Plant regeneration, cryopreservation, and genetic transformation of napiergrass (Pennisetum purpureum Schum)


Material Information

Plant regeneration, cryopreservation, and genetic transformation of napiergrass (Pennisetum purpureum Schum)
Physical Description:
xii, 175 leaves : ill. ; 29 cm.
Wan, Chun-Hua, 1963-
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1994.
Includes bibliographical references (leaves 149-174).
Statement of Responsibility:
by Chun-Hua Wan.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002028120
notis - AKL5731
oclc - 33032696
System ID:

Full Text







Chun-Hua Wan 1994

Dedicated to My Parents
Shu-Qing Wan and Wan-Fen Hu


I wish to thank Dr. Indra Vasil, chairman of my Ph.D. supervisory committee, for

his support and inspiration during the course of the work of this dissertation. Also,

thanks are extended to other members of the committee: Drs. R. Fed, C. Hannah, R.

Smith, and S. Schank.

I wish to acknowledge Dr. Vimla Vasil for her advice and encouragement

throughout the course of the research. Thanks are also due to other members of Dr.

Indra Vasil's laboratory for their discussion and help over the years of this project.

I wish to thank my family, both at home and in China, for their love and support

which has made the completion of this work possible.


ACKNOW LEDGMENTS ........................................................................................... iv


LIST OF TABLES ................................................................................................... viii

LIST OF FIGURES ................................................................................................... ix


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

REVIEW OF LITERATURE ...................................................................................... 2
Plant Regeneration of Grasses..................................................................... 2
Genetic Transformation of Grasses ............................................................ 22

MATERIALS AND METHODS ................................................................................ 46
Plant Regeneration ..................................................................................... 46
Cryopreservation......................................................................................... 54
Genetic Transformation............................................................................... 57

RESULTS ............................................................................................................ 62
Plant Regeneration from Leaf Cultures....................................................... 62
Plant Regeneration from Inflorescence Cultures......................................... 64
Plant Regeneration from Embryogenic Suspension Cultures...................... 72
Plant Regeneration from Cryopreserved Cell Cultures................................ 78
Plant Regeneration from Protoplast Cultures.............................................. 79
Genetic Transformation via Bombardment of Leaf and Inflorescence Calli.... 87
Genetic Transformation via Bombardment of Suspension Cultures.......... 106
Genetic Transformation via Electroporation of Protoplasts ....................... 110

DISCUSSION........................................................................................................ 139
Callus Cultures.......................................................................................... 139
Suspension, Cryopreserved Cell, and Protoplast Cultures........................ 142
Genetic Transformation............................................................................. 145

REFERENCES ..................................................................................................... 149

BIOGRAPHICAL SKETCH.................................................................................... 175



APH 3' II:

2,4-Dichlorophenoxy acetic acid
2,4,5-Trichlorophenoxy acetic acid
Abscisic acid
Alcohol dehydrogenase
S-aminoethyl L-cysteine dihydrodipicolinate synthase
Acetolactate synthase
Aminoglycoside 3'-phosphotransferase II
Large and compact cell clusters
Casein hydrolysate (acid, type I)
Chlorophenoxy acetic acid
Dimethyl sulfoxide
Deoxyribonucleic acid
Embryogenic callus
5-Enolpyruvyl-shikimate-3-phosphate synthase
Fresh weight
Glutamin synthetase
Hygromycin phosphotransferase
Indole-3-acetic acid
2[N-morpholino]ethanesulfonic acid
Milligram per liter
Naphthalene-1-acetic acid
Nopaline synthase
Neomycin phosphotransferase II
Phosphinothricin-N-acetyl transferase
Packed cell volume
Polyethylene glycol
Phosphoenol pyruvate
Plant growth regulators
Restriction fragment length polymorphism
Ribonucleic acid
Revolutions per minute

Note: Refer to Table 1 for abbreviations of nutrient media.

S: Segment
S3P: Shikimate-3-phosphate
SC: Small embryogenic cell dumps
SE: Standard error
SORB: Sorbitol
T-DNA: Transfer DNA
TTC: 2,3,5-Triphenyl tetrazolium chloride
V: Volts
VIR: Virulence
X-glu: 5-Bromo-4-choloro-3-indolyl-B-D-glucuronic acid


Table 1. Nutrient in media used in the present study........................................ 49

Table 2. Callus production in milligram from leaves of Lot 1 plants in 1989........ 65

Table 3. Embryogenic activity and induction frequency of somatic
embryogenesis of leaf explants from Lot 1 in 1989 ............................ 65

Table 4. Plant regeneration from call derived from leaf explants from Lot 1
in 1989 ....................................................................................... ........ 66

Table 5. Production of callus and induction of somatic embryogenesis from
young leaves of PI 300086 ................................................................. 67

Table 6. Summary of results on initiation of embryogenic callus cultures from
PP13 inflorescences ........................................................................... 68

Table 7. Summary of protoplast yields from L89, S89, 17, and cryopreserved
17 suspensions.................................................................................... 83

Table 8. Development of somatic embryogenesis after bombardment of leaf
explants .................................................................................... .......... 9 1

Table 9. Transient gene expression in leaf cultures ......................................... 92

Table 10. Inhibition of callus growth by Basta..................................................... 95

Table 11. Summary of plant survival under selection ......................................... 99

Table 12. Summary of stable transformation from bombardment of S89
suspensions..................................................................................... 111

Table 13. Summary of experiments on electroporation of protoplasts isolated
from cryopreserved suspensions...................................................... 137

Table 14. PEG treatment of protoplasts isolated from a cryopreserved
suspension three weeks after its reestablishment ............................ 138


Fig. 1. Schematic representation of pMON8678, pMON19606, pBARGUS,
and pAHC25 plasm ids ........................................................................... 58

Fig. 2. Size range of immature inflorescences used for culture ........................ 69

Fig. 3. Callus produced by an inflorescence segment four weeks after culture... 69

Fig. 4. Growth curves of C91 callus at 4-, 9-, and 14-day subculture intervals .... 73

Fig. 5. Necrosis (%) of C91 callus at 4-, 9-, and 14-day subculture intervals....... 73

Fig. 6. Plant regeneration from C91 callus at 4-, 9-, and 14-day subculture
intervals.......................................................................................... ....... 74

Fig. 7. Regrowth curves of cryopreserved S89 suspensions............................ 80

Fig. 8. TTC assay of cryopreserved S89 suspensions...................................... 80

Fig. 9. Plant regeneration from cryopreserved 17 suspensions......................... 81

Fig. 10. Protocallus formation from 17 protoplasts .............................................. 84

Fig. 11. Plant regeneration from 17 protocallus................................................... 88

Fig. 12. Growth of callus C91 on Basta-containing media.................................. 96

Fig. 13. PAT assay of plants (TO-14) from bombardment of callus................... 101

Fig. 14. Southern blot analysis of plants from bombardment of calli................. 103

Fig. 15. Natural tolerance of S89 suspension to Basta..................................... 105

Fig. 16. Natural tolerance of S89 suspension to Glyphosate............................ 105

Fig. 17. Southern blot analysis of Basta-resistant suspensions (B1-3) from
bom bardm ent....................................................................................... 112

Fig. 18. PAT assay of Basta-resistant suspensions (B1-3) from bombardment... 114

Fig. 19. Southern blot analysis of Glyphosate-resistant suspensions (G1-3)
from bom bardm ent (I)........................................................................... 116

Fig. 20. Southern blot analysis of Glyphosate-resistant suspensions (G1-3)
from bombardment (II).......................................................................... 118

Fig. 21. Southern blot analysis of Glyphosate-resistant suspensions (G1-3)
from bombardment (III)......................................................................... 120

Fig. 22. Southern blot analysis of Glyphosate-resistant suspensions (G1-3)
from bombardment (IV) ........................................................................ 122

Fig. 23. Protoplast culture after treatment with Basta....................................... 124

Fig. 24. Callus derived from Basta-treated protoplast cultures ......................... 124

Fig. 25. Culture of electroporated protoplasts in the presence of Basta........... 124

Fig. 26. Transient gene expression after electroporation of protoplasts........... 129

Fig. 27. Protoplast viability after electroporation............................................... 129

Fig. 28. Electroporation of protoplasts at different densities............................. 130

Fig. 29. PAT assay of Basta-resistant suspensions (P1-7) from
electroporation of protoplasts............................................................... 131

Fig. 30. Southern blot analysis of Basta-resistant suspensions (P1-7)
from electroporation of protoplasts....................................................... 133

Fig. 31. Plant regeneration from electroporated 17 protoplasts......................... 135

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



Chun-Hua Wan

August 1994

Chairman: Indra K. Vasil
Major Department: Plant Molecular and Cellular Biology

Plants of napiergrass or elephantgrass (Pennisetum purpureum Schum.) were

regenerated from embryogenic calli derived from leaf and inflorescence segments,
embryogenic suspensions, cryopreserved cell cultures, and protoplasts. Stable genetic

transformation was achieved following gene delivery via biolistic bombardment of

suspension cultures and electroporation of protoplasts. The induction of somatic

embryogenesis in segments of young leaves was found to be under precise

developmental control. This allowed a zone of cells competent for somatic

embryogenesis to be mapped along the length of young leaves. Somatic

embryogenesis was also initiated in segments of immature inflorescences. Factors

important for the development of embryogenic suspension cultures were investigated.

The starting callus tissues and gradual change of liquid medium were found to be

critical for establishing embryogenic suspension cultures, which provided a critical

source of protoplasts for the regeneration of normal mature plants.

Suspension cells were cryopreserved in liquid nitrogen after pregrowth in 6%

mannitol and freezing at 0.5C per minute to -40C in 0.5 M sorbitol and 5% DMSO.

Plants were regenerated and established in soil from cryopreserved cells after 2.5

years of cryogenic storage.

Optimal conditions for gene transfer were established on the basis of transient
expression of uidA, the gene that encodes fl-glucuronidase (GUS), following the

delivery of plasmid DNA by biolistics to callus and suspension cells, and by

electroporation to protoplasts. Two selectable marker genes were used for selection of

stable transformants: the bar gene of Stremptomyces hygroscopicus that encodes

phosphinothricin acetyl transferase (PAT) and confers resistance to the herbicide

Basta, and a maize mutant EPSPS gene that encodes a variant of 5-

enolpyruvylshikimate-3-phosphate synthase and confers resistance to the herbicide

Glyphosate. Integration of foreign DNA into the genome of napiergrass was confirmed

by Southern blot analysis and enzyme assays in three cell lines obtained by

bombardment of suspensions and electroporation of protoplasts. The genes, uidA

and/or nptll (a gene encoding neomycin phosphotransferase or NPT II) which were

cotransferred on the same plasmids with the selectable marker genes, were

cointegrated into the genome as confirmed by Southern blot analysis. However, GUS

activity was found only in one of the transformed cell lines. NPT II activity was not

detected in the transformed line containing the gene fragment. The level of PAT

activity remained unchanged in the transformed suspensions for one year under

selection. GUS activity was greatly reduced during one year maintenance on selection



The technology of gene transfer followed by plant regeneration from single
transformed somatic cells permits access to an unlimited gene pool for genetic
improvement of crops, and also provides opportunities to study structure, function, and
expression of genes. Efficient regeneration from callus, suspension, and protoplast
cultures is a prerequisite for obtaining transgenic plants. Accordingly, the objectives of
the present study were 1) to elucidate factors that control somatic embryogenesis, 2) to
investigate conditions necessary for the establishment of embryogenic suspension
cultures and for regeneration of mature plants from protoplasts, 3) to develop
techniques for long-term storage of embryogenic suspension cultures, and 4) to
attempt stable genetic transformation by protoplast electroporation as well as
microprojectile bombardment of napiergrass.
Napiergrass or elephantgrass (Pennisetum purpureum Schum.), a member of

the Gramineae, was selected as the experimental species because it is an important
renewable biomass energy crop (Smith 1984, Prine and Woodard 1993) and a quality
forage for ruminants (Sollenberger et al. 1988).


Plant Reaeneration of Grasses

The Gramineae include more than 10,000 species and all the cereal crops, such
as wheat, rice, maize, oats, barley, millet, sorghum, rye, and provide most of the world's
sugar, forage and fodder, as well as bamboos, canes, and reeds (Heywood 1993).
World agriculture has historically centered on domestication and improvement of
grasses for food and nutrition (Borlaug and Dowswell 1988).
Traditional methods for plant improvement in yield and quality rely on breeding,
selection, and clonal propagation. Such methods can be complemented and
supplemented by novel technologies of molecular and cellular biology-biotechnology
(Tanksley et al. 1989, Bech and Ulrich 1993, Visser and Jacobsen 1993, Vasil 1993).

The regeneration of plants from cell and tissue cultures is an essential part of
biotechnology. The earliest attempts to initiate and regenerate plants from cell cultures

of grasses relied largely on the experience with dicotyledonous species that most often

require combinations of auxins and cytokinins for optimal results (I. K. Vasil and Vasil
1986). This resulted in only transient and at best sporadic plant regeneration. Since
the early 1980's, recognition of factors such as the use of strong auxins and suitable
explants has produced a dramatic improvement in regeneration in this important group
of plants. These have led to the establishment of long-term regenerable callus, cell
suspension, and protoplast cultures, which have proven valuable for obtaining

transgenic plants of many important species such as wheat (V. Vasil et al. 1992), rice
(Toriyama et al. 1988), maize (Fromm et al. 1990), oats (Somers et al. 1992), barley

(Wan and Lemaux 1994), sorghum (Casas et al. 1993), sugarcane (Bower and Birch

1992), and rye (Castillo et al., personal communication), fescue (Wang et al. 1992),

and orchardgrass (Horn et al. 1988a).

Plant Regeneration from Callus Cultures

Callus culture is the most efficient method to regenerate plants of grasses. Two

pathways of plant regeneration have been described in literature: organogenesis, which

leads to formation of shoot primordia with open vascular systems, and embryogenesis,

which forms somatic embryos with closed vascular systems (I. K. Vasil and Vasil 1986,

Morrish et al. 1987). All publications before 1980 reported only organogenesis. In

most of these cases, plant regeneration was transient and caused by the derepression

of pre-formed shoot primordia present in the explants (Dunstan et al. 1979). Since

1980 when somatic embryogenesis was first convincingly demonstrated in the

Gramineae (Vasil and Vasil 1980, Dale 1980, Brettell et al. 1980), it has been found to

be the most common pathway of plant regeneration in vitro for this group of plants

(Vasil 1987).

Somatic embryoaenesis

Somatic embryogenesis has a significant advantage over organogenesis in that,

unlike adventitious shoot meristems formed in organogenesis, the somatic embryos,

like their zygotic counterparts, arise from single cells either directly or after the

formation of a mass of proembryogenic cells. Although a few studies have suggested

the possibility of multicellular origin (Wernicke et al. 1982), the single-cell origin of

somatic embryos has been confirmed in detailed histological studies of dicots

(McWilliam et al. 1974, Tisserat et al. 1979) and monocots (Vasil and Vasil 1982a, Ho

and Vasil 1983a, Conger et al. 1983, Magnusson and Bornman 1985, Lu and Vasil

1985) and is substantiated by the fact that no chimeras have been found in the

regenerants from embryogenic callus cultures (Morrish et al. 1987) and in populations

of transgenic plants (Gordon-Kamm et al. 1990, Vasil et al. 1993, Goto et al. 1993).
The use of plant growth regulators such as 2,4-D and culture of young and
immature tissues/organs such as leaves, inflorescences, and embryos readily allowed
the induction of embryogenic callus in most grass species (I. K. Vasil and Vasil 1986,

Morrish et al. 1987). Production of such a callus was rare and inefficient from mature
tissues (Nabors et al. 1983, Abe and Futsuhara 1985). Two types of call are
commonly produced in tissue culture of grasses: embryogenic and nonembryogenic.
The embryogenic callus originates from specific foci in the explants. In immature
embryos it is initiated from cells adjoining the procambial strand in the scutellum (Vasil
and Vasil 1982a, Lu and Vasil 1985), in young inflorescences from meristematic cells in
the floral primordia (Botti and Vasil 1984) or parenchymatous cells near the vascular
tissue, and in the basal portion of young leaves from cells of the lower epidermis and
mesophyll near vascular bundles (Haydu and Vasil 1981, Lu and Vasil 1981 a, Conger
et al. 1983, Ho and Vasil 1983a). It is not known how and why these cells selectively
express somatic embryogenesis. The two types of calli are morphologically
distinguishable. The nonembryogenic callus is soft, unorganized, and contains
sparsely-cytoplasmic, vacuolated, and large cells devoid of any prominent metabolic
reserves. The embryogenic callus is most often compact, organized, and white to pale

white in color, and contains large numbers of small, richly cytoplasmic, starch-

containing meristematic cells (Vasil and Vasil 1981 a, b, 1982, Ho and Vasil 1983a, Botti

and Vasil 1983, 1984, Lu and Vasil 1985, Karisson and Vasil 1986a). It is not
uncommon to see isolated sectors of embryogenic callus arising in the masses of
nonembryogenic callus. These sectors of embryogenic callus possibly originate from a
few original groups of embryogenesis-competent cells scattered amongst
nonembryogenic cells (Vasil 1987), rather than from the conversion of the

nonembryogenic cells (Nabors et al. 1983). Indeed, conversion of nonembryogenic

cells into embryogenic cells has never been convincingly demonstrated. The
embryogenic callus proliferates more slowly, especially during initial phases of growth,
than the nonembryogenic callus, requiring selective transfer of the embryogenic callus
at each subculture in order to maintain its embryogenic nature. In some species, a
friable and fast growing embryogenic callus type is produced either directly from
explants or after selection from the white compact embryogenic callus: Type II callus of
maize (Green 1982, Armstrong and Green 1985, Vasil et al. 1984a, 1985, Tomes and
Smith 1985, V. Vasil and Vasil 1986, Shillito et al. 1989, Morocz et al. 1990), oat
(Bregitzer et al. 1989), and sorghum (Wei and Xu 1990), and type C callus of wheat
(Redway et al. 1990a) and teosinte (Zea diploperennis) (Pedrosa 1994).
Growth and development of embryogenic cells is greatly Influenced by the
nature and amount of plant growth regulators, particularly 2,4-D, present in the nutrient
medium (Vasil 1987). At low levels of 2,4-D (1-2 mg/L), the meristematic activity of
competent cells is maintained and stimulated, resulting in the formation of embryogenic
callus. Increased concentrations (e.g., 5 mg/L or higher) are frequently toxic (V. Vasil
and Vasil 1986). Reducing the level of 2,4-D in the medium to about 0.1 mg/L allows
formation of many somatic embryos, but also causes the irreversible conversion of
some embryogenic cells to enlarged and vacuolated nonembryogenic cells that may
serve as nurse cells (Vasil and Vasil 1982b, Karisson and Vasil 1986a). Such a
dynamic feature of growth and development of the embryogenic callus permits
manipulation of somatic embryogenesis in vitro. For instance, Morocz et al. (1990)
used a 3-4 week subculture interval to maintain freshly isolated embryogenic callus
cultures, which contained embryoids at all stages of development (from green and/or
whitish scutellar-like structures to small shoots and roots). However, when the callus
was subcultured more frequently (every 10-15 days) the resulting cultures became
highly friable and homogenous with only early developmental stages of embryoids.

