Experiments concerning the molecular evolution of the allotetraploid Pennisetum purpureum (napiergrass)

MISSING IMAGE

Material Information

Title:
Experiments concerning the molecular evolution of the allotetraploid Pennisetum purpureum (napiergrass)
Physical Description:
vi, 179 leaves : ill., photos. ; 28 cm.
Language:
English
Creator:
Ingham, Lynwood D., 1956-
Publication Date:

Subjects

Subjects / Keywords:
Pennisetum purpureum -- Evolution   ( lcsh )
Molecular evolution   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 167-178).
Statement of Responsibility:
by Lynwood D. Ingham.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001706501
notis - AJB8756
oclc - 25479632
System ID:
AA00003329:00001

Full Text









EXPERIMENTS CONCERNING THE MOLECULAR EVOLUTION OF THE
ALLOTETRAPLOID PENNISETUM PURPUREUM (NAPIERGRASS)


















By

LYNWOOD D. INGHAM JR.


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

UNIVERSITY OF FLORIDA


1990
























This dissertation is dedicated to my mother Verna, my

brother Jimmy, my sisters Lynda, Beth, and Linda, my

stepfather Jim, my grandmother Thelma, my grandfather Howard

and my great aunt and uncle Verna and Vincent. Besides my

blood family, this dissertation is dedicated to my spiritual

families who are teaching me to live life on life's terms and

to strive for my potential. I also dedicate this dissertation

to Dave who has consistently been there for me and has been

guiding me through the steps.












ACKNOWLEDGEMENTS


First I acknowledge my major professor Curt Hannah for

giving me a home, the means, and the opportunity to become an

independent thinker; we look forward to our continued

interaction over the next couple of years. Thanks go to Wayne

Hanna for our continued collaboration and also for setting an

example for me. Many thanks go to my committee members Rob

Ferl, Don McCarty, Bill Gurley, and Francis Davis for their

continued support. Special thanks go to Rob Ferl for his

confidence in me. Special thanks go to Don McCarty for the

continued interaction between him and his lab and myself.

Special thanks go to Bill Gurley for getting me excited about

science and leading me into plant molecular biology. Thanks go

to Eduardo Vallejos, along with Brian Bournival, for teaching

me something about isozymes and for our interaction. Thanks go

to Ernie Almira for his collaboration in the sequencing core.

Thanks go to Kathy Taylor for our conversations about science

in general. Thanks go to Janine Shaw, Maureen Clancy, John

Baier, Dina St. Clair, Mrinal Bhave, and the rest of the

Hannah lab members. Thanks go to Sam Camp, Beth Laughner,

Caprice Simmons, Chris Carson and the rest of the people I've

had the pleasure of interacting with here in Fifield Hall.


iii













TABLE OF CONTENTS



ACKNOWLEDGEMENTS . . .

ABSTRACT ... . . v

CHAPTER 1 INTRODUCTION . .. 1

CHAPTER 2 MATERIALS AND METHODS . 19

DNA Isolation, RNA Isolation, Digestion, and
Fractionation ....... .. .... 19

Southern Blotting, RNA Blotting, Probe Labeling,
and Hybridizations .. .. ... 22

Genomic and cDNA Cloning ............. 25

Sucrose Synthase Isolation and Antibody Production 25

Copy Number Determination .. ...... 25

DNA Sequencing .. . 27

Isozyme Electrophoresis and Staining ...... 28

CHAPTER 3 RESULTS OF PROGENITOR SYUDY ...... .30

CHAPTER 4 RESULTS OF SUCROSE SYNTHASE COMPARISONS .. 95

CHAPTER 5 DISCUSSION ...... ... .... 150

REFERENCE LIST .. ..... ..... 167

BIOGRAPHICAL SKETCH ................ 179













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

EXPERIMENTS CONCERNING THE MOLECULAR EVOLUTION OF THE
ALLOTETRAPLOID PENNISETUM PURPUREUM (NAPIERGRASS)

By

Lynwood D. Ingham JR.

December, 1990


Chairperson: L. Curtis Hannah
Major Department: Horticultural Sciences

This study was undertaken to obtain molecular evidence

concerning the identification of the progenitors of the

allotetraploid Pennisetum purpureum Schumach (napiergrass)

with particular emphasis on sucrose synthase coding sequences

of P. purpureum and P. alaucum (L.) R. Br. (pearl millet).

Published cytological and crossability evidence suggests that

P. glaucum (AA) is one of the progenitors of P. purpureum

(A'A'BB). The data presented here suggest that P. squamulatum

may share in common both the A' and B genomes of P. purpureum.

Average or consensus sequences were determined for the 140 bp

and 160 bp Kvn I families of tandemly arrayed repetitive

sequences from P. purpureum and P. scuamulatum. Average

sequences were determined for the P. glaucum 140 bp Kpn I

family and the diploid P. hohenackeri Hochst. ex Steud. 160 bp

KDn I family. P. purpureum and P. scuamulatum share similar

v







restriction fragment length polymorphisms (RFLP). Isozyme

phenotypes discussed here show a close relationship between

P. purpureum, P. squamulatum, and P. glaucum. The combination

of the evidence presented here with previously published

evidence including crossability, isozyme phenotypes, and

cytoplasmic genome similarities suggests that P. purpureum

evolved from the interspecific hybridization of P. glaucum

with P. squamulatum. The other possibility is that P.

purpureum and P. sauamulatum share common progenitors and that

the interspecific hybridizations which gave rise to these

plants occurred at approximately the same time period. P.

purpureum diverged from P. squamulatum at a later date than

their divergence from P. Qlaucum. The major sucrose synthase

proteins in P. purpureum and P. glaucum root appear slightly

smaller than the 88,000 mol wt maize sucrose synthase

proteins. A sucrose synthase genomic sequence from P.

purpureum was shown to have similar gene structure as compared

to the Sus gene from maize. The cDNA encoded by this P.

purpureum sucrose synthase gene as well as a P. glaucum

sucrose synthase cDNA was sequenced. These two cDNAs exhibit

98.6% homology from exon 5 through the polyadenylation site in

the P. purpureum cDNA. Another genomic sequence from P.

purpureum having similarity to the 5' end of Shl has been

cloned. High copy number poly A RNA of less than 2.0 kb

having similarity only to the 5' end of sucrose synthase

coding sequences was also identified.












CHAPTER 1
INTRODUCTION



Tandemly arrayed repetitive sequences appear to evolve in

concerted evolution (Dover 1982). These repeated DNA sequences

remain homogenous within a species over evolutionary time. The

rate of fixation of the sequence is greater than the rate of

divergence (Dover 1982; Arnheim 1983). Molecular drive results

in the fixation of a particular repeated sequence. This

fixation of variants in a population is not predicted by the

genetics of natural selection and genetic drift (Dover 1982).

Two major mechanisms have been proposed to participate in

concerted evolution; unequal crossing over (Smith 1976) and

gene conversion (Ohta 1977; Nagylaki and Petes 1982).

Mechanisms allowing the proliferation of tandemly arrayed

repetitive sequences have been postulated. Tandemly arrayed

repetitive sequences are generally located in the constitutive

heterochromatin of chromosomes (John and Miklos 1979; Appels

et al. 1981; Bedbrook et al. 1980). These heterochromatic

regions of chromosomes have low recombinational rates and

hence reduced rates of unequal crossing-over (Miklos 1985).

Stephan (1989), using computer programming, suggests that the

crossing-over rate in heterochromatin is too low to cause the

proliferation of tandemly arrayed repetitive sequences.
1







2

Slipped-strand mispairing, or replication slippage, may be a

major mechanism behind concerted evolution of repeated

sequences (Levinson and Gutman 1987). Based upon computer

modeling, Stephan (1989) has proposed that the creation of

nucleotide periodicities involves not only unequal crossing-

over, but an amplification process as well. This amplification

process may consist of replication slippage and rolling circle

replication. The evidence for rolling circle replication is

the appearance of extrachromosomal circular satellite DNAs

(reviewed by Walsh 1985) suggesting an extrachromosomal

amplification process such as rolling circle replication

(Hourcade et al. 1973).

The function of repetitive DNA is an unresolved topic.

Repetitive DNA has been termed selfish DNA. Doolittle and

Sapienza (1980) referred to prokaryotic transposable elements

and eukaryotic middle repetitive sequences as selfish DNA; and

suggested that selfish DNA promotes its own spread through a

population at the expense of the evolutionary fitness of the

population. Orgel and Crick (1980) stated that selfish DNA

arises as a sequence spreads through a genome while making no

significant contribution to the phenotype. These authors

referred to selfish DNA as repetitive DNA, introns, and

segments of DNA between genes. Dover (1980) questioned the

concept that selfish DNA is not involved in phenotypic

selection; however, he did agree that selfish DNA may not have

an intrinsic function. He suggested that selfish DNA may give







3

some specific and common form to closely related genomes. Many

workers have suggested varied functions of repetitive

sequences. Bostock (1980) proposed that the association of

constitutive heterochromatin with satellite DNA, or tandemly

arrayed repetitive sequences, suggests possible functions of

these repetitive elements. He classified potential functions

into four broad categories: chromosome structure, cell

metabolism, homologous pairing of chromosomes, and promotion

of chromosome rearrangements or recombination. Miklos (1985)

has reviewed proposed functions of tandemly arrayed repetitive

sequences. The functions which have been proposed include

determination of centromere strength, chromosome pairing,

recombination, three dimensional architecture of the nucleus,

genomic reorganization, and speciation. Also mentioned is a

second class of proponents suggesting that these repeated

sequences are byproducts of molecular mechanisms involved in

DNA replication and recombination. In spite of advances in

molecular biology, the role of repetitive sequences continues

to elude investigators.

In plants, tandemly arrayed repetitive sequences are

often termed highly repetitive sequences. Highly repetitive

sequences have been classified as those sequences with a copy

number greater than 100,000 per genome (McIntyre et al. 1988).

There are, however, examples of moderately repetitive

sequences which are tandemly arranged (Zhao et al. 1989;

Hallden et al. 1987; Shmookler-Reis et al. 1981). These







4

tandemly arrayed repetitive sequences are often termed

satellite DNA (Ganal et al. 1986; Peacock et al. 1981) because

they may be separated from the bulk of nuclear DNA by

equilibrium sedimentation in density gradients (Kit 1961). The

term satellite DNA is a misnomer since not all tandemly

arrayed repetitive sequences are AT rich and thus cannot be

separated in a density gradient (Koukalova et al. 1989).

Repeat sizes of tandemly arrayed repetitive sequences in

plants are usually in the 160 to 200 base pair range

(Schweizer et al. 1988; Junghans and Metzlaff 1988; Dennis and

Peacock 1984). Repeat sizes in the 350 base pair range have

also been reported (Ganal and Hemleben 1986; Ganal and

Hemleben 1988). Tandem repetitive elements in constitutive

heterochromatin are commonly methylated (Deuming 1981;

Shmookler-Reis 1981; Ganal and Hemleben 1988). The percentage

of plant genomes consisting of characterized repetitive

element families varies: the highly repeated satellite DNA of

Cucumis sativus composes 20-30% of the nuclear DNA (Ganal et

al. 1986), the GC rich satellite DNA in Scilla siberica

accounts for 20% of the nuclear genome (Deumling 1981), the

Hind III family of Brassica campestris accounts for 15% of the

genome (Lakshmikumaran and Ranade 1990), the 375 bp tandemly

repeated sequences of Allium cepa constitutes 4% of the genome

(Barnes et al. 1985), and the Bam HI family in Nicotiana

tabacum accounts for 2% of the nuclear genome (Koukalova et

al. 1989). Variation in DNA content between closely related







5

species is at least partially due to differences in the amount

of tandemly arrayed repetitive sequences (Bennett et al. 1977;

Geever et al. 1989).

In plants there are few examples of characterized

tandemly arrayed repetitive sequences. Most of the examples

include Scilla siberica (Deumling 1981), Vicia faba (Kato et

al. 1984), Allium cepa (Barnes et al. 1985), Triticum species

(Rayburn and Gill 1988), Oryza species (Zhao et al. 1989),

Secale cereale and Secale silvestre (Bedbrook et al. 1980).

Secale cereale and Agropyron cristatum (Xin and Appels 1988),

Lycopersicon esculentum and Solanum acaule (Schweizer et al.

1988), Hordeum vulgare (Junghans and Metzlaff 1988), Zea mays

and its relatives (Dennis and Peacock 1984), Cucurbita species

(Leclerc and Siegel 1987; Ganal and Hemleben 1986), Cucumis

sativus L. (Ganal and Hemleben 1988), Nicotiana tabacum

(Koukalova et al. 1989), and several species within the family

Brassicaceae (Hallden et al. 1987).

Tandemly arrayed repetitive sequences in plants have been

utilized for various purposes; most of these involve

investigating the relatedness of species. For example, in the

family Brassicaceae, a 175 bp sequence from Brassica napus

hybridized to DNA from six genera within that family. One of

the species hybridizing to this repeat was in a different

tribe than Brassica napus (Hallden et al. 1987). In Cucumis

sativus L. there are three types of tandemly arrayed

repetitive sequences of around 180 bp (Ganal et al. 1986).







