Scheffler Immo E. Mitochondria

Подождите немного. Документ загружается.

the phenomenon for the present discussion has been the well - documented

involvement of mitochondria or mitochondrial functions.

It has become useful to subdivide the process of apoptosis into several

phases or stages. The ﬁ nal phase encompasses a number of characteristic

changes such as cell shrinkage, chromatin degradation and nuclear fragmenta-

tion, and loss of plasma membrane integrity. It appears that these changes are

essential for rapid recognition of the dying cell by its neighbors, which phago-

cytose the remnants and efﬁ ciently eliminate any debris that would otherwise

cause an inﬂ ammatory response by the immune system. As a consequence,

“ dead ” cells from apoptosis are not readily seen in vivo in intact tissue, and

the description of this terminal phase rests largely on observations in tissue

culture.

The ﬁ rst phase includes the description of the various stimuli or signals

that induce apoptosis in different cells. The stimulus may be an external one,

received by a cell surface receptor, or an internal one, as the result of the action

of a drug, toxin, or radiation damage. Of course, drugs, toxins, and radiation

are also ultimately external inﬂ uences, but the difference between these two

mechanisms becomes apparent during the following phase: the signal trans-

duction, which leads to changes in metabolic state, induction of various enzymes

(proteases, nucleases), and the morphological changes preceding cell death.

Needless to say, these two signaling pathways may converge at some level, but

at the same time the above description should not give the impression of a

simple, universal and invariable pathway operating in all cells under all condi-

tions. Deprivation of growth factors (e.g., interleukin - 3 for hematopoietic

progenitor cells) induces apoptosis; that is, unoccupied receptors can stimulate

apoptosis. Failure to stimulate germinal center B cells by antigens causes them

to die. Transforming growth factor β and tumor necrosis factor can trigger

apoptosis in various cell types. Many more examples could be given (247, 250) .

It should be clear that these signal transduction pathways include most, if not

all, of the following known reactions, intermediates, and interactions: tyrosine

kinases, phosphatases, steroid receptors, inositol phosphates, transcriptional

activators and factors, and so on.

There is usually a signiﬁ cant time interval (several hours) between the

delivery of the ﬁ rst stimulus and the observation of morphological changes in

the targeted cells. It is also during this interval that thresholds or checkpoints

are reached beyond which the process becomes irreversible. Many details of

events occurring during this effector phase remain elusive and are the subject

of a very active research.

As mentioned above, the nematode, C. elegans , continues to be a useful

model system, and it was in this model system that the ﬁ rst genes speciﬁ cally

associated with apoptosis were isolated (ced1, ced - 2, ced - 3, . . . , ced - 10, etc.).

Many of these genes were subsequently shown to be required for the disposal

of the dead cell(s), and hence their detailed description will be omitted here.

The Ced - 3 gene, however, was identiﬁ ed to encode a cysteine protease, and

many different homologues have since been found in mammalian cells. The

MITOCHONDRIA AND APOPTOSIS 391

392 MITOCHONDRIAL MUTATIONS AND DISEASE

ﬁ rst was interleukin 1 β converting enzyme (ICE), and hence this group of

aspartate - speciﬁ c cysteine proteases was initially referred to as ICE x , where

x is a distinguishing subscript. More recently, this family of proteases has been

referred to as caspases. In a simpliﬁ ed view, apoptosis is accompanied by

extensive proteolytic cleavage, and there is in fact a cascade of such cleavages,

since these ICE enzymes exist as inactive precursors (pro - caspase) that must

be cleaved after a speciﬁ c aspartate residue before they can be assembled into

heterotetramers and active proteases. The other targets of these proteases are

still largely unknown, although it was intriguing that one of these proteases

(CPP32) can cleave the poly - (ADP ribose) polymerase (PARP), an enzyme

involved in DNA repair. However, PARP knockout mice develop normally.