Factors Controlling Induction of Somatic Embryogenesis
Besides 2,4-D-like growth regulators and immature tissue/organs as explants
(see the previous section), the genotype, developmental stage of explants, and
environmental influence of donor plants are some of the other important factors
regulating the initiation of embryogenic callus.
Genotypic variation in culture response is a well-known phenomenon. Genetic
control of plant regeneration from tissue cultures has been indicated in various studies.
Willman et al. (1989) suggested that expression of somatic embryogenesis in maize

tissue cultures was under control of a single gene or a block of genes. Tomes and
Smith (1985) and Hodges et al. (1986) reported that a few nuclear genes were
responsible for the control of maize regeneration. Peng and Hodges (1989) presented
evidence that regeneration ability of rice tissue cultures was under control of both
nuclear and cytoplasmic genes. Ma et al. (1987) found that the ability to form
regenerable callus of sorghum was inherited as a dominant trait among different
genotypes. Mathias and Fukui (1986) found that the 4B chromosome of one of the four

wheat varieties studied had a significant effect on in vitro morphogenesis. Rakoczy-
Trojanowska and Malepszy (1993) concluded that the ability to regenerate plants from
inflorescence cultures of rye was controlled by numerous loci and had a recessive
character. These reports indicate the possibility of improving tissue culture response by
plant breeding (e.g., Morocz et al. 1990). However, it does not explain the fact that

genotypes once considered nonregenerable (e.g., maize, Green and Phillips 1975; rice,
Kyozuka et al. 1988) were later found regenerable (Duncan et al. 1985, Biswas and
Zapata 1992).
The use of explants at defined stages of development has been identified as a
critical factor in establishing totipotent cultures from embryos (Vasil and Vasil 1981a, Lu
et al. 1983, 1984), inflorescences (Rangan and Vasil 1983, Botti and Vasil 1984, Boyes

and Vasil 1984), and leaves (Wemicke and Brettell 1980, Lu and Vasil 1981a, Haydu
and Vasil 1981, Conger et al. 1983, Ho and Vasil 1983a, Wemicke and Milkovits 1984,
Wenzler and Meins 1986). Response of explants from well-nourished plants was
different in culture from those of nutrient-deficient plants (Duncan et al. 1985). Cultures
from plants grown in summer were known in some cases to behave differently than
those collected from plants grown in the cool season (Rines and McCoy 1981, Hanzel
et al. 1985, Ma et al. 1987). Many genetic differences in culture response could be
circumvented by varying nutrients and growth regulators in the culture medium (Duncan
et al. 1985).
The difference in tissue culture response observed among genotypes, at
various stages of development, and donor plants from different environmental
conditions has been suggested to be a physiological phenomenon caused by varied
hormonal status of the explants (Vasil 1987, Bhaskaran and Smith 1990). This
viewpoint is supported by several lines of evidence. In tissue culture of young leaves of
napiergrass, Rajasekaran et al. (1987b) showed that the younger parts of leaves, which
are competent for somatic embryogenesis, contain higher levels of endogenous
indoleacetic acid (IAA) and abscisic acid (ABA), while the more mature parts of the
leaves, which do not form embryogenic calli, contain a substantially lower level of
endogenous IAA and ABA. The relationship of hormonal metabolism to competence
for somatic embryogenesis was further demonstrated by Rajasekaran et al. (1987c)
who found that treatment of plants with fluridone, an inhibitor of ABA biosynthesis,
inhibited somatic embryogenesis in leaf cultures. Such inhibition could be overcome by
exogenous ABA in the culture medium. The varying degrees of embryogenic
competence along the length of napiergrass leaves were not found to be associated
with changes in cell cycle or DNA content (Taylor and Vasil 1987), and levels of DNA
methylation (Morrish and Vasil 1989). Dolezelova et al. (1992) also reported no
variations in the cell cycle and nuclear DNA content among maize tissues differing in

embryogenic competence. These findings enhance the concept that expression of
somatic embryogenic competence is largely a physiological phenomenon (I. K. Vasil
and Vasil 1986, Vasil 1987, Bhaskaran and Smith 1990). It is inferred, therefore, that
all genotypes are capable of producing totipotent cultures, and that the correct
meristematic explants (i.e., right developmental stage from right donor plants) plus
initial exposure to correct in vitro culture conditions (i.e., nutrients, growth regulators,

incubation methods) are critical.

Long-term Cultures of Embrvoaenic Calli
Successful induction of somatic embryogenesis in vitro does not always result in
long-term embryogenic cultures. Lu et al. (1982,1983) reported that it was not possible

to maintain the white compact embryogenic callus (type I) obtained in immature maize
embryo cultures for more than three subcultures (9-10 weeks) due to rapid and
extensive differentiation of somatic embryos. Redway et al. (1990a) found that the
white compact callus produced in cultures of wheat immature embryos was also difficult
to subculture as it became brown and ceased to grow on maintenance medium. In leaf
culture of napiergrass, Chandler and Vasil (1984) were able to maintain embryogenic

cultures for over 6 months by selection and transfer of only the white compact callus;
however, for each subculture 15-50% of callus pieces became necrotic. In general, the

potential of long-term culture of the compact embryogenic callus is limited. The friable
embryogenic callus (type II), however, can be maintained in culture for long periods of
time. Friable callus is produced at very low frequency and is described only in maize
(Green 1982, Armstrong and Green 1985, Vasil et al. 1984a, 1985, Tomes and Smith
1985, V. Vasil and Vasil 1986, Kamo and Hodges 1986, Shillito et al. 1989, Prioli and
Sondahl 1989), oat (Bregitzer et al. 1989), wheat (Redway et al. 1990a), and sorghum
(Wei and Xu 1990). Shillito et al. (1989) reported in maize that only two out of one

thousand immature embryos plated in culture produced the friable callus. Bregitzer et

al. (1989) found in oat that only 10% of the embryogenic call that survived 30 weeks of
subculture became uniformly friable. Redway et al. (1990a) found that in wheat the
frequency was as low as 0.5% of plated embryos producing the friable callus.
Armstrong and Green (1985) and Bregitzer et al. (1989) reported in maize and

oat, respectively, that formation of friable callus could be induced directly from cultured
immature embryos. Among the factors important for the induction were nutrients (e. g.,
L-proline, N6 or MS basal salts) in culture medium, developmental stages of embryos,
and genotypes. For instance, Armstrong and Green (1985) found in maize that N6
medium (Chu et al. 1975) was superior to MS basal salts (Murashige and Skoog 1962).
Also, addition of L-proline to the medium and use of the proper developmental stage of
embryos were important for direct induction of friable callus. However, the effect of N6

salts and L-proline was not found beneficial by others in maize (V. Vasil and Vasil
1986) and the presence of L-proline was toxic to oat friable embryogenic cultures
(Bregitzer et al. 1989). In addition, Bregitzer et al. (1989) reported that in one of the
three oat cultivars tested, the friable callus that was produced directly from embryo
cultures could not be maintained in culture beyond 30 weeks after its initiation.
The friable callus more often developed from the compact callus over a long

period of time by selective subculture. Vasil et al. (1984a) reported that the addition of

casein hydrolysate and reduction of sucrose to 2% in the maintenance medium caused

a gradual morphological transformation during subculture of the opaque and white
compact callus produced from cultured immature maize embryos. Although the
compact callus grew rather slowly and often turned brown upon subculture, they found
that about 25-30% of the callus pieces became increasingly less opaque and white and
appeared soft and somewhat transparent This soft callus culture maintained its
capacity for plant regeneration over a year. Kamo and Hodges (1986) found in maize
that the friable callus that was embryogenic and fast growing could only be identified 3
months after selective subculture. Because the friable callus developed from maize

inbred A188 declined in growth rate and turned brown 6-9 months after its initiation,
and the callus from inbred B73 and hybrid A188 x B73 grew well for over one year,
Kamo and Hodges (1986) suggested that B73 might contain genes that contribute to
long-term maintenance in culture. In wheat, two types of embryogenic call that slightly
differed in their morphology could be produced from cultures of immature embryos and
inflorescence (Redway et al. 1990a): Off-white, compact, and nodular embryogenic

callus and white compact embryogenic callus. The white compact callus turned brown
and ceased to grow on maintenance medium. The off-white compact callus became
less organized, less compact, and formed more soft nonembryogenic callus upon
transfer to maintenance medium. Therefore, it was necessary to select and transfer
only the more compact and organized pieces of callus at each subculture. The
inflorescence-derived callus became completely brown and necrotic after two
subcultures. The immature embryo-derived callus, however, could be subcultured for a
long period of time. After more than five months of subculture, a friable embryogenic
callus was obtained in a few cultures. After further subculture, a fast-growing culture
that consisted almost entirely of friable callus only was developed. Such a culture
maintained its capacity for plant regeneration for over 19 months. An interaction
between seasonal/environmental and genotypic factors in controlling the induction of

the friable embryogenic callus was also suggested in the study of Redway et al.
(1990a). During 1987 the friable callus was only observed in one of the six cultivars

tested and occurred at a very low frequency (0.5%). However, during 1988 the friable
callus was obtained in seven out of eight cultivars tested and at a much higher

frequency (approximately 10%). The development of the friable callus is also
described for oat Bregitzer et al. (1989) cultured immature embryos of three oat
cultivars. Embryogenic calli were selected and subcultured every other week. Initially
(the first 12 weeks) the call remained compact, but 18.2% became necrotic after
transfer. Continued selective subculture for a long period of time (over 30 weeks since

culture initiation) resulted in increased friability and less necrosis. Finally, an extremely

friable and uniformly embryogenic callus developed in about 10% of the selected
cultures in two of the three cultivars. The friable culture maintained its ability to
regenerate plants for over 78 weeks after its establishment.
It is unknown at present whether the friable callus can be initiated or selected in
other species of grasses. From the available information reviewed above, it is indicated
that the production of the friable callus in grasses is influenced by various factors,
including nutrition (Vasil et al. 1984a, Armstrong and Green 1985, V. Vasil and Vasil
1986, Kamo and Hodges 1986), developmental stage of explants (Armstrong and

Green 1985, Redway et al. 1990a), environmental influence of donor plants (Redway et
al. 1990a), and genotype (Kamo and Hodges 1986, Bregitzer et al. 1989, Redway et al.
1990a). Further study is needed to investigate these factors in other grass species.

Plant Regeneration from Embryogenic Cell Suspensions and Protoplasts
Embryogenic cell suspension cultures have been described as suspensions
which are finely dispersed, are free of any callus pieces, or organized tissues,
meristems and meristemoids, are fast growing, and are comprised mostly of groups of
small, richly cytoplasmic and starch-containing meristematic cells that form somatic
embryos upon decrease of the 2,4-D concentration (I. K. Vasil and Vasil 1986). The
importance of such cultures is in the fact that they are thus far the only reliable source
of totipotent protoplasts (Vasil and Vasil 1992) and can be used as target cells for

biolistic transformation (Gordon-Kamm et al. 1990, Fromm et al. 1990).
Before the concept of embryogenic suspension cultures in grasses was
developed in early 1980's (Vasil and Vasil 1980, 1981b, Lu et al. 1981a, b, Vasil et al.
1983), research on protoplasts of grasses followed the model approach used for dicot
mesophyll protoplasts. Since the first report of success in obtaining plants from
cultured mesophyll protoplasts of tobacco (Takebe et al. 1971), tens of thousands of

variations of nutrient media and other culture conditions were tried without success in

finding a suitable method for the culture of mesophyll protoplasts of the Gramineae
(Galston 1978, Potrykus 1980). This led to suggestions about the existence of a mitotic

block in grass protoplasts and speculation that they were constitutionally incapable of
mitotic division (Flores et al. 1981). However, sustained cell divisions could be
obtained in protoplasts isolated from cell cultures of many grass species including

barley (Koblitz 1976), rice (Cai et al. 1978), pearl millet (Vasll and Vasil 1979), and
maize (Potrykus et al. 1979). The totipotency of protoplasts in the Graminae was
demonstrated by Vasil and Vasil (1980) who cultured protoplasts of pearl millet
(Pennisetum amedcanum) isolated from a cell suspension culture that itself could form

somatic embryos and plants on regeneration media (Vasil and Vasil 1981b). The
concept of developing an embryogenic cell suspension for totipotent protoplast cultures
was further extended to Guineagrass (Panicum maximum) (Lu et al. 1981a, b) and

napiergrass (Vasil et al. 1983). Changes in nutrient media and culture conditions were

not critical. The single most important factor in the success of these experiments was
the use of embryogenic cell suspensions, which grow rapidly, for the isolation of

protoplasts. The failure of the protoplast-derived plantlets to grow in soil, which was

reported in the earlier studies, was later overcome by the use of highly embryogenic

cell suspensions and/or feeder-layers or agarose-embedding techniques for protoplast

cultures in cereals such as rice (Fujimura et al. 1985, Kyozuka et al. 1987), maize

(Rhodes et al. 1988, Shillito et al. 1989, Prioli et al. 1989), wheat (Vasil et al. 1990,

Qiao et al. 1992, Chang et al. 1991), sorghum (Wei and Xu 1990), barley (Jahne et al.
1991, 1992), and grasses such as sugarcane (Srinivasan and Vasil 1986, Chen et al.
1988), meadow fescue (Wang et al. 1993), tall fescue (Takamizo et al. 1990), creeping
bentgrass (Terakwa et al. 1992), orchardgrass (Dalton 1988), Agrostis alba (Asano and

Sugiura 1990), and Poa pratensis L. (Nielsen et al. 1993).

Establishment of embrvoaenic suspension cultures
The general method for establishment of embryogenic cell suspensions remains

the same, even today, as originally proposed (Vasil and Vasil 1980, Lu and Vasil

1981b, Lu et al. 1981b, Vasil and Vasil 1981b, 1982a, Ho and Vasil 1983, Vasil et al.

1983). It involves careful selection for small embryogenic cells at each subculture. It

applies well to the white compact type of embryogenic callus as starting material for

establishment of a embryogenic suspension (Vasil and Vasil 1984a). Dalton (1988)

and Yan et al. (1991) adapted the method for fescue and ryegrass, and barley,

respectively, by directly culturing embryos in liquid medium and hence reducing the

time required for establishing cell suspensions. In maize and wheat, the use of

homogenous fast growing, friable embryogenic callus has been found to be critical

(maize: Green 1983, V. Vasil and Vasil 1986, Rhodes et al. 1988, Shillito et al. 1989,

Prioli et al. 1989, Morocz et al. 1990; wheat Redway et al. 1990a, Vasil et al. 1990,

Chang et al. 1991, Qiao et al. 1992, Ahmed and Sagi 1993).

The time needed to develop a cell suspension varies. With the friable

embryogenic callus as the starting material, it can be from a few weeks to a few

months. For instance, in maize the time period from placing the friable callus in liquid

medium to an established stable embryogenic cell suspension was two weeks by

Morocz et al. (1990) and five weeks by V. Vasil and Vasil (1986). For wheat 2-4

months were required (Redway et al. 1990, Chang et al. 1991). With the compact

embryogenic callus, 2-9 months (an average of 4-5 months) is needed, as reported for

napiergrass (Vasil et al. 1983), orchardgrass (Horn et al. 1988b), rice (Wang et al.