6

Also present in this species is a 360 bp repetitive element

that contains a 180 bp segment related to the above three

types of sequences. There is a sequence in the related species

Cucumis melo similar to the cucumber 360 bp repeat (Ganal and

Hemleben 1988). Major tandemly arrayed repetitive sequence

families from two Cucurbita species, C. epgo and C. maxima

were cloned and sequenced. These repetitive elements exhibited

no cross homology. However, the C. pepo repeat was present in

low amounts in the C. maxima genome and vice versa indicating

a common ancestor contained both of these sequences prior to

speciation (Ganal and Hemleben 1986). The major tandemly

repeated sequences in maize heterochromatin were shown by

hybridization and sequencing to be more closely related to

teosinte than to Tripsacum dactyloides thus suggesting maize

was derived from teosinte (Dennis and Peacock 1984). Tandem

repeats from Lycopersicon esculentum and Solanum acaule were

used to identify hybrids from protoplast fusions between these

two species. The tomato repeat is also present in other

Lycopersicon species (Schweizer et al. 1988). A common 350 bp

repeat found in Secale cereale and Agropyron cristatum

terminal heterochromatin was reported to be an example of

parallel amplification and not an indication of relatedness

between these two species. The S. cereale and A. cristatum

species differed at their rDNA intergenic spacer regions, 5S

DNA loci, and also differed in morphological characters,

chromosome pairing, and isozyme patterns (Xin and Appels







7

1988). In the Oryza species genome specific tandemly arrayed

repetitive sequences are specific based upon hybridizational

intensities; these probes can be used for classifying unknown

species (Zhao et al. 1989). Most other published work on plant

tandemly arrayed repetitive sequences involves only the

characterization of those families.

Another repetitive sequence family used in this

evolutionary study is ribosomal DNA, or rDNA. Structural

variation in rDNA between species can provide useful

information concerning the points of divergence of related

species (Springer et al. 1989; Burr et al. 1983; Cordesse et

al. 1990). Ribosomal DNA repeat structure has been well

characterized (Long and Dawid 1980; Appels and Honeycutt

1986). The 17S, 5.8S, and 26S rDNA genes are transcribed as

one unit and then cleaved to yield mature rRNA (Perry 1976).

Between each set of rDNA genes is the intergenic spacer, or

IGS (Miller and Beatty 1969). Ribosomal DNA varies in repeat

size within and between species (Rogers and Bendich 1987a;

Choumane and Heizmann 1988; Molnar et al. 1989; Cordesse et

al. 1990). Heterogeneity of rDNA repeat size is present within

many species (Jorgensen et al. 1987; Rogers and Bendich 1987;

Molnar et al. 1989; Springer et al. 1989; Cordesse et al.

1990) but absent from others (Varsanyi-Breiner et al. 1979;

Kavanagh and Timmis 1988; Springer et al. 1989). These repeat

sizes range from 7.8 to 18.5 kb (Varsanyi-Breiner 1979; Yakura

et al. 1983). Length variation of the rDNA repeat within a







8

species and/or within a genus has been accounted for by the

number of subrepeats within the IGS (Yakura et al. 1984;

McMullen et al. 1986; Choumane and Heizmann 1988; Cordesse et

al. 1990). In Hordeum species the number of rDNA repeat sizes

is equal to the number of nucleolar organizer regions (NOR) or

rDNA loci (Molnar et al. 1989). There is only one NOR in Vicia

faba but the rDNA repeat length shows considerable

heterogeneity; hence in this case the number of rDNA loci is

not equal to the rDNA repeat length heterogeneity (Rogers and

Bendich 1987b). Allotetraploids in Oryza species may exhibit

less rDNA repeat size polymorphisms than their diploid

progenitors. This homogenization process of the rDNA loci on

nonhomologous chromosomes may occur relatively rapidly

(Cordesse et al. 1990). Triticum aestivum exhibits co-

evolution of rDNA repeats on nonhomologous chromosomes as well

(Dvorak 1982).

The KDn I tandemly arrayed repetitive sequence families

in several Pennisetum species were characterized, along with

RFLP analysis, to identify relationships between nuclear

genomes of P. purpureum and related species P. Qlaucum and P.

squamulatum. There are several lines of evidence suggesting a

relatedness amongst the above three species. P. glaucum

(2n=2x=14) is considered to be a progenitor of P. purpureum

(2n=4x=28) (Jauhar 1981). When P. glaucum (AA) is crossed with

P. purpureum (A'A'BB), the resulting triploid interspecific

hybrid (A'AB) contains seven bivalents and seven univalents







9

suggesting a close relationship between these two species

(Harlen 1975; Jaurar 1981). P. squamulatum (2n=6x=54) has been

described as an autoallohexaploid (Patil et al. 1961), an

allohexaploid (Rangaswamy 1972), and a segmental allohexaploid

(Jaurar 1981). An autoallohexaploid contains two different

genomes, one of which is in the tetraploid state. An

allohexaploid is composed of three different genomes. A

segmental allohexaploid exhibits segmental homology between

the constituent genomes. P. glaucum can also be crossed with

P. squamulatum at a high frequency (Dujardin and Hanna 1989)

resulting in sexual and apomictic hybrids (Dujardin and Hanna

1983). Napiergrass and P. squamulatum as well can be crossed

(Hanna, unpublished). Double cross hybrids between P. glaucum,

P. purpureum, and P. squamulatum indicate a relatedness

between the genomes of these three species (Dujardin and Hanna

1984). It is likely that P. scruamulatum or P. purpureum genes

fully and partially restored male fertility in cytoplasmic

male sterile P. glaucum interspecific hybrids thus indicating

common genes are present (Dujardin and Hanna 1990). The

monomorphic prolamine seed proteins in P. glaucum are present

in P. squamulatum and some accessions of P. purpureum (Lagudah

and Hanna 1990). Isozyme banding patterns also suggest that P.

squamulatum has partial homology to the genomes of P. glaucum

and P. purpureum (Lagudah and Hanna 1989). Significant

relationships exist between P. purpureum and P. squamulatum







10

mitochondrial DNAs (mtDNA); both of these mtDNAs are related

to P. glaucum mtDNA (Chowdury and Smith 1988).

This study focuses on the close homology of the Kpn I

tandemly arrayed repetitive sequence families between P.

purpureum, P. squamulatum, and P. glaucum. Also addressed is

the similarity between P. purpureum and P. squamulatum in rDNA

RFLPs, the common RFLPs when DNA from these two species are

probed with napiergrass genomic and cDNA clones, and similar

isozyme phenotypes. The unexpected results of this study

suggest that P. purpureum is more closely related to P.

squamulatum than it is to P. claucum. When all the evidence is

considered, the genomes of the hexaploid P. squamulatum may

include the genomes present in the tetraploid P. purpureum. P.

purpureum and P. squamulatum may have had common progenitors

and arose from independent interspecific hybridization events

at similar times. Or, napiergrass (A'A'BB) may have evolved

from a cross of P. squamulatum and P. glaucum with both

gametes in the reduced state prior to fertilization. The above

cross would yield an interspecific hybrid composed mostly of

chromosomes from P. squamulatum; this model fits the data

presented in this paper.

Another set of experiments involved with studying the

molecular evolution of P. purpureum included the cloning and

sequencing of sucrose synthase coding sequences. By comparing

the sequences of sucrose synthase genes within and outside of







11

the genus Pennisetum the divergence of P. purpureum was

further elucidated.

Sucrose synthase (UDP-glucose: D-fructose-2-glucosyl

transferase, EC 2.4.1.13) catalyzes the reversible reaction

below:

UDP-glucose + fructose <-----> Sucrose + UDP

(Cardini et al. 1955). The isolation of this enzyme first

occurred in wheat germ (Cardini et al. 1955). Originally

Cardini et al. (1955) suggested that sucrose synthase led to

the formation of sucrose in vivo. By 1957 Turner and Turner

proposed that the function of sucrose synthase is to supply

UDP-glucose as a glucosyl donor in the synthesis of starch.

Experimental evidence supporting the proposed role of sucrose

synthase in the degradation of sucrose was initially obtained

in sweet corn (De Fekete and Cardini 1964) and in rice (Murata

et al. 1966). The experimental evidence was the same for both

sweet corn and rice; glucose-C14 from sucrose was incorporated

into starch while the nucleotide diphosphate (NDP), UDP, was

shown to be the NDP of choice for the enzyme degrading the

sucrose. Since sucrose synthase utilizes UDP in the

degradation reaction of sucrose, the authors of both papers

suggested that sucrose synthase degrades sucrose instead of

synthesizing sucrose. Chourey and Nelson (1976) offered the

first genetic evidence supporting the role of sucrose synthase

in starch biosynthesis by showing that shl (a sucrose synthase

deletion mutant was used in this experiment) maize kernels are







12

deficient in starch. Detailed kinetic and substrate

specificity studies of sucrose synthase from various sources

(De Fekete and Cardini 1964; Avigad 1964; Avigad et al. 1964;

Milner and Avigad 1964, 1965; Avigad 1967; Grimes et al. 1970;

Murata 1971; Delmer 1972; Murata 1972; Sharma and Bhatia 1980)

suggest that sucrose synthase degrades sucrose mainly in

storage tissues. This sucrose degradation occurs at sites

lacking in invertase, with intense growth, and with high

sucrose concentration (Avigad 1982). The UDP-glucose product

from sucrose synthase degradation of sucrose can be used also

for cell wall synthesis and storage compounds (Feingold and

Avigad 1980). Franck (1979) referred to D-glucose as the major

"key building block" used for synthesis of natural products.

Pontis (1977) suggested that for heterotrophic organisms whose

viability depends on the consumption of organic carbon, most

organic constituents have, at one time, been part of a sucrose

molecule.

Carbon fixation leading to sucrose biosynthesis differs

between C3 and C4 photosynthetic types of plants. In C3 plants

the reductive pentose phosphate pathway occurs in the stroma

of the chloroplast; CO2 and inorganic phosphate are converted

to triose phosphate and exported to the cytosol for conversion

to sucrose (Sicher 1986). With C4 plants, characteristically

tropical grasses, there is an added cycle of carbon fixation

The first cycle of carbon fixation occurs in the mesophyll

cells. In NADP-malic enzyme (NADP-ME) C4 plants,







13
phosphoenolpyruvate is carboxylated to yield oxaloacetate in

the cytosol of the mesophyll cells. This oxaloacetate is

converted into malate in the chloroplast; the malate is

transported to the chloroplast of the bundle sheath cells. The

malate is then decarboxylated to yield pyruvate which cycles

back to the mesophyll chloroplast to be converted to

phosphoenolpyruvate. The CO2 from malate then enters the

reductive pentose phosphate pathway. Sucrose is synthesized in

the cytosol of the bundle sheath cells of C4 type

photosynthetic plants (Rees 1987).

The free energy of hydrolysis of sucrose is conserved by

the formation of UDP-glucose. The free energy of sucrose

hydrolysis is -7000 calories (cal). This free energy is

conserved in the formation of UDP-glucose which has a free

energy of hydrolysis of the alpha-D-glucopyranosyl phosphate

bond of -8000 cal. This energy of the glycosidic linkage of

sucrose could be used for the synthesis of other saccharides,

starch, callose, or cell wall polysaccharides (Neufeld and

Hassid 1963).

Evidence for the direction of the sucrose synthase

reaction includes a comparison to sucrose phosphate

synthetase, tissue specificity, and acceptor concentration

(fructose). Both sucrose synthase and sucrose phosphate

synthetase are present in photosynthetic tissue; however, very

low activities of sucrose synthase have been detected in

leaves (Claussen et al. 1985). In nonphotosynthetic tissue,







14

sucrose synthase is present while sucrose phosphate synthetase

is relatively low in abundance (Avigad 1982). Because of the

differential expression of these enzymes a conclusion was

reached that sucrose synthase activity is high only in

nonphotosynthetic plant tissue (Delmer and Albersheim 1970).

Not much attention has been focused on the role of sucrose

synthase in leaves because of the above conclusion (Claussen

et al. 1985). As discussed, sucrose synthase in

nonphotosynthetic tissue such as storage tissue and sink

organs likely degrades sucrose yielding UDP-glucose for

cellular biosynthesis (Avigad 1982). The direction of the two

reactions in photosynthetic tissue could be either yielding

sucrose in the forward direction or degrading sucrose in the

reverse direction. Synthesis of sucrose by sucrose synthase

depends on the availability of free fructose; fructose is

barely detectable in intact leaves while present in storage

tissues, roots, developing seeds, and exudates (Feingold and

Avigad 1980; Avigad 1982). Fructose in the cytoplasm may

rapidly be phosphorylated by hexokinase or fructokinases

(Turner et al. 1977; Copeland et al. 1978). This

phosphorylated form of fructose will serve as a substrate for

the sucrose phosphate synthetase reaction (UDP- glucose +

Fructose 6-phosphate <-----> sucrose 6-phosphate + UDP + H+)

thus competing with the sucrose synthase reaction leading to

sucrose synthesis (Avigad 1982). The sucrose 6-phosphate

formed from sucrose phosphate synthetase is hydrolyzed to free







15

sucrose by sucrose phosphatase; this irreversible pathway

could account for the large accumulation of sucrose in many

plants (Neufeld and Hassid 1963). In photosynthetic tissue

sucrose synthesis does not appear to be facilitated by sucrose

synthase. Instead sucrose synthase may degrade sucrose and

yield UDP glucose for starch synthesis (Turner and Turner

1957; Neufeld and Hassid 1963; Bucke and Coombs 1974; Jenner

1980; Liu and Shannon 1981). Cobb and Hannah (1988) have shown

that shl kernels grown on reducing sugars contain sucrose;

this is evidence that sucrose synthase is not necessary for

sucrose synthesis and that this enzyme has a role in sucrose

degradation instead.