The ICE proteases are constitutively expressed in many cells, and their activa-

tion is a post - translational event, although some additional transcriptional

activation following the death stimulus may also occur. It is apparent, there-

fore, that there must be an upstream signal for post - translational activation

which remains to be more clearly deﬁ ned in mammalian cells. Certain broad -

range protease inhibitors can prevent apoptosis in response to common stimuli,

but other highly speciﬁ c and restricted inhibitors may interfere with the signal-

ing pathway initiated from only a limited number or a unique inducer. Apop-

tosis induction may have “ common ” as well as “ private ” pathways.

The ced - 9 mutation in nematodes led to the characterization of the corre-

sponding nematode gene and the potential functional identiﬁ cation of the

human homologue Bcl - 2. The Bcl - 2 protein was localized in the outer mito-

chondrial membrane (although see below), and this ﬁ nding was one of the ﬁ rst

links to mitochondria. The mammalian Bcl - 2 gene was initially implicated in

tumorigenesis because it was found to be the location of a breakpoint (18q21)

in a reciprocal translocation found persistently in human B - cell tumors (251 –

254) . Bcl - 2 protein became overexpressed under the inﬂ uence of an immuno-

globin promoter. A whole family of bcl - 2 - like proteins (Bcl - 2, Bcl - x

, Mcl - 1,

Bﬂ - 1, A1, . . . ) were subsequently deﬁ ned and characterized as inhibitors of

apoptosis by transfection studies and overexpression in various cells (254) .

Other family members have also been localized in the outer mitochondrial

membrane (255) . In addition, factors promoting apoptosis have also been

found: Bax, Bcl - x

, Bad, Bak, Bik, . . . and so on (see below).

Before discussing Bcl - 2 and mitochondria in more detail, some other

approaches related to the study of apoptosis and mitochondria must be intro-

duced. Active, healthy mitochondria exhibit a transmembrane potential ( ΔΨ

)

across the inner mitochondrial membrane which is of fundamental importance

in the discussion of the mechanism of oxidative phosphorylation (see Chapter

5 ). A pH and electrochemical gradient across the inner membrane is set up as

the result of electron transport through the electron transport chain, with

excess positive charge on the outside — that is, in the intermembrane space.

The discovery was made some years ago that certain lipophilic dyes such as

rhodamine 123, 3,3 ′ - dihexyloxacarbocyanine iodide (DiOC6(3)), 5,5 ′ ,6,6 ′ -

tetrachloro - 1,1 ′ ,3,3 ′ - tetraethylbenzimidazolcarbocyanine iodide (JC - 1), and

chloromethyl - X - rosamine (CMXRos) (see Chapter 5 for a more detailed

description) can accumulate in the matrix. Under controlled conditions the

ﬂ uorescence intensity can be correlated with the magnitude of the membrane

potential ( ΔΨ

), and these dyes have therefore been used as convenient and

powerful indicators of mitochondrial ΔΨ

. When such dyes were applied to

cells undergoing apoptosis, a fall in ΔΨ

was observed as one of the relatively

early events in many different cell types, prior to DNA fragmentation. In fact,

the collapse of the transmembrane potential was correlated in time with the

point of no return — that is, the point at which apoptosis could no longer be

reversed by withdrawal of the stimulus (256) . The breakdown of the ΔΨ

can

be shown to involve the so - called mitochondrial “ permeability transition ”

(PT) (257) . Brieﬂ y, it is believed to be the result of the opening of a PT pore,

which can be measured electrophysiologically by the opening of a “ megachan-

nel, ” large enough to allow solutes of up to 1500 Da to pass through along a

concentration gradient. This PT pore is hypothesized to be regulatory with

diverse functions such as voltage sensing, thiol sensing, redox sensing, pH

sensing, sensing of adenine nucleotide levels, and regulation of calcium homeo-

stasis. Normally, only a brief opening or transient assembly of a small number

of such pores may be required, but the massive opening of many such pores

will collapse the ΔΨ

, stop oxidative phosphorylation, arrest the import of

proteins into mitochondria, and induce leakage of cytochrome c out into the

cytosol (see below). Thus, a long - term, massive opening of the PT pores can

be expected to do permanent damage to mitochondria, and the PT was pos-

tulated to be a central, rate - limiting event in the apoptotic cascade (255 – 257) .

