“mg; .
4—“ «fag-f“ .

pd: ‘- 3‘-

.3.

u—J'q‘r‘

’14..
4, .A

.53.;ué4-Ji

:__:‘~

‘5‘»: .

..

A’s-0‘ -

n W‘

'2"

4941‘

...,

6.3.11

" £5Mm~

amakfhna

at

I‘L'é €3.34;

{224'

1"

' \ n
.w
" .9 -

-.

 

 

"2.:
p

gr“

 

 

 

424;,

 

v
I...

u._
.f" *4

7

v

 

1.45 '
".
v 3

: . H5
1333‘?
fin-57‘

3 r
4%44344’44

)
I. v:
243$

' if? i:
S‘M‘EE
#

 

l tfc’ SIS

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

THE ROLE OF HOMOLOGOUS RECOMBINATION
IN THE PRODUCTION OF DEBILITATING
MITOCHONDRIAL DNA REARRNGEMENTS IN PLANTS

presented by

ELAINE MARIE PALUCKI

has been accepted towards fulfillment
of the requirements for

PLANT BIOLOGY AND CELL

PH'D' degree in AND—MOLECULAR BIOLOGY

 

W

Major professor

Date Orv-£01 93,4610“
0 I ‘

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

. J 41934; 21 913299“

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p. 15

 

THE ROLE OF HOMOLOGOUS RECOMBINATION IN THE PRODUCTION OF
DEBILITATING MITOCHONDRIAL DNA REARRANGEMENTS IN PLANTS

By

Elaine Marie Palucki

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Plant Biology
Department of Cell and Molecular Biology

2003

ABSTRACT
THE ROLE OF HOMOLOGOUS RECOMBINATION IN THE PRODUCTION
OF DEBILITATING MITOCHONDRIAL DNA REARRANGEMENTS IN
PLANTS
By

Elaine Marie Palucki

Recombination plays a significant role in causing mitochondrial disorders such as
cytoplasmic male sterility as well as in the maintenance and expression of plant
mitochondrial genomes. Despite this knowledge, the proteins involved and the
mechanism by which recombination occurs in plant mitochondria remain uninvestigated.
In bacteria the best studied mechanism of homologous recombination requires the RecA
protein. These studies addressed whether the E. coli RecA protein could play a role in
generating rearrangements that cause mitochondrial disorders. Over-expression of the E.
coli RecA in plant mitochondria was anticipated to increase the number of mitochondrial
DNA (mtDNA) rearrangements. An increase in mtDNA rearrangements was expected to
result in the appearance of one or a combination of the following phenotypes:
variegation, male sterility, improper leaf expansion and overall diminished growth or

yield.

To test the hypothesis, constructs were created in which the E. coli RecA gene or
its dominant negative derivative were ligated behind a mitochondrial targeting sequence
311d a strong promoter. Agrobacterium—mediated transformation was used to integrate
these genes into the nucleus of Arabidopsis thaliana (Columbia and chm lines) as well as

Nicotiana tabacum. In these studies, over-expression of E. coli RecA in Arabidopsis did

not result in phenotypic manifestation of plant mitochondrial syndromes. In contrast
phenotypic characteristics of mitochondrial syndromes were apparent in Nicotiana, where
multiple homoetic-like changes were seen during floral development. These changes
were expressed as a mosaic on each plant, with normal flowers appearing next to those
with developmental abnormalities. Once these abnormalities were induced, their
expression and transmission was independent of the presence of the E. coli RecA
transgene. T1 plants showed additional characteristics of plant mitochondrial
dysfunction including: infertility, dwarfism, changes in leaf expansion and variegation.
To assess whether these abnormalities were maternally inherited, a back-cross population
was generated. Maternal inheritance was confirmed because all the progeny exhibited
some level of floral abnormality as well as infertility. Additional support for the
mitochondrial basis of these changes came in the identification of a mtDNA RFLP in

nine of the backcross plants.

The production of the plants provides a new resource in which to study the roles
of mitochondrial segregation and threshold effects of a mixed population of mtDNA
molecules. The transgenic plants are phenotypically novel compared to other plant
mitochondrial mutants in that the manifestations of mitochondrial disorders are not
uniformly presented. The appearance of floral abnormalities as a mosaic, in conjunction
with the number and variability of the floral phenotypes, suggest that active segregation
is occurring in the plants, and proper development of floral organs is responsive to

thresholds of functional mitochondria.

ACKNOWLEDGEMENTS

I would like to thank first and foremost Dr. Barbara B. Sears, for her gentle
guidance, academic expertise and thoughtful encouragement, as well as for her
willingness to allow me to explore an area that was not the laboratory’s main focus. I
would also like to thank my committee members Dr. Helmet Bertrand, Dr. Kenneth

Keegstra and Dr. Michael Thomashow for all their help and input over the years.

Without the help of many scientists, this work would not have been possible. I
thank members of the Thomashow laboratory for their aid in learning transformation
techniques, in specific Dan Zarka. Many laboratories were invaluable resources for the
protein analyses described: the Hammerschmidt, Keegstra, McIntosh and Osteryoung
laboratories. In specific hearty thanks to Roxy Nickels, Rosemary McAndrew, Diane
Jackson-Constan and Andrea de la Rocha for the time they spent advising me. Many
scientists at Michigan State University have aided in my professional growth both as role
models and mentors and to whom I am deeply grateful: Dr. David Douches, Dr.
Raymond Hammerschmidt, Dr. Kenneth Sink and Dr. Carolyn Malmstrom. In addition, I
would like to thank Dr. Carolyn Mahnstrom for not only welcoming me to her laboratory,

and for her belief in my abilities as a science communicator.

Recognition must also be given to past and present Sears laboratory members.

Many undergraduates contributed greatly to this dissertation, specifically Trinh T. Tran,

Julianna Fratos-Matos, and Sandra Bertrand. In specific, I thank Kristi Nichols, not only

iv

for her tireless hours of greenhouse screening, pollen counting and DNA isolation, but
also for the gifl of her boundless and selfless friendship that has been an irreplaceable
support to me. To fellow graduate students Lara Stoike Steben and Elizabeth Graffey, I
also offer great thanks for all I learned from them. My time as a graduate student was
enriched in every possible way by the deep friendship, generous love, and scientific
excellence of Monica Guha Majumdar for whom I thank completely. I have become a

better person by her example.

Lastly, I thank my family. Thanks to Jeannie Hoffman-Censits who has been dear
and special friend. Many thanks to my father for his valuable, constant support and my
brother whose enthusiasm for science is contagious. For my sister Laura Palucki-Blake
there are not enough words to describe my appreciation for her love, support, and endless
faith in my abilities. Finally, I thank David A. Rudd for being my confidante, my

compass, and atlas for this journey and never letting me lose my way.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. x
LIST OF FIGURES ........................................................................................................... xii
CHAPTER 1
Characteristics of Plant Mitochondria and Their Genomes ................................................ 1
Comparison of plant and animal mitochondria ....................................................... 1
Plant mitochondrial genome structure ..................................................................... 3
Role of recombination in the production of debilitating mitochondrial syndromes
in plants ................................................................................................................... 7
Amplification of molecules produced by recombination plays a role in phenotypic
presentation ........................................................................................................... ll
Mechanisms and proteins involved in homologous recombination ...................... 15
Hypothesis ............................................................................................................. l 8
Significance ........................................................................................................... 19
CHAPTER 2
Material and Methods ....................................................................................................... 20
Experimental Organisms and Growth Conditions
Arabidopsis thaliana ................................................................................. 20
Tobacco ..................................................................................................... 20
Other organisms ........................................................................................ 22
Transformation
E. coli and Agrobacterium ........................................................................ 22
Arabidopsis ............................................................................................... 24
Tobacco ..................................................................................................... 24
Kanamycin Seed Assays ....................................................................................... 25
Plant Phenotypic Analyses .................................................................................... 26
Pollen Germination Assays ................................................................................... 26
Polymerase Chain Reaction .................................................................................. 27
Cloning of Plasmids .............................................................................................. 29
Ligation ................................................................................................................. 31
Recombination Assays .......................................................................................... 32
DNA Isolation
Plasmid DNA ............................................................................................ 33
Chlamydomonas DNA .............................................................................. 35
Plant DNA ................................................................................................. 35
Restriction Endonuclease Di gestions .................................................................... 36
Protein Purification
E. coli ........................................................................................................ 36
Tobacco ..................................................................................................... 36

vi

CHAPTER 2

Material and Methods (continued)
Gel Electrophoresis
Agarose Gels ............................................................................................. 37
Polyacrylamide Gels ................................................................................. 38
Gel Transfer and Blotting
DNA Transfer ........................................................................................... 38
Southern Blotting ...................................................................................... 39
Protein Transfer ........................................................................................ 40
Western Blotting ....................................................................................... 41
CHAPTER 3
Creation of RecA Constructs and Plant Transformation ................................................... 43
Introduction ........................................................................................................... 43
Results
Construction of clones with wild-type and dominant negative forms of
E. coli recA preceded by a mitochondrial transit peptide sequence ......... 4-4
Recombination assays to verify the activity of the cloned modified RecA
gene ............................................................................................................ 46
Confirmation of initial transformation ...................................................... 49
Confirmation of the T1 plants ................................................................... 54
Confirmation of the BCI plants ................................................................. 57
Discussion ............................................................................................................. 57
Conclusions ........................................................................................................... 60
CHAPTER 4
Phenotypic and Genetic Analyses of Transgenic Plants in Which the E. coli RecA
Protein is Expressed in the Mitochondria ......................................................................... 61
Introduction
Mitochondrial disorders in maize and Arabidopsis .................................. 61
Cytoplasmic male sterility ........................................................................ 64
Uniquely Expressed CMS: Alloplasmic lines in Nicotiana and Daucus ..65
Summary ................................................................................................... 70
Goal and Hypothesis ............................................................................................. 71
Expected Phenotypes
Transgenic plants expressing wild-type RecA ......................................... 71
Transgenic plants expressing dominant negative RecA ........................... 72
Results
Phenotypic changes in Arabidopsis lines expressing various E. coli
recA genes ................................................................................................. 74
Phenotypic changes in Nicotiana tabacum lines expressing the E. coli
recA genes ................................................................................................. 76
Phenotypic Analysis of T1 progeny from line 12 ...................................... 84
Inheritance of floral abnormalities ................................................ 84

vii

CHAPTER 4
Phenotypic and Genetic Analyses of Transgenic Plants in Which the E. coli RecA
Protein is Expressed in the Mitochondria (continued)

Results
Phenotypic Analysis of T1 progeny from line 12
Analysis of overall growth ............................................................ 92
Analysis of reproductive ability .................................................. 102
Summary of T1 progeny analyses ............................................... 105
Production of a Back-cross (BC!) population ........................................ 107
Analysis of the BC 1 population ............................................................... 109
Assessment of male sterility through pollen germination assays ........... 124
Summary of BC! analysis ....................................................................... 133
Discussion ........................................................................................................... 133
Conclusion .......................................................................................................... 144
CHAPTER 5
Molecular Basis for Abnormal Floral Development and Male Sterility
Introduction ......................................................................................................... 145
The location of gene deletions and their roles in mitochondrial
dysfunction .............................................................................................. 147
The organization of chimeric 0173' and their role in mitochondrial
dysfimction .............................................................................................. 150
Fertility restorer’s and their relation to CMS-associated ORFs .............. 155
Summary ................................................................................................. 157
Goal and hypothesis ............................................................................................ 158
Results
Rationale for probes chosen for mitochondrial DNA analysis studies ...158
Investigation of affected T1 plants .......................................................... 160
Investigation of afi‘ected BCl plants ....................................................... 161
Discussion ........................................................................................................... 166
Conclusion .......................................................................................................... 169
CHAPTER 6
Conclusions and Future Directions
Summary of results .............................................................................................. 170
Relationship of my observations to models of CMS action and mitochondrial
signaling .............................................................................................................. 171
Future directions ................................................................................................. 183

viii

APPEN

Prczein

LIIER

APPENDD(
Protein Localization Experiments ................................................................................... 185

LITERATURE CITED ................................................................................................... 192

ix

Table
Ithlt‘
Itbli
Itbh
Ilbll

itbl

Iabl
iibl
Iabi
lrbl

iabl

Izbl

lab]

Irbl

Irbl
Irbl
Iihi
Tihl

szi

LIST OF TABLES

Table 2.1: Murashige and Skoog Media .......................................................................... 21
Table 2.2: Bacterial, Agrobacterial and Chlamydomonas strains ................................... 23
Table 2.3: Primers used in polymerase chain reaction experiments ................................. 28
Table 2.4: Plasmids used in experiments ......................................................................... 34
Table 2.5: Antibodies used for Western blot analyses .................................................... 42
Table 3.1: Effect of RecA plasmids on recombination rate of pDK8 plasmid in strain
HBlOl .............................................................................................................. 48
Table 3.2: Number of transgenic Arabidopsis plants recovered ...................................... 50
Table 3.3: Kanamycin seedling assay for resistance ........................................................ 53
Table 3.4: Kanamycin seedling assay for resistance ........................................................ 56
Table 4.1: Summary of Arabidopsis thaliana transformation and screening data ........... 75

Table 4.2: Characteristics of T0 Nicotiana tabacum transformed with the wild-type
RecA construct .............................................................................................. 77

Table 4.3: Description of Abnormalities seen in T0 Tobacco carrying the wild-type
RecA construct ................................................................................................ 78

Table 4.4: Comparison of averaged data fiom T1 plants both with and without RecA
expression ...................................................................................................... 88

Table 4.5: Data fi'om individual Tl progeny in the line 12pet, ordered by % abnormal
flowers and categorized by the presence or absence of the transgene ......... 103

Table 4.6: Seed Set fi'om Pollination of T1 plants with wild-type pollen ..................... 108

Table 4.7: Developmental abnormalities in BC] plants lacking the RecA u'ansgene ....1 10

Table 4.8 (a-e): Complete data for pollen germination and viability ........................... 114
Table 5.1 Mitochondrial genes identified in flowering plants ..................................... 146
Table 5.2: Summary of known gene deletions and their recombination sites ............... 148

iMdJ

hmil

Table 5.3: Summary of known CMS causing chimeric genes and their restorers
of fertility ...................................................................................................... 152

Table 5.4 Probes using in Southern blot analysis of T1 and BC. plants ......................... 159

xi

LIST OF FIGURES

Figure 1.]: Different segmental arrangements of the mitochondrial genome of
Arabidopsis thaliana result fiom recombination at two large repeats ............. 5

Figure 3.1: Schematic representation of the T-DNA region of the wild type RecA clone
in the Agrobacterium expression vector pBIlZl .......................................... 45

Figure 3.2: Agarose gel confirmation of tobacco To transformation ............................... 52

Figure 3.3: Agarose gel confirmation of T. plants with and without the transgene ........ 55

Figure 3.4: Agarose gel confirmation of tobacco BC] populations ................................. 58
Figure 4.1: Comparisons of wild-type and abnormal flowers in the T. progeny ............ 80
Figure 4.2: Comparisons of wild-type and abnormal flowers in the T1 progeny ............. 81
Figure 4.3: Correlation between the percentage of abnormal flowers on a plant and the
amount of infertility ...................................................................................... 83
Figure 4.4: Overview of the genetic analyses .................................................................. 85
Figure 4.5: Number of plants in each T1 line grouped by the overall amount of floral
abnormality expressed .................................................................................. 86
Figure 4.6: Average trends for the T. plants, in terms of height, number of flowers,
percentage of abnormal flowers and percentage of infertility ..................... 89
Figure 4.7: This is an example of a markedly abnormal flower from the BC;
generation ..................................................................................................... 91

Figure 4.8: These are examples of abnormal leaf phenotypes seen in the T1
generation ...................................................................................................... 93

Figure 4.9: A markedly abnormal leaf fiom the T1 generation ....................................... 94

Figure 4.10: Comparison of the percentage of floral abnormality with the percentage of
abnormal leaves per plant in the line 12pet both with and without
expression of the E. coli RecA transgene ................................................... 96

Figure 4.1 1: Comparison of the percentage of floral abnormality with the percentage of

abnormal leaves per plant in the line 12stig both with and without
expression of the E. coli RecA transgene .................................................... 99

xii

Figure -

figure!

Figure .

Figure

figure

figure

figure

Figure

Figure 4.12: Number of plants in each T1 line grouped by the overall amount of
infertility they expressed .......................................................................... 104

Figure 4.13: Correlation between the percentage of abnormal flowers on a plant and the
level of infertility ................................................... _ ...... 106

Figure 4.14: Number of plants in each BC; line grouped by the overall amount of floral
abnormality expressed .............................................................................. 112

Figure 4.15: Number of plants in each BC] line grouped by the overall amount of
infertility they expressed .......................................................................... 113

Figure 4.16: Averages for control Sarnsun and the BC 1 plants, in terms of the percentage
of abnormal flowers and percentage of infertility .................................... 117

Figure 4.17: Correlation of the percentage of abnormal flowers found per plant and the
level of infertility per plant ....................................................................... 1 18

Figure 4.18: Comparison of the number petaloid converted flowers found out of the
total number of abnormal flowers ............................................................ 120

Figure 4.19: Average amount of pollen germination for individual plants in the BC]

lines .......................................................................................................... 125
Figure 4.20: Average amount of pollen produced by individual plants in the BC,
lines .......................................................................................................... 127
Figure 4.21: Correlations of pollen quantity, pollen germination, and level of
infertility per plant ................................................................................... 131
Figure 4.22: Correlation test of the percentage of floral abnormality, pollen
quantity and pollen germination ability .................................................... 132
Figure 5.1:
(A) Agarose gel of selected plants fiom the lines 31 and 44 digested with EcoR I
(B) Southern blot probed with the cosmid clone 7G1 .............................................. 163
Figure 5.2: Summary of BC. plants that show the ~2.9 kb mitochondrial RF LP ........ 165
Figure A.l: Western blots of leaf tissue.
(A) AOX antibody
(B) RecA antibody ..................................................................................... 186

xiii

Figure

figure

Figure A.2: Western blots of proteins isolated from leaves and seedlings of
transformed and untransformed tobacco
(A) COXII
(B) HSP70 .............................................................................................. 188

Figure A.3: Western blots of proteins isolated fi'om leaves and seedlings of
transformed and untransformed tobacco: RecA ........................................ 189

xiv

”7?;

11:.

iii:

CHAPTER 1

Characteristics of Plant Mitochondria and Their Genomes

Proper growth and development of plants involves the co-ordinate firnction of
three distinct genetic compartments: the nucleus, plastids and mitochondria. In these
interactions, nuclear genes play important roles in regulation of organelle development
and firnction, since the majority of plastid and mitochondrial proteins are nuclearly
encoded and imported into the organelles after synthesis. Aspects of organelle
metabolism can impact nuclear gene expression as well, with calcium, heme,
phorphyrins, oxidation/reduction components and ATP acting as signals in cross talk
between the chloroplasts and mitochondria as well as with the nucleus (Allen et a1.
1993, Rodermel 2001, Susek and Chory 1992). Knowing that proper cellular firnction
is linked to organelle function, it is useful to know the overall characteristics of
organelles. This dissertation will explore the role of mitochondria in plant
development; with this chapter highlighting the significant characteristics of the plant
mitochondrion, in specific, its genome as it relates to overall plant development. The

folIOWin g chapters will contain the relevant background and introduction for the

Specific tOpics they cover.

Comparison of plant and animal mitochondria
Despite the fact that all mitochondria have a functional similarity as the main

Si “3 0f energy transduction, certain features distinguish plant mitochondria from their

eukaryotic counterparts. For example, in terms of their respiratory chain, plant
mitochondria are more complex than mammalian mitochondria: having multiple,
active, non-ATP producing, alternative NAD(P)H dehydrogenases which bypass
complex I (Rasmusson et a1. 1998, Moller 1986) as well as an alternative oxidase
pathway which bypasses ATP producing complexes 1H and IV (Sideow and Umbach
1995)- In addition, plant mitochondria differ in their capability to export and import
several metabolites such as: oxaloacetate, aspartate, a-ketoglutarate, malate and
glutanuate (Douce and Neurburger, 1989). However, the most striking of the
differences between angiosperm mitochondria and their animal counterparts are found

in mitochondrial genome size and organization.

Animal and fungal mitochondrial genomes are known to be composed
typically of a 16-150 kb, stable, circular structure. All mammalian mtDNA are similar
in size ( l 6.6 kb) as well as gene content and order, containing a total of 37 genes: 22
tRNAS, 2 ribosomal RNA genes and 13 genes involved in the respiratory chain and
oxidative phosphorylation (Schon, 2000). While gene content and overall genome size
are highly stable, animal mitochondria are known to evolve quickly in terms of

individual gene sequences.

In contrast, plant mtDNAs are more complex and at least 10-100 times larger
than their animal counterparts (reviewed by Gillham 1994). In fact, flowering plants
have the largest known mitochondrial genomes, ranging in size from 208 kb in

B’aSSiCO hirta (Palmer and Hebron 1987) to 2400 kb in muskrnelon (Ward et al.

1981)- In addition to their large size, angiosperm mitochondrial genomes are less
compact in their gene organization than either animal genomes or those of lower plants
such as the bryophytes (Oda et al. 1992). In addition, even closely related angiosperm
species can have mitochondrial genomes of vastly different size. In the
cucurbitaceae, genomes show a nine fold variation in size: 330 kb (watermelon,
Citrillus lanatus L.), 800 kb (squash, Cucubita pepo L.), 1500 kb (cucumber, Cucumis
sativus L.) and 2400 kb (melon, C. melo L.). There is no evidence that the larger
cucurbit mtDNAs possess more coding regions, gene duplications or accumulation of
introns (Ward et a 1981, Stern and Newton 1985, Havey et al. 1997). In general, 50-
70% of the plant mitochondrial genome is made up of these noncoding sequences
(including introns), and 90% of the total sequences are considered to be unique from
known sequences (Breiman and Galun 1990). Nuclear and chloroplast genome sizes
in the same species do not parallel the size variations in the mtDNAs (Palmer 1982,

Havey et al. 1998).

Plant mitochondrial genome structure

Physical mapping studies on plant mitochondrial genomes have been
performed and consistently yield a circular structure for these genomes (Iwahashi et a1.
1 992; Janska and Mackenzie 1993, Fauron et a1. 1995, Kubo et al. 1995, Moeykens et
al. 1995, Shikanai et a1. 1998). Along with determining the sizes and potential gene
ol'gimization, mapping studies also identified classes of large (1-10 kb) repetitive

elemfllts that are hallmarks of all plant mtDNAs (reviewed by Hanson and Folkerts

L.‘

5+

“5

1992)- As defined by Stern and Palmer (1984), these repeats are sequences present in
a mininmm of two copies relative to other sequences in the genome. These repeats
can be recovered in four genomic environments as a result of homologous
recombination between the copies, and can be located in both coding and noncoding

regions of the genome (Stern and Palmer 1984).

Using these repeats as a basis, Palmer and Shields (1984) proposed a
multipartite model for the structure of plant mitochondrial genomes in which a single
“master circle” DNA molecule undergoes both inter- and intrarnolecular homologous
recombination to create a population of smaller genomic circles. In this model,
recombination between direct repeats will convert the master chromosome into two
subgenornic molecules, each containing one copy of the repeat. Recombination
between inverted repeats will give rise to different isomers of the main mitochondrial
circle, by inverting the segment in between them. In spite of genome complexity, any
rcpeat has only four possible genomic environments based on the sequences that
adjoin it- Both PCR and Southern blotting of repeat units have been used to
demOnStrate that in naturally occurring mitochondrial populations, the alternative
orientations of each repeat unit can be identified and that these molecules exist in

equal stoichiometry with one another (Klein et a1. 1994).

To illustrate the master circle model, the 372-kb mtDNA of Arabidopsis
thaliana is shown in Figure 1.1. When mapping this genome, Klein et al. (1994)

identified the presence of two pairs of repeats of 4.2-kb and 6.4-kb respectively that

 

figure 1.1: Different segmental arrangements of the mitochondrial genome of
Arabidopsis thaliana result from recombination at two large repeats. The two
repeat units are represented by straight arrows or arrows with closed diamonds
at the end. Molecules 1, 2 and 3 represent the complete genome in size (372
kb) and gene content. Molecules 4 and 5 represent subgenomic circles (138 kb
and 243 kb in size respectively). Individual regions of the genome between
repeats were assigned letters, so that changes in organization could be seen.
This figure is modified fiom Klein et al. 1994.

male

Recon

mrhi

T0611]

.CIII

$La-
...3

we .-
Lilli

can

no;

103

enable the assembly of alternative versions of the mitochondrial genome.
Recombination between identical repeats can create three versions of the master circle
in which segments have different orientations (1,2,3), and two subgenomic molecules
of 138 and 234-kb (4,5). These studies also showed that the sequences around each
repeat occurred in equal stoichiometries, suggesting that recombination across repeats
occurs freely and without loss of any of the resulting molecules. The interconversions
depicted in Figure 1 are the simplest of those that can be imagined: because each
mitochondrion contains many copies of the genomes, recombination between repeats
could involve more than one unit of the genome. For example, recombination
involving one repeat on molecule 1 and the identical repeat on molecule 3 would lead
to a mtDNA dimer containing four copies of each repeat. From recombinations
involving this dimer, other complex variants could be generated.

With the multipartite model as a guide, experiments were performed to
visualize the presence of these subgenorrric molecules using electron microscopy (EM)
and Pulse Field Gel Electrophoresis (PFGE). The initial results of these studies were
unexpected, Instead of finding a population of circular molecules representing the
master Circle and discrete subgenomic sizes, only 0-5% of the total mtDNA of B. hirta,
tobacco or Chenopodium album was found to be circular (Bendich 1993, Backert et a1.
1 997)- Rather, a portion of land plant mitochondrial DNA appears in PF GE analysis as
a smear of linear molecules 40-250 kb in size regardless of the size of the

1"mitochondrial genome, or of the expected subgenomic circles (Scissum-Gunn et al.

1 998, Bendich 1993, Oldenberg and Bendich 1998).

 

Of the total mtDNA from liverwort or from two species of Brassica, 50-90%
remains bound in the well of the pulse field gel. Electron microscopy indicates that
the well-bound DNA is a mass of complex branching structures. These branching
structures are often rosette-like and larger than the predicted size of the unit genome of
the mitochondrion. Backert et al. (1997) analyzed the mtDNA of C. album in PFGE to
determine if there was single-stranded DNA present. In fact, the well-bound DNA
was especially rich in single-stranded DNA, including linear molecules that were
single-stranded at one or both their ends, sigma-like molecules, and entirely single-
Stranded circular DNAs. These studies generated the hypothesis that this single-
stranded DNA represents recombination and replication intermediates of the
mitochondrial genome. Conceivably, a high rate of recombination between repeat
units contributes to the heterogeneity and broad range in sizes of the population of
plant mtDNA. While neither circular master chromosomes nor subgenomic molecules
of distinct size classes were identified in any of these studies, PFGE results do suggest
that recombination plays a role in generating the physical structure of plant mtDNA.
In addition, it appears that this structure is perhaps even more complicated than the

mutltipartite model predicts.

R019 of recombination in the production of debilitating mitochondrial syndromes of
plants

Compared to marmnalian systems, where over 200 mitochondrial diseases have
been identified, plants exhibit a relatively small number of syndromes caused by the

l"fitmhondria (Wallace 1999). In fact only two plant mitochondrial syndromes are

naturally occurring, well documented and researched: cytoplasmic male sterility
(CMS) and non-chromosomal stripe (NCS) (reviewed by Khvorostov et a1. 2000,
Hanson 1991). In contrast to animal mitochondrial diseases, in which many are
caused by point mutations (Schon 2000, Wallace 1999), the mutations that generate
the CMS and NCS phenotypes are caused by active recombination of the genome. In
the following paragraphs, the molecular basis for both cytoplasmic male sterility and

the non—chromosomal stripe mutants are explored.

Cytoplasmic male sterility (CMS), a condition in which the plant is unable to
produce functional pollen, has been characterized in over 150 plant species, including
a number of crops: Phaseolus vulgaris, Brassica napus, beet, carrot, maize, onion,
petunia, tobacco, rice, rye, sorghum, sunflower and wheat (reviewed in Schnable and
Wise 1998). In contrast to CMS, the NCS mutants of maize are characterized by
white, yellow, or light green leaf striping/variegation, and occasionally overall poor
BYOWth and sterility (reviewed by Newton 1995). Although not as well researched,
PhenOIYpes similar to NCS have been known to occur in the chm mutant lines of
Ar abidopsis thaliana that in addition to variegation, also produce phenotypically

distorted leaves (Sakarnoto et. a1 1996).

Molecular characterization of these disease-like syndromes has indicated that
reconlbit'ration occurring outside of the large repeat units can have a dramatic impact
On plant mitochondrial genomes and their expression. At the heart of these disorders

are mDNA rearrangements that disrupt the normal expression of mitochondrial genes

01- create chimeric genes whose expression disturbs organelle fimction (Rottman et al.
1987, Kadowaki et a1. 1990, Mackenzie and Chase 1990, Hartmann et al. 1994, Senda
et al- 1998). Occasionally, the debilitating mtDNA rearrangements are generated by
recombination between repeats larger than 1-kb (Conklin and Hanson 1994), but more
often, the repeats that are implicated are of medium (100-450 bp) or short (6-40 bp)
length (Marienfeld and Newton 1994, Sakamoto et al. 1996, Kadowaki et al. 1990).
For example, the NCS3, NCSS, and NCS6 mutants of maize have chimeric genes that
resulted from rearrangements between repeats of 6-36 bp (Marienfeld and Newton,
1994, Hunt and Newton 1991, Newton et al. 1990). A more specific discussion of the
nature of these mitochondrial syndrome—causing mutations and their effect on overall

mitochondrial firnction will be addressed in Chapter 5.

'I‘he large and complex nature of plant genomes allows them the space, and
active recombination provides the opportunity by which new coding sequences may
come into existence (Budar et al. 2003). The short repeat sequences that participate in
Chimeric gene production are common and littered throughout higher plant
mitochondrial genomes (Andre et al. 1992, Conklin and Hanson 1994). For example,
in the Arabidopsis mtDNA, many genes contain duplications that could be
recombinationally active: 4% of the sequence complexity is composed of 144
duplications of 30-560 bp, where the repeats are 90% identical (Unseld et al. 1997).
In cuclumber short 30-53 bp repetitive DNA motifs which were degenerate,
Overlapping and showed no homology to any sequences in the database have been

estimated to account for >13% of the mitochondrial genome (Lilly and Havey 2001).

In fact, it has been proposed that mitochondrial genomes are much like a sea of

repetitive units, in which coding sequences appear as islands (Lilly and Havey 2001 ).

Small et al. (1989) proposed that these small repeats could contribute to the
formation of an entire class of rare substoichiometric recombinant molecules. These
molecules could act as a reservoir of sequence rearrangements that, upon
amplification, could result in altered gene expression. In this model, variations of the

multipartite structure might be used as an amplification mechanism to modulate the

amount of a gene product required during a specific stage of plant development.

In addition recombination at short repeats could cause deletion of portions of

genes. Also, recombination at short-medium repeats could disconnect a gene from its
Pmper promoter region changing its transcriptional activity as well. For example, the
maize cox2 gene is known to lie immediately downstream from a recombinationally
a-<=tive 0.7 kb direct repeat that is present in two copies in the master chromosome
(LUpold et al. 1999a, 1999b). The promoters for the con gene exist upstream from
both repeats and are named A and B based on their genomic context; Region A
c'Dntains two promoters and region B contains three promoters. Transcriptional assays
for promoter strength have shown that promoter region B is more active than A (4:1
1"3»tio); and quantitative PCR shows that con is more ofien located in the genomic
arrangement with the B promoter region (6:1). Through these data, Lupold et al.

( 1 999b) concluded that the use of promoters responds to their genomic context and

that there is a role for recombination in regulating gene expression.

10

1
?

 

Amplification of molecules produced by recombination plays a role in phenotypic
presentation
While small- to medium-sized repeats are adequate substrates for
recombination in the generation of debilitating mtDNA rearrangements and are formd
copiously throughout plant mitochondrial genomes, their presence alone may not be
enough to cause phenotypic production of mitochondrial syndromes. It has been
shown that, during overall development, there are a variety of rearranged molecules
present at a low level that do not affect the overall phenotype of the plant. However,
these molecules are available for amplification (random or selective) at different times
and stages of development, presenting a mechanism by which plant mitochondrial
genomes could modulate gene expression (Bonhomme et al. 1992, Vitrat et al. 1992,
Fla et al. 1995, Yesodi et al. 1995, Suzuki et al. 1996, Guitierres et al. 1997, Laser et

al- 1997),

A study by Janska et al. (1998) on a cytoplasmically male sterile line of the
c(>1'nmon bean (Phaseolus vulgaris) is an excellent example of this type of gene-
e)Kpression control. In the common bean, a chimeric gene called pvs-orj239 is

suspected to be the cause of sterility because it is transcribed actively in both
Vegetative and reproductive tissues (Chase 1994), but accumulates only in the
reproductive tissue. The pvs-orf239 sequence disappears from the genome when
SPOntaneous reversion to fertility occurs or when the nuclear fertility restorer allele FR

is introduced (Mackenzie and Basset 1987, Mackenzie and Chase 1990, Janska and

11

Mackenzie 1993). Janska et al. (1998) hypothesized that, if a population of randomly
rearranged molecules is present in the mitochondria, then the pvs-orj239
rearrangement should be present at low levels in fertile plant mitochondria. The
reverse should also be true, and the fertile gene arrangement should be found at low
levels within sterile plant mitochondria. Results of this study strongly support the
hypothesis: the transition from the progenitor mitochondrial configuration to the pvs-
077239 involves the amplification of the sterility—inducing gene arrangement and
suppression of the progenitor arrangement. These results were further supported by
the finding that fertile revertants still contained the pvs-orj239 sequence, but at
substoichiometric levels. It was suggested that the selection and/or maintenance of
tllis sequence could be because it is included on a subgenomic circle that carries a
necessary gene. In a separate study, Arrieta-Montiel et al. (2001) used the pvs-orj239
sequence as a mitochondrial genetic marker and demonstrated that it was universally
present in 105 wild and 36 cultivated lines of Phaseolus. The pvs-orfi39 sequence
Was substiochiometric in all but ~10% of the lines examined. Their results also
identified a putative progenitor sequence in Phaseolus glabelus. The nature of this
Sequence suggests that the present-day pvs-orf239 most likley originated fiom a
tI‘lnncated substoichiometric sequence that went through one or two recombination

e‘Ients to produce the present day sequence.
Investigations with mitochondrial mutants of other plants also have indicated

thal differential amplification of a pre-existing set of subgenomic molecules can be a

llleans of altering mitochondrial gene expression. In Arabidopsis thaliana, the

12

 

ti

maternal distorted leaf phenotype (MDL) is a good example of this process. The
MDL phenotype was found to be due to three independent mtDNA rearrangements
that affected the 17233-17211 6 operon. PCR experiments showed that wild-type plants
contained a low level of the rearranged tps3-tpll6 operon. The reverse was also true:
MDL plants contained the wild-type gene arrangement at low levels and amplified
levels of the rearranged sequence. From these results, it was suggested that the loss of
the CHM protein severely reduces the abundance of the master genome in favor of
recombinant subgenomic molecules. Sakamoto et al. (1996) hypothesized that CHM
is required for maintaining the master chromosome and/or eliminating unusual

molecules that result in physiological deficiencies.

In addition to producing aberrant phenotypes, these types of subgenomic shifts
can occur in wild-type plant tissues. For example, Kanawaza et al. (1994) found
through tissue culture studies, that during the transition from green plant to callus and
back to green plant, the population of mitochondrial molecules changed in tobacco: a
specific rearranged restriction fragment carrying the atp6 gene accumulated in the
dedifferentiated callus cells at high levels; this same fragment was found in only low
levels in the regenerated green plants. Changes in subgenome stoichiometry were also
observed in the CMSH mitochondrial mutant of N. sylvestris, in which amplification

leads to cytoplasmic male sterility (Lelandais 1998).

