco, E. Arreghini, F. Raffaldi, F. Vecchio, G. Barbuzzi, M. Biasini and F. Sala. 1987. Somaclonal variation versus chemically induced mutagenesis in tomato (Lycopersicon esculentum L). Theor. Appl. Genet. 74:733-738. Gengenbach B.G., J.A. Conelly, D.R. Pring and M.F. Conde. 1981. Mitochondrial DNA variation in maize plants regenerated during tissue culture selection. Theor. Appl. Genet. 59:161-167. Genovesi A.D. and G.B. Collins. 1982. In vitro production of haploid plants of corn via anther culture. Crop Sci. 22:1137-1144. Ghosh A. and V.N. Gadgil. 1979. Shift in ploidy of callus tissues: a function of growth substances. Indian J. Exp. Biol. 17:562-564. Gmitter F.G. Jr., X. Ling, C. Cai and J.W. Grosser. 1991. Colchicine-induced polyploidy in Citrus embryogenic cultures, somatic embryos, and regenerated plantlets. Plant Sci. 74:135-141.

PAGE 321

314 Gnanadesikan R. 1977. Methods for Statistical Data Analysis of Multivariate Observations. John Wiley and Sons, New York. Gobel E., P.T.H. Brown and H. Lorz. 1986. In vitro culture of Zea mays L. and analyses of regenerated plants. In "Nuclear techniques and in vitro culture for plant improvement", pp. 21 -27. International Atomic Energy Agency, Vienna, Austria. Goeschel J.E. and S.J. Kays. 1975. Concentration dependencies of some effects of ethylene on etiolated pea, peanut, bean and cotton seedlings. Plant Physiol. 55:670-677. Good N.E., W.A. Andreae and M.W. van Ysselstein. 1956. Studies on 3-indoleacetic acid metabolism. II. Some products of the metabolism of exogenous indoleacetic acid in plant tissues. Plant Physiol. 31 :231 -235. Good N.E., R. Hangarter and P. Carlson. 1982. The use of conjugates of indoleacetic acid as auxins in tissue culture. In E.D. Earle and Y. Demarly (eds.): "Variability in Plants Regenerated from Tissue Culture", pp. 1 40-1 47. Praeger, New York. Goodman R.M., H. Hauptli, A. Crossway and V.C. Knauf. 1987. Gene transfer in crop improvement. Science 236:48-54. Gordon-Kamm W.J., T.M. Spencer, M.L Mangano, T.R. Adams, R.J. Daines, W.G. Start, J.V. O'Brien, S.A. Chambers, W.R. Adams Jr., N.G. Willetts, T.B. Rice, C.J. Mackey, R.W. Krueger, A.P. Kausch and P.G. Lemaux. 1990. Transformation of maize cells and regeneration of fertile transgenic plants. Plant Cell 2:603-61 8. Gould A.R. 1984. Control of the cell cycle in cultured plant cells. CRC Crit. Rev. Plant Sci. 1:315-344. Gould A.R. 1986. Factors controlling generation of variability in vitro. In I.K. Vasil (ed.): "Cell Culture and Somatic Cell Genetics of Plants, Volume 3, Plant Regeneration and Genetic Variability", pp. 549-569. Academic Press, New York. Grambow H.J. and B. Langenbeck-Schwich. 1983. The relationship between oxidase activity, peroxidase activity, hydrogen peroxide and phenolic compounds in the degradation of indole-3-acetic acid in vitro. Planta 157:131-137. Gray D.J. and B.V. Conger. 1985. Influence of dicamba and casein hydrolysate on somatic embryo number and culture quality in cell suspensions of Dactylis glomerata (Gramineae). Plant Cell Tissue Organ Cult. 4:123-133. Gray D.J., B.V. Conger and G.E. Hanning. 1984. Somatic embryogenesis in suspension and suspension-derived callus cultures of Dactylis glomerata. Protoplasma 122:196-202. Green C.E. 1977. Prospects for crop improvement in the field of cell culture. HortScience 12:131-134. Green C.E. 1982. Somatic embryogenesis and plant regeneration from the friable callus of Zea mays L. In A. Fujiwara (ed.): "Plant Tissue Culture 1982', pp. 107108. Maruzen, Tokyo.

PAGE 322

315 Green C.E. 1983. New developments in plant tissue culture and plant regeneration. In A. Hollaender, A.I. Laskin and P. Rogers (eds.): "Basic Biology of New Developments in Biotechnology", pp. 1 95-209. Plenum Press, New York. Green C.E., C.L Armstrong and P.C. Anderson. 1983. Somatic cell genetic systems in corn. In K. Downey, R.W. Voellmy, F. Ahmad and J. Schultz (eds.): "Advances in Gene Technology: Molecular Genetics of Plants and Animals", pp. 147-157. Academic Press, New York. Green C.E. and R.L. Phillips. 1975. Plant regeneration from tissue culture of maize. Crop Sci. 15:417-421. Guzman Mejia R. 1 978. Una nueva localidad para el teosinte Zea perennis y primer reporte de Zea mexicana para Jalisco. Bol. Informativo del Inst, de Bot., Univ. de Guadalajara 6:9-10. Haard N.F. 1977. Physiological role of peroxidase in postharvest fruits and vegetables. In R.L. Ory and A.J. St. Angelo (eds.): "Enzymes in Food and Beverage Processing", pp. 143-171. American Chemical Society, Washington, DC. Hahne B. and F. Hoffmann. 1986. Cytogenetics of protoplast cultures of Brachycome dichromosomatica and Crepis capillaris and regeneration of plants. Theor. Appl. Genet. 72:244-251. Hangarter R.P. and N.E. Good. 1981. Evidence that IAA conjugates are slow-release sources of free IAA in plant tissues. Plant Physiol. 68:1424-1427. Hangarter R.P., M.D. Peterson and N.E. Good. 1980. Biological activities of indolacetylamino acids and their use as auxins in tissue culture. Plant Physiol. 65:761-767. Hanna W.W., C. Lu and I.K. Vasil. 1984. Uniformity of plants regenerated from somatic embryos of Panicum maximum Jacq. (Guinea grass). Theor. Appl. Genet. 67:155-159. Hanning G.E. and B.V. Conger. 1982. Embryoid and plantlet formation from leaf segments oWactylis glomerata L. Theor. Appl. Genet. 63:155-159. Hanson M.R. 1984. Stability, variation and recombination in plant mitochondrial genomes via tissue culture and somatic hybridisation. Oxf. Surv. Plant Mol. Cell Biol. 1 :33-52. Hanzel J.J., J.P. Miller, M.A. Brinkman and E. Fendos. 1985. Genotype and media effects on callus formation and regeneration in barley. Crop Sci. 25:27-31 . Harris R.J. 1 975. A Primer of Multivariate Statistics. Academic Press, New York. Hashim Z.N., W.F. Campbell and J.G. Carman. 1990. Morphological analyses of spring wheat (CIMMYT cv. PCYT-10) somaclones. Plant Cell Tissue Organ Cult. 20:9599. Haydu Z. and I.K. Vasil. 1981 . Somatic embryogenesis and plant regeneration from leaf tissues and anthers of Pennisetum purpureum. Theor. Appl. Genet. 59:269-273.