In simple terms, it made sense to think that turning off the power supply was

a logical ﬁ rst step in programmed cell death. A study of respiration deﬁ cient

mammalian cell mutants has shown that such cells die within a day when they

are switched from a high glucose medium to medium containing galactose. In

the absence of oxidative phosphorylation, cells are sustained by a high rate of

glycolysis (30) which is not possible with galactose (see Section 1.2 ).

On the other hand, several experiments in cell free systems suggested a

more complex picture. Isolated nuclei can be induced to undergo apoptosis

when exposed to cytoplasmic extracts from cells in the process of apoptosis.

Controversy existed about whether such extracts can be free of mitochondria

or whether enrichment with mitochondria accelerates the process. Subse-

quently, the presence of “ damaged ” mitochondria was shown to be required,

and it was hypothesized that such mitochondria release an apoptogenic protein

(apoptosis inducing factor, AIF). This protein has been characterized as an

∼ 50 - kDa intermembrane protein, with protease or protease activating activity

but no nuclease activity, which can induce apoptosis in isolated nuclei more

efﬁ ciently than ceramide, another mitochondrial constituent apparently

released by such organelles (255) .

A major breakthrough from in vitro studies of characteristic events associ-

ated with apoptosis (DNA fragmentation in isolated nuclei, activation of cas-

pases) was the discovery of the crucial role of cytochrome c, initially based on

MITOCHONDRIA AND APOPTOSIS 393

394 MITOCHONDRIAL MUTATIONS AND DISEASE

the fractionation of the cytoplasm and further demonstrated by immuno -

depletion and reconstitution experiments (258, 259) . These and subsequent

studies have suggested that the cytochrome c is involved in triggering the

cleavage and activation of DEVD - speciﬁ c caspases. In short, cytochrome c is

now known to interact with Apaf - 1, leading to the formation of the multimeric

“ apoptosome ” in the presence of ATP or dATP (260, 261) . The apoptosome

activates the initiator caspase, caspase 9, which subsequently cleaves and acti-

vates the effector caspases 3 and 7. Many more studies have since deﬁ ned

cytochrome c - dependent activation of apoptosis and cytochrome c - indepen-

dent pathways, the latter requiring proteins such as AIF and endonuclease G.

These proteins and others (Smac, Omi, adenylate kinase - 2) are found in the

intermembrane space of mitochondria.

A central question that was raised immediately following these observa-

tions was, How do these proteins get released into the cytosol? Much potential

confusion was generated by attempts to link the mechanism of cytochrome c

release to the breakdown of the ΔΨ and the permeability transition (PT). The

release of cytochrome c requires the permeabilization of the outer membrane,

while the PT occurs in the inner membrane. A very clear formulation of the

problem and a partial resolution of this problem can be found in the authorita-

tive recent review by Green and Kroemer (262) . The authors clearly distin-

guish between mitochondrial outer - membrane permeabilization (MOMP) and

the PT causing a breakdown of ΔΨ . Which one comes ﬁ rst? It depends.

In one scenario the PT is associated with an equilibration/ﬂ ux of ions and

water, causing the mitochondria to swell and leading eventually to the rupture

of the outer membrane. The PT - associated MOMP could be prevented by

speciﬁ c inhibitors such as bongkrekic acid and cyclosporin A, which target the

adenine nucleotide transporter (ANT) and the protein cyclophilin D (CyP - D,

a peptidyl - prolyl - cis - trans isomerase), respectively. Thus, ANT and cyclophilin

D became postulated constituents of the PT pore, although this pore has has

not been isolated and well - hdeﬁ ned even today. In current thinking the PT

pore is opened by Ca

overloads, and prolonged opening causes cell death by

necrosis (263) . CyP - D - knockout mice develop normally and show no protec-

tion against a range of apoptotic stimuli, suggesting that the PT pore does not

play a role in most forms of apoptosis (263) . There is also recent evidence for

a PT pore forming in the absence of ANT, and a participation of the porin

(VDAC) channel in the outer membrane has been considered (262) .