Plant stress such as anaerobiosis, chilling, and wounding

have been shown to alter sucrose synthase gene expression as

has sucrose concentration. Springer et al. (1986) called the

maize Shl sucrose synthase an anaerobic protein; however,

McElfresh and Chourey (1988) showed that Shl mRNA is induced

by anaerobiosis but the translation of that message is not

induced. Potato sucrose synthase mRNA is also induced by

anaerobiosis (Salanoubat and Belliard 1989). Wheat sucrose

synthase activity increases after chilling shock while sucrose

phosphate synthetase and invertase activity were not altered

(Calderon and Pontis 1985). Wounding in potato resulted in a

decrease in sucrose synthase mRNA levels while increased

sucrose concentrations led to an increase in sucrose synthase

mRNA (Salanoubat and Belliard 1989). Increased sucrose







16

synthase activity in cotton leaves corresponds to increased

sucrose concentrations (Hendrix and Huber 1986).

Sucrose synthase is a tetramer; the subunits are

generally around 90,000 molecular weight (mol wt). The maize

Shl sucrose synthase subunit is 88,000 mol wt (Su and Preiss

1978). The maize Sus sucrose synthase subunit is the same size

as the Shl sucrose synthase subunit and these proteins were

given the size of 87,000 mol wt by Echt and Chourey (1985).

The sucrose synthase proteins from maize were determined to be

360,000 mol wt (Tsai 1974). Sucrose synthase isolated from

wheat germ is 370,000 mol wt while sucrose synthase isolated

from wheat leaf tissue is 380,000 mol wt (Larsen et al. 1985).

Soybean nodules contain a sucrose synthase of 400,000 mol wt

with a subunit mol wt of 90,000 (Morell and Copeland 1985).

The sucrose synthase found in peach fruit is 360,000 mol wt,

made up of identical subunits of 87,000 mol wt (Moriguchi and

Yamaki 1988). Mung bean seedlings have a sucrose synthase of

380,000 mol wt and subunits of 94,000 (Delmer 1972). Rice

seeds have a sucrose synthase of 410,000 mol wt and a subunit

size of 100,000 mol wt (Nomura and Akazawa 1973).

There are only a few examples of sucrose synthase genes

or cDNAs which have been sequenced. Maize is the biological

system most explored not only in sucrose synthase protein but

also for sucrose synthase structural genes and cDNAs. The

sucrose synthase gene of maize, Shi, has been sequenced as

well as the cDNA encoded in that gene. Hence the gene







17

structure of Sh1 was determined (Werr et al. 1985; Sheldon et

al. 1983). Sh has 16 exons with the first and second introns

being 1014 bp and 511 bp respectively. Based upon RNA analysis

the Shl gene is predominantly expressed in the endosperm as

compared to other tissues (Springer et al. 1985). The Shl

sequences are the only published sucrose synthase genomic

sequences (Geiser et al. 1982; Sheldon et al. 1983; Werr et

al. 1985; Zack et al. 1986; Baler and Hannah unpublished).

However, a maize Sus sucrose synthase genomic clone has been

partially sequenced along with a cDNA clone encoded by that

structural gene (McCarty et al. 1986; Shaw and Hannah

unpublished). The Sus gene has only 15 exons; when compared to

the Shl gene structure intron 15 is not present. This sucrose

synthase gene, Sus, is largely constitutively expressed

(Springer et al. 1985; McCarty et al. 1986; Hannah and McCarty

1988).

Besides the two maize sucrose synthase cDNAs a potato

sucrose synthase cDNA has also been sequenced. This potato

cDNA includes the entire protein coding region (Salanoubat and

Belliard 1987). Another dicot sucrose synthase cDNA, from

soybean, has been partially sequenced; this 500 bp cDNA has

similarity with exons 14, 15, and 16 of the maize Shl gene

(Thummler and Verma 1987). Two wheat sucrose synthase cDNAs

have been partially sequenced (Marana et al. 1988). These two

wheat cDNAs including most of exon 13, and exons 14, 15, and

16 were compared to the maize Shl sequence. One of the wheat







18

sucrose synthase cDNAs was more similar to the Shl sequence

than to the other wheat cDNA sequence.

In the work presented here, Pennisetum purpureum sucrose

synthase and the sucrose synthase subunits are shown to be

similar in size to the maize counterparts. Two putative P.

purpureum sucrose synthase genomic sequences were identified

and cloned; one of these genomic clones was partially

sequenced and compared to the Shl and Sus gene structures. Two

nearly full length cDNA clones were selected from a leaf cDNA

library. One of these which is coded for in the partially

sequenced genomic clone was sequenced. The sequenced sucrose

synthase genomic and cDNA clones from P. purpureum are more

homologous to the Sus gene from maize than to the Shl gene;

the other sucrose synthase cDNA appears to have equal

similarity to either of the maize sucrose synthase genes based

upon cross hybridization. An additional gene from P. purpureum

was cloned. It has similarity with the 5' end of the Shl gene

from maize. This unidentified genomic clone may code for an

abundant 1.9 kb mRNA which hybridizes to the 5' end of sucrose

synthase structural genes. Also, a P. glaucum sucrose

synthase cDNA of 2.5 kb was sequenced and shown to have more

similarity to the Sus gene of maize than to the Shl gene. The

sequenced P. purpureum sucrose synthase cDNA clone is compared

to the P. glaucum sucrose synthase cDNA as well as to the

maize sucrose synthase cDNA from the Sus locus.













CHAPTER 2
MATERIALS AND METHODS



DNA Isolation. RNA Isolation. Digestion, and Fractionation



All Pennisetum species were supplied by Dr. W.W. Hanna

(Coastal Plain Exp. Station, Tifton, Georgia) except for P.

purpureum (PI300086) which came from Dr. S.C. Schank

(University of Florida). DNA was isolated from mature leaf

tissue of plants (Rivin et al. 1982). Approximately 10 g to 20

g of leaf tissue was used. The Kpn I monomers used as probe

DNA and for the copy number determination were isolated from

3% polyacrylamide gels by diffussion into TE [10 mM Tris-Cl

(pH 7.8) and 1 mM EDTA (ethylenediamine-tetracetic acid)] over

night at 40 C. Insert DNA from recombinant plasmids was

isolated from agarose gels by electrophoresis into wells

containing gel buffer and 10% glycerol (Maniatis et al. 1981).

Nuclear DNA was isolated from mature leaf tissue

following an unpublished procedure from Dr. Eduardo Vallejos.

Twenty grams of tissue was cut into small pieces with scissors

and immersed into five volumes of ice cold 100 mM Tris-HC1 (pH

8.0), 10 mM EDTA, 500 mM sucrose, and 0.1% mercaptoethanol.

The leaf tissue was homogenized in a Waring blender for 60

seconds. The homogenate was filtered through cheese cloth.

19







20

Triton X-100 was added to a final concentration of 0.5%. The

samples were spun at 2000 rpm for 15 mins at 40 C. The pellet

was resuspended at 1 ml/10 g of tissue in 100 mM Tris-HC1 (pH

8.0) and 20 mM EDTA. An equal volume of the above solution

with 100 ug/ml of lysozyme was added. This suspension was

mixed then incubated at 650 C for 30 mins with occasional

inversion. After incubation the suspension was extracted once

with phenol:chloroform:isoamyl alcohol (50:48:2) and

centrifuged at 5,000 rpm. The supernatant was reextracted with

chloroform:isoamyl alcohol (96:4) and centrifuged. The DNA in

the supernatant was then cesium chloride purified (Maniatis et

al. 1982).

Plasmid minipreparations followed the procedure below.

Fifty ml cultures of cells were grown overnight in LB medium

with 100 ug per ml of ampicillin. The cells were chilled,

centrifuged, and resuspended in 580 ul of 25% sucrose and 50

mM Tris-HCl (pH 7.8). Oak Ridge tubes were kept on ice through

the following steps up to the extractions. A 100 mg/ml

lysozyme solution was made with the above sucrose solution and

200 ul was added to the cell suspension. After five minutes

475 ul of 500 mM EDTA was added. After five minutes 1.25 ml of

0.1% Triton X-100, 50 mM Tris-HCl (pH 7.8), and 10 mM EDTA was

added. After 20 minutes the cell debris was centrifuged out at

15,000 rpm for 30 mins. The supernatent was extracted with

phenol:chloroform:isoamyl alcohol and centifuged at 5,000 rpm

for 5 mins, then reextracted with chloroform isoamyl alcohol.







21
The supernatant was precipitated with 1/10 volume 3 M sodium

acetate and two volumes ethanol. The precipitant was

centrifuged at 10,000 rpm for 15 mins. The pellet was

resuspended in TE and reprecipitated with 0.5 volume of 7.5 M

ammonium acetate and two volumes of ethanol. The precipitant

was pelleted, washed with 70% ethanol, dried, and resuspended

in TE.

Lambda DNA minipreparations involved scraping off the top

agarose from two petri plates exhibiting confluent lysis. The

top agarose was shaken at 100 rpm at room temperature in 25 ml

of SM [50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 8 mM MgSO4, and

0.01% gelatin] for three hours. At this point a standard

procedure was followed (Maniatis et al. 1982).

RNA was isolated from P. purpureum mature leaf tissue, P.

glaucum immature seed, and maize kernel using procedure for

maize (McCarty 1986).

DNA digestions for genomic, pUC 19, and M13 mpl9 samples

were carried out according to the suppliers procedure.

Restriction endonucleases were from Bethesda Research

Laboratories (BRL) except Bst NI which came from New England

Biolabs.

DNAs were fractionated in 3% polyacrylamide along with

123 bp markers (BRL). A Bio-Rad 16 cm gel apparatus was used.

Gel widths were 1.5 cm. The buffer system employed was TBE

(100 mM Tris base, 100 mM Boric acid, and 20 mM EDTA). The 3%

polyacrylamide gels contained 0.1% bis-acrylamide, 0.0005%







22

TEMED, and 0.12% ammonium persulfate. Samples were

electrophoresed at 80 volts from 5 to 7 hours. After

electrophoresis, the DNA was stained with ethidium bromide (5

ug/ml) for 15 minutes. The gels were illuminated with an

Ultra-violet Products, Inc. transilluminator and photographed

with a Polaroid MP-4 Land Camera.

DNAs were fractionated in 0.8% to 1.25% agarose. BRL H4

and H5 gel electrophoresis units were used with TBE buffer.

Samples were electrophoresed at varying voltage depending on

the length of the run and the sample migration distance. After

electrophoresis the gels were stained and photographed as

above.

RNAs were fractionated in 1.2% agarose/formaldehyde gels

at 50 volts using a BRL H4 electrophoresis unit. The MOPS

buffer and procedure is described in Maniatis et al. (1982).



Southern Blotting, RNA Blotting. Probe Labeling, and
Hybridizations



DNA was blotted onto nylon membrane (Amersham Hybond N or

Hybond N+) following Maniatis' procedure (Maniatis et al.

1982) except the blotting was performed on foam rubber in a

tray instead of the wick method. When Hybond N was used, the

transfer buffer was 20X SSPE [3.6 M NaC1, 200 mM NH2PO4 (pH

7.7), and 20 mM EDTA]. After blotting onto Hybond N the DNA

was crosslinked to the nylon by 8 minutes of UV irradiation.







23

When the DNA was blotted onto Hybond N+ the transfer solution

was 400 mM NaOH.

RNA was blotted onto Amersham Hybond. After

electrophoresis the gels were soaked in 20X SSPE for 10

minutes and then blotted identically to a Southern blot using

20X SSPE. This was followed by crosslinking and hybridization.

Two types of hybridization procedures were used. In the

first, the prehybridization solution was 5X Denhardt's, 5X

SSPE, 0.1% SDS (sodium lauryl sulfate), and 150 ug/ml of

denatured and sheared salmon sperm DNA. The hybridization

solution was 2X Denhardt's, 5X SSPE, 0.1% SDS, and 75ug/ml

salmon sperm DNA (Maniatis et al. 1982). In the second

hybridization procedure, the prehybridization and

hybridization solutions were the same: 500 mM Na2HPO4 (pH 7.2),

1% BSA (bovine serum albumin), and 7% SDS (Church and Gilbert

1984). For both procedures prehybridization was approximately

one hour and hybridization extended overnight. The

temperatures for the above procedures were from 600 C to 680

C as indicated in the figure legends. Probe labeling involved

a nick translation kit or a random primer kit (BRL) following

the suppliers procedure. Hybridizations were carried out in a

hot air incubator with rotation. Washing conditions included

a five minute wash at room temperature in 2X SSPE and 0.1% SDS

followed by at least two washes in 0.07X SSPE or 0.1X SSPE and

0.1 % SDS as indicated in the figure legends. Washing







24

temperatures were at the same temperatures as the

hybridizations.

After washing, the filters were dried. The distribution

of alpha [32P]-labeled DNA on the blots was determined by

autoradiography using Kodak X-OMAT AR film and intensifying

screens at -700 C for various lengths of time. The

autoradiographs were developed using a Konica QX-60A

developer.