PAGE 323

316 Heinz D.J., M. Krishnamurthi, L.G. Nickell and A. Maretzki. 1977. Cell, tissue and organ culture in sugar cane improvement. In J. Reinert and Y.P.S. Bajaj J. (eds.): "Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture", pp. 3-17. Springer-Verlag, Berlin. Heinz D.J. and G.W.P. Mee. 1971 . Morphologic, cytogenetic and enzymatic variation in Saccharum species hybrid clones derived from callus tissue. Amer. J. Bot. 58:275-262. Heinz D.J., G.W.P. Mee and LG. Nickell. 1969. Chromosome numbers of some Saccharum species hybrids and their cell suspension cultures. Amer. J. Bot. 56:450-456. Heslop-Harrison J. 1957a. The experimental modification of sex expression in flowering plants. Biol. Rev. 32:38-90. Heslop-Harrison J. 1957b. The sexuality of flowers. New Biol. 23:9-28. Heslop-Harrison J. 1959. Growth substances and flower morphogenesis. J. Linn. Soc. Lond., Botany, 65:269-281. Heslop-Harrison J. 1961. The experimental control of sexuality and inflorescence structure in Zea mays L. Proc. Linn. Soc. Lond. 172:108-123. Hibberd K.A. and C.E. Green. 1982. Inheritance and expression of lysine plus threonine resistance selected in maize tissue culture. Proc. Nat. Acad. Sci., USA, 79:559-563. Hinman R.L. and J. Lang. 1965. Peroxidase catalyzed oxidation of indole-3-acetic acid. Biochemistry 4:1 44-1 48 Ho W. and I.K. Vasil. 1983. Somatic embryogenesis in sugarcane (Saccharum officinarum L). I. The morphology and physiology of callus formation and the ontogeny of somatic embryos. Protoplasma 1 1 8:1 69-1 80. Holbrook L.A. and S.J. Molnar. 1984. Response in tissue culture of undeveloped ears from young plants. Maize Genet. Coop. News Lett. 58:161-162. Holm R.E. and F.B. Abeles. 1968. The role of ethylene in 2,4-D induced growth inhibition. Planta 78:293-304. Holtzer H. and N. Rubenstein. 1977. Binary decisions, quantal cell cycles, and cell diversification. In L. Nover and K. Mothes (eds.): "Cell Differentiation in Microorganisms, Plants and Animals", pp. 424-437. Fisher, Jena/Elsevier, Amsterdam. Huang A.E. and N.F. Haard. 1977. Properties of IAA oxidase from ripening tomatoes. J. Food Biochem. 242:2470-2473. Hubbard E.T., J.P Cook, M.D. Hollingsworth and C.R. Cowley. 1984. In vitro culture, regeneration and analysis of seven inbred corn lines. Agron. Abstr. 76:72.

PAGE 324

317 Hunault G. 1979. Recherches sur le comportement des fragments d'organes et des tissus de Monocotyledones cultives in vitro. V. Discussion. Rev. Cytol. Biol. Veget. 2:259-287. Hunsinger H. and K. Schauz. 1987. The influence of dicamba on somatic embryogenesis and frequency of plant regeneration from cultured immature embryos of wheat {Triticum aestivum L). Plant Breeding 98:1 1 9-1 23. litis H.H., J.F. Doebley, R. Guzman and B. Pazy. 1979. lea diploperennis (Gramineae): a new teosinte from Mexico. Science 203:1 86-1 88. International Union of Biochemistry Nomenclature Committee. 1984. Enzyme Nomenclature 1 984. Academic Press, New York. Irvine J.E. 1984. The frequency of marker changes in sugarcane plants regenerated from callus culture. Plant Cell Tissue Organ Cult. 3:201 -201 . Irvine J.E., M. Fitch and P.H. Moore. 1983. The induction of callus in sugarcane tissue cultures by selected chemicals. Plant Cell Tissue Organ Cult. 2:141-149. Jackson J.A. and P.J. Dale. 1988. Callus induction, plant regeneration and an assessment of cytological variation in regenerated plants of Lolium multiflorum L. J. Plant Physiol. 132:351-355. Jackson M.H. and D.J. Campbell. 1979. Effects of benzyladenine and gibberellic acid on the responses of tomato plants to anaerobic root environments and to ethylene. New Phytol. 82:331-340. Jefferies R.L. 1980. The role of organic solutes in osmoregulation in halophytic higher plants. In D.W. Rains, R.C. Valentine and A. Hollaender (eds.): "Genetic Engineering for Osmoregulation: Impact on Plant Productivity for Food, Chemicals, and Energy", pp. 135-154. Plenum Press, New York. Jha S. and S. Sen. 1987. Nuclear changes and organogenesis during callus culture of Urginea indica Kunth., Indian squill. Cytologia 52:433-438. Joarder O.I., N.H. Joarder and P.J. Dale. 1986. In vitro response of leaf tissues from Lolium multiflorum A comparison with leaf segment position, leaf age and in vivo mitotic activity. Theor. Appl. Genet. 73:286-291 . Johnson S.S., R.L. Phillips and H.W. Rines. 1987. Possible role of heterochromatin in chromosome breakage induced by tissue culture in oats {Avena sativa L). Genome 29:439-446. Jones J.B. and T. Murashige. 1974. Tissue culture propagation of Aechmaea fasciata Baker and other bromeliads. Comb. Proc. Intl. Plant Prop. Soc. 25:1 17-126. Jordan M.C. and E.N. Larter. 1985. Somaclonal variation in triticale (x Triticosecale Wittmack) cv. Carman. Can. J. Genet. Cytol. 27:151-157. Kamada H., T. Kiyosus and H. Harada. 1988. New methods for somatic embryo induction and their use for synthetic seed production. In Vitro Cell Dev. Biol. 24:71 A (abstract).

PAGE 325

318 Karp A. 1 989. Can genetic instability be controlled in plant tissue cultures? Intern. Assoc. Plant Tiss. Cult., Newslett. 58:2-11. Karp A. and S.W.J. Bright. 1985. On the causes and origins of somaclonal variation. Oxf. Surv. Plant Mol. Cell Biol. 2:199-234. Karp A. and S. A. Maddock. 1 984. Chromosome variation in wheat plants regenerated from cultured immature embryos. Theor. Appl. Genet. 67:249-255. Karp A., R.S. Nelson, E. Thomas and S.W.J. Bright. 1982. Chromosome variation in protoplast-derived potato plants. Theor. Appl. Genet. 63:265-272. Karp A., S.H. Steele, S. Parmer, M.G.K. Jones, P.R. Shewry and A. Breiman. 1987. Relative stability among barley plants regenerated from cultured immature embryos. Genome 29:405-412. Kitto S.L and J. Janick. 1985. Hardening treatment increased survival of syntheticallycoated asexual embryos of carrot. J. Amer. Soc. Hort. Sci. 1 10:283-286. Kobayashi S. 1987. Uniformity of plants regenerated from orange (Citrus sinensis Osb.) protoplasts. Theor. Appl. Genet. 74:10-14. Konsler J.V. and L.J. Grabau. 1989. Ethephon as a morphological regulator for corn. Agron. J. 81 :849-852. Krikorian A.D., S.A. Staicu and R.P. Kan. 1981. Karyotype analysis of a daylily clone reared from aseptically cultured tissues. Ann. Bot. 47:121-131. Krishnamurthi M. 1982. Disease resistance in sugarcane developed through tissue culture. In A. Fujiwara (ed.): "Plant Tissue Culture 1982', pp. 769-770. Maruzen, Tokyo. Kumar A.S., O.L. Gamborg and M.W. Nabors. 1988. Plant regeneration from cell suspensions cultures of Vigna aconitifolia. Plant Cell Rep. 7:1 38-1 41 . Lapitan N.L.V., R.G. Sears and B.S. Gill. 1984. Translocation and other karyotypic structural changes in wheat and rye hybrids regenerated from tissue culture. Theor. Appl. Genet. 68:547-554. Larkin P.J. 1 987. Somaclonal variation, history, method and meaning. Iowa State J. of Research 61 : 393-434. Larkin P.J., P.M. Banks, R. Bhati, R.I.S. Brettell, P.A. Davies, S.A. Ryan, W.R. Scowcroft, L.H. Spindler and G.J. Tanner. 1989. From somatic variation to variant plants: mechanisms and applications. Genome 31 :705-71 1 . Larkin P.J., S.A. Ryan, R.I.S. Brettell and W.R. Scowcroft. 1984. Heritable somaclonal variation in wheat. Theor. Appl. Genet. 67:443-455. Larkin P.J. and W.R. Scowcroft. 1981. Somaclonal variation-a novel source of variability from cell cultures for plant improvement. Theor. Appl. Genet. 60:197214.