The more dominant mechanism for MOMP nowadays excludes the PT and

events at the inner membrane, and a breakdown of ΔΨ probably is a later

event following cytochrome c release and caspase activation (Figure 7.2 ). In

this mechanism the anti - apoptotic Bcl - 2 proteins (Bcl - 2, Bcl - x

, Bcl - W. . . . ) and

the pro - apoptotic Bax family members (Bax, Bak, Bad, Bid . . . ) play a very

prominent role. Two problems must be addressed: (1) What constitutes the

channel in the outer membrane through which proteins of considerable size

( > 50 kDa) can escape from the IMS to the cytosol? (2) What mechanisms

induce the formation of such channels that are obviously not present in normal

mitochondria? The literature on these questions is enormous, and it is a chal-

lenge to review for the uninitiated. Controversies and/or uncertainties still

abound. The interested reader should start with some recent reviews (260, 261)

and plunge into the literature. Based on in vitro studies with isolated outer

membranes or vesicles made from mitochondrial lipids, it is now believed that

oligomeric assemblies of Bax or Bak, in combination with active Bid (tBid),

form the channel (264, 265) . The possibility of a participation of porin (VDAC)

has also been considered. However, before the monomeric, cytosolic Bax or

Bad can insert into the outer membrane and form multimers, it must be acti-

vated by interaction with activated Bid (tBid, formed by caspase - 8 - mediated

cleavage of Bid) and by other mechanisms such as phosphorylation and protein

modiﬁ cations by reactions with ROS (266) . The anti - apoptotic factors interact

with Bax or Bad to prevent activation, membrane insertion, and aggregation

Figure 7.2 (A) Schematic view of signaling in apoptosis. In this view the mitochon-

drial permeability transition takes a central position on which multiple early pathways

converge. (B) (Left) In this view the outer membrane permeabilization occurs before

the collapse of the membrane potential (see reference 262 for a recent review). (Right)

A less - deﬁ ned mechanism leading to OMP and a permeability transition involving the

inner membrane.

Mitochondrial

permeability

transition

Multiple inducers of apoptosis

Collapse of

ΔΨ

Activation of

caspase

cascade

Chromatin condensation

DNA fragmentation

Nuclear fragmentation

Selective proteolysis

Cell shrinkage

MITOCHONDRIA AND APOPTOSIS 395

396 MITOCHONDRIAL MUTATIONS AND DISEASE

into channels. Bax has a C - terminal membrane - spanning region that is folded

into a hydrophobic groove in the monomer. Upon activation, this C - terminual

region becomes exposed for insertion into the outer - membrane. A truncation

of the Bcl - 2 protein deleting its C - terminal transmembrane (anchor) sequence

makes the protein incapable of suppressing apoptosis induced by several

agents. Furthermore, this anchor sequence can be replaced by the correspond-

ing anchor sequence from the yeast outer - membrane protein Mas70 (now

Tom . . . ), and Bcl - 2 targeting and function are restored. An important experi-

ment has shown that Bcl - 2 can block apoptosis even in cells that lack mito-

chondrial DNA and are unable to carry out oxidative phosphorylation (267) .

This is not to say, however, that they lack a membrane potential, since such

cells and mutant cells completely defective in mitochondrial protein synthesis

(31, 33) are perfectly capable of mitochondrial protein import.

High - resolution structures for several of these proteins are available (see

reference 261 for a summary). A similarity to the pore - forming domains of

diphtheria toxin and the bacterial colicins became apparent, supporting the

conclusions from the studies performed with liposomes. The anti - apoptotic

members have four conserved BH domains, named after the deﬁ ning region

in Bcl - 2 — that is, Bcl - 2 homology domains (BH1, BH2, BH3, BH4); Bax/Bad

family members lack the BH4 domain, and Bid is a BH3 - only protein. The

Bax/Bak-mediated

outer membrane

permeabilization

(OMP)

Omi

Smac

Cyt c

APOPTOSOME

APAF-1

procaspase-9

active executioner caspases

Apoptosis-inducing

signals

BH3-only proteins

Bax

Bak

BH-123 proteins

anti-apoptotic

Bcl-2 proteins

mitochondrial

involvement

in necrosis

Figure 7.2 (Continued)

relative ratios and the interaction of these proteins determine the susceptibil-

ity to MOMP. Many studies with transfected cells and transgenic mice support

such a view. Bid (and other BH3 - only members) serve as “ upstream sentinels ”

ready to receive and interpret death and survival signals (260) .

