Download Evolutionary origin and consequences of uniparental mitochondrial

Human Reproduction, Vol. 15, (Suppl. 2), pp. 102-111, 2000
Evolutionary origin and consequences of uniparental
mitochondrial inheritance
Rolf F.Hoekstra1
Laboratory, of Genetics, Department of Plant Sciences, Wageningen
University, Wageningen, the Netherlands
!
To whom correspondence should be addressed at: Laboratory of Genetics,
Department of Biomolecular Sciences, Wageningen University, Dreijenlaan 2,
NL-6703 HA Wageningen, the Netherlands. E-mail: [email protected]
In the great majority of sexual organisms,
cytoplasmic genomes such as the mitochondrial genome are inherited (almost) exclusively through only one, usually the
maternal, parent. This rule probably
evolved to minimize the potential spread
of selfish cytoplasmic genomic mutations
through a species. Maternal inheritance
creates an asymmetry between the sexes
from which several evolutionary consequences follow. Because natural selection
on mitochondria operates only in females,
mitochondrial mutations may have more
deleterious effects in males than in females.
Strictly uniparental inheritance creates
asexual mitochondrial lineages that are vulnerable to mutation accumulation (Muller's
ratchet). There is evidence that over evolutionary time mitochondrial genomes have
indeed accumulated slightly deleterious
mutations. Mutation accumulation in
animal mitochondrial genomes is probably
slowed down mainly by two processes: a
severe reduction in germline mitochondrial
genome copy number at some point in the
life cycle, enabling more effective elimination of mutations by natural selection, and
occasional recombination between maternal
and paternal mitochondrial genomes following paternal leakage.
102
Key words: evolution/maternal inheritance/
mitochondria/mitochondrial
DNA/mtDNA
mutations
Introduction
Because mitochondria cannot be synthesized
de novo, these organelles must originate from
pre-existing ones and they must be passed on
from one generation to the next. A striking
and almost universal feature of mitochondrial
inheritance is uniparental transmission: in
many eukaryotic organisms the offspring from
a sexual cross receives mitochondria from
only one of the two parents (for a recent
review, see Birky, 1995). In animals and
flowering plants, it is the maternal parent that
transmits the mitochondria. The vast difference in the amount of cytoplasm contributed
by the male and female gamete to the zygote
almost guarantees effective uniparental inheritance, but active elimination of male-derived
mitochondrial DNA from zygotes has been
demonstrated (e.g. Kaneda et al., 1995).
Uniparental mitochondrial inheritance has also
been shown in some species lacking sexual
gamete dimorphism, such as the unicellular
green alga Chlamydomonas reinhardtii, in
which the zygotes are formed by fusions
© European Society of Human Reproduction & Embryology
Uniparental mitochondria! inheritance
between two equally sized gametes that differ
in so-called + and - mating type. The mitochondrial DNA derived from the + mating
type gamete is degraded in the early zygote
(Gillham, 1994).
Uniparental inheritance of mitochondrial
DNA is in sharp contrast to the familiar
biparental inheritance of nuclear DNA. From
an evolutionary perspective these two contrasting inheritance patterns have far-reaching
consequences. Uniparental inheritance results
in asexual mitochondrial DNA lineages, while
biparental inheritance creates sexual lineages
due to meiotic recombination of paternally
and maternally derived DNA.
This review will first discuss ideas about
the functional importance and evolutionary
origin of uniparental mitochondrial inheritance, and then it will concentrate on some of
the consequences of the male-female asymmetry in inheritance and of the general lack of
recombination between animal mitochondrial
genomes.
Functional significance of uniparental
mitochondrial inheritance
Several authors have noted that the mixing of
cytoplasm when gametes fuse ought to produce the potential for a rapid spread of deleterious cytoplasmic elements through a sexual
population (Grun, 1976; Eberhard, 1980;
Cosmides and Tooby, 1981). The argument
assumes a lack of precise control of intracellular replication and subsequent segregation of
the cytoplasmic DNA at gametogenesis. Such
strict control is not known to exist in presentday organisms, and indeed would presumably
be difficult to establish for multi-copy genomes whose replication is not synchronized
with that of the cell. Without control but with
biparental transmission, a mutant germline
mitochondrial genome with a replication
advantage over the wild-type genome would
out-compete the latter and gain transmission
to a majority of the gametes. This segregation
advantage, if great enough, could outweigh
negative effects on the host's fitness and spread
through the population. This is a scenario in
evolution that would lead to maladaptation
instead of adaptation.
Natural selection is thus expected to favour
genetic variants that minimise the possibilities
for spread of such deleterious mitochondrial
mutations. One possible mechanism is to
restrict the role of transmitting mitochondria
to one of the parents.