PAGE 326

319 Larkin P.J. and W.R. Scowcroft. 1983a. Somaclonal variation and eyespot toxin tolerance in sugarcane. Plant Cell Tissue Organ Cult. 2:1 1 1-121 . Larkin P.J. and W.R. Scowcroft. 1983b. Somaclonal variation and crop improvement. In T. Kosuge, CP. Meredith and A. Hollaender (eds.): "Genetic Engineering of Plants", pp. 289-31 4. Plenum Press, New York. Lat J.B. and M.M. Lantin. 1976. Agronomic performance of sugar cane clones derived from callus tissue. Philippine J. Crop Sci. 1 :1 17-123. Lazar M.D., T.H.H. Chen, LV. Gusta and K.K. Kartha. 1988. Somaclonal variation for freezing tolerance in a population derived from norstar winter wheat. Theor. Appl. Genet. 75:480-484. Lazzeri P.A., D.F. Hildebrand and G.B. Collins. 1987. Soybean somatic embryogenesis: effects of hormone and culture manipulations. Plant Cell Tissue Organ Cult. 10:197-208. Lazzeri P.A. and H. Lorz. 1988. In vitro genetic manipulation of cereals and grasses. Adv. Cell Cult. 6:291-235. Le Rudulier D., A.R. Strom, A.M. Dandekar, LT. Smith and R.C. Valentine. 1984. Molecular biology of osmoregulation. Science 224:1 064-1 068. Lebart L, A. Morineau and K.M. Warwick. 1984. Multivariate Descriptive Statistical Analysis. John Wiley & Sons, New York. Lee M. and R.L Phillips. 1987. Genomic rearrangements in maize induced by tissue culture. Genome 29:122-128. Lee M. and R.L. Phillips. 1988. The chromosomal basis of somaclonal variation. Ann. Rev. Plant Physiol. Plant Mol. Biol. 39:413-437. Lee T.T., A.N. Starratt and J.J. Jevnikar. 1982. Regulation of enzymic oxidation of indole-3-acetic acid by phenols: structure-activity relationships. Phytochemistry 21:517-523. Linacero R. and A.M. Vazquez. 1986a. Somatic embryogenesis and plant regeneration from leaf tissues of rye (Secale cereale L). Plant Sci. 44:21 9-222. Linacero R. and A.M. Vazquez. 1986b. Somaclonal variation in plants regenerated from embryo calluses in rye (Secale cereale L). In W. Horn, C.J. Jensen, W. Odenbach and O. Schieder (eds.): "Genetic Manipulation in Plant Breeding", pp. 479-481 . Walter de Gruyter, Berlin. Litz R.E. and R.A. Conover. 1983. High frequency somatic embryogenesis from Carica suspension cultures. Ann. Bot. 51 :683-686. Liu M.C. and W.H. Chen. 1976. Tissue and cell cultures as aids to sugarcane breeding. 1 . Creation of genetic variation through callus culture. Euphytica 25:393-403. Liu M.C, Y.J. Huang and S.C Shih. 1972. The in vitro production of plants from several tissues of Saccharum species. J. Agric. Ass. China, New Series 77:52-58.

PAGE 327

320 Lorz H. and P. Brown. 1986. Variability in tissue culture derived plants possible origins; advantages and drawbacks. In W. Horn, C.J. Jensen, W. Odenbach and 0. Schieder (eds.): "Genetic Manipulation in Plant Breeding", pp. 513-534. Walter de Gruyter, Berlin. Lorz H., E. Gobel and P. Brown. 1988. Advances in tissue culture and progress towards genetic transformation of cereals. Plant Breeding 1 00:1-25. Lorz H. and W.R. Scowcroft. 1983. Variability among plants and their progeny regenerated from protoplasts of Su/su heterozygotes of Nicotiana tabacum. Theor. Appl. Genet. 66:67-75. Lowe K., D.B. Taylor, P. Ryan and K.E. Paterson. 1985. Plant regeneration via organogenesis and embryogenesis in the maize inbred line B73. Plant Sci. 41:125-132. Lu C. and I.K. Vasil. 1981a. Somatic embryogenesis and plant regeneration from leaf tissues of Panicum maximum Jacq. Theor. Appl. Genet. 59:275-280. Lu C. and I.K. Vasil. 1982. Somatic embryogenesis and plant regeneration in tissue cultures of Panicum maximum Jacq. Amer. J. Bot. 69:77-81 . Lu C., I.K. Vasil and P. Ozias-Akins. 1982. Somatic embryogenesis in Zea mays L Theor. Appl. Genet. 62:109-112. Lu C., V. Vasil and I.K. Vasil. 1983. Improved efficiency of somatic embryogenesis and plant regeneration in tissue cultures of maize (Zea mays L). Theor. Appl. Genet. 66:285-289. Lu C.C., L. Currah and E.B. Peffley. 1989. Somatic embryogenesis and plant regeneration in diploid Allium fistulosum x A. cepa Fi hybrid onions. Plant Cell Rep. 7:696-700. Maddock S.E. 1985. Cell culture, somatic embryogenesis and plant regeneration in wheat, barley, oats, rye and triticale. In S.W.J. Bright and M.G.K. Jones (eds.): "Cereal Tissue and Cell Culture", pp. 131-175. Martinus Nijhoff/W. Junk, Dordrecht, The Netherlands. Maddock S.E., V.A. Lancaster, R. Risiott and J. Franklin. 1983. Plant regeneration from cultured immature embryos and inflorescences of 25 cultivars of wheat {Triticum aestivum). J. Exp. Bot. 34:915-926. Maddock S.E., R. Risiott, S. Parmar, M.G.K. Jones and P.R. Shewry. 1985. Somaclonal variation in the gliadin patterns of grains of regenerated wheat plants. J. Exp. Bot. 36:1976-1984. Magoja J.L. and I.G. Palacios. 1987. Early expression of heterosis in diploperennial teosinte-maize hybrids. Maize Genet. Coop. News Lett. 61 :63. Magoja J.L. and G. Pischedda. 1986. Introgression of teosinte germplasm in maize: a method to improve heterosis and variability. Maize Genet. Coop. News Lett 60:82-83.