In summary, studies of plant tissue culture and mitochondrial mutants have

yielded results consistent with the idea that mtDNA molecules could be maintained at

13

substoichiometric levels when not expressed, but become transcriptionally and
translationally active when amplified above threshold levels. This means that in
addition to regulation of gene expression by transcriptional, post-transcriptional and
translational processes within plant mitochondria (Smart et al. 1994, Conley and
Hanson 1994, Sarria et al. 1998, Mulligan et al. 1991, Menassa et al. 1999), an
additional level of control could be attained through recombination followed by

genome amplification.

If plant mtDNA truly exists as a rare master circle and various subgenomic
molecules, the necessity of maintaining all genes to assure mitochondrial firnction
could have driven the evolution of a system for modulation of the ratios of the various
genomes. As opposed to the spindle structure, which assures equal partitioning of
sister chromatids during mitosis, no such precise mechanism is thought to exist for
segregating mtDNA to daughter mitochondria (Birky 1983, 1991). Without such a
process, random sorting-out of the various different subgenomes should occur, but the
resultant gene deficiencies would severely impact plant cell viability. Recognition and
counting of the subgenomes, with subsequent amplification of any in low copy number
would provide a solution to this problem. Other alternatives are also conceivable: if
the master mtDNA circle always replicates with copies segregated to daughter
mitochondria, no gene functions would be lost. Alternatively, continual fusions of
mitochondria, with mixing of their genomes and fiequent recombination, would also

reduce the possibility of gene loss by vegetative segregation.

l4

I 'r

Mechanism and proteins involved in homologous recombination

With the evidence that maintenance and expression of plant mitochondrial
genomes are affected by recombination and amplification, it is surprising to note that
the proteins that trigger rearrangements or operate in replicational amplification have
not been identified. Recombination is the primary mechanism by which molecules for
amplification are produced. In specific, homologous recombination is usually
mentioned as the process responsible for the mtDNA reanangements in plant
mitochondria. While this pathway has not been examined experimentally in plant

mitochondria, it has been thoroughly studied in the bacterium E. coli.

The mechanism for homologous recombination has been elucidated most fully
through studies of the RecBCD pathway (reviewed by Camerini-Otero and Hsieh
1995, Kowalczykowski and Eggleston 1994). Initiation of recombination requires
nicking of DNA or a double strand break, followed by the production of single-
stranded molecules through the action of the RecBCD protein, which acts as both a
helicase and a nuclease. Next, a presynaptic filament is formed as the RecA protein
associates with the single-stranded DNA. RecA is the central protein involved in
homologous recombination since it participates in both strand pairing as well as the
recognition of sequence homology. When RecA is inactivated by mutation,
homologous recombination is drastically reduced, demonstrating this protein’s pivotal
role in the process (Camerini-Otero and Hsieh 1995). The association of RecA on the
single-stranded DNA allows for the pairing of DNA regions with significant amounts

of homology, followed by an exchange of DNA strands, producing a region of

15

heteroduplex DNA and a Holliday junction. This Holliday junction can move by
branch migration that increases the region of heteroduplex DNA. In the final step of
homologous recombination, a complex of three proteins (RuvABC) recognizes the
Holliday junction and cleaves the exchanged DNA segments from one another.
Sequence homology is central to the genetic exchange in homologous recombination.
The level of sequence homology required for proper pairing and exchange has been
well documented: in E. coli the minimum length of homology is 20 bp (Watts et al.
1985, Shen and Huang 1986). The mechanism for recombination at repeats smaller
than 20 bp does not require RecA and is termed illegitimate recombination (Allgood
and Silhavy 1988). In contrast the minimum length of DNA for homologous

recombination in mammals this number is increased to 200 bp (Liskay 1987).

The function and sequence of the RecA protein is evolutionarily conserved
(Horii et al. 1980, Sancar et al. 1989). Homologs and orthologs have been identified in
countless other bacteria and in eukaryotes (reviewed by Stassen et al. 1997, Vergunst
and Hooykaas 1999), including the yeast Rad51 protein, which has been characterized
in depth. Two homologues of RecA have been found in Arabidopsis, one of which is
targeted to the chloroplast (Cerutti et al. 1992). A chloroplast protein
immunologically related to RecA and induced by DNA damaging agents was detected
in pea and Chlamydomonas reinhardtii, and a RecA-like strand transfer activity was
present in the stromal extracts of pea (Cerutti et a1. 1992, Binet et al. 1993).
Mammalian mitochondria also harbor a protein related to RecA (Thygarajan et al.

1996). The presence of RecA-like proteins within organelles suggests that

16

homologous recombination probably occurs there, possibly for the primary reason of

recombinational repair of DNA damage (Sears 1998).

In a study that clearly shows the importance of RecA in plant chloroplasts,
Cerutti et al. (1995) used the knowledge that particular missense mutants and N-
terminal truncations of RecA can inhibit the activity of the wild-type protein in E. coli.
The ability to have a dominant negative impact is characteristic of proteins that
function as multimers; in this case, the N-terminus of RecA is involved in protein-
protein interactions, and the alterations prevent the formation of an active filament
(Lauder and Kowalczykowski 1993). One such dominant negative mutant of RecA
lacks the first 42 amino acids (AN42RecA, Horii et al. 1992); it was tested by Cerutti
et al. (1995) to see if it would interfere with the endogenous homologous
recombination system in the chloroplast of Chlamydomonas reinhardtii.
Transformation by bioloistic bombardment was used to integrate the wild-type recA
and AN42recA alleles of E. coli into the inverted repeat of the chloroplast. Both of
these alleles and a null allele that contains an internal deletion (int-recA), were
engineered to be under the control of plastid regulatory elements. Under normal
conditions, the expression of these alleles did not interfere with any essential
component of cell growth and survival. An in vivo assay based on homologous
recombination restoring the frmction of the chIL gene was used to evaluate whether
expression of the recA alleles affected chloroplast recombination. While the control
int-recA transformants showed the same level of recombination as non-transformed

cells, the AN42recA transformants showed a S-fold decrease in homologous

17

recombination and wt-recA transformants showed a lS-fold increase in plastid DNA
recombination. These results strongly support the existence of a RecA-pathway for

homologous recombination in chloroplasts.

While the role of RecA has been explored in chloroplasts, the mechanism by
which recombination occurs in plant mitochondria is unknown, despite the large
number of recombinationally produced mtDNA mutants. Recently, a mitochondrially
targeted RecA gene product was identified in Arabidopsis thaliana through bio-
inforrnatics (Brent Neilsen, Brigham Young University personal communication), but
its role in recombination and production of debilitating disease phenotypes has yet to

be explored.

Hypothesis

The main hypothesis of this research is that the homologous recombination
pathway functions not only to produce a dynamic population of mitochondrial
subgenomes but also contributes to aberrant mitochondrial DNA rearrangements. To
test this hypothesis, the goal was to create transgenic plants in which the levels of
homologous recombination were stimulated or inhibited by expression of a bacterial
RecA gene or its dominant negative allele. Specifically, this dissertation will address
whether over-expressing the E. coli RecA protein produces the phenotypes of

cytoplasmic male sterility, variegation, distorted leaf shape and overall stunted growth.

18

Significance

It has become increasingly apparent that the dynamic nature of the plant
mitochondrial genome depends upon recombination. By testing the hypothesis that
homologous recombination (mediated by a RecA-type protein) is responsible for
aberrant mtDNA rearrangements, this investigation will increase our understanding of
the genetic processes that lead to abnormal mitochondrial function. As reviewed in
this Introduction, in at least some cases, mitochondrial gene expression can be
controlled by recombinational activity, followed by amplification of recombined
molecules in a population. It is important to understand the firndamental genetic
mechanisms that underlie this phenomenon, since mitochondrial gene expression has

ramifications for plant growth and development.

19

CHAPTER 2

Materials and Methods

Experimental Organisms and Grth Conditions

Arabidopsis thaliana

This study used Arabidopsis thaliana L. variety Columbia and the chm line
(Redei 1973). They were grown under standard conditions (Katavic et al. 1994) under
continuous light at 26°C. Plants used for transformation were treated as described by
Clough and Bent 1998, including clipping after the first flowers appeared to induce more
bolting. T2 plants were generated fi'om seeds of the T1 plants. Briefly, seeds were
surface-sterilized by a five minute rinse with 95% ethanol, followed by two lO-minute
rinses with 50% bleach + 1 pl of 10% SDS (1 ml), and then six subsequent rinses with
sterile water. About 100 seeds were then sowed on 100 x 15mm Petri plates containing
Murashige and Skoog media (MS; Table 2.1, Murashige and Skoog 1962) or MS + 80
rig/pl of kanamycin, and placed under continuous light at 26°C. Kanamycin-resistant

plants were transferred to soil and grown as described above.

Tobacco

Seeds of Nicotiana tabacum variety Turkish Sarnsun were a gift fi'om Dr.
Kenneth Sinks’ laboratory. Initial plants used for leaf disc transformation (described
below) were from tissue culture stocks maintained on MS medium. Seeds fi'om the T,

and BC 1 progeny were surface-sterilized as described above. About 100 seeds were then

20

Table 2.1: Murashige and Skoog Medium“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Component Final Concentration (mg/L)
NH4NO3 1650.0
KNO3 1900.0
MgSO4 - 7 H20 370.0
MnSO4 - 4 H2O 22.3
ZnSO4 - 4 H20 8.6
CuSO4 - 5 H20 0.025
CaCl2 - 2 H20 440.0
KI 0.83
CoCl2 - 6 H20 0.025
KH2PO4 170.0
H3303 6.2
NaMoO4 - 2 H20 0.25
NaF e EDT A 27.85
Nicotinic acid 0.5
Pyridoxin-HCl 0.5
Thiarnin HCl 1.0
Glycine 2.0
Myoinositol 100.0

 

 

 

 

*To each liter, 20 g sucrose and 8 g of agar was added and the pH was adjusted to 5.8

with 1M KOH.

21

sowed on Petri plates containing MS or MS + 80 rig/pl of kanamycin under continuous
light at 26°C. Once seedlings had their first true leaves, they were transferred to soil and

maintained in the growth room under continuous light. At the 6-leaf stage, plants were

transferred to 8 inch pots and maintained in the greenhouse.

Other organisms:

The E. coli, Agrobacterium tumefaciens and Chlamydomonas reinhardtii strains
used in this research can be found in Table 2.2. Unless noted to be different, bacteria
were grown overnight at 37°C in Luria-Bertani medium (LB; 10 g tryptone, 5 g yeast
extract, and 5 g NaCl in l L), either as 5 ml cultures or on agar plates. Agrobacterium
cultures for transformation were grown in either 5 ml or 250 m1 of LB at 26°C overnight.

Chlamydomonas cultures were grown using standard procedures decribed by Harris

(1939)

Transformation

E. coli and Agrobacterium
Washed E. coli cells were prepared for electroporation according to instructions
from Biorad. One to two microliters of ligated DNA was added to 40 ul of cells and

stored on ice. Cells were transferred to an electroporation cuvette and shocked with 25
uF, 2.5 kV, 200 Q (Biorad Pulser) of electricity. One ml of LB was added to the cells
and then they were incubated for one hour (37°C), followed by plating on selective

Table 2.2: Bacterial, Agrobacterial and Chlamydomonas strains

22

 

 

 

 

 

 

 

Strain Organism Experiment Key Feature
DHSa E. coli Cloning Highly transformable
Recombination Assays
HBlOl E. coli Recombination Assays Recombination deficient
(RecA minus)
GM48 E. coli Cloning Contains E. coli RecA
gene
GV3101 Agrobacterium Transformation Contains binary vectors
cc3455 Chlamydomonas Cloning Contains E. coli
PCR ANRecA

 

 

 

 

23

 

media The plates were incubated at 37°C for a minimum of 16 h. and potential
transformants were verified by plasmid isolation and DNA digestion. To verify that

electroporation was successful, an undigested plasmid control was performed with every

experiment.

Arabidopsis.

Arabidopsis plants were transformed using the floral dip method of Clough and
Bent (1998). Agrobacterium strain GV3101 with the binary vectors pB1121, pBIlZl-
RecA, pBIlZl-ANRecA, (250 ml in LB) was grown, centrifuged and resuspended in 500
ml of ARABIDIPT" solution (5% sucrose, 0.05% Silwet L77; OSI specialties).
Arabidopsis plants that had begun to bolt were dipped in ARABIDIPT“ solution for five
seconds. Pots of plants were placed on their side, covered in saran wrap and placed
overnight in a growth chamber (26°C). Then plants were placed upright in the growth
room, grown to maturity and seeds were harvested by standard methods. Transformation
was confirmed by seed assays for kanamycin resistance and polymerase chain reaction

experiments described later in this chapter.

Tobacco

Tobacco plants were transformed using a tissue culture based leaf disc protocol.
Liquid cultures (5 ml) of Agrobacterium containing the strain GV3101 with the binary
vectors pB1121, pBIl21-RecA, pBIlZl-ANRecA were grown. Dilutions (1:20) of these
cultures were made with MS medium for transformation. Young expanding leaves of

plants grown on an agar medium were removed and cut into sterile 0.5 X 0.5 um

24

explants. These explants were submerged in the bacterial suspension for five minutes
and then blotted briefly on sterile filter paper. Five leaf explants were placed upside
down and incubated (26°C, continuous light) on co-cultivation media (MS + 1 mg/l BAP
and 0.1 mg/l 1AA). After two days, explants were placed right side up on MS + timentin
(300 ug/l) + kanamycin (100 ug/l) media and incubated under the same conditions. The
timentin selects against the bacteria, and the kanamycin selects for the transformants. As
the shoots appeared, they were removed from the explants and sub-cultured to induce
rooting. At the 4-6 leaf stage, explants with generous roots were moved to 2 x 2” inch
multipots and grown under continuous light. Transformation was confirmed by PCR.
Plants confirmed by PCR to have the RecA gene or ANRecA constructs were transferred
to the geenhouse and grown to maturity. The progeny of these plants were also tested
for the presence of the transgenes by germinating seeds on medium containing

kanamycin.

Kanamycin Seed Assays

At least 100 seeds were sown on MS media + kanamycin (80 ug/l). Seeds were
allowed to grow for two weeks (Arabidopsis) or three weeks (tobacco). Seedlings that
were green and growing were considered kanamycin-resistant in Arabidopsis. In
tobacco, all seedlings had green cotyledons, but plants that were resistant to kanamycin
continued to grow and produce true leaves. Therefore, only plants that produced at least
three true leaves were scored as transformants. Under the assumption that true

transformants should generate a 3:1 ratio (transformed: untransformed) in the first

25

generation progeny, the number of kanamycin-resistant and sensitive plants were

subjected to a Chi-square test.

_BLant Phenotypic Analyses

During the three to four months that the tobacco plants were in the greenhouse,
multiple traits were monitored and measured. For T1 plants, average height was
measured (cm) after flowering had ceased and the number of stems that each plant
produced was also assessed. Flowers and leaves were visually inspected and assessed
every other day and were tagged by abnormality type that they expressed. Flowers were
generally tagged at stage 12 (Koltunow et al. 1990). When crosses were performed,
flowers were emasculated at stage 11. Seed pods were allowed to mature on the plant
and then harvested. All the pods were counted and opened to determine the number of
pods that were fallow. Seeds were quantified per flower by weighing them in Eppendorf
tubes (data not presented). All the data on phenotypic characteristics were analyzed in

Excel or through SAS.

Pollen Germination Assays

Pollen germination assays were developed using the protocol of Kearns and
Inouye (1993) as a basis. Anthers were removed from plants at floral stage 12 (Koltunow
et al. 1990), when pollen was being released, and placed in Eppendorf tubes with 40 ul of

pollen germination medium (50 g sucrose, 0.05 g KNO3, 0.05 g H3BO3, 0.15 g

26

Ca(NO3)2, 0.10 g MgSO4, in 500 ml). Pollen grains were mixed to insure even
resuspension and 10 ul aliquots were removed and placed on glass slides. Pollen was
incubated overnight in germination chambers made of Petri plates and moistened filter
paper to provide high humidity and promote germination. For each anther, a minimum of
two 10 ul samples were counted under the microscope and any sign of a pollen tube was
considered an indicator of germination ability. Counting was facilitated by the use of a

grid-marked cover slip. For each plant, the pollen of a total of five anthers was counted.

Polymeraae ChainRgrction

Polymerase Chain Reactions (PCR) were performed according to the
manufacturer’s instructions (Invitrogen). 1 pl of undiluted DNA or 1 ul of a 1:10 dilution
of total plant DNA was used as a template in each reaction; for all other sources of DNA,
1 u] was used as a template. In a 25 pl total volume, the following constituents were
added to produce a final concentration of: IX PCR buffer, 0.2 mM each dNTP, 1.5 mM
MgCl2, 5 pmole primers, and 2.5 U of Taq polymerase. All PCR assays included a
negative control to which no DNA was added and control reactions with DNA lmown to

amplifiy. The protocols, primers, and the expected fragment sizes are found in Table 2.3.

Program EPl consists of these steps: denaturation at 94°C for 3 minutes, followed
by these steps: 92°C for l min, 58°C for 1 min, 72°C for 2 min, repeated 29 times.
Protocols were completed by a 72°C extension for 10 min Program MZl consists of
these steps: denaturation at 94°C for 3 minutes, followed by these steps 94°C for l min,

50°C for l min, 72°C for 2 min, repeated 29 times. Protocols were completed by a 72°C

27

Table 2.3 Primers used in Polymerase Chain Reaction Experiments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Name Sequence 5’-3’ Primer Program Size &
Paired gene
with
WtRecA
RecAl CGGGTCGACAGGAGTAAAAATGGGATCCGACG
Sall Baml-II l 100 bp
RecA3 EPl
ATCT CT ACCGGATCCCTIT CACT G G ANRecA
RecA2 BamHI
978 bp
RecA3 ACATCTAGAT'I‘AAAAATCT’I‘CGTTAG’IT
Xbal
NA for seq. NA
RecA4 TCI‘GTTCGTCTCGACATCCC
WtRecA
429 bp
RecA5 CCI‘GCACGCTCAGCGTGGT EPl
ANRecA
303 bp
Target3 WtRecA
1200 bp
EPl
RecA6 AGAGCTCGATTAAAAATC’ITCGTTAGTT ANRecA
SacI l 100 bp
[3 F r
Targetl TACGAATTCGAGCI‘ATGGCI‘T Atpase
EcoRI Target 2 EP1 targeting
sequence
Target2 TGCGGATCCAGCGTACTGTACG 193 bp
BamHI
RecA5 EPl See above
Target3 TACGAACCCGGGCTATGGCIT
Small RecA6
ATP21- CT GTGCI'I‘ ATTATGGAACT EPl Atp2-l
F ATP21-
ATP21- CCGACAGCAGATGGGATACGAC R 702 bp
R

 

Underlined nucleotides are the engineered restriction endonuclease cut sites discussed in

the text.

28

 

extension step for 10 min All programs were run in an MJ Research PTC-200

thermocycler in 0.2 ml thin walled tubes.

To analyze the PCR reactions, 12 pl of the total 25 u] were analyzed by agarose
gel electrophoresis as described below. When PCR products were digested for cloning,

the reactions were scaled up to 50 ul and between 12-25 pl were used in digestions.

Cloning of Plasmija

RecA constructs were created in three vectors: pACYl84 was used to check the
function of RecA in E. coli, while two binary vectors pMIT-GSyl (Hemon et al. 1990)
and pBIlZl (Clonetech) were used to create the construct ultimately used for
Agrobacterium transformation. The E. coli RecA and its AN42 derivative were
amplified by PCR using primers engineered with restriction endonuclease cut sites
available in these vectors. The primer sequences and their cut sites are shown in Table
2.3. PCR products were cloned first into the vector pACYl84 because of the way in
which the mitochondrial targeting sequence was to be joined to the RecA gene. To
achieve in-frame insertion of the RecA gene, a BamH I site was generated by the PCR
primer, but this changed the identity of the second and third amino acids. To ensure that
this would not abolish or decrease the function of the protein, the pACY-RECA plasmid

was used for recombination assays as described below.

To create the pACY-RECA and pACY-AN42RECA constructs, PCR was used to
retrieve the genes from wild-type E. coli (GM48) and from a Chlamydomonas (cc3455)

line that contained the AN42RecA gene (Cerutti et al. 1995). Primers Recl (Sal I and

29

BamH I sites) and RecA3 (Xba I) were used to amplify the wild type RecA while primers
RecA2 (BamH I) site and RecA3 were used for the AN42RecA. The plasmid pACYl84
confers both chloramphenicol and tetracycline-resistance and the location of Sal I, BamH
I and Xba I sites are such that digestion with these enzymes removes part of the
tetracycline gene. Therefore proper clones could be identified by growth on

chloramphenicol and not tetracycline, with verification by restriction digestions.

A two step cloning process was required to obtain the construct which contained
both the B-Fr ATPase mitochondrial targeting sequence (Nicotiana plumbagnifolia,
Hemon et al. 1990, Boqu and Chua 1985) and the appropriate RecA genes, because cut
sites that were amenable for joining the targeting sequence to RecA were found in the
vector pBIl21. Oligonucleotides used for PCR are listed in Table 2.3. The 193-bp B-Fl
ATPase mitochondrial targeting sequence was obtained through PCR using the primers
targetl (EcoR I site) and target2 (BamH I site) and the pMIT-GSyl plasmid as template.
Next, the vector pMIT-GSyl was digested with EcoR I and BamH I to remove the entire
glutamine synthase insert fiom the T-DNA region. Ligation was used to insert the 193 bp
B-F. ATPase mitochondrial targeting sequence into this vector, creating a new vector
containing the targeting sequence in the T-DNA region (ptarget). Next PCR products of
the RecA and AN42RecA genes were generated using the same combinations of primers
used for the pACYl84 cloning described above. These constructs and the vector ptarget
were digested with BamH I and Xba I and then ligations were performed to insert each

individual construct of the RecA genes into the ptarget vector. These reactions created

30

the vectors ptarget-RecA and ptarget-AN42RecA. Due to the location of the cut sites

however, these genes were not in the proper orientation for expression in Agrobacterium.

To transfer the newly created, mitochondrially—targeted RecA constructs to the
vector pB1121, primers were developed with appropriate restriction endonuclease sites:
target3 (Sma I) and RecA6 (Sac I). In the vector pBIlZl, a Sma I site is found 5’to the
CaMV promoter and a Sac I site is situated 5’ to the NOS terminator; digestion with these
enzymes removes the GUS gene for replacement by the mitochondrially—targeted RecA
constructs. PCR was used to amplify the genes with the proper cut sites, followed by
restriction digests and ligation of the pBIlZl vector and PCR products. These vectors are
known as pBIl2l-RecA and pBIl21-ANRecA. pBIl21 is illustrated in Chapter 3. All
plasmid clones were verified by restriction digestions and PCR and transferred to

Agrobacterium via electroporation for use in transformation experiments.

Ligation

Ligations were performed via the manufacturer’s protocol (Invitrogen). A
minimum ratio of 3:1 (insertzvector) was used in all experiments and contained
appropriate amounts of plasmid DNA and product to achieve this PCR ratio. Ligation
reactions were incubated at room temperature overnight. A set of parallel incubations
was always performed as negative controls for the ligation procedure using doubly
digested vector both with and without ligase enzyme. After incubation, DNAs were
precipitated with 1 ul glycogen, 0.1 volume of 5M NaCl and 2 volumes ethanol for one

hour at -20°C. After centrifugation and two 70% ethanol rinses, ligation products were

31

resuspended in 5 ul water. Agarose gel electrophoresis of 2 ul was used to examine the

ligation products prior to transformation.

Recombination Assays

The plasmid pDK8 (Y amamoto et al. 1996) in E. coli was used to assess the
fimction of the RecA gene when its second and third amino acids had been changed due
to the addition of a BamH I site. This plasmid contains two defective neomycin (neo)
genes, one with an amino end deletion and the other with a carboxyl end deletion. These
genes are in direct orientation to one another. This plasmid contains an ampicillin-
resistance gene and when recombination occurs a fimctional neo gene, which confers
kanamycin resistance, is produced. Hence the recombination frequency can be measured
as the frequency of kanamycin-resistant cells out of the total number of arnpicillin
resistant cells. HBlOl recombination deficient E. coli were transformed such that they
contained the plasmid pDK8 alone or in tandem with either pACY-RECA or pACY-

AN42RECA plasmids.

Cells were grown in 5 ml cultures with 50 ug/ul arnpicillin for ~5 h (ODwo of 0.4-
0.6). Serial dilutions of bacteria were made fi'om 10'l to 103 and 500 pl of each dilution
was plated on LB + 50 ug/ul arnpicillan and on two LB + 50 ug/pl kanamycin (two
replicates for each). Bacterial colony numbers from these plates were then used to find
an appropriate dilution at which the number of antibiotic resistant colonies could be
clearly counted. This procedure was repeated on ten independent lines carrying the

plasmid(s) and an average recombination rate was determined.

32

For each trial plasmid, DNA was isolated and the presence of pDK8 and/or pDK8
with pACY-RECA or pACY-AN42RECA plasmids was verified. Once it has
recombined, the pDK8 plasmid has a different restriction digestion profile when out with
EcoR I. If it has not recombined, it should yield 3 bands (7.0, 6.0 and 2.3 kb), and after
recombination only one band (5.5 kb). Distinct banding patterns also exist for the pACY-
RECA (3.0, 1.7 kb) or pACY-AN42RECA (2.6, 1.7 kb) plasmids when digested with
EcoR 1. Hence the presence of the plasmids and the recombination status of the target

plasmid were assessed by restriction digestion with this enzyme.

DNA Isolation

Plasmid DNA

Plasmid DNA was isolated via the alkaline lysis procedure described by
Sambrook et al. (1989). The supernatant from the plasmid extraction was subjected to a
phenolzchloroform (25:25) extraction prior to ethanol-precipitation. Washed and dried
plasmids were resuspended in 10 to 50 pl of sterile water and stored at 4°C. A list of the
plasmids used and their characteristics can be found in Table 2.4 and their reference

information is found when they are mentioned in the text.

33

Table 2.4: Plasmids used in experiments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Plasmid Host Experiment Key Features
Used Organism
pACY 1 84 E. coli Recombination Chloroamphenicol resistance;
Assays Able to be co-maintained with
pDK8
Had appropriate restriction sites
to accept RecA constucts
pDK8 E. coli Recombination Ampicillin resistance;
Assays Contains the neomycin gene
(neo) separated such that
recombination will generate
kanamycin resistance
pMIT-GSy E. coli Cloning Contains targeting sequence of [3
PCR Fl-ATPase
pBIlZl E. coli Transformation 358 CaMV promoter
Agrobacterium Provides kanamycin resistance
pBIlZl-RecA E. coli Transformation 35S CaMV promoter
Agrobacterium Provides kanamycin resistance
and
RecA construct
pB1121- E. coli Transformation 35S CaMV promoter
ANRecA Agrobacterium Provides kanamycin resistance
and
ANRecA construct
pmt-sylsa8 E. coli Southern probe Random mtDNA partials (SalI
digest)
TB#5-1 E. coli Southern probe 26S rDNA (13 kb)
TB#6-1 E. coli Southern probe SS and 18S rDNA, rp116
H10/59 E. coli Southern probe randomly cloned mtDNA from
Oenothera
H2/1 E. coli Southern probe cox]
H1/23 E. coli Southern probe alpha su-A TPase
7G1 E. coli Southern probe nad4L, nad5c
0RFs:106, I61, 139, I 10, 25,
122c
39E9 E. coli Southern probe rrnZ6, nad5de, nad9 ,rplI 6,
rps3ab, trnN, cob
ORFs: 153a, 107a ,131, 315, 206,
I43, 121a, 167, 116
26F4 E. coli Southern probe trnD, trnS, atp9, nad1 b, nad1c,
atp6-2, atpl
ORFs: 262, 105a, 251, 106fi 152.
106a, l I 1 b

 

34

 

Chlamydomonas DNA

A small amount of cells were scraped off petri plates and placed in 0.5 ml TEN
buffer (lOmM Tris-HCl, lOmM EDTA, 150 mM NaCl), resuspended by vortexing and
then centrifuged (10 sec). Cell pellets were then resuspended in 150 pl water with 300 pl
of SDS-EB buffer (2% SDS, 400 mM NaCl, 40 mM EDTA) and extracted with an equal
volume of chloroform:isoamyl alcohol (24:1). The aqueous phase was removed and the
DNA precipitated with 2 volumes of 100% ethanol (4°C, 30 min) and concentrated by
centrifugation (10 minutes) and 2 washes in 70% ethanol. DNA was dissolved in ~40 pl

of water.

Plant DNA

DNA was prepared from Arabidopsis and tobacco using a modification of the
method of Doyle and Doyle (1987). One gram of leaf tissue was ground in liquid
nitrogen and then resuspended in 3 ml CTAB solution (1.4 M NaCl, 0.1 M Tris-HCI pH
8.0, 20 mM EDTA, 2 % cis-trimethylammonium bromide, 2% polyvinyl pyrrolidone).
Extracts were incubated (1 h, 65°C) followed by addition of 10 pg/pl RNase A and
incubation at 37°C for 15 minutes. One phenolzchloroform and a minimum of two

chloroformzisoamyl alcohol extractions were performed on each sample. DNA was
precipitated by the addition of a 0.2 volume of 5M NaCl (final concentration of 2M) and

0.75 volume of 100% isopropanol. DNA was concentrated by centrifugation (8,000 rpm,
10 rrrin), rinsed twice in 70% ethanol, resuspended in 100-400 pl of sterile water and

stored at —20°C.

35

Restriction Endonuclefi Digstions

Restriction endonuclease digestions were performed following the manufacturers’
instructions. In general, <1 pg of DNA (plasmid/PCR products) was digested for cloning
experiments, whereas 2-10 pg of total plant DNA was used for Southern blotting.
Digestions were performed in 50—200 pl volumes, that in addition to DNA, contained:
1X appropriate salt buffer, one unit of enzyme and water. Digestions were incubated
overnight at the appropriate temperature and then precipitated with 1 pl glycogen, 0.1
volume of 5M NaCl and 2 volumes of ethanol at —20°C for one hour. After
centrifugation, digested DNA was resuspended in a rrrinimum of 15 pl and analyzed via

agarose gel electrophoresis.

Protein Purification

 

E. coli

Proteins were isolated fi'om E. coli DHSa cells. 5 ml cultures were grown to an
OD600 of 0.7-1.0. One milliliter of cells was centrifirged (500 g) and pellets were
resuspended in 2X sample buffer (6X= 7m] 4X Tris-HCI pH 6.8, 3 ml glycerol, 1 g SDS,
0.93g DTT, 0.0012 g bromophenol blue). 6 pl of the sample was run as a control for

western blots with the RecA antibody.

Tobacco
Proteins were isolated via a modification of the green leaf gradient mitochondrial

isolation protocol of Boutry et al. (1984). Five grams of green leaves from plants grown

36

in tissue culture or ten grams of etiolated seedlings were ground in washed sand and 50
ml grinding buffer (0.33 M sucrose, 50 mM Tris-HCl pH 8.0, 1 mM EGTA, 10 mM
KH2PO4, 0.2 % BSA, 0.04% B-mercaptoethanol) and then filtered through two layers of
miracloth. Extracts were spun at 5,700 rpm for 30 s. The supernatant was removed and
this step was repeated twice. Next, mitochondria were pelleted by centrifugation (12
min, 15,000 rpm) and resuspended in 400 pl of green-leaf resuspension buffer (0.4 M
mannitol, lOmM KH2PO4 0.2% BSA). Resuspended mitochondria were layered on
Percoll gradients (13%, 26% and 50%) and centrifirged (30 min, Beckrnan SW40,
20,000 rpm). Two milliliters of mitochondria were removed from the interface of the
26% and 50% percoll, resuspended in 8 ml of mitochondrial resuspension buffer (250
mM sucrose and 30 mM MOPS pH 6.8) and centrifirged (20 min, 11,000 rpm). Pellets
were resuspended in 1.5 ml mitochondrial resuspension buffer and centrifuged (15 min,
12,000 rpm). Pellets were dissolved in 20 pl mitochondrial suspension buffer and 4 pl of
standard loading dye and the entire sample was run on polyacrylamide gels. (one gram of
tissue typically yielded 0.9 ug/ pl protein). The entire extraction protocol was performed

at 4°C.

Gel Electrophoresis

Agarose gels
DNA was separated on agarose gels that were between 0.8%-1.5% depending on
the size of the expected fiagrnents. In general, gels for confirming DNA isolations and

those of PCR fi‘agments were 0.8% agarose (w/v), while gels of endonuclease digested

37

plant DNA used for Southern blotting were 1.0% (w/v) gels. All gels were made with 1X
TBE and run in 1X TBE + ethidiurn bromide at ~35 V for 2-8 hours. 2. DNA cut with
BstEII (American Allied Biochemicals) or the 123-kb marker (Invitrogen) were used to
assess size of the fragments on most gels. Agarose gels were documented using the

BioRad Imaging system.

Polyacrylamide gels

Proteins were separated on 12% polyacrylamide separating gels (1.5 M Tris pH
8.8, 10% SDS, 305 acrylarnide, 10% APS, TEMED) with 4 % stacking gels (0.5 M Tris
pH 6.8, 10% SDS, 30 % acrylarnide, 10% APS, TEMED). Gels were run using 1 X
running buffer (5X = 15 g/l Tris, 72 g/l glycine, 5 g/l SDS) for 5 hours at 40-60 volts.
Protein sizes were determined by comparison to 2 pl of pre-stained SDS-PAGE standards
(Bio-Rad). Protein samples were heated at 90°C for 10 min before being loaded onto

gels.

Gel Transferfi and Blotting
DNA transfer

DNA was prepared for transfer to a membrane by a modification of the protocol
found in Sarnbrook et al. (1989). Gels of restriction digests were incubated for 15 to 30
minutes in 0.2 N HCl to facilitate transfer of large DNA fragments. This was followed
by half hour incubations in denaturant (0.5M NaOH, 1M NaCl) and neutralization (0.5
Tris pH 7.4, 1.5 M NaCl) solutions, respectively. Nylon membranes (Osmonics Inc. 0.22

pm) were prepared for the transfer by pre-soaking in 2X SSC (20X SSC = 3M NaCl, 0.3

38

M Na-citrate). DNA transfer was accomplished by a standard stacking blot procedure
(Sambrook et al. 1989), in which DNA movement is by osmosis. Transfer was allowed
to occur for a minimum of 12 h, after which the membrane was dried for 1 h and exposed
to UV light (Stratagene Co., UV Stratalinker 1800). Membranes were either stored in

plastic wrap or used immediately for Southern blotting.

Southern blotting

All Southern blotting was performed through DIG chemi-luminescent Southern
hybridizations and a modification of the BMB Company’s protocol. Plasmids used as
probes are listed in Table 5.4 and are referenced in Chapter 5. Probes were made by
mixing 10 ng to 3 ug of cut plasmid DNA with 2 pl of 10X hexanucleotide mix, 2 pl 10X
DIG nucleotide mix, 1 pl Klenow enzyme and water to a final volume of 20 pl. Water
and plasmid were boiled first for 10 min, centrifirged and then the other components were
added. Reactions were incubated (minimum of 4 h at 37°C) afier which the probe was
precipitated (l h, -20°C) with 1 pl of glycogen, 2.5 pl of 4M LiCl2 and 75 pl of cold
ethanol. Probes were centrifuged (15 min, 12,000 rpm) and rinsed twice in 70% ethanol,
dried and resuspended in 50 p1 TE + 0.1 % SDS. Probes were boiled for 10 min prior to
use. For use, 25 pl of probe was added to 50 ml prehybridization solution. Probes were

either used immediately or stored at -20°C. Probes were reused for 2-3 months.

Membranes were prepared for hybridization with a minimrun of two hours

incubation in 50 m1 of prehybridization solution (5X SSC, 0.1% n-laurelsarcosine, 1.0%

SDS, 100 ug/pl sonicated salmon sperm DNA, 1.5 % BMB blocking reagent at 65°C).