Genomic and cDNA Cloning



Two genomic libraries and two cDNA libraries were

constructed. The first genomic library utilized the vector

lambda gtwes; the insert DNA was 5 kb to 10 kb Eco RI P.

purpureum (PI300086) DNA. Both vector arms and insert DNA were

isolated from sucrose gradients and cloned using standard

techniques (Maniatis et al. 1982). Promega Corporation

Packagene Extract was used to package the ligation products.

Genomic clones were plated and screened using Amersham

nitrocellulose (Maniatis et al. 1982). The second genomic

library was constructed with 6 kb to 11 kb Bam HI P.

purpureum (PI300086) DNA isolated from an agarose gel and

cloned into EMBL3 prepared according to the suppliers

procedure (Promega Corporation). The same packaging, plating,

and screening procedures were followed except Amersham Hybond

N+ was used. To construct the two cDNA libraries a BRL cDNA







25

synthesis kit was used following the suppliers directions. The

same packaging, plating, and screening procedures were

followed with Amersham Hybond N*.



Sucrose Svnthase Isolation and Antibody Production



Sucrose synthase was isolated by Dr. D.R. McCarty (Echt

and Chourey 1985) for use in antibody production. Isolated

protein was used in eliciting an antibody response in a

rabbit. Anti-Shi antibody staining followed a procedure by

Bio-Rad.



Copy Number Determination


The percentage of the genomes composed of the Kpn I

repetitive sequence families was based upon hybridization of

K&n I monomers to 100 nanograms (ng) of nuclear DNA and

standard amounts of isolated KDn I monomers (5, 10, 15, 20,

25, and 30 ng). A dot blot apparatus was used (Schliester and

Schull). The experiments were replicated three times. The P.

purpureum (PI 300086) dot blots contained a combination of 140

bp and 160 bp KEn I monomers as standards to compare with 100

ng of nuclear DNA; the probe DNA was the same as the DNA used

as the standards. In determining the copy number of the Kpn I

tandemly arrayed repetitive sequence families in P. purpureum

the size of the repetitive element used in the calculation was







26

150 bp. The P. glaucum dot blots were composed of genotype

Tift 18DB nuclear DNA compared to genotype Tift 23 Kvn I

monomers; the probe DNA was the same monomers as the

standards. After hybridization of the probes to the dot blots,

washing, and autoradiography the radioactive areas of the dot

blots were counted in a scintillation counter. Each of the

three experiments for both sets of samples was treated

separately. A standard curve was drawn for each experiment and

the percentage of the genome hybridizing to the probe was

determined from the curve. An average value with a standard

deviation was calculated for both sets of experiments. The

average value was used to determine the copy number.

To determine the copy number from the percentage of the

genome(s) composed of these repeats the following procedure

was used. The average weight of a nucleotide pair is 618

g/mole. In one picogram there are 1.6181 X 10-15 moles. When

the number of moles in one picogram is multiplied by

Avagodro's number (6.022 X 102 nucleotide pairs/mole) the

value for the number of base pairs in one picogram is 9.7443

X 108. This value of base pairs per picogram was used to

determine the size of the genome(s) in base pairs from the

published sizes of the P. purpureum and P. glaucum genomes

which were given in picograms. The size of the genome(s) in

base pairs was multiplied by the percentage of the genome(s)

which hybridized the KEn I probes to identify the number of

base pairs which hybridized the probe. This figure was then







27

divided by the size of the Kpn I monomers to determine the

copy number per cell of these tandemly arrayed repetitive

sequences.



DNA Sequencing



The KDn I repetitive sequence monomers from P. purprueum,

P. alaucum, P. sauamulatum, and P. hohenackeri were sequenced

with the Singer dideoxy method (Sanger et al. 1977). These

monomers were originally cloned into pUC19 for sizing and then

subcloned into M13mpl9. Then M13mpl9 subclones were either

sequenced in both directions or sequenced twice by John Baier.

An IBI Gel Reader was utilized with the subsequent sequences

analyzed with IBI DNA/Protein Sequence Analysis software. The

Kpn I monomers from P. setaceum and P. villosum were sequenced

in both directions directly in pUC19 by Dr. Ernesto Almira at

the ICBR DNA Sequencing Core Facility using a Genesis 2000 DNA

Analysis System. This method involved fluorescent chain-

terminating dideoxynucleotides (Prober et al. 1987).

The two sucrose synthase cDNA clones, pc309 and pcPMSS,

were originally subcloned into pUC19. These Eco RI inserts

were then subcloned into pGEM-7Zf(+) from Promega Corporation.

Deletions of these two clones proceeded in either direction

using the Erase-a-Base System from Promega. Deletions in the

T7 direction were performed with both cDNA clones being double

digested with Aat II (3' overhang for inhibition of







28

exonuclease III) and Xho I. Deletions in the SP6 direction

involved the enzymes Nsi I (3' overhang) and Bam HI. Deletions

of these clones were selected from minipreped plasmid DNA and

sequenced at the ICBR sequencing core in collaboration with

Dr. Ernesto Almira. The gaps in the sequencing data for the

two clones were filled by using 20 base oligonucleotides as

primers to sequence through areas where deleted subclones were

not obtained. The individual strands for the two clones were

compared using IBI DNA/Protein Analysis software.

Approximately 98% of the bases in the two strands aligned. The

most common ambiguous bases resulted from C/T and A/G

mismatches. A common location for ambiguous bases was when an

A or G followed a T which gave a strong signal. The corrected

sequences of pc309 and pcPMSS were compared; bases not

aligning between the two sucrose synthase cDNAs were further

reviewed. Additional sequence comparison between the

Pennisetum clones and maize sucrose synthase cDNAs utilized

the above IBI software.



Isozyme Electrophoresis and Staining



Starch gel electrophoresis, as described by Shields et

al. (1983), was employed to fractionate the isozymes. Protein

extraction was performed with leaf tissue frozen in liquid N2,

the buffer was 0.1M Tris-Cl (pH 7.8), 10mM DTT

(dithiothreitol) and 20% glycerol. The ratio of leaf weight to







29

buffer was 1 gm per ml. Wicks saturated with extract were

stored at -700 C. Two buffer systems were used for gels and

electrodes. The first involves citric acid as the electrode

buffer and histidine as the gel buffer (Flides and Harris

1966); MDH (malate dehydrogenase) was the only isozyme stained

for with this buffer system. The second buffer system was used

for PGI (phosphoglucoisomerase), APS (alkaline phosphatase),

GOT (glutamate oxaloacetate transaminase), and ADH (alcohol

dehydrogenase). This buffer system included boric acid as the

electrode buffer and tris citrate as the gel buffer (Ridgway

et al. 1970). Staining for MDH, PGI, APS, and ADH was

performed as in Vallejos (1983). Staining for GOT is from

Bournival et al. (1989).












CHAPTER 3
RESULTS OF PROGENITOR STUDY



Total DNAs from six P. purpureum genotypes were digested

with Kpn I, electrophoresed in 3% polyacrylamide, and stained

with ethidium bromide (Fig. 1). All six genotypes contain 140

bp and 160 bp Kpn I families of tandemly arrayed repetitive

sequences indicating a species wide pattern. P. purpureum

genotypes N-16 (lane 3) and Merkeron (lane 4) exhibit a

greater proportion of 160 bp Kpn I monomers compared to 140 bp

KDn I monomers. P. purpureum genotype PI300086 has similar

amounts of 160 bp and 140 bp Kpn I monomers (lane 2).

To compare the allotetraploid P. purpureum to additional

Pennisetum species (Fig. 2A), total DNAs from P. hohenackeri

(lane 2), P. squamulatum (lane 3), P. purpureum (lane 4), and

P. glaucum (lane 5) were digested with Kpn I and

electrophoresed as above. P. squamulatum and P. purpureum

share a common digestion pattern, both containing 140 bp and

160 bp repetitive sequence families. The diploid pearl millet

contains a 140 bp Kpn I family while the diploid P.

hohenackeri contains a 160 bp family. P. hohenackeri is not

considered to be a closely related species of P. purpureum

based upon phenotype, isozyme banding patterns (Lagudah and

Hanna 1989), and additional evidence presented in this paper.

30






Fig. 1. Restriction endonuclease digestion pattern of P.
purpureum DNAs. P. purureum total DNAs (7.5 ug)
were digested with Kpn I, electrophoresed in 3%
polyacrylamide, and stained with ethidium bromide.
The 140 bp and 160 bp Kpn I family monomers are
labeled. Lane 1 is 123 bp markers. L. purPure. u
genotypes are PI300086 (lane 2), N16 (lane 3),
Merkeron (lane 4), N166 (lane 5), N137 (lane 6),
and N138 (lane 7).















1 2 3 4 5 6 7


160 bp
140 bp






Fig. 2. Restriction endonuclease digestion pattern of
Pennisetum DNAs. Total DNAs (7.5 ug) from selected
Pennisetum species were digested with Kpn I,
electrophoresed in 3% polyacrylamide, and stained
with ethidium bromide. The 140 bp and 160 bp Egn I
family monomers are labeled. For A, lane 1 is 123
bp markers. Total DNAs are P. hohenackeri (PS156)
(lane 2), P. scuamulatum (PS26) (lane 3), .
purpureum (PI300086) (lane 4), and ,. alaucum Tift
23 (lane 5). For B, lane 1 is 123 bp markers.
Total DNAs are P. purpureum (N16) (lane 2), F.
Purp~reum (N16) X P. glaucum Tift 23A (lane 3),
and P. Qlaucum Tift 23A (lane 4).


























*\234


1 23 4 5













160 bp ....
140 bp


160 bp
S140 bp







35

However, the appearance of the 160 bp eKn I family of P.

hohenackeri indicated further investigation was necessary in

identifying the relationship between the 160 bp Kpn I families

of the species depicted in Fig. 2A.

Fig. 2B displays the Kpn I digestion pattern of P.

purpureum (N16) (lane 2), P. glaucum Tift 23A X P. purpureum

(N16) (lane 3), and P. glaucum Tift 23A (lane 4). In P.

purpureum half of the chromosomes belong to the A' genome

while the other half belong to the B genome. The proposed

progenitor of P. purpureum, P. glaucum (A genome designation),

has a 140 bp Kpn I family; this suggests the 140 bp Kpn I

family of P. purpureum is located within the A' genome. Hence

the 160 bp Kpn I family of P. purpureum may be a marker for

the B genome. This experiment addressed the question of the

Kpn I families being genome markers in P. purpureum. The

triploid interspecific hybrid has the genome designation of

AA'B; the A complement from P. glaucum and the A'B complement

from P. purpureum. The Kpn I digest of the triploid shows a

combination of bands present in both parents. There is a

redistribution of 140 bp and 160 bp Kpn I families as seen by

ethidium bromide staining intensity of the triploid as

compared to P. purpureum. The P. purpureum parent has a

predominance of the 160 bp Kon I monomer over the 140 bp KPn

I monomer. The triploid has a predominance of the 140 bp Kpn

I families as would be expected since 14 of the 21 chromosomes

in the triploid hybrid are either A or A' from P. glaucum or






36

P. purpureum respectively. This redistribution pattern of Kpn

I families indicates that the 140 bp Kpn I family of P.

purpureum may be an A' genome marker while the 160 bp Kpn I

family may be a B genome marker.

To investigate cross hybridization of these Kpn I

families DNA from the Pennisetum species in Fig. 2A were

digested, electrophoresed, blotted, and probed with nick

translated 140 bp Kn I sequences from P. glaucum (Fig. 3).

Total DNAs from P. hohenackeri (lane 1), P. sauamulatum (lane

2), P. purpureum (lane 3), and P. glaucum (lane 4) and nuclear

DNAs from P. purpureum (lane 5) and P. glaucum (lane 6) were

probed. Probing of nuclear DNA showed that the Kpn I families

are of nuclear origin. Stringent washing conditions were used

indicating a strong similarity between these families of

tandemly arrayed repetitive sequences. The 140 bp Kpn I

monomers from P. glaucum used as a probe cross hybridized with

the other 140 bp Kpn I families and also with the 160 bp Kpn

I families. This cross hybridization suggested that these

sequences have a common origin.

Fig. 4 shows Kpn I families of additional genotypes of

species already observed and also additional Pennisetum

species. Total DNAs were digested, electrophoresed in 3%

polyacrylamide, and stained with ethidium bromide. Lane 1 is

a second P. squamulatum genotype, PS24; this genotype compares

with P. squamulatum genotype PS26 seen in Fig. 2A except there

is less DNA hence the repeats are not as visible. Lane 2 and






Fig. 3. Cross hybridization of the KEn I families from the
Pennisetum species shown in Fig. 2. Total DNAs (2
ug) from Pennisetum species were digested with En
I, electrophoresed in 1.75% agarose, Southern
blotted, probed at 650 C with nick translated
pearl millet IB23 140 bp KEn I monomers,and washed
at 650 C in 0.07X SSPE and 0.1% SDS. Concatomers of
three repeats from the 140 bp and 160 bp gn I
families are labeled as 420 bp and 480 bp
respectively. Total DNAs are P. hohenackeri (PS156)
(lane 1), P. suamulatum (PS26), (lane 2), P.
purpureu (PI300086) (lane 3), and EP alaucum Tift
23 (lane 4). Nuclear DNAs are E p.urpureum
(PI300086) (lane 5) and P. alaucum Tift 18DB (lane
6).