PAGE 328

321 Maloney M.M., J.F. Hall, G.M. Robinson and M.C. Elliott. 1983. Auxin requirements of sycamore cells in suspension culture. Plant Physiol. 71 :927-931 . Mangelsdorf P.C. 1974. Corn. Its Origin, Evolution, and Improvement. The Belknap Press of Harvard University Press, Cambridge, Massachusetts. Mangelsdorf P.C, LM. Roberts and J.S. Rogers. 1981. Crosses of Zea diploperennis with corn. Maize Genet. Coop. News Lett. 55:19-21. McCann A.W., G. Cooley and J. van Dreser. 1988. A system for routine plantlet regeneration of sunflower {Helianthus annuus L.) from immature embryo-derived callus. Plant Cell Tissue Organ Cult. 14:108-110. McClintock B. 1 978. Mechanisms that rapidly reorganize the genome. Stadler Genet. Symp. 10:25-48. McCoy T.J. and R.L. Phillips. 1982. Chromosome stability in maize (Zea mays) tissue cultures and sectoring in some regenerated plants. Can. J. Genet. Cytol. 24:559-565. McCoy T.J., R.L. Phillips and H.W. Rines. 1982. Cytogenetic analysis of plants regenerated from oat {Avena sativa) tissue cultures: high frequency of partial chromosome loss. Can. J. Genet. Cytol. 24:37-50. McDaniel J.K., B.V. Conger and E.T. Graham. 1982. A histological study of tissue proliferation, embryogenesis, and organogenesis from tissue cultures of Dactylis glomerata L. Protoplasma 110:121-1 28. McDonnell R.E. and B.V. Conger. 1984. Callus induction and plantlet formation from mature embryo explants of Kentucky bluegrass. Crop Sci. 24:573-578. Mehta U., I.V.R. Rao and H.Y. Mohan Ram. 1982. Somatic embryogenesis in bamboo. In A. Fujiwara (ed.): "Plant Tissue Culture 1982", pp. 109-1 10. Maruzen, Tokyo. Meijer E.G.M. and D.C.W. Brown. 1987. Role of exogenous reduced nitrogen and sucrose in rapid high frequency somatic embryogenesis in Medicago sativa. Plant Cell Tissue Organ Cult. 10:1 1-19. Meins F. 1983. Heritable variation in plant cell cultures. Ann. Rev. Plant Physiol. 34:327-348. Metakovsky E.V., A.Y. Novoselskaya and A. A. Sozinov. 1987. Problems of interpreting results obtained in studies of somaclonal variation in gliadin proteins in wheat. Theor. Appl. Genet. 73:764-766. Miller P.D. 1982. Maize pollen: collection and enzymology. In W.F. Sheridan (ed.): "Maize for Biological Research", pp. 279-293. University Press, University of North Dakota, Grand Forks, North Dakota. Mitra G.C. 1985. Somaclonal variation as a strategy for plant improvement. Cell Chromosome Res. 8:59-68. Morel G. 1960. Producing virus-free cymbidiums. Amer. Orchid Soc. Bull. 29:495-497.

PAGE 329

322 Morel G. 1963. La culture in vitro du meristeme apical de certaines Orchidees. C.R. Acad. Sci. Paris 256:4955-4957. Morel G. 1 964a. Tissue culture-a new means of clonal propagation of orchids. Amer. Orchid Soc. Bull. 33:473-478. Morel G. 1964b. La culture in vitro du meristdme apical. Rev. Cytol. Biol. Veg. 27:304314. Morel G. 1965a. Clonal propagation of orchids by meristem culture. Cymbidium Soc. News 20:3-1 1 . Morel G. 1965b. Eine neue Methode erbgleicher Vermehrung: die Kultur von Triebspitzen-Meristemen. Die Orchidee 16:165-176. Morrish F.M., W.W. Hanna and I.K. Vasil. 1990. The expression and perpetuation of inherent somatic variation in regenerants from embryogenic cultures of Pennisetum glaucum (L.) R. Br. (pearl millet). Theor. Appl. Genet. 80:409-416. Morrish F., V. Vasil and I.K. Vasil. 1987. Developmental morphogenesis and genetic manipulation in tissue and cell cultures of the Gramineae. Adv. Genet. 24:431499. Morrison D.F. 1967. Multivariate Statistical Methods. McGraw-Hill, New York. Moss G.I and J. Heslop-Harrison. 1968. Photoperiod and pollen sterility in maize. Ann. Bot. 32:833-846. Murashige T. and F. Skoog. 1962. A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant. 15:473-497. Murashige T. and R. Nakano. 1 967. Chromosome complement as a determinant of the morphogenetic potential of tobacco cells. Amer. J. Bot. 54:963-970. Murata M. and T.J. Orton. 1984. Chromosome fusions in cultured cells of celery. Can. J. Genet. Cytol. 26:395-400. Nafziger E.D., LM. Wax and CM. Brown. 1986. Response of five winter wheat cultivars to growth regulators and increased nitrogen. Crop Sci. 26:767-770. Nault L.R. 1980. Maize bushy stunt and corn stunt: a comparison of disease symptoms, pathogen host ranges and vectors. Phytopathology 70:659-662. Nault L.R. and W.R. Findley. 1981. Zea diploperennis: a primitive relative offers new traits to improve corn. Ohio Report on Research and Development in Agriculture, Home Economics and Natural Resources 66:90-92. Nault L.R, R.E. Gingery and D.T. Gordon. 1980. Leafhopper transmission and host range of maize rayado fino virus. Phytopathology 70:709-712. Nault L.R., D.T. Gordon, V.D. Damsteegt and H.H. litis. 1982. Response of annual and perennial teosintes (Zea) to six maize viruses. Plant Disease 66:61 -62.

PAGE 330

323 Nitsch C. 1977. Culture of isolated microspores. In J. Reinert and Y.P.S. Bajaj (eds.): "Applied and Fundamental Aspects of Plant Cell, Tissue and Organ Culture", pp. 268-278. Springer Verlag, New York. Norberg O.S., S.C. Mason and S.R. Lowry. 1988. Ethephon influence on harvestable yield, grain quality, and lodging of corn. Agron.J. 80:768-772. Norberg O.S., S.C. Mason and S.R. Lowry. 1989. Ethephon alteration of corn plant morphology. Agron. J. 81 :603-609. Novak F.J., M. Dolezelova, M. Nesticky and A. Piovarci. 1983. Somatic embryogenesis and plant regeneration in lea mays L. Maydica 28:381-390. Nuti-Ronchi V., M.A. Caligo, M. Nozzolini and G. Luccarini. 1984. Stimulation of carrot somatic embryogenesis by proline and serine. Plant Cell Rep. 3:210-214. Oono K. 1 978. Test tube breeding of rice by tissue culture. Trop. Agric. Res. Series, Ministry Agric. Forest. (Japan) 1 1 : 109-1 24. Oono K. 1 985. Putative homozygous mutation in regenerated plants of rice. Mol. Gen. Genet. 198:377-384. Orton T.J. 1980. Chromosomal variability in tissue cultures and regenerated plants of Hordeum. Theor. Appl. Genet. 56:101-112. Orton T.J. 1983. Experimental approaches to the study of somaclonal variation. Plant Mol. Biol. Rep. 1 :67-76. Orton T.J. 1984. Somaclonal variation: theoretical and practical considerations. In J. P. Gustafson (ed.): "Gene Manipulation in Plant Improvement", pp. 427-469. Plenum, New York. Orton T.J. 1985. Genetic instability during embryogenic cloning of celery. Plant Cell Tissue Organ Cult. 4:159-169. Orton T.J. 1987. Genetic instability in celery tissue and cell cultures. Iowa State J. Res. 61 :481 -498. Ozias-Akins P. and I.K. Vasil. 1983a. Proliferation of and plant regeneration from the epiblast of Triticum aestivum (wheat; Gramineae) embryos. Amer. J. Bot. 70:1092-1097. Ozias-Akins P. and I.K. Vasil. 1983b. Improved efficiency and normalization of somatic embryogenesis in Triticum aestivum (wheat). Protoplasma 1 1 7:40-44. Palacios I.G. and J.L. Magoja. 1987. Relevant traits and heterosis of diploperennial teosinte-maize hybrids. Maize Genet. Coop. News Lett. 61 :66-67. Palacios I.G. and J.L. Magoja. 1988. Heterosis for dry matter and protein production per plant in diploperennial teosinte-maize hybrids. Maize Genet. Coop. News Lett. 62:81-82. Papenfuss J.M. and J.G. Carman. 1987. Enhanced regeneration from wheat callus cultures using dicamba and kinetin. Crop Sci. 27:588-593.