1989), and barley (Jahne et al. 1991a). During this period, selective enrichment of the

cultures by smaller groups of small, round, richly cytoplasmic and starch-containing

cells is enforced at each subculture of 3-7 days. The enrichment is accomplished by

slowly draining off old liquid medium and replacing it with fresh medium at certain

dilution ratios at each subculture (Vasil and Vasil 1984b). The draining of the
supernatant medium removes most of the large, vacuolated, and nondividing cells
which are sloughed off into the medium along with other debris. However, in some
cases it was also possible to produce embryogenic cell suspensions by filtering with
sieves for certain sizes of cell groups and subculturing only the selected ones in fresh
liquid medium. For instance, with the friable embryogenic wheat callus, Qiao et al.
(1992) established embryogenic cell suspensions in two months after repeated
filtration. The filtrate containing cell dusters of 30-1000 pm diameter was subcultured
weekly in a 1:3 dilution with fresh medium. During the establishment of rice
embryogenic cell suspensions, Abdullhah et al. (1986) replaced liquid medium at
weekly intervals without reducing the cell density and sieved the suspensions through a
500 pm nylon mesh when necessary to remove large dumps of cells.
It is common to initiate a cell suspension by placing call in a liquid medium of
the same composition as the callus initiation or maintenance medium (Wei and Xu
1989) or placing them in Murashige and Skoog's (1962) medium supplemented with
2,4-D (1.0-2.5 mg/L), sucrose (3%), and inositol (100 mg/L) with or without 5-10%
coconut milk (Vasil and Vasil 1984b). However, some reports showed special nutrient

requirements for suspension cultures. Ozawa and Komamine (1989) found that rice
calli which were initiated and maintained on N6 medium or R-MS medium (a modified
MS medium) did not proliferate in liquid R-MS medium but grew vigorously when the R-

MS medium was supplemented with 300 mg/L casein hydrolysate (CH) or 300 mg/L CH
and 25 mM L-proline. By contrast, the calli grew vigorously in liquid N6 medium even in
the absence of proline and/or CH. Biswas and Zapata (1992) reported that seed-
derived rice calli initiated and maintained on MSCH2 medium (MS salts + 1 mg/L 2,4-D
+ 1.0 mg/L CH) could not grow in several liquid media, including liquid MSCH2. The
culture became either rooty or necrotic and eventually died. However, in a modified

liquid B5 medium (Gamborg et al. 1968), some cell dusters continued to grow without

root formation, but the growth rate of the cells was low and could not be further
improved. Transfer of these cell clusters to liquid MS1 medium (MS salts + 1.0 mg/L
2,4-D) resulted in vigorous growth with large aggregated cell groups. Occasionally,
small groups of round, densely cytoplasmic cells with thin wall appeared. Growth and
dissociation of the cellular aggregates was further improved when they were transferred
to liquid N6P2 medium (N6 salts +2 mg/L 2,4-D + 50 mM L-proline). After a few weeks
of weekly subculture, fast-growing cell lines consisting of highly cytoplasmic and small

groups of cells were obtained in the liquid N6P2 medium. In barley, Jahne et al.
(1991a) initiated anther callus on different solid media (based on MSm salts and L3
salts, both modified versions of MS). When calli derived from each of the media were
placed in a liquid medium, they all readily dispersed and started to grow immediately.
However, after 8 weeks of culture, suspension lines initiated with call from L3 media
continued to grow, while suspensions started with call from MSm media gradually
ceased growth. In addition, the genotype was also found to play a role in initiation of

their barley suspensions. The frequency of successful suspension initiation was
strongly dependent on barley genotypes, Igri, Princesse, Baronesse, and Gimpel. With
Igri, embryogenic cell suspensions could be relatively easily established compared to
the other genotypes. Cell lines of Igri could regenerate green plants in a longer period
of time, while cell lines of Princesse and Baronesse could regenerate only albinos and

this occurred for a short period of time (four months). Gimpel was the best genotype in

anther culture response with the highest yield of regenerated green plants, but for

suspension cultures it was inferior to Igri. These examples illustrate that each cell
culture system may have an unique requirement for adaptation of growth in liquid

Maintenance of embryogenic cell suspensions

After a long period (often 4-5 months) of careful selection and manipulation,

embryogenic cell suspensions that are finely dispersed and fast-growing can be
established in all major cereals and grasses (Vasil and Vasil 1992). These cell

suspensions, often referred to as cell lines, are generally maintained by subculture at a

1:5 or 1:7 dilution (suspension inocula: fresh medium) every 4-5 days in liquid media

containing 1.0-2.0 mg/L 2,4-D (Vasil and Vasil 1984b).

The established embryogenic cell suspensions consist predominantly of small
dusters (ca. 10-100 cells) of densely cytoplasmic cells, as originally described for pearl

millet (Vasil and Vasil 1982b) and later for orchardgrass (Horn et al. 1988b), maize
(Shillito et al. 1989, Prioli and Sondahl 1989), barley (Jahne et al. 1991a), and wheat

(Redway et al. 1990a, Chang et al. 1991, Qiao et al. 1992). Sometimes, these small

cell clusters tended to grow into larger aggregates and selective subculture and filtering

were used to maintain a high frequency of small aggregates of embryogenic cells (Prioli

and Sondahl 1989). The established cell suspension can contain other cell types too.

V. Vasil and Vasil (1986) found in maize that a proportion (10-20%) of the embryogenic

cells in each subculture always became nonembryogenic by rapid elongation, formation

of thick-walls, loss of starch and cessation of cell division activity. Rhodes et al.

(1988a) found that in their regenerable maize cell suspensions cell clusters associated

to form cell clumps as large as several thousand cells per aggregate and with irregular

shapes. Redway et al. (1990a) found that the wheat embryogenic cell suspension

contained 5-10% large, elongated and highly vacuolated nonembryogenic cells. In

most cases, no organized structures were observed in cell suspensions, However,

Horn et al. (1988b) considered the small clusters of densely cytoplasmic cells as
representing very early developmental stages of somatic embryos in their orchardgrass

cell suspension. Biswas and Zapata (1992) reported that in their established rice cell

suspensions, small globular embryos that developed at the surface of the aggregates

were continuously released into the medium as free-floating structures. The somatic
embryos up to the globular or the early scutellar stage in established cell suspensions
were also seen in other grass species (Vasil and Vasil 1981b, Lu and Vasil 1981b, Ho
and Vasil 1983b). Further development of these embryo structures did not occur in
liquid medium.
During maintenance of the cell suspensions, however, the embryogenic nature
of the cell lines is lost over time, often in a few months (Datta et al. 1990a, Jahne et al.
1991a) and sometimes in 1-2 years (Ozawa and Komamine 1989, Biswas and Zapata
1992, Funatsuki et al. 1992).

Plant regeneration of cell suspensions
Transfer of embryogenic cell suspensions to agar-media containing low
concentrations of 2,4-D is a general method for plant regeneration (V. Vasil and Vasil
1986). In some cell lines, special conditions such as alternate subculture intervals,
changes in the composition and concentration of salts and growth regulators are
required. For instance, Biswas and Zapata (1992) obtained only a few green shoots

when rice cells were transferred to low 2,4-D media. In comparing different media, they

found that although MS based media did not allow plant regeneration to occur, N6
based media supported shoot regeneration. The regeneration of shoots could be
further improved when the basal N6 medium was reduced to half strength and both
auxin (NAA) and cytokinin (BA) were included in the regeneration medium. These
results were in agreement with those of Ozawa and Komamine (1989), who also
reported that their rice cell suspensions could regenerate plants only when subcultured
every three days (not seven days).

Protoplast culture and regeneration

The availability of potent commercial cell wall-degrading enzymes in 1960s

made it possible for the first time to obtain large numbers of viable plant protoplasts

(Cocking 1960, Takebe et al. 1968). Leaves, calli, and nonmorphogenic cell

suspensions of grasses can readily release protoplasts upon enzyme digestion (Morrish

et al. 1987). However, protoplasts isolated from those sources either do not divide or

form only nonregenerable callus. Since the concept of embryogenic cell suspensions

for protoplast culture was developed, the embryogenic cell suspensions have been the

only source for totipotent protoplasts (Vasil and Vasil 1992). In maize, and in other

monocots in general, embryogenic callus cultures, no matter how friable, do not

normally yield totipotent protoplasts, although the protoplasts form callus (Imbrie-

Milligan and Hodges 1986, Shillito et al. 1989).

Culture of protoplasts often does not involve sophisticated manipulation of

media and culture conditions, if the donor cell suspensions are highly embryogenic and

well adapted to liquid culture conditions (Vasil 1987). Protoplast culture medium most

often comprises either KM-8p basal medium (Kao and Michayluk 1975) or its

modifications (Vasil and Vasil 1980, Jahne et al. 1991b) and low concentrations of 2,4-

D with or without other auxins and cytokinins. At a plating density of 0.5-5 x 106 the

first division of protoplasts takes place within 3-5 days of culture. In three weeks,

protocolonies visible to naked eyes are seen (Morrish et al. 1987). The protocolonies

and their further development in culture resemble the cell aggregates found in donor

suspension cultures. Development of somatic embryos and regeneration of plants can

be obtained using culture conditions identified for the donor cell suspensions. This

protocol for protoplast culture and regeneration is essentially the same as originally

used (Vasil and Vasil 1980, Lu et al. 1981, Vasil et al. 1983) and has produced mature

plants from protoplast cultures of all major cereals and grasses (Vasil and Vasil 1992).

In the past few years, the use of nurse cells (cells from a cell suspension) in
protoplast culture has been shown to be either crucial (Kyozuka et al. 1987) or

beneficial (Rhodes et al. 1988a, Shillito et al. 1989, Prioli and Sohndahl 1989, Jahne et
al. 1991b). Kyozuka et al. (1987) mixed rice protoplasts with molten agarose medium
and placed the solidified agarose blocks containing protoplasts into liquid medium that
contains cells from a rice cell suspension. This so-called agarose-bead method
increased the protoplast plating efficiency (the frequency of colony formation from total
protoplasts plated) from 0 to 4%. Shillito et al. (1989) used the same technique in
maize and improved plating efficiency from 0.003% to 0.054%. Rhodes et al. (1988a)
reported that the plating efficiency could reach 10%, representing a 100-fold increase,
when maize protoplasts were grown on filters directly over a feeder layer of nurse cells
compared to no nurse cells. In barley, Funatsuki et al. (1982) found that without
feeding, protoplasts formed few or no colonies, but with feeding plating efficiencies
were up to 10%.
At present, it is difficult to define the importance of feeder cells for protoplast
culture. For instance, in rice, maize, and wheat, feeder cells were shown to play an
essential role in recovering protocolonies in some culture systems (Kyozuka et al. 1987,

Rhodes et al. 1988a), but in others they were not needed for efficient colony formation
and plant regeneration (Fujimura et al. 1985, Abdullah et al. 1986, Prioli and Sondahl
1989, Vasil et al. 1990). Nevertheless, it is believed and, in some cases, demonstrated

that the plating efficiency, in general, can be improved via nurse cell culture (Jahne et
al. 1991b, Funatsuki et al. 1992).

Variability and Uniformity in Regenerated Plants
Genetic fidelity of plants regenerated in culture is a basic requirement for clonal
propagation and genetic transformation. However, variability, both epigenetic and

genetic, has been widely reported in plant tissue culture (Bayliss 1980, Larkin and

Scowcroft 1981, D'Amato 1985). In grasses, both relative uniformity (Chen et al. 1981,
McCoy and Phillips 1982. Hanna et al. 1984, Swedlund and Vasil 1985, Armstrong and
Green 1985, Karlsson and Vasil 1986b, Rajasekaran et al. 1987a, Shenoy and Vasil
1992, Pedrsa 1994, Chowdhury and Vasil 1993, Valles et al. 1993) and variability
(Heinz et al. 1977, Oono 1978, Orton 1980, McCoy et al. 1982, Karp and Maddock

1984, Lapitan et al. 1984, Larkin et al. 1984, Earle and Gracen 1985, Cooper et al.
1986, Gobel et al. 1986, Karp 1989) of cultured cells and/or regenerated plants have
been documented.
Detailed studies of cultured cells have revealed that embryogenic cell cultures
have much less variability than nonembryogenic and organogenic types of cell cultures
(Swedlund and Vasil 1985, Kartsson and Vasil 1986b). It Is believed that the
embryogenic cells physiologically and cytologically resemble cells of developing young
embryos and meristems and are considered to be in a continuing "embryogenic phase"
(1. K. Vasil and Vasil 1986). Physiological and developmental constraints impose strict
accuracy and bipolarity of mitosis on cells in the embryogenic phase. Such constraints
are diminished during cellular differentiation in the conversion of embryogenic to
nonembryogenic cells in vitro and therefore are responsible for the variability observed
in nonembryogenic cell cultures. The variability and stability of regenerated plants is

also correlated with the mode of plant regeneration. In all reports of variability,

organogenesis was involved in plant regeneration. In cultures in which both

organogenesis and embryogenesis occurred (Maddock et al. 1983, Karp and Mddock
1984), the observed variation was considered to result from plants regenerated via
organogenesis (Maddock 1985). Plants regenerated via somatic embryogenesis
exhibited no variability or low variability at morphological and/or cytological levels
(Hanna et al. 1984, Swedlund and Vasil 1985, Armstrong and Green 1985). Recent
surveys at biochemical and molecular levels also failed to detect any variation among
plants regenerated via embryogenesis. Shenoy and Vasil (1992) and Pedrosa (1994)

showed absence of variation in different isozymes in plants regenerated via somatic
embryogenesis of napiergrass and teosinte, respectively. By restriction fragment
length polymorphism (RFLP) analysis of mitochondrial, plastid, and nuclear genomes,
no variability could be found in plants from embryogenic cultures of napiergrass
(Shenoy and Vasil 1992), sugarcane (Chowdhury and Vasil 1993), meadow fescue
(Valles et al. 1993), and wheat (Chowdhury and Vasil 1994). The variations seen in
plants derived from somatic embryos were dearly transient in nature and were
considered to have resulted from epigenetic changes rather than genetic changes. For
instance, teosinte plants regenerated via somatic embryogenesis showed considerable
amounts of variation in morphology, including dwarfing and conversion of male flowers
in the tassel to females (Pedrosa 1994). Such variability disappeared completely in the
sexual progenies or after treatment of the abnormal plants with gibberellic acid. It is
believed that there is a strong selection in favor of normal cells during the process of

somatic embryogenesis with the result that only normal or nearly normal plants can be
regenerated (Swedlund and Vasil 1985, 1. K. Vasil and Vasil 1986).

Concluding Remarks
Long-term callus culture and plant regeneration from protoplasts have been
achieved in tissue culture of monocots. Identifying special types of embryogenic callus
is critical for the maintenance of the long-term cultures. Embryogenic cell suspensions

are the only reliable source of totipotent protoplasts. It is rare, difficult, and time
consuming to establish long-term callus culture (friable type of embryogenic callus) as
well as embryogenic suspensions. Cryogenic technology has been used for
cryopreservation of embryogenic cell cultures. Cell suspensions could be quickly
reestablished when the capacity of the original cell suspensions dropped in prolonged
culture (Shillito et al. 1989, Gordon-Kamm et al. 1990). The reestablished cell
suspensions were used to release protoplasts capable of regenerating fertile plants

(Shillito et al. 1989) or used for biolistic bombardment to produce fertile transgenic
plants (Gordon-Kamm et al. 1990). Therefore, coupled with cryogenic technology, cell
and tissue culture promises to provide a system for genetic manipulation of grasses.

Genetic Transformation of Grasses

The transformation of grasses, like any other group of plants, Is a complex
process involving gene delivery, gene expression, selection and regeneration of
transformed cells, and transmission of transgenes to progenies. Monocots, particularly
grasses, present a unique challenge in that they are not the natural host of
Agrobactedum and are very recalcitrant to cell culture techniques.

Methods of Gene Transfer
Aarobacterum and others. Agrobactedum is a soil-borne bacterium that can
cause crown gall and hairy root disease by a process in which part of the bacterial Ti

(tumor inducing) or Ri (root inducing) plasmid transfer DNA (T-DNA) is integrated into

the nuclear genome of plant cells (Ream 1989). With Ti plasmids, the transformed
plant cells express the inserted genes that encode enzymes involved in biosynthesis of
phytohormones and opines (amino acid derivatives). The result is tumor growth in
plants and supply of opine nutrients for the bacterium. A similar situation occurs with Ri
plasmids, resulting in hairy root growth. The Agrobacterium-mediated plant

transformation system takes advantage of the gene transferring ability of
Agrobacterium and has been used to transform over 30 plant species, all of which are

dicots (Corbin and Klee 1991, Rithie et al. 1993).
Most dicot plants produce specific phenolic compounds at wound sites. The
perception of these compounds by Agrobacterium results in the induction of a battery
of virulence (viW) gene operons of the plasmids. The products of the vir genes are then

engaged in the processing, mobilization, and transfer of T-DNA into plant cells (Ream
1989). Little is known, however, about the process of T-DNA integration into host
chromosomes (Corbin and Klee 1991).
Monocot species, however, do not usually form tumors in response to
Agrobactedum (Ream 1989, Corbin and Klee 1991). Nevertheless, over a dozen
monocot species, including grasses, have reportedly been transformed by
Agrobactedum (Hooykaas-Van Slogteren et al. 1984, Suseelan et al. 1987, Graves and

Goldman 1986, 1987, Feng et al. 1988, Conner et al. 1988, Dommisse et al. 1990,

Deng et al. 1990, Conner and Dommisse 1992). These reports based their claims on
the induction of swellings or tumors and/or detection of opines, which does not satisfy
the rigid definition of transformation that requires DNA analysis (Potrykus 1990). Other
researchers provided results of Southern blot analysis but failed to demonstrate gene
integration into high molecular weight DNA in their putatively transformed cell cultures
(Bytebier et al. 1987., Schafer et al. 1987, Rained et al. 1990, Kuehnle and Suggii
1991). Gould et al. (1991) even reported DNA hybridization in both putatively

transformed maize plants and their progenies, but evidence of hybridization to high
molecular weight DNA (i. e., undigested DNA in Southern blots) in their only one
"transformed" plant, as well as in its progeny, was lacking. Further study is needed in
order to understand what factors are important in order to achieve transgenic plants of
grasses by the Agrobactedum technology.
In the effort to transform monocots with Agrobacterium, the work of Grimsley et
al. (1987) was significant in that it provided evidence suggesting that Agrobactedum is
capable of transferring T-DNA into cells of monocot species. The techniques they
developed involved engineering partially or completely duplicated genomes of maize
streak virus into T-DNA of a A. tumefaciens strain, which was then used to inoculate
maize seedlings. DNA transfer was monitored by the appearance of viral symptoms on

the recipient plants. This method, called agroinfection, resulted in systemic spread of

the virus, suggesting T-DNA transfer to maize cells where a functional virus was
released, replicated, and spread systematically. The T-DNA transferring ability to
monocots was later demonstrated as being comparable to dicots (Grimsley 1989). No
evidence, however, of gene integration in the plant genome was obtained.
In exploring the method of gene transfer to monocots by Agrobactedum, several

techniques have recently been developed in order to avoid possible artifacts such as
expression of marker genes harbored on the T-DNA by Agrobacteium (Ritchie et al.
1993) and do novo biosynthesis of opines by nontransformed plant cells (Christou et al.

1986). Castle and Morris (1990) developed a system where a promoter such as
mannopine synthase was used to prevent expression of GUS fusion gene in bacteria.
Janssen and Gardner (1989) altered the ribosome binding site of the GUS fusion gene

from a prokaryotic to a consensus eukaryotic sequence to discourage translation of
GUS protein by Agrobactedum. Vancanneyt et al. (1990), taking advantage of the fact
that prokaryotes do not splice mRNA, introduced introns of plant genes into the GUS
fusion gene on T-DNA. Using these techniques, it has been recently found that
transient expression of genes delivered by Agrobactedum to maize is strongly
dependent on the vir genes of the Ti plasmid, plant genotypes (Ritchie et al. 1993,

Shen et al. 1993), and plant tissue types (Ritchie et al. 1993). These results suggest a

potential of Agrobactedum for transformation of monocots.

Potrykus (1990, 1991) discussed the limitations of other transformation
approaches, including microinjection, macroinjection, incubation of dry seeds or
embryos in DNA, pollen transformation, pollen tube pathway, electrophoresis, liposome
fusion, liposome injection, and microlaser. No indications of success have been
reported by any of these methods, which are therefore not discussed.