The examination of mice in which either the Bcl - 2, the Bcl - x, or the Bax

genes were knocked out has yielded a complex picture. Bcl - 2 – / – mice are born

as viable pups but die at a few weeks of age. Bcl - x – / – mice die at ∼ day 13 of

embryonic development, due to massive cell death in the central nervous

system among other tissues. Bax – / – mice appear healthy, with lymphoid

hyperplasia and sterility of males due to lack of sperm cell production. Overall,

such studies support a lineage - speciﬁ c role for these genes, but details remain

to be clariﬁ ed, and for the moment no signiﬁ cant relationship to mitochondrial

function or activity can be seen from such studies in intact mice. It may have

come as a surprise that live, normal - looking mice were born.

A new development that remains to be fully explored promises to shed new

light on the mechanism of IMS protein release in the process of apoptosis. Up

till now, Bax and Bak proteins were thought to be involved exclusively in

apoptosis, by relocating from the cytosol and from a position on the ER to

induce MOMP. A poineering new study by Karbowski et al. (268) has impli-

cated these proteins in the normal and highly dynamic process of mitochon-

drial fusions. These results and some related investigations by Delivani et al.

(269) indicate that Bax and Bad coordinate both fusion and ﬁ ssion of mito-

chondria in healthy cells. Karbowski and colleagues found that Bax/Bak

double knockout mice have fragmented but still functional mitochondria. In

contrast, Delivani et al. showed that overexpression of the anti - apoptotic

protein Bcl - xL (or the nematode homologue CED - 9) in mammalian cells

promoted mitochondrial fusion. A working model that will undoubtedly

become more complex in the future has Bax/Bak act on the factor Mnf2 which

is required for fusion. The absence of Bax/Bak in DKO mice can be compen-

sated for by overexpression of Mnf2. At the same time, Bax/Bak are postulated

to inhibit a factor that is involved in ﬁ ssion, a process that proceeds unchecked

in the absence of these pro - apoptotic factors.

It may be too early to dismiss the idea that Bax monomers form aggregates

in the outer membrane that constitute the channel for the escape of cyto-

chrome c and other IMS proteins. The channels demonstrated in liposomes

may be in vitro artifacts, and so far they have been shown to allow small ions

and ﬂ uorescent dyes to pass through, but the passage of a 50 - kDa protein may

be a greater challenge. Since we do not understand the process of mitochon-

drial fusion and ﬁ ssion in much detail (see Chapter 3 ), we cannot readily

explain the release of the IMS proteins during apoptosis by invoking mecha-

nisms active in fusion. However, it should be noted that fusion, ﬁ ssion, and

MOMP all involve major perturbations of the outer membrane (and the inner

membrane). Under conditions where normal fusion and ﬁ ssion occur, the

process may be sufﬁ ciently orderly/controlled to prevent the loss of any pro-

teins from the IMS, but excess recruitment of pro - apoptotic Bax may lead to

MITOCHONDRIA AND APOPTOSIS 397

398 MITOCHONDRIAL MUTATIONS AND DISEASE

large enough perturbations to allow leakage of such proteins. The process

could be autocatalytic as downstream caspases become activated.