PAGE 331

324 Papes D., V. Garaj-Vrhovac, S. Jelaska and Kolevska-Pletikapic. 1983. Chromosome behaviour in cultured cell populations of higher plants. In P.E. Brandham and M.D. Bennett (eds.): "Kew Chromosome Conference II", pp. 155-163. George Allen and Unwin, London. Pareddy D.R. and J.F. Petolino. 1990. Somatic embryogenesis and plant regeneration from immature inflorescences of several elite inbreds of maize. Plant Sci. 67:211-219. Pence V.C. and J.L Caruso. 1984. Effects of IAA conjugates on morphogenesis and callus growth from tomato leaf discs. Plant Cell Tissue Organ Cult. 3:1 01 -1 1 0. Peschke V.M. and R.L Phillips. 1991. Activation of the maize transposable element Suppressor-mutator (Spm) in tissue culture. Theor. Appl. Genet. 81 :90-97. Peschke V.M., R.L. Phillips and B.G. Gengenbach. 1986. Transposable element activity in progeny of regenerated maize plants. In D.A. Somers, B.G. Gengenbach, D.D. Biesboer, W.P. Hackett and C.E. Green (eds.): "Book of Abstracts. VI. Intl. Cong. Plant Tissue and Cell Culture", p. 285. University of Minnesota, Minneapolis/St. Paul. Peschke V.M., R.L. Phillips and B.G. Gengenbach. 1987. Discovery of transposable element activity among progeny of tissue culture derived maize plants. Science 238:804-807. Pescitelli S.M., J.C. Mitchell, A.M. Jones, D.R. Pareddy and J.F. Petolino. 1989. High frequency androgenesis from isolated microspores of maize. Plant Cell Rep. 7:673-676. Petolino J.F. and A.M. Jones. 1986. Anther culture of elite genotypes of maize. Crop Sci. 26:1072-1074. Phillips G.C. and J.F. Hubstenberger. 1987. Plant regeneration in vitro of selected Allium species and interspecific hybrids. HortScience 22:124-125. Phillips G.C. and K.J. Luteyn. 1983. Effects of picloram and other auxins on onion tissue cultures. J. Amer. Soc. Hort. Sci. 108:948-953. Phillips R.L, S.M. Kaeppler and V.M. Peschke. 1990. Do we understand somaclonal variation?. In H.J.J. Nijkamp, L.H.W. van der Plas and J. van Aartrijk (eds.): "Progress in Plant Cellular and Molecular Biology", pp. 131-141. Kluwer Academic Publishers, Dordrecht, The Netherlands. Pijnacker L.P., K. Sree Ramulu, P. Dijkhuis and M.A. Ferwerda. 1989. Flow cytometric and karyological analysis of polysomaty and polyploidization during callus formation from leaf segments of various potato genotypes. Theor. Appl. Genet. 77:102-110. Pinfield N.J., J.O. Sanchez-Torres and C.N. McDermott. 1984. The modifying effect of gibberellic acid and kinetin on the growth abnormalities induced by 2,4-D in marrow seedlings and the possible involvement of endogenous ethylene. Plant Growth Regul. 2:99-109.

PAGE 332

325 Pischedda G. and J.L. Magoja. 1987. Potential use of diploperennial teosinte germplasm for maize improvement. Maize Genet. Coop. News Lett. 61 :65. Pischedda G. and J.L. Magoja. 1988. Diploperennial teosinte introgressed population of maize: variation within Si derived lines. Maize Genet. Coop. News Lett. 62:84. Pischedda G. and J.L Magoja. 1990. More about maize introgressed with diploperennial teosinte germplasm. Maize Genet. Coop. News Lett. 64:74. Pittenger T.H. and E.F. Frolik. 1950. Preparing slides for determining percentage of pollen abortion. Maize Genet. Coop. News Lett. 24:60-61 . Preil W. 1 986. In vitro propagation and breeding of ornamental plants: advantages and disadvantages of variability. In W. Horn, C.J. Jensen, W. Odenbach and 0. Schieder (eds.): "Genetic Manipulation in Plant Breeding", pp. 377-403. Walter de Gruyter, Berlin. Prioli L.M., W.J. Silva, P. Arruda and M.R. Sondahl. 1985. In vitro regeneration capacity of corn and teosinte genotypes. In R.R. Henke, K.W. Hughes, M.J. Constantin and A. Hollaender (eds.): 'Tissue Culture in Forestry and Agriculture", pp. 342343. Plenum Press, New York. Prioli L.M., W.J. Silva and M.R. Sondahl. 1984. Tissue culture and plant regeneration in diploid perennial teosinte. J. Plant Physiol. 117:185-190. Qureshi J.A., P. Hucl and K.K. Kartha. 1992. Is somaclonal variation a reliable tool for spring wheat improvement? Euphytica 60:221 -228. Radojevic L. 1985. Tissue culture of maize Zea mays «Cudu». I. Somatic embryogenesis in callus tissue. J. Plant Physiol. 1 1 9:435-441 . Rajasekaran K., M.B. Hein, G.C. Davis, M.G. Carnes and I.K. Vasil. 1987b. Endogenous plant growth regulators in leaves and tissue cultures of Napier grass (Pennisetum purpureum Schum.). J. Plant Physiol. 130:13-25. Rajasekaran K., M.B. Hein and I.K. Vasil. 1987a. Endogenous abscisic acid and indole3-acetic acid and somatic embryogenesis in cultured leaf explants of Pennisetum purpureum Schum.: effects in vivo and in vitro of glyphosate, fluridone and paclobutrazol. Plant Physiol. 84:47-51 . Rajasekaran K., S.C. Schank and I. K. Vasil. 1986. Characterization of biomass production, cytology and phenotypes of plants regenerated from embryogenic callus cultures of Pennisetum americanum x P. purpureum (hybrid triploid Napiergrass). Theor. Appl. Genet. 73:4-10. Rao I.U., I.V.R. Rao and V. Narang. 1985. Somatic embryogenesis and regeneration of plants in the bamboo Dendrocalamus strictus. Plant Cell Rep. 4:191-194. Rapela M.A. 1984. Indirect somatic (nonzygotic) embryogenesis in tissue cultures of maize. Maize Genet. Coop. News Lett. 58:1 10-1 12. Rapela M.A. 1985. Organogenesis and somatic embryogenesis in tissue cultures of Argentine maize (Zea mays L). J. Plant Physiol. 121 :1 19-122.