Protoplasts. Protoplasts, which are cells without cell walls, were obvious

target cells to demonstrate that Agrobactedum was not needed for gene transfer to

plants (Davey et al. 1980). With the knowledge of reversible membrane
permeabilization by chemical compounds such as polyethylene glycol (PEG) or poly-L-
omithine (PLO) or by electrical shocks, Draper et al. (1982), Krens et al. (1982), and
Fromm et al. (1985) demonstrated direct DNA transformation of protoplasts by PEG
and electroporation. After further refinement of the techniques (Shillito et al. 1985,

Negrutiu et al. 1987), both PEG and electroporation of protoplasts resulted later in
transgenic plants such as maize (Rhodes et al. 1988b, Golovkin et al. 1993), rice
(Peng et al. 1992, Toryama 1988), and forage grass fescue (Wang et al. 1992, Ha et
al. 1992).
The exact mechanism of how DNA molecules bypass the membrane and get
into the cell is not dear. It is believed that molecule diffusion across the membrane

takes place at the pores that are formed at certain regions of the membrane during
chemical/osmotic or electric shock of protoplasts (Van Wert and Saunders et al. 1992).
Recently, at least for electroporation, researchers have found evidence in support of an
active process-called electroosmosis as being the dominant mechanism for molecular
exchange (Sowers 1989, Dimitrov and Sowers 1990, Van Wert and Saunders 1992).
Nevertheless, for the best efficiency of gene delivery in protoplast transformation, a

given species or even a cell line requires modification of well established protocols
such as PEG molecular weight, electric field strength, and pulse duration (Vasil et al.
Although electroporation was initially used to deliver DNA across membranes in
the absence of cell walls (Fromm et al. 1985), recent evidence suggests that it may
transfer DNA into cells even in the presence of cell walls. Dekeyser et al. (1990)
reported that gene expression was obtained after electroporation of Intact leaf
segments of rice, maize, wheat, and barley. Abdul-Baki et al. (1990) demonstrated
uptake and expression of genes after electroporation of germinating tobacco pollen

grains. Even more recently, fertile transgenic maize plants were produced after

electroporation of partially digested embryos or finely chopped call (D'Halluin et al.

Biolistics. Biolistics, the term coined from biological ballistics, is the technique
which employs high velocity microprojectiles to deliver substances into cells and tissues
(Sanford et al. 1987, Klein et al. 1987, Sanford 1988). The concept of shooting
biological materials into living targets has long been used by plant pathologists. For
instance, MacKenzie et al. (1966) developed an air blast device to deliver viral particles
suspended in water directly to maize seedlings for mechanical transmission of virus.
For the same purpose, Dean (1960) even included silica powder in his viral suspension
for airblast of virus to sugarcane seedlings. In the biolistic process (Sanford et al.
1987, Klein et al. 1987), plasmid DNA is first coated by CaCI2 and spermidine
precipitation onto microprojectiles (tungsten or gold particles 0.4-2.0 ;i in diameter)
which are then accelerated to high velocity by gun powder discharge in order to
penetrate cell walls and cell membranes (Sanford 1988). Thousands of transient
transformation events, as monitored by the GUS blue spot assay, can be achieved by a
single shot and up to 10% conversion frequency from transient to stable transformation
has been recorded with tobacco (Russell et al. 1992).

Recently, many new versions of the biolistic device have been developed.
These include the Biolistic PDS-1000/He device (Kikkert 1993), the electric discharge
or ACCELLTM technology (McCabe and Christous 1993), the particle inflow gun (Vain
et al. 1993), the airgun device (Oard 1993), and the microtargeting apparatus (Sautter

Gene Expression
For the purpose of genetic transformation, the success of gene transfer
depends on successful gene expression. The elucidation of molecular mechanisms of

the crown gall disease at the beginning of 1980's encouraged researchers to try to
introduce genes of bacteria, animals, and plants into tobacco via Agrobacterium
(Herrera-Estrella et al. 1983). Gene expression failed because, as realized later,
transcription machineries of prokaryotes and eukaryotes are very different and It is the
5' flanking sequence (the promoter sequence) that is responsible for controlling when
and where the gene transcript is produced and the 3' flanking sequence that is
responsible for the transcript stops. As had been proven successful in animal cell
transformation (Colbere-Garapin et al. 1981), three research groups demonstrated in
1983 that bacterial genes, CAT and NPT II, could be expressed and functional protein
could be synthesized in transformed plant cells when the coding sequences were
placed under transcriptional control of nos promoter, which is a plant constitutive
promoter (Bevan et al. 1983, Fraley et al. 1983, Herrera-Estrella et al. 1983). The work
of Bevan et al. (1983) and Fraley et al. (1983) also showed that resistance to antibiotics
could be used for selection of transformants. This marked the beginning of plant
genetic engineering.
Although the nos and the cauliflower mosaic virus 35S promoters (Guilley et al.
1982) have been among the most widely used promoters, many attempts have been
made to search for a "stronger" promoter for monocots. These include isolation of
promoters of monocot genes such as AdhI (Dennis et al. 1984), Act1 (McElroy et al.
1990, 1991), Ubil (Christensen et al. 1992), use of introns (Callis et al. 1987, Vasil et
al. 1989), and synthesis of promoters that contain enhancer elements (Last et al.
1991). These "improved" monocot promoters were claimed better than the "standard"
35S promoter on the basis of transient gene expression, and therefore, expected to
have an enhanced effect on selection of stable transformants (McElroy et al. 1990,
1991, Last et al. 1991, Taylor et al. 1993). Although these monocot promoters have
yielded transgenic plants such as rice (Cao et al. 1992), sugarcane (Bower and Birch
1992), maize (Fromm et al. 1990), and wheat (Vasil et al. 1992, 1993, Weeks et al.

1993), it is not readily explainable why the frequency of stable transformation remains
the same for 35S promoter and Act1 and Ubil promoters (Cao et al. 1992, Vasil et al.

1993). Kuai et al. (1993) reported that in experiments on PEG-mediated transformation
of protoplasts isolated from fescue suspension cultures, transient gene expression was
greatly influenced by cell age (time after suspension subculture) whereas the frequency
of stable transformation was proportional to protoplast plating efficiency. This suggests
a possible difference in mechanisms which regulate transient expression and which
control stable expression.

Selectable Markers

Since the first introduction of selectable markers in plants (Bevan et al. 1983,
Fraley et al. 1983), the use of a substance toxic to plant cells and a resistance gene
that confers resistance to transformed cells has been an essential step in the
transformation process. This is because even with the most efficient system, the
efficiency of transformation, measured by transformed cells vs. the total cells targeted,
is very low. Without selection, it can be difficult and costly, if not impossible, to identify

a transgenic plant among the large population of the resulting regenerants (Christou

Among the selectable markers available at present, the choice for a particular
species under investigation is important Kanamycin was successful in selection for
maize transgenic plants (Rhodes et al. 1988b, Omirulleh et al. 1993). However, the
use of kanamycin prevented selected rice calli from regenerating plants while geneticin
(G418) did not interfere with plant regeneration and allowed recovery of transgenic rice
plants (Toriyama et al. 1988, Peng et al. 1990). Similarly, hygromycin was successfully
employed to produce transgenic plants such as maize (Walters et al., 1992), rice
(Shimamoto et al. 1989, U et al. 1992, Datta et al. 1992), and orchardgrass (Horn et al.

1988a), but hygromycin-selected wheat callus appeared to have lost its capacity for

plant regeneration while phosphinothricin-resistant callus regenerated transgenic wheat
plants (Vasil et al. 1993). At the level of transgenic plants, choice of the selection
marker system can also make a difference. For instance, all transgenic potato plants

selected with the AEC-DHPS system appeared morphologically normal (Per et al.
1993), but some of the transgenic tobacco plants produced in the same way displayed
morphological abnormalities (Shaul and Galili 1992). It is, therefore, advisable that for
each given species or even a particular genotype under study (Peri et al. 1993), a
particular selection system should be evaluated.

Antibiotics. Aminoglycoside antibiotics, including kanamycin, neomycin, G418
(a gentamycin derivative), paromomycin, and hygromycin, are the best known group of

antibiotics used for plant transformation. Their target of action is on the protein
synthesis machinery (Kors 1991, Wilmink and Dons 1993). Kanamycin, neomycin,
G418, and paromomycin bind to the small ribosomal subunit and therefore inhibit the
initiation of translation. Hygromycin binds to a ribosomal binding site of an elongation
factor and interferes with peptide chain elongation. The gene that confers resistance to
kanamycin, G418, neomycin, and paromomycin was cloned from Tn5 of E. coil (Bevan

et al. 1983). It codes for aminoglycoside 3'-phosphotransferase II (APH 3' II), also

called neomycin phosphotransferase II (NPT II). The enzyme phosphorylates a specific
hydroxyl group of the antibiotic, using the y-phosphate group of ATP, and therefore
abolishes the binding ability of the antibiotic to ribosomes. Hygromycin can be
inactvated in the same way by hygromycin phosphotransferase (HPT), which is
encoded by the hpt gene of E. coli (Blochinger and Diggelmann 1984). Although
monocot species display much higher natural tolerance to antibiotics (Potrykus et al.
1985, Hauptmann et al. 1988, Dekeyser et al. 1989) than dicot species (Fraley et al.
1983, Christou et al. 1988), transgenic plants such as rice (Toriyama et al. 1988) and
maize (Rhodes et al. 1988b) have, nevertheless, been selected with antibiotics.

Methotrexate. Methotrexate, a growth inhibitor, has been successfully used to
produce transformed cell lines of Panicum maximum (Hauptmann et al. 1988) and rice
(Meijer et al. 1991) and more recently transgenic plants of maize (Golovkin et al. 1993).
This compound inhibits the enzyme dihydrofolate reductase, blocking nucleotide
biosynthesis and resulting in cell death. A mutant mouse dihydrofolate reductase gene
(dhf,) encodes an enzyme with a reduced affinity for methotrexate and therefore
confers resistance (Eichholtz et al. 1987). Meijer et al. (1991) questioned the suitability
of methotrexate as the selective agent for transformation based on the complex
Southern hybridization patterns of their resistant rice lines and unusual CsCI banding of
their DNA isolates. But nevertheless, Golovkin et al. (1993) have obtained transmission
of the gene to transgenic progenies.

Herbicides. Herbicides have increasingly become popular as selective agents
for transformation of monocots. Among the different classes of herbicides, including
inhibitors of amino acid synthesis, lipid synthesis, photosystems I & II, mitosis, and
auxin-type inhibitors (Holt 1993), those that target acetolactate synthase (ALS), 5-
enolpyruvyl-shikimate-3-phosphate synthase (EPSPS), or glutamine synthetase (GS)
have been successfully used in producing transgenic plants such as rice (Li et al.
1992b), maize (Fromm et al. 1990, Gordon-Kamm et al. 1990), and wheat (Vasil et al.
Chiorsulfuron is the active ingredient in the herbicide Glean marketed by
DuPont Co. Like other members of sulfonylurea and members of imidazolinone and
triazolopyrimidine herbicide classes, chlorsulfuron specifically inhibits the bacterial and
plant ALS, the enzyme that catalyses the first reaction of biosynthesis of branched-
chain amino acids isoleucinee, leucine, and valine) (Holt 1993). The mutant
Arabidopsis ALS gene, csr 1-1, confers resistance to chlorsulfuron by encoding an ALS

with a reduced affinity for the herbicide (Haughn et al. 1988). Fertile transgenic maize

and rice plants selected in chlorsulfuron-containing media after transformation have

been obtained by using csr 1-1 in combination with the 35S promoter (Fromm et al.

1990, Li et al. 1992b)

Basta is a nonselective herbicide manufactured by Hoechst Co. that contains

20% L-phosphinothricin (also called PPT or glufosinate) as its active ingredient

Another PPT-based herbicide is bialaphos (L-phosphinothrinyl-L-alanyl-L-alanine) that

is synthesized by fermentation of Streptomyces hygroscopicus and sold commercially

by Meiji Seika Ltd. under the trade name Herbiace. The tri-peptide compound is

believed to be transported across the plasma membrane and rapidly releases L-

phosphinothricin due to peptidase cleavage (De Block et al. 1987, Padgette et al.

1989). Basta, PPT, and bialaphos have all been successfully used in selection for

transgenic plants such as maize (Gordon-Kamm et al. 1990, Fromm et al. 1990, Koziel

et al. 1993), rice (Christou et al. 1991, Cao et al. 1992), wheat (Vasil et al. 1992, 1993,

Weeks et al. 1993), and oat (Somers et al. 1992).

PPT is an analog of glutamate and acts as a competitive inhibitor of GS.

Because PPT can mimic partly the transition state of the reaction catalyzed by GS and

because of its tight binding to the enzyme, GS is irreversibly inhibited by PPT (Padgette

et al. 1989). In plants, glutamine synthetase plays a pivotal role in nitrogen metabolism

and ammonia assimilation. It is the only enzyme in plants that can detoxify ammonia

released by nitrogen reduction, amino acid catabolism, and respiration. Tachibana et

al. (1986a, b) have shown that in the treatment of PPT, it is the accumulation of

ammonia rather than the depletion of glutamine that causes the death of plant cells.
A mechanism of PPT detoxification was found in Streptomyces. Two genes

were cloned, bar from S. hygroscopicus (Murakami et al. 1986, Thompson et al. 1987)

and pat from S. viddochromogenes (Wohlleben et al. 1988). Both genes encode

phosphinothricin-N-acetyl transferase (PAT), which catalyzes the conversion of the

toxic L-phosphinothricin to the nontoxic N-acetyl-L-phosphinothricin in the presence of
acetyl CoA as a cofactor. Transformation with bar first resulted in transgenic plants of
tobacco (De Block et al. 1987) and later maize, rice, wheat and others (see the review
section in production of transgenic plants in grasses).
Glyphosate or N-phosphonomethylglycine is the active compound In the
herbicide RoundupTM produced by Monsanto Co. It specifically inhibits EPSPS in the
shikimate pathway, blocking biosynthesis of aromatic amino acids and resulting in cell
death (Comai et al. 1985). Because Glyphosate is mobile in the phloem and tends to
accumulate in the apex of stem and root, it affects especially meristematic and apical
cells of plants (Comal et al. 1989).
In the reaction catalyzed by EPSPS, shikimate-3-phosphate (S3P) and
phosphoenol pyruvate (PEP) are reversibly condensed to form 5-enolpyruvylshikimate-
3-phosphate (EPSP) and inorganic phosphate. This reaction takes place in plastids.
Glyphosate is a competitive inhibitor with respect to PEP but an uncompetitive inhibitor
with respect to S3P (Padgette et al. 1989). Although the mechanism for metabolic
inactivation of the herbicide has not been identified, resistance to Glyphosate has been
reported through EPSPS overproduction as well as mutation of EPSPS with a reduced
affinity to Glyphosate. In a Glyphosate-tolerant Petunia hybride cell culture,

Steinrucken et al. (1986) found EPSPS was overproduced approximately 20-fold and
this was due to a approximately 20-fold amplification of the EPSPS gene (Shah et al.
1986). Mutant EPSPS genes of bacterial and plant origin were also found to confer
resistance (Comai et al. 1983, Padgette et al 1989). Comal et al. (1985) and Fillatti et
al. (1987) introduced via Agrobactedum to tobacco and tomato, respectively, the
Salmonella typhimuium mutant EPSPS gene (am A) without a chloroplast transit
sequence and obtained transgenic plants with Glyphosate resistance. Padgette et al.
(1989) found the level of Glyphosate tolerance of transgenic tobacco and other plants
expressing mutant EPSPS genes with a chloroplast transit sequence to be clearly

superior to that of cytosol-targeted transgenic plants. The work of Padgette et al.
(1989) suggests that correct cellular targeting may be important In monocots, the use
of Glyphosate was only reported in the production of transformed wheat call (Vasil et
al. 1991). No transgenic monocot plants have yet been produced by using Glyphosate
as the selective agent
Other herbicides have been used as selective agents in reports on the
production of transgenic tobacco plants resistant to bromoxynil (Stalker et al. 1988),
atrazine (Cheung et al. 1988), and 2,4-D (Streber and Willmitzer 1989). Similar studies
have not been reported for monocot species.

Production of Transgenic Plants in Grasses
In the past four years, progress in transformation of grasses has been very
rapid. This is due to the increased number of species with established efficient
protoplast culture systems and the use of biolistic technology, the later in particular.
Although stably transformed ceil lines have been reported for a number of grasses
such as Panicum maximum (Hauptmann et al. 1988), sugarcane (Rathus and Birch

1992, Chowdhury and Vasil 1992), sorghum (Hagio et al. 1991), wheat (Vasil et al.
1991, Zhou et al. 1993), and barley (Lazzeri et al. 1991), the focus of this review is on
those reports in which transgenic plants and evidence for transmission of transgenes
are obtained.

Protoplast transformation
Protoplast transformation, or transformation via electroporation or PEG-
treatment of protoplasts, was the first direct DNA transformation method that resulted in
transgenic plants of monocots such as rice (Todyama et al. 1988, Zhang et al. 1988),
maize (Rhodes et al. 1988b, Golovkin et al. 1993), and orchardgrass (Horn et al.

1988a). It continues to be the method of choice for those species, such as rice, where

efficient protoplast regeneration systems are available (Goto et al. 1993).
Vasil et al. (1988) compared the difference between electroporation and PEG
treatment with protoplasts of Panicum maximum and found that although PEG-treated
protoplasts had a higher plating efficiency, electroporation resulted in slightly higher
transient gene expression. By electroporation of fescue protoplasts, Ha et al. (1992)
obtained 3-9 resistant cell lines per 106 protoplasts treated. Similarly, by PEG
treatment of fescue protoplasts, Wang et al. (1992) had 1-10 resistant cell lines per 106
protoplasts treated. Thus, the two methods, electroporation and PEG, do not appear to
differ significantly in the efficiency of stable transformation.