39

After prehybridization, membranes were hybridized overnight (minimum of 12 h) at
65°C. After membranes had been hybridized they were washed twice (65°C, 15 min)
with 25 ml of 2X SSC, 0.1% SDS, followed by two washes (65°C, 15 min) with 25 ml of
0.5X SSC, 0.1% SDS. Membranes were then incubated at room temperature for five
minutes in Genius One solution (1M Tris pH 7.5, 150 mM NaCl), followed by a two-hour
blocking wash in Genuis Two (1% non-fat milk in Genius One). Next detection of the
probe was facilitated by a 30 minute incubation with anti-DIGoxigenin-AP FAB
fiagments (BMB, 1:10,000 dilution in Genius Two). Membranes were subjected to two
15 minute washes (room temperature) with Genius One and a single five minute wash
with Genuis Three solution (100 mM Tris-HCl, 100 mM NaCl, pH 9.5). Membranes
were then placed in acetate sheets and covered with 1 ml of CDP-STAR (BMB;1:100
dilution in Genius Three). Membranes were exposed to fihn for a minimum of one hour
to a maximum of overnight, depending on probe and DNA amounts, at 37°C. Blots were

often stripped of their probes by two washes (0.2M NaOH, 0.1% SDS) at 37°C; they

were either reused immediately or stored in 2X SSC at 4°C.

Protein transfer

The polyacrylamide gels were prepared for transfer by a 15 min incubation in
transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, 0.005% SDS). Nylon
membranes (Millipore, Immobilon-P, 0.45 pm) were prepared by a 10 min incubation in
methanol followed by three washes with sterile water and a five minute incubation in

transfer buffer. Proteins were electrically transferred using the Biorad Trans-Blot and the

manufacturers instructions (30V, 4°C, minimum of 12 hours).

40

Western blotting

Western blotting was performed via a modification of the method recommended
by the Pierce Company. Membranes were incubated first in 50 ml of 1X TBS-Tween
(20X=48 g Tris, 175 g NaCl, 20 ml Tween-20 in one liter) for five minutes, followed by
a 1 hour blocking wash in 1X TBS-Tween, 5% dried milk and 0.5% Tween-20. Next,
membranes were incubated with primary antibodies for 1 hour; dilutions were made in
1X TBS-Tween and are listed in Table 2.5. Incubations with the secondary antibody
(HRP conjugated anti-mouse antibody, Pierce) were for 1 hour and dilutions were made
in 1X TBS-Tween and are listed in Table 2.5. Each of these two incubations was
immediately followed by six ten-minute washes in 1X TBS-Tween. After the final wash,
detection was achieved through incubating membranes in SuperSignal West Pico
Substrate for five minutes as suggested by the manufacturers instructions. Membranes
were exposed to fihn for 2 to 30 minutes depending on antibody and protein

concentrations. Blots were stripped by the manufacturer’s direction.

41

Table 2.5: Antibodies used for Western blot analyses

 

 

Antibody to Mono- or Dilution Source/Reference
Polyclonal Used
E. coli RecA monoclonal 1:100,000 Leinco Technologies (R147)

Mouse anti-RecA
Arm321 (Thrgq-GIU127)

 

Maize HSP70 monoclonal 1:5,000 Gift from Dr. Christine Chase
(PM004)

original source: Lund et al. 1998

 

 

 

 

 

 

 

Tobacco AOX monoclonal 1:5000 Gift from Dr. Lee McIntosh
COX2 monoclonal 1:5000 Gift From Dr. Lee McIntosh
Mouse IgG (H + L) secondary 1:200,000 Pierce Company
antibody Goat Anti-Mouse
HRP Conjugated

 

 

42

CHAPTER 3

Creation of RecA Constructs and Plant Transformation

INTRODUCTION

The RecA protein is the first known member in the family of proteins that have
the capacity to catalyze DNA strand-exchange reactions. RecA is a multi-functional
protein in bacteria, but is primarily described as a DNA-dependent ATPase and an ATP-
dependent DNA binding protein. In E. coli, activities of RecA include recombinational
processing, induction of the SOS response and bypass of mutagenic lesion (Lusetti and

Cox 2002).

Bacterial homologues of RecA are highly conserved not only in their function but
also in their polypeptide sequence. When the E. coli RecA sequence is compared to the
sequence of 67 bacterial RecA homologues, the percentage of identical amino acids
residues ranges fi‘orn 49% (Mycoplasma pulmonis) to 100% (Shigella flexneri).
Additionally, among the sequences for RecA homologues that were examined, many of
the non-identical residues retained chemical similarity. Through these alignments, the
core domain of the RecA protein has been identified to encompass amino acids 34-269.
This domain contains the regions involved in monomer-monomer formation, ATP-
binding and DNA binding. Both the N—terminal and C-terrninal of the RecA protein are
more variable in sequence structure than this core domain (reviewed by Lusetti and Cox
2002). These features were considered in the initial experiments that created the

mitochondrially—targeted E. coli RecA and AN42RecA constructs. Procedures for

43

selecting the transformants and the verification of their composition are also described in

this chapter.

RESULTS

Construction of clones with wild-type and dominant-negative forms of E. coli RecA
preceded by a mitochondrial transit peptide sequence

The wild type RecA gene was retrieved by PCR from E. coli. The N-terrninal
deletion variant, which lacked the first 42 amino acids, was PCR amplified from C.
reinhardtii that aheady carried the construct (Cerutti et al. 1995). The primers used for
PCR retrieval were engineered with restriction enzyme cut sites that would allow for
cloning into the Agrobacterium tumefaciens vector pBIlZl . As described more fully in
Chapter 2, this involved a two-step cloning process that joined the Nicotiana
plumbagnifolia B-Fr ATPase targeting sequence and E. coli RecA genes in a modified
version of pROK2 vector, pMIT-GSyl first (Boutry and Chua 1985; Hemon et al. 1990),
followed by PCR amplification of the entire constuct for placement into the vector
pB1121. A schematic representation of the wild-type RecA clone can be seen in Figure
3.1 . Amplified DNA constructs were checked via restriction digestions, PCR and

sequencing (data not shown).

——-’ —-P

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RB target3 _. L8
358
- Nos-P NPTII Nos-T CaMV Target RecA Nos-Tl
‘ 4—RecA6 V
Hindlll BamHI Sac! EcoRl
Smal

Figure 3.1 Schematic representation of the T-DNA region of the wild-type RecA clone
in the Agrobacterium expression vector pB1121. The NPTII gene is expressed with the
NOS promoter and terminator providing a means of selection through kanamycin
resistance. The RecA clones have been fused to the nritchondrial targeting sequence
from the B—F 1 ATPase, with expression driven by the cauliflower mosaic 358 viral
promoter. Arrows above show the direction of transcription, LB and RB in this diagram
refers to the left and right borders of the T—DNA region. The important enzyme cut sites
for insertion and the primers used for PCR verification are indicated.

45

Recombination assays to verifii the activity of the cloned modified RecA gene

The cloning of the wild-type RecA construct included altering the second and
third codons to create a BamH I cut site so that the targeting sequence could be added and
the reading flame maintained. This modification changed residues from alanine and
isoleucine to glycine and serine, respectively. Because of this, it was necessary to verify
that the altered protein retained activity. This was monitored through homologous
recombination assays performed in E. coli as described by Yarnarnoto et al. (1996). The
in vivo assays use a plasmid, pDK8, which contains the gene that confers resistance to
ampicillin, and two partial neomycin genes (kanamycin resistance). One of the partial
neomycin resistance genes has a 248-bp arnino-terminal deletion, while the second has a
283-bp carboxy-terminal deletion. This leaves 506-bp of homology through which
recombination can act. If homologous recombination occurs, the resulting plasmid
changes in size, loses an EcoR I cut site and gains kanamycin resistance. Therefore the
amount of homologous recombination that occurs in a population of cells can be
measured through the determination of the proportion of colonies resistant to ampicillin
and the proportion that was resistant to both arnpicillin and kanamycin. The

recombination events can be verified through restriction digestions.

To use the pDK8 plasmid to test the ability of the altered RecA to catalyze
recombination, the PCR-generated RecA gene, without the mitochondrial targeting
sequence, was cloned into the vector pACYC184 to produce a construct called
pACYRecA. The vector carries chloramphenicol resistance and belongs to a different

plasmid compatibility group than pDK8. The pACYRecA plasmid was introduced into

46

E. coli cells of strain HBlOl, which were then transformed by electroporation with the
plasmid pDK8. Transformants were screened for both arnpicillin (pDK8) and
chloramphenicol (pACYRecA) resistance and ten independent lines were then tested for
the frequency with which kanamycin resistant cells were produced. For each trial,
plasmid minipreps were performed to insure that both plasmids were present in each

colony and to assess the plasmid structure by restriction digestion.

Results for these recombination assays are shown in Table 3.1. The E. coli strain
HBlOl is a recA’ mutant strain and, therefore, deficient in homologous recombination.
Trials were done to determine the rate of recombination in the orginal HB 1 01 strain, with
the progenitor plasmid pACYC184 as controls to compare with the pACYRecA (contains
the E. coli RecA gene) and pACYANRecA (N—terminal deletion RecA). The presence of
pACYRecA was found to increase recombination roughly ISO-fold, showing that in spite

of the two-codon alteration, the RecA protein was still active.

The activity of the dominant negative mutant of RecA had been previously tested
by Cerutti et al. (1995) to see if it would interfere with the endogenous homologous
recombination system in the chloroplast of Chlamydomonas reinhardtii. An in vivo assay
based on homologous recombination restoring the function of the chlL gene was used to
evaluate whether expression of the dominant negative RecA affected chloroplast
recombination. The AN42recA transformants showed a 5-fold decrease in homologous
recombination and suggested that a RecA based homologous recombination mechanism

exists in chloroplasts. In the HBlOl cells used, no firnctional RecA protein was present

47

Table 3.1 Effect of RecA plasmids on recombination frequency of pDK8 plasmid in E.

 

 

 

 

 

coli HBlOl.
Resident plasmid Recombination frequency Standard deviation
(per 105 cells)
None 0.86 0.45
pACYl84 1.16 0.78
pACY-RecA 161 .00 1 8.00
PACY-AN42RecA l .04 0.34

 

 

 

 

48

 

for which the dominant negative RecA could interact with and, therefore the lack of

activity by the dominant negative RecA was an expected result.

Confirmation of initial transformation

The transformation of Arabidopsis was achieved through the floral dip method
described in Chapter 2. The greatest number of transformants (16) was achieved with the
wild-type E. coli RecA in the Columbia variety of Arabidopsis, while lower numbers of
transformants were recovered from the three allelic chm lines (Table 3.2). In contrast,
only the Columbia line yielded transformants when the dominant-negative RecA was
used. For each plant, DNA was isolated and PCR was performed to verify the presence

of the transgenic RecA construct (data not shown for Arabidopsis).

Tobacco transformation was accomplished as described in Chapter 2 by the
Agrobacterium-mediated leaf disc method and subsequent plant regeneration. Individual
plants were regenerated from callus and all the regenerated plants represented a To
population. In total, twenty-one plants (representing sixteen independent calli) were
recovered that carried the E. coli RecA gene; while seventeen plants (representing fifteen
independent calli) had integated the dominant negative gene. Plants were labeled by
their callus number and when calli produced more than one plant they were distinguished
by letter. This is in contrast to Arabidopsis, in which To transformants were gown from

seed.

49

Table 3.2 Number of transgenic Arabidopsis plants recovered.

 

 

 

 

 

 

 

 

 

 

Plant Line Type of E. coli RecA # of transformants
transgene recovered
Columbia Wild-type l6
Dominant 5
negative
chm-1 Wild-type 3
Dominant 0
negative
chm-2 Wild-type 4
Dominant 0
negative
chm-3 Wild-type 2
Dominant 0
negative

 

 

 

 

50

Individual tobacco plants were regenerated on kanamycin medium, providing the first
method of screening for nuclear integration. In addition, DNA was isolated from these
plants and PCR was used to screen for the presence of the nuclear transgene. For PCR, 3
reaction which amplifies the nuclear ath-I gene was run as a control to assure that the
DNA was amplified by PCR Since ath-I is present in two copies in the genome
(Lalanne et al 1998b), this reaction produces two bands of ~700 bp. After DNA was
proven to be arnplifiable, PCR was performed with the primers target3 and RecA6
(primer sequence in Table 2.3). Amplification with these primers produces bands of
either 430 bp (W tRecA) or 300 bp (AN42RecA). For all plants, PCR confirmation of the
presence of the RecA construct was confirmed in three separate PCR reactions before a
plant was considered to be a transformant. PCR reactions from a selection of To plants

are shown in Figure 3.2.

In addition to confirmation by PCR, verification of the transformation event was
achieved by testing for kanamycin resistance. Seeds fi'om the To transformants were
plated on kanamycin, germinated and the seedlings scored for resistance or sensitivity. In
A grobacterium—mediated transformation, the entire T-DNA region integrates into a site.
In the case of tobacco leaf discs, a single event usually occurs and the resulting plant
tissue will be hemizygous. Since To plants set seed through self-pollination, the T.
progeny should follow a 3:1 Mendelian ratio for the inheritance of the nuclear transgene.
The genotype of the parent can be confirmed by assaying the T1 seedlings and performing
a Chi-square analysis. A p value geater than 0.05 supports the hypothesis that the data

matches a 3:1 ratio. Table 3.3 shows Chi-square results for the To plants 10B, 12, 16 and

51

.0 LO
8‘“ 09—“- r~-E
EiaEEEzeéaes

738
A 615

 

D
MW
neg
Wt +
AN +
Sam
W138
wt4
wt10b
wt12
AN1
(AN-4
AN?
AN15
Sam
Wt +

 

Figure 3.2 Agarose gel confirmation of tobacco To transformation.
On each gel PCR products from plants are labeled by the callus
number from which they were derived. A letter indicates a sample
fi'om one of several plants derived from the same callus. The
abbreviation Sam refers to control plants (Nicotiana tabacum var
Turkish Samsun) that underwent tissue culture but were not
transformed. The abbreviation neg refers to the PCR negative
control reaction and the abbreviation Wt+ and AN+ are plasmid
controls for amplification with the RecA primers. Molecular
Weight marker (MW) lane contains the 123 bp ladder from
Invitrogen.

(A) PCR with control atp2—1 primers

(B) PCR with primers target3 and RecA6

52

Table 3.3 Kanamycin-seedling assay for resistance. The T0 tobacco plants were allowed

to self-pollinate to produce the seed. Seeds from the wild-type control (Samsun) were
germinated on either kanamycin or non-selective medium (MS).

 

 

 

 

 

 

 

 

 

 

 

 

 

Total #
To plant seeds Kanamycin Kanamycin X2 p—value
germinated Resistant Sensitive
Sam 176 0 176 0 >099
(Ran)
Sam 114 -- -- -- All gew
(MS)
10b 130 88 42 4.1 0.10-0.25
12 377 293 74 1.5 0.90-0.95
16 83 59 24 0.6 0.95-0.97
AN7 97 65 32 3.5 0.75-0.90

 

53

 

AN7. These data are representative of the data for all To plants and confirmed that the

parent (To) plants were hemizygous for the Ti insertion.

Confirmation of the T1 plants

To tobacco transformants showed phenotypic abnormalities that will be discussed
in Chapter 4. It was necessary to evaluate the inheritance and expression of these
phenotypic abnormalities both in the presence of the transgene and without the transgene.
To do this, seeds from To plant 12 were germinated on non-selective medium and DNA
was isolated from 75 plants in an attempt to identify ~20 plants with the transgene and
~20 plants without the transgene. Each DNA sample was subjected to PCR with primers
for the native atp2-I gene and the RecA transgene. Figure 3.3 shows a representative
PCR gel showing products fi'orn plants in the line 12pet. Plants 31, 42, 44, and 45 do not
contain the RecA construct. As with the To plants, three PCR reactions were analyzed

before the RecA status of a plant was considered definitive.

To confirm these results, seeds fi'om randomly selected plants and from plants
that were to be used in the back-cross studies were collected and tested for the
segegation of kanamycin resistance. Table 3.4 shows these results. Since the To plant
was heterozygous (hemizygous), T1 progeny could be heterozygous, homozygous
dominant or homozygous recessive in relationship to the nuclear transgene. These
possibilities can be distinguished by the numbers of kanamycin resistant seedlings in
these assays. Chi-square analyses of the ratios indicate that each of these genotypes was

found in the T1 populations. Plants that were deduced by PCR to lack the transgene, were

54

gag E
QQOFFNSWONQ
Echmvsr V'VVU)

 

 

Figure 3.3: Agarose gel confirmation of T1 plants with and without the
transgene. On each gel, plants are labeled by their greenhouse number
and are from line 12pet. The abbreviations Sam, neg, Wt+, and AN+ are
all as previously described. Molecular Weight marker (MW) lane is the
123 bp ladder from Invitrogen.

(A) PCR with ath-I primers

(B) PCR with primers target3 and RecA6

55

Table 3.4 Kanamycin-seedling assay for resistance. Seeds from the T] tobacco plants
from the line 12pet were assayed.

 

 

 

 

 

 

 

 

 

Total #

T. seeds Kanamycin Kanamycin x2 p-value Tl plant
plant germinated Resistant Sensitive genotype

30 62 61 I 0.156 0.95-0.9 RR

3 1 72 0 72 0 >099 rt

41 107 74 33 1.78 0.10-0.25 Rr

42 23 0 23 0 >099 rr

44 148 1 147 0.06 >0.99 rr

45 87 1 86 .011 >099 rt

46 163 127 36 0.8 0.25-0.5 Rt

47 34 33 1 0.27 0.5-0.75 RR

 

 

 

 

 

 

 

 

 

56

found to be completely kanamycin sensitive confirming this designation. As with the
previous assays, Sarnsun seeds were plated on MS or kanamycin media as controls (data

not shown).

Confirmation of the BC 1 plants

To analyze the inheritance of the developmental aberrations, flowers on various
T 1 plants that lacked the nuclear transgene were crossed by pollinating them with the
Sarnsun line as the male parent. Seeds that were gown from these plants (31, 42, 44,
45) were designated BC1 for the first back cross generation. Ten progeny plants from
each of the four lines were gown, except for line 44, from which 12 plants were gown.
PCR was performed on all these plants to confirm that the RecA construct was not
present; a subset of the results is shown in Figure 3.4. None of these plants contained the
RecA construct. These plants were not examined by the seed assay for kanamycin

resistance.

DISCUSSION

The goal of the experiments described in this chapter was to create a
mitochondrially targeted RecA construct to use for plant transformation. In addition the
experiments verify the presence and/or absence of the transgene in the T0, T1 and BCl

plants.

57

 

Sam

 

 

 

MW
neg
Wt+
AN+
Sam
7

8

1

8

9
10

2

2

3
Sam
\Nt+

 

Figure 3.4 Agarose gel confirmation of tobacco BC] populations. Plants
on the gel are labelled by the T. parent plant and then a number from 1-10.
The abbreviations Sam, neg, Wt+, and AN+ are all as previously
described. Molecular Weight marker (MW) lane contains the 123 bp
ladder from Invitrogen.

(A) PCR with atp2-I primers

(B) PCR with primers target3 and RecA6

58

To maintain an intact reading frame for the E. coli RecA gene, while joining it to
the B-Fl ATPase mitochondrial targeting sequence, required altering the second and third
amino acids of the RecA protein. Before this construct was used in plants, the activity of
the modified RecA protein was tested through an in vivo recombination assay in E. coli.
When expressed from a plasmid in RecA-deficient E. coli cells, the modified protein was
able to catalyze recombination of the plasmid pDK8 at levels 150 times higher than in
cells that did not contain the RecA plasmid. The modified protein was considered to be
functional. This is consistent with observations of conserved regions of the E. coli RecA
protein (Lusetti and Cox 2002). Two separate studies have compared the protein
sequences of 67 bacterial strains and have found its core domain to be from amino acid
34-269 (Roca and Cox 1997, Karlin and Brocchere et al. 1996). The second and third
codons were not conserved among bacteria in either of these studies. The effectiveness
of the dominant negative RecA construct in inhibiting recombination was not tested,
because my construct did not alter the amino acid sequence of the protein that had been
shown to decrease recombination in E. coli and the Chlamydomonas chloroplast (Cerrutti

et. al., 1995)

After kanamycin selection for transformation, all the To transformed plants
(Arabidopsis and Nicotiana) were checked for the presence of the RecA constructs by
PCR amplification. After the plants had set seed, the results of the PCR were confirmed
by seed assays for kanamycin resistance and Chi—square analysis. In total, five
AN42RecA (Columbia), sixteen wild-type RecA (Columbia) and nine wild-type RecA

(chm) plants independent transformants in Arabidopsis were recovered. In Nicotiana,

59

twenty-one independent transformants with the E. coli RecA and seventeen transformants

with the dominant negative gene were recovered.

No phenotypic changes were seen in the Arabidopsis transformants and therefore
no firrther analyses were performed on these plants. However Nicotiana transformants
and their resulting Ti and BC1 generations were recovered and genotyped. Phenotypic

and molecular analyses of these plants will be described in Chapters 4 and 5.

CONCLUSIONS

The experiments in this chapter show that the RecA gene can accept changes in its

first two amino acids and remain fimctional. In addition, it verified the production of
transgenic Arabidopsis and tobacco lines carrying both the E. coli RecA and ANRecA

constructs.

60

CHAPTER 4

Phenotypic and Genetic Analyses of Transgenic Plants in which the E. coli RecA
Protein is Expressed in the Mitochondria

INTRODUCTION

In most eukaryotes, proper mitochondrial biogenesis and function is a
requirement because mitochondria produce the majority of the cell’s energy. In humans,
when mitochondria function improperly, a wide variety of defects can occur including
neurological disorders, degenerative diseases, diabetes, blindness and cancer (Wallace
1999, Schon 2000). These clinical problems involve a diverse array of tissues, which are
similar in that proper functioning of the tissue requires large amounts of energy
(reviewed by Wallace 1999, Schon 2000). While plants do not display mitochondrial
diseases in the clinical sense, improper mitochondrial frmctioning does produce
recognizable phenotypic syndromes. The following review will highlight the role of
mitochondrial energy production in plant gowth and development by exploring known

plant mitochondrial disorders.

Mitochondrial Disorders in Maize and Arabidopsis

Proper mitochondrial function in plants is required for growth, respiration,
development and especially reproduction (Nakajima et al. 1997). Plant mitochondrial
disorders are hallmarked by a distinct set of characteristics that affect reproduction and
gowth specifically. In maize, the mitochondrial nonchromosomal stripe mutants (NCS)

have been extensively characterized (Newton and Coe 1986). NCS plants were named for

61

the most prominent of their abnormalities: the development of variegated leaf sectors in
vegetative tissues. This is expressed most notably as leaf striping caused by abnormal
arrest of chloroplast deve10pment (NCSZ, Roussel et al. 1991). In addition to this
variegation, NCS plants also display poor overall growth and a decreased seed yield
(Conley and Hanson, 1995). Molecularly, the NCS phenotypes have been correlated with
deletions in the mitochondrial electron transport gene con (N CSS, NCS6 Newton et al.
1990), ribosomal genes rps3 and rp116 (NCS3, Hunt and Newton 1991) and nad4 (NCS2,
Mareinfeld and Newton 1994, Baker and Newton 1995). When homoplasmic, each of
these mutations is considered lethal as shown by the inability of the plants to live outside
of tissue culture. Therefore individual NCS plants are heteroplasmic in regards to their
mitochondrial DNA (Newton and Coe 1986, Baker and Newton 1995). In each case, the
severity of the phenotypes is correlated to the amount of altered mtDNA present (Gu et

al. 1993).

In Arabidopsis thaliana the nuclear allele, chloroplast mutator (chm), acts on
mtDNA to create maternally inherited mutants whose phenotypes are very similar to
those in the NCS family. Whereas individual NCS plants show a mixture of phenotypic
characters, chm mutants fall into two classes, which are inherited independently of one
another. In the first class white, yellow or light geen variegation is produced in the
leaves beginning as early as the cotyledon stage (Redei, 1973, Redei and Plurad, 1973).
The second class of mutants is characterized by an overall asymmetric gowth and
uneven leaf surface that generates leaves with a rough or ragged appearance (Redei 1973,

Redei and Plurad 1973 Sakamoto et al. 1996). These phenotypic characteristics caused

62

them to be named maternal distorted leaf (MDL). These plants also tend to have more
secondary shoots than wild-type Arabidopsis. In addition to the alterations in leaf shape
and overall gowth, MDL plants also display reduced fertility expressed as low seed set
per capsule, reduced total pollen per flower and elevated levels of defective pollen
(Mourad and White, 1992). In contrast these are not features of the variegated plants,

which show normal amounts of seed and pollen production (Redei, 1973).

As with NCS plants, both chm and MDL mutations appear to result from
mitochondrial DNA rearrangements or deletions in specific genes. In variegated plants, a
rearrangement was found that involved a region of DNA containing 01125 and exon c of
the nad5 gene (Martinez-Zapater et al., 1992). This rearrangement was found prior to
complete sequencing of the mitochondrial genome of Arabidopsis, which revealed that
the “rearrangement” is actually present in substoichiometric amounts in the master DNA
chromosome and available for amplification to produce the aberrant phenotypes. In
MDL plants, three types of mtDNA changes were detected: a deletion of the majority of
the gene rps3, a partial deletion of the gene rp116, and a rearrangement in which the rps3-
ml] 6 operon was inserted downstream of the gene atp9. Each of these changes was
found to be due to recombination at an ll-bp repeat unit (Sakamoto. 1996). As a
consequence, normal rps3 and rpll 6 transcripts were no longer detectable in MDL plants.
In addition, quantitative PCR analysis of these mutant plants showed that the degee of
MDL phenotype expressed was correlated with the abundance of the MDL-specific
rearrangements, along with the concomitant disappearance of the wild-type DNA

arrangement (Sakamoto et al. 1996). These findings led to two alternative hypotheses.

63

The first of these assumes that the CHM protein is involved in maintaining a high level of
master chromosomes that contain the complete mitochondrial genetic information. The
second postulates that the CHM protein is responsible for eliminating unusual molecules
that were created by aberrant recombination and cause physiological deficiencies

(Sakamoto et a1. 1996, Martinez-Zapater et a1. 1992).

Cytoplasmic Male Sterility

While NCS and chm represent species-specific mitochondrial disorders, the most
frequently found mitochondrial disorder in plants is that of cytoplasmic male sterility
(CMS). Occurring in over 150 plant species, CMS is defined as a maternally inherited
trait characterized by an inability to produce viable pollen (Laser and Iestem 1972,
Mackenzie et al. 1994). In most of the CMS types, the deficiency in pollen function is
due to an impairment of normal microsporogenesis or garnetogenesis (Kaul 1998) as
opposed to impairment of overall anther development. CMS is thought to occur more
fl'equently than other disorders because the development of the male gametophyte is a
highly energy intensive process in which optimal firnction of the mitochondria is crucial

(Bergnan et al. 2000, Hanson 1991, Levings 1993, Elkonin and Tymov 2000).

For several plant species, CMS-associated DNA sequences have been located in
the mitochondrial genome and are often due to chimeric open reading frames that were
created through recombination (Schnable and Wise 1998). Examples include: urfl3 in
maize (Levings 1990), S-pcf in petunia (Young and Hanson 1987), 0411522 in sunflower

(Kohler et al. 1991) and wheat 013956 (Song and Hedgecoth, 1994). In sterile anthers,

the transcript abundance of the chimeric open reading flames is geater than in fertile
anthers, and chimeric proteins produced fl'om the transcripts can be found (Breiman and
Galun 1990). In some of these systems, nuclear genes that restore fertility exist, and such
restoration is associated with a decrease in transcription of the chimeric genes and
chimeric protein production (reviewed in Elkonin and Tymov 2000, Breiman and Galun

1990).

Uniquely expressed CMS: Alloplasmic lines in Nicotiana and Daucus

In most plant species CMS is singularly expressed as an alteration in pollen
development and viability, a phenotype denoted here as classical CMS. However, in the
genera Daucus and Nicotiana, the definition of CMS has been expanded to include
altered floral architecture. In particular, an alteration in starnen and petal development
gives rise to abnormal phenotypes that resemble those caused by homeotic floral
mutations (Linke et al. 2003, Bonnet et al. 1991, Bergnan et al. 2000, Bomer et al.
1995). In plants, homeotic mutations are ones where the development of a particular
floral organ is misdirected to become the organ found in the next inner or outer whorl.
The mechanism for floral development is described by the ABC model in which the
expression of different combinations of MADS box transcription factors leads to the
differentiation of the four basic floral structures (Bowman et al. 1989). However, the
homeotic CMS phenotypes occupy a separate niche in the known pathways of floral
development since they result fl'orn mitochondrial dysfunction as opposed to nuclear gene

changes (Szkclarazyk et al. 2000).

65

In Daucus, both forms of CMS, homeotic and classical, are known to occur.
Classical sterility occurs in a line called brown anther (Welch and Grimball 1947, Banga
et al. 1964, Struckmeyer and Simon 1986, Michalik et al. 1971, Nothnagel et al. 2000),
which is characterized by brown deformed stamens. Physiologically, this is due to
tapetum degadation, with pollen development being blocked at anthesis in the late stages
of meiosis (Michalik et al. 1971, Nothnagel et al. 2000). Molecularly, this phenotype has
been linked to the production of a unique l7-kD mitochondrial protein that is synthesized

specifically in brown anther lines and not in normal fertile lines (Scheike et al. 1992).

Homeotic flower conversions related to CMS were first documented in wild
carrot (Daucus carota carota) in the early 1950’s and were referred to as petaloid
(Thompson 1961, McCollum 1996). The petaloid nomenclature was used to identify
multiple changes in stamen morphology including those that have been transformed into
petals, petal-like, bract-like or carpeloid structures (Thompson 1961, Eisa and Wallace
1969, Straub 1971, Chader and Frese 1981, Erikson et a1. 1982, Kitagawa et al. 1994).
These phenotypes were caused by an incompatible interaction between nucleus and
cytoplasm and could be mimicked through the production of cytoplasmic hybrids:
depending on the type of cytoplasm, different expressivities of petaloid phenotype could
be achieved. For example, when D. martinus is used as a cytoplasmic donor in crosses
with D. carota, cybrids with white-geen petals are produced (as opposed to naturally
occurring white petals) and the stamens appear as petals or spoon-like structures. In
contrast, when D. gadecaei provides the donor cytoplasm, petal color is unchanged but

instead petals are reduced in size. Anthers in this cybrid do not form except as

66

rudimentary filaments. In cybrids with D. gumifer cytoplasm, flowers were produced
that had no petals, no stamens, and a severe reduction in sepals. Interestingly, while all
the other floral organs were affected in these plants, the central gynocieum was left
unmodified. To date molecular studies on these and other types of petaloid CMS carrot
lines have found two distinct mitochondrial changes that are linked to these phenotypes.
In multiple alloplasmic petaloid lines, Szklarczyk et al. (2000) have found a rearranged
version of the atp9—1 gene associated with CMS. In a separate study of seven different
CMS petaloid lines, Nakajirna et al. (2001) identified a rearranged version of the oer
gene region (termed ortB-CMS) that is only found in petaloid CMS lines. In these lines,
CMS has been correlated to the expression of an orfB-CMS specific protein that

accumulates in flowers of CMS plants but not in the leaves.

While Daucus CMS lines represent one source of florally abnormal CMS lines,
the number of Nicotiana cytoplasms that cause male sterility is extraordinary; in fact, the
only genus with a comparable number of sterile cultivars is Solanum (Kofer et al. 1991).
The availability of these distinct CMS lines led to the conclusion that Nicotiana materials
reveal a flrller extent of mitochondrial involvement in flower development than other
species and, therefore, offers a favorable material for the study of cytoplasmic control of
floral development (Bonnet et al. 1991, Kofer et al. 1991). Classical CMS, characterized
by production of non-functional pollen by morphologically normal stamens, is known to
occur in Nicotiana (Hart 1965, Sand and Christoff 1973, Gerstel 1980, Edwardson 1970).

However, in the majority of Nicotiana CMS varieties, the lack of firnctional pollen

67

production is accompanied by changes in corolla length, corolla adnation and stamen

morphology (Kofer et al. 1991 , Bonnet et al 1991).

The homeotic phenotypes of Nicotiana CMS are diverse and show that
cytoplasmic effects are exhibited throughout the developmental spectrum of andro genesis
and floral development (Rosenberg and Bonnet 1983). Normal tobacco floral
development culminates in a flower composed of five fused sepals, with five pink petals
flrsed into a corolla, five stamens and a bilobed stigma. In contrast to normal
development, alloplasmic cybrids exist in which no stamens are produced or in which
stamen development is arrested at the primordial stage, yet corolla development is normal
(cytoplasm fl'om N. sauvelons, Glimelius et a1 1981, Kofer et al. 1991). Cybrids with N.
undulata cytoplasm show the formation of normal stamen filaments but, instead of
anthers, pink petal-like structures with empty pollen sacs are formed. In this line a
shortening of the corollas is seen concomitantly. The distortion of anther structm'e is
most complete in cybrids of N. bigelovii, in which stamen filaments show no anther or
pollen sacs but are instead light pink in color and exhibit flinged petal-like tips (Kofer et
al. 1991). In addition to these changes, alloplasmic lines exist where unfirsed or split
corollas are seen and distorted anthers are produced which bear stignatoid tissues at their
tips (Bonnet et al. 1991). Lastly, in cybrids with Hyoscyamus, geen flowers with no
corolla or stamens are produced. In addition to this severe phenotype, other less affected
cybrids were found that contained stamens with a range of petaloid conversion (flom

pink filaments with normal anthers to complete conversion of anthers to petals),

68

conversion of sepals to petals and a modification of the segnentation of the stigma

(Zubko et al. 2001).

In each of the above studies, the homeotic floral traits were maternally inherited
and stably transmitted over the subsequent generations (Zubko et al. 1996, Kofer et al.
1991, Bonnet 1991). In some cases, it has been seen that backcross progeny fl'orn
petaloid plants exhibited a variety of different forms of petaloidy from the parent plant
(Kofer et al. 1991). The homeotic CMS lines contain a novel set of mtDNA flagments
when compared to either its cytoplasmic donor or nuclear donor. In general, the extent of
mtDNA alteration correlated with the severity of the male-sterile phenotype, as measured
by both the degee of splitting of the corolla and stamen morphology (Kofer et al. 1991,
Belliard et al. 1979). As opposed to other forms of CMS, the homeotic phenotypes have
not been associated with any specific DNA flagnents or rearrangements. In addition, in
contrast to carrot, no CMS specific protein production has been found to be due to novel
mtDNAs, although transcriptional activity of 0112 74 is increased in some lines (Bergman
et al. 2000). However, in cases where fertility has been restored through nuclear genes,
the appearance of normal stamens was accompanied by mtDNA changes that were both
qualitative and quantitative in nature. Therefore, it has been proposed that a genetic
mechanism behind the regulation of stamen development is mitochondrial recombination

followed by differential amplification (Kofer et al. 1991).

69

Summary

Since mitochondria are the main source of energy in the cell, it is no surprise that
cells with mitochondrial defects exhibit developmental abnormalities. In both plants and
animals, the tissues most affected by mitochondrial dysfimction tend to have the highest
oxidative energy requirements; in plants, this appears to be especially true of the tissues
involved in male gametogenesis. The importance of mitochondrial function is
demonstrated by the fact that improper firnctioning leads to the production of phenotypes
such as: leaf variegation (NCS, chm), improper leaf expansion and development (MDL),
overall poor gowth and yield (NCS, MDL), cytoplasmic male sterility and improper
floral development. At the heart of the majority of these abnormalities are mtDNA

rearrangements caused by recombination.