PAGE 333

326 Rappaport L. 1980. Plant growth hormones: internal control points. Bot. Gaz. 141:125130. Ray P.M. 1958. Destruction of auxin. Ann. Rev. Plant Physiol. 9:81-1 18. Redenbaugh K., D. Slade, P. Viss and J.A. Fujii. 1987. Encapsulation of somatic embryos in synthetic seed coats. HortScience 22:803-809. Redway F.A., V. Vasil, D. Lu and I.K. Vasil. 1990a. Identification of callus types for longterm maintenance and regeneration from commercial cultivars of wheat {Triticum aestivum L). Theor. Appl. Genet. 79:609-617. Redway F.A., V. Vasil and I.K. Vasil. 1990b. Characterization and regeneration of wheat (Triticum aestivum L.) embryogenic cell suspension cultures. Plant Cell Rep. 8:714-717. Rhodes C.A., C.E. Green and R.L Phillips. 1986. Factors affecting tissue culture initiation from maize tassels. Plant Sci. 46:225-232. Richey F.D. and G.F. Sprague. 1932. Some factors affecting the reversal of sex expression in the tassels of maize. Amer. Nat. 66:433-443. Ryan S.A. and W.R. Scowcroft. 1987. A somaclonal variant of wheat with additional /3-amylase isozymes. Theor. Appl. Genet. 73:459-464. Sacristan M.D. 1971. Karyotypic changes in callus cultures from haploid and diploid plants of Crepis capillaris (L.) Wallr. Chromosoma 33:273-283. Sacristan M.D. and G. Melchers. 1969. The karyological analysis of plants regenerated from tumorous and other callus cultures of tobacco. Mol. Gen. Genet. 105:317333. Sala F. and M.G. Biasini. 1987. Heritable variability induced by the in vitro culture and genetic improvement of cultivated plants. G. Bot. Ital. 120:43-54. Sallee P.J. 1982. Prefixation and staining of the somatic chromosomes of corn. In W.F. Sheridan (ed.): "Maize for Biological Research", p. 119. University Press, University of North Dakota, Grand Forks, North Dakota. Saunders J.W. and E.T. Bingham. 1975. Growth regulator effects on bud initiation in callus cultures of Medicago sativa. Am. J. Bot. 62:850-855. Schaeffer G.W. 1982. Recovery of heritable variability in anther derived doubledhaploid rice. Crop Sci. 22:1160-1164. Schaeffer G.W., FT. Sharpe Jr. and P.B. Cregan. 1984. Variation for improved protein and yield from rice anther culture. Theor. Appl. Genet. 67:383-389. Schaffner J.H. 1927. Control of sex reversal in the tassel of Indian corn. Bot. Gaz. 84:440-449. Scheunert E.U., Z.B. Shamina and H. Koblitz. 1978. Karyological features of barley callus tissues cultured in vitro. Biol. Plant. 20:305-308.

PAGE 334

327 Schobert B. 1977. Is there an osmotic regulatory mechanism in algae and higher plants? J. Theor. Biol. 68:17-26. Schuller A., G. Reuther and T. Geier. 1989. Somatic embryogenesis from seed explants of Abies alba. Plant Cell Tissue Organ Cult. 1 7:53-58. Scowcroft W.R. 1985. Somaclonal variation: the myth of clonal uniformity. In B. Hohn and E.S. Dennis (eds.): "Genetic Flux in Plants", pp. 217-245. Springer Verlag, New York. Seagull R.W. 1989. The plant cytoskeleton. CRC Crit. Rev. Plant Sci. 8:131-167. Seal H.L. 1964. Multivariate Statistical Analysis for Biologists. J. Wiley, New York. Sears R.G. and E.L. Deckard. 1982. Tissue culture variability in wheat: callus induction and plant regeneration. Crop Sci. 22:546-550. Sears R.G., A.C. Guenzi and B.S. Gill. 1984. Somaclonal variation in wheat. Agron. Abstr. 76:87. Semal J. 1986. Somaclonal Variations and Crop Improvement. Martinus Nijhoff, Dordrecht, The Netherlands. Sengupta J., S. Jha and S. Sen. 1988. Karyotype stability in long-term callus derived plants of Crepis tectorum L. Biol. Plant. 30:247-251 . Sequira L. and L. Mineo. 1966. Partial purification and kinetics of indoleacetic acid oxidase from tobacco roots. Plant Physiol. 41 : 1200-1 208. Shenoy V.B and I.K. Vasil. 1992. Biochemical and molecular analysis of plants derived from embryogenic tissue cultures of Napiergrass (Pennisetum purpureum K. Schum.). Theor. Appl. Genet. 83:947-955. Sheridan W.F. 1982. Anther culture of maize. In W.F. Sheridan (ed.): "Maize for Biological Research", pp. 389-396. University Press, University of North Dakota, Grand Forks, North Dakota. Singh B.D. and B.L. Harvey. 1975. Cytogenetic studies on Haplopappus gracilis cells cultured on agar and in liquid media. Cytologia 10:347-354. Singh R.J. 1986. Chromosomal variation in immature embryo derived calluses of barley (Hordeum vulgare L). Theor. Appl. Genet. 72:710-716. Skeene K.G.M. and M. Barlass. 1983. Studies on the fragmented shoot apex of grapevine. J. Exp. Bot. 34:1271-1280. Skirvin R.M. and J. Janick. 1976. Tissue culture-induced variation in scented Pelargonium spp. J. Amer. Soc. Hort. Sci. 101:281-290. Skoog F. and CO. Miller. 1957. Chemical regulation of growth and organ formation in plant tissues cultured in vitro. In Symp. Soc. Exp. Biol. XI. The biological action of growth substances, pp. 1 1 8-1 31 .

PAGE 335

328 Smith J.S.C., M.M. Goodman and R.N. Lester. 1981. Variation within teosinte. I. Numerical analysis of morphological data. Econ. Bot. 35:187-203. Smith R.L, M.K.U. Chowdhury and D.R. Pring. 1987. Mitochondrial DNA rearrangements in Pennisetum associated with reversion from cytoplasmic male sterility to fertility. Plant Mol. Biol. 9:277-286. Smith S.M. and H.E. Street. 1974. The decline of embryogenic potential as callus and suspension cultures of carrot (Daucus carota L.) are serially subcultured. Ann. Bot. 38:223-241 . Sokal R.R. and F.J. Rohlf. 1969. Biometry. W.H. Freeman, San Francisco. Sondahl M.R., D.A. Evans, LM. Prioli and W.J. Silva. 1984. Tissue culture regeneration of plants in lea diploperennis, a close relative of corn. Bio/Technology 2:455458. Songstad D.D., W.L Petersen and C.L. Armstrong. 1992. Establishment of friable embryogenic (Type II) callus from immature tassels of Zea mays (Poaceae). Amer. J. Bot. 79:761-764. Sozinov A., S. Lukyanyuk and S. Ignatova. 1981. Anther cultivation and induction of haploid plant in triticale. Z. Pflanzenzucht. 86:272-285. Springer W.D., C.E. Green and K.A. Kohn. 1979. A histological examination of tissue culture initiation from immature embryos of maize. Protoplasma 1 01 :269-281 . Steel R.G.D. and J.H. Torrie. 1980. Principles and Procedures of Statistics: A Biometrical Approach. McGraw-Hill, New York. Stewart C.R., C.J. Morris and J.F. Thompson. 1966. Changes in amino acid content of excised leaves during incubation. II. Role of sugar in the accumulation of proline in wilted leaves. Plant Physiol. 41 :1 585-1 590. Stimart D.P. 1986. Commercial micropropagation of florist flower crops. In R.H. Zimmerman, R.J. Griesbach, F.A. Hammerschlag and R.H. Lawson (eds.): 'Tissue Culture as a Plant Production System for Horticultural Crops", pp. 30131 6. Martinus Nijhoff, Dordrecht, The Netherlands. Stuart D.A. and S.G. Strickland. 1984a. Somatic embryogenesis from cell cultures of Medicago sativa L. I. The role of amino acid additions to the regeneration medium. Plant Sci. Lett. 34:165-174. Stuart D.A. and S.G. Strickland. 1984b. Somatic embryogenesis from cell cultures of Medicago sativa L. II. The interaction of amino acids with ammonium. Plant Sci. Lett. 34:175-181. Stuber C.W., J.F. Wendel, M.M. Goodman and J.S.C. Smith. 1988. Techniques and scoring procedures for starch gel electrophoresis of enzymes from maize (Zea mays L). North Carolina Agricultural Research Service Technical Bull. 286. North Carolina State University, Raleigh, North Carolina.