1 2 3 4


480 bp
420 bp


I


5 6




I

I


::b


U;
m

Si
S)


S
Ul


w






Fig. 4. Restriction endonuclease digestion pattern of
additional Pennisetum DNAs. Total DNAs were digested
with jKn I, electrophoresed in 3% polyacrylamide,
and stained with ethidium bromide. All samples were
7.5 ug except P. scuamulatum which was 2.5 ug. The
140 bp and 160 bp En I family monomers are labeled.
Total DNAs are P. scuamulatum (PS24) (lane 1), P.
glaucum ssp. monodii (PS470) (lane 2), P_. laucum
ssp. monodii (PS34) (lane3), P_ setaceum (PS49)
(lane 4), P. setaceu (PS247) (lane 5), and P_
villosum (PS249) (lane 6). Lane 7 is 123 bp markers.














1 2 3 4 5 6 7


160 bp
140 bp ---







41

3 are two genotypes of P. glaucum ssp. monodii. These are

similar in digestion pattern to the pearl millet genotypes

seen in Fig. 2. In lanes 4 and 5 are two genotypes of P.

setaceum (Forsk.) Chiov. P. setaceum, a polymorphic species

often characterized as a triploid, also has a 160 bp Kvn I

family of repeated DNA. P. villosum R. Br. ex Fresen. is seen

in lane 6, this polymorphic species has an intensely staining

160 bp Kpn I family present in the digest. P. villosum

pachytene chromosomes when stained with acetocarmine or aceto-

orcein exhibit terminal heterochromatic knobs and centromeric

bands (Juahar 1981). These large heterochromatin regions may

contain the intensely staining 160 bp Kpn I band seen in lane

6.

In Fig. 5 total DNA from 11 Pennisetum species and

interspecific hybrids was digested with Kpn I,

electrophoresed, Southern blotted, and probed with 140 bp Kpn

sequences from P. glaucum. P. glaucum ssp. monodii is in lane

1; no difference has been observed between the repetitive

sequence families of P. glaucum ssp. monodii and pearl millet

as shown by hybridization or electrophoresis in 3%

polyacrylamide, including the appearance of a 170 bp Kpn I

family visible only by hybridization. Lanes 2-4 are the

genotypes seen in Fig. 2B, P. purpureum (N16), P. Qlaucum Tift

23A, and the P. glaucum X P. purpureum triploid respectively.

A redistribution of the Kpn I families at the 420 bp and 480

bp positions in the triploid when compared to the P. purpureum






Fig. 5. Cross hybridization of the Kpn I families
of additional Pennisetum species and interspecific
hybrids. Total DNAs from Pennisetum species and
interspecific hybrids were digested with KZn I,
electrophoresed in 1.75% agarose, Southern blotted,
probed at 650 C with random primer labeled pearl
millet IB23 140 bp Kpn I monomers, and washed at 650
C in 0.07X SSPE and 0.1% SDS. All samples were 2 ug
except oriental which was 6 ug. Concatomers of
three repeats from the 140 bp and 160 bp Kpn I
families are labeled 420 bp and 480 bp respectively.
Total DNAs are P. qlaucum ssp. monodii (PS34) (lane
1), P. urpureu (N16) (lane 2), alaucum Tift 23A
(lane 3), E. laucum Tift 23A X purpureum (N16)
(lane 4), EP alaucum (H165) X P. sauamulatum (PS262)
(lane 5), P. purpureum Merkeron X P sauamulatum
(PS262) (lane 6), EP sauamulatum (PS26) (lane 7), E.
setaceum (PS247) (lane 8), P. villosum (PS249) (lane
9), P. orietale (PS15) (lane 10), and P. flaccidum
(PS396) (lane 11).












1 2 3 4 5 6 7
1234567





- cf *8 *





S18.8 1
n gll/


8 9 10 11






400
00"
,-,8


ma.. A&


ww


480 bp
420 bp


pwo


Wwoq p -C- _







44

The triploid in lane 4 has more hybridization of the probe at

420 bp versus 480 bp as expected because there are twice as

many A\A' chromosomes as B chromosomes. A P. glaucum (H165) X

P. scuamulatum (PS262) hybrid is shown in lane 5. The

additional A chromosomes from P. glaucum have combined with

the chromosomes from P. squamulatum to increase the amount of

the 140 bp KDn I family as compared to the 160 bp KDn I

family; this is indicated by the hybridization intensity of

the probe at the 420 bp region versus the 480 bp region. Lane

6 is an interspecific hybrid between P. purpureum Merkeron X

P. squamulatum (PS262). P. squamulatum (PS26) is in lane 7.

The Southern blot resulting in the autoradiograph in Fig. 5

was exposed for two periods of time to observe comparable

amounts of hybridization between lanes 1-7 and lanes 8-11.

Lanes 8-11 represent a longer exposure. P. setaceum (PS247) is

shown in lane 8. Cross hybridization between the pearl millet

probe and P. setaceum as well as the rest of the Pennisetum

species indicates similarity between these KPn I families. P.

villosum (PS249) was characterized in Fig 4 as having an

intensely staining 160 bp KDn I family; in lane 9 this

Pennisetum has strong hybridization signals further suggesting

a large proportion of the P. villosum genome is comprised of

this repeat family. The quantity of P. orientale L.C.Rich. DNA

in lane 10 is three times the amount of DNA in the other

lanes; hence, there is less of the 160 bp Kpn I family than is

indicated by the hybridizational intensity. Also present in P.







45

orientale is a tandem repeat slightly larger than the band

seen at 480 bp. Lane 11 is P. flaccidum Griseb., the bands of

hybridization are barely visible.

The appearance of the 160 bp Kpn I family of tandemly

arrayed repetitive sequences in most of the Pennisetum species

characterized here indicates that this repetitive sequence

family is common within the genus. Of all the Pennisetum

species observed, only P. glaucum (pearl millet and ssp.

monodii) does not have a detectable 160 bp Kpn I family. The

140 bp Kpn I family is present in the closely related species

P. purpureum, P. squamulatum, and P. glaucum (including pearl

millet and ssp. monodii). P. alaucum and P. purpureum are in

the Pennisetum section Penicillaria while P. squamulatum is in

the section Heterostachya (Jauhar 1981). Additional Pennisetum

species, especially in the section Penicillaria, may contain

140 bp Kpn I families. Over all, species characterized in this

paper represent four of the five sections within the genus

Pennisetum; all of these sections have species containing the

160 bp Kpn I family of repeats. Moreover, the species

containing these repeat families are from the X=7 and the X=9

groups of the genus.

A Hind III family of repetitive sequences was identified

in P. squamulatum. The family is not apparent in P. purpureum

(Fig. 6). Hind III was used to digest total DNA from two

genotypes each of P. squamulatum and P. purpureum followed by

electrophoresis in 3% polyacrylamide and visualization with







Fig. 6. Restriction endonuclease digestion pattern of
two Pennisetum species. Total DNAs (7.5 ug) were
digested with Hind III, electrophoresed in 3%
polyacrylamide, and stained with ethidium bromide.
The 275 bp Hind III repeats are labeled. Total DNAs
are P. squamulatum (PS24) (lane 1), P. souamulatum
(PS26) (lane 2), P. urpureu (PI300086) (lane 3),
and P. Durureum (N16) (lane 4). Lane 5 is 123 bp
markers.












1 2 3 4 5


275 bp --







48

staining. The repeat appears to exist only in monomeric form

as indicated by the lack of tandem arrays seen in the

fractionated DNA. The presence of this repeat suggests that P.

squamulatum contains genetic information not present in P.

purpureum. An alternative possibility is that P. purpureum may

have this Hind III family in an undigestible form as cryptic

DNA.

To further elucidate the tandem arrangement of the Kpn I

families, DNAs from P. glaucum (Fig. 7, Lanes 1-3), P.

purpureum (lanes 4-6), P. squamulatum (lanes 7-9), and P.

hohenackeri (lanes 10-12) were digested for 2, 5, and 60

minutes. The DNAs were then electrophoresed, Southern blotted,

and probed with random primer labeled pearl millet 140 bp Kpn

I sequences. As digestion proceeds for each sample there was

a decrease in the larger concatomers of repeats and an

increase in the smaller concatomers. These results demonstrate

the tandem repeat arrangement of these families.

In Fig. 8 methylation of the P. purpureum Kpn I families

of repeats is demonstrated. P. purpureum DNA was digested with

Eco RII (lane 1), then redigested with Eco RII (lane 2) to

attempt to completely digest the DNA. In lane 3 P. purpureum

DNA was digested with the Eco RII methylation insensitive

isoschizimer Bst NI and then redigested (lane 4). These

digests were electrophoresed, blotted and probed with random

primer labeled pearl millet 140 bp KDn I sequences. As can be

seen, the methylation sensitive Eco RII did not cleave the KPn







Fig. 7. Incomplete and complete DNA digestions from selected
Pennisetum species. Total DNAs were digested for 2
min, 5 min, and 60 min with Kpn I, electrophoresed
in 1.25% agarose, Southern blotted, probed at 650 C
with pearl millet IB23 140 bp gn I monomers, and
washed at 650 C in 0.1X SSPE and 0.1% SDS. All
samples were 2 ug; aliquots for each species came
from the same reaction mixture. Concatomers of two
repeats from the 140 bp and 160 bp ~En I families
are labeled 280 bp and 320 bp respectively. Lanes 1
to 3 are glaucum Tift 23 (2 min, 5 min, and 60
min respectively). Lanes 4 to 6 are E. purpureum
(PI300086) (2 min, 5 min, and 60 min respectively).
Lanes 7 to 9 are P. sauamulatum (PS26) (2 min, 5
min, and 60 min respectively). Lanes 10 to 12 are P.
hohenackeri (PS156) (2 min, 5 min, and 60 min
respectively).











1 2 3 4 5 6 7 8 91011 12


320 bp ,
280 bp


me04a






Fig. 8. Hybridization pattern of P. purureum DNA digested
with isoschizomers to detect methylation of eKn I
families. Total DNA from P purpureum (PI300086) was
digested with either Eco RII or Bst NI,
electrophoresed in 2% agarose, Southern blotted,
probed at 650 C with nick translated P. qlaucum IB23
140 bp Kpn I monomers, and washed at 650 C in 0.07X
SSPE and 0.1% SDS. All lanes contained 2 ug of DNA.
Bands of hybridization which migrate at 560 bp and
980 bp are labeled. Lane 1 is Eco RII digested DNA.
Lane 2 is DNA digested along with the DNA in lane 1,
extracted, precipitated, and redigested with Eco
RII. Lane 3 is Bst NI digested DNA. Lane 4 is DNA
digested along with the DNA in lane 3, extracted,
precipitated, and redigested with Bst NI.











1 2


S8980 bp




560 bp


34







53

I repeat families to the extent that Bst NI did. The

restriction endonuclease isoschizomers used here test for

methylation at the trinucleotide CNG, N being any nucleotide.

This experiment indicates that the KDn I families of tandemly

arrayed repetitive sequences within the P. purpureum nuclear

genomes are heavily methylated.

The percentage of the purpureum genotype PI300086

(Fig. 9, panel A) and P. glaucum Tift 23 (Fig. 9, panel B)

genomes comprised of the Kpn I repeat families was determined

along with copy numbers per cell. Various amounts of either P.

purpureum 140 bp and 160 bp Kpn I monomers or P. glaucum 140

bp KEn I monomers were used to create a standard curve. The

nuclear DNA of either species was then compared to the

standard curve created for each of the species. A dot blot

apparatus was used to facilitate the binding of samples to the

membrane for probing. The dot blot for the P. purpureum

samples was probed with 140 bp and 160 bp Kpn I monomers from

P. purpureum. The dot blot for the P. glaucum samples was

probed with 140 bp Kpn I monomers from pearl millet. For

napiergrass and pearl millet, 100 nanograms of nuclear DNA

(lane 1) was compared to various amounts of monomers; the

various amounts were 5, 10, 15, 20, 25, and 30 nanograms in

lanes 2-7 respectively. For P. purpureum genotype PI300086,

16.4 +/- 1.5 percent of the nuclear genomes are composed of

Kpn I repetitive element hybridizing sequences. The P. glaucum

genotype Tift 23 has 19.9 +/- 0.8 percent of the nuclear























































00 v
tna to- )0
0 00 -4 r.r
0 1 to 0 0
trz -4 V A I


:44 0 s 10 C c
ON >4c ,4a -. ,.-
r -*E*4 S4


a) ta 0t
tP> 4 14 44 Z W 4

Aq 0 i
04.




A



'.40 4 g1 4.



$44 44 c
0 to 4
cu













00

U 444
4' 0.9












rz.- I0S0
(a C O g





















40 z z 041r41
A C
41 0,0~ c0"
A 44 A rr
14 C- 144 a. 0 4 0


Q) r 4) 0 e


r- Ul ?A) c
f-4 1040 r C



to in ii

U~00 r- 0 r
C4 0 triv 90
COC 010V r4













NS







C,
Cla



CY0







56

genome composed of KDn I repetitive element hybridizing

sequences. Reassociation kinetics have indicated that about

20% of the pearl millet genome is comprised of the highly

repetitive class of sequences (Patankar et al. 1985); the 140

bp &pn I tandemly arrayed repetitive sequences are thus the

major highly repetitive DNA sequences in P. glaucum. The KDn

I families of P. purpureum are likely the major repetitive

sequence families within the nuclear genomes as well. Based

upon a mole of nucleotide pairs averaging 618 grams, one

picomole of DNA equals 9.74 X 108 bp. P. purpureum nuclear DNA

is approximately 5.78 picograms (Taylor and Vasil 1987). Thus

there are 6.2 X 106 copies of &Kn I repeats per cell. P.

glaucum, which contains approximately 4.9 picograms per cell

(Patankar et al. 1985), has 6.8 X 106 copies of Kpn I repeats

per cell.