Among various research groups, the reported transformation efficiencies differ
in a range of 0-10 fold. Rhodes et al. (1988b) and U et al. (1992a) obtained a
frequency of 5% and 6% in maize and rice, respectively, while Shimamoto et al. (1989)
reported only 0.1-0.6% in rice. Similarly, Peng et al. (1992) obtained 8-26
transformants per 106 rice protoplasts treated, and Golovkin et al. (1993) and Wang et
al. (1992) observed 2-3 and 1-10 transformants per 106 protoplasts treated,
respectively, for maize and fescue. The importance of protoplast plating efficiency of a

cell line for transformation is not clearly demonstrated in these reports. The maize cell
line used by Golovkin et al. (1993) had a plating efficiency of 0.5-8.0% but the
transformation efficiency was 2-3 transformants per 106 protoplasts treated, while the
rice cell line used by Datta et al. (1992) had a plating efficiency of 0.5-18.0%, yet the

transformation efficiency was 8-26 transformants per 106 protoplasts treated.
All of the reports (over a dozen) on protoplast transformation used the 35S
promoter, except Omirulleh et al. (1993) who used a wheat a-amylase promoter with
enhancers. With this modified monocot promoter, Omirulleh et al. (1993) obtained 10-
20 resistant call per 106 protoplasts treated, a transformation efficiency still within the
range reported for the 35S promoter.

Antibiotics (kanamycin, G418, and hygromicin), methotrexate, and the herbicide
PPT were among the selectable markers used in protoplast transformation. In all

selection protocols reported, selective agents at effective concentrations (i.e., those
that were just enough to kill or severely inhibit growth of control cells) were added at
various stages of protoplast culture. U et al. (1992a) did not start selection until rice
protocalli were 0.5-1.0 mm diameter, but the individual calli were transferred for

selection at a concentration that was 4-5 times higher than that used by others to kill
nontransformed control rice cells that were at the multicellular stage of two weeks after
protoplast culture (Shimamoto et al. 1989, Datta et al. 1990). Although early
application (e.g., five days after DNA delivery to maize protoplasts, Omirulleh et al.
1993) of a lethal level of selection was effective in eliminating nontransformed escapes,
when the lethal level of selection was imposed to the culture later (e.g., 21 days after
maize protoplast transformation, Golovkin et al. 1993), the nontransformed call
somehow survived the selection. This may be a very common phenomenon because
cells at different stages may have different levels of tolerance to a certain concentration
of a selective agent Small multicellular colonies are more sensitive and are killed by a
lower concentration of a selective agent than the large colonies. Regenerated plantlets

may also have a different level of sensitivity to the agent compared to the callus from
which the plants were regenerated. Increase of the concentration of the selective

agent 4-5 fold from the effective level did not seem to impair the capacity of cultures for
plant regeneration but effectively eliminated the nontransformed escapes (e.g., rice,

Shimamoto et al. 1989, Datta et al. 1990, Peng et al. 1992, U et al. 1992). Therefore,
these data seem to be in favor of selection with higher concentrations. Using higher
concentrations of selective agents in protoplast transformation to eliminate escapes is
also supported in two other reports. In protoplast transformation of orchardgrass, some
transformed cell lines maintained under the lethal concentration of selection appeared

to be a mixture of transformed and nontransformed cells, because from at least one of

the transformed lines nontransformed escape plants could also be regenerated (Hom

et al. 1988a). In transformation of maize protoplasts, Golovkin et al. (1993) found that

the first set of protocalli grown out from the effective concentration of the selective

agent were likely to be transformed and that resistant cells from cultures under

prolonged selection were likely to be nontransformed. It is not uncommon that selected

calli were propagated under the same selection pressure before transfer to

regeneration media (Wang et al. 1992, Ha et al. 1992, Horn et al. 1988a, Toriyama et

al. 1988, Shimamoto et al. 1989, Datta et al. 1990). However, prolonged culture under

selection was also reported to impair plant regeneration (Datta et al. 1992). Therefore,

it remains to be further demonstrated that a short exposure of protoplast cultures to a

high selection pressure (e.g., 4-5 fold higher than the effective level for controls) can be

safe and effective.

As discussed earlier about PPT, its target of action is glutamine synthase and its

effect is to block the synthesis of glutamine. Some researchers removed glutamine

from protoplast culture media during selection at the expense of reduced plating
efficiencies (Datta et al. 1992, Wang et al. 1992). The presence of the amino acid in

the selection medium may not harm the purpose of selection by PPT, as demonstrated

in maize selection (Omirulleh et al. 1993). This result reinforced the findings by

Tachibana et al. (1986a, b) that it is the accumulation of ammonia in the cell, rather

than the lack of glutamine, that causes cell death by PPT.

Not all resistant lines remained regenerable after selection. U et al. (1992a)

reported only 30% of resistant lines were regenerable. Although U et al. (1992a)

reported full fertility for each transgenic plant, male sterile or even fully sterile

transgenic plants were not uncommon in the TO transgenic plants (Rhodes et al.

1988b). The male sterility problem could be solved by cross pollination using pollen

from wild type plants (Golovkin et al. 1993, Omirulleh et al. 1993, Peng et al. 1992).

Thus, it is advisable to germinate some seeds around the time when resistant call are
transferred for plant regeneration so that wild type pollen can be available.

In conclusion, DNA delivery to protoplasts by electroporation and PEG followed
by protoplast culture has been proven to be an effective way to introduce genes into
grass species. The requirement for an efficient regeneration system of protoplasts that

is associated with this technology, however, is the determining factor as to how
successful this transformation technology can be in grasses. Consequently, this
technology of transformation can not be fully exploited until our understanding of
protoplast culture is significantly advanced.

Biolistic transformation
Protoplast transformation outlined in previous sections demonstrated the

potential of direct DNA transfer (Toriyama et al. 1988, Zhang et al. 1988, Rhodes et al.
1988b, Horn et al. 1988a). However, the demand for an efficient regeneration system
for protoplasts has forced many people to seek alternatives. In the past three years,
the biolistic technology has been proven to be the premier alternative and is likely to be
the method of choice especially for very recalcitrant species like wheat Today
transgenic plants have been produced by bombardment in five monocot species and
many more species have been stably transformed. This review will focus on those

reports that produced transgenic plants.

The list of transgenic monocot plants produced with biolistics by October, 1993
includes twelve reports covering five species. The targets for bombardment include all
types of tissues young embryos of maize (Koziel et al. 1993), rice (Christou et al.
1991, U et al. 1993), and wheat (Weeks et al. 1993, Vasil et al. 1993); primary callus
directly produced from embryos of rice (U et al. 1993) and wheat (Vasil et al. 1993);
young calli that have been maintained for only 2-4 months in culture of rice (U et al.

1993) and wheat (Vasil et al. 1993); old call maintained over 12 months of maize

(Walters et al. 1992); a special type of callus of wheat (Vasil et al. 1992), and cell
suspensions of maize (Kordon-Kamm et al. 1990, Fromm et al. 1990), rice (Cao et al.
1992), and oat (Somers et al. 1992). It is possible therefore that all cell types,
especially those young cells capable of mitotic division, are competent for integrative
transformation. This is in contrast to the previous hypothesis that only some cells are
competent for integrative transformation while others have lost the cellular mechanism
for integration of foreign DNA (Portrykus 1990, 1991).

The most widely used promoter is the 35S promoter despite the fact that the
35S promoter produced much lower transient expression in monocots compared to
dicots (Hauptmann et al. 1987). Other monocot promoters such as Act1, Ubil and
Emu have also been successful. In a comparison of the 35S and Act1 promoters,
transient expression with the Act1 promoter was estimated 50-200 fold higher than that
with the 35S promoter in rice (McElroy et al. 1990). However, the 35S and Actl

promoters were equally effective in selection efficiency after bombardment (Cao et al.
1992). Similarly, the Ubil promoter was 10-45 fold more active than the 35S promoter
in transient assay of several monocot species (Christensen et al. 1992, Taylor et al.
1993). But, the Ubil promoter was not found superior to the 35S promoter in selection
for wheat transformants (Vasil et al. 1993). Therefore, "strong" promoters with high
transient gene expression have not been proven to be correlated to a high efficiency of
stable transformation.
Antibiotics and herbicides were used as selectable markers with PPT being the

most popular agent. As in protoplast transformation, some markers interfered with
plant regeneration (Vasil et al. 1993). Selection protocols vary among the reports, from
starting selection immediately after bombardment (Walters et al. 1992, Weeks et al.
1993) to starting 1-14 days after bombardment (Gordon-Kamm et al. 1990, Fromm et
al. 1990, Koziel et al. 1993, Bower and Birch 1992, Cao et al. 1992. U et al. 1993,
Somers et al. 1992, Vasil et al. 1992, 1993). Stepwise increase of selection pressure

(Walters et al. 1992, Fromm et al. 1990, Bower and Birch 1992, U et al. 1993, Vasil et

al. 1992, 1993) as well as a constant sublethal to lethal level of selection (Gordon-

Kamm et al. 1990, Fromm et al. 1990, Koziel et al. 1993, Cao et al. 1992, U et al. 1993,

Somers et al. 1992) have been used. Selection protocols should be designed so that
the rare transformed cells can proliferate independently of the rest of the cell population
and still maintain the ability to regenerate. Commonly, selection pressure has been
applied one to two weeks after DNA delivery. But Koziel et al. (1993) applied the full
selection pressure one day after bombardment and others started a low selection

pressure immediately after bombardment (Weeks et al. 1993, Walters et al. 1992).

Some researchers used full selection pressure from the start to the end of selection,
including during plant regeneration. Others used stepwise increase until resistant lines
were identified, then they were maintained at a sublethal to lethal level of selection. It
is likely that the development of a selection protocol for a particular culture system will

be a central part of the research in biolistic transformation, because no rules can be

generalized for all species, and it is vital that the capability of plant regeneration be
maintained throughout the selection process. Vasil et al. (1993) found in wheat that

after bombardment of embryos or callus, if using a high level of PPT to start selection,
there were more resistant lines that were not regenerable compared to those that were

from the selection protocol using a low starting concentration. Weeks et al. (1993)

found that with 1 mg/L bialaphos, wheat callus growth was inhibited 27% in three

weeks and 67% in six weeks. So it was six weeks before they could evaluate the effect

of a selection. Therefore, the transformed calli were maintained under this selection in
a chimeric stage with the nontransformed cells. Green sectors produced during
selection were then transferred for shooting and rooting under further selection for

transformed plants. Other systems exist which seem to be simpler. U et al. (1993)
reported that resistant rice call were evident after 2-3 weeks of selection at a sublethal
level and the removal of the resistant call from their explants often revealed brown

scars at the base of the resistant calli. Therefore, the resistant and nonresistant call
were dearly separated physically and physiologically. After further rounds of selection,
the majority of the surviving resistant call were found to be transformed. Selection
during plant regeneration further ensured recovery of only transformed plants. In all

instances, no chimeras were found in the resulting transgenic plants. Therefore,
microprojectiles are capable of delivering DNA into single embryogenic cells for nuclear
genomic integration. Furthermore, the single-cell origin of somatic embryogenesis is

The ratio of conversion from transient expression to stable incorporation of
microprojectile-delivered DNA in dicots ranges from around 1% in cotton (Finer and
McMullen 1990) to 5-10% in tobacco (Klein et al. 1988a, Russell et al. 1992). In maize
(Gordon-Kamm et al. 1990), the ratio was less than 0.1%: each filter of cells had at
least 1000 GUS fod and yet on the average, only one stable transformant was
recovered. In wheat (Weeks et al. 1993), the surface of each bombarded embryo was
covered with blue spots, and yet only 13 transformed cell lines were recovered from a
total of 6248 embryos bombarded.
In several studies where resistant calli were selected prior to plant regeneration,
it was shown that only about 30% of the selected calli were regenerable and an even

lower percentage of the calli were able to produce fertile plants (Gordon-Kamm et al.

1990, Vasil et al. 1992, 1993). As encountered in protoplast transformation, male
sterility of TO transgenic plants was not uncommon in biolistic transformation. (Gordon-
Kamm et al. 1990, Somers et al. 1992, Vasil et al. 1992, 1993).

Integration and Transmission of Transqenes
Foreign genes can be transferred to cells via various ways as discussed in

previous sections. However, little is known about how the introduced DNA molecules
are integrated into host chromosomes. Analysis of stably transformed cells, especially

transgenic plants, has revealed some general features of the integration events shared
by all direct DNA transformation methods.

1. Copy number per genome varies among transformation events.

One or a few copies of a gene per genome have been reported in protoplast
transformation of maize (Rhodes et al. 1988b, Golovkin et al. 1993), orchardgrass
(Hom et al. 1988a), and rice (Toriyama et al. 1988, Shimamoto et al. 1989, Peng et al.
1992, U et al. 1992a, Maijer et al. 1991, Goto et al. 19993), as well as in blolistic
transformation of maize (Gordon-Kamm et al. 1990, Spencer et al. 1992, Koziel et al.
1993), sugarcane (Bower and Birch 1992), and wheat (Weeks et al. 1993). However,
high copy number has also been reported. Meijer et al. (1991) estimated 50-100 gene
copies in transgenic rice plants transformed by PEG treatment of protoplasts;

Shimamoto et al. (1989) observed up to ten copies in their transgenic rice plants.
Similarly, in plants transformed by biolistics, wheat had up to 35 copies (Weeks et al.
1993) and maize 5-20 copies (Gordon-Kamm et al. 1990, Walters et al. 1992).
Hayashimoto et al. (1990) studied transgenic rice plants produced by PEG and

observed insertions of 2-5 copy concatemers at 2-4 loci in the genome. Similar
situations have been reported in dicot species. In transgenic tobacco plants produced

by protoplast electroporation, plasmid molecules appeared to form both head-to-head

and head-to-tail concatemers (Riggs and Bates 1986). Tandem repeats of T-DNA were
also reported in Agrobactedum-mediated transformation of tobacco (Kuhlemeier et al.
1987, Weishing 1988). Thus, high copy numbers can be caused at least in part by
concatemer formation prior to integration in various transformation procedures.
However, It is not known at present why sometimes a low copy number (1-2) was
produced and at other times high copy numbers occurred. Such a result of low vs. high
copy numbers had been seen in different transgenic lines produced even in the same
experiment (Gordon-Kamm et al. 1990).

2. Location and number of the integration sites per genome are random.
Spencer et al. (1992) reported mendellan inheritance of the transgene bar in
two of the four transgenic maize lines produced by biolistics. Two bands were detected
in these lines when Southern blots were probed with bar. However, only one band

cosegregated with PAT activity in a 1:1 ratio in R1 backcrossed progeny and in a 3:1
ratio in the selfed R2 progeny, indicating that two original integration sites occurred in
different chromosomes and only one of the two integration events produced a
functional copy of bar. Similar inheritance of transgenes was reported in biolistic
transformation of maize (Fromm et al. 1990, Walters et al. 1992, Koziel et al. 1993),
rice (Li et al. 1993), oat (Somers et al. 1992), and wheat (Vasil et al. 1992, 1993,
Weeks et al. 1993), and protoplast transformation of rice (Shimamoto et al. 1989, Datta
et al. 1990, Peng et al. 1992, U et al. 1992a, Meijer et al. 1991). In the case of
integration of multiple functional copies on different chromosomes, nonmendelian ratios
of inheritance would be expected, as reported by Shimamoto et al. (1989) for rice
protoplast transformation where segregation ratios larger than 3:1 were observed in
selfed progenies. Non-mendelian segregation can also be due to the site of insertion,
either through unstable integration that results in the loss of DNA inserts in all or some
of the gametes, or by affecting gamete viability in some manner. In one of the four
transgenic maize lines studied by Spencer et al. (1992), bar and GUS genes

cotransferred on separate plasmids were cointegrated into the same chromosome in

close linkage. One of the three GUS bands in RO plants was lost in the R1 progeny,
indicating an unstable integration site. Transmission of both bar and GUS bands from
the R1 to the R2 was impossible via male gametes, and only two of 34 R2 progeny
received the genes via female gametes in outcrossing, suggesting an event of poor
transmission. Similar results were reported by Walters et al (1992) for biolistic
transformation of maize. Loss of expression and lack of mendelian segregation of

transgenes was also reported in biolistic and Agrobactedum-mediated transformation of

dicotyledonous species (Tomes et al. 1990, Budar et al. 1986, Chyi et al. 1986,

Czemilofsky et al. 1986, Heberle-Bors et al. 1988)

3. Duplication of DNA inserts can occur.

Datta et al. (1990) showed that one of their transgenic lines obtained by PEG-

treatment of protoplasts produced ten selfed progenies, all of which were hygromycin

resistant and had a gene integration pattern identical to the parent, indicating that the

primary transformant was homozygous. They proposed that because the starting

material for protoplast isolation and transformation was a haploid microspore-derived

cell suspension, the recovered fertile primary transgenic plants could have been

homozygous diploid. Spencer et al. (1992) also recovered primary homozygous

transgenic plants of maize after biolistic transformation. The transgenic plants from

one of the transformed callus lines did not segregate in the R1 progeny for PPT

resistance and bar-hybridizing fragments. The possibility that the transformation was

nonnuclear (maternally inherited) was ruled out in analysis of R2 plants. The R1 plants

were outcrossed as male and female parents. The resulting R2 population had a 1:1

ratio of PPT resistance and sensitivity. The bar-hybridizing fragment cosegregated with

PPT resistance. The data therefore support the hypothesis that a single nuclear

integration event occurred in the transformation process and was replicated in the

callus, resulting in RO plants that were homozygous for the introduced DNA. A possible

mechanism proposed for the observation was gene conversion (Undegren 1953). In

the transformation event, integration of plasmid DNA and the replication event may

have been concomitant, both the result of the same nick in chromosomal DNA.