PAGE 336

329 Sun C.S., S.C.Wu, C.C. Wang and C.C. Chu. 1979. The deficiency of soluble proteins and plasmid ribosomal RNA in the albino pollen plantlets of rice. Theor. Appl. Genet. 55:193-197. Sun Z., C. Zhao, K. Zheng, X. Qi and Y. Fu. 1983. Somaclonal genetics of rice, Oryza sativa L. Theor. Appl. Genet. 67:67-73. Suprasanna P., K.V. Rao and G.M. Reddy. 1986. Plantlet regeneration from glume calli of maize (Zea mays L). Theor. Appl. Genet. 72:120-122. Sutter E. and R.W. Langhans. 1 981 . Abnormalities in chrysanthemum regenerated from long term cultures. Ann. Bot. 48:559-568. Swedlund B. and R.D. Locy. 1988. Somatic embryogenesis and plant regeneration in two-year old cultures of Zea diploperennis. Plant Cell Rep. 7:144-147. Swedlund B. and I.K. Vasil. 1985. Cytogenetic characterization of embryogenic callus and regenerated plants of Pennisetum americanum (L.) K. Schum. Theor. Appl. Genet. 69:575-581 . Syono K. 1965. Changes in organ forming capacity of carrot root callus during subculture. Plant Cell Physiol. 65:403-419. Taliaferro CM., S.M. Dabo, E.D. Mitchell, B.B. Johnson and B.D. Metzinger. 1989. Morphologic, cytogenetic, and enzymatic variation in tissue culture regenerated plants of apomictic old-world bluestem grasses (Bothriochloa sp.). Plant Cell Tissue Organ Cult. 19:257-266. Tang Y.W. and J. Bonner. 1947. The enzymatic inactivation of indoleacetic acid. I. Some characteristics of the enzyme contained in pea seedlings. Arch. Biochem. 13:11-25. Taniguchi K. and R. Tanaka. 1 989. Induction of shoot primordia from calli of Crepis capillaris (2n=6) and regeneration of plantlets. Japan. J. Genet. 64:199-208. Taylor M.G. and I.K. Vasil. 1987. Analysis of DNA size, content and cell cycle in leaves of Napier grass (Pennisetum purpureum Schum.). Theor. Appl. Genet. 74:681686. Thomas M.R. and K.J. Scott. 1985. Plant regeneration by somatic embryogenesis from callus initiated from immature embryos and immature inflorescences of Hordeum vulgare. J. Plant Physiol. 121:159-169. Thomas R.L. and J.J. Jen. 1980. Preparation of a homogeneous tomato fruit peroxidase. Prep. Biochem. 10:581-596. Thomas R.L., C.V. Morr and J.J. Jen. 1980. Oxidation of indoleacetic acid by a homogeneous tomato fruit peroxidase. J. Food Biochem. 4:235-246. Tomes D.T. 1986. Cell culture, somatic embryogenesis and plant regeneration in maize, rice, sorghum and millets. In S.W.J. Bright and M.G.K. Jones (eds.): "Cereal Tissue and Cell Culture", pp. 175-203. Martinus Nijhoff/W. Junk, Dordrecht, The Netherlands.

PAGE 337

330 Toncelli F., G. Martini, G. Giovinazzo and V. Nuti-Ronchi. 1985. Role of permanent dicentric systems in carrot somatic embryogenesis. Theor. Appl. Genet. 70:345-348. Torrey J.G. 1 967. Morphogenesis in relation to chromosomal constitution in long-term plant tissue cultures. Physiol. Plant. 20:265-275. Trigiano R.N. and B.V. Conger. 1987. Regulation of growth and somatic embryogenesis by proline and serine in suspension cultures of Dactylis glomerata. J. Plant Physiol. 130:49-55. Tulecke W., G. McGranahan and H. Ahmadi. 1988. Regeneration by somatic embryogenesis of triploid plants from endosperm of walnut, Juglans regia L. cv. Manregian. Plant Cell Rep. 7:301-304. Umbeck P.F. and B.G. Gengenbach. 1983. Reversion of male sterile T-cytoplasm maize to male fertility in tissue culture. Crop Sci. 23:585-588. Vallejos C.E. 1983. Enzyme activity staining. In S.D. Tanksley and T.J. Orton (eds.): "Isozymes in Plant Genetics and Breeding, Part A", pp. 469-516. Elsevier, Amsterdam. Valles M.P., Z.Y. Wang, P. Montavon, I. Potrykus and G. Spangenberg. 1993. Analysis of genetic stability of plants regenerated from suspension cultures and protoplasts of meadow fescue (Festuca pratensis Huds.). Plant Cell Rep. 12:101-106. Vasil I.K. 1982. Somatic embryogenesis and plant regeneration in cereals and grasses. In A. Fujiwara (ed.): "Plant Tissue Culture 1982", pp. 101-104. Maruzen, Tokyo. Vasil I.K. 1983a. Regeneration of plants from single cells of cereals and grasses. In P.F. Lurquin and A. Kleinhofs (eds.): "Genetic Engineering in Eukaryotes", pp. 233-252. Plenum Publishing Corporation, New York. Vasil I.K. 1983b. Toward the development of a single cell system for grasses. In "Cell and Tissue Culture Techniques for Cereal Crop Improvement", pp. 131-144. Science Press, Beijing. Vasil I.K. 1985. Somatic embryogenesis and its consequences in the Gramineae. In R.R. Henke, K.W. Hughes, M.P. Constantin and A. Hollaender (eds.): 'Tissue Culture in Forestry and Agriculture", pp. 31 -47. Plenum Press, New York. Vasil I.K. 1986. Cell Culture and Somatic Cell Genetics of Plants, Vol. 3: Plant Regeneration and Genetic Variability. Academic Press, New York. Vasil I.K. 1987. Developing cell and tissue culture systems for the improvement of cereal and grass crops. J. Plant Physiol. 128:193-218. Vasil I.K. 1988. Progress in the regeneration and genetic manipulation of cereal crops. Bio/Technology 6:397-402. Vasil I.K. 1990. The realities and challenges of plant biotechnology. Bio/Technology 8:296-301 .