7.6 FUNGAL SENESCENCE

Many fungi appear to be able to divide mitotically for an almost unlimited

period of time, growing into large clones whose size may be limited only by

physical parameters. In the wild, Armillaria bulbosa colonies in excess of 0.5 -

km diameter have been found. On the other hand, some strains in the labora-

tory exhibit the phenomenon of senescence, that is, they cannot be serially

transferred indeﬁ nitely in culture. In a few prominent cases the phenomenon

has been investigated genetically. Since mitochondrial genomes were found to

play a large role in the process, it has been tempting to draw parallels with

aging in humans; but as our understanding has advanced, the resemblance is

only at the physiological level: the inability of cells to undergo further mitotic

divisions and a generalized cellular decay at the end. The best - studied exam-

ples have been expertly reviewed by Grifﬁ ths (270, 271) , and some salient

features are appropriate for the present chapter. Although nuclear genes can

play a role, a common theme is that senescence in fungi is associated with

cytoplasmic inheritance. A triggering event appears to be an alteration in the

mitochondrial genome which eventually leads to a severe mitochondrial

degeneration and dysfunction and hence senescence.

There has been a renewed interest in using yeast as a model system for

life - span research, in connection with calorie restriction and the role of the

gene SIR2 (encoding a NAD - dependent deacetylase (272) . A full discussion

is beyond the scope of this book. It should be noted that there are two life

spans to be considered for yeast: (a) a replicative life span limiting the number

of times a given mother cell can produce buds and (b) a chronological life

span, which is a measure of the capacity of stationary cultures to remain viable

over time. Piper (273) discusses the relevance of studies with yeast to our

understanding of aging in more complex organisms.

The existence of mitochondrial plasmids in many organisms, and fungi in

particular, has already been discussed in Section 4.1.6 . Even at relatively high

copy numbers, they may act as relatively innocuous molecular parasites,

expressing functions primarily devoted to their own autonomous replication,

but they can also contribute to the phenomena of senescence and hypoviru-

lence. In Neurospora crassa and Neurospora intermedia , plasmids isolated

from wild strains have been shown to be responsible for senescence. The kalilo

plasmid ( ∼ 9 kb) isolated originally in Hawaii causes N. intermedia to stop

growing, and in the terminal stages, abnormalities in cytochrome spectra and

content, deﬁ ciencies in large ribosomal subunits, and loss of cyanide - sensitive

respiration are observed. It is now clear that the plasmid can insert at one of

several locations in the mitochondrial DNA. There may be more than one

insertion in the process of senescence, and an mtDNA population may evolve

until at death a predominant insert is observed. A similar picture is seen with

the maranhar plasmid isolated in an N. crassa strain in India, as well as with

other plasmids found in a large sample of natural isolates (274) . While there

appears to be no doubt that the kalilo and maranhar plasmids are the respon-

sible agents, other plasmids in Neurospora are neutral. The challenge therefore

is to understand why some plasmids can and do insert themselves in the

mtDNA. According to Grifﬁ ths, the “ initial insertion event is probably best

viewed as a mistake. ” The mistake, however, is irreversible and ultimately

lethal. Does it have any biological signiﬁ cance, especially outside of the

laboratory?

Historically, senescence in the fungus Podospora anserina was one of the

ﬁ rst to attract attention, since all natural isolates died, but they each exhibited

distinct and reproducible life spans. Cytoplasmic inheritance was also observed,

but heterologous, extragenomic mitochondrial plasmids cannot be detected.

Instead, as senescence progresses, circular DNA molecules derived from the

mtDNA are observed. The triggering event is unclear, but the net effect is that

one of several different segments of mtDNA (labeled α , β , γ , etc.) becomes

excised, associated as head - to - tail multimers, and circularized to form a series

of senDNAs. At the same time, intact mtDNA molecules disappear. The most

common sequence is found in a - senDNA, which consists of monomers identi-

cal in sequence to the ﬁ rst intron of the CO1 gene (cytochrome oxidase). It

may be signiﬁ cant that this group II intron encodes a protein with homology

to reverse transcriptase, but generalizations of the mechanisms for senescence

in Podospora anserina must include an explanation for the β - , γ - , δ - , etc.

senDNAs, which are not restricted to introns and do not encode obvious func-

tions (270) . Curiously, a senescing culture can be rejuvenated by treatment

with ethidium bromide, as if the intercalator were capable of selecting for

wild - type mtDNA molecules, since the sen plasmid present at the time of

addition is eliminated. When ethidium bromide is removed again, senescence

recurs in association with the appearance of a different senDNA type.