70

;-u_.___

 

GOAL AND HYPOTHESIS

The overall goal of this study is to investigate the impact of homologous
recombination on mitochondrial flmction in plant gowth and development. It is
hypothesized that a protein homologous to RecA plays a role in generating the
mitochondrial DNA rearrangements that are phenotypically manifested by the
mitochondrial disorders described previously. To test this hypothesis, transgenic plants
were generated (see Chapters 2 and 3) which over-expressed the E. coli RecA protein or
its dominant-negative derivative in their mitochondria. This chapter will highlight the
phenotypic screening of these transgenic plants for mitochondrial disease phenotypes, as
well as the requisite genetic studies to ensure that this phenotypic expression was

mitochondrial in origin.

EXPECTED PHENOTYPES:

Transgenic plants expressing wild type RecA

It is expected that when the E. coli RecA protein is over-expressed in either
Arabidopsis or Nicotiana tabacum mitochondria, there will be an increase in the number
of DNA rearrangements. Since mitochondrial DNA rearrangements have been the cause
of all known plant mitochondrial disorders, this increase should result in the appearance
of one or a combination of these phenotypes: variegation, male sterility, improper leaf

expansion and overall diminished gowth or yield. Although not previously seen in

71

Arabidopsis, it is possible that sterility could be manifested in flower development as

homeotic floral conversions as they are known to occur in Nicotiana.

In the chm line of Arabidopsis, which already shows marked phenotypic
expression of mitochondrial disorders, an increase in mitochondrial DNA recombination
is not expected to be measurable by an increase in phenotypic display. This is because
the penetrance of chm in homozygous lines is already at 100%. However, the increased
recombination and rearrangement activity of the over-expressed RecA, in concert with
previous chm mutations, could cause an increase in lethality among the plants. The

lethality could then be an indicator of increased recombination activity.

Transgenic plants expressing dominant negative RecA

It is expected that when the E. coli dominant negative RecA protein is over-
expressed in either Arabidopsis or Nicotiana tabacum mitochondria, there will be a
decrease in recombination and DNA rearrangements. Since there are no known
mitochondrial disorders that stem flom the disappearance of recombination, in these
plants no detrimental phenotypic differences are expected. However, recombination is
thought to play a significant role in maintaining subgenomic populations within
mitochondria and it is not known how important these population dynamics are to gowth
and development. Therefore, if decreasing recombination affects this balance,

mitochondrial disorders may arise.

72

Expression of the dominant negative RecA could be expected to alleviate the
mutant mitochondrial syndromes observed in the chm line. It has been shown that in
these lines the debilitating phenotypes are caused by an active system of recombination
and amplification of rearranged molecules. Since the dominant negative RecA should
decrease recombination, it should also decrease the expression of the phenotype.
Therefore chm plants with this construct should show less variegation, distorted leaves
and sterility, when compared to chm plants. For these studies mitochondria from
Columbia plants were introduced by crossing into chm/chm nuclear lines both with and
without the dominant negative transgene. This allows for a direct comparison of

phenotypic effects.

73

RESULTS

Phenotypic changes in Arabidopsis lines expressing various E. coli recA genes

The transformation and subsequent confirmation of transformation in Arabidopsis
were described in Chapter 3. The geatest number of transformants (16) was achieved
with the wild-type RecA in the Columbia variety of Arabidopsis, while lower numbers of
transformants were recovered from the three allelic chm lines (Table 4.1). In contrast,
only the Columbia line yielded transformants when the dominant-negative RecA was
used. In all of the transformants, when T1 progeny from these lines were examined, no
plants showed the expected phenotypes or any other abnormalities when compared to the

wild-type.

Columbia plants expressing E. coli RecA gew comparably to wild-type
Columbia plants. These plants did not show any of the changes in leaf expansion,
coloration, overall height or fertility that would indicate mitochondrial dysfunction.
Similarly, chm lines with E. coli RecA expression do not show any enhancement of the
chm phenotype or increased lethality. No expected phenotypic changes were predicted
for the Columbia plants carrying the dominant-negative RecA and none were found.
Although not exhaustively sought, no chm lines with this transgene were recovered. In
general, it is recogiized that for each of these lines only a small number of progeny were

analyzed and that perhaps a larger screening would have produced a notable phenotype.

74

Table 4.1: Summary of Arabidopsis thaliana transformation and screening data

 

 

 

 

 

 

 

 

 

 

 

 

Number of
# of plants
Plant Line E. coli transformants Number of plants showing
RecA recovered screened aberrant
phenotypes
Columbia Wild-type 16 150* 0
Dominant 5 50 0
negative
chm-1 Wild-type 3 50 0
Dominant 0 Not applicable
negative
chm-2 Wild—type 4 50 0
Dominant Not applicable
negative
chm—3 Wild-type 2 50 0
Dominant 0 Not applicable
negative

 

 

 

 

 

 

* 50 plants screened from each of three randomly chosen transformant lines. All other

numbers represent plants flom the same line

75

Phenotypic changes in Nicotiana tabacum lines expressing the E. coli recA genes

As described in Chapter 2, tobacco transformation was accomplished via the
Agrobacterium-mediated leaf disc method and subsequent plant regeneration. Therefore,
individual plants could be regenerated from the same transformed callus and all the plants
regenerated represented a To population of plants. This is in contrast to the Arabidopsis
transformants in which To transformants were gown from seed. From the Nicotiana
transformations, twenty—one plants (representing sixteen independent calli) were
recovered that carried the E. coli RecA gene; while seventeen plants (representing fifteen
independent calli) had integated the dominant negative gene (molecular confirmation

found in Figure 3.2).

In both the dominant negative RecA transformants and the wild-type RecA
transformants, To regenerated plants did not show abnormalities in terms of overall
growth, variegation or leaf development. However, upon flowering, the wild-type RecA
transformants produced flowers with homeotic abnormalities. Tables 4.2 and 4.3 outline
the general characteristics of all the thecA transformants. Although not included in the
tables, control plants (Samsun) that underwent transformation with the vector pB1121 (no
RecA insert) were taken through the regeneration procedure in parallel. These control
plants, hereafter referred to as Sarnsun, did not exhibit any floral or other phenotypic

abnormalities.

Of the twenty-one thecA To plants (which were derived from the

aforementioned 16 independent events), only one individual showed completely normal

76

W W a

TT. 7. _iilnill ; ,Lliiiw. .

 

Table 4.2: Characteristics of T0 Nicotiana tabacum transformed with the wild-type
RecA construct. Letters represent plants derived from the same callus.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Callus/Plant # flowers # abnormal % abnormal % fallow seed
Number total flowers flowers pods
1 61 12 20 10
2 A 45 5 l 1 13
B 59 14 24 39
3 A 34 34 100 100
B 44 4 9 l l
4 71 0 0 21
5 50 17 34 10
6 12 1 8 0
7 57 2 4 24
8 53 3 6 7
9 52 3 6 40
10 A 56 1 2 12
B 104 13 13 26
C 39 2 5 33
D 43 2 5 28
1 1 36 16 44 36
12 32 23 72 31
13 92 2 2 25
14 58 1 2 12
15 62 3 5 15
16 64 5 8 2

 

 

 

 

 

 

 

77

 

lal

C01

9.1 ,, I

h

Table 4.3: Description of abnormalities seen in To tobacco carrying the wild-type RecA
construct.

 

Callus/Plant Number

# Abnormalities

Types of Abnormalities

 

2

Petaloid stamens“
Fused flowers

 

2

Petaloid stamens
Fused flowers

 

Petaloid stamens
All petaloid stamens;
(huved corolla

 

All petaloid stamens

 

w>

 

Petaloid stamens
Curved corolla

 

 

01-h

MO N—t

Unfused Sepals
Petaloid stamens
All petaloid stamens
Curved corolla
Fused flowers

 

Curved corolla

 

Petaloid stamens

 

Petaloid stamens
Increased floral (flan number"

 

\O GONG

N N------

Petaloid stamens
Fused flowers

 

Petaloid stamens

 

Nr—r

Petaloid stamens
All petaloid stamens

 

Petaloid stamens

 

DO w>

 

Curved corolla

 

ll

Petaloid stamens

 

12

\o—pdr—ua

Petaloid stamens
All petaloid stamens
Fused stamens

Stignatoid stamens

Petaloid stamens & curved corolla
Stigmatoid & petaloid starrrens
Fused flowers and petaloid stamens
Short stigna
Stigmatoid stamens and crn'ved corolla

 

l3

Petaloid stamens

 

14

Curved corollas

 

15

Fused stigmas
Fused corollas

 

l6

 

 

or low—_-

 

Unfused Sepals
Petaloid stamens
Stigmatoid stamens

 

a:
it

one, two, three or four of five stamens can be affected
10 petals, 8 anthers and single conjoined stigna of 8 locules in single flower.

78

 

floral development (plant 4). All the other plants showed a mosaic of both normal and
abnormal flowers produced intermittently throughout their flowering branches. The
flaction of abnormal flowers on each plant varied, running the gamut from 2% (plants
10A, 13 and 14) to completely abnormal (100%, plant 3A). It should be noted that even
plants produced from the same callus did not necessarily show similar levels of

abnormality.

The types of abnormalities seen in the flowers differed in each line as well. A
typical tobacco flower consists of “five fused sepals, a five lobed gametopetalous corolla
whorl of five epipetalous stamens and a bicarpelate ovary” (Rosenberg and Bonnett,
1983, Figure 4.1A). The abnormalities which were noted in the wild-type RecA
transformants include: complete fusion of flowers at their base, both complete (all
stamens converted) and incomplete petaloidy, a marked curvature of the corolla,
increased floral organ number, lack of fusion of the sepals and changes in stigma height
and architecture as well as the presence of stignatoid tissue at the tips of anthers. (Table
4.3 and Figures 4.1, 4.2). In general, individual plants produced between 1-3 types of
abnormal flowers in concert with the normal flowers. The petaloid phenotype was the
most common (16/20 affected plants). Only one plant produced a uniform phenotype
(line 3A, 100% completely petaloid flowers), while plant 12 was notable in that it

produced nine distinct floral phenotypes.

Along with floral abnormality, a second characteristic that indicates mitochondrial

dysfimction would be that of cytoplasmic male sterility. To assess sterility, the flaction

79

 

Figure 4.]: Comparisons of wild-type and abnormal flowers in the TI progeny. Images in
this dissertation are presented in color.

(A) wild-type tobacco flower

(B) thecA transformant flower with no viable pollen produced

(C) thecA transformant flower with only 3 stamens and 4 petals

(D) thecA transforrnant flower with 4 stamens and l petaloid stamen

(E) thecA transforrnant flower with 3 petals and 3 petaloid anthers

80

 

 

Figure 4.2: Comparisons of wild-type and abnormal flowers in the T1 progeny. Images in
this dissertation are presented in color.

(A) wild-type tobacco flower

(B) thecA tmnsformant flower with 10 petals, 10 stamens and 2 stigmas fiJsed
(C) thecA transformant flower with 5 stamens, l petaloid and 2 stigmatoid
(D) thecA transformant flower with 4 stamens, l stigmatoid

(E) WtRecA transformant flower, fused by corolla

(F) lateral view of the flower in E.

81

 

of fallow seed pods produced after unassisted self-fertilization was quantified for each
plant (% fallow, Table 4.2). Despite the abnormality of the flowers, seed production was
not universally decreased in the To wild-type RecA transformants when compared to
Samsun plants. On average, Samsun plants were 29.0 i 16.0% fallow, with a range from
10-40% fallow seed-pods. Individual thecA transformed plants ranged from 0—100%
fallow, with the average infertility being 24 i 21%. Therefore, these plants were not
considered to be more infertile than Samsun. Comparisons were also made between the
amount of fallow seed-pods and the percentage of floral abnormality each plant exhibited
(Figure 4.3). These results showed a correlation (12=O.5187). As noted in Table 4.2,
while some plant lines were largely abnormal and also markedly infertile (lines 3A, 11
and 12), other lines existed in which abnormality was low and infertility was greater than
average (line 9). The reverse was true as well, as plants with lower than average
infertility were found that had high levels of abnormal flowers (line 5). Taken together,
the appearance of homeotic-like floral abnormalities and sterility in these plants met the

criteria of a mitochondrial disease phenotype in the genus Nicotiana.

Changes in floral development were not seen in the dominant negative To
transformants. Fractions of fallow flowers in the dominant negative lines were very
similar to those for control plants (data not shown). Overall this fits with the general
expectation set forward, but no fiirther studies were performed to address this issue. It
was expected that decreasing recombination would not have a detrimental effect on
overall growth and development and, therefore, no mitochondrial disease phenotypes

would be seen in these plants.

82

 

Correlation of the percentages of abnormal flowers and
fallow seed pods

—I

ossssés

% fallow pods

 

0 10 20 30 4O 50 60 70 80 90 100 110

% abnormal flowers
R2 = 0.5187

 

 

 

Figure 4.3: Correlation between the percentage of abnormal flowers on a plant and
the amount of infertility. Infertility is defined by the number of fallow seed pods
produced. Plants are fi'om the To lines and each diamond represents an individual
plant. R2 value calculated using the best fit line in Excel.

83

Phenotypic Analysis of T1 progeny front Line 12

Inherita_n_c; of floral abnormalities

To evaluate the inheritance and stability of the floral abnormalities in the
transgenic RecA plants, seeds from two different abnormal flowers (one petaloid and one
stigmatoid) from the To plant 12 were grown. These seeds were derived from self-
pollinations and T1 progeny in this analysis were separated by the presence or absence of
the thecA transgene. In this chapter, the lines will be designated as 12pet (fi'om the
petaloid flower) or 12stig (the stigmatoid flower) with or without RecA. The heritage of

these plants is depicted in Figure 4.4.

This study was performed with these distinct goals: 1) To determine whether
progeny from each type of flower would exhibit identical abnormalities as the parent
flower and/or the original To plant; 2) To determine if additional developmental
abnormalities would occur in the continued presence of the transgene; 3) To determine if,
once produced, the appearance of developmental abnormalities would be independent of
the presence of the transgene; 4) To determine whether the inheritance pattern of the

developmental abnormalities was maternal.

Figure 4.5 (A, B) illustrates the percentage of abnormal flowers in individual
plants with and without the presence of the wild-type RecA gene. If a mitochondrial
defect is responsible for the floral abnormalities, it was anticipated that individual

progeny plants would appear uniform for the abnormality that was seen in the parental

84

Self-pollinated (plant 12)

T1 tiiiiiit iiiiitiii

w'th RecA Without RecA

. ‘E
l

Back-crossed
to Samsun (plant 31)

 

ifiiiiiiii

Figure 4.4: Overview of the genetic analyses. Individual flowers from To plants were
allowed to self-pollinate and seeds from those flowers were grown and screened for the
presence or absence of the RecA transgene (T1 generation). Individual flowers on plants
that lacked the transgene were then backcrossed using Samsun pollen to produce a back—
cross generation that continued to lack the transgene (BC1 generation). Images in this
dissertation are presented in color.

85

(A) Data from the line 12stig

 

 

Line 123tig

#ot' plants per class
O -h NO! th 01 or N on

 

1to1011to 21to 3110 41lo 51in 61to 71to 81to 91to
20 30 4O 50 60 70 80 90 100

abundance of abnormal flowers (56)

 

 

(B) Data from the line 12pet

 

 

Line 12pet

# plants per class
0 A N w # U! 03 N

 

1t01011to 21to 31to 41to 51to 61m 71!!) 81b 91m
20 30 40 50 60 70 80 90 100

abundance of abnormal flowers (SS)

 

 

Figure 4.5: Number of plants in each T1 line grouped by the overall amount of floral
abnormality expressed. Plant classes were created in blocks of 10% to organize the data;

colors refer to the presence (black) or absence (white) of the RecA transgene

86

plant (100% of the flowers would have petaloid anthers). Although it would also be
reasonable for the T1 progeny to show an abnormality level that was the same as the
parental plant (the T 0 plant 12 had 72% abnormal flowers, hence each of the progeny
plants would also be 72% abnormal). As can be seen (Figure 4.5, summarized in Table
4.4), these expectations were not met. In fact, a tremendous variety in the percentage of
floral abnormality can be seen from plant to plant. For example, in 12pet plants,
abnormality levels range from 2-85% in the presence of the transgene and between 1-
72% when the gene is removed. Similarly, in the 12stig plant lines, abnormality levels
range fi'om 2-59% in the presence of the transgene and between 3-51% when the gene is
removed. No unaffected plants were found in any of the lines in regard to floral

development.

While the ranges for the lines are quite varied, when data for each line were
averaged and statistically analyzed, more distinct conclusions could be made (Table 4.4,
Figure 4.6). For the 12pet and 12stig lines, the average percent of floral abnormalities
was similar both in the presence and absence of the transgene (Table 4.4, Figure 4.6), yet
these values were much lower than that of the progenitor plant from which they came
(72%). When tested using Anova analysis, it was found that the average percentage of
floral abnormality in each of the lines was statistically different firm that of the Samsun
plants (p <0.0001). Statistical analyses also showed that plants with and without the
nuclear transgene present did not have different variances of floral abnormalities fi'om
each other (Tukey’s, Student-Newman-Keul’s [SNK], Fisher’s LSD [LSD] and Scheffe’s

tests; a = 0.05). In addition, the type of flower from which the seeds were derived did

87

Table 4.4: Comparison of averaged data from T. plants both with and without RecA

 

 

 

 

 

 

 

 

 

 

 

expression.
Feature Samsun T0 plant 12 Line 12stig 1 Line 12pet]
w/ RecA w/o RecA w/ RecA w/o RecA
(15)* (15) (18) (26) (19)
# flowers 62 i 18 32 72 :1: 39 76 :t 56 99 d: 39 93 d: 47
height 88 :1: 12 not 81 i 11 78 d: 12 79 d: 26 75 at 26
(cm) measured
%fallow 30¢ 8 31 99 21:1 99i0.5 96¢? 91 :l: 14
range 19-47% NA 96-100% 99-100% 74-100% 47-100%
% 0 72 21¢ 14 22i15 29:1:23 31:23
abnormal
range NA NA 2-59% 3-51% 1-72% 2-85%

 

 

 

 

 

 

 

 

* (N) = number of plants

88

 

 

Average Percentage of
Abnormal Flowers

88888

10

 

Samsun 12 stig
plant line

12 pet

 

 

 

Number of Flowers

 

Samsun 12 stig
plant line

 

 

 

Average Percentage of Fallow

 

 

 

Seed Pods
120
100
80
32 60
40
20
0 l
Samsun 12 stig 12 pet
plant line

Average Height of Plants
120
100

height (cm)
8

 

Samsun 12 stig
plant line

12pet

 

 

 

Figure 4.6: Average trends for the T. plants, in terms of height, number
of flowers, percentage of abnormal flowers and percentage of infertility.
Infertility was defined as the number of fallow seed pods out of the total
number of pods produced. Colors refer to the presence (black) or absence

(white) of the RecA transgene.

89

not affect inheritance of the abnormalities, as both the stigmatoid and petaloid flowers
showed statistically similar ranges of overall abnormality. While the percentage of
abnormal flowers was increased in all the T1 plants, the total number of flowers per plant

was not statistically different from Samsun (p = 0.0641, Figure 4.6).

A second prediction for the T, plants was that the type of floral abnormalities in
the progeny would mimic those of the parental flower from which they were derived.
Thus it was thought that progeny from the petaloid flower would display the petaloid
phenotype in all its flowers, while those of the stigmatoid flower would produce
stigmatoid progeny. Instead, as in the progenitor plant, mosaic flowering stalks were
produced. Normal flowers were interspersed on each plant along with a variety of floral
abnormalities. Progeny in either line, petaloid or stigmatoid, had the ability to produce
any of the phenotypes that had been seen in the progenitor plant 12, whether or not the
transgene was present. (refer to Table 4.3). In addition, novel phenotypes were seen
including: decreases in stamen and petal numbers (reduced to three, Figure 4.1E),
changes in corolla fusion (Figure 4.2E-F), conversion of sepals to petals and additional
petals produced on the exterior of the corolla. In lines with and without the transgene,
abnormal phenotypes became more complex as combinations of previously seen
phenotypes occurred: flowers that were both firsed together and petaloid, flowers in
which petals were unfused but stamens were petaloid or where petaloid and stigmatoid
anthers occurred together (Figure 4.2C). Flowers were also produced in which organs

were converted to floral buds. Figure 4.7 shows a flower in which conversion to buds has

90

 

Figure 4.7: A markedly abnormal flower from the T1 generation. In addition to having a
non-fused corrolla and the conversion of multiple sepals to petals, this flower has no
stigma and only 4 stamens two of which are converted to petals. In addition to these
problems with whorl definition, this flower has produced new floral buds in the same
whorl where the sepals would appear. Images in this dissertation are presented in color.

(A) normal tobacco flower

(B) view of abnormal flower from top

(C) view of abnormal flower fi'om lefi side
(D) view of abnormal flower from right side.

91

occurred in the whorls that contain petals, stamens and the ovary making it appear as a

fusion of four flowers rather than a single flower.

Aim/SE of Overall Growth

In addition to the appearance of novel floral abnormalities, new growth and leaf
phenotypes began to appear in the T1 progeny fiom either type of progenitor flower,
independent of the transgene’s presence. For example, plants began to produce
individual leaves that were not completely expanded (Figure 4.8B-E) or unusually shaped
(Figure 4.8F—J). The shape of the leaf was often distorted due to the presence of multiple
veins (4.86), as opposed to the single main vein typical of a normal tobacco leaf (4.8A).
Leaves in this generation also began to show small patches of variegation (Figure 4.8D).
Just as flowers showed fusion to one another in this generation (Figure 4.2E, F), so did

leaves. Figure 4.9 shows a fusion of two leaves at the midvein.

Identical to the way the floral abnormalities are expressed, malformed leaves
appear sporadically on the plant and are intermittent with normal leaves. The presence of
plants in which multiple stems were produced firm the base of the plant or even fiorn the
center of an existing stem was also noted. The last characteristic that was analyzed in
these lines was height, which was similar to floral abnormalities in the manner in which it
was expressed: the lines 12pet and lZsti g had varied ranges for overall height (data not
shown). However the average height in each of these lines was not significantly different

fiom that of the Samsun plants (anova p= 0.2596, table 4.4, Figure 4.6).

92

 

Figure 4.8: These are examples of abnormal leaf phenotypes seen in the T1 generation.
All leaves were taken from approximately the same position on the plant and size
differences are therefore not due to maturity of the leaves. Images in this dissertation are
presented in color.

(A) wild-type leaf

(B & C) drastically dwarfed and non-expanded leaves

(D) dwarfed leaf with variegation at the tip

(E) non-expanded leaf (equally non-expanded from the midvein)

(F) non-expanded leaf (only one-side effected)

(G) leaf with two mid-veins present

(H) leaf with overall distorted shape

(1) leaf with crescent shaped appearances, found normally below floral meristem
(J) leaf altered to have cup—shaped morphology

93

 

Figure 4.9: A markedly abnormal leaf from the T1 generation (with RecA transgene), in
which fusion of two leaves has occurred at the midvein. Images in this dissertation are
presented in color.

(A) Top view of leaf

(B) Bottom view of the leaf
(C) Side view of leaf

94

Since the multiple phenotypes are likely to be due to mitochondrial dysfunction, I
expected that there would be a correlation between changes in height, the percentage of
abnormal leaves, the number of stems and the amount of floral abnormality per plant.
Floral abnormality was used as the base for comparison because it was the prominent
phenotype seen in the To and Tr plants. Pair-wise comparisons of the extent and severity
of developmental abnormalities are presented graphically in Figures 4.10 and 4.11. Both
of these figures provide comparisons of data from progeny in which the RecA transgene
was present or absent. In the plant line 12pet, no correlation was apparent between the
percentage of abnormal flowers and the prevalence of any other developmental
abnormality, when the RecA transgene was not present (Figure 4.10, B, D, F). In distinct
contrast, plants containing the transgene show a marked increase in the abundance of

abnormal leaves as the amount of floral abnormalities increase (Figure 4.10A).

The correlation noted for line 12pet with the transgene is identical for the 12stig
plants. In this line, plants with the transgene also showed an increase of the percentage of
abnormal leaves as floral abnormality increased (Figure 4.11A). For plants without
RecA, no correlation existed between the percentage of floral abnormality with height or
the number of stems produced. Weak correlations were found for plants with RecA

between the extent of floral abnormality and reductions in height (Figure 4.11C).

Additionally a negative correlation was seen to exist between the extent of floral

abnormality and leaf deformation without the transgene (Figure 4.11E)

95

 

(A) Abnormal flowers vs abnormal leaves

 

% abnormal leaves

0 20 40 60 80 1 00
% abnormal flowers R2 = 0,1733

 

 

 

(B) Abnormal flowers vs abnormal leaves

03‘!)
00

N
O

 

% leaf abnormality
A
O

O

0 20 40 60 80
2 _
% abnormal flowers R ' 0'00

 

 

 

Figure 4.10: Comparison of the percentage of floral abnormality with the percentage of
abnormal leaves per plant in the line 12pet both with (A) and without (B) expression of
the E. coli RecA transgene.

96

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(C) Abnormal flowers vs height
1 20 .
E 100 9H1; ‘
8, 80 ’s a , o .
is" 60 s
3 40
-= 20
0 l l j I
0 20 40 60 80 1 00
% abnormal flowers R2 = 0.0639
(D) Abnormal flowers vs height
150 -
s l
3 100 I; g c. . . H
E " ’ ' o "—1
or
.5 50
.C
0 l I l
0 20 40 60 80
% abnormal flowers R2 = 0.0267

 

 

 

Figure 4.10: Comparison of the percentage of floral abnormality with the height of the
plant in the line 12pet both with (C) and without (D) expression of the E. coli RecA
transgene.

97

 

(E) Abnormal flowers vs number of stems

 

 

 

 

In
E
B
In
“6
=11:
0 50 1 00
% abnormal flowers R2 = 0.0804
(F) Abnormal flowers vs number of stems
5
g 4
2 3
(0
Io- 2
.2 1
0
0 20 40 60 80

% abnormal flowers R2 = 0.0011

 

 

 

Figure 4.10: Comparison of the percentage of floral abnomality with the percentage of
number of stems per plant in the line 12pet both with (E) and without (F) expression of
the E. coli RecA transgene.

98

 

(A) Abnormal flowers vs abnormal leaves

% abnormal leaves

 

20 40 60 80
96 abnormal flowers R2 _ 0.2226

0

 

 

 

(B) Abnormal flowers vs abnormal leaves

 

% abnormal leaves

20 4o 60
R2 = 0.0662

0

% abnormal flowers

 

 

 

Figure 4.11: Comparison of the percentage of floral abnormality with the percentage of
abnormal leaves per plant in the line 12stig both with (A) and without (B) expression of
the E.coli RecA transgene.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(C) Abnormal flowers vs height
120 - ~—---—-~~-—W~-Wm _WWCHH,
100 w , . ‘
E 80 o. «0'. .- t“;
o ..
5') 60
3 4o
20
0 I l l
0 20 40 60 80
% abnormal flowers R2 -.- Q1139
(D) Abnormal flowers vs height
100
E 80 ‘ ‘ . ‘ , . .
.._. 60
g, 0
E 40
20
0 I l I l l
0 10 20 30 40 50 60
% abnormal flowers R2 = 0 0017

 

 

 

Figure 4.11: Comparison of the percentage of floral abnormality with the height of each
plant in the line 12stig both with (C) and without (D) expression of the E.coIi RecA
transgene.

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(E) Abnormal flowers vs number of stems
5
2“ °
0 3
TI; 2 e: e e g
:u: 1 Teen: e s cc: e e
0 I I I
0 20 40 60 80
% abnormal flowers R2 = 0.0379
(F) Abnormal flowers vs number of stems
5 Wu ___.__. _ __ _ ---_ __ ________.
4 e
E
d) 3 :
"13' 2 ,
=11:
1 e e e see c e e e e e
0 fl l I I I
0 10 20 30 40 50 60
% abnormal flowers R2 = 0.0062

 

 

 

Figure 4.11: Comparison of the percentage of floral abnormality with the number of
stems per plant in the line 12stig both with and without expression of the E.coli RecA
transgene.

101

An_alys_is of reproductive ability

As with the original transformants, the T. plants were allowed to set seed
naturally through unassisted self-pollination. However, unlike the To plants, the T.
progeny plants both with and without the nuclear transgene showed high levels of
infertility as measured by the inability to set seed through self-pollination (Table 4.5,
Figure 4.12). While infertility levels in the progenitor plant (31%) and the Samsun
controls (30 i 8%) were similar, the average infertility among lines both with (12pet: 91¢
14% and 125tig: 99 i 1.2%) and without RecA (12pet: 96 i 7.4% and 12stig: 99 i
0.51%) was quite high (Table 4.4, Figure 4.6). These averages differ significantly from

infertility levels in Samsun plants grown at the same time (anova p < 0.0001).

Infertility levels were particularly high in the individual plants derived from the
stigmatoid flower. In plants that did not have the transgene, only one of fifteen plants
showed any seed set at all (2% fertile, plant 12stig # 8). In lines with the transgene, 14
of the 18 plants were 100% self-infertile and the other 4 plants were > 96% self-infertile.
In the progeny from the petaloid flower, a more varied range of infertility was seen both
with and without the presence of the nuclear transgene (Table 4.5, Figure 4.128). Despite
the range of infertility, the majority of T. plants were greater than 90% infertile. Neither
the effect of the transgene nor the type of flower fi'om which the line was derived had any

statistically significant effect on infertility (Scheffe’s test 0:005).

102

Table 4.5: Data fiorn individual T. progeny in the line 12pet, ordered by % abnormal
flowers and categorized by the presence or absence of the transgene. Bold numbers
indicate plants used for generating the back-cross progeny.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Plants Without the recA Transgene Plants with the recA Transgene
Plant # % Abnormal Self-Sterility Plant # % Abnornnl Self-Sterility
flowers flowers
28 1 93 7 2 93
19 5 100 9 4 100
31 6 100 41 4 100
40 6 100 14 6 97
43 7 100 46 7 100
18 10 100 37 13 99
39 11 100 49 16 100
16 13 99 5 25 86
42 21 100 20 26 85
6 25 92 12 26 89
32 29 74 47 29 97
8 40 86 48 31 100
23 41 100 27 33 97
17 46 82 38 35 100
44 46 100 22 40 70
45 51 100 10 38 100
3 63 100 2 39 84
25 65 97 13 43 47
l 73 97 15 48 58
4 58 92
30 59 68
26 59 93
29 60 99
1 1 64 99
24 68 100
21 86 100
Avg. 29 :h 23% 96 :t 7% Avg. 31 :1: 23% 91 :1: 14%

 

 

 

 

 

 

 

 

 

103

(A)Data from the line lZstig

 

Line 1Zstig

L‘l NoRecA
I RecA

    

# plants per class
8

110101110 2110 3110 4110 5110 6110 7110 8110 9110
20 30 40 50 60 70 80 90 100

so Infertility

 

 

 

(B) Data fi'om the line 12pet

 

Line 12pet

N
O

_s
01

[Jno RecA
IRecA

# plants per class
or 8

 

O

110 1110 2110 31104110 5110 6110 711081109110
102030405060708090100

°/. Infertility

 

 

 

Figure 4.12: Number of plants in each T. line grouped by the overall amount
of infertility they expressed. Infertility was defined as the number of fallow
seed-pods out of the total number of pods produced. Plant classes were
created in blocks of 10% to organize the data; colors refer to the presence
(black) or absence (white) of the RecA transgene.

104

In the individual T. 12pet plants, where the range of infertility for individual plants was
more variable, the amount of infertility was not correlated with the amount of floral
abnormality.

As with the To transformants, no correlation existed between the abundance of
abnormal flowers per plant and the extent of infertility (Figure 4.13, Table 4.5). Some
plants had relatively high levels of infertility yet low percentages of abnormal flowers
(e.g. plants 19, 28, 31 fiom table 4.5). Other plants had high levels of floral abnormality
and infertility (e.g. plants 1, 21, 24, 25, 44, 45), as well as plants with low levels of
infertility and high levels of floral abnormality (e.g. plants 13, 22, 15, 30). The presence

or absence of the transgenic RecA did not influence these results (Table 4.5, Figure 4.13).

Smanmf T. progeny analyses

T. progeny both with and without the RecA transgene were grown fiom flowers
that were either petaloid or stigmatoid in phenotype. All of the progeny plants showed
some level of floral abnormality and the appearance of the floral abnormalities occurred
independently of the presence of the thecA transgene. The presence of the transgene
did not lead to a statistically significant difference in the abundance of developmental
abnormalities. Taken together, these observations allow several conclusions to be made:
1) After their initial appearance, the developmental abnormalities are transmitted to the
progeny independently of the transgene 2) The mosaicism of the floral traits in the

progeny fits the expectations for vegetative segregation of a non-Mendelian trait.

105

(A) Plants without the nuclear RecA transgene

 

Correlation of Infertility and Floral Abnormality in T,
plants without RecA

 

'/s fallow seed pods

0 10 20 30 40 50 60 70 80 90 100
96 abnormal flowers R2 = 0.0187

 

 

 

(B) Plants with the nuclear RecA transgene

 

Correlation of Infertility and Floral Abnormality in T.
plants with RecA

1 20
1 00
80
60
40
20

 

% fallow seed pods

0 20 40 60 80 100
% abnormal flowers

 

R2 = 0.0335

 

 

Figure 4.13: Correlation between the percentage of abnormal flowers on a plant and the
level of infertility in plants without (A) or with (B) the RecA transgene. Infertility is
defined by the number fallow seed pods produced. Plants are the T. generation (line 12
pet) and each diamond represents an individual plant. R2 value calculated using the best-

fit line in Excel.

106

l“

In terms of the phenotypes seen, individual T. plants were similar to the To individuals
that produced mosaic flowering branches with multiple types of abnormalities. This was
contrary to what was expected to occur, which was that the progeny plants would
transmit only the floral abnormality of its parent flower (i.e. be 100% petaloid or
stigmatoid in phenotype). This mosaic phenotype included the appearance of additional
novel developmental abnormalities in the T. generation, independent of the transgene’s
presence. Secondary phenotypic changes were found in these plants such as appearance
of abnormalities in leaf shape, size and overall plant height. Lastly, it was found that the
T. progeny plants had a significantly higher level of infertility than the Samsun lines,
regardless of the type of progenitor flower fi'om which they were derived and presence or

absence of the nuclear transgene.

Production of the Back-cross (BCI) Population

Crosses with wild-type Samsun pollen were made with plants in T. lines for two
reasons: to ensure that there would be progeny from the self-sterile T. plant lines and to
ascertain the inheritance pattern of the floral abnormalities. Table 4.6 shows the number
of plants in each of the four lines that were crossed, as well as the total numbers of
successful crosses. Almost all attempted crosses were successful in seed production.
Hence, male fertility was affected in these plants, but female fertility was not. Crosses
were equally successfirl independent of the morphology of the flowers, as the overall
phenotype of the flower did not affect its ability to be pollinated. A few unsuccessful

reciprocal crosses (5) were performed concurrently.

107

Table 4.6: Seed set fi'om pollination of T. plants with wild-type pollen. Numbers in

parenthesis are percentages out of the total number of plants.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Plant Transgene # # of # total Number of crosses producing seed
Line plants plants flowers Total Normal Abnormal
total crossed crossed number flowers flowers
Itotal /total
12stig RecA 18 14 41 38 23/25 12/13
(78%) (93%) (92%) (92%)
none 15 11 20 18 10/10 8/10
(74%) (90%) (100%) (80%)
12pet RecA 26 11 35 34 21/22 13/13
(42%) (97%) (95%) (100%)
none 19 9 29 29 21/21 8/8
(47%) (100%) (100%) (100%)

 

108

 

Analysis of the BC. population

To determine whether the floral abnormalities and infertility were maternally inherited,
four plant lines were chosen fi'om the 12pet T. progeny that did not have the nuclear
transgene. These lines (31, 42, 44 and 45) were selected because each line displayed
100% self-sterility and levels of floral abnormalities varying fi'om low to relatively high
(see Table 4.5). In each of these lines, flowers that were normal in phenotype were
crossed with wild-type pollen fiom Samsun plants. It is these seeds that gave rise to the
BC. progeny. These seeds, which are part of the BC. generation, were used to analyze
maternal inheritance since the parental plant did not contain an active transgene,
eliminating the chance of nuclear gene effects. Since the pollen donors never contained
the transgene and had only normal flowers, any floral abnormalities seen in progeny

should be due to maternal inheritance.