4. Cotransformation with genes harbored on different plasmids is possible.

Cotransformation with genes on separate plasmids has been demonstrated in
protoplast transformation (e.g., rice, Shimamoto et al. 1989, Peng et al. 1992, Meijer et
al. 1991, Fujimoto et al. 1993) as well as biolistic transformation (e.g., maize, Gordon-
Kamm et al. 1990, Walters et al. 1992, Koziel et al. 1993, and rice, U et al. 1993).
Given the fact that the offspring of transgenic plants such as maize (Gordon-Kamm et
al. 1990), rice (Goto et al. 1993), and wheat (Vasil et al. 1993) did not show
abnormalities, the success of the cotransformation would encourage the use of genes
of interest being cotransferred on a separate plasmid with a selectable marker gene,
which was thought to be able to integrate on a locus unlinked to that of the gene of
interest, therefore to allow removal of the marker gene after meiosis in the progeny.
Such a system of marker gene removal would present a great advantage for
transformation of open-pollinated species such as oats (Somers et al. 1992), for which
removal of marker genes such as those for herbicide resistance from transgenic plants
is necessary in order to prevent spread of the marker genes into the population of wild

weeds in the field. However, the advantage of the system has not yet been
demonstrated in grasses, because genes cotransferred on separate plasmids have
been reported to be very closely linked, as in cases of biolistic transformation of maize

(Spencer et al. 1992, Walters et al. 1992, Koziel et al. 1993) and protoplast
transformation of rice (Goto et al. 1993, Fujimoto 1993). How widespread the linkage is
between genes cotransferred by direct transformation in transgenic plants needs to be
further tested with a larger number of independent transformants. Studies in dicots
with Agrobactenum-mediated transformation, nevertheless, suggested that integration

of different T-DNAs could occur to the same locus (De Block and Debrouwer 1991) as
well as to unlinked loci (McKnight et al. 1987).
The frequency of coexpression of two genes on the same or separate plasmids
is very different In protoplast transformation of rice (Goto et al. 1993) and biolistic

transformation of maize (Gordon-Kamm et al. 1990, Spencer et al. 1992), the
frequencies of cointegration of two genes on separate plasmids were 100% (12/12
lines) and 77% (30/39), respectively, and the frequencies of coexpression of both
genes were 33% (4/12) and 18% (7/39), respectively. When two genes were on the
same plasmid, however, the coexpression frequency was much higher. Ha et al.
(1992) reported a 67% (4/6) coexpression frequency in protoplast transformation of
fescue, and Somers et al. (1992) reported a 75% coexpression frequency in biolistic
transformation of oat. All these data are in agreement with transformation studies of
dicots (Uchimiya et al. 1986, Schocher et al. 1986, Tagu et al. 1988, Damm et al. 1989,
Christou and Swain 1990).

Concluding Remarks
In the past three years, protocols for production of transgenic plants have been
established for major cereal and grass crops. In rice, the protocol relies on the efficient
regeneration of protoplasts in which DNA is delivered by electroporation (Fujimoto et al.
1993, Goto et al. 1993). In maize (Koziel et al. 1993), wheat (Vasil et al. 1992,1993),
and barley (Wan and Lemaux 1993), the success of transformation is based on biolistic
delivery of DNA to cultured immature embryos or callus with which a selection system is
developed to allow recovery of transgenic plants. It should be noticed, however, that in
each of these reports tremendous manpower and expertise in cell and tissue culture is
involved, especially in developing a system for a new species. It can be anticipated
that similar techniques will be proven effective in producing transgenics of other


Plant Regeneration

Plant Materials

Plants of Napiergrass (Pennisetum purpureum Schum.) used in this study were

grown in the field at three sites (Lots 1, 2, and 3) in Gainesville, Florida. Lot 1 was

located on the University of Florida campus which was 10 miles south of Lots 2 and 3.

Lots 2 and 3 were half a mile apart at the Horticultural Experimental Unit, located in the

northwest of Gainesville.

Three genotypes-PP13, PI 300086, and S41 were obtained from the collections

and breeding lines of Dr. S. C. Schank, University of Florida. PP13 and PI 300086 are

tetraploids of Napiergrass. S41, a triploid, is a hybrid to pearl millet (Pennisetum

glaucum (L.) K. Schum.). All plant materials had been maintained in Lot 1 for several

years before initiation of the present study. PP13 was moved to Lot 2 by transplanting

underground tissues in the spring of 1991, and in the early spring of 1993 they were

again moved in the same manner to Lot 3.

All plants used for DNA extraction and enzymatic assays were maintained in

greenhouse. They were fertilized once per growth cycle with Osmocote a slow-

release fertilizer (17-6-10 plus minor elements). They were watered 1-3 times per week

depending on the season; with hot summers watering was more frequent

Medium Preparation and in vitro Culture Conditions

Murashige and Skoog (1962) (MS) based media were prepared by mixing 4.3

g/L MS salt mixture (Sigma), 5 mLuL of MS vitamin stock (200 x), 3% (w/v) sucrose, 5%
(v/v) coconut water, growth regulators, and 0.8% (w/v) agar or 0.2% (w/v) gelrite. MS
vitamin stock (200 x) contains 20 mg/L thiamine-HCI, 100 mg/L nicotinic acid, 100 mg/L

pyridoxine-HCI, 400 mg/L glycine-HCI, and 20 g/L myo-inositol. Coconut water was
prepared by incubating the water of fresh coconuts from a grocery store at 75C for
one hour followed by filtration through Whatman # 4 filter paper. IAA (1 mg/mL) and
2,4-D (0.1 mg/mL) stocks were prepared by dissolving the chemicals in a few drops of
95% ethanol and then adding to H20. BA, kinetin, and zeatin, all at 1 mg/mL, were
prepared by dissolving the chemicals in 1 N NaOH and adding to H20. All stocks were
stored at -20*C. Kao and Michayluk's 8P medium (1975) were prepared by mixing
growth regulators, glucose and/or sucrose to lx of KM stock (2x). KM stock (2x) was
made of 200 mUL of 8p macro stock (10x), 20 mULL of 8p micro stock (100x), 20 mULL

of 8p vitamins stock (100x), 200 mL/L of organic acids stock (10x), 200 mUL of sugars
and sugar alcohols (10x), 200 mg/L of myo-inositol, 8 mLJL of folic acid stock (0.1
mg/mL), 500 mg/L of enzymatic hydrolyzate casein, and 10% (v/v) of coconut water.
All stocks were prepared according to Kao and Michayluk (1975) and stored at -20C.

All media were adjusted to pH 5.8 before filter sterilization (for 8PN and V-2%
only) or before autoclaving at 125C and 22 psi for 22-25 minutes (Table 1). Cultures

in Petri dishes were sealed with Parafilm* (American National CanTM, USA) and

incubated in the dark at 27C. Cell suspension cultures were incubated in Erlenmeyer
flasks (125-250 mL) on a gyratory shaker at 150 rpm in the dark at 27C.

Leaf Culture

Young leaf explants were obtained and cultured according to the procedures
modified from Haydu and Vasil (1981). Shoots were collected from about 6.5-foot-tall
plants. The 1-2 outermost layers of leaves were removed and the remaining apical part
of the shoot, which consisted of two visible nodes and several younger furled leaves,

was immersed in 95% ethanol for 30 seconds. After further removal of several outer
leaves aseptically, the innermost 3-4 tightly furled leaves were sequentially cut into 2
mm-thick transverse segments, starting at the base. For each shoot, leaf segments
were assigned sequential Arabic numerals, with the basal segment being S1. Seven to

10 consecutive leaf segments, with the cut surface of the basal end in contact with the
culture medium, were placed in one Petri dish (100 x 15 mm) containing approximately
25 mL of PP medium (Table 1). Cultures were monitored with a dissecting microscope.
One month after culture initiation, the following parameters were recorded: Fresh
weight of callus, embryogenic activity (defined as the number of sites with formation of
white compact embryogenic callus per leaf segment in tissue culture), and the induction
frequency of somatic embryogenesis for a leaf segment (defined as percentage of leaf
segments from a particular position along the length of the leaf that produce
embryogenic callus). Calli harvested from each explant were divided in half one month
after culture initiation. One half was transferred and subcultured at biweekly intervals
on fresh PP medium in a Petri dish (100 x 95 mm). The other half was transferred to
MSO.1 and then MSO medium (Table 1) for plant regeneration (see below). Cultures of
leaf segments were initiated in 1989 and 1990 with plants from Lot 1, and repeated in

1992 and 1993 with plants from Lots 2 and 3, respectively.

Young leaves of PI 300086 were cultured in the same manner as for PP13
described above. MS medium was used for induction of somatic embryogenesis
containing 3% (w/v) sucrose, 5% (v/v) coconut water, 0.2% (w/v) gelrite, and the
following concentrations of growth regulators: 0, 5, 12, 23, and 46 mM of 2-CPA, 3-

CPA, 4-CPA, 2,4-D, 2,4,5,-T, picloram, dicamba, chloramben, silvex, or coumarin.
Plants were randomly chosen from the field for tissue culture. For each plant, five leaf
segments (5-15 mm from the leaf base) were cultured as described for PP13 leaves.
The induction frequency of somatic embryogenesis (or callus formation) for a medium

was defined as the percentage of plants producing embryogenic callus.

00 1 1

wom W)

I 'E 9
V |

Cd o

V- T-

n o
co oi


710001000 10 10 0

00CGO00 0O00 f0 00
co o o 00G c
Wo -:

0 w (0 (0 000 0 w
22 22 22W

W) 9
S 2 2










I l
I 8


1 a 8






S ^C
82 Ea2

Inflorescence Culture

Inflorescences of PP13 were obtained and cultured according to the method
modified from Wang and Vasil (1982). Shoots were collected from field grown plants in

October of 1989-1993. The shoots were cut at the last visible node and immersed in
95% ethanol for 30 seconds after removal of the outmost furled leaves. The
inflorescences were dissected out under aseptic conditions and cut sequentially into 2
mm-thick transverse segments, starting at the base of the inflorescence and up to 12

cm toward the tip. The age of inflorescences ranged mainly from 5 mm to 5 cm. For
each inflorescence, the segments were numbered sequentially with the basal segment
being SI. Seven to 10 consecutively numbered segments, with the cut surface of the
basal end in contact with the culture medium, were placed in a Petri dish (100 x 15 mm)
containing approximately 25 mL of either PP medium or other media (Table 1, see
results). The induction frequency of somatic embryogenesis, defined as percentage of
inflorescence segments producing embryogenic callus, was recorded one month after
culture initiation. Calli harvested one month after culture initiation were subcultured on
PP medium at various intervals or transferred to MSO.1 and then MSO for plant
regeneration (see below).

The culture procedure for S41 inflorescence followed that described for PP13.
The media used for PI 300086 leaves that contained 5 mM 2,4,5-T or 5 mM 2,4-D were
used for inflorescence culture of S41. Young inflorescences (0.8-7.0 cm in length)
were used. Embryogenic calli obtained one month after culture initiation were
subcultured biweekly in the same medium for maintenance.

Establishment and Maintenance of Cell Suspension Cultures
Both white compact embryogenic callus and soft translucent nonembryogenic
callus obtained from young leaves or immature inflorescences on PP medium one

month after culture initiation were used for initiation of cell suspension cultures. The

calli were harvested and placed directly in 250 mL Edenmeyer flasks containing 30 mL
of MS2.5C (Table 1). Calli produced from each inflorescence segment were equally
divided in half. One half was transferred to MSO.1 and then MSO for plant regeneration
(see below). The other half was used for initiation of suspension cultures. Uquid
cultures in the flasks were kept on a gyratory shaker at 150 rpm in the dark at 27C
without subculture for the first month. Varying amounts of the medium were replaced
every 3-7 days for the following 2-3 months: old medium and elongated thick-walled
cells in the surface layer of the medium were removed and replaced with the same
amount or a few additional mL of fresh medium. With time, the rate of growth and the
number of small, thin-walled, and richly cytoplasmic cells (embryogenic cells) increased.
The small groups of embryogenic cells were preferentially selected, often by pipetting
the lower portion of the suspension a few seconds after a vigorous shaking of the flask,
and transferred to varying amounts of fresh medium in a new flask. The resulting
cultures, which were composed largely of embryogenic cells, were maintained by
transferring 4 mL of embryogenic cells from the lower portion of the flask to 30 mL of

fresh medium every week.

Cell suspension lines were also established by microscopic selection of
individual small dumps of embryogenic cells. A selected embryogenic cell clump was

placed in a 250 mL Erienmeyer flask containing 30 mL of MS2.5C and allowed to
propagate for 5-6 weeks before subculture. The resulting suspension was subcultured
weekly by transferring 4 mL of cells to 30 mL of MS2.5C.

To establish a cell suspension culture from cryopreserved cells, cells in an
polypropylene ampoule stored in liquid nitrogen were thawed in 40C water and plated
on MS2.5C solid medium. Calli obtained 10-14 days after plating were transferred to a
flask containing 30 mL of MS2.5C liquid medium.
For plant regeneration from a suspension, 4 mL of embryogenic cells were

transferred 3-4 days after subculture to a Petri dish (100 x 15 mm) containing

approximately 25 mL of MS2.5C solid medium (i.e., MS2.5C liquid medium solidified by
2 mg/L gelrite). Excessive liquid was removed, either by pipette or by evaporation for
30 minutes in a sterile hood, before sealing the dish. The call were transferred one
week later to MSO.1 and then MSO for plant regeneration (see below).

Isolation. Culture, and Reaeneration of Protoplasts from Cell Suspensions
Cells from suspensions 3-4 days after subculture were used for isolation of
protoplasts. Approximately 2 mL of the cells were incubated in a Petri dish (100 x 15
mm) containing 20 mL of an enzyme solution for 2-6 hours in the dark at room
temperature on a gyratory shaker at approximately 60 rpm. Enzyme solutions were
prepared by dissolving enzymes, as indicated in Table 24 in results, in the protoplast
washing solution (0.6 M mannitol, 0.7 mM NaH2PO4*H20, 5 mM CaCl2-2H2O, 5 mM
myo-inositol, and 3 mM MES, pH 5.8). The enzyme mixture used by Vasil et al. (1983)
was also tested. After incubation, the protoplast and enzyme mixture was filtered
successively through Miracloth (Calbiochem Corp., USA), 50 pm and 25 pm stainless
steel filters and collected in 15 mL polystyrene centrifuge tubes (Coming Inc., USA).
Protoplasts were pelleted and washed with the washing solution three times by
centrifugation at 800 rpm (100 x g) for 3 minutes. Protoplast yield was determined after

the second wash with a corpuscle counting chamber (Hausser Scientific, USA) by
placing into the chamber one drop of an unknown density of a protoplast suspension,
counting under a microscope the number of the protoplasts in the 16 smallest squares,
and multiplying the number by 5000 per mL to yield the protoplast density in the
solution. Protoplasts were cultured in 8PN medium (Table 1). The protoplasts were
plated at a density of 2 x 106 protoplasts/mL either in a shallow layer of liquid medium
(1.5 mL in 60 x 15 mm or 0.75 mL in 35 x 10 mm Costar Petri dishes, Costar, USA) or
by mixing equal volumes of the protoplast suspension with 0.6% (w/v) molten

SeaPlaque (FMC Corp., USA) agarose (final concentration 0.3%). Culture dishes were

sealed with Parafilm* and incubated in the dark at 27C. Dilution of the cultures was
made with V-2% medium (Table 1). One week after protoplast culture, 0.5 mL of V-2%
was added to the dish (60 x 15 mm). Two weeks after protoplast culture, 1.0 mL of V-
2% was added to the dish (60 x 15 mm). Three weeks after protoplast culture, 1.5 mL
of V-2% was added. Four weeks after protoplast culture and then every week, old
medium was removed and 1.5 mL of V-2% was added. Protoplast-derived colonies
(protocolonies), visible to the naked eye after four weeks of protoplast culture, were
transferred to MSO.1 and then MSO for plant regeneration (see below).

Chromosome Count
Translucent white root tips of three plants from suspensions and three plants
from protoplasts were collected and pre-treated at 4C in a saturated solution of 1,4-
dichlorobenzene for two hours, followed by fixation in absolute alcohol : glacial acetic
acid (3:1) for two hours. The fixed roots were hydrolyzed in 5 N HCI for 22 minutes and
pre-stained in 4% FeNH4(SO4)2 for two hours. After thorough washing with running tap
water for 30 minutes, the roots were stained in 0.2% hematoxylin (Sigma) for 30
minutes. Stained root tips were squashed with two drops of 45% acetic acid for

microscopic examination. Chromosomes were scanned under a microscope at

magnification of 200 x and viewed at 1000 x (oil immersion). The count for each plant
was established from 3 well-spread cells.

General Procedure of Plant Reaeneration from Callus
To regenerate plants, calli were transferred to MSO.1 (Table 1) and cultured
under a 16-hour photoperiod with cool white fluorescent tubes at 27C. After 2-4
weeks of incubation, germinated embryoids were transferred to a Magenta box
(Magenta Corp., USA) containing 50 mL of MSO (Table 1) and cultured under the same
condition. The young plants recovered in 3-4 weeks on MSO were transplanted to a

small pod (approximately eight cm in diameter) containing soil made of MetromixO

Vegetable Plug Mix amended with 20% (v/v) Perlite and were kept under 16-hour

photoperiod at room temperature. For the first two weeks in the soil, the plants were

covered with dear plastic bags to maintain high humidity. After the plants were fully

established, usually in 4-5 weeks, they were transferred to one-gallon plastic containers

containing the same soil mixture and maintained under greenhouse conditions.


Cryopreservation of Napiergrass cell suspensions followed the general

procedures for cultured cells (Withers 1985) modified for grasses by Gnanapragasam

and Vasil (1990, 1992a). The effect of different pregrowth treatments and various

cryoprotectants on the survival rate of cryopreserved cells were studied with a

nonregenerable cell suspension, S89. A combination of the pregrowth treatments and

cryoprotectants was then applied to a regenerable cell suspension, 17 (see results).

Prearowth Conditions

Eight to 12 mL of 7-day-old suspension cultures were transferred to 30 mL of

MS2.5C supplemented with either 6% mannitol or 6% sorbitol. The cells were

harvested for cryopreservation four days after culture.

Crvoprotectant Treatment

Ten mL of the pregrown suspension was transferred to a 15 mL centrifuge tube

and centrifuged at 100 x g for three minutes. The pelleted cells were adjusted to eight

mL of packed cells in a total of ten mL aliquot (80% PCV). The ten mL aliquot was

mixed well, transferred to a 125 mL Erlenmeyer flask and placed on ice. An equal

volume of an ice-cold double-strength cryoprotectant solution was added to the flask

gradually in a period of 60 minutes with constant agitation. The cryoprotectant solution
consisted of either 5% (v/v) DMSO and 0.5 M sorbitol (solution I) or 0.5 M glycerol, 0.5

M DMSO, and 1.0 M sucrose (solution II) and was prepared in double strength and filter
sterilized. After the addition of cryoprotectants, the suspension was kept on ice for an
additional one hour on a gyratory shaker at 100 rpm before freezing.