PAGE 338

331 Vasil I.K. and V. Vasil. 1986. Regeneration in cereal and other grass species. In I.K. Vasil (ed.): "Cell Culture and Somatic Cell Genetics of Plants, Vol. 3: Plant Regeneration and Genetic Variability", pp. 121-150. Academic Press, New York. Vasil V., A.M. Castillo, M.E. Fromm and I.K. Vasil. 1992. Herbicide resistant fertile transgenic wheat plants obtained by microprojectile bombardment of regenerable embryogenic callus. Bio/Technology 10:667-674. Vasil V., C. Lu and I.K. Vasil. 1983. Proliferation and plant regeneration from the nodal region of Zea mays L. (maize, Gramineae) embryos. Amer. J. Bot. 70:951-954. Vasil V., C. Lu and I.K. Vasil. 1985. Histology of somatic embryogenesis in cultured immature embryos of maize (Zea mays L). Protoplasma 127:1-8. Vasil V., F. Redway and I.K. Vasil. 1990. Regeneration of plants from embryogenic suspension culture protoplasts of wheat (Triticum aestivum L). Bio/Technology 8:429-434. Vasil V. and I.K. Vasil. 1980. Isolation and culture of cereal protoplasts. II. Embryogenesis and plantlet formation from protoplasts of Pennisetum americanum. Theor. Appl. Genet. 56:97-99. Vasil V. and I.K. Vasil. 1981. Somatic embryogenesis and plant regeneration from tissue cultures of Pennisetum americanum and P. americanum x P. purpureum hybrid. Amer. J. Bot. 68:864-872. Vasil V. and I.K. Vasil. 1982. The ontogeny of somatic embryos of Pennisetum americanum (L.) K. Schum.: In cultured immature embryos. Bot. Gaz. 143:454465. Vasil V. and I.K. Vasil. 1984a. Induction and maintenance of embryogenic callus cultures in the Gramineae. In I.K. Vasil (ed.): "Cell Culture and Somatic Cell Genetics of Plants, Vol. 1 , Laboratory Procedures and Their Applications", pp. 36-42. Academic Press, New York. Vasil V. and I.K. Vasil. 1984b. Preparation of cultured tissues for scanning electron microscopy. In I.K. Vasil (ed.): "Cell Culture and Somatic Cell Genetics of Plants, Vol. 1 , Laboratory Procedures and Their Applications", pp. 738-743. Academic Press, New York. Vasil V. and I.K. Vasil. 1986. Plant regeneration from friable embryogenic callus and cell suspension cultures of Zea mays L. J. Plant Physiol. 124:399-408. Vasil V., I. K. Vasil and C. Lu. 1984. Somatic embryogenesis in long-term callus cultures of Zea mays L. (Gramineae). Amer. J. Bot. 71 :1 58-1 61 . Vian W.E. 1976. The effectiveness of picloram as an auxin source compared to 2,4-D in wheat callus culture medium. Agron. Abstr. 68:65. von Arnold S. 1 987. Improved efficiency of somatic embryogenesis in mature embryos oiPicea abies. J. Plant Physiol. 128:233-244. von Arnold S. and A. Wallin. 1988. Tissue culture methods for clonal propagation of forest trees. Intern. Assoc. Plant Tiss. Cult., Newslett. 56:2-13.

PAGE 339

332 Vyas L.N., S. Sharma and S.L. Jain. 1975. Reversal of the effect of 2,4-dichlorophenoxy acetic acid by gibberellic acid, kinetin and potassium nitrate. Biochem. Physiol. Pflanzen 167:361-365. Wagenkencht A.C. and R.H. Burris. 1950. Indoleacetic acid inactivating enzymes from bean roots and pea seedlings. Arch. Biochem. 25:30-53. Wang A.S. 1 987. Callus induction and plant regeneration from maize mature embryos. Plant Cell Rep. 6:360-362. Wellhausen E.J., LM. Roberts and E. Hernandez. 1952. Races of Maize in Mexico. Bussey Inst, of Harvard University, Cambridge, Massachusetts. Wenzler H. and F. Meins. 1986. Mapping regions of the maize leaf capable of proliferation in culture. Protoplasma 1 31 :1 03-1 05. Wernicke W. and R. Brettell. 1980. Somatic embryogenesis from Sorghum bicolor leaves. Nature 287: 1 38-1 39. Wernicke W., R. Brettell, T. Wakizuka and I. Potrykus. 1981. Adventitious embryoid and root formation from rice leaves. Z. Pflanzenphysiol. 103:361-365. Wernicke W. and L. Milkovits. 1984. Developmental gradients in wheat leaves Responses of leaf segments in different genotypes cultured in vitro. J. Plant Physiol. 115:49-58. Wernicke W., I. Potrykus and E. Thomas. 1982. Morphogenesis from cultured leaf tissue of Sorghum bicolor the morphogenetic pathways. Protoplasma 1 1 1 :53-62. Wersuhn G. and U. Dathe. 1983. Karyotypic variability of in vitro cultures of plants. Biol. Zentralbl. 102:551-557. Westerhof J., F.A. Hakkaart and J.M.A. Versluijs. 1984. Variation in two Begonia x hiemalis clones after in vitro propagation. Sci. Hortic. 24:67-74. Wetherell D.F. and D.K. Dougall. 1976. Sources of nitrogen supporting growth and embryogenesis in cultured wild carrot tissue. Physiol. Plant. 37:97-103. Wiersma D.W., E.S. Oplinger and S.O. Guy. 1986. Environment and cultivar effects on winter wheat response to ethephon plant growth regulator. Agron. J. 78:761764. Wiley E.O. 1981. Phylogenetics. The Theory and Practice of Phylogenetic Systematics. John Wiley and Sons, New York. Wilkes H.G. 1967. Teosinte: The Closest Relative of Maize. Bussey Inst, of Harvard University, Cambridge, Massachusetts. Wilkinson T.C. and S.A. Thompson. 1987. Genotype, medium and genotype x medium effects on the establishment of regenerable callus. Maydica 32:89-1 05.

PAGE 340

333 Yang S.F. and H.K. Pratt. 1978. The physiology of ethylene in wounded plant tissues. In G. Kahl (ed.): "Biochemistry of Wounded Plant Tissues", pp. 595-622. Walter de Gruyter, Berlin. Zagorska N.A., Z.B. Shamina and R.G. Butenko. 1974. The relationship of morphogenetic potency of tobacco tissue culture and its cytogenetic features. Biol. Plant. 16:262-274. Zehr B.E., M.E. Williams, D.R. Duncan and J.M. Widholm. 1987. Somaclonal variation in the progeny of plants regenerated from callus cultures of seven inbred lines of maize. Can. J. Bot. 65:491-499. Zheng K.L, S. Castiglione, M.G. Biasini, A. Biroli, C. Morandi and F. Sala. 1987. Nuclear DNA amplification in cultured cells of Oryza sativa L. Theor. Appl. Genet. 74:65-70. Zimmermann E.S. and P.E. Read. 1986. Micropropagation of Typha species. HortScience 21:1214-1216.

PAGE 341

BIOGRAPHICAL SKETCH Luis Filipe Galant Moreira Pedrosa was born on December 13, 1946 in Lisbon, Portugal. He completed with honors both elementary and high school education as Graduate Valedictorian, in Figueira da Foz, Portugal. He entered the University of Lisbon in 1965 and graduated with distinction in 1970 after earning a Bachelor of Science degree in Biology. He became Assistant Lecturer at the Botany Department of the University of Lisbon in 1 970, where he initiated his studies of tissue culture. His work was interrupted by his military service, from 1971 to 1974, which was spent at the National Hydrographic Institute in Lisbon doing research in biological oceanography and primary productivity of the Portuguese coastal waters. In 1974 he moved to Funchal, Madeira Island, to help create and oversee a research facility dedicated to support the local floriculture industry. His major role was associated with the micropropagation and greenhouse facilities. For the outstanding work produced during the five years that followed he was granted the Medal of Honor of the City of Funchal. After resuming his position as Assistant Lecturer at the University of Lisbon in 1 979, he then moved to Gainesville, Florida in 1 982 to begin his doctoral studies at the Botany Department of the University of Florida. He is the single father of Luis David Pedrosa, born in 1 984. 334

PAGE 342

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Indra K. Vasil, Chairman Graduate Research Professor of Botany I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. k I Di orris H. Williams, Cochairman Professor of Botany I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Thomas J. Sheehj Professor Emeritus of Horticultural Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 0, (J/Ja/c^. Henry C. Aldrich Professor of Microbiology and Cell Science

PAGE 343

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. This dissertation was submitted to the Graduate Faculty of the Department of Botany in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 1 993 William Louis Stern Professor of Botany Dean, Graduate School