It was expected that if floral abnormalities are maternally inherited, then all the
progeny from a given line would show some level of abnomality. The same expectation
is true for infertility; if this trait is inherited maternally then all the progeny should show
an increased production of fallow seed-pods when compared to the Samsun plants. In
contrast, if the traits are nuclear, the progeny should show segregation by Mendelian

ratios.

The level of abnormality in each line followed the pattern seen previously in the

T. progeny. Individual plants produced a variety of different floral abnormalities

interspersed with flowers that were normal in appearance. In each line the percentage of

109

Table 4.7: Developmental abnormalities in BC. plants lacking the RecA transgene. All
lines were derived from crossing 12pet T. plants with wild-type Samsun plants as the

pollen donor. The tabulation of fallow seed-pods excluded flowers used for crosses.

 

Samsun
N=10

Line 31
N=10

Line 42
N=10

Line 44
N=12

Line 45
N=10

 

Average %
fallow seed

pods

22 ¢ 9.0

84 :1: 28.0

98 ¢ 4.9

96 :l: 8.3

92 ¢19

 

Range of
fallow seed
pods

9-31 %

10-100%

84-1 00%

71-100%

46-] 00%

 

% fallow seed
pods on the

T. parent

plant

NA

1 00%

1 00%

1 00%

1 00%

 

 

Average % of
floral

abnormalities

15¢ 15.9

47 i 34.1

35 :1: 30.3

38 :1: 28.4

 

Range of %
of floral

abnomalities

NA

1- 47%

9-95%

2-94%

1 0-99%

 

% floral
abnormalities
on the T.

 

 

parent plant

NA

6%

 

 

21%

 

6%

51%

 

 

 

110

abnormal flowers varied greatly from plant to plant (Table 4.7); to illustrate this in terms
of the population of plants, they were grouped into classes based on the percentage of
abnormality per plant. Using this artificial grouping it can be seen that no pattern exists
between or within the four lines (Figure 4.14). The fact that every plant of the BC.
generation exhibits some level of abnormality, indicates that the floral abnormalities have
been maternally inherited. Furthermore the mosaic appearance of the trait is indicative of

vegetative segregation, another hallmark of non-Mendelian inheritance.

The trait of infertility also appeared in a mosaic pattern among BC. plants, with
most plants having some fertile flowers and mainly infertile flowers. Although derived
from T. lines that were 100% self-sterile, the progeny plants displayed a range of
infertility (summarized in Table 4.7). As with floral abnormality, plants were grouped in
classes to enable a comparison of the distribution of infertility between lines derived fiom
the different T. progenitor plants. Despite the presence of a range of infertility, in each
line the majority of plants were completely self-sterile (Figure 4.15, averaged data shown
in Figure 4.16). An Anova analysis coupled with tests for similarity showed that the
infertility in each of the BC. lines was statistically different fiorn the Samsun lines (p <
0.0001). Unlike the To and T. analysis, in the BC. plants, it appears that abnormality and
infertility are correlated to some degree: as the percentage of floral abnormality
increased, so did the level of infertility (Figure 4.17 11:0.0942). As with floral
abnormality, the fact that every plant produced some level of infertility points to a

conclusion that this trait was likely to be a maternally inherited extra-nuclear one.

111

names—0 sch. snaffle—Q 50 8‘ x N 3 rl m kale inn
. flaw-.020 5.1.21 antafln ..

 

Line 31

 

 

 

# of plants per class

IOU-Lalo

a.-

Line 44

 

110 11 21 31 41 51 61 71 81 91
W b b b b b b b b b
20 so 40 50 80 70 N I) 100

percentagebyclass

 

 

 

7
36
fl
55
84
*2
£3
‘62
at1
0
110112131415181718191
10101010101olotototo
2030405060708090100
percentagebyclass
Line42
83.5
g 3
32.5
32
0
E1!)
a
E1
'605
a

O

11011 21 31 41 51 61 71 81 91
10101010101010101010
2030405000703090100

peroentagebyclass

 

 

 

 

plants per class

Line45

 

110 11b21103110411051m81lo711081ln 9110
10 20 30 40 50 50

90100

percentagebyclass

 

 

Figure 4.14: Number of plants in each BC. line grouped by the overall amount of
floral abnormality. Plant classes were created in blocks of 10% to organize the data.

 

Li)». {his a.

CF av.
.. In 2
..xarvrasr. _ . r
a. . ; l
a
a . . \ “on
..5 .4. .l 7

 

 

Line 31 Line 44

   

 

 

 

 

 

 

a
7 in
0
i. 2
0 0
h 5 I—
8 8
3 4 in
C
a 3 E?
‘1 '0.
'6 2 '5
it 1 11:
o o
110 1110 2110 3110 4110 5110 6110 71m 8110 91m 110 11b2110 31104110510) B110 71b81b 9110
40 50 60 7 1 10 20 30 40 50 60 70 1
Class class
Line42 .
Llne45
8
10 3 7
9 g 6
%8 s. 5
a7 8 4
8° ‘3 3
£5 -“-’ 2
24 a
n3 3‘ 1
32 °
1
0

110111021033110411051108110711081109103
102030405080700000100

 

 

 

 

 

 

Figure 4.15: Number of plants in each BC. line grouped by the overall amount of
infertility. Infertith defined as the number of fallow seed-pods out of the total number
of pods produced. Plant classes were created in blocks of 10% to organize the data.

113

Table 4.8 (a-e): Complete Data for Pollen Germination and Viability

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a)
Samsun % % pollen Average number # anthers
Fallow germination of pollen grains sampled
(plant number
(25)
1 31.4 51¢2 2140¢605 5
2 29.7 47 :t 22 1796 ¢ 896 5
6 10.4 46¢ 13 1705¢355 5
8 29.6 44 ¢ 16 1378 ¢ 684 5
9 9.6 42 ¢ 13 1569 ¢ 784 5
0))
Line 31 % % pollen Average number of # %
Fallow germination pollen grains anthers abnormal
(plant number) sampled flowers
(65)
1 100 0.2 ¢ 0.4 1022 ¢ 735 6 10
2 100 2.4 :t 5.9 5.6 i 2.9 6 6
3 100 1.6 :t 3.0 253 ¢ 268 6 10
4 100 0.0 i 0.0 3.4 i: 3.7 6 42
5 84 1.1:t1.7 1125¢523 6 11
6 10.4 9.5 :t 9.3 1106 ¢ 541 6 5
7 100 0.7 ¢ 2.1 6.5 ¢ 5.8 9 47
8 80.8 7.0 i 6.0 1521 ¢ 802 6 4
9 99 0.0 i 0.0 8.0 i 3.6 6 16
10 100 0.0 :l: 0.0 3.8 :1: 3.8 6 2

 

 

 

 

 

 

 

 

114

 

 

(C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Line 42 % % pollen Average number # %
Fallow germination of pollen grains anthers abnormal
(plant number) sampled flowers
(62)
1 100 0.0 d: 0.0 41 ¢ 82 6 12
2 99 0.4 ¢ 1.1 166 ¢ 238 6 28
3 95 0.0 i 0.0 304 ¢ 169 6 21
4 84 1.0 3: 1.8 355 ¢ 429 7 37
5 97 7.7 i 16.2 305 ¢ 246 6 85
6 100 0.0¢0.0 7.8¢ 13 6 61
7 100 1.8 ¢ 4.4 664 ¢ 468 6 25
8 100 0.0 :l: 0.0 113 :t 258 6 95
9 99 1.7 ¢ 3.8 167 i 274 6 94
10 100 0.8 ¢ 1.4 771 :1: 370 6 9
(d)
Line 44 % % pollen Average number # %
Fallow germination of pollen grains anthers abnormal
(plant number) sampled flowers
(72)
1 100 0.0 ¢ 0.0 8 ¢ 7 6 46
2 90 0.33 :t 0.69 556 ¢ 514 6 3
3 100 0.0¢0.0 56¢ 71 6 21
5 100 1.9 i 0.0 26 ¢ 35 6 53
6 71 1.1¢1.3 819¢566 6 11
7 94 5.6 ¢ 8.8 587 ¢ 413 6 14
9 100 0.0 ¢ 0.0 2.7 ¢ 3.8 7 25
10 100 0.0 ¢ 0.0 445 ¢ 326 6 94
11 100 0.0 :l: 0.0 74 i 99 6 20
12 100 0.0 ¢ 0.0 30 ¢ 56 6 91
13 99 5.9 i 11.4 149 ¢ 324 6 20
14 94 0.0 ¢ 0.0 40 ¢ 54 6 17

 

 

 

 

 

 

 

 

115

 

 

(e)

 

 

 

 

 

 

 

 

 

 

 

 

Line 45 % Fallow % pollen Average number # %
germination of pollen grains anthers abnormal
(plant number) sampled flowers
(63)
1 100 1.2 ¢ 2.8 447 as 382 6 26
2 100 0.1¢0.22 183¢ 128 6 44
3 99 0.6 :1: 1.5 207 i 363 6 40
4 100 0.0¢0.0 18¢17 6 99
5 100 0.0 ¢ 0.0 3 at 2.0 6 16
6 70 4.4 i 8.1 979 ¢ 530 6 10
7 91 4.4 ¢ 5.6 1279 ¢ 870 8 72
8 98 0.6 d: 1.5 75 i 167 6 37
9 47 9.3 :1: 12.4 1170 i 594 6 10
10 90 0.5 :1: 1.2 426 :t 440 8 25

 

 

 

 

 

 

 

 

ll6

 

 

Average Percentage of Abnormal Flowers

 

 

Samsun 31 42 44 45
Plant Llne

 

 

 

Average Percentage of Fallow Seed Pods

 

Samsun 31 42 44 45
Plant Line

 

 

Figure 4.16: Averages for control Samsun and the BC. plants, in terms of the
Del"montage of abnormal flowers and percentage of infertility. Infertility was
defined as the number of fallow seed-pods out of the total number of pods
pro(luced.

117

 

[ff Corrslaflonofthepsrcontageabnonnalflowsrsandthepsrcsntagsot
fallowssodpods

120

 

% Fallow

40

20

 

0

0 20 40 60 80 100 120
96 Abnormal R2 = 0.0942]

ngn 4.17: Correlation of the percentage of abnormal flowers found per

plant and the level of infertility per plant. Infertility was defined as the
number of fallow seed pods out of the total number of pods produced.
Diamonds indicate data for an individual plant in the four BC. lines (31, 42,
44, 45).

 

118

In addition to CMS and floral abnormalities, similar leaf changes to those seen in
the T. plants were found in these BC. progeny. As in the T. progeny, height and flower
number of the BC. plants varied, but when the data are analyzed with anova the lines (10
not show a statistically significant difference fi‘om the Samsun plants (p=0.1405 and
p=0.3744 respectively). The BC. population was similar to that of the T. in that a variety
of abnormal flowers were generated fi'om flowers that were normal in appearance. The
diversity of floral phenotypes included: the conversion of one to five of the stamens to
petals or stigmas, decreases in petal number, decreases in stamen number, splitting of the
corolla, fusion of flowers and the conversion of sepals to petals. In addition
combinations of any or all of these individuals abnormalities were found. When
quantified, all but six of the 42 plants had greater than 50% of the flowers which showed
conversion of anthers to petals as their primary type of abnormality. This phenotype was
scored whether a single stamen was converted or the conversion occurred in tandem with
other abnormalities (Figure 4.18A). For plants in the lines 42, 44, and 45, as the fraction

of abnormal flowers increased so did the occurrence of petaloidy among those flowers

(Figure 4.18B)

It should be mentioned that the possibility of paternal inheritance could not be
completely ruled out by these studies: only five reciprocal crosses were attempted with
the T. progeny; but 30 crosses with pollen fiom many different BC. progeny were
attempted and none led to seed set. This result will be addressed in depth in the section

on pollen viability. Despite the failure of reciprocal crosses due to male-infertility, a few

119

(A)

 

 

 

 

 

 

Line31
150
100
Be
50
0
P 10 8 6 2 1 3 5 9 4 7
plantnumber
Line 42
150
100
38
50
0
P 10 1 3 7 2 4 6 5 9 8
plantnumber

 

 

 

Figure 4.18: Comparison of the number of petaloid converted flowers out of the total
number of abnormal flowers. (A) Total percentage of abnormal flowers is shown in white
and the number of petaloid flowers out of the total percentage of abnormal flowers is
shown in black. (B) The correlation of these data for each individual line is shown.
Numbers refer to the plant number and the P indicates the parent plant fi'om the T.
generation.

120

re,q

(A)

 

 

 

 

 

 

Line 44
120
100
80
a? 60
40
20
0
P 26714111339151210
plantnumber
Line45
150
100
50
0
P 6 9 5 1 1 8 3 2 7 4
plant number

 

 

 

Figure 4.18: Comparison of the number of petaloid converted flowers out of the total
number of abnormal flowers. (A) Total percentage of abnormal flowers is shown in white
and the number of petaloid flowers out of the total percentage of abnormal flowers is
shown in black. (B) The correlation of these data for each individual line is shown.
Numbers refer to the plant number and the P indicates the parent plant from the T.
generation.

121

 

Line 31

 

 

 

 

 

 

E 150

E .

.o 100 3 ° ° I I
111 M
33 50 ‘ .

% O

2.2 0 * . ' ' 'A

o 10 20 3° 4° 50
% abnormal total R2 = 0,0465

 

 

Line 42

150
100

 

%API%abnormal
or
o

% abnormal total R2 = o 2013

 

 

 

Figure 4.18: Comparison of the number of petaloid converted flowers out of the total
number of abnormal flowers. (A) Total percentage of abnormal flowers is shown in white
and the number of petaloid flowers out of the total percentage of abnormal flowers is
shown in black. (B) The correlation of these data for each individual line is shown.
Numbers refer to the plant number and the P indicates the parent plant fiom the T.
generation.

122

(B)

 

Line 44

1 50
100
50

 

% APl%abnormal

% abnormal total R2 = o 3955

 

 

Line 45

150
100

 

%APPrbabnonnal
or
o

0 20 40 60 80 100120
% abnormal total R2 = 0 5

 

 

 

Figure 4.18: Comparison of the number of petaloid converted flowers out of the total
number of abnormal flowers. (A) Total percentage of abnormal flowers is shown in white
and the number of petaloid flowers out of the total percentage of abnormal flowers is
shown in black. (B) The correlation of these data for each individual line is shown.
Numbers refer to the plant number and the P indicates the parent plant fi'om the T.
generation.

123

factors led me to conclude that the developmental abnormalities are due to defective

mitochondria inherited from the maternal parent.

Assessment of Male Sterility through Pollen Germination Assays

Although the percentage of fallow seed-pods allows one assessment of infertility
in the BC. plants, to conclude that this infertility is due to pollen dysfunction requires that
pollen viability in each of these plants be analyzed. One way in which pollen viability
can be assessed is by the ability of a pollen grain to germinate and develop a pollen tube.
Hence, to further characterize the nature of the infertility, the BC. plants were tested for
pollen germination in vitro (with five anthers of normal appearance sampled from each
plant for these assays). For each plant, the average percentage of pollen that germinated
was determined and used as an indicator of male sterility. In addition, these assays were

used to quantify the amount of pollen produced by the BC. plants.

All 42 BC. progeny plants were analyzed through these assays for pollen
germination rate and pollen amount. As controls, five randomly chosen Samsun plants
were analyzed. Each of the five Samsun plants showed greater than 40% pollen
germination (Table 4.8, Figure 4.19). In contrast, none of the 42 BC. plants showed
greater than 10% germination and the majority of the plants (39/44) produced pollen with
<5% germination. When tested for statistical significance by anova, the pollen
germination of the BC. progeny were found to differ significantly from that of the

Samsun plants (p< 0.0001). Tests for similarity such as Tukeys, LSD, SNK and

124

 

 

 

 

 

Line 31
80.00
8
1;; 60.00
s
g 40.00
0:
O) 20.00
a?
0.00
mammmeszas7149m
plant number
Line 42
80.00
.0 70.00
B 60.00
g 50.00
E 40.00
g 30.00
20.00
39 10.00
0.00
81828688895 7 9 410213 6 8
plant number

 

 

 

Figure 4.19: Average amount of pollen germination for individual plants in the BC.
lines. Germination was assayed using five anthers per individual plant. S= Samsun
control plants (black). Nurnbers=plant number in that individual line (white).

125

 

Line 44

80.00
70.00
60.00
50.00
40.00
30.00
20.00
1 0.00

0.00

% germlnated

 

818286888913 7 5 6 213 8 9101114
plant number

 

 

 

Line 45

80.00
70.00
60 .00
50.00
40 .00
30.00
20 .00
1 0 .00

0.00

% germinated

 

81828688899 6 713 8102 4 5
plantnumber

 

 

 

Figure 4.19: Average amount of pollen germination for individual plants in the BC.
lines. Germination was assayed using five anthers per individual plant. S= Samsun
control plants (black). Nurnbers=plant number in that individual line (white).

126

 

Line 31

3000.00
2500.00
2000.00
1 500.00
1 000.00
500 .00
0.00
-500.00

 

average # of pollen grains

 

plant number

 

 

Line 42

5 3000.00
.=, 2500.00
0.

'5 w 2000.00

C ..

* E 1500.00
3 on

8 1000.00
0

>

N

“2:: l l " u . -- -- --

-500.00

 

plant number

 

 

 

Figure 4.20: Average amount of pollen produced by individual plants in the BC 1 lines.
Pollen numbers were assayed using a minimum of five anthers per individual plant. S=
Samsun control plants (black). Numbers=plant number in that individual line (white).

127

 

Line 44

3000.00
2500.00
2000.00
1 500.00
1 000.00
500.00
0.00
-500.00

average # of pollen grains

 

plant number

 

 

Line 45

3000.00
2500.00
2000.00
1 500 .00
1 000.00
500.00
0.00
-500.00

 

average # of pollen grains

 

plant number

 

 

Figure 4.20: Average amount of pollen produced by individual plants in the BC. lines.
Pollen numbers were assayed using a minimum of five anthers per individual plant. 8:
Samsun control plants (black). Nurnbers=plant number in that individual line (white).

128

Scheffe’s place the BC. plants in multiple separate categories from Samsun in terms of

germination ability (a=0.05).

Unlike the germination data, the quantity of pollen grains produced by each of the
individual plants was quite varied (Table 4.8, Figure 4.20). For the Samsun plants, the
average amount of pollen ranged fi'om 1378 d: 684 through 2140 :t 605 grains per anther.
In the BC. plants the range was larger; for example, in the plants derived from the T.
plant 31, the average amount of pollen production ranged from 0.5 at 2.9 to 1521 i 802
grains/anther. An anova analysis was performed comparing each BC. plant to the others
and the Samsun controls. This analysis reveals a significant difference in number of
pollen grains per anther between individual plants (p < 0.0001), however tests for
similarity such as Tukeys, LSD, SNK’s and Scheffe’s all created complex and different
groupings for the plants. While no overall pattern emerged in regard to pollen amount,
when the most conservative analysis was used (Scheffe’s test (1:005) 18 of the 42 BC.
plants produced comparable amount of pollen to the Samsun plants. These data suggest
that, as with the floral phenotypes, a mosaic pattern exists at the organ level as well, in
which different anthers have different abilities for producing pollen and producing

functional pollen.

Since there was a large variability in the amount of pollen produced per anther
fiom the individual plants, the possibility that the amount of pollen produced could be
related to the pollen viability (as indicated by the pollen germination assays) was

assessed. To analyze whether there was a relationship between these two variables, the

129

values were compared graphically. (Figure 4.21A). In general, as pollen numbers
increased, so did the ability of that plant’s pollen to germinate (12:0.3615). Nonetheless,
these germination abilities are still significantly lower than those for Samsun plants with
similar quantities of pollen grains. The quantity of pollen was also compared to the
ability of each plant to set seed (Figure 4.218). This correlation showed that as the
number of pollen grains increased there was an increase in the number of pods that
contained seed (r2==0.377l). In addition, pollen germination ability was compared to the
amount of seed set (Figure 4.21C). These data are consistent with the other comparisons:
as the germination ability of pollen increases, the percentage of fallow seed pods

decreases (#0475).

These analyses were then extended to test for a correlation between the
prevalence of floral abnormalities and amount of pollen produced or the ability of the
pollen to germinate (Figure 4.22). When the percentage of floral abnormality per plant
was compared with the ability of that plant’s pollen to germinate, only a slight negative
correlation was seen as shown in panel A (r2=0.0232). Yet, in general, when plants had
more abnormal flowers, less pollen germination was observed. Similarly, panel B
(Figure 4.22) shows that a slight negative correlation exists between pollen quantity and
floral abnormality (r2=0.0905); as the percentage of floral abnormality per plant

increased, fewer pollen grains were produced by normal appearing anthers.

130

(A)

 

Comparislon of pollen quanltity and

 

germination ability
10
c
.2 8
2 o
E 4
8 2
a!
0
0 500 1000 1500 2000
"“"b" R2=0.3615

 

 

 

(B)

 

Comparislon of pollen quantity and the
percentage of fallow seed pods

% fallow

 

0 500 1000 1500 2000
#of Ion
"a R2 = 0.3771

 

 

 

(C)

 

Comparision of pollen germination ability and
the percentage of fallow seed pods

96 fallow
seséé

 

20
0 e
0 2 4 6 8 10
’5 W'“‘”" R2 = 0.4715

 

 

 

Figure 4.21: Correlations of pollen quantity, pollen germination, and level of infertility
per plant. Infertility was defined as the number of fallow seedpods out of the total number
of pods produced. Data is for individual plants in one of the four BC. lines (31, 42, 44,

45).

131

(A)

 

Comparision of the percentage of floral
abnomality and germination ability

‘lo germlnatlon

 

0 20 40 60 80 100 120
96 abnormal

 

 

 

(B)

 

Comparision of the percentage of floral
abnormality and quantity of pollen

 

pollen number

0 20 40 60 80 100 120

% abnormality R2 _ 0 0905

 

 

 

Figure 4.22: Correlation test of the percentage of floral abnormality, pollen quantity and
pollen germination ability. Diamonds indicate data for individual plants in one of the
four BC. lines (31,42,44,45).

132

Summary of the BC. analysis:

All the individual BC. plants had pollen with statistically lower germination
ability than did the Samsun plants. In addition over half the plants produced less pollen.
The amount of pollen produced by a plant correlated directly with the pollen’s ability to
germinate and with the ability of the flower to set seed. In addition a slight correlation
existed between the percentage of abnormal flowers per plant and the amount of pollen
produced and the ability of that pollen to germinate. Taken in concordance with the data
that each of the BC. plants exhibits some level of infertility and abnormality, these results

suggest that these plants may be considered cytoplasmically male sterile.

DISCUSSION

The genetic analyses were performed to determine the effect of expressing the E.
coli RecA protein in plant mitochondria. The RecA protein is a key component in the
process of homologous recombination in E. coli, and it is assumed that a naturally
occurring RecA-like protein exists in plant mitochondria, as homologues have been found
in animal and fimgal mitochondria Knowing this, the expectation was that over-
expression of this protein would increase mtDNA recombination. Increased
mitochondrial DNA recombination could lead to an increase in aberrant recombination
and subsequent formation of DNA rearrangements. In this study novel transgenic plants
that over-expressed the E. coli RecA protein in their mitochondria were made in an
attempt to induce the phenotypic appearance of mitochondrial disease-like syndromes.

Since the mitochondrial syndromes are most frequentally phenotypically displayed as

133

 

cytoplasmic male sterility, variegation, and homeotic-like floral abnormalities, it is
predicted that over-expression of E. coli RecA protein in wild-type Arabidopsis and

Nicotiana would induce the appearance of those mitochondrial disease-like phenotypes.

These studies found that the expression of E. coli RecA in Arabidopsis did not
result in phenotypic manifestation of plant mitochondrial syndromes in either the T0 or T.
populations that were screened. There are a few potential reasons for this. First, perhaps
expressing RecA in these plants is lethal, creating mitochondrial DNA alterations that can
not be compensated for. Second, perhaps the small repeats present in the Arabidopsis are
not amenable to recombination by RecA. It is conceivable that too small a number of
progeny were examined to truly see an efl‘ect. In addition no changes were seen with the
transgenic dominant-negative RecA lines in tobacco. This result is in concordance with
what was predicted, as it is unknown the effect decreasing naturally occurring

recombination would have on mitochondrial fimction.

In contrast in Nicotiana, 20 of 21 initial transformants with the wild-type E. coli
RecA gene were found to have aberrant floral development. Individual flowers displayed
homeotic-like conversions of anthers to petals, anthers to stigmas and other abnormalities
indicative of improper floral whorl definition. The production of abnormal flowers
occurred as a mosaic on each plant, with both affected and normal flowers appearing
sporadically over the flowering stalks. The mosaicism was mimicked on the population

level as well: individual To plants had variable levels of abnormal floral production when

134

compared to one another, even when they were derived from the same transformed

callus.

While floral development can be affected by changes in mitochondrial gene
expression, floral development is best understood to be regulated by a set of nuclear
homeotic MADS box transcription factors. Given this information and the fact that the
presence of RecA in Arabidopsis did not induce a phenotypic response, the possibility
needed to be considered that one of these nuclear homeotic genes might have been
disrupted in the transformation process. Two factors argue against this being likely.
First, the floral mutations were seen in all but one of the independently transformed lines.
Since the integration of the Agrobacterium T-DNA region in the plant nucleus is known
to occur through random insertion, it is statistically unlikely that the same set of nuclear
floral development genes would be disrupted in each of these independent integrations.
Secondly, it is known that in terms of phenotypic response, nuclear gene inactivations are
normally “all or nothing” responses. Hence a nuclear gene inactivation would have
created plants in which all flowers had the same phenotype, rather than creating a diverse

mixture and range of floral abnormalities on each plant.

When progeny fiom one of these initial transformants was analyzed, statistically
similar levels of floral abnormalities were seen in the T. plants both with and without the
RecA transgene present. These plants continued to show a range of phenotypic responses
typical for a mitochondrial disorder. The fact that removal of the transgene does not

eliminate the production of floral developmental abnormalities indicates that once the

135

mitochondrial defects are generated, they are able to persist in the absence of the
transgene. Analysis of the BC. population (generated by backcrossing with wild-type
Samsun pollen) confirmed this result. All the BC. progeny continued to show a mosaic

pattern of floral developmental abnormalities.

In both the T. and BC. populations a large range of floral abnormalities were seen
concomitantly with the appearance of novel leaf developmental abnormalities and
infertility. This diverse assemblage of phenotypic response is typical of mitochondrial
syndromes as their behavior is understood in mammalian systems. In humans,
mitochondrial diseases encompass a large array of clinical problems that are molecularly
and genetically complex and usually display a unique pattern of inheritance (Wallace
1999). To understand both the etiology and the pathogenesis of mitochondrial disease
requires an understanding of the ways in which mitochondria behave genetically as

members of a population.

As has already been established, the first way in which mitochondrial genetics
differs from that of the nucleus is that mitochondria generally are inherited uniparentally,
most often maternally. In conjunction with this each cell contains a multiplicity of
mitochondria and each of these mitochondria contain multiple copies of their genome.
This creates a nested genetic population of mitochondria and mitochondrial DNA within
each cell. Mutations can arise in either a population of mitochondria or in population of
mtDNAs that result in the presence of two or more mtDNA genotypes within a cell,

organ or individual; a situation defined as heteroplasmy. If the mutations are pathogenic,

136

the proportion of mutated molecules in each heteroplasmic population affects the severity
of the biochemical defect and its subsequent phenotypic appearance. The effect of
mutational load on phenotypic response is not necessarily linear in nature. During the
life of an organism the mutational load of aberrant mtDNA in a cell is affected by
vegetative segregation; this is due to differences in both mtDNA replication and
segregation between cell and tissue types. In the same vein, different cells have different
minimal energy requirements (thresholds), and the levels of heteroplasmy and the
dynamics of mitotic segregation play a critical role in determining a disease’s phenotypic

presentation (Schon, 2000, Wallace 1999).

In human diseases, it has been repeatedly demonstrated that the same mtDNA
mutation can produce markedly different symptoms among the members of the same
family due to variation in the percentage of mutant mtDNAs each individual inherits. For
example a point mutation in the atp6 gene of humans is known to cause both Leigh’s
syndrome and the clinical phenotype NARP (Neurogenic muscle weakness, Ataxia, and
Betinus Eigmentosa). When the mutation is lower in abundance (< 75%) among the
mtDNAs, the phenotype that is manifested is NARP; whereas when this mutation is
present in > 95% of the mtDNAs, it causes early onset of Leigh’s disease. Leigh’s
disease produces more debilitating clinical symptoms including: ataxia, hypotonia,
spasticity, developmental delay optic atrophy and opthaloplegia, due to the higher
amount of mutated DNA present. In the same manner, the disease’s LHON (Leber’s
flereditary _thic Neuropathy) and dystonia are due to an identical point mutation in the

nad6 gene, but mutational load drives the production of two diverse phenotypes. LHON

137

is marked by onset in midlife of sudden blindness (due to death of the optic nerve), while
dystonia is a generalized movement disorder characterized by impaired speech, mental
retardation and short stature. Similar to the relationship of NARP and Leigh’s disease,
LHON patients have a lower percentage of mutant mtDNAs than patients with dystonia.
In classifying human mtDNA diseases, the actual DNA mutation is always used as a
marker for the disease, since the same phenotype may be produced by different mutations
and the same mutations can cause markedly different phenotypes (reviewed by Wallace

1999)

Another factor in disease presentation, besides the overall percentage of mutant
mtDNA is the local concentration of these molecules. Patients who have Kearns-Sayer
syndrome (KSS), Pearson’s syndrome (PS) and Progressive External Opthalrnoplegia
(PEO) all harbor the identical large-scale deletion of mtDNA. KSS is manifested as a
multi-systemic disorder, PEO only affects skeletal muscle and PS is predominantly a
hematopoetic disease. This is explained by the fact that in mammalian embryos there is
no mtDNA replication until the onset of germ layer differentiation, so the segregation
pattern of the mutant mtDNAs among germ layers results in varying mutational loads in
different tissues. If the deleted mtDNAs segregate into all the germ layers the phenotypic
response is KSS; if the segregation is non-uniform, the result is PS, and if segregation is
specifically to the muscle, the resulting phenotype is PEO. In addition to the type of
tissue in which the mutant mitochondria are localized, the age of the tissue can also be

significant, as mitochondrial mutations tend to increase with age.

138

In all of these syndromes, rapid shifts in the levels of heteroplasmy within a
pedigree result in a range of phenotypes (from asymptomatic to fatal). The offspring of
a single heteroplasmic mouse can have different levels of heteroplasmy individually, but
the mean heteroplasmy will be approximately equal to the proportion of mutated mtDNA
in the mother (Lightowers et al. 1997, Chinnery et al. 2000). What is the cause of this
imperfect transmission? There is currently no way to tell to what extent this vegetative
segregation of mutant mtDNAs is due to stochastic partitioning as opposed to preferential
replication (Birky 2001). However, studies in both yeast and human cell cultures have
shown that neither size of the mtDNA nor the number of replication origins are factors
for preferred mtDNA replication. Human cells in culture will replicate mtDNA
molecules until a specific mass of DNA exists in the cell (Diaz et al. 2000). This
phenomenon creates the opportunity for a greater number of deleted molecules to exist in
a cell, relative to wild type (full sized) mtDNAs (Tang et al. 2000; Diaz et a1. 2000).
Another theory proposes that increased multiplication of mitochondria with mutant
genomes can be explained if it is based on amount of respiration occurring (Tang et al.
2000). In this theory, it would take more mutant mitochondria to supply a sufficient level
of ATP for the cell and therefore mitochondria that contain mutant mtDNA are multiplied

to higher numbers to meet this need (Bertrand et al. 1986, Diaz et a1 2000).

Animal mitochondrial diseases can be used as a model for understanding the
developmental abnormalities that result from the expression of the E. coli RecA in
mitochondria of transgenic Nicotiana plants. RecA is known to be a crucial protein in the

process of homologous recombination. One of the distinguishing features of plant

139

mitochondrial biology is that active recombination is used to subdivide the genome into a
complex population of substoichiometric molecules (Arrieta-Montiel et al. 2001). This
creation of rearranged mtDNAs and their subsequent maintenance at low levels in the
population allows them to be available for selective amplification at different stages of
development to affect gene expression. In fact, in bean the presence of the CMS
inducing pus-071239 rearrangement has been characterized in 141 different wild and
cultivated species. It was detected in 100% of the species examined, the majority of
which maintained the rearrangement at a substoichiometric level (less than 1 copy in
every 100 cells of young seedlings or leaves). However the very presence of the
sequence suggests that it is efficiently transmitted and cannot be irreversibly lost from
these lines or that it keeps being created de novo at low frequency. It also provides
firrther evidence that a sequence can become fixed in a genome by a rearrangement that

locates it to a non-dispensable molecule (Arrieta—Montiel et al. 2001).

Since recombination is the first step for this mechanism, over-expressing the
RecA protein could cause an increase in naturally occurring recombination and in
aberrant recombination that would generate mtDNA rearrangements. Interaction of
recombination and amplification has the potential to generate a population of mtDNAs
unlike that which would normally exist, resulting in an impairment of mitochondrial
ftmction. How might the presence of these amplified rearrangements affect plant
development? Since floral development and especially male gametogenesis are stages of
plant development which place exceptional demands on mitochondrial activity (Bergman

et al. 2000), it would be expected that changes in the mtDNA population in these tissues

140

would correlate with alterations in fertility and floral morphology. Therefore in plants
with greater recombinogenic potential, like those with transgenic RecA, there should be
greater amounts of pollen inviability and abberant flower development. While this efl'ect
was found, its expression was not uniform, most likely due to the effects of sorting out

and threshold levels of functional mitochondria required by specific tissues and organs.

The presence of active recombination in floral meristematic tissues followed by
rapid sorting out of the mtDNAs into different tissues provides a means to rationalize the
presence of normal flowers and homeotic-like flowers on the same plant. Normal flowers
would have received a lower percentage of abnormal mtDNAs, while flowers with
homeotic conversions would be likely to have higher quantities of aberrant mtDNA
molecules. This segregation may occur in the specific organs as well, with the threshold
needs of the organ and the nature of the mitochondrial mutation both potentially linked to
the phenotypic manifestation. If the early developing floral meristem is deficient in
functional mitochondria, an impact would be observed on sepal or petal deve10pment,
with fusion of flowers and corollas; however if the minimum threshold is breeched at a

later stage, after petal formation, perhaps only anther formation would be affected.

In these studies, it appears that the stamen is the most responsive flower organ to
the threshold effect caused by mitochondrial dysfimction. For example, in both the T.
and BC. generations, many flowers contained a mixture of anther phenotypes within the
same flower including: normal appearing fertile anthers, normal appearing infertile

anthers, petaloid anthers or stigmatoid anthers. The most common alteration seen was

141

that of infertility, followed by the homeotic-like conversion of the anthers to petals. It is
hypothesize that each physical manifestation is reflective of the amount of aberrant
mtDNA in the anther: anthers that produce inviable pollen would appear in response to
lower quantities of aberrant mtDNA than would the anthers that undergo organ
conversion. This would be due to the fact that production of pollen is so energy intensive
that even at low levels of mitochondrial impairment, it cannot proceed properly.
However, it is also tempting to hypothesize that the presence of different aberrant

mtDNA rearrangements could cause the alternative phenotypes.