Freezing. Thawing. Regrowth. and Plant Regeneration

One mL of the cryoprotected suspension was aliquoted into a 1.2 mL
polypropylene, screw-cap cryogenic ampoule (InterMed Nunc, Denmark) and kept on
ice before freezing. The ampoules were placed in CryoMed Model 1010 Micro
Computer Programmable Freezer Unit (CryoMed, USA) and frozen at a rate of 0.5"C
per minute to -40C. The temperature was held at -40"C for 45 minutes before the
ampoules were transferred to liquid nitrogen in a liquid nitrogen storage tank (CryoMed,

USA). After a period (2-30 months) of storage in liquid nitrogen, ampoules were
removed from the storage tank and immediately placed in a 40C water bath and
swirled constantly. The ampoules were transferred to room temperature once the ice
had disappeared. The surface of the ampoules was wiped with 95% ethanol. Cell
suspensions were poured out into Petri dishes (100 x 15 mm) containing MS2.5C solid

medium with or without Whatman #4 filter paper. Excessive liquid was removed with a
pipette or by leaving the dish open under the laminar flow hood for 30-45 minutes.
Dishes were sealed with Parafilm* and cultured in the dark at 27C. After 7-10 days in

culture, cells were transferred to MS0.1 and then MSO for plant regeneration (see
below) or to a 250 mL flask containing 30 mL of MS2.5C to reestablish a cell
suspension culture.

Viability Assays of Crvooreserved Cells
The viability of cryopreserved cells was determined by the TTC (2,3,5-triphenyl
tetrazolium chloride) reduction assay and the regrowth potential test (Steponkus and

Lanphear 1967, Gnanapragasam and Vasil 1992). For the TTC assay, the thawed
suspension in an ampoule was transferred to a 15 mL centrifuge tube containing three
mL of 0.6% TTC solution containing 0.6% (w/v) TTC in 0.05 M Na2 HPO4 -KH2PO4
buffer (pH 7.4) and 0.05% (v/v) Triton x-100. The cells were kept at room temperature
in the dark for 16 hours followed by centrifugation at 100 x g for three minutes. After
two washes with double distilled water, the cell pellets were resuspended in seven mL
of 95% ethanol followed by 5-minute incubation in a water bath at 80C. After cooling

to room temperature, the volume of each tube was adjusted to ten mL using 95%
ethanol. The absorbance of the supematent was measured at 530 nm using a
Beckman DU-40 Spectrophotometer (Beckman, USA). The viability was expressed as
percent survival calculated by comparing the absorbance of cryopreserved cells to that
of the control (S89 suspension cells). For the regrowth potential test, thawed
suspensions in ampoules were transferred onto preweighed wet Whatman filter paper
(8.5 cm Whatman # 4) on MS2.5C solid medium in Petri dishes (100 x 15 mm). For the

control, ten mL of a 20% PCV culture was obtained from a 4-day-old suspension

culture of S89, and transferred to a sterile flask. One mL of the cells was pipetted after
vigorously shaking the flask and plated on a filter paper (8.5 cm Whatman # 4) in
MS2.5C solid medium. The filters and their cells were weighed again 30 minutes after
plating and every two days thereafter. Growth curves were obtained by plotting the
fresh weight of the culture against time. The rate of growth was determined according
to the growth curve by identifying a period of time when growth was linear. The percent
survival was then calculated by comparing the rate of growth for cryopreserved cells
and controls for the same period of time (day 6 to day 12 in Fig. 8).

Genetic Transformation

Plasmid DNA

Four DNA plasmids designed in different laboratories were used in the present
study (Fig. 1). pMON8678 (6.5 kp) contained a single uidA gene, which encodes p-

glucuronidase (GUS). It was constructed by inserting the 0.6 kb cauliflower mosaic
virus (CaMV) 35S RNA promoter (E35S) containing a duplication of the -90 to -300
region (Kay et al. 1987), a 0.58 kb fragment containing the first intron from a maize
alcohol dehydrogenase gene (ADH) (Callis et al. 1987), the coding sequence of the
Escherichia coli gene for GUS (Jefferson et al. 1986), and the 3' termination sequence

from the nopaline synthase (NOS) gene (Fraley et al. 1983) into pUC19 (Yanisch-
Perron et al. 1985). pMON19606, a triple expression vector (11.1 kb), was identical to
pMON8678, except that it included two additional genes: a variant of the maize 5-
enolpyruvylshikimate-3-phosphate synthase (EPSPS) gene conferring resistance to the
herbicide Glyphosate that was connected to the enhanced 35S promoter and the 3' NOS
termination sequences, and the neomycin phosphotransferase (NTP II) gene conferring
resistance to the antibiotic kanamycin (Beck et al. 1982) that was under control of the 35S

promoter (without the upstream duplication) and 3' NOS termination sequences (Vasil et al.
1991). pBARGUS (or referred to herein as Cu4), a dual expression vector of 9.1 kb,
contained uidA and bar of Stremptomyces hygroscopicus that encodes for
phosphinothricin acetyl transferase (PAT) and confers resistance to the herbicide Basta
(Thompson et al. 1987). Both genes, uidA and bar, were identical, with Adhl intron 1
and 3' NOS termination sequences, but differed in their promoters; uidA was driven by
the Adhl promoter and bar was under control of the 35S promoter (Fromm et al. 1990,
Vasil et al. 1993). pAHC25 (or C25, 9.8 kb) consisted of the uidA and bar genes, each
under control of the maize Ubil promoter (Christensen et al. 1992, Vasil et al. 1993).

pMON8678 (8.5 kb)


-I 6.5kb

pMON19606 (11.1 kb)

.--- 1 EPSPS AaH E3s -- S NPT I3I--I38 ADH GUS N
10.8kb-- 1.8kb -- 2.9kb 5.6kb

pBARGUS (9.1 kb)

CaMlV 355 Adhi BAR Nos 3' Nos 3'
promoter Intron 0.6 kb 0.3 kb 0.3 kb
0 43 k 0.5 kb

1.8 kb

Adhi Adhi
Intron promoter
0-5 kb 1.2 kb

PAHC25 (9.8 kb)

~us Jn





Fig. 1. Schematic representation of pMON8678, pMON19606, pBARGUS, and
pAHC25 plasmids (from Vasil et al. 1991, 1993).

Plasmid DNA was propagated in E. coli DH5-a, purified by ultracentrifugation,

sterilized in absolute ethanol, and dissolved in the regular TE buffer (pH 8.0) according

to protocols of Sambrook et al. (1989).

Biolistic Bombardment of Callus and Suspension Cultures

Explants of leaves and inflorescences or pieces of callus were transferred to

Petri dishes (60 x 15 mm) containing PP medium one day prior to biolistic

bombardment followed by incubation in the dark and 27*C for 24 hours. For

suspensions, approximately 0.5 g of cells obtained three days after subculture were

pipetted onto a 5.5 cm Whatman #4 filter paper on solid MS2.5C medium in a Petri dish

(60 x 15 mm) followed by incubation in the dark and 27C. Cultures were bombarded
with the DuPont PDS-1000 (gun powder) or PDS-1000/He (helium) apparatus (DuPont,

USA) using protocols described previously (Taylor and Vasil 1991).

Electroporation and PEG Treatment of Protoplasts

Protoplasts were isolated as described previously and electroporated with

plasmid DNA using a Cell-PoratorTm (BRL, USA) or treated with 18-20% polyethylene

glycol (PEG, 6000) according to Vasil et al. (1988). Treated protoplasts were cultured

as described previously. Viability of the protoplasts was estimated two days after

electroporation visually under the microscope with reference to controls (cultures of

protoplasts with no treatment).

Natural Tolerance Test of Cell Cultures to Basta and Glvphosate
Sensitivity of embryogenic callus to Basta was estimated using embryogenic
callus C91 derived from inflorescence (see results). Six pieces of the callus were

transferred to Petri dishes (100 x 15 mm) containing PP medium supplemented with 0,

1, 2, 4, 5, 10, or 20 mg/L of Basta. Five dishes were used per concentration of Basta.

The callus was then subcultured every 14 days. The number of callus pieces died (%

necrosis) was recorded seven days after the transfer. Fresh weight (mg) was

measured and number of plants formed per plate was counted 36 days after culture.

Natural tolerance of suspensions to Basta and Glyphosate was tested using the

2-month-old suspension S89 (see results). Four mL of suspension obtained 7 days

after subculture was transferred to 30 mL of MS2.5C and MS2.5 (Table 1). For Basta,

cells were transferred to MS2.5C containing 2.5, 5.0, 10.0, 20.0, 40.0, or 80.0 mg/L of

Basta. For Glyphosate, cells were transferred to MS2.5 containing 0.01, 0.10, 0.50,

1.00, or 10.00 mM of Glyphosate. The suspensions were cultured in the dark at 27"C

on a gyratory shaker at 125 rpm and subcultured weekly in the same medium. At each

transfer, four mL of cell suspension was obtained from the middle layer of flasks

immediately after vigorous shaking. Cell growth was monitored by measuring the

changes in packed cell volume (PCV) and fresh weight (FW). PCV was determined by

pipetting 4-10 mL of cell suspension into a graded centrifuge tube and centrifuging for

five minutes at 800 rpm; it was expressed as mL of packed cells per ten mL of cell

suspension. FW was determined by removing the supematant and weighing the

packed cells in the preweighed centrifuge tube; it was expressed as gram of cells per

ten mL of cell suspension. Each treatment had 3 flasks as replicates. PCV and FW

were measured three times per replicate.

To test the sensitivity of protoplasts to Basta, protoplasts were isolated from

suspension S89 (see results) and cultured as described previously. Two million

protoplasts were plated in 1.5 mL of 8p/v medium in a COSTAR dish (65 x 15 mm).

Basta was added at a final concentration of 1.0, 5.0, 10.0, and 20.0 mg/L one week

after protoplast culture, or it was added two weeks after protoplast culture along with

the dilution medium V-2% that contained 10 mg/L Basta. Protoplasts were also

electroporated (6 x 106 protoplasts per electroporation at 1180 pIF, 100 V, and 15 pg

pBARGUS) and cultured with addition of Basta (final concentration of five and ten

mg/L) three weeks after protoplast culture.

Enzymatic Assays and Southern Blot Analysis

Fluorometric detection and histochemical staining for GUS activity were carried

out according to Jefferson (1987). PAT activity was determined as described

previously (Chowdhury and Vasil 1992). NPT II assay followed the protocol described

by Peng et al. (1993). For Southern blot analysis, total DNA was isolated according to

Dellaporta et al. (1983), digested with restriction enzymes, fractionated on 0.8%

agarose, transferred to nylon membrane (Hybond, Amersham), and cross-linked with

ultraviolet irradiation. The membrane filter was hybridized with NP-labelled DNA

restriction fragments prepared by the random priming method of Feinberg and

Vogelstein (1983), washed at 65C in 0.5x SSPE and 0.1% SDS (Sambrook et al.

1989), and exposed to Kodak XAR5 x-ray film at -80C.


Plant Reaeneration from Leaf Cultures

Developmental Phase of Leaf Cells Mapped for Somatic Embrvoaenesis
Leaves of grasses display basipetal development Meristematic cells are
confined to leaf bases. Cells at the leaf tip develop earlier and reach maturity first
Therefore, along the length of a young leaf, cells of all developmental phases are
present, from undifferentiated meristematic cells at the base to highly differentiated and
mature cells at the tip. Experiments were designed such that the in vitro response of a
leaf explant could be studied according to its position along the length of the leaf and
the developmental phase conducive for the induction of somatic embryogenesis could
be localized (or mapped).
Because napiergrass is a perennial species, several experiments were
conducted with plants from the same stock (PP13) over several years and from

different locations. In 1989 and 1990, over one hundred plants from Lot 1 were used
for leaf culture. In nearly each case, only explants from a specific region near the base
of the leaf produced calli. To further characterize the in vitro response of the leaf,
seven plants were randomly chosen in 1989 and studied in detail. Each explant was
evaluated for its callus production, frequency of somatic embryogenesis induction,
embryogenic activity (the number of sites at which white compact embryogenic callus
was produced per explant four weeks after culture initiation), and capacity for plant
regeneration. Although the length of the three innermost furled leaves used for tissue

culture was over 200 mm, segments obtained from a region 14-26 mm above the leaf

base (S7-S13) produced the largest amount of call, 100% induction frequency of
somatic embryogenesis, the highest embryogenic activity, and the largest number of
regenerated plants (Tables 2, 3, and 4). Callus production from segments outside this
region declined. Explants from beyond 80 mm above the leaf base did not produce
any callus (data not shown). Therefore, this highly responsive region (14-26 mm above
the leaf base) was called the developmental zone of somatic embryogenesis.

Tissue culture response of plants obtained in 1992 and 1993 from newly
transplanted fields (Lot 2 and Lot 3), declined significantly, probably due to the fact that
the plants were not yet fully established after the transplantation. However, for the
responsive plants, defined developmental zones of somatic embryogenesis remained
unchanged (data not shown).

All responsive leaf explants developed call in a similar fashion. The surface of
the cut ends became swollen on the third day after culture initiation, followed by
appearance of yellowish callus tissue by the end of the first week, The callus tissue
grew to cover the entire surface of the explant A white compact embryogenic callus
began to emerge as small spots on the surface of the yellowish callus tissue
approximately 14 days after culture initiation. In 3-4 weeks after the initiation of culture,

many embryoids at different stages of development were seen scattered on the surface

of the embryogenic callus tissue. Occasionally, formation of shoots or roots was

observed At this time, three types of callus tissue could be identified: 1) white
compact embryogenic callus, 2) soft, friable yellowish-brownish nonembryogenic callus,
and 3) soft nonembryogenic gelatinous callus.
Embryogenic callus could be maintained by selective biweekly subculture for
several months without losing its capacity of regenerating plants. However, when the
nonembryogenic call were subcultured, necrosis often occurred.

Induction of Embrvoaenic Callus from New Genotypes

Because of potential commercial importance of cultivars PI 300086 and S41,
attempts were made to establish tissue cultures from explants obtained from near the
leaf base in media containing different growth regulators. Although 2,4-D, 2,4,5-T,
silvex, as well as chloramben at various concentrations induced callus formation in
PI 300086, white compact embryogenic callus was formed only in media containing
2,4-D and 2,4,5-T (Table 5). In leaves of S41, white compact embryogenic callus was
found only in two out of five dishes containing 2,4-D, and three out of five dishes
containing 2,4,5-T (data not shown). From leaf explants of both genotypes, the
embryogenic callus was produced in the same manner as from PP13, but at a lower
frequency (Table 5) and with a much lower embryogenic activity (data not shown).

Plant Reaeneration from Inflorescence Cultures

Embrvoaenic Callus Cultures Initiated from Seaments of Inflorescences

Various frequencies of somatic embryogenesis were observed from segments
of inflorescences using different media. With N6-based media, 5-40% of the cultured
inflorescences produced embryogenic callus (Table 6). With the PP medium used for
leaf cultures (Haydu and Vasil 1998), all inflorescences produced embryogenic callus
(Table 6). Furthermore, all segments from each inflorescence produced embryogenic
callus on PP medium. Although most inflorescences cultured ranged from 0.5 cm to 5

cm in length (Fig. 2), inflorescences up to 10 cm in length were also able to produce
embryogenic callus.

On PP medium, swelling of spikelet primordia and spikelets was observed one
week after culture initiation. Two types of callus could be identified two weeks after
culture: Translucent, soft friable nonembryogenic callus and white compact
embryogenic callus (Fig. 3). The calli appeared to be derived from the surface of the



V- .e







8 t oo2S

!A~s' N





S ^a 8 W a

a C4


- N w U 0 P.

3 I"



SU) N O00

F- -Ml0 0 10 0

- *-0 0-

*'*o. aS U

N F- U)

N Q S U) U) q

'~ ~ 2~ U) 2

U) 0 F- N 0 U

U 4 00 P -
- N U) U) S F











v a


00 0000

9 oomoooo

a owoobo

66 mlaob
So *o"boo d
o '- o b o o S0

b b Nobooo b I
S~h o ooboo g

b o C4 ob q

| 5 0o -ba C

0 o n 20 w 1

S? o oooho

SS cowob bo 0

era ococo o 0







-- V- 3 8



Coo0000 0000 o000 RV

388o g8 gSSgo set-



W) N0 CM i n 10 co < w- IV N qr qr 0 to to

w! s NV-as R! s w VRV RT .

Table 6. Summary of results on initiation of embryogenic callus cultures from PP13
inflorescences. Embryogenesis induction frequency (%) was defined as the
percentage of inflorescences producing embryogenic callus.

Total # Embryogenesis
Year of Inflorescences Induction
Culture Medium Cultured Frequency (%)

1982 N6 706

1990 N6 21 5
N6C 15 40
PP 1 100
I-MS2-CH 4 0
MS2 4 0
MS2-Glu-CH 2 0

1991 PP 182 100
PP-2 71 85b
MS2-5 72 85b

1992 PP 151 100

a Reported by Wang and Vasil (1982).
b Estimated percentage.

Fig. 2. Size range of immature inflorescences used for culture. x 1.4.

Fig. 3. Callus produced by an inflorescence segment 4 weeks after culture.
E: embryogenic callus. NE: nonembryogenic culture. x 13.


rachis and spikelets. Little or no callus was formed on the cut ends of segments.
Three to four weeks after culture, calli could be harvested for subculture. Prolonged
culture led to regeneration of plants from the embryogenic callus. When separated
from the nonembryogenic callus for biweekly subculture, the white compact
embryogenic callus continued to produce both white compact embryogenic callus and
soft nonembryogenic callus and maintained the capacity for plant regeneration for over

nine months. When the friable nonembryogenic callus was subcultured, either as a
callus piece separated from the compact embryogenic callus or as a piece attached to
the embryogenic callus, necrosis occurred with a frequency of 30-50% at each
subculture. Consequently, when subcultured in this manner, calli were lost 3-4 months
after culture initiation.