To compare the average sequences (Grellet et al. 1986) of

the 140 bp and 160 bp Kpn I families of P. purpureum

(PI300086), P. glaucum Tift 23, P. squamulatum (PS26), and P.

hohenackeri (PS156), either four or five individual KDn I

monomers isolated from individual bands from a 3%

polyacrylamide gel were sequenced. The 140 bp Kpn I average

sequences from P. squamulatum and P. glaucum are compared to

the P. purpureum average sequence in Fig. 10. This is sequence

data from one individual plant of each species. Where there is

an n in the P. purpureum or P. squamulatum average sequences

the bases of two of the four sequences at that position were






Fig. 10.


Comparison of the average sequence (Grellet et al.
1986) of P. urureum (PI300086) to the average
sequences of P. suamulatum (PS26) and P. glaucum
Tift 23. Average sequences of P. purpureum and P.
squamulatum are composed of sequences of four
individual Kpn I family monomers. The P. alaucum
average sequence is composed of five individual
sequences. The three sequences are aligned for
comparison purposes. The P. squamulatum and P.
glaucum sequences are compared to the P. urpureum
average sequence. Differences between the sequences
are designated by an asterisk above the position of
the base difference in the sequence being compared
to the P. purpureum average sequence. In the P.
purpureum and P. suamulatum average sequences the
appearance of an n is due to their being two bases
equally represented at that position; if one of
these two bases is the same as the base at the
identical position in a sequence being compared then
it is not considered a difference between the two
sequences.






























Comparison Of Average 140 bp Ken I Sequences

1 10 20 30 40 50 60

. purureumn

GGTACCCCGAAATAGTGCATTCAGGnCCGAAACACAAGTTTTGCATCTTTITACGTGCCG
C
T
. sauamulatum

GGTACCCCGAAAtAGTGCATTCAGGCCCGAAACACAAGTTTTGCATCTTTTTACGTGCCn
A
G
P. laucum

GGTACCCCGAAATAGTGCATTCAGGCCCGAAACAAGTTTTGCATCTTTTTACCTGCCG



70 80 90 100 110 120

E Durpureum

AAGGTTAGCGAAATGCTCCGAAACACTCCCAACATCATTTTGGTTCCAATGGAGTAGAA


E. sauamulatum
*
AAGGTTAGCnAAATGCTCnGAAACACTCCCAAACATCATTTTGGGTCCAATGGAGTAGAA
A C
G T
P. alaucum

AAGGTTAGCGAAATGCTCCGAAACACTCCCAAACATCATTTTGGGTCCAATGGACTAGAA



130

. Durpureum

TGGATGCTTCGCAACTTT


P. sauamulatum

TGGATGCTTCACAACTTn
C
T
STGGlaucum

TGGATGCTTCAAAAACTTC







59

one particular base and the bases in the other two sequences

at that position were of another particular base. The

asterisks in Fig. 10 represent bases in the aligned sequences

that differ from the P. purpureum average sequence. If there

is an n at one of the positions in the P. purpureum average

sequence or the other aligned sequences, and one of the two

bases underneath the n is in common, then this is not

considered a difference in the comparison.

The individual sequences used in this comparison among

these three species varied in length, sequence, and AT

content. The average sequences are depicted as 138 bp in

length. But these Kpn I families are termed 140 bp families

because individual fragments vary in size. The P. purpureum

average sequence is 55.4% AT with individual sequences

differing from the average by 4.3% to 8.0%. The average

sequence for P. squamulatum is 54.3% AT with individual

sequences differing from the average by 1.5% to 6.5%. The P.

glaucum average sequence is 54.3% AT with individual sequences

differing from the average by 8.0% to 10.19%.

The sequences in Fig. 10 which are compared to the P.

purpureum average sequence exhibit close similarity. There are

only two positions which distinguish the P. souamulatum 140 bp

Kpn I average sequence from that of P. Durpureum. The P.

qlaucum average sequence varies from the P. purpureum average

sequence at 4 positions; two of these four differences are the

same differences as that of P. sauamulatum. The conclusion







60

from these data is that the P. purpureum, the P. sauamulatum,

and the P. qlaucum 140 bp KDn I families are closely related

to each other.

The average 160 bp KEn I sequences of P. squamulatum and

P. hohenackeri are compared to the average sequence of P.

purpureum in Fig. 11. The major difference between the 140 bp

and 160 bp Kpn I families is an 18 bp segment located between

base numbers 130 and 147. This segment of the sequence is

underlined. A couple of individual 160 bp KEn I fragments from

both P. setaceum and P. villosum were sequenced as well to

determine how similar these repeat families are to the 160 bp

KEn I families of the above species. The P. setaceum and P.

villosum sequences are not average sequences; hence they are

not compared to the P. purpureum average sequence. Differences

between the two individual sequences for these two species are

denoted by an n; the actual bases observed are listed below

the n. The format up of Fig. 11 is similar to that of Fig. 10.

The variable position in the P. purpureum average sequence at

base number 48 was different in all four individual fragments

sequenced.

As in the case of the 140 bp KDn I fragments, the 160 bp

Kpn I individual fragments varied from each other in length,

sequence, and AT content. The AT content of the P. purpureum

average sequence is 56.1% while the individual sequences

varied from the average by 4.5% to 14.1%. The P. squamulatum

average sequence is 55.8% AT while the individual sequences






Fig. 11. Comparison of the average sequence (Grellet et
al. 1986) of P. ourureum (PI300086) to the
average sequences of E. scuamulatum (PS26) and E.
hohenackeri (PS156). All average sequences are
composed of four individual KEn I family monomers.
Also shown in this figure are sequences from two
individual Epn I family monomers from P. setaceua
(PS247) and EP. villosum (PS249); differences
between these two sequences are represented as an
n with the two bases below the n being from the
two sequences. The sequences are aligned so that
the insertion and the deletion present in the P.
hohenackeri average sequence can be compared to
the other average sequences. Differences between
the average sequences are designated by an
asterisk above the position of the base difference
in the average sequence being compared to the P.
purpureum average sequence. In the average
sequences the appearance of an n is due to their
being two bases equally represented at that
position; if one of these two bases is the same as
the base at the identical position in a sequence
being compared then it is not considered a
difference between the two sequences. All four
individual sequences were different at position 48
in the e purpureum average sequence. Bases from
position 130 to 148 are underlined in all six
sequences. This 18 bp underlined region represents
the major difference between the 140 bp and 160 bp
En I family sequences.









62



Comparison Of Average 160 bp Kpn I Sequences

10 20 30 40 50 60

Pe ourpureum

GGTACCCCGAAATAGTGCATTCAGGTCCGAAACACAAGATTTGCAGCnTTTTACnTGCCG
A
G
P. sauamulatum
** *
GGTACCCCnTAATAGTGnATTCAnGCTCGAAACACAAGTTTTGCATCTTTTTACATGCCn
A C A A
G G G G
F. hohenackeri
** *
GGTACCCCGAAATAGTGCATTCAGGCTCGAAACACAAGTTTTGCATCGTTTTACnTGCCG
A
G
. setaceum

GGTACCCCnAAATAnTGCATTCAnACnCGAAACnCAAGTTTTGCATnATTTTACGTGCCG
G A G C G C
A G A T A T
Svillosum

GGTACCnnGAAATnGTGCATTnAGGCCCGAAnCAnnAGTTTTGnnnCnTTTnACnTGCnG
TT A T A AA CCG T G A G
CC T C G TT TGT G T G C

70 80 90 100 110 120

P ourpureum

AAGGTTTGCGAAnTGCTCCnAAACACTCCCAAACATCATTTTGGGTCnAATGGAGTAGAA
A A C
T G T
. smuamulatum

AAGGTTAGCGAAATGCTCCnAAACACTACCAAACATCATTTTGGGTCTAATGGAGTAGAA
A
G
P. hohenackeri

AAGGTTn CGAAATGCTCCGAAGCACnGCCAAACATCGTTTTGGGTCTAATGGAGTAGAA
C A
T T
P. setaceum

AAGGTGTGCGAAATnCTCCGAAACACTCCCAnACAATnTTTTGGGTCCAATnGAGTAGAn
G C G T A
A A A G
P. villosum

AAGnnnTGnGAAATGCTCnGAAACACTnnCAAACATCATTnnGGGTCTAATGGnGTnnAA
CTA C T AT TG G G-
GCT G C TC CC A AC

130 140 150

P ouroureumn

TGGATGCTTTTGTTGCGAAAnCATTTTCnCAACTTC
C G
T T
E_ sauamulatum

TGGATCTTTTGTTnGAAACCATTTTCnCAACTTC
C A
T G
. hohenackeri

CTGnATGCTTTCGTTGCGAAACCATTTTCGnnnnTTC
C A---
G CAAn
P. setaceum

nTGGnTGTTTTCGTTnnGAAACCATTTTCGCAACTTC
A CT
- A TC
F. villosum

TGGnTGATTTnnnTnCGAAACCATTTTCnCnACTTn
T TGT G AT C
G CCC A GA T






63

varied from the average by 3.8% to 13.5%. The AT content for

the P. hohenackeri average sequence is 52.9% while the

individual sequences varied from the average by 1.3% to 13.5%.

The AT contents for the two P. setaceum individual sequences

are 56.1% and 59.4%; these two sequences differ from each

other by 9.6%. The AT content for the two P. villosum

individual sequences are 58.3% and 53.8% while the sequences

differ by 21.8%.

Comparison of the average sequences of P. squamulatum and

P. hohenackeri to P. purpureum indicates that the 160 bp Kpn

I family of P. sauamulatum is the more closely related. As

seen in the aligned average sequence comparison, P. purpureum

differed from P. sauamulatum by nine bases. These are

indicated by asterisk. The P. hohenackeri average sequence

differs from the P. purpureum average sequence at 15 positions

including a deletion at base 68 and an insertion at base 121.

P. hohenackeri had two individual sequences with an A at

position 151 without any bases corresponding to positions 152-

154 as compared to the P. purpureum average sequence; the

other two individual sequences had CAAC and CAAG at positions

151-154. All four P. hohenackeri individual fragments

sequenced were included in the average sequence. These

sequence data suggest that the P. purpureum (PI300086) and P.

squamulatum (PS26) Kpn I families diverged after the

divergence from the P. hohenackeri (PS156) Kpn I family.






64

Grellet et al. (1986) have constructed an average

sequence of a radish (Raphanus sativus) tandemly arrayed

repetitive sequence family of 177 bp. This sequence was shown

to have three subrepeats of about 60 bp; it was also shown to

have limited similarity to other plant tandem repeats and

animal alphoid sequences. Subrepeats within the P. purpureum

average 160 bp KDn I sequence were observed (Fig. 12). These

subrepeats are aligned according to conserved regions within

the individual subrepeats. The base numbers of the subrepeats

are relative to the average 160 bp KEn I sequence of P.

purpureum in Fig. 11. There are five conserved regions of

three bases or more indicated by the solid lines underneath

the individual subrepeats. As can be observed, these

subrepeats may have undergone rearrangements involving

insertions, deletions, and substitutions. Similarity between

the P. purpureum 160 bp Kpn I subrepeats and a maize satellite

DNA monomer (Peacock et al. 1981) is depicted in Fig. 12 as

well. Four of the five conserved regions in the P. purpureum

subrepeats are also present in the segment of the maize

monomer below the subrepeats. Those regions conserved between

the P. purpureum and maize sequences are underlined below the

maize sequence. When maize DNA was digested with Kpn I,

electrophoresed, Southern blotted, and probed with pearl

millet 140 bp Kpn I sequences no hybridization was detected.

Also probed with pearl millet 140 bp Kpn I monomers was

sorghum and several members of the Saccharum complex; again,






Fig. 12.


Subrepeats within the P. purpureum average 160
bp Kpn I sequence were aligned according to
conserved regions present in these subrepeats. These
conserved regions present in the subrepeats are
underlined below each subrepeat. Below the EP
purpureum subrepeats is a maize satellite DNA
monomer segment (Peacock et al. 1981). The regions
of similarity between the P., urpureun
subrepeats and the maize satellite DNA monomer
segment are underlined below the maize sequence. The
maize sequence is aligned such that the positioning
of the conserved regions between the species can be
compared to the conserved regions from the
subrepeats.
























Aligned Subrepeats Within the Pennisetum purpureum Average
160 bp Kon I Sequence; Similarity to a Maize Repeat is
Also Shown



Bases 85-142

ACTCCCAAA CATCATTTTGGGT Cn AATGGAGTA GAATGGATGCTTTT GTTGCGAAAn
C C
T T

Bases 143-33

CATTTTCnCAACTTC GGTACCCCGAAATAGTGCATTCAGGTCCGAAAC
G
T

Bases 34-84

ACAAGATTTGCAGCnTTTTACnTGCCGAA GGTTTGC GAAn TGC TCCnAAAC
A A A
G T G





Maize Satellite DNA Monomer Segment Bases 68-139

ACACCTACGGATTTTTGACCAAGAAATGGTCTCCACCAGAAAATCCAA


GAATGTGATCTAGGCAAGGAAAC






67

no hybridization was detected indicating a lack of KDn I

hybridizing sequences within the genomes of these species

(data not shown). The P. purpureum average 160 bp KDn I

sequence is composed of subrepeats; the conserved regions

within these subrepeats are also present in a maize satellite

DNA monomer. The P. purpureum and maize repeats in Fig. 12 may

have had a common origin.