Just as anthers appear to be more responsive than other floral organs, it appears
that flowers are more likely to be affected than other plant organs. However in both the
T. and BC. generations, leaves and overall plant growth was seen to be impacted as well.
These phenotypes may have appeared later for two reasons. First, since these organs also
contain chloroplasts and hence are able to share the burden of energy production, a higher
level of mutant mtDNA may be required to change the phenotype. Secondly, if all the
mutant mtDNA is localized to a specific tissue, how sorting out occurs developmentally
will definitely shape how the syndromes are presented. As with the mammalian diseases
Kearns-Sayer syndrome, Pearson’s syndrome and Progressive External Opthalmoplegia it
is differential segregation of mtDNA at the level of tissue layers that cause the variations
in phenotypic expression. In this scenario, overall growth would be affected if the
aberrant mtDNA is in all plant tissues at a high concentration, while leaf development

would be affected if a high concentration of abberrant mtDNA is segregated from cell

142

lines produced by the apical meristem, while flowers would be altered when primary

segregation affects the threshold levels of mutant DNA fi'om the floral meristem.

While in general, the use of animal mitochondrial disease pathology is an
amenable model for the results of this study, one aspect of the plant mitochondrial
syndrome pathology is not reconcilable with this analogy. In animal systems, if the
mother carries mutant mtDNAs her progeny will each display a particular level of mutant
mtDNA, distinct from each other and her. However, when the levels of mutant mtDNA
are averaged for all the progeny, the percentage of mutation reflects the overall
composition of the original parent (Lightowers et al. 1997, Chinnery et al. 2000).
Although not examined at the mtDNA level, this is not true of the phenotypic responses
for the E.coli RecA transgenic plants in either the T. or BC. populations. While these
plants show various levels of phenotypic response comparable to each other and to the
parent, when averaged they do not necessarily show a similar level to the parent plant
(nor do they show a consistent increase or decrease). These results have also been found
by others in alloplasmic tobacco lines, in which back cross progeny exhibit a wider
variety of petaloid forms than does the parent plant (Kofer et al. 1991). I believe these
results reflect the fact that genome rearrangements and amplification are involved in the
production of debilitating mitochondrial syndromes as well as in normal growth and

development.

143

CONCLUSION

This chapter describes the phenotypic analyses of transgenic plants carrying an E.
coli RecA protein targeted to plant mitochondria. It was expected that if recombination
plays a role in generating aberrant mitochondrial syndromes, over-expression of the
RecA protein would cause phenotypes typical to plant mitochondrial syndromes such as
cytoplasmic male sterility. A variety of maternally inherited phenotypic responses were
found in the transgenic lines such as CMS, variegation, aberrant leaf development and
homeotic floral conversions. The inheritance of these phenotypes, in addition to being
maternal and showing vegetative segregation, can be understood in the context of the

inheritance of mtDNA syndromes in animal models.

144

CHAPTER 5

Molecular Basis for Abnormal Floral Development and Male Sterility

INTRODUCTION

Although the majority of mitochondrial proteins are encoded by nuclear DNA and
translated on cytoplasmic ribosomes, a number of proteins are encoded by the
mitochondrial genome. Typically the angiosperm mitochondrial DNA contains a subset
of the genes that encode proteins for electron transport and oxidative phosphorylation
(complexes 1, III, IV and V), as well as those encoding the mitochondrial translational
apparatus (rRNAs, tRNAs and ribosomal proteins; Brown 1999, Bonen and Brown 1993,
see Table 5.1). Regulatory mechanisms must exist to ensure the co-operative expression
of the genes enoded by the mitochondrial and nuclear genomes, however interaction of
the two genetic systems is not well defined. Some information on the role of
mitochondrial gene expression in proper mitochondrial function has been obtained

through the characterization of mutants.

As noted in Chapter 1, there are only two well-researched classes of plant
mitochondrial mutants: cytoplasmic male sterile and non-chromosomal stripe mutants.
However the wealth of individual mutants within each category provides an excellent
starting point for the analysis of the effects of mitochondrial gene alteration.
Characteristically, only genes coding for proteins involved in the electron transport
pathway, ATP formation or the mitochondrial ribosome have been altered in these

syndromes (Table 5.1). Molecularly, mitochondrial dysfirnction is associated with

145

Table 5.] Mitochondrial genes identified in flowering plants. Gene names shown in
bold have been determined to be associated with the mitochondrial syndromes CMS,
NCS or chm.

 

 

 

 

 

 

 

 

 

 

 

 

 

Function [ Gene names
Complex I nadl, nad2, nad3, nad4, nad4L, nadS, nad6, nad'l
Complex III cob
Complex IV coxI, coxII, coxIH
Complex V atpA, atp6, atp9
Transfer RNAs C, D, E, F, G, H, I, K, M, M, N, P, Q, S, V, W, Y
Ribosomal RNAs rrn26, rm18, rm5
Ribosomal proteins rps3, rps7, rple, rpslB, rpsl4, rpsl9, rpll6
RNA maturase -related mat-r
ORF

 

 

(Reviewed by Schnable and Wise 1998, Budar et al. 2003, Conley and Hanson 1995,
Breiman and Galun 1990, Elkonin and Tymov 2000)

146

specific gene alterations created by active recombination that causes deletions or the
formation of novel chimeric open-reading frames. It is the goal of this chapter to
highlight commonly seen mitochondrial gene alterations and discuss how these mutations

are thought to disrupt mitochondrial firnction.

The location of gene deletions and their roles in mitochondrial dysfunction:

Deletion mutants are rare in plant mitochondria. In total eight deletion mutants
have been described in the literature and will be further described below. At the
molecular level, each of these deletions is due to rearrangements at short (6-102 bp)
repeats. Specifically, these deletions affect genes of complex 1, complex IV or the

mitochondrial translational apparatus (see Table 5.2).

Five distinct non-chromosomal stripe mutants are known to exist and for each one
the precise gene alteration that causes the phenotype has been completely characterized.
NCSZ plants lack a fully assembled respiratory complex I due to a truncation of the nad4
gene (Marienfeld and Newton 1994, Karpova and Newton 1999). The NCS3 and NCS4
mutants are due to different deletions at the 5’ end of the mitochondrial rps3 gene, which
is commonly co-transcribed with the 7p]! 6 gene. Due to these mutations, both NCS3 and
NCS4 plants show reduced levels of protein synthesis (Hunt and Newton 1991, Newton
et al. 1996). Deletions in these two genes are also known to cause the maternal distorted
leaf phenotype that has arisen in the chm lines of Arabidopsis (Sakamoto et al. 1996, see
Chapter 4). Lastly, the NCSS and NCS6 plants carry different 5’ deletions of the cox2
gene, which encodes a subunit of complex IV (Lauer et al. 1990, Newton et a]. 1990). In

each of these lines, if the deletion is homoplasmic, it is lethal to kernel development

147

Table 5.2: Summary of known gene deletions and their recombination sites. Information

for the repeat involved in NCS6 was unavailable (NA).

 

 

 

 

 

 

 

 

 

 

 

Plant Disorder Genes Region of Sequence
involved homology
(bp)
Maize NCSZ nod4 16 CAGAACAAAAGGGAAG
NCS3 77733 12 CTGGGGTGGGGC
77111 6
NCS4 77933 16 ACCCCAACGTTAGACT
rpl I 6
NCSS coxII 6 TCCTGC
NCS6 coxII NA NA
Arabidopsis MDL 77233 1 l GGAACAAAATC
ml] 6
Tobacco CMSI nod 7 Rec1=102 Not shown due to length
CMSI] nod 7 R602: 65

 

 

 

 

 

148

 

(Yamato and Newton 1999) and, therefore, NCS plants can only survive and be
propagated when heteroplasmic populations of mitochondria are present in a tissue
(Newton and Coe 1986, Gu et al. 1993, Marienfeld and Newton, 1994).

The CMSI and CMSI] mutants of Nicotiana sylvestris are near homoplasmic
mutants that have large deletions in their mtDNA, due to two recombination events
involving repeats of 102 bp and 65 bp respectively (Chetrit et al. 1992). These deletions
remove the nad7 gene (Pla et al. 1995) and the upstream region of the nod] gene
(Lelandais et al. 1998, Gutierres et al. 1999). In addition to the lack of NAD7 and NADl
proteins, complex I formation in these plants is defective for NAD9 (Gutierres et al.
1997). The effect of these gene deletions in leaf tissue is that plants show reduced
glycine oxidation and a lack of rotenone-sensitive NAD(P)H dehydrogenase activity
(Sabar et al. 2000). These are the only known CMS causing mutations that do not involve

chimeric gene formation.

The fact that deletion mutants are able to survive despite severe alterations to
complex I and complex IV of the electron transport chain is credited in large part to their
heteroplasmic nature (Newton and Coe 1986, Gu et al. 1993, Mareinfeld and Newton
1994). Karpova et al. (2002) have pointed out that the combined action of multiple
NAD(P)H dehydrogenases and the alternative oxidase pathway (AOX) should allow
plant mitochondria to be more tolerant of respiratory defects than their animal
counterparts. However, these pathways bypass ATP production, making them inefficient
in terms of energy generation. Although that would be disadvantageous, the presence of

these enzymes could serve to maintain normal levels of metabolites, reduce reactive

149

oxygen species (AOX) and introduce electrons to the ubiquinone pool (NADPH
dehydrogenases). In fact, an investigation by Karpova et al. (2002), on the activity of
various AOX genes in the different NCS plants showed not only increased AOX activity
but also differential activity of the multiple AOX isoforms. For example, the activity of
AOX is increased in tassels of all NCS mutants; while the NCS2 mutants, which have
deletions in complex I, show an increase in the putative redox-regulated AOX2 activity.
Complex IV NCS mutants (NCS6) show increased AOX3 activity only, while the
ribosomal mutants NSC3/4 have increases in both AOX2 and AOX3. These results
suggest that not only are the alternative pathways involved in NCS mutant survival, but
that when mitochondrial dysflmction is sensed, the signal is able to convey information
about the location of the lesion and elicit a specific physiological response (Karpova et al.

2002).

As with the NCS mutants, survival of CMSI and CMSII plants depends on the
activation of alternative pathways: the activation of internal and external NAD(P)H
dehydrogenases. These complex I defective mutants have been shown to have increased
alternative respiration and activity of the rotenone—insensitive NAD(P)H dehydrogenases
(Gutierres et al. 1997, Sabar et al. 2000). The role of these pathways was examined in
leaf tissue; however, the level of this effect in the tissue in which it is phenotypically

manifested (male gametophytes) has yet to be studied.

The organization of chimeric 0'73; and their role in mitochondrial dysfitnction:

Cytoplasmic male sterility occurs in many plant species and it is predominantly

associated with the formation of chimeric opening reading flames (ORFs; Schnable and

150

Wise 1998) rather than gene deletions. Although each ORF is unique in nature, they all
have notable common characteristics. First these ORFs have been derived through the
combination of a known gene or gene region with a previously unknown/uncharacterized
sequence. The creation of each of these chimeric open reading flames has been shown to
rely on multiple recombination events, as suggested by the model of Small et al. (1989).
Below, a few of the well-researched and distinctive CMS-associated ORFs will be

discussed. A summary of known chimeric genes can be found in Table 5.3.

Prior to the epidemic of southern corn leaf blight in 1970, the male sterile T
(Texas)-cytoplasm maize system was used to produce 85% of hybrid seed in the USA
(Schnable and Wise 1998). The use of CMS-T as an agricultural tool justified the focus
of investigations aimed at studying the genetic and molecular mechanisms underlying
both male sterility and fertility restoration. Sterility in CMS-T is caused by the presence
of the 13 kDa URF13 protein (Dewey et al. 1987). This protein assembles at the inner
mitochondrial membrane and is thought to be a mitochondrial pore forming protein. The
URF13 protein is generated by transcription of a chimeric gene encoding the 5’ portion of
the atp6 gene firsed to 777126 gene fragments and then the 14er 3 sequence (Dewey et a1
1986). The URF13 protein accumulates in many tissues of CMS-T maize plants,
however it only produces severe effects in the tapeta] cell layers of anthers, which are
seen to undergo premature degeneration causing pollen abortion (Warmke and Lee 1977).
In addition to its role in sterility, the susceptibility to fungal pathogens is absolutely

linked to the male sterility in CMS-T. Interestingly, when transgenic tobacco plants are

151

Table 5.3: Summary of known CMS causing chimeric genes and their restorers of

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fertility.
Plant ORF Native genes Protein size Restorer genes and
name involved function

Maize urfl3 atp6, rrn26 13 kDa Rfl= decreases

(CMS-T) transcription of 147173
Rf2= aldehyde
dehydrogenase

Petunia pcf atp9-1 , 00x11 25 kDa Rf= decreases
transcription
of pcf

Phaseolus pvs-orf239 atpA, cob 27 kDa Fr= copy number

vulgaris suppression of
mitochondrial
subgenome
carrying orf

Sunflower orfl-1522 atpA 16 kDa Rf=decreases

(PETl) transcription

Radish orf125 17 kDa Rf=decreases

@osenil) transcription

Rice (Bo) orf79 atp6 NA Rfl= transcript
processing
lediting

Sorghum orfl 07 atp9 NA RB: transcript

(A3) processing

Wheat orfl56 coxII 7 kDa Rf= affects translation
of protein

Brassica ortZ24 atp6 26 kDa prl=decreases

napus transcription

(polima)

Brassica orfl38 19 kDa

napus

(ogura)

Brassica or1222 nad5c NA

napus

(napus)

Carrot orfB-CMS orfl?(atp§), urf 25 kDa none

Carrot Petaloid atp9-I , urf NA none

Tobacco orf274 ath None found Rf=reduces
transcription

 

152

 

made to express URF13 mitochondrially the resulting plants are toxin sensitive but not

male sterile (Von Alhnan et al. 1991, Chaumont et al. 1995, Schnable and Wise 1998.)

In petunia, the presence of a fused gene named pcf (for petunia CMS-associated
fused gene) is responsible for infertility. The pcf gene is composed of the 5’ upstream
sequences of the first membrane spanning domain of atp9, sequences coding for the first
2 exons of coxII and an unidentified reading flame, 1015', which has no homology to any
known sequences (Young and Hanson 1987, Pruitt and Hanson 1991, Nivison et al.
1994). Transcription of the pcf locus is tissue specific, with pcf transcripts 4-5 times
more abundant in anthers relative to leaves (Young and Hanson 1987). Although the
three genes are co-transcribed, RNA transcripts are processed such that a 25 kDa sterility
related protein results from the translation of only the 14er region. This protein (URF S)
has characteristics of a peripheral membrane protein and is produced in sterile plants
(Nivison and Hanson 1989), however it is present in mitochondria of both vegetative and
reproductive tissues (Nivison and Hanson 1989, Conley et al. 1991, Conley and Hanson
1994, Nivison et al. 1994). To explore the role of PCF in sterility, transgenic tobacco
plants were made that expressed the 25 kDa protein encoded by 1015' under the control of
both the 35S CaMV promoter and the TA29 tapetal-specific promoter (Koltunow et al.
1990). When transgenic tobacco plants that expressed URFS were made, they were male
fertile mimicking results when transgenic plants were made with CMS-T. This led Wintz
et al. (1995) to hypothesize that expression of the pcf gene in early meiosis might be

critical for the sterility to occur.

153

Cytoplasmic male sterility in the common bean (Phaselous vulgaris) is associated
with the mitochondrial sequence pvs—orfl39 (Abad et al. 1995, J anska et al. 1998). This
gene sequence encodes the 5’ portion of the atpA gene (atpl in other species) fused to the
open reading frame 239, and is co-transcribed with the cob gene. Translation of pvs-
073‘239 produces a 27 kDa protein that only accumulates in reproductive tissue (Abad et
al. 1995) and is rapidly degraded in vegetative tissues (Sarria et al. 1998). This protein is
specifically associated with the microspore cell wall. Unlike CMS-T and pcf, when
transgenic tobacco were made to express nuclear copies of mitochondrially targeted

orj239, the resulting plants were male sterile (He et al. 1996).

While 28 different CMS cytoplasms are known to exist in sunflower (Horn et al.
1996, Horn 2002), the best characterized (PETl) was generated from an inter-specific
cross between Helianthus petiolaris and H. annus (Leclerq 1969). PETl CMS is
associated with the expression of a novel open reading flame, 011322, created by a
recombination event involving inversion and insertion of the orfl 22 sequence 3’ to the
atpl gene (Kohler et al. 1991, Laver et al. 1991). The first 18 amino acids of ORF522
are identical to those of 0713, which may be the plant homologue of yeast and
mammalian ATP8, a subunit of the Fo-F. ATPase (Gray et al. 1998). Translation of
orfl-I522 produces a hydrophobic 15 kDa protein (Horn et al. 1991, Laver et al. 1991,
Moneger et al. 1994) that is bound to mitochondrial membranes (Horn of al. 1996). The
method of action for CMS in PETl plants has been deduced to resemble the natural

action of programmed cell death in the tapetum of microspores (Balk and Leaver 2001).

154

 

When compared to one another, the unidentified DNA sequences of the CMS-
associated ORFs do not have any major similarities. The only common feature appears
to be that they often encode large hydrophobic domains (Levings 1993, Schnable and
Wise 1998). In fact, there are only two examples of CMS-associated genes that exhibit
sequence similarity. In B. napus the 07722 (napus cytoplasm) is 79% similar at the
amino acid level to 071224 (polima cytoplasm) (L’Homme et al. 1997). The 071707
found in the A3 mtDNA of sorghum and the 3’ region of 07179 flom the rice CMS-Bo
cytoplasm also exhibit a high level of sequence similarity (Tang et al. 1996). In
contrast, the identifiable gene regions involved in these chimeric ORFs predominantly
involve portions of genes coding for subunits of the ATP synthase and not other
mitochondrially encoded genes (reviewed by Schnable and Wise 1998, Conley and
Hanson 1995, see Table 5.3). Transcripts originating flom these altered reading flames
are translated into unique proteins that are flequently associated with the mitochondrial
membrane. Through the use of comparative physical mapping, and comparison of gene
expression patterns between sterile, fertile and restored plants, it has been shown that the
CMS-associated ORFs are responsible for pollen infertility. In addition for almost all of
the known CMS systems, nuclear restorer genes exist that can alter expression or function

of the chimeric ORF.

Fertility restorers and their relation to CMS-associated ORFs
In ahnost all cases of CMS, nuclear genes that suppress the cytoplasmic
dysfunction exist (Schnable and Wise 1998). The nature of these fertility restorers

provides clues to how CMS-associated ORFs may affect pollen formation. In the

155

 

majority of cases fertility restorers act to decrease expression of the RNA or protein level
of the CMS-associated ORF. For example, in the rice CMS-Bo cytoplasm, nuclear
restorers act by changing processing and editing of the 07179 transcript to restore fertility
(Kadowaki et al. 1990). In petunia the FR gene acts to decrease transcript abundance of
pcf (Pruitt and Hanson 1991, Nivison and Hanson 1989). The FR gene has been cloned
recently and encodes a mitochondrially-targeted protein of 14 repeats of a 35 amino acid
pentatricopeptide repeat motif (PPR proteins, Bentolila et al. 2002). These proteins have
been shown to have roles in RN A/protein and protein/protein interactions in other species

(Wise and Pring 2002).

Only one of the known fertility restorers causes any structural changes in the
CMS-associated cytoplasmic genes. Both the spontaneous reversion to fertility and
nuclear fertility restoration in Phaseolus vulgaris were originally thought to occur due to
the disappearance of the 210 kb subgenomic molecule that contains pvs-071239
(MacKenzie et al. 1988, Janska and Mackenzie et al. 1993, Mackenzie and Chase 1990).
More recent studies have shown, however, that the DNA molecule is maintained at low
levels in both fertile and fertility-restored plants and this molecule is amplified in CMS-
affected tissues (Janska and Mackenzie 1993, Janska et al. 1998, Arrieta-Monteil et al.

2001).

Multiple fertility restorers exist in the CMS-T system of maize. Rf8, Rf“ and Rfl

are all known to have effects on T477113 transcript accumulation (Wise et al. 1996,

Kennel] et al. 1987). These changes in transcript accumulation are correlated with an

156

 

80% decrease in abundance of the URF13 protein (Dewey et al. 1987, Kennel et al. 1987,
Kennell and Pring 1989). However fertility restoration with Rfl is not complete and
relies on the action of a second restorer, R12. Rf2 has been cloned, sequenced and
encodes an aldehyde dehydrogenase gene (Ciu et al. 1996, Liu et a1. 2001). R12 is the
only known restorer gene that does not directly affect either the DNA, RNA or protein
amount of a sterility-associated ORF. Instead it is thought to be a “compensatory
restorer” whose action compensates for the metabolic dysfirnction caused by URF13
protein. The role of Rf’Z will be firrther discussed in Chapter 6 in the section on the effect

of chimeric CMS gene expression on pollen development.

Summary:

One of the ways mitochondrial gene expression has been explored is through the
investigation of cytoplasmic male sterile and non-chromosomal stripe mutants. In these
plants either gene deletions or chimeric gene formation occurs, causing rrritochondrial
function to be disturbed. In either case recombination appears to play a role in generating
the deletion or chimeric gene formation. Analyses of these mutants show that the genes
affected most flequently belong to complex 1, complex IV, the ATP synthase or the

mitochondrial translation apparatus.

157

GOAL AND HYPOTHESIS

Knowing the types of genome changes that commonly occur in mitochondrial
disorders, the goal of the research presented in this chapter was to locate a change in the
mitochondrial genome of plants showing phenotypic effects on floral development and
fertility. These plants were generated by the over-expression of the E. coli RecA protein
in the mitochondria. Since RecA is a key factor in homologous recombination, its over-
expression should increase recombination and the potential for aberrant mtDNA
rearrangements. It was hypothesized that these plants would contain at least one
mitochondrial DNA change that could be correlated to the phenotypic alterations found.
It was expected that each of the plants could contain a distinct/different altered

mitochondrial region.

RESULTS

Rationale for probes chosen for mitochondrial DNA analysis studies

Although research has been done on the organization of the mitochondrial
genome of Nicotiana tabacum, a complete clone library was not available for use during
this study. Instead individual gene probes and larger cosmid clones containing potential
genes of interest were used from the Sears laboratory clone library. A complete list of
plant mtDNA clones and their reference names from the Sears clone library are listed in
Table 5.4. The clones anticipated to be the most useful fit one of two criteria: (1) they

had previously been seen to be altered in CMS, NCS or chm plants (example 5S, 18S

158

Table 5.4 Probes using in Southern blot analysis of T. and BC. plants

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Probe From Size Genes on plasmid] cosmid Reference
name (plant) (kb)
pmt- Nicotiana 16 Random mtDNA: SalI Aviv et al. 1984
sylsa8 sylvestris digest
TB#5-1 Maize 13 26S rDNA (13 kb) Gifl fl‘om Sam
Levings
TB#6—l Maize 16 5S and 18S rDNA, 777116 Chao et al. 1983
H10/59 Oenothera NA Mixed chloroplast and Shuster and Brennike
mitochondrial partial 1987
digests
H2/1 Oenothera NA 00x1 Gift from Axel
Brennike
Hl/23 Oenothera 6.4 alpha su-A T Pose Gift flom Axel
Brennike
7G1 Arabidopsis 25 nad4L, nad5c
07132106, 161,139, 110, 25,
1220
39E9 Arabidopsis 39 "7726, nad5de, nad9 ,7p116, Klein et al. 1994
rps3ab, trnN, cob
0713': 153a, 1070 ,131, 315, Unseld et al. 1997
206, 143, 1210, 167, 116
26F4 Arabidopsis 26 trnD, trnS, atp9, nadlb,
nadIc, atp6-2, atpl
0719: 262, 1050, 251, 106]:
152, 1060, 1110

 

159

 

and 7p]! 6 genes flom maize); (2) cosmid clones that contain large segments of DNA
provided the ability to scan the genome on a broad scale. For example three cosmid
clones from Arabidopsis thaliana of 25 kb, 26 kb and 39 kb covering 25% of the

Arabidopsis mtDNA were used (Unseld et al. 1997, Klein et al. 1994).

Investigation of aflected T 7 plants

Since the maternal inheritance of the phenotypic abnormalities had not yet been
confirmed, only a small scale investigation for mtDNA rearrangements was performed on
the T. progeny. In these studies, 16 T. plants derived flom To parent 10B were used,
while eight plants from the To parent 12 were used. All progeny were derived from self-
crosses and carried the recA gene in the nucleus. Since recA is still present in these
plants, and its activity should still be present, it is possible that rearrangements were
actively accruing. After PCR had been performed to characterize plants for the presence
or absence of the RecA gene, five plants without the nuclear RecA transgene were also

analyzed.

The total DNA of 16 progeny plants fl'om the line 10B was digested with the
restriction enzymes EcoR I and Hind III and Southern blotting was performed as
described (Chapter 2). The Hind III digested plants were analyzed with the probes pmt-
sylsa8, TB#6-l, H2/l and Hl/23. While the EcoR I digests were probed with pmt-sylsa8,
TB#5-1, TB#6-l and 1110/59. The total DNA from eight progeny plants flom line 12
were only digested with EcoR I and were probed with pmt-sylsa8, TB#5-1, and HID/59.

The total DNA flom five plants in line 12 which didn’t contain the nuclear RecA were

160

digested with EcoR I and probed with the Arabidopsis cosmid 761. This probe was
chosen specifically because a pilot study with this enzyme and probe on three highly

morphologically affected plants from the T. line 108 had shown an RF LP.

In each of these investigations, no mtDNA changes were found when T. plants
were compared to the wild-type Samsun plants (data not shown). It is recognized that
these studies used a very restricted number of plants, enzymes and probes and the results
do not eliminate the possibility that an mtDNA alteration has occurred. A more thorough
investigation with a larger number of probes, a larger population size and more restriction
enzymes could yield mtDNA polymorphisms. In addition, only DNA flom leaves was
used for the analyses, while the majority of phenotypic alterations were detected in the

floral organs and pollen.

Investigation of aflected BC 1 plants

As explained in Chapter 4, T. plants that did not contain the nuclear RecA
transgene were crossed with pollen fi'om Samsun plants to create a backcross generation
(BC.) that would not contain the nuclear gene, but would carry maternally inherited
mitochondria that had been exposed to the transgene. Four lines of BC. progeny were
generated fl'om these crosses and were named afier the T. parent plants (31, 42, 44, 45).
These plants were all derived flom the same To plant (12; see Figure 4.4). All the BC.
lines included ten plants, except line 44, which contained 12 plants. Each of these 42

plants showed the expected matemally-inherited phenotypic abnormalities, and was,

therefore, analyzed for a correlated mitochondrial DNA change. For these analyses,

161

DNA flom all the plants were digested with the enzymes BamH I, EcoR I and Hind III.
In each case DNA fi'om all the plants digested with EcoR I was analyzed using the
Arabidopsis probes 7Gl, 39E9 and 26F4 with these two exceptions: plants in line 42
were not analyzed with 26F4 and plants in line 45 were not examined with probe 39E9.
All the DNA samples digested with either BamH I and Hind III were probed with 7G1

but not with 39E9 or 26F4.

No DNA polymorphisms were found for any plants when the restriction enzymes
BamH I and Hind III were used. No DNA polymorphisms were found for any plants
when the EcoR I digest DNA was examined with either the 24F4 or the 39E9 probe.
However when EcoR I was used for analysis with the probe 7G1, a polymorphism was
identified in nine of the 42 plants. A ~2.9 kb band of was found in the plants 31-7, 31-8,
31-10, 42-8, 42-9, 44-10, 44—12, 45-2 and 45-3 (Figure 5.13). This RFLP was confirmed
in three separate Southern blots for each of these plants and was never seen in any of
sixteen different Samsun plants that were used as controls. In each case, the appearance
of the band shows variable stoichiometry flom plant to plant, with plant 44—12 showing
the most intense RFLP band and other plants like 31-8 and 31-10 showing a less intense
RFLP pattern. The genes and open reading flames that were recognized by this cosmid
and could be affected are listed in Table 5.4. This probe contains two genes flom
complex I (nad4L and nad5) and six open reading frames with unidentified firnctions.
Taking the results of previous work on mitochondrial dysfunction into account, it would

be tempting to surmise that the complex 1 genes might be effected.

162

“512345678 812345518

- L
at

5'3 A p

3.0 » .

3 b
43 r l I ' 4 ‘
3.6 o I l ' ' 3_,: ’

23D

 

1.9 e 23 ’
19 ’
1.4 r
I
13 1.4 p
13 ’.
03"
o
0.7. '

 

Figure 5.1:

(A) Agarose gel of EcoR I digested DNA from selected plants in the lines 31 and 44 . In
the lane marked MW is a lambda DNA marker (BstE H digest), sizes are as indicated.
Lanes marked S refer to control DNA from wild-type tobacco (Nicotiana tabacum var
Turkish Samsun). Plants are in orderl=3l (T. parent), 2=31-6, 3=3l-7, 4=31-8, 5=31-10,
6=44~10, 7:44-12.

(B) Southern blot probed with the cosmid clone 7G1. The black arrow indicates the
mitochondrial DNA RFLP that was identified in nine of the BC. plants.

163

However, hybridizations with sub-clones of the cosmid clone need to be performed to

truly pinpoint the affected gene(s) or gene regions.

The relationship between phenotypic expression of floral abnormalities and
sterility as it relates to plants that harbor the ~2.9 kb rearrangement is shown in Figure
5.2. In general, for each of the lines, plants that display the RFLP have the highest levels
of both floral abnormalities and sterility for their particular lineage (plants 31-7, 42-8, 42-
9, 44-10, 44-12). The fact that the two plants with the lowest amount of floral
abnormality in line 31 contain the RF LP is in contradiction to this trend, however these
plants presented high sterility levels that could be correlated to its presence. This same
reasoning could be used to explain the presence of the RFLP in plants 45-2 and 45—3,
which do not show high (>SO%) levels of floral abnormalities, but still are markedly

sterile.

While the presence of this RFLP is not directly correlated to phenotypic
manifestation, the lack of appearance in any of the sixteen control samples, in various
repetitions, led me to believe that this RF LP is not a naturally occuning subgenorrric
form. The inability to correlate strong phenotypic expression with the presence of the
RFLP, taken along with the variable stoichiometry of the RFLP itself, could be attributed
to sampling bias. All the DNA samples used in this study are from leaf tissue, yet the
phenotypic appearance of abnormalities is consistently manifested in the floral organs, a
circumstance that will be expanded upon in the discussion section of this chapter. The

possibility exists also that

164

 

eeeeeenee

% ab flowers
% sterile 100

meager.—

%ab flowers 9
%sten‘le 100 100 95 100 99 84 100 97 99 100

.4... assesses his:

%abflowers 10
%sterile 47 170 100 90 100 98 99 100 91 100

..Neahhhkh‘ihhk‘eee

%abflowersS
%sterile 90 71 94 94 99 100 100 100 100 100 1:) 100

 

 

 

 

Figure 5.2: One To plant (12) produced plants 31, 42, 44, and 45 among its T. progeny.
Progeny plants florn these lines are shown here as the BC.. Plants are organized by
increasing amounts of abnormality. Underneath each plant is the level of floral
abnormality it presented and its sterility as assessed by the number of fallow seed-pods
produced. The eight plants with white pots are those that contain the 2.9 kb mitochondrial
RFLP with the probe 761 and their sterility and abnormality data is in bold, and their
identification numbers in order of presentation are 31-10, 31-8, 31-7, 42-9, 42-8, 45-3,
45-2, 44—12, 44-10. Images in this dissertation are presented in color.

165

these phenotypes are due to alternative mtDNA changes that were not detected in this

limited screen of the mitochondrial genome.

DISCUSSION

The goal of the research presented in this chapter was to analyze the
mitochondrial DNA of plants that contained the E. coli RecA gene over-expressed in
their mitochondria and to correlate a mtDNA change with the phenotypic expression of
floral abnormality and sterility in these plants. While no changes were seen in a limited
screen of the T. plants, nine of the 42 BC. plants were found to contain an RFLP when
probed with the Arabidopsis cosmid clone 761. In general, the plants that had the RF LP

displayed high levels of both floral abnormality and sterility for their particular lineage.

This screening of the mtDNAs florn the T. and BC. lines was limited in that only
a small number of enzymes and probes were used and, therefore, not all of the possible
mtDNA changes that exist could have been identified. In fact it is foreseen, given the
fact that RecA stimulates recombination at repeat units and that most mitochondrial
genomes contain multiple repeats of various sizes, that numerous different mtDNA
changes could have occurred in these plants. Despite the phenotypic similarities, many
different gene changes can be responsible for their appearance, as is evidenced by the
variety of genes affected in other CMS systems. In order to fully characterize the effects
of over-expressing RecA, a much more thorough search of the genome for mtDNA

changes should be performed. Once all the various changes in each of plants have been

166

 

characterized, RNA and protein production would be checked for alterations in floral
tissue before the RFLP could be truly correlated to the appearance of sterility and floral
abnormality. Therefore, the identification of the ~2.9 kb RF LP in these plants represents
merely the first step in the process of identifying the effect that over-expressing E. coli
RecA has on mitochondrial DNA organization and its subsequent phenotypic
manifestations. In addition to its identification, this region should be finely mapped to
see what specific gene/gene regions are affected and if there is a repeat unit involved in

the RFLP formation.

The research to finely map the potential RF LP is particularly necessary because
analyses with two other restriction enzymes (BamH I and Hind HI) did not identify any
DNA changes. It is unusual to find an RFLP in digests with only one enzyme, as

rearrangements normally affect the digestion patterns of other enzymes as well.

Although the presence of this potential RFLP in only nine of the 42 BC. plants
and the variability of its stoichiometry in those plants is a detriment to providing a firm
correlation of its expression with the phenotypes, it is not an unexpected result. The
mosaic phenotypic presentation of floral abnormalities and sterility in the BC. plants
strongly suggests that amplification and sorting out of mitochondria and mitochondrial
DNA is actively occurring (Chapter 4). As reviewed previously in other systems,
amplification of these types of aberrant genes occurs in certain tissues. In fact, it is often
the threshold level of a mtDNA change that drives the production of a specific
phenotype. The Southern blotting in these analyses was performed using total DNA

isolated flom leaf tissues. Leaf tissues are relatively low in their mitochondrial numbers

167

and also are not the primary tissues expressing a phenotype through which mitochondrial
dysfunction was assessed. It is quite conceivable that leaf tissues in these plants
contained low levels of this rearrangement that were amplified or segregated out of the

floral meristem to cause the phenotype.

With this in mind, it is suggested that a more sensitive assay for the
rearrangement would be necessary. For example, once the RFLP is finely mapped, a
quantitative PCR based assay could be designed to identify levels of the rearrangement in
floral tissues compared to leaf tissues. This would allow the establishment of a more
cogent link between the phenotype and the presence of the mtDNA RF LP. Although it
presents an initial challenge in terms of molecular characterization, the mosaic phenotype
of the transgenic lines provides a unique chance to explore how changes in mitochondrial
DNA population dynamics affect the process of floral development. Unlike other CMS
systems in which mitochondrial dysfunction affects pollen and floral development
universally on the plant, these lines display phenotypes that vary over space and
developmental time. This makes them a unique resource in plant mitochondrial genetics
in that they offer the ability to follow both segregation and amplification as phenotypic

dysfunctions are generated during all stages of the floral developmental process.

168

CONCLUSIONS

This chapter describes the molecular analyses of transgenic plants carrying E. coli
RecA targeted to plant mitochondria. It was expected that, if recombination caused the
matemally-inherited floral abnormalities and sterility that were seen in these plants, a
mitochondrially based DNA change could be correlated to these phenotypes. In nine of
the 42 BCl plants examined a mitochondrial RFLP was identified. In general, this RFLP
was found in plants that had high levels of both floral abnormality and sterility.
However, further research needs to be performed to identify and correlate expression of
this RFLP with the phenotypes to establish its role in the production of floral

abnormalities and mitochondrial dysfunction.

169

 

CHAPTER 6

Conclusions and Future Directions

SUMMARY OF RESULTS

Previous studies of plant mitochondrial DNA have shown that the dynamic
structure of its mitochondrial genome is dependent on recombination and amplification of
recombined molecules. By testing the hypothesis that homologous recombination induced
by the RecA protein is involved in the production of aberrant mitochondrial DNA
rearrangements, this research strove to understand this frmdamental genetic process. In
addition the effect that changes in mitochondrial genome organization and subsequent

mitochondrial firnction could have on plant development were observed.