Effect of Subculture Intervals on Maintenance of Embryooenic Callus

An embryogenic callus line, named C91, was initiated and selected from call
produced with several plants on PP medium in 1991 and maintained by transfer of only
the embryogenic callus at biweekly intervals. The effect of subculture intervals on fresh
weight (FW), % necrosis (percentage of callus pieces dying at each subculture), and %
regeneration (percentage of callus pieces producing shoots at each subculture) was
studied three months after callus initiation. Six equal-sized pieces (approximately 7 mm

diam.) of the embryogenic callus were selected two months after callus initiation and
placed in a Petri dish (100 x 15 mm) containing PP medium. Five dishes were included

in the following three subculture regimes: 4-, 9-, and 14-day subculture intervals. At
each subculture, calli were weighed, divided, and transferred to the same medium. No
difference was found in fresh weight increase among all three subculture intervals with
each having approximately 20-fold increase in a period of 26 days (Fig. 4). The
percentage of necrosis began to increase after the second subculture in all subculture
regimes and reached 24%, 34%, and 20% after 26 days with 4-, 9-, and 14-day interval

regimes, respectively (Fig. 5). Subculture at 4-day intervals suppressed plant
regeneration with only 1-3% of callus pieces producing shoots (Fig. 6). However, when
subcultured at 9-day and 14-day intervals, up to 36% and 47%, respectively, of the call
were found to have produced shoots (Fig. 6). Therefore, in the case of nonselective
subcultures, calli could be lost due to necrosis and plant regeneration. Although

frequent subculture (i.e., 4-day intervals) could keep plant regeneration at a minimum
rate (Fig. 6), it did not reduce the percentage of necrosis occurring at each subculture
(Fig. 5).

Plant Reaeneration from Embrvoaenic Suspension Cultures

Susooensions Established with Embrvoaenic Callus from Leaves

Callus obtained three weeks after culture of PP13 leaves was placed in liquid
medium. During the first three months of subculture, no difference was found among
the three flasks of liquid cultures initiated. Few or no groups of embryogenic cells were
seen. After three months, embryogenic cells started to accumulate in one of the three
flasks. Selective subculture of the embryogenic cells in this flask resulted in a change

of morphology of the liquid culture. The culture became less viscous and the number

of embryogenic cells increased. In the next two months, a cell suspension resulted that
was not viscous and consisted almost entirely of embryogenic cells. This suspension,
designated as line L89, was maintained by transfer of 4 mL of embryogenic cells to 30
mL of medium every week. Cell clusters in the other two flasks remained undissociated
and little growth of embryogenic cells was observed.
Plant regeneration was induced from L89 approximately 7 months after initiation
of the suspension. Cells (4 mL) were plated on solid MS medium containing a
combination of 2,4-D (0 and 0.5 mg/L), BA (0.5, 1.0, and 2.0 mg/L) or zeatin (0.25, 0.5,
and 1.0 mg/L), ABA (0, 0.5, 1.0, and 1.5 mg/L), with or without CH (300 mg/L) and



44 1000-
U- 5W0-

. .

o N V 0 0 NV00


indtervais M indtervals ------0..... d"intervals

Fig. 4. Growth curves of C91 callus at 4-, 9-, and 14-day subculture intervals.
Embryogenic C91 callus was obtained 3 months after callus initiation and
subcultured (see text for details). Mean of five replicates.

0 30-

S 25-
20 "

a .. .. .i.


-- 4-d4 Intervals "---. 9-d intervals .... ....... 14- intervals

Fig. 5. Necrosis (%) of C91 callus at 4-, 9-, and 14-day subculture intervals.
Embryogenic C91 callus was obtained 3 months after callus initiation and
subcultured (see text for details). Mean of five replicates.

,, '<


-- 4-d intervals ------ 9-d Intervals --........ ........ 14-d Intervals

Fig. 6. Plant regeneration from C91 callus at 4-, 9-, and 14-day subculture intervals.
Embryogenic C91 callus was obtained 3 months after callus initiation and
subcultured (see text for details). Mean of five replicates.

ou -
20 -
15 -
10 -

I I I 1 1 A 1 A i A w & i

I- --

proline (15 mM). White and compact callus with organized structures was formed with
varying frequencies on all media tested. Only on the media containing 0.5 mg/L 2,4-D,
0.5 mg/L ABA, and 0.5-1.5 mg/L BA or 0.25-1.5 mg/L zeatin germination of somatic

embryos was observed. Green plantlets were obtained after the germinating somatic
embryos were transferred to MS medium containing no growth regulators. The
plantlets, however, did not survive in soil.

During the maintenance of L89, nonembryogenic cells gradually increased 11
months after initiation of the suspension. In the next 1-2 months, the nonembryogenic
cells began to dominate the culture and few embryogenic cels remained.

Suspensions Established with Embrvoaenic Callus from Inflorescences
The development of suspension culture was found to be different when call
from different segments of an inflorescence were used for initiation. In an experiment
with a 46-mm-long inflorescence, call derived from each of the 23 segments (each 2
mm thick) were placed in a flask containing liquid medium. Cultures in six out of 23
flasks did not grow and became black in the first 1-2 months. In approximately three
months, cultures in four of the 23 flasks developed roots. Removal of the roots from
the cultures at each subculture did not prevent development of new roots. These

cultures were later discarded when the flasks consisted mainly of roots. Four to five

months after initiation of suspensions, continued subculture produced two suspension
cultures, 17 and 117, which consisted mainly of small embryogenic cell clumps (SC).
The rest of the cultures remained largely undissocated and consisted mainly of large
and compact cell clusters (BC). The development of suspensions from initial call was
not found to be related to the capacity of the calli for direct plant regeneration nor the
position of the explants along the length of the inflorescence. The BC cultures did not
regenerate plants when transferred to regeneration media. The 117 suspension, which
was whitish, non-viscous, fast growing, and consisted mainly of embryogenic cells, also

did not regenerate plants when transferred to regeneration media. The 17 suspension
was yellowish, slightly viscous, and fast growing. It contained approximately 80%
small, thin-walled, and richly cytoplasmic cells that were in small cell groups
(approximately 400 pm in diam.) and 20% elongated cells or cell aggregates that had
little cytoplasmic contents. During the period 6-7 months after suspension initiation,
cells transferred to solid MS2.5C quickly grew into a yellowish layer of callus. A few
foci of organized structures could be seen on the surface of the culture. After transfer
of the cells to MSO.1, distinct somatic embryos could be identified in approximately one
week. Plants were germinated from the embryos 3-5 weeks after transfer of the callus
with somatic embryos to MSO medium. Plantlets recovered were established first in the
greenhouse and later in the field. The morphology and growth of the regenerants,
including the flowering date and pollen maturation, were similar to the control plants
from which the tissue culture was initiated (see below in protoplast cultures).

Calli collected from all segments of different ages of four inflorescences (3, 4, 8,
and 25 mm in length, respectively) were also used to initiate suspensions. With calli
from the 8-mm-long inflorescence, roots developed six weeks after initiation of the
suspension. In a period of five months, one suspension (127) was developed that was

fast growing, whitish in color, slightly viscous, and consisted predominantly of
embryogenic cells, while the other two consisted mainly of elongated cells and large

compact cell dusters. Cells from suspension 127 did not regenerate plants, however,
when transferred to regeneration media.

To test whether gradual change of liquid medium during the initial growth of
callus in liquid medium affects the development of suspensions, call from 3-4
sequential segments of one of three inflorescences (48, 32, and 9 mm in length,
respectively) were used in combination to initiate a suspension culture. Approximately

three weeks after culture initiation, subculture at 3-7 day intervals was initiated with

replacement of 3-4 mL of old medium with 4-6 mL of fresh medium. This practice of

subculture led to development of cultures that appeared slightly yellowish, very viscous
and began to accumulate embryogenic cells about two months after initiation of the
liquid cultures. The accumulation of embryogenic cells discontinued after a subculture

that involved excessive removal (approximately 10 mL) of the old medium and addition

of a large amount (approximately 25 mL) of fresh medium. In the following months of

continued subculture, no finely dispersed suspension cultures were produced.

An additional cell suspension was initiated according to Vasil et al. (1983).
Approximately 1 g of the white compact embryogenic callus, obtained four weeks after
culture initiation, was collected from different explants of inflorescences. The callus

was sliced into small pieces and placed into a 125 mL Erienmeyer flask containing 15
mL of MS2.5C liquid medium. Subculture was started two weeks after initiation of the

suspensions and continued weekly by removing and replenishing 4-8 mL of medium.
This suspension, designated as 189, was used three months after initiation to produce
the cell line S89 by microscopic selection (see below). Seven months after initiation,

the resulting suspension was nonviscous and consisted of 10-20% elongated

nonembryogenic cells and 80-90% embryogenic cells that were in aggregates of
various sizes. Transfer of 189 cells to regeneration media did not result In plant

In summary, a total of 43 cell suspensions was initiated with inflorescence-

derived calli. In a period of 5-6 months after initiation of suspension, five cell lines (17,

117, 127, 189, and S89) were established that were finely dispersed and consisted

mainly of embryogenic cells. Only one cell line, 17, regenerated mature plants.

Suspensions Established after Microscopic Selection
To develop a cell suspension that could rapidly release a large amount of
protoplasts, individual cell clumps (120-840 ipm in diameter) with dense cytoplasm were

selected from liquid cultures and transferred to 30 mL of liquid medium. Two such cell

dumps, 460 and 480 pm in diameter, were obtained from the inflorescence-derived
liquid culture 189 three months after its initiation (see above). In approximately seven
weeks, two cell suspensions were established that were both fast growing, nonviscous,
whitish to yellowish in color, and consisted mostly (over 95%) of embryogenic cells.
The 460 upm-derived suspension, designated as S89, was chosen to be maintained as
a cell line. Plant regeneration experiments with S89 only produced white compact
callus with somatic embryo structures. No plants were recovered.

Plant Reaeneration from Crvooreserved Cell Cultures

Crvooreservation of Cell Suspensions
The effect of different combinations of pregrowth treatments and cryoprotectant
solutions on cell line S89 is shown in the regrowth curves of the cryopreserved cultures
(Fig. 7). Pregrowth in 6% mannitol and cryoprotectant treatment with 0.5 M sorbitol and

5% DMSO produced the highest survival rate (33%) based on regrowth potential. It
had 86% survival rate based on the TTC assay (Fig. 8).

Plant Reaeneration from Crvopreserved Cell Cultures

Pregrowth in 6% mannitol and cryoprotection with 0.5 M sorbitol and 5% DMSO
was applied to the embryogenic cell line 17 seven months after culture initiation. This

combination led to the cryopreservation of the suspension, from which plants could be

routinely regenerated and established in the greenhouse after 2.5 years of cryogenic
storage (Fig. 9).

Plants were also regenerated and established in soil routinely from suspensions
(C17) reestablished from the cryopreserved 17 cells. CI7 suspensions resembled the
original 17 cell suspension in morphology, such as being slightly viscous and consisting

mainly (80%) of small groups of small, thin-walled, and cytoplasmic rich cells. Green

plants could be regenerated and established in the greenhouse two months after

transfer of C17-denved callus to regeneration media (data not shown). However, there
were large amounts of callus which formed only somatic embryos that did not
germinate to form plants. In addition, there were albino plants regenerated from the
cryopreserved suspensions, which did not occur in the original 17 suspension. In
addition, protoplasts isolated from C17 suspensions were also capable of plant
regeneration (see below). Because the original suspension 17 lost its ability for plant
regeneration within one year after initiation, the capacity of plant regeneration between

the 17 and CI17 suspensions was not able to be compared.

Plant Reaeneration from Protoolast Cultures

Protoplasts Isolated from Susoensions
Protoplast quantity (Table 7) and quality (see below) obtained from suspensions
varied among different cell lines. Cells from 1-2 month-old L89 yielded a low amount of
protoplasts even after overnight digestion (Table 7). Several enzyme mixtures,
including that of Vasil et al. (1983), released few or no protoplasts. With a combination

of 2% (w/v) of cellulase 'Onozuka' RS (Yakult Inc., Japan) and 0.1-1.5% (w/v) of
pectolyase Y-23 (Seishin Pharmaceutical Co., Ltd., Japan), 0.1-0.3 x 106 protoplasts
per mL of the enzyme mixture were obtained. Most cells remained undigested during
10 h incubation.

Cells from 1-12 month-old S89, on the other hand, readily released protoplasts.
During a three-hour digestion, the enzyme mixture containing 0.5% cellulase RS and
0.05% pectolyase Y23 produced an average of 3 x 106 protoplasts per mL of the
enzyme mixture (Table 7). Most cell clusters of the suspension were digested. Uttle
spontaneous fusion of the released protoplasts (5%) was observed. Using the same





-S- Mu


Fig. 7. Regrowth curves of cryopreserved S89 suspensions. Cells were pregrown in
the presence of mannitol (Mann.) and sorbitol (Sorb.) and cryopreserved in
solution I (1) and II (11). Control was one-year-old S89 suspensions (see text
for details). Mean of five replicates.





c 0.6

0.2 +

0 -



Fig. 8. TTC assay of cryopreserved S89 suspensions. Cells were cryopreserved by
pregrowth in 6% mannitol and cryoprotection in 0.5 M sorbitol and 5%
DMSO. Control was one-year-old S89 suspensions (see text for details).
Mean of five replicates SE.


Fig. 9. Plant regeneration from cryopreserved 17 suspensions. A: Cell storage tank.
x 0.19. B: Calli from cryopreserved cells plated on solid MS2.5C with
organized embryogenic callus (arrowhead). x 2.2. C: Germinating somatic
embryo on solid MSO. x 25. D: Plantlet on solid MSO from cryopreserved 17
cells. x 19. E: Plant in soil in the greenhouse. x 0.5.


Table 7. Summary of protoplast yields from L89,
suspensions. Each number represents
containing at least two replicates.

S89, 17, and cryopreserved 17
the average of an experiment

Cell Une Enzymes Yield
(x10s protoplastu/mL enzyme mix)

2% Cellulase RS
+ 0.1% Pectolyase Y23

2% Celulase RS
+ 0.5% Pectolyase Y23

2% Cellulase RS
+ 1.5% Pectolyase Y23

2% Cellulase RS +
1% Pectolyase Y23 +1% Pectinase

2% Cellulase RS + 0.5% Pectolyase
Y23 + 0.5% Pectinase

2% Celulase RS
+ 1% Macerozyme-RIO

2% Cellulase R8
+ 0.5% Macerozyme-R200

2% Celulase RS
+ 1% Driselase

2% Cellulase RS
+ 1% Pectinse

2% Celulase RS
+ 1% Rhozyme

2% Cellulase RS
+ 1% Rhozyme Rohn & Hass

That of Vasil at al. (1983)

0.5% Cellulase RS
+ 0.05% Pectolyase Y23

0.5% Cellulase RS
+ 0.05% Petolyase Y23

Cryopreserved 17

0.5% Cellulase RS
+ 0.05% Pectolyase Y23

Fig. 10. Protocallus formation from 17 protoplasts. A: Cells from the embryogenic cell
suspension culture. B: Freshly isolated protoplasts. x 390. C-F: First few
divisions of protoplasts after cell wall regeneration. x 390. G: Cell colonies
formed from protoplasts. x 10.





enzyme mixture ( 0.5% cellulase RS and 0.05% pectolyase Y23), cells from 17
(Fig. 10A) and cryopreserved 17 suspensions were also readily digested (Fig. 10B).
However, released protoplasts tended to fuse spontaneously with one another, leading
to a lower yield (Table 7). After 2-3 hours of digestion, an average of 0.3 x 106
protoplasts per mL of the enzyme mixture were recovered.
Protoplasts obtained from different cell lines were morphologically similar.
Nearly all isolated protoplasts were small in size (23.2 1.5 Upm diam.) and contained
very dense cytoplasm (Fig. 1013).

Protoplast Culture and Plant Regeneration
Protoplasts isolated from L89, S89, 17, and cryopreserved 17 suspensions
regenerated cell walls and started the first cell division within the first week in liquid
culture. Continued cell division was not found in L89-derived protoplasts in liquid or
solid media. Cells from L89 became elongated and contained little cytoplasm.

In agarose culture, protoplasts from S89 and 17 suspensions yielded zero to a
few colonies per culture dish one month after culture, compared to hundreds to
thousands of protocolonies in liquid culture. Protoplasts in liquid culture sedimented on
the bottom of the dish and aggregated especially around dish corners. Consequently,
the plating efficiency could not be assessed.

In liquid culture of 17 protoplasts, the most visible change within the first two
days was the formation of many oval shaped cells. The first cell division was observed
4-5 days after culture in different experiments. Many two-celled structures were found
by 6-7 days (Fig. 10C). At this time, unequal cell division that resulted in two daughter
cells of different size was frequently encountered; equal cell division was also
observed. The next few divisions took place in the following two days. Three- to four-
celled structures were observed and cell colonies with 4-16 cells were also found (Figs.

10D-F). Continued cell division in the following weeks resulted in colonies of cells that

were small, thin walled, richly cytoplasmic, and contained many starch grains. The cell
colonies were compact with no visible intercellular spaces (Fig. 10G). The culture at
this time resembled the suspension from which the protoplasts were isolated. Colonies
became visible to the naked eye 3-4 weeks after protoplast culture. No colonies were
found in the agarose culture, however. The large 17 protoplast-derived colonies were
harvested four weeks after protoplast culture. Transfer of the colonies to MSO.1
medium resulted in continued cell growth and differentiation, leading to formation of
somatic embryos and germination of the embryos in three weeks (Fig. 11A). After
transfer to MSO medium, green plants were observed in 1-2 weeks (Fig. 11B).
Fourteen regenerants were chosen for transfer to soil. AH 14 plants grew to maturity in
the field (Fig. 11C). The plants in the field were morphologically normal and produced
flowers and anthers identical to the control plants from which the original tissue culture
was initiated. Chromosome count revealed the normal number 2n=28 (data not

The culture of protoplasts from S89, 127, and cryopreserved 17 suspensions
followed a similar course of cell growth and differentiation described for 17. However,
S89 protoplast-derived colonies did not regenerate plants. A few protocolonies derived
from 127 protoplasts regenerated only shoots. Green plants were obtained from

protoplasts of the cryopreserved 17 suspensions (see below). But they did not continue

growth after formation of the first few leaves and roots.

Genetic Transformation via Bombardment of Leaf and Inflorescence Calli

Effect of Biolistic Bombardment on the Development of Embrvooenic Callus
After having identified the leaf cells competent for somatic embryogenesis as
targets for direct gene delivery, the effect of biolistic bombardment on callus formation

was studied. Seven or eight leaf segments per dish, obtained from the region between

Fig. 11. Plant regeneration from 17 protocallus. A: Organized embryogenic callus
from 17 protocallus (Fig. 10G). x 11. B: Green plantlets regenerating from
17 protoplast-derived embryogenic callus. x 2. C: Flowering plants from 17
protoplasts and suspensions (row D: plants regenerated from 17
protoplasts, row M: plants regenerated from 17 suspensions seven months
after suspension initiation, row 2: control plants).