The origin of plant tandemly arrayed repetitive sequences

are transfer RNAs according to Benslimane et al. (1986). The

repeats may have arisen from a tRNA progenitor by reverse

transcription followed by rolling circle replication. When a

GenBank search was conducted with the entire P. purpureum

average 160 bp Kpn I sequence the strongest similarity was

found with a putative tobacco chloroplast tRNAArg gene (Deno

and Sugiura 1984). To determine if the subrepeats of the P.

purpureum average 160 bp Kpn I sequence have similarity to

tRNAs or any other sequence a GenBank search was performed

with bases 85 to 142 from the P. purpureum average 160 bp Kpn

I sequence (the top subrepeat in Fig. 12). This GenBank search

detected sequences which are similar to only part of the 160

bp Kpn I average sequence represented as a subrepeat verses

the entire average sequence. The result of the GenBank search

with the subrepeat differed from the GenBank search with the

entire 160 bp KEn I average sequence. The strongest similarity

was found to a maize mitochondrial 5 kb repeat (Houchins et

al. 1986) with no close similarity to any tRNAs. For both













Sc o a 0 4j r r r. r a 0
0 W1 0 0W m c 41 W1 .Go 41 we 0 0
44 r A (a 0 o 1 k o to 39C v4--
o,-. to -.4 r. r o44) x 4 a) m m ) 1 01
4.o wmwmA~joooa4;;mwwc
(0 4i4) d&: 0 C r-4 S.4 >i MA C 0
0) 0 940 Cl 0, t 0- 4 l ~ r- l ,4 -P4
04-4 4) 41 0 4)4 0 0 0-4 to u41 4
4) r. 0 O O W H M MM 41 0) -,4 -r4
14 0 H -- U A f t
MdiU d) 0 X0 : 1-4 -0 0 to to 00
k O o trO Q~n .T' OEO .'4r t

A1Q~C Q1
to to0 0 41 4j.4J
0 0'44 11 U 0 4) 0 to*
41to4)o>.4)AC a 0C
0 X: X: 1 0.0 1 j 4
~,oww~4 WCUlEl..r

O.0El4J *Op4~ L4uw, 0I~ro C3o T *)jUSa
4 E- 4-) 0 U J
-4j~ -4 0 9 .

-4 mowoc0
41G 3 1 4) to c C IA (A w
M C% 4 01 El0WC 34 0 0 go
v4 El .0 41 00 El 0El a 4
V'4 to a .
:jC I ( 141 4 0 04


0 0 r C) a 4J -4 W to A
Cn r' -,4 'A ~




P4 go 4 .0 41
r l-4 0 0 P4 0
10.0 9)- U 0 0 04 V0 414
00, 4.1 4) 0 0 M 4J V ja
to 0 W 4 C a r- (A V

U 0 S--44% :1 #a 4)
> q- 41 4.) $4 -e4
to 4 W 'a 0 F. M t
-4 tv c 41 am) 0 C i to 0 -,4
2 to rq to 1 r 0 $4 -P
( QRV ) 4).z C X 0 0kr
0 q) 4+r ) -r4 X 4) 0 m r. ou 4j 41 >4

a 0 & r,. -r4 'A r,4 OC 0 4j 0 0.
04C 04 to to0 V $4 M I w W

4) to1 0 4 0) V 4) 0 0 H (a W 14k
4 0 0404 0 > 0S o k 0 01-4 9094 0 >% 0r
E-4 A 4J A tP4J 4 0 W "0 (A 0-4 #0 -A ) W 040




t;r
.r4























t)


U

0)
4 )
00




O4
0 0
Q O
0 c








Q) 0








0
U 0

S0
3
CO


O
O
0
3 H
0)
10


IE4


E-4
o I4


XU
CO
o




S E-
U






S *


C
0


0
Co *


E-4










U










4J
m











40
41










0


U
4) E-





O 1


2U
O
40 A
a4 E


4C 0
10
4
2


I
o I
0 1 a
EI *
UI E*


o I
0






O I


E-4
E- *



o E-





E-4c
o -
r0 I*
E- 4c
o u


E-U* U




0 IU U
E-i *E44
a 0


C,.
I
XEs


0)
m >*















0a
a


4
0)









&







70

GenBank searches FASTA was used (Pearson and Lipman 1988) with

a K-tuple size of three.

A P. purpureum subrepeat (from base 85 to 142 of the

average sequence) is compared to the tRNAArU sequence mentioned

above in Fig. 13. The asterisk between the two sequences

indicate bases in common. Thirty two of the 57 bases

comprising the P. purpureum subrepeat in Fig. 13 agree with

the bases in the tobacco tRNA coding region; this amounts to

a 56.1% conservation of bases. The subrepeat including bases

85 to 142 of the P. purpureum average sequence when aligned to

the tRNA has two regions of the possible tRNA progenitor

missing. The first missing region is the T loop side of the

anticodon loop (bases 49-63) seen as a gap represented with

dashed lines in the subrepeat sequence; the second is the T

loop side of the aminoacyl stem which would extend past the B

box region of the subrepeat sequence.

Part of the conserved regions include sequences having

homology to the RNA polymerase III internal promoter elements:

A box and B box. The region of the P. purpureum subrepeat

having homology to the A box has a dashed line above it; this

region is also compared to the RNA polymerase III A box

consensus sequence (Joyce et al. 1988) below the sequence

comparison. The A box in the P. purpureum subrepeat has

homology to 7 out of 7 conserved bases within the RNA

polymerase III A box consensus sequence. The B box in the P.

purpureum subrepeat (also indicated by a dashed line above the







71
subrepeat sequence) has homology to 8 out of 8 conserved bases

within the RNA polymerase III B box consensus sequence with an

insertion of one base, this is shown below the sequence

comparison. It appears that the subrepeats composing the Kvn

I families of _P purpureum and the other Pennisetum species

characterized arose from a tRNA progenitor; the presence of

the transcriptional control elements within these subrepeats

is fairly conclusive. The maize satellite DNA segment in Fig.

12 also has regions similar to the A box and B box, but with

limited sequence identity.

Ribosomal DNA RFLPs are depicted in Fig. 14. Total DNA

from seven Pennisetum species and four interspecific hybrids

(Fig. 14A) were digested with Sst I, electrophoresed, blotted,

and probed with a maize rDNA repeat containing the coding

regions for the 17s, 5.8s, and 26s rRNA genes (McMullen et al.

1986). The variable bands are interpreted as differences in

the size of the intergenic spacer (IGS) region of the rDNA

repeats. The data are compatible with the idea that the size

of the variable band is due to the number of subrepeats within

the IGS. This figure (14A) displays the rDNA RFLPs within and

between Pennisetum species with an emphasis on the

similarities between P. purpureum and P. squamulatum

hybridization patterns. Lanes 2-7 are six genotypes of P.

purpureum; rDNA repeat length heterogeneity is seen only in

genotype N16. All of the P. purpureum genotypes have a rDNA

repeat size of 9.1 kb when all three major bands are totaled;














- 0 04M) W W-4 0 a 0 O 0 0 r. O 0I I oc R
4-i --4 1I A -10 0
~4.4 '"4 --q 0 jr-4G MuIr W N .0 Q-rq4) a (
.,4 41 0 v v m 0 z rq 0 :1 r-4 r-4 r-I f-4 C 0 V C: r-4 r-4 r-4
001 r a u x ` w-%M4 ad 0 20 -~0
t.)o (U 4 -U E-4 W to MJto
4 4 41d 3)
(D ~ xP a 0 0 4jo ~r
4) c to 4) 4j r tn 0 a, 44
4.) C 41 4.) 01-4 0 $4 4) 0 to r .
.r' 4) ) :1 u 4J4J 0 0 ---


4 3t 2 w'G w ID -t-I 0 0 4 awm -
Z % at 04
to 404c 4
to 4 441 N 136 4) :3 to

l 4 o3 0 0 to 4)





I-) A 13. 41 4)4)4 4
0 to1 '0
4) '44b 0 w t9ro ad aat c
41 a (o0 w o D (
to4 Cn 4-) :3 0 4jr4c
:3 t m o. 0 0C
0 .r.4 -- 0
n 4104 4J 41' 0 C:
X m l 0 '.4 L r-
-41

x m 0:4 R to A 4) 4)
IzA m 0 v 4 ou
w oc IA 0 N m IA ~ f~ I
-P4 0% Z ..MLn 4) ,f fa (
to Z PLrl 4j l m r -4

OD -4 rito 1o E4 4)r-4 a -
r- X X 4;4 o %0 ad $
t~to 000 0 o
4J 4) 4) :3 $4 :3( r. 3 ) Im 0-0
r-4 0 0,00
S 0 M 0 44
v ~amc 0 r4 34 41rl0 D 04 0 j
r4 M U Q) 4 ) $4 in = g -, $4 W (0 :

fa r-4 ty F4
a 0 z $4, c I
a, c o Ol--,, w0
r-4 $4 0 C 4j V 0 0 .4) to 1 5
to o.* o 0 :3 0 t
a 0c ,C 0 )t 04, 4 0 W Y

0 W0 -. 0%clty U t W-r
.0 Ait 0 4) a UVO 4 1010 4 do i t V U





t;t
rkl





















4%.

:::
V t
*~' 4, a


9 N


40C4
N


C',
""a
r 494 N
\>/s




4 -
9 S




Ir
cm
4,




4 "
V "
30 P
*0 C
9 N


*~r


tI


* I
* I
II I

II I
I I
II
I I
*




Sl


I Ia






74

genotype N16 also has a repeat of 8.6 kb. P. purpureum, an

allotetraploid, has two bivalents associated with the

nucleolus during diakinesis of meiosis (Jauhar 1981). The P.

purpureum hybridization pattern indicates there has been

homogenization of rDNA repeat length between nonhomologous

chromosomes. In lanes 8 and 9 are two related pearl millet

genotypes; lanes 10 and 11 are two genotypes of P. glaucum

ssp. monodii. The hybridization patterns between the

cultivated and wild subspecies of P. glaucum are the same

although the hybridization intensities differ. There are two

major repeat sizes in the P. glaucum genotypes. These are 8.0

kb and 7.8 kb. This diploid species, which contains one

bivalent associated with the nucleolus during diakinesis of

meiosis (Jauhar 1981), exhibits greater repeat length

heterogeneity than the allotetraploid P. purpureum. This rDNA

repeat length heterogeneity exists on homologous chromosomes.

Lanes 12 and 13 are two genotypes of the apomictic hexaploid

P. squamulatum. As can be seen, there is one major band of

hybridization representing the IGS region, the major repeat

has a length of 9.1 kb. This hybridization pattern is

identical to P. purpureum except that P. squamulatum has two

minor bands of 4.0 kb and 2.8 kb. There are three bivalents

associated with the nucleolus during diakinesis of meiosis in

P. squamulatum (Jauhar 1981). Again, as in P. purpureum, there

appears to be a homogenization of rDNA repeat length between

nonhomologous chromosomes. One P. hohenackeri genotype was






75

characterized (lane 14). P. hohenackeri contains an Sst I site

not present in the rDNA repeats of other Pennisetum species

observed. The P. hohenackeri rDNA repeat may have diverged

prior to divergence of the other Pennisetum species shown

here. When the four major bands of hybridization are totaled,

P. hohenackeri appears to have a rDNA repeat length of 8.2 kb.

The interspecific hybrid P. alaucum Tift 23A X P. purpureum

(N16) hybridization pattern is in lane 15. Only one of the two

repeat sizes in the P. purpureum parent N16 is present in the

triploid. This indicates the P. purpureum (N16) repeat length

heterogeneity seen in lane 3 exists on homologous chromosomes

and not on nonhomologous chromosomes because homologous

chromosomes segregate during meiosis. Other than the lack of

the 8.6 kb repeat from the P. purpureum parent, the

hybridization pattern is additive between the parents. Lane 16

is the interspecific hybrid P. glaucum Tift 239DB X P.

squamulatum (PI319196) while lane 17 is P. glaucum (H165) X P.

squamulatum (PS26). Only the P. squamulatum (PS26) parent is

shown in this figure; however, the hybridization pattern of

the above P. glaucum X P. sauamulatum hybrid is the

expectation of the combination of any of the P. glaucum or P.

squamulatum genotypes present in this figure. A P. purpureum

Merkeron X P. squamulatum (PS262) hybrid is shown in lane 18.

This hybridization pattern displays comigration of the major

repeat sizes of these two species indicating very similar if

not identical sized rDNA repeats. The next four lanes of








Pennisetum species (lanes 19, 22, 23,and 24) are P. flaccidum,

P. villosum, and two genotypes of P. setaceum respectively.

The major repeat sizes of P. flaccidum are 8.2 kb and 7.9 kb.