If recombination via RecA plays a role in generating mitochondrial DNA
rearrangements, it was expected that the effects of those rearrangements could be
witnessed phenotypically as variegation, cytoplasmic male sterility, aberrant leaf
development and/or dwarfisrn. In these studies transgenic tobacco plants that contained
the E. coli RecA gene targeted to their mitochondria showed phenotypes characteristic of
tobacco plants alloplasmic for mitochondria, homeotic-like changes during floral
development. However, unlike the cybrid plants, which show a uniform phenotype in all
flowers, abnormalities in my transgenic plants were expressed as a mosaic with normal
flowers appearing next to those with developmental abnormalities. It was concluded that

the presence of RecA stimulated recombination at multiple sites creating rearrangements

170

 

that, when amplified and segregated out, affected mitochondrial function to generate the

phenotypes.

RELATIONSHIP OF MY OBSERVATIONS TO MODELS OF CMS ACTION
AND MITOCHONDRIAL SIGNALING

The most commonly affected genes involved in the plant mitochondrial disorders
NCS and CMS involve the electron transport chain, the F.-F0 ATPase or the translation of
mitochondrial messengers (Levings 1993; Tables 5.1, 5.2, 5.3). However any
mitochondrial gene could conceivably be involved if there is something universal or
different about the way the mitochondrial gene is expressed or the way it fimctions in the
anther during pollen development (Levings 1993). In fact, the mechanisms of CMS are
still unclear even when the chimeric genes are well characterized, as most CMS-associated
genes are constitutively expressed in both vegetative and floral organs, but only affect the
anther and formation of pollen (Schnable and Wise 1998, Conley and Hanson 1995). In
addition, the CMS phenotypes are not uniform in all plants. In general the first
cytological abnormalities appear in the pollen tapetal cell layer, followed by abortion of
the microspore (Budar et a1. 2003). In tobacco, carrot (Kitagawa et al. 1994) and

Plantogo overall floral morphology can be altered as well (Van Damme et a1. 1983).

It is puzzling why the CMS-associated mitochondrial dysfunction interferes solely
in pollen production, because mitochondrial electron transfer, ATP formation and
mitochondrial translation should be essential functions in every phase of plant growth and

development. Typically, the lack of a generalized physiological effect on plant growth is

171

 

explained with the logic that mitochondria are dispensable in most plant tissues since
alternative sources of ATP production (glycolysis) are available especially in cells with
chloroplasts (photophosphorylation). In addition to these functions, the cyanide-resistant
alternative oxidase pathway present in mitochondria allows plant tissues to continue
limited electron transport, NAD+/NADH cycling and production of carbon skeletons
without the use of the mitochondrial electron transport chain or ATPase functions (Conley
and Hanson 1995, Moore and Sideow 1991). These alternative biochemical pathways for
energy transduction and production of crucial metabolites may allow development to

continue properly in tissues that do not require high levels of ATP.

The specific detrimental effect on pollen development by mitochondrial gene
mutations is rationalized by the fact that the energy demand on anther cells is greatest at
that stage of development and the function of the alternative pathways cannot compensate
at this time. This is supported by the fact that there is a 40-fold increase in mitochondria
per cell in the tapetal layer of maize anthers and a 20-fold increase in sporogenous tissues;
such large increases in mitochondrial number are not evidenced in any other maize tissues
(Warmke and Lee 1977). In addition, expression levels of nuclear and mitochondrial
genes encoding subunits of respiratory enzymes change strikingly during pollen
maturation (Moneger et al. 1992, Conley and Hanson 1994, Smart et al. 1994, Lalanne et
al. 1998, Wen and Chase 1999). With this in mind, it was hypothesized that the chimeric
gene products in CMS plants somehow interfere with the normal physiology of the
mitochondria. Since these chimeric genes include portions of normal mitochondrial

polypeptide sequences they could decrease the efficiency of respiration or ATP production

172

by interacting with the properly expressed subunit or if they share control regions with the
naturally occurring genes they could compete for transcription factors. Since pollen
development requires extensive increases in mitochondrial numbers and energy, if energy
production is altered even slightly, pollen development could be affected. The stumbling
block for this model of CMS action is that it can not be explain why other energy-
dependent developmental stages that lack functioning chloroplasts (i.e. germination, root

development) do not seem to be affected in these plants.

The model in which CMS is caused by a dysfunctional mitochondrial metabolism
is supported by the formation of transgenic beet plants that have anti-sense expression of
the mitochondrial pyruvate dehydrogenase gene. These plants are male sterile and it is
suggested that the inhibition of pyruvate dehydrogenase activity in the anther tapetum may
prevent the conversion of pyruvate to acetyl-CoA. As a result, the TCA cycle can no
longer operate at a sufficient rate to meet all the requirements of tapetal cells. The
formation and degeneration of various tissues during pollen development as well as the
regulation and synthesis of anther specific transcripts may impose high demands for
energy and key biosynthetic intermediates produced by the mitochondria. Under the
conditions, the TCA cycle would need to operate flrlly, since it is an important source for
many of the intermediates required for biosynthetic pathways in addition to its role in

oxidative energy production (Y ui et al. 2003).

A second hypothesis that exists for the pollen-specific action of CMS is that the

chimeric CMS-gene products interact with an unknown anther specific factor. This

173

interaction then triggers a deleterious cascade of events that leads to the abortion of
microspores (reviewed in Wise et al. 1999). This model was originally developed by
Flavell (1974) in regards to CMS-T, where the URF13 protein is known to be present in
roots, shoots, leaves, tassels and ears but is only toxic to anthers (Levings, 1993). An
anther-specific factor is expected to interact with URF13 to change the permeability of the
inner mitochondrial membrane, destroying mitochondrial activity and causing cell death.
The flmdarnental problem for this model to date is that no anther-specific compound has

been found to interact with URF13 (Budar et al. 2003).

The recent cloning of the Rf2 restorer gene of CMS-T plants adds a new
dimension to our understanding of CMS action. The Rf2 gene encodes a mitochondrial
aldehyde dehydrogenase (Cui et al. 1996, Liu et al. 2001), a protein whose main firnction
in mammals is the detoxification of ethanol-derived acetyladehyde (Lindahl and Peterson
1991). There are two hypotheses on how the action of Rf2 could restore fertility. In the
metabolic hypothesis, it is assumed that URF13 acts to alter basic mitochondrial fimction
and the production of additional aldehydes by R12 restores fertility through a catalytic or
detoxifying action. In the interaction hypothesis, it is suggested that the Rf2 protein either
directly or indirectly interacts with URF13 to eliminate its deleterious effects. For
example, Rf2 could catalyze the oxidation of an aldehyde component of the irmer
mitochondrial membrane thereby altering the interaction between URF13 and the
membrane where it naturally accumulates (Budar et al. 2003). While it provides a greater
scope of how mitochondrial biosynthetic and/or redox reactions might play a role in CMS

dysfunction, the function of Rf2 does not provide a greater understanding of whether it is

174

a disruption of energy production or an interaction between CMS and anther specific

proteins that generates the CMS phenotype.

While the work on CMS-T does not provide definitive support for either F lavell’s
(1974) interaction or the metabolic deficiency hypothesis, research performed on the
homeotic-like CMS plants in carrot and tobacco support the role of energy metabolism as
being important to phenotypic expression. These studies (F arbos et al. 2001, Berelebide et
al. 2002, Bergman et al. 2000, Linke et al. 2003) show that co-operative genetic
interactions exist between nuclei and mitochondria during flower development of
Nicotiana and Daucus. Ultimately improper mitochondrial fimction interferes with the

expression of known nuclear MADS box genes that implement proper floral development.

Normal floral development has been most actively studied in Arabidopsis and
found to be controlled by the action of homeotic MADS-box transcription factors
classified as A, B, or C firnctional genes (Coen and Meyerwitz 1991, Okamuro et a1.
1993). These organ identity genes are transcribed fl'om an early stage in floral meristem
division and their specific overlapping actions demark each of the four floral whorls. The
A-function genes (apetalaI and 2: APl AP2) are expressed predominantly in whorls 1 and
2, whereas the B-function genes (pistillata: PI, apetela3: AP3) are expressed in whorls 2
and 3, while whorls 3 and 4 are where C function genes (agamous, AG) are expressed.
Normal formation of sepals requires only the genes of the A family, while joint expression
of A and B genes creates petals. Stamen production also involves co-expression, this time

of B and C genes, while only C genes need be present for carpel development. The A and

175

C genes are antagonistic to one another and when either one is not expressed, homeotic
conversions of floral organs occur (reviewed in Lohman and Wei gel 2002). In addition to
these basic genes, a fourth class (B, sepetella) of genes is required for petal, stamen and
carpel formation to occur properly (Theisen and Saedler 2001). In addition, multiple
upstream genes have been recognized in this signaling pathway, such as leofiv and
superman. Leafy is a central player in the establishment of floral meristem identity and
the regulation of flower homeotic gene expression (Okamuro et al. 1993), while the
superman gene is a cadastral gene that is involved in controlling the boundary division
between whorls 3 and 4 as well as some aspects of cell division (Bowman et al. 1992,

Meyerowitz et al. 1995).

In a study by F arbos et al. (2001), the link between mitochondrial dysfunction and
the nuclear genetics of flower production was explored. This study used an alloplasmic
male sterile line (created by fusion of the N. tabacum nucleus with N. repanda cytoplasm)
in which abnormal flowers are produced. These flowers exhibit several anomalies: short,
less pigmented petals, 4-5 shortened stamen filaments with shriveled anthers capped with
stigmatoid structures and fusion of stamen to the carpel (Farbos et al. 2001). Molecular
characterization of the mtDNA showed that a novel reading frame in which atpl is co-
transcribed with an upstream ORF (0712 74) is responsible for the phenotype, as this co-
transcript accumulates only in male-sterile and not male fertile lines. In addition, direct
measurements showed that alloplasmic sterile cells had an ATP + ADP pool that was 55%
smaller than those of male fertile cells. The overall ratio of ATP/ADP in these cells was

also significantly lower, reflecting a less active energy metabolism (Bergman et a1. 2000).

176

The floral abnormalities are uniformly expressed in these plants and therefore were
hypothesized to be due to a slower development caused by the low levels of cellular ATP,
The change in ATP/ADP ratios was hypothesized to affect the synchrony of the floral

genetic signaling cascade (Farbos et al. 2001).

When expression patterns of organ identity genes were investigated in the
alloplasmic tobacco plants, expression patterns were comparable to wild-type in regards to
NFL (leafii homologue, Kelly et a1. 1995), NtDEF, NtGLO (B genes, Hansen et al. 1993)
and NAG] (C gene, Kempin et al. 1993; Farbos et al. 2001). However, it did appear that
coordination between the cadastral genes and organ identity genes might be altered,
suggesting that the mitochondrial dysflmction might be affecting the expression of the
superman (SUP) gene. In Arabidopsis, superman is initially activated by leafiz and
regulated by the presence of AP3, PI and AG. To analyze this possibility, male-sterile
transgenic tobacco that over-expressed the Arabidopsis SUP gene were made. These
plants displayed partial restoration of floral development; the boundaries between whorls
3 and 4 were restored to normal morphology, stamen morphology also improved and
functional pollen was produced. This work, therefore, supports a model in which
mitochondrial energy production is involved in some respect with the floral developmental
signaling cascade. It was suggested that the low ATP/ADP levels produced could have
acted as a stress signal therefore deregulating the expression of the SUP gene (F ar'bos et

a1. 2001, Berelebide et al. 2002).

177

 

Other alloplasnric lines also show flower development to be uniformly altered due
to changes in MADS box gene expression. In carrot, carpeloid CMS flowers with known
mitochondrial gene changes were found to contain altered transcript patterns for the B
class homologues DcMADS2 and DcMADS3. While these genes were properly induced,
they showed a decrease in transcription relative to fertile carrots. The production of
carpeloid flowers, in correlation with the reduced expression of the late B-ftmction activity
suggests that a threshold of B gene expression is required for stamen formation to occur
properly (Linke et al. 2003). Similarly, homeotic flower modifications in alloplasmic
cybrids of Nicotiana and Hyoscyamus show a significant decrease in transcripts of the
GLOBOSA gene, providing support for cytoplasmic influence on the expression of B
activity genes (Zubko et al. 2001). Additional support for the role of mitochondria in
MADS box gene signaling can be found in photoperiod-sensitive, pistilloid wheat plants
where the expression of the class B homologue WAP3 was greatly reduced (Murai et al.

2002)

Taken together these studies suggest that a disturbed interaction of nuclear and
cytoplasmic gene products leads to an impaired expression of MADS box genes,
especially the B-firnction and SUPERMAN genes, which in turn results in flower
malformation and male sterility. While the regulatory feedback mechanisms of plastid to
nuclear signaling are known to involve redox-regulated phosphorylation steps and
metabolites such as porphyrins, sugars and reactive oxygen species (Rodermel 2001), very

little information exists about the signals between mitochondria and the nucleus that

178

 

regulate transcription in plant cells. One signaling molecule that is known to be active in

this pathway is heme (Leon et al. 1998, Susek and Chory 1992).

Recent data suggest a role for mitochondrial reactive oxygen species (ROS) as a
component for stress-induced mitochondria to nucleus signaling A recent study by
Maxwell et a1. (2002) on tobacco cell cultures, explored the effects of several oxidative
stress-causing agents [antimycin A (AA), salicylic acid (SA) and hydrogen peroxide
(11202)]. These studies found that treatments with these molecules led to a rapid rise in
intracellular ROS levels. Disruption of normal mitochondrial function was caused which
altered gene expression of seven cDNAs showing similarity to genes induced in
programmed cell death (Maxwell et al. 2002). This finding supports the idea that the
mitochondria may act as a central depot in the cell where diverse stress stimuli are

integrated (Lam 1999) and relayed to bring about changes in gene expression.

In the Maxwell et al. (2002) study, assays were performed with bongkrekic acid, a
known inhibitor of the mitochondrial permeability transition pore (PT pore), which is
critical to animal programmed cell death (PCD). When cells were pre-treated with
bongkrekic acid, the gene inductions seen with AA, SA and H202 were blocked (Maxwell
et al. 2002). The opening of this pore results in the collapse of the electrochemical
transmembrane potential and uncouples oxidative phosphorylation, resulting in a drop in
ATP formation leading to cell death (Desagher and Martinou 2000). Outside of apoptosis,
there is an emerging view that the PT pore can Open transiently allowing for the

movement of a number of small molecules in and out of the mitochondrion; those

179

 

molecules may be important to normal signaling events that occur between the
mitochondria and cytosol. This also correlates with a theory by Schon (2000) about the
nature of cells affected by nritochondrial mutations. The animal cells most affected by
mitochondrial diseases are those containing excitatory cells, and not those that only
require high ATP levels to function. For example, there are few mitochondrial diseases of
the liver, which is highly oxidative but not excitatory. The essential nature of excitatory
cells is due to the import and/or export of molecules across the plasma membrane. Schon
(2000) incorporates this with the observation that pumps and channels require ATP and
are used in ATP compartrnentalization within the cell. He hypothesizes that there is a
relationship between ATP use and ATP compartrnentalization that is important to
phenotypic display of mitochondrial disorders. However, it is also possible that this
phenomenon could relate to nritochondrial firnction through a change in the movement of

signaling molecules through the PT pore.

This information may be relevant to our conception of CMS action and its
relationship to mitochondrial signaling. Certain CMS proteins, like URF13 and
ORFH522, are known to be localized to mitochondria membranes and perhaps interact
with other proteins to induce pollen degeneration, perhaps via a cell death-signaling
pathway. In addition to its localization, expression of ORFH522 has been shown to play a
role in programmed cell death (Balk and Leaver 2001). Also lending support to the
potential relationship between URF13 and signaling via ROS is the fact that the restorer

Rf2 is an aldehyde dehydrogenase. The action of Rf2 is likely to change the concentration

180

 

of metabolites, ROS and signaling molecules in the cell that may alter the status of the

mitochondria and, therefore, change gene expression to restore fertility.

The data of Maxwell et al. (2002) suggest that the mitochondrion, in addition to
being the “powerhouse of the cell”, can serve as an intermediary in intracellular stress
signaling. It does so both by interpreting stress signals followed by initiating the proper
response to that stress. These responses could be subtle adjustments in cell function
through alterations in nuclear gene expression as well as such severe responses as
triggering cell death. Hence, changes in mitochondrial function caused by mitochondrial
DNA rearrangements could be read as a stress signals. The response to these stress
signals could be altered gene expression of MADS box genes or the initiation of

programmed cell death.

Studies in animals have begun to explore the phenomenon of tissue based
threshold effects and mitochondrial segregation on disease presentation/signaling. In fact
work is underway to introduce pathogenic mtDNA mutations in mouse germ line cells to
permit direct investigation of how stochastic fluctuation of heteroplasmic mtDNA
mutations affect energy production, oxidative stress and apoptosis (Wallace 1999). In
plant mitochondrial syndromes, this aspect of segregation and threshold effects has not
been explored. Amplification and segregation of mtDNA subgenomes have been
observed in plant tissues, however those processes have not been implicated in changes in
phenotypic presentation. With the exception of non-chromosomal stripe mutants, this is

likely attributable to the fact that the most common mitochondrial disorder in plants

181

 

(CMS) presents a uniform phenotype in pollen and floral tissues. Similarly tobacco and
carrot homeotic alloplasmic lines have 100% phenotypically affected flowers. The
transgenic plants that were produced are, therefore, phenotypically novel in that the
manifestations of mitochondrial disorders are not unifonnly presented. The mosaic nature
of the appearance of the floral abnormalities, in conjunction with the number and
variability of these types of mutations, suggest that active segregation is occurring in these
plants.

If active segregation and threshold levels are responsible for this phenomenon, it is
hypothesized that normal flowers would have received a lower percentage of abnormal
mtDNAs, while flowers with homeotic conversions would be likely to have higher
quantities of mutant mtDNA molecules. Individual tissues appear to be sensitive to
threshold as well, with the stamen being the most responsive floral organ to the threshold
effect caused by mitochondrial dysfimction. Referring to the observations in Chapter 4,
both the T. and BC. generations had many flowers with a mixture of anther phenotypes.
The most common developmental defect was infertility but that was followed by the
homeotic-like conversion of the anthers into petals. This led to the hypothesis that each
physical manifestation is reflective of the amount of aberrant mtDNA in the particular
tissue, with an infertile phenotype appearing at lower quantities of aberrant mtDNA than
the phenotypes that cause organ conversion. This would be due to the fact that production
of pollen is so energy intensive that even at low levels of mitochondrial impairment, it
cannot proceed properly. The effect of mitochondrial segregation and its impact on
threshold applies to the specific organs as well, with the threshold needs of the organs and

the nature of the mutated mitochondrial gene(s) playing a role in phenotypic

182

manifestation. The gene that is mutated is considered to be important in light of the
results found with the NCS mutants and induction of the alternative oxidase (Karpova
2002), in which the site of a mutation was found to affect nuclear gene expression. These
observations are interpreted such that

the effect of the threshold amount of aberrant mtDNA is greater in organs, such as
flowers, which do not have chloroplasts to aid in energy production, than in other organs

of the plant.

FUTURE DIRECTIONS

In the scope of known CMS action, the plants generated by this research provide a
new material to explore how segregation of mitochondrial DNA molecules might play a
role in gene expression as it relates to pollen and floral development. To really explore
these possibilities will involve a continuation of the groundwork laid by these initial
studies. First, a thorough search for all potential mitochondrial DNA-based changes that
RecA could have caused should be performed, to determine if the 2.9 kb RF LP is the only
change in these plants or if other mtDNA alterations also are present. Secondly, these
changes should be finely mapped and the amounts of their DNA, transcript and protein
levels in floral tissues should be quantified. After these types of comparisons have been
made, then specific effects of these transcripts and/or proteins on homeobox genes and
mitochondrial metabolites can be examined. Current studies on carrot and tobacco have

only investigated stigmatoid floral abnormalities; the abundance of homeotic-like flowers

183

in my transgenic plants allow for the inter-relationship between mitochondrial function

and other genes in the developmental cascade to be explored.

In addition there are aspects of mitochondrial genome organization that were not
explored in this research but could also be followed in these plants. For example
transgenic lines that contain the AN42 deletion mutant of RecA were also created, and the
effect of decreasing recombination on mtDNA structure and flmction are not known.
Since observations of plant mtDNA structure through pulse field gel electrophoresis have
been interpreted to indicate that massive recombination complexes exist in vivo, my
transgenic materials (both the E. coli RecA and AN42 deletion mutant of RecA) could be

used to test these interpretations.

184

 

111'

Appendix

Protein Localization Experiments

Introduction

Chapter 3 described my efforts to verify that the transgenic plants had integrated
the nuclear transgene. To assure that the nuclear transgene was translated and transported
to the mitochondria, protein localization experiments were undertaken. This protein

localization work is presented here.

Results

Mitochondrial proteins were isolated fi'om leaves of tissue culture grown plants or
from etiolated seedlings in an attempt to localize the mitochondrially-targeted E. coli
RecA protein. Proteins were run on polyacrylamide gels and western blotted with three
mitochondrial antibodies in addition to a monclonal antibody for RecA. Monoclonal
antibodies were chosen for two mitochondrial proteins that are known to be naturally low
in abundance (HSP70 and AOX) and one that should be highly abundant (COXII). As a
control for RecA, protein isolations from E. coli were run on each gel. As a control for

mitochondrial proteins, voodoo lily mitochondrial protein isolates were used.

Figure A.1 shows a western blot of leaf proteins that was reacted with the AOX
antibody and with the RecA antibody (data not shown for HSP70). It can be seen that
only the control protein isolates from Voodoo Lily and E. coli react with the antibodies as
expected. It was a concern in these experiments that the amount of mitochondrial

proteins was not large enough for detection, either because of the low abundance of the

185

 

 

ESL—73
%982...‘
AOX
DEN?)
>9"
RecA § ”’2‘”

37 kD

 

 

Figure A.1: Western blots of leaf tissue. WT12 refers to pooled T. progeny from the
progenitor plant 12. The abbreviation Sam refers to control plants, VD for Voodoo lily

protein extracts. Expected protein sizes are marked with an arrow.

(A) AOX antibody (1:5000 dilution)
(B) RecA antibody (1:100,000 dilution)

186

proteins in mitochondria or because the isolations did not contain enough mitochondria.
To solve this problem, an antibody for a higher abundance mitochondrial protein
(COXH) was obtained for use as a control. Etiolated seedlings, which should be a tissue

enriched for mitochondria, were used in addition to leaf tissue.

Figure A.2 shows a representative western blot using the COXH and HSP70
antibodies. While mitochondrial proteins were detectable for the high abundance protein
COXH, from both the Samsun and transformed seedlings, it was only detectable in leaves
of the Samsun plants. However, when antibodies for the low abundance protein HSP70

were used, only Voodoo lily showed a protein.

Figure A.3 shows these same protein isolations when probed with the RecA
antibody. Unlike the previous blots with the RecA antibody, which had shown no
reactions with plant mitochondrial protein isolates, many nonspecific bands were found
with these protein isolates. Even in E. coli and Voodoo lily (which previously showed
no interaction to the antibody), extra background banding was found. In these blots faint
bands at about 40 kD can be seen in both the Samsun control and transgenic plant lines.
Taken with the high levels of background reactions however, these bands were not
interpreted as representing the RecA protein, but rather are non-specific banding with

other proteins in the sample.

187

 

 

seeds leaves

 

 

      
 

2' ._
8 3 E g: E 5: f3
11.1 9 8 2 8'3 g in
Cox2 ~-- “ ' A
a
c «— - —- ~31 _1 ..__ 29110
‘
Hsp70

0 <———70I<D

Figure A.2: Western blots of proteins isolated fi'om leaves and seedlings of transformed
and untransformed tobacco. Samples are as described in Fig. Al . Expected protein sizes
are marked with an arrow.

(A) COXH antibody (1:5000 dilution)
(B) HSP70 antibody (1:5000 dilution)

188

 

 

 

 

= a? -—

8. 5' g 91 g e 8

UJ > a) g (I) g Lu

.. _ f

’4‘- ..‘I'. a

5“ ~ H has?

..... a... «- -~ ‘— 37KD

 

Figure A.3: Western blots of proteins detected by the E. coli RecA antibody. Proteins
were isolated fi'om leaves and seedlings of transformed and untransformed tobacco. All
transgenic plants are T1 progeny from the progenitor plant 12. The abbreviation Sam
refers to control plants, VD Lilly for Voodoo lily protein extracts. Expected protein sizes
are marked with an arrow. The RecA antibody was a 1:100,000 dilution.

189

 

Discussion

Mitochondrial protein isolations and subsequent Western blots did not provide
conclusive evidence to localize the RecA protein to the mitochondria. While Western
blotting could recognize a high abundance protein such as COXH, extracts did not have
enough protein so that low abundance proteins could be identified (AOX or HSP70).
Therefore the inability to detect proteins known to be low in abundance meant that
detection of the transgenic protein could be problematic. Initial attempts with leaf tissues
were unable to detect the RecA protein. Attempts to localize RecA in etiolated seedling
tissues, which should be enriched for mitochondria, were plagued by background

interactions.

Although expression in the transgenic plants is driven by the 358 CaMV
promoter, the E. coli RecA protein might be a low abundance protein, like AOX or
HSP70. Even in low amounts it could still stimulate recombination because the minimal
length in E. coli that is required for nucleoprotein filament formation and strand exchange
is about 10 monomers of RecA (covers ~40 nucelotides; Kowalczkowsld et al. 1994).
In fact, the only study in which RecA has been detected in mitochondria by Western
blotting was one in which a mitochondrially targeted E. coli RecA was expressed in
transfected human lung cells (Paul et al. 2001). In these studies the amount of imported
RecA protein was thought to be about 100 molecules per mitochondrion, a level that was

3 times higher than the value necessary to detect RecA in E. coli.

190

 

 

In addition to the potential problems with the abundance of the protein, finding a
tissue that is amenable to these experiments is important. Tissues such as etiolated
seedlings that should have provided increased numbers of mitochondria, were plagued by
problems with backgron interactions with the antibody. Given the limitations of
detection, plant tissue that has a high number of mitochondria such as tobacco cell
suspension cultures, may be necessary to provide adequate mitochondria for successful
immuno-detection. These lines were not originally established because cell culture lines
are naturally prone to mtDNA reorganization and hence would not have been useful for
the other analyses in this study. I did not pursue creating cell cultures as part of my
dissertation as they would have taken nine months to grow to the stage in which they
would be useful. These data, however, leave open the possibility that the nuclear
transgene is not transcribed, translated or properly imported into the mitochondria.
Although unable to identify a RecA protein in the plant mitochondrial extracts, the
phenotypic display and inheritance of abnormalities discussed suggest that mitochondria

are indeed affected by this transgene.

191

 

 

LITERATURE CITED

Abad AR, Mehrtens BJ, MacKenzie S (1995) Specific expression in reproductive tissues
and fate of a nritochondrial sterility-associated protein in cytoplasmic male sterile
bean. Plant Cell 22905-912.

Allen JF (1993) Control of gene expression by redox potential and the requirement for
chloroplast and mitochondrial genomes. J Theor Biol 165:609-631.

Allgood ND, Silhavy TJ (1988) Illegitimate recombination in bacteria. In: Genetic
Recombination (R Kucherlapati, GR Smith, eds), American Society of
Microbiology, Washington DC, pp. 309-330.

Andre C, Levy A, Walbot V (1992) Small repeated sequences and the structure of plant
mitochondrial genomes. Trends Genetics 8:128-132.

Arrieta—Montiel M, Lyznik A, Woloszynska M, Janska H, Tohme J, MacKenzie S (2001)
Tracing Evolutionary and Developmental Implications of Mitochondrial
stoichiometric shifting in common bean genetics. 158:851-864.

Aviv D, Arzee-Gonen P, Bleichman S, Galun F (1984) Novel alloplasmic Nicotiana
plants by donor-recipient protoplast fusionzcybrids having N. tabacum or N.
slyvestris chrondriomes from alien species. Mol Gen Genet 196:244—253.

Backert S, Lurz R, Oyarzabal 0A, Bomer T (1997) High content size and distribution of
single stranded DNA in the mitochondria of Chenopodium album (L.). Plant Mol
Biol 33:1037-1050.

Baker, R. F. and Newton K. J. (1995). Analysis of defective leaf sectors and aborted
kemals in NCS2 mutant maize plants. Maydica 40: 89-98.

Balk J, Leaver CJ (2001) The PETl-CMS mitochondrial mutation in sunflower is
associated with premature programmed cell death and cytochrome c release. Plant
Cell 13 (8): 1803-1818.

Banga O, Petit J, van Bennekem J L, (1964) Genetical analysis of male sterility in carrot
Daucus carota L. Euphytica 19:263-269.

Belliard G, Vedel F, Pelletier G (1979) Mitochondrial recombination in cytoplasmic
hybrids of Nicotiana tabacum by protoplast fusion. Nature 281 :401-401.

Bendich AJ (1993) Reaching for the ring: the study of mitochondrial genome structure.
Curr Genet 24:279-290.

192

 

 

Bentolila S, Alfonso AA, Hanson MR (2002) A pentatricopeptide repeat-containing gene
restores fertility to cytoplasmic male-sterile plants. Proc Natl Acad Sci USA 99:
10887-10892.

Berelebide A, Hemould M, Farbos I, Glimelius K, Mouras A (2002) Restoration of
stamen development and the production of functional pollen in an alloplasmic
CMS tobacco line by ectopic expression of the Arabidopsis thaliana SUPERMAN
gene. Plant J 29:607-615.

Bergman, P., Edqvist, J., Farbos, I and Glimelius, K. (2000). Male-sterile tobacco
displays abnormal mitochondrial atpl transcript accumulation and reduced floral
ATP/ADP ratio. Plant Mol. Biol 42: 531-544.

Bertrand H, Collins RA, Stohl LL, Goewert RR, Lambowitz AM (1980) Deletion
mutants of Neurospora crassa mitochondrial DNA and their relationships to the
stop/start growth phenotype. Proc Natl Acad Sci USA 77:6032-6036.

Binet M-N, Osman M, J agendorf AT (1993) Genomic nucleotide sequence of a gene for
Arabidopsis thaliana encoding a protein homolog of Escherichia coli RecA. Plant
Phys 103:673-674.

Birky CW Jr. (1983) Relaxed cellular controls and organelle heredity. Science 222: 468-
475.

Birky CW Jr. (1991) Evolution and population genetics of organelle genes. In: Evolution
at the Molecular Level (RK Selander, AG Clark, TS Whitham, eds) Sinauer
Associates, Sunderland, MA, pp. 112-134.

Birky (2001) The inheritance of genes in mitochondria and chloroplasts: laws mechanism
and models. Ann Rev Genetics 35: 125-148.

Bonen L, Brown GG (1993) Genetic plasticity and its consequences: perspectives on
gene organization and expression in plant mitochondria. Can J Bot 71:645-660.

Bonhomme S, Budar F, Lancelin D, Small 1, Defiance C (1992) Sequence and transcript
analysis of the Nc02.5 Ogura-specific fragment are correlated with cytoplasmic
male sterility in Brassica cybrids. Mol Gen Genet 235:340-348.

Bonnet HT, Kofer W, Hakansson G, Glimelius K (1991) Mitochondrial involvement in

petal and stamen development studied by sexual and somatic hybridization of
Nicotiana species. Plant Sci 80:119-130.

193

 

 

Bomer T, Linke B, Nothnagel T, Scheike R, Schutlz B, Steinbom R, Brennike A, Stein
M, Wricke G (1995) Inheritance of nuclear and cytoplasmic factors affecting male
sterility in Daucus carota. In U Kluck and G Ericke (eds). Genetic mechanisms
for hybrid breeding. Advances in plant breeding 18:111-112. Blackwell Science,
Berlin.

Boutry M, Chua N-H (1985) A nuclear gene encoding the B-subunit of the rrritochondrial
ATP synthetase in Nicotiana plumbagnifolia. EMBO 4(9) 2159-2165.

Boutry M, Faber AM, Charbonnier M, Briquet M (1984) Microanalysis of plant
mitochondrial protein synthesis products. Plant Mol Biol 3:445-452.

Bowman, J. L., Symth, D. R. and Meyerowitz, E. M. (1989) Genes directing flower
development in Arabidopsis. Plant Cell 1:37-52.

Bowman JL, Sakai H, Jack T, Weigel D, Mayer U, Meyerowitz EM (1992)
SUPERMAN, a regulator of floral hometoic genes in Arabidopsis. Development
114:599-615.

Breiman A, Galun E (1990) Nuclear-mitochondrial inter-relation in angiosperms. Plant
Sci 7 1 :3-19.

Brown GG (1999) Unique aspects of cytoplasmic male sterility and fertility restoration in
Brassica napus. J Hered. 90(3) 351-356.

Budar F, Touzet P DePaepe R (2003) The nucleo-mitochondrial conflict in cytoplasmic
male sterilities revisited. Genetica 117:3-16.

Camerini-Otero RD, Hsieh P (1995) Homologous recombination proteins in prokaryotes
and eukaryotes. Ann Rev Genet 29:509-553.

Cerutti H, Osman M, Grandoni P, J agendorf AT (1992) A homolog of Escherichia coli
RecA protein in plastids of higher plants. Proc Nat Acad Sci 89:8068-8072.

Cerutti H, Johnson AM, Boynton JE, Gillham NW (1995) Inhibition of chloroplast DNA
recombination and repair by dominant negative mutants of Escherichia coli
RecA. Mol Cell Biol 15:3003-3011.

Chada ML, Frese L (1981) Observations on a petaloidy type of male sterility in carrots
(Daucus carota L.) Gartenbauwissenschafl 46:21-24.

Chao S, Sederoff RR, Levings CS (1983) Partial sequence analysis of the SS to 188

ribosomal RNA gene region of the maize mitochondrial genome. Plant Physiol
71 :190-193.

194

 

 

Chase CD (1994) Expression of CMS unique and flanking mitochondrial DNA sequences
in Phaseolus vulgaris L. Curr Genet 25:245-251.

Chaumont F, et al (1995) Targeting the maize T -uer 3 product into tobacco mitochondria
confers methomyl sensitivity to mitochondrial respiration. Proc Natl Acad Sci
USA 92:1167-1171.

Chetrit P, Rios R, De Paepe R, Vitart V, Gutierres S, Vedel F (1992) Cytoplasmic male
sterility is associated with large deletions in the mitochondrial DNA of two
Nicotiana sylvestris protoclones. Current Genetics 21 :131-137.

Chinnery PF, Thorburn DR, Samuels DC, White SL, Dahl H-H, Turnbull DM,
Lightowlers RN, Howell N (2000) The inheritance of mitochondrial DNA
heteroplasmyzrandom drift, selection or both? Trends in Genetics 16:500-505.

Clough SJ, Bent AF (1998) Flora] dip: a simplified method for Agrobacterium-mediated
transformation of Arabidopsis thaliana. Plant J 16:735-43.

Coen ES, Meyerowitz EM (1991) The war of the whorls: genetic interactions controlling
flower development. Nature 353:31-37.

Conklin CL, Hanson MR (1994) Recombination of plant mitochondrial genomes. In:
Paszkowski J (eds) Homologous recombination and gene silencing in plants.
Kluwer Acadmic Publ. Dorchecht Boston, London, pp 61-81.

Conley CA, Hanson MR (1993) A truncated recombination repeat in the mitochondrial
genome of a Petunia CMS line. Curr Genet 23:477-482.

Conley CA, Hanson MR (1994) Tissue specific expression in plant mitochondria. Plant
Cell 6:85-91.

Conley CA, Hanson MR (1995) How do alterations in plant mitochondrial genomes
disrupt pollen development? J Bio and Biomem. 27: 447-457.

Cui X, Wise RP, Schnable PS (1996) The R12 nuclear restorer gene of male sterile T-
cytoplasm maize. Science 272:1334-1336.

Desagher S, Martinou J -C (2000) Mitochondria as the central control point of apoptosis.
Trends Cell Biol 10:369-377.