Numerous studies have sought to identify strains with different life spans,

and to identify nuclear or mitochondrial mutations responsible for this varia-

tion. Details can be found in the reviews by Grifﬁ ths (270, 271) and by Bertrand

(275) . For example, a short - lived strain dies only on solid medium, but not in

liquid culture. After repeated transfers in liquid cultures, immortal cultures

could be obtained. These were female sterile and incapable of transmitting the

immortality as males. Mitochondrial mutations were likely responsible for the

altered phenotype. A publication on Podospora anserina (276) can serve for

an interested reader as an entry point into the recent literature.

In conclusion, senescence in Podospora anserina can be linked to mitochon-

drial DNA instability and mitochondrial energy metabolism; but here also the

triggering mechanism, the transition from the nonsenescent (juvenile) state to

the senescent state, remains to be elucidated. As in the case of the Neurospora

plasmids, the true biological signiﬁ cance of the phenomenon of this induced

aging remains another puzzle.

FUNGAL SENESCENCE 399

400 MITOCHONDRIAL MUTATIONS AND DISEASE

7.7 CYTOPLASMIC MALE STERILITY IN PLANTS

Cytoplasmic male sterility (CMS) has been presented brieﬂ y in a different

context in Chapter 4 . The integration of speciﬁ c plasmids into mtDNA can

cause CMS in maize (277) . Another genetic mechanism, independent of the

existence of extragenomic plasmids, will be illustrated here with a select

number of examples. It should be noted that the phenomenon has been

observed in more than 150 plant species. It is widely exploited commercially

as an expedient method to produce pollen - sterile plants (278) . Some authors

have referred to it as a “ molecular arms race between two conﬂ icting genomes

with different modes of inheritance (279, 280) . The involvement of the mito-

chondrial genome in all cases points to a crucial role of mitochondrial

function(s) in the development of the male reproductive system and pollen

formation in particular. The following discussion will give the reader a ﬂ avor

of the kinds of molecular mechanisms being studied. It cannot be exhaustive,

but the reader will ﬁ nd much up - to - date information in an expert review by

Chase (281) .

The exceptionally large mitochondrial genomes in plants have already been

discussed (Chapter 4 ) in terms of their high recombination rate, thought to

include a population of large circular mtDNA molecules, each representing a

portion of the genome. Speciﬁ c “ recombination repeat ” sequences typically

greater than 2 kb are responsible for homologous recombinations capable of

permuting the genome without mutating the coding regions. Occasionally,

illegitimate (?) recombination events involving much shorter repeat sequences

can occur, and they may be responsible for the evolutionary changes in the

plant mt genomes. Even more rarely, such recombination events can create

new chimeric genes responsible for CMS. The search for such new sequence

arrangements in the large plant mtDNA is a formidable task.

L ’ Homme et al. (282) have thrown considerable light on the situation in the

oilseed rape (canola) species Brassica napus , and the results may exemplify

the general nature of the genetic changes associated with CMS (Figure 7.3 ).

The fertile B. napus plant has a cytoplasm (mitochondrial genome) referred

to as cam ; two infertile strains (male sterility) have cytoplasms (mitochondrial

genomes), nap and pol . The cam cytoplasm may have originated from the

diploid B. campestris progenitor of the amphidiploid B. napus . The nap cyto-

plasm makes plants male - sterile only when certain nuclear Rf s (restorer of

fertility) are missing, but the pol cytoplasm makes most B. napus cultivars male

sterile. The whole mtDNA sequence is not yet available, but many segments

containing known genes have been identiﬁ ed and physical maps of restriction

fragments are available. The task was to ﬁ nd differences between the cam, nap ,

and pol mt genomes.

Several lines of evidence have pointed at the vicinity of the atp6 gene as

the region differing in the cam, pol , and nap mtDNA. Detailed restriction

mapping, sequencing, and the analysis of chimeric transcripts found in the

various strains has yielded a picture of the arrangement of sequences in the