The Two major repeats are 8.5 kb and 8.1 kb for the P.

villosum (lane 22). The first P. setaceum has major repeats of

8.3 kb and 8.1 kb. With the second P. setaceum (lane 24) the

major rDNA repeats are 8.4 kb and 8.1 kb. There exists little

intraspecific heterogeneity of rDNA repeat length or RFLPs

within these species except for P. purpureum (N16) and the two

P. setaceum species. However, there is intragenic

heterogeneity of rDNA repeat length between all these

Pennisetum species with the exception of P. purpureum and P.

squamulatum. The Sst I RFLPs between P. purpureum and P.

squamulatum may indicate a close relationship between these

rDNA repeat families. In maize and maize relatives rDNA repeat

length variation is not phylogenetically informative (Zimmer

et al. 1988). However, in Lisianthius species rDNA length

variation can be used as an indication of relatedness (Systma

and Schaal 1985).

P. purpureum and P. squamulatum rDNA repeats were further

compared. Total DNAs were digested with Bst NI,

electrophoresed, blotted, and probed with the same maize

repeat used to probe the Sst I digest (Fig. 14B). Lanes 2 and

3 are two P. purpureum genotypes; lanes 4 and 5 are two P.

squamulatum genotypes; lane 6 is P. glaucum; lane 7 is P.

hohenackeri. The RFLPs differ among all four of these








Pennisetum species. While purpureum and P. sauamulatum have

similar Sst I rDNA RFLPs, Bst NI digestion indicates sequence

divergence has occurred. The rDNA RFLPs of Bst NI digested

Pennisetum DNAs do not indicate as close a similarity between

P. purpureum and P. sauamulatum rDNA repeats as the Sst I

hybridization pattern indicates.

To further compare the nuclear genomes of P. purpureum,

P. squamulatum, P. glaucum, and P. hohenackeri RFLPs were

observed when the above Pennisetum DNAs were probed with

cloned sequences from napiergrass. For Fig. 15A, B, and C,

DNAs were digested with Eco RI, electrophoresed, blotted, and

probed. All three panels in Fig. 15 have the same order of

samples. The first three lanes are different genotypes of P.

purpureum; lanes 4 and 5 are two genotypes of P. squamulatum;

lane 6 is P. hohenackeri, lanes 7 and 8 are two P. glaucum

genotypes; and lane 9 is a P. glaucum X P. Durpureum triploid

hybrid. The DNAs in Fig. 15A were probed with a genomic

sucrose synthase clone from P. purpureum. The RFLP

similarities between P. purpureum and P. squamulatum are

evident. These similarities between the hybridization patterns

are marked by the three arrows. Only the P. squamulatum RFLPs

are similar to P. purpureum. P. glaucum and P. hohenackeri

exhibit no common RFLPs when compared to P. purpureum. In Fig.

15B a close relationship between the genomes of P. purpureum

and P. scuamulatum is evident. The probe for this panel was a

random cDNA clone from P. purpureum leaf tissue. This clone
















i-I *4)' 00 *l 0 *

0 V0 0 wc
E-4 0 01 W -0
A n Of 4 F-4 r.


0
r
U,0
00




O 0



0%
0
~ U

to,


DUI~

W4 01UF~Q 0 W


W4 0
rC.


















S I


m u
IS
SW
'U







a ,m n S o




t t


*- *
*,I




t t t






80

contains a sequence which is repeated many times within the

nuclear genome. The top arrow in Fig. 15B designates a band of

hybridization seen in the P. purpureum, P. squamulatum, and P.

glaucum genotypes. The bottom arrow in Fig. 15B is adjacent to

a common band of hybridization in the P. purpureum and P.

squamulatum genotypes. A random genomic clone was used to

probe the DNAs in Fig. 15C; no common RFLPs were detected in

any of the Pennisetum species. Fig. 15 displays additional

evidence indicating similarity between the nuclear genomes of

P. purpureum and P. squamulatum; at the same time though,

there are differences as well. As with the rDNA repeat

comparisons, P. hohenackeri shows no relatedness to P.

purpureum or P. squamulatum.

Isozyme phenotypes (Figs. 16-20) were observed for

various Pennisetum species and interspecific hybrids to

identify common zones of activity. In all five of the isozyme

gels shown, MDH (malate dehydrogenase), PGI

(phosphoglucoisomerase), APS (alkaline phosphatase), GOT

(glutamate oxaloacetate transaminase), and ADH (alcohol

dehydrogenase) the samples are in the same order except that

in the ADH figure there are seven P. alaucum samples instead

of six. The direction of electrophoresis is from the bottom to

the top. The MDH banding pattern is shown in Fig. 16. Lane 4

is P. hohenackeri while lanes 5-7 are various genotypes of P.

setaceum, these plants are closely related phenotypically.

These two species exhibit a differential MDH banding pattern














v41041IA-- 4 %4 41) 4

to g) 041 F4 N -W
~k~~ad -
i 0 4) ~O w
$4'b441 FA 'A a(~4
go W 41 %-- r C F-I $4 M
$4 0 g? o 0-4)4jv
04 P-4 f-4 4w0
b. r( b4


$401

$414.3c U41
0


04 0 to 0M


*r-4
a u i a 'Mar
41e4 W
4-) $4 4J rz




(64 4I ) r. :1

5:...4 4)


09-44 4
0Z V V (m
V 41 0 0Q)
r. 0 r4 M t
-H A 0 04-r4


p C.,?


o C
N.TI 0
0) I


r4 .4






82






4

\


+ ,
+ 00
4
r
\ -
o o
a





0

o o






Nlq
V.-~


0/ "~
4..
09o 04







83
as compared to the other Pennisetum species. All Pennisetum

species and hybrids exhibit four zones of activity. It should

be mentioned that three MDH loci have been reported in pearl

millet (Tostain et al. 1987). The difference may be related to

the buffer systems. Or, it is possible the top bands are from

one locus with one band being a ghost band. The ghost band,

which migrates differentially, could be due to an alteration

in protein structure.

The PGI banding patterns of these Pennisetum species are

shown in Fig. 17. P. flaccidum (lanes 2 and 3), P. hohenackeri

(lane 4), and P. setaceum (lanes 5-7) all contain the slowest

migrating PGI zones of activity among the various species. The

above three Pennisetum species, along with P. orientale in

lane 1, have distinctly different PGI phenotypes as compared

to the three related Pennisetum species: P. squamulatum, P.

purpureum, and P_. laucum. The slowest migrating PGI zone of

activity present in the P. purpureum genotypes (lanes 11-13)

comigrates with a zone of activity present in the pearl millet

genotypes. The P. squamulatum PGI phenotypes (lanes 8-10) are

similar to the P. Durpureum phenotypes (lanes 12-14) except

for the P. Durpureum zones of activity which comigrate with

the P_ glaucum zones of activity purpureum N14 in lane 13

appears to have three alleles of the PGI gene as seen by the

appearance of more than three zones of activity. All of the P.

claucum PGI phenotypes (lanes 14-19) are similar, the two

bands may be from one locus with the top band being a ghost





















to E
041 04.
:3k~(


- -f 4-9P4Of4)P4
-04 O-% C gJao
0 to go laI


04. 1.4
lz)?
C'~
$4~
00.

04 45


H4.4) 14 41 r



_V 4 04.1 4),o
0410)
N 0 04 -A -A

$44. 0 4j0 4-4 0I4)
H 4V1 41 4)
0 r 0r- $ t
04 4 A0 Q-r


-Cd 1*1
*,%(4 01
o _

C'


w 4) 04
- A



-0 Cu)



(1)



P404
0




AlI


I-4 -4





















4-


\ 0
U)


< cm
\ o
4\
10, 1;



CO
1Q








%
(0
p 10

11
C -
., -o




0- CO




4 9,*
0A
00 '.
0-c-
0 T


4
eo,
*11.


;I


I;I


~t"







86

band. One PGI locus has been identified in pearl millet

(Tostain et al. 1987). The P. glaucum X P. squamulatum

interspecific hybrids (lanes 20-22) and the P. purpureum X P.

squamulatum hybrids (lanes 23-25) have the same PGI banding

patterns. The similarity in both sets of interspecific hybrids

suggests that P. squamulatum has a gene(s) involved in the

posttranslational modification of PGI. This is because the

slowest migrating zones of activity in P. purpureum and P.

glaucum are more likely altered in mobility rather than not

expressed in the P. squamulatum interspecific hybrids. This

would explain the difference between the slowest migrating

zone of activity present in the P. squamulatum genotypes and

hybrids as compared to P. purpureum. The PGI banding patterns

show a close relatedness between P. purpureum and P.

squamulatum.

The APS banding patterns are shown in Fig. 18. P.

purpureum (lanes 11-13) and P. squamulatum (lanes 8-10) have

only one comigrating zone of activity; the fastest migrating

zone of activity. The two bands in the P. squamulatum

phenotypes may be from one locus with one band being a ghost

band. It does not appear that the zone of activity present in

the P. glaucum genotypes (lanes 14-19) is present in the P.

purpureum or P. sauamulatum phenotypes. All three of the

closely related Pennisetum species, P. glaucum, P. purpureum,

and P. squamulatum, have different APS isozyme banding

patterns. The P. qlaucum X P. squamulatum interspecific















041 04&r.4Nv O0 coo~
S- i Pe--4 v-4 -4 0IU 14

180P14 0 o- 4) 0 40 W .O
u 4, %%, ,r.r4 Pl
a0) to I k 04 is 41 '1 c

* a Irr~() U
04 g 44
SI00 w
Mo~~o i~n 3C 4 C



00O~ 41gQ~
41 0 .0Inr40 M




W4JI4O U -4Y6 Y 00 C 4
00 U 0,
O uwl $4 to


c 0 C 4 0&
dtl 140 4) Cla U O
0)l %4 4(





0r r-4
qfl44 41 0to
0n
c,



v~JE NH~, Q

iIdiirj
0%,. r-4 o 1 cc
0 ol A 41 0~ m go (
q r(T4












'V-4
-4 V C c to (n
0 ~ r P4 C4UI~
04 410 0 go (n %.o 0 a~cu
r-4, 4 a rj -I 4j %0r-
Qp, kf. P-4 00r
$4 ON C ,4 ca 0

V.4 j2, $4 0



w"o C2$4t ]
la m 910 x om


CUA) 4 14 1 0 0
1:0 4 ? C q vV-

r-4I
t;a
-V4ara ~ f~
ra + Q dQ








88










4




+ to

04
0' C38'i

cv
CYC
+ 'a



1V

0
OCY


cri

co



C/) IL)a
0f
oo
0O
C* O'J
Sa









6'G (Q
COO Q


V~0 1.l

CY






89

hybrids (lanes 20-22) exhibit zones of activity present in

both parents and form intergenic heteropolymers. The P.

purpureum X P. squamulatum hybrids (lanes 23-25) are similar

to the napiergrass parent indicating the fastest migrating

bands in both parents comigrate. APS isozyme phenotypes show

extensive variation between Pennisetum species.

The GOT banding patterns (Fig. 19) show comigrating zones

of activity present in all the Pennisetum species except P.

glaucum for the fastest migrating band. The slowest migrating

zone of activity present in the P. purpureum genotypes may

comigrate with zones of activity present in the P. glaucum

genotypes (lanes 14-19). P. hohenackeri (lane 4) has the same

GOT isozyme phenotype as P. purpureum (lanes 11-13). Both of

the interspecific hybrid crosses involving P. squamulatum

(lanes 20-22 and lanes 23-25) suggest heterozygosity in the P.

squamulatum parent. Pearl millet has two GOT loci

characterized (Tostain et al. 1987); one of these GOT isozymes

may comigrate with a zone of activity present in the P.

purpureum phenotypes.

ADH banding patterns were also observed (fig. 20). The

three related Pennisetum species, P. squamulatum (lanes 8-10),

P. purpureum (lanes 11-13), and P. glaucum (lanes 14-20), have

three zones of activity in common P.. purpureum has additional

zones of activity as well. P. flaccidum (lanes 2 and 3) has a

similar phenotype as P. alaucum. In the P. alaucum X P.

squamulatum interspecific hybrids (lanes 21-23) the common

















g a 4a






go4

' 0 0
t# k




A4 44
4)k

w 0 00




OJ
r 4


* o .u
4j 0 t
00 ^


0< 4. of


4) 4 0



E-c4 V 4J 4)
0 c 0 4

SOC- O
U-^l X 0


Wm *






C-




(N .Nm


C r u
ri i l


- -
Cl.'




















+ ul
4C-

(G.V
4,1
+ CY



c.o



C**. i r:.L :~~i... ~~







cot
'Ile



U -)
I0
(G,,
crcI
~ S

4o 0* (

#% *. c
U' O

C,, I.
C' *o :
Q' -l



















oa3 0 ,,,
S0 4 1

0 I0 *1













= to 44 .
Sa 0









. J00
M O



C 0Q.1


ou
4)


4.5 o



to &0 0to


v- PA 4 0 4 4)A
FJ OcI I OQr-


0,o

0 r
a 0


04
-M IA

c- c
N.0 <0

c 'S.-
ni E:<-


c
SIo
0


uljg


*IN *c
Io4- r4


















\


"a,













I:
0


+ Co
4 0




0'


co
(V)



0
'0
CM

0(


I'-
Co
I-
00,







'0

4 O' CO
o4
9







41, s


^o 0
Co

<-~





'0 T
4)
*0 C'
C,
0' 0
0
0'~ '


i4








4IE


It


4


o,
i







94

zones of activity are evident, these likely represent two

homodimers and one heterodimer. The P. purpureum X P.

squamulatum hybrids (lanes 24-26) have similar phenotypes as

the P. purpureum parent and further show the comigration of

the three zones of activity common to the parents.