Dewey RE, Levings HI CS, Timothy DH (1986) Novel recombinations in maize

mitochondrial genome produce a unique transcriptional unit in Texas male sterile
cytoplasm. Cell 44:439-449.

195

 

 

Dewey RE, Timothy DH, Levings 111 CS, (1987) A nritochondrial protein associated with
cytoplasmic male sterility in the T cytoplasm of maize. Proc Natl Acad Sci USA
84:5374-5378.

Diaz, F. Bayona-Bafaluy, M. P., Rana, M., Mora, M., Hao, H., and Moraes, C. T. (2002).
Human mitochondrial DNA with large deletions repopulates organelles faster

than full-length genomes under relaxed copy number control. Nucleic Acids
Researh 30: 4626-4633.

Douce R, and Neurburger M (1989) The uniqueness of plant mitochondria. Ann. Rev.
Plant Physiol. Plant Mol Bio 40:371-417.

Doyle J J , Doyle J L (1987) A rapid DNA isolation procedure for small quantities of fresh
leaf tissue. Phytochem Bull 19:11-15.

Edwardson, J. R. (1970). Cytoplasmic male-sterility. Bot. Rev. 36, 341-420.

Eisa HM, Wallace, DH (1969) Morphological and anatomical aspects of petaloidy in
carrot Daucus carota L. J Am Soc Sci 94:545-548.

Elkonin LA, Tymov VS (2000) Studying genetic control of cytoplasmic male sterility in
plantszstate of the problem and current approaches. Russian Journal of Genetics
36 (4) 347-358.

Erikson EH, Garment MB, Peterson CE (1982) Structure of cytoplasmic male sterile and
fertile carrot flowers. J Am Soc Hort Sci 107: 698-706.

Farbos I, Mouras A, Bereteride A, Glimelius K (2001) Defective cell proliferation in the
floral meristem of alloplasmic plants of Nicotiana tabacum leads to abnormal
floral organ development and male sterility. Plant J 26:131-142.

Fauron C, Casper M, Gao, Y, Moore B (1995) The maize mitochondrial genome:
dynamic, yet firnctional. Trends Genetics 11:228-235.

Flavell R (1974) A model for the mechanism of cytoplasmic male sterility in plants, with
special reference to maize. Plant Sci Lett 3:259-263.

Folkerts 0, Hanson MR (1991) The male sterility-associated pd gene and the normal
atp9-1 gene in Petunia are located on different mitochondrial DNA molecules.
Genetics 129:885-895.

Gerstel, D. U. (1980) Cytoplasmic male sterility in Nicotiana (a review). N.C. Agric
Res Serv Tech Bull 263: 1—31.

196

 

Gillham N (1994) Organelle Genes and Genomes. Oxford University Press New York.

Glimelius K, Chen K, Bonnett HT (1981) Somatic hybridization in Nicotiana:
segregation of organellar traits among hybrid and cybrid plants. Planta 153:504-
5 10.

Gray W et a1 (1998) Genome structure and gene content in protests mitochondrial DNAs.
Nucleic Acids Reseach 26:865-878.

Gu J, Miles D, Newton KJ (1993) Analysis of leaf sectors in the NCS6 mutant of maize.
Plant Cell 5:963-971.

Gutierres S, Sabar M, Lelandais C, Chetrit P, Diolez P, Degand H, Boutry M, Vedel F,
Kouchkovsky YD, Paepe RD (1997) Lack of mitochrondrial and nuclear-
encoded subunits of complex I and alteration of the respiratory chain in Nicotiana
sylvestris mitochrondrial deletion mutants. Proc Natl Acad Sci USA 94:3436-
3441.

Gutierres S, Conbettes B, De Paepe R, Mirande M, Lelandais C, Vedel F, Chetrit P
(1999) In the Nicotiana sylvestris CMSII mutant a recombination mediated
change 5’ to the first exon of the mitochondrial nadl gene is associated to the lack
of the NADH-ubiquinone oxidoreductase (complex I) NADl subunit. Eur J
Biochern 261 :361-370.

Hansen G, Estruch JJ, Sommer H, Spena A (1993) NTGLO: a tobacco homologue of the
GLOBOSA floral homeotic gene of Antirrhinum majus: a cDNA sequence and
expression pattern. Mol Gen Genet 239:310-312.

Hanson, MR. (1991). Plant mitochondrial mutations and male sterility. Annu Rev
Genet 25: 461-486.

Hanson MR, Folkerts O (1992) Structure and firnction of higher plant mitochondrial
genomes. Int Rev Cyt 141:129-172.

Hanson MR, Wilson RK, Bentolila S, Kohler RH, Chen H-C (1999) Mitochondrial gene
organization in Petunia male fertile and sterile plants. J Hered 90(3):362-368.

Harris EH (1989) The Chlamydomonas Sourcebook: A comprehensive guide to biology
and laboratory use. Academic Press, San Diego.

Hart GE (1965) Studies on extrachromosomal male sterility in Nicotiana. Thesis,
University of California at Berkley.

197

 

 

 

Hartmann C, Recipen H, J ubier M-F, Valon C, Delchier-Basin E (1994) Mitochondrial
DNA variability detected in a single wheat regenerant involves a rare
recombination event across a short repeat. Curr Genet 25:456-464.

Havey MJ (1997) Predominant paternal transmission of the mitochondrial genome in
cucumber. J Hered 88:232-235.

Havey MJ McCreight JD, Rhodes B, Taurick G (1998) Differential transmission of the
cucurbit organellar genomes. Theor Appl Genet 97:122-128.

He SC, Abad AR, Gelvin SB, Mackenzie SA (1996) A cytoplasmically male sterility-
associated mitochondrial protein causes pollen disruption in transgenic tobacco.
Proc Natl Acad Sci USA 11763-11768.

Hemon P, Robbins MP, Cullimore JV (1990) Targeting of glutamine synthetase to the
mitochondria of transgenic tobacco. Plant Mol Biol 15:895-904.

Horii T, Ogawa T, Ogawa H (1980) Organization of the RecA gene of Escherichia coli.
Proc Nat Acad Sci USA 77:313-317.

Horii T, Ozawa N, Ogawa T Ogawa H (1992) Inhibitory effects of N- and C-terminal
truncated Escherichia coli RecA gene products on functions of the wild-type RecA
gene. J Mo] Biol 223: 105—114.

Horn R, Hohler RH, Zetsche K (1991) A mitochondria] 16 kD protein is associated with
cytoplasmic male sterility in sunflower. Plant Mo] Biol 17:29-36.

Horn R, Husledt J EG, Horstmeyer A, Hahnan J Zetsche K, Freidt W (1996) The CMS-
associated 16 kDa protein encoded by 01111522 in PET] cytoplasm is also present
in other male sterile cytoplasms of sunflower. Plant Mol Biol 30:523-538.

Horn R (2002) Molecular diversity of male sterility inducing and male fertile cytoplasms
in the genus Helianthus. Theor App] Genetics 104:562-570.

Hunt MD, Newton KJ (1991) The NSC3 mutation: genetic evidence for the expression of
ribosomal protein genes in Zea mays mitochondria. EMBO J 10: 1045-1052.

Iwahashi M, Nakazono M, Kanno A, Sugino K, Ishibashi T, Hirai A (1992) Genetic and
physical maps and a clone bank of mitochondrial DNA from rice. Theor Appl
Genet 84:275-279.

J anska H, Mackenzie SA (1993) Unusual mitochondrial genome organization in
cytoplasmic male sterile common bean and the nature of cytoplasmic reversion to
fertility. Genetics 135:869-879.

198

 

(E4; .

J anska H, Sarria R, Woloszynska M, Arrieta-Montie] M, Mackenzie SA (1998)
Stoichiometric shifts in the common bean mitochondrial genomes leading to male
sterility and spontaneous reversion to fertility. Plant Cell 10:1163-1180.

Kadowaki K-I, Suzuki T, Kazarna S (1990) A chimeric gene containing the S1 portion of
the atp6 gene is associated with CMS of rice. Mol Gen Gen 224: 10-16.

Kanazawa A, Tsutsumi N, Hirai A (1994) Reversible changes in the composition of the
population of mtDNAs during dedifferentiation and regeneration of tobacco.
Genetics 138:865-870.

Karlin S, Brocchere L (1996) Evolutionary conservation of recA genes in relation to
protein structure and function. J Bacteriol 178:1881-94

Karpova OV, Newton KJ (1999) A partially assembled complex I in NAD4-deficient
mitochondria of maize. Plant J 17:511-521.

Karpova OV, Kuzrrrin EV, Elthon TE, Newton KJ (2002) Differential expression of
alternative oxidase genes in maize mitochondrial mutants. Plant Cell 14:3271-
3284.

Katavic V, Haughn, Reed D, Martin M, Kunst L (1994) In planta transformation of
Arabidopsis thaliana. Mol Gen Genet 245:363-370.

Kaul MLH (1998) Male sterility in higher plants. Heidelberg: Springer Verlag.

Kelly AJ, Bonnlander MB, Meeks-Wagner DR (1995) NFL, the tobacco homolog of
FLORICAULA and LEAF Y, is transcriptionally expressed in both vegetative and
floral meristems. Plant Cell 7:225-234.

Kempin SA, Mandel MA, Yanofsky MF (1993) Conversion of perianth into reproductive
organs by ectopic expression of the tobacco flora] homeotic gene NAGl . Plant
Physio]. 103:1041-1046.

Kennel] JC, Wise RP, Pring DR (1987) Influence of nuclear background on transcription
of a maize mitochondrial region associated with Texas male sterile cytoplasm.
Mol Gen Genet 210-399-406.

Kennel] JC, Pring DR (1989) Initiation and processing of atp6, T -urfl 3 and 011221
transcripts fiom mitochondria of T cytoplasm maize. Mol Gen Genet 216:16-24.

Kearns CA, Inouye DW (1993) Techniques for Pollen Biologists. University Press of
Colorado, Niwot Co. pp101-106.

Khvorostov IB, Ivanov MK, Dymshits GM (2002) The nritochondrial genome of higher
plants: Gene expression regulation. Molecular Biology 36: 314-321.

199

 

 

Kitagawa, J. Posluszny, U., Gerrath, J. M., and Wolyn, D. J. (1994). Developmental and
morphological analyses of homeotic male sterile and fertile carrot flowers. Sex
Plant Reprod. 7:41-50.

Klein M, Eckert-Ossenkopp U, Schmiedeberg I, Brandt P, Unseld M, Brennicke A,
Schuster W (1994) Physical mapping of the mitochondrial genome of Arabidopsis
thaliana by cosmid and YAC clones. Plant Journal 6:447-455.

Kofer, W., Glimelius, K., and Bonnet H. T. (1991). Modifications of mitochondrial DNA
cause changes in floral development in homeotic-like mutants of tobacco. Plant
Cell 3: 759-769.

Kohler RH, Horn R, Loss] A, Zetche K (1991) Cytoplasmic male sterility in sunflower is
correlated with the transcription of a new open reading flame with the atpA gene.
Mo] Gen Genet 227:369-376.

Koltunow AM, Truettner J, Cox KH, Wallroth M and Goldberg RB (1990) Different
temporal and spatial gene expression patterns occur during anther development.
Plant Cell 2:1201-1224.

Kowalczykowski SC, Eggleston AK (1994) Homologous pairing and DNA strand
exchange proteins. Ann Rev Biochern 63:991-1043.

Kowalczykowski SC, Dixon DA, Eggleston AA, Lauder SD, Rehrauer WM (1994)
Biochemistry of homologous recombination in E. coli. Microbiol Rev 58:401-
465.

Kubo T, Satoh Y, Muro T, Kinoshita T, Mikami T (1995) Physical and gene organization
of mitochondrial DNA fi'om the fertile cytoplasm of sugarbeet (Beta vulgaris L.).
Curr Genet 28: 235-241.

L’Homme et a] (1997) Brassica nap cytoplasmic male sterility is associated with
expression of a mtDNA region containing a chimeric gene similar to pol CMS-
associated 011224 gene. Curr Genet 31:325-335.

Lalanne E, Mathieu C, Vedel F, De Paepe R (1998) Tissue-specific expression of genes
encoding isoforms of the mitochondria] ATPase beta subunit in Nicotiana
sylvestris. Plant Mol Biol 38: 885-888.

Lalanne E, Mathieu C, Vedel F, DePaepe R (1998b) Three genes for the atp2 family fiom
diploid tobacco Nicotiana sylvestris: nsatp2.2.1 encoding the mitochondrial
ATPase [3 two isoform and two related pseudogenes nastp 2.2.2 and nsat 2.2.3.
Plant Physiol 118:331.

200

 

Lam E, Pontier D, del P020 0 (1999) Die and let live-programmed cell death in plants.
Curr Opin Plant Sci 2:502-507.

Laser BS, Mohr S, Odenback W, Oettler G, Kuck U (1997) Parental and novel copies of
the 01125 gene in the hybrid crop plant triticale-predominant transcriptional
expression of the matema] gene copy. Curr Genet 32:337-347.

Laser KD, Lestem NR (1972) Anatomy and cytology of rrricrosporogenesis in
cytoplasmic male sterile angiosperms. Bot Rev 38:425-454.

Lauder SD, Kowalczykowski SC (1993) Negative co-dominant inhibition of recA protein
function. Biochemical properties of the RecA], RecA13, and RecA56 proteins
and the effect of the RecA56 protein on the effect of the wild-type recA function
in vitro. J Mol Biol 234: 72-86.

Lauer M, Knudsen C, Newton KJ, Gabay—Laughnan S, Laughnan JR (1990) A partially
deleted mitochondria] cytochrome oxidase gene in the NSC6 abnormal growth
mutant of maize. New Biologist 2:179—186.

Laver HK, Reynolds SJ, Moneger F, Leaver CJ (1991) Mitochondrial genome
organization and expression associated with cytoplasmic male sterility in
sunflower (Helianathus annuus) Plant J l(2):185-193.

Leclerq P (1969) Une sterilite male cytoplasmique chez le toumesol. Ann Arne] Plant
19:99-106.

Lelandais C, Albert B, Gutierres S, Paepe RD, Godelle B, Vedel F, Ceuit P (1998)
Organization and expression of the mitochondrial genome in the Nicotiana
sylvestris CMSII mutant. Genetics 150:873-882.

Leon P, Arroyo A, MacKenzie SA (1998) Nuclear control of plastid and mitochondrial
development in higher plants. Annu Rev Plant Phys Plant Mol Biol 49:453-480.

Levings, C. S. III, (1990). The Texas cytoplasm of maize: cytoplasmic male sterility and
disease susceptibility. Science 250: 942-947.

Levings 111 CS (1993) Thoughts on cytoplasmic male sterility in ems-T maize. Plant Cell
5: 1285-1290.

Lightowlers RN, Chinnery PF, Tumbul] DM, Howell N (1997) Mammalian

mitochondria] genetics: heredity, heteroplasmy and disease. Trends in Genetics
13: 450-454.

201

 

Lilly J W, Havey MJ (2001) Small repetitive DNAs contribute significantly to the
expanded mitochondrial genome of cucumber. Genetics 159:317-328.

Lindahl R, Peterson DR (1991) Lipid aldehyde oxidation as a physiological role for class
3 aldehyde dehydrogenases. Biochem Pharmacol 41:1583-1587.

Linke B, Nothnagel T, Bomer T (2003) Flower development in carrot CMS plants:
mitochondria affect the expression of MADS box genes homologous to GLOBOSA
and DEFICIENS. Plant J 34:27-37.

Liskay RM, Letsou A, Stachelek J L (1987) Homology requirement for efficient gene
conversion between duplicated chromosomal sequences in mammalian cells.
Genetics 115: 161-167.

Liu F, Cui X, Horner HT, Weiner H, Schnable PS (2001) Mitochondrial aldehyde
dehydrogenase activity is required for male fertility in maize. Plant Cell 13:1063-
1078.

Lohmann JU, Weigel D (2002) Building beauty: the genetic control of floral patterning.
Developmental Cell 2:135-142.

Lund AA, Blum PH, Bhattramakki D, Elthon TE (1998) Heat-stress response of maize
mitochondria. Plant Physiol 1 16: 1097-1 1 10.

Lupold DS Caoile AGFS Stern DB (1999)a The maize mitochondrial cox2 gene has five
promoters in two genomic regions including a complex promoter consisting of
seven overlapping units. J. Biol Chem 274(6): 3897-3903.

Lupold DS Caoile AGF S Stern DB (1999)b Genomic context influences the activity of
maize mitochondrial 00x2 promoters. Proc Nat] Acad Sci 96:11670-11675.

Lusetti SL, Cox MM (2002) The bacterial RecA protein and the recombinatriona] DNA
repair of stalled replication forks. Annu Rev Biochem 71:71-100.

McCollum GD (1996) Occurrence of petaloid stamens in wild carrot Daucus carota from
Sweden. Econ Bot 20:362-367.

Mackenzie SA, Bassett MJ (1987) Genetics of fertility restoration in cytoplasmic male
sterile Phaseolus vulgaris L. Theor Appl Genet 74: 642-645.

Mackenzie SA, Pring DR, Basset MJ, Chase CD (1988) Mitochondrial DNA
rearrangement associated with fertility restoration and cytoplasmic reversion to
fertility in cytoplasmic male sterile Phaseolus vulgaris. Proc Nat] Acad Sci USA
85:2714-2717.

202

 

Mackenzie SA, Chase CD (1990) Fertility restoration is associated with loss of a portion
of the mitochondrial genome in cytoplasmic male sterile common bean. Plant Cell
2: 905-912.

Mackenzie, S., He, S. and Lyznik, A. (1994). The elusive plant mitochondrion as a
genetic system. Plant Physiol 105, 775-780.

Marienfeld JR , Newton KJ (1994) The maize NCS2 abnormal growth mutant has a
chimeric nad4-nad 7 mitochondrial gene and is associated with reduced complex I
function. Genetics 138:855-863.

Martinez-Zapater, J. M., Gil P., Cape], J ., and Somerville, C. R. (1992). Mutations at the
Arabidopsis CHM locus promote rearrangements of the mitochondrial genome.
Plant Cell 4: 889-899.

Maxwell DP, Nickels R, McIntosh L (2002) Evidence of mitochondrial involvement in
the transduction of signals required for the induction of genes associated with
pathogen attack and senescence. Plant Journal 29(3) 269-279.

Menassa R, L’Homme Y, Brown GG (1999) Post-transcriptional and developmental
regulation of a CMS-associated mitochondrial gene region by a nuclear restorer
gene. Plant Journal 17: 491-499.

Meyerowitz EM, Bowman J L, Brockman LL, Drewa GN, Jack T, Sakai H, Medrano LJ
(1995) Role of SUPERMAN in maintaining Arabidopsis floral whorl boundaries.
Nature 378: 199-203.

Michalik B (1971) Genetic studies on male sterility in carrot (Daucus carota L.) Plant
Breed Acclimatization Seed Prod 15:445-474.

Moeykens CA, Mackenzie SA, Shoemaker RC (1995) Mitochondrial genome diversity in
soybean: repeats and rearrangements. Plant Mol Biol 29:245-254.

Moller IM (1986) Membrane bound NAD(P)H dehydrogenases in plant mitochondria.
Physiol Plant 67:517-520.

Moneger F, Mandaron P, Niogret MF, Freyssinet G, Mache R (1992) Expression of

mitochondrial genes during microsporogenesis in maize. Plant Physiol 99:396-
400.

Moneger F, Smart CJ, Leaver CJ (1994) Cell-specific regulation of gene expression in
mitochondria during anther development in sunflower. Plant Cell 6:811-825.

Moore AL, Sideow JN (1991) The regulation and nature of cyanide resistant alternative
oxidase of plant mitochondria. Biochim Biophys Acta 1059: 121-140.

203

 

Mourad, G. S., and White, J. A., (1992). The isolation of apparently homoplasdic
mutants induced by a nuclear recessive gene in Arabidopsis thaliana. Theor Appl
Genet 84, 906-914.

Mulligan RM, Leon P, Walbot V (1991) Transcriptional and post-transciptional
regulation of maize mitochondrial gene expression. Mol Cell Biol 11: 533-543.

Murai K, Takumi S, Koga H, Ogihara Y (2002) Pistillody, homeotic transformation of
stamens into pistil-like structures caused by nuclear cytoplasm interaction in
wheat. Plant J 29: 169-181.

Murashige T, Skoog F (1962) A revised medium for rapid growth and bioassays with
' tobacco tissue cultures. Physiol Plant 15:473-479.

Nakajima Y, Yamamoto T, Oeda K (1997) Genetic variation of mitochondria] and
nuclear genomes in carrots revealed random amplified polymorphic DNA
(RAPD). Euphytica 95:259-267.

Nakajima Y, Yamamoto T, Muranaka T, Oeda K (2001) A novel orfB-related gene od
carrot mitochondrial genomes that is associated with homeotic cytoplasmic male
sterility (CMS). Plant Mol Biol 46:99-107.

Newton KJ, Coe EHJ (1986) Mitochondrial DNA changes in abnormal growth (non-
chromosomal stripe) mutants of maize. Proc Natl Acad Sci USA 83:7363-7366.

Newton KJ, Knudsen C, Gabay-Laughnan S, Laughnan JR (1990) An abnormal growth
mutant of maize has a defective mitochondrial cytochrome oxidase gene. Plant
Cell 2:107-113.

Newton KJ (1995) Abberrant growth phenotypes associated with mitochondrial genome
rearrangements in higher plants. Levings CS and Vasil IK “The Molecular
Biology of Plant Mitochondria. DordrechtzKluwer Academic, pp 585-596.

Newton KJ, Mariano JM, Gibson, CM, Kuzmin E, Gabay-Laughnan S (1996)
Involvement of the S2 episomal sequences in the generation of NCS4 deletion
mutation in the maize mitochondria. Develop. Genetics 19:277-286.

Nivison HT, Hanson MR (1989) Identification of a mitochondrial protein associated with
cytoplasmic male sterility in petunia. Plant Cell 1:1121-1130.

Nivison HT, Sutton CA, Wilson RK, Hanson MR (1994) Sequencing, processing and

localization of the petunia CMS-associated mitochondrial protein. Plant J 5:613-
623.

204

 

 

Nothnagel, T., Straka, P. and Linke, B. (2000) Male sterility in populations of Daucus
and the development of alloplasmic male-sterile carrot lines. Plant Breeding
119:145-152.

Oda K, Yamamoto K, Ohta H, Nakamura Y, Takemura M (1992) Gene organization
deduced from the complete sequence of liverwort Marcham'a polymorphia

mitochondrial DNA: a primitive form of the plant mitochondrial genome. J Mol
Biol 223:1-7.

Okamuro JK, den Boer BGW, Jofirku KB (1993) Regulation of Arabidopsis flower
development. Plant Cell 5:1 183-1193.

Oldenberg D, Bendich AJ (1998) The structure of mitochondrial DNA fiom the liverwort
Marchantia polymorpha. J Mol Bi01276: 745-758.

Palmer JD (1982) Physical and gene mapping of chloroplast DNA fiom Atriplex
triangularis and cucumis sativa. Nucleic Acids Res 10:1593-1605.

Palmer JD, Shields CR (1984) Tripartite structure of the Brassica campestris
mitochondrial genome. Nature 307: 437-439.

Palmer J, Hebron J (1987) Unicircular structure of Brassica hirta mitochondrial genome.
Curr Genet 1 1:565-570.

Paul R, Dalibart R, Lemoine S, Lestienne (2001) Expression of E. coli RecA targeted to
mitochondria of human cells. Mutation Res 486:11-19.

Pla M, Mathieu C, De Paepe R, Chetrit P, Vedel F (1995) Deletion of the last two exons
of the mitochondrial nad 7 gene results in the lack of the NAD7 polypeptide in a
Nicotiana sylvestris CMS mutant. Mol Gen Genet 248:79-88.

Pruitt KD, Hanson MR (1991) Transcription of the petunia mitochondria] CMS-
associated pcf locus in male sterile and fertility restored lines. Mol Gen Genet
227:348-355.

Rasmusson AG, Heiser V, Irgang KD, Brennicke A, Grohmann L (1998) Molecular
characterization of the 76 kDa iron sulfirr protein subunit of potato mitochondrial
complex 1. Plant Cell Physiol 39:373-381.

Redei, G. P. (1973). Extra-chromosomal mutability determined by a nuclear gene locus in
Arabidopsis. Mutat Res 18: 149-162.

Redei, G. P., and Plurad, S. B. (1973). Hereditary structural alterations of plastids
induced by a nuclear gene in Arabidopsis. Protoplasma 77: 361-3 80.

205

 

Roca AI, Cox MM (1997) RecA portein: Structure fimction and role in recombinational
DNA repair. Prog Nucleic Acids Res Mol Biol 56: 129-223.

Roderrnel S (2001) Pathways of plastid to nucleus signaling. Trends Plant Sci 6:471-478.

Rosenberg SM Bonnett HT (1983) Floral organogenesis in Nicotiana tabacum: A

comparision of two cytoplasmic male sterile cultivars with a male fertile cultivar.
Am J Bot 70:266-275.

Rottrnan WH Breas T, Hodge TP, Lonsdale DM (1987) A mitochondrial gene is lost via
homologous recombination during reversion of CMS-T maize to fertility. EMBO
J 6: 1541-1546.

Roussel, D. L., Thompson D. L. Pallardy, S. G., Miles D., and Newton K. J. (1991).
Chloroplast structure and function is altered in NCSZ maize mitochondria]
mutant. Plant Physiol 96: 232-238.

Sabar M, De Paepe R, de Kouchkovsky Y (2000) Complex I impairment, respiratory
compensations and photosynthetic decrease in nuclear and mitochondrial male 3
sterile mutants of Nicotiana slyvestris. Plant Physiol 124:1239-1250.

Sakamoto W, Kondo H, Murata M, Motoyoshi F (1996) Altered mitochondrial gene
expression in a maternal distorted leaf mutant of Arabidopsis induced by
chloroplast mutator. Plant Cell 8: 1377-1390.

Sambrook J, Fritsch E, Maniatis T (1989) Molecular Cloning: A Laboratory Manual.
Cold Spring Habor Press, Cold Spring Harbor NY.

Sancar A, Stachelek C, Konigsberg W, Rupp WD (1989) Sequences of the recA gene and
protein. Proc Nat Acad Sci USA 77:26] 1-2615.

Sand, S. A. and Christoff, G. T. (1973). Cytoplasmic-chromosomal interactions and
altered differentiation in tobacco. J Hered 64: 24-30.

Sarria R, Lyznik A, Vallejos CE, Mackenzie SA (1998) A cytoplasmic male sterility-
associated rrritochondrial peptide in common bean is post-translationally
regulated. Plant Cell 10:1217-1228.

Scheike, R, Gerold, E., Brennicke, A., Mehring-Lemper, M., and Wricke, G. (1992).
Unique patterns of of mitochondrial genes, transcripts and proteins in different
male sterile lines of Daucus carrota. Theor Appl Genet 83:419-427.

Schnable PS, Wise RP (1998) The molecular basis of cytoplasmic male sterility and
fertility restoration. Trends Plant Sci 3:175-180.

206

 

Schon EA, ( 2000) Mitochondrial genetics and diseases. Trends Biochem. Sci. 25:555-
560.

Scissum-Gunn KD, Gandhi M, Backert S, Nielson BL (1998) Separation of different
conformations of plant mitochondrial DNA molecules by field inversion gel
electrophoresis. Plant Mol Biol Report 16:219-229.

Sears, BB. (1998) Replication, recombination, and repair in the chloroplast genetic
system of Chlamydomonas. In: Molecular Biology of Chloroplasts and
Mitochondria in Chlamydomonas. Kluwer Academic Publ. Advances in
Photosynthesis Series (J .D. Rochaix, M. Goldschmidt-Clermont, and S. Merchant,
Eds), Chapter 7, pgs. 115-138.

Senda M, Onondera Y, Mikami T (1998) Recombination events across the atpA-
associated repeated sequences in the mitochondrial genomes of beets. Theor App
Genet 96:964-968.

Shen P, Huang HV (1986) Homologous recombination in Escherichia coli: dependence
on substrate homology. Genetics 112: 441-457.

Shikanai T, Kaneko H, Nakata S, Harada K, Watanabe K (1998) Mitochondrial genome
structure of a cytoplasmic hybrid between tomato and wild potato. Plant Cell
Repl7: 832-836.

Schuster , Brennike A (1987) Plastid, nuclear and reverse transcriptse sequences in the
mitochondrial genome of Oenthera: is genetic information transferred between
organelles via RNA? EMBO J 6: 2857-2863.

Sideow JN, Umbach AL (1995) Plant mitochondrial electron transfer and molecular
biology. Plant Cell 7:821-831.

Small 1, Suffolk R, Leaver CJ (1989) Evolution of plant mitochondrial genomes via
substoichiometric intermediates. Cell 58:69-76.

Smart CJ, Moneger F, Leaver CJ (1994) Cell-specific regulation of gene expression in
mitochondria during anther development in sunflower. Plant Cell 6:811-825.

Song J, Hedgecoth C (1994) A chimeric gene (orf256) is expressed as a protein only in
cytoplasmic male sterile lines of wheat. Plant Mol Biol 26:535-539.

Stassen NY, Logsdon JM Jr, Vora GJ, Offenberg HH, Palmer JD, Zolan ME (1997)
Isolation and characterization of rad51 orthologs from Coprinus cinereus and
Lycopersicon escuIentum, and phylogenetic analysis of eukaryotic recA
homologs. Curr Genet 31:144-57.

207

 

Stern DB, Newton KJ (1985) Mitochondrial gene expression in Cucurbitaceaezconserved
and variable features. Current Genetics 9:395-405.

Stern DB, Palmer JD (1984) Recombination sequences in plant rrritochondrial genomes:
diversity and homologies to known plant mitochondrial genes. Nuc Acids Res
12:6141-6157. '

Struab, J. (1971) Uberweibchen aus sortenkreuzungen bei mannlich sterilen mohren.
Naturwissenschafien 58: 57-58.

Stucloneyer E, Simon D (1986) Anatomy of fertile and male sterile carrot flowers from
different genetic sources. J Am Soc Hort Sci 111:965-968.

Susek RE, Chory J (1992) A tale of two genomes: role of a chloroplast signal in
coordinating nuclear and plastid genome expression. Aust J Plant Physiol 19:
387-399.

Suzuki T, Kawano S, Sakai A, Hirai A, Kuroiwa T (1996) Variability of mitochondrial
subgenomic molecules in the meristematic cells of higher plants. Genes Genet
Syst 71:329-333.

Szkclarazyk M, Oczknowski M, Augustyniak H, Bomer T, Linke B, Michalik B (2000)
Organization and expression of mitochondrial atp9 gene from CMS and fertile
carrots. Theor Appl Genet 100:263-270.

Rousell DL, Thompson DL, Pallardy SG, Miles D, Newton KJ (1991) Chloroplast
structure and function is altered in the NCS2 maize mitochondrial mutants. Plant
Phys 96:232-238.

Tang HV, Pring DR, Shaw LC, Salazr RA, Muza FR, Yan B, Sheere K (1996) Transcript
processing internal to a mitochondrial open reading frame is correlated with
fertility restoration in male—sterile sorghum. Plant J 10:123-133.

Tang Y, Schon EA, Wilichowski E, Vazquez-Memije ME, Davidson E, King MP (2000)
Rearrangements of human mitochondrial DNA (mtDNA): new insights into the
regulation of mtDNA copy number and gene expression. Mol Biol Cell 11:1471-
1485.

Theisen G, Saedler H (2001) F loral quartets. Nature 409:469—471.

Thompson, DJ (1961) Studies on the inheritance of male sterility in the carrot Daucus
carota L. var sativa. Proc Am Soc Hort Sci 78:332-338.

Thyagarajan B, Padua RA, Carnpbel] C (1996) Mammalian mitochondria possess
homologous DNA recombination activity. J Biol Chem 271 :27536-27543.

208

Unseld M, Marienfeld JR, Brandt P, Brennicke A (1997) The mitochondrial genome of
Arabidopsis thaliana contains 57 genes in 366,924 nucleotides. Nature Genetics
15:57-61.

Van Damme JMM, (1983) Gynodioecy in Plantago Ianciolata L. 11 Inheritance of 3 male
sterility types. Heredity 50:253-273.

Von Allrnan J -M et a1 (1991) Transfer of methomyl and HmT-toxin sensitivity fiom T-
cytoplasm maize to tobacco. Mol Gen Genet 229:405-412.

Vergunst AC and Hooykaas, PJJ (1999) Recombination in the plant genome and its
application in biotechnology. Crit Rev Plant Sci 18:1-31

Vitrat MV, De Paepe R, Mathieu C, Chetrit P, Vedel F (1992) Amplification of
substoichiometric recombinant mitochondrial DNA sequences in a nuclear, male

sterile mutant regenerated fiom protoplast culture in Nicotiana sylvestris. Mol
Gen Genet 233: 193-200.

Wallace DC (1999) Mitochondrial diseases in man and mouse. Science 283:1482-1487.

Ward BL Anderson KS, Bendich AJ (1981) The mitochondrial genome is large and
variable in a family of plants. Curr Genet 12:55-67.

Warmke HE, Lee SLJ (1977) Mitochondrial degeneration in Texas cytoplasm male-
sterile corn anthers. J. Hered 68:213-222.

Watts VM, Ingles CJ, Urdea MS, Rutter WJ (1985) Homology requirements for
recombination in Escherichia coli. Proc Nat Acad Sci USA 82: 4768-47 72.

Welch J E, Grimball E (1947) Male sterility in the carrot. Science 196:593.

Wen L, Chase CD (1999) Pleiotropic effects of a nuclear restorer of fertility locus on
mitochondrial transcripts in male fertile and S male sterile maize. Curr Genet 35
(5):521-526.

Wintz H eta] (1995) Expression of the CMS-associated urfS sequence in transgenic
petunia and tobacco. Plant Mol Biol 28:83-92.

Wise RP, Dill CL, Schnable PS (1996) Mutator induced mutations of the rfl nuclear
fertility restorer of T-cytoplasm maize alter the accumulation of T-ur173
mitochondrial transcripts. Genetics 143:1383-1394.

Wise RP, Bronson CR, Schnable PR (1999) The genetics, pathology and molecular
biology of T-cytoplasm male sterility in maize. Adv Agron 65:79-13].

209

 

Wise RP, Pring DR (2002) Nuclear-mediated mitochondrial gene regulation and male
fertility in higher plants: Light at the end of the tunnel. Proc Natl Acad Sci USA
99: 10240-10242.

Yamamoto K, Takahashi N, Fujitani Y, Yoshikura H, Kobayashi I (1996) Orientation
dependence in homologous recombination. Genetics 143:27-36.

Yamato KT, Newton KJ (1999) Heteroplasmy and homoplasmy for maize mitochondrial
mutants: A rare homoplasmic nad4 deletion mutant plant. J Heredity 90:369-373.

Yesodi V, Izhar S, Gidoni D, Tabib Y, Firon N (1995) Involvement of two different urf-s
related mitochondrial sequences in the molecular evolution of the CMS-specific
S-pcf locus in petunia Mol Gen Genet 248:540-546.

Young EG, Hanson MR (1987) A firsed mitochondrial gene associated with cytoplasmic
male sterility is developmentally regulated. Cell 50:41-49.

Yui R, Iketani S, Mikani T, Kubo T (2003) Antisense inhibition of mitochondrial
pyruvate dehydrogenase Ela subunit in anther tapetum causes male sterility.
Plant J 34:57-66.

Zubko MK, Zubko EI, Patskovsky Yu-V, Khvedynich OA, Fisahn J, Gleba YY,
Schneider O (1996) Novel homeotic CMS patterns generated in Nicotiana via
cybridization with Hyoscyamus and Scopolia. J Exp Bot 47:1101-1110.

Zubko MK, Zubko E1, Ruban AV, Adler K, Mock HP, Misera S, Gleba YY, Grimm B
(2001) Extensive developmental and metabolic alterations in cybrids Nicotiana

tabacum + Hyoscyamus niger are caused by complex nucelo-cytoplarnsic
incompatibility. Plant J 25:627-639.

210