Evolution of uniparental mitochondrial
inheritance
The above-mentioned argument has formed
the basis for mathematical population genetic
models showing that uniparental mitochondrial inheritance could have evolved to minimize cytoplasmic mixing and hence to prevent
the spread of deleterious mitochondrial
mutations (Hoekstra, 1990; Hastings, 1992;
Hurst and Hamilton, 1992; Law and Hutson,
1992). The models take as a starting point a
population in which mitochondria are inherited
from both parents, and then analyse under
which conditions a mutation enforcing uniparental mitochondrial inheritance will increase in
frequency and eventually become established.
The basic argument is as follows. With
biparental mitochondrial inheritance, a 'selfish' mitochondrial mutation (i.e. a mutation
that is competitively superior in mitochondrial
replication but deleterious to the host organism) can be further transmitted by all descendants who have inherited this mutation. In this
way the mutation can come to be established
in the whole population. In contrast, under
uniparental (maternal) mitochondrial inheritance this mutation can only be transmitted
further by the daughters who received this
mutation from their mother, and again the next
generation only by the daughters of these
daughters. Thus, the mutation will remain
103
R.F.Hoekstra
restricted to a segment of the population
formed by individuals descended along the
particular female line that originates from the
female in which the mutation originally arose.
It can never reach individuals descended from
other females. Moreover, because the mutation
lowers fitness of individual carriers, or hosts,
natural selection will act to remove this segment of the population, leading to extinction
of the mutant, selfish genome.
Uniparental cytoplasmic transmission
and male-female dimorphism
Mechanisms of uniparental mitochondrial
transmission require a fundamental sexual
asymmetry in fusing gametes, because it is
compulsory that one of them (and not the
other) transmits mitochondria to the zygote.
This asymmetry is revealed by male-female
gamete dimorphism among so-called anisogamous species (characterized by very small
male gametes and large female gametes) and
by mating-type dimorphism in isogamous
species (where all gametes are the same size).
Isogamy is seen mainly in some groups of
protists and fungi, and is discussed further
below.
It is possible that sexual dimorphism
evolved for reasons other than the regulation
of uniparental transmission of cytoplasmic
DNA and that mechanisms regulating uniparental transmission were based on pre-existing
sexual asymmetry. However, it is also possible
that sexual dimorphism evolved primarily to
control uniparental genomic transmission.
This possibility has been suggested (Hoekstra,
1987) and further analysed (Hurst and
Hamilton, 1992). The latter authors make the
interesting empirical observation that the taxa
that exhibit a fertilization mechanism in which
the cytoplasm of the mating partners is not
mixed (a group of ciliates among the protists,
and the mushrooms among the fungi) also
exhibit a lack of sexual dimorphic (male104
female) differentiation. (These species are
nevertheless sexual, and there is the advantage
that mating can occur with any other individual, rather than with just the 50% of the
species with the opposite sex.) Hurst and
Hamilton considered this phenomenon to support their hypothesis that the integrity of
organellar inheritance is primary in compelling
the evolution of sexual dimorphism.
However, irrespective of the precise causal
connection between the evolution of sexual
dimorphism and the evolution of uniparental
cytoplasmic inheritance, the asymmetries
involved have important evolutionary consequences.
Uniparental (maternal) mitochondrial
inheritance can be harmful for male
fertility
A genetic element such as the mitochondrial
genome that is transmitted by only one sex is
subjected to natural selection forces different
to those that affect a genetic element, such as
the nuclear genome, which is transmitted by
both sexes. Whether a mitochondrial mutation
is favoured or not by natural selection will
depend only on its effects in females, not on
its effects in males. This is caused by the fact
that the mutation is only transmitted to a
future generation by females, not by males.
Because males are a 'dead end' for mitochondria, mitochondrial effects in males have no
influence on the fate of particular mitochondrial genotypes. Therefore a mitochondrial
mutation with beneficial effects in females but
deleterious effects in males will be favoured
by selection and is expected to spread through
a population. An instructive example of such
a situation is provided by the phenomenon of
male sterility in plants.
Of angiosperm species (the flowering
plants), - 5 % exhibit the condition of gynodioecy, or male sterility. At the phenotypic level,
male sterility manifests as the coexistence
Uniparental mitochondrial inheritance
within the population of (normal) hermaphroditic plants and plants that produce only seeds
but no viable pollen. Much genetic and population genetic research on this trait has been
carried out in natural populations of Plantago
and Thymus, as well as in several agricultural
crops. At first sight, the presence of appreciable frequencies of plants deficient in their
male function is puzzling. These plants can
only reproduce as females, and are expected to
have a lower fitness than their hermaphroditic
conspecifics, which can function both as male
and as female. Selection ought to eliminate
the genotypes that cause male sterility. The
explanation of the conundrum is that, in almost
all cases where the genetics of male sterility
has been investigated, the male sterile phenotype is caused by a mitochondrial mutation;
elimination of the male function does not
affect the transmission prospects of mitochondrial genes (as long as there is no overall
shortage of pollen in the population). Indeed,
a mitochondrial mutation that eliminates the
male function and at the same time somewhat
enhances fitness of the female hosts will
be selectively favoured, because the fate of
mitochondrial genes is affected only by their
female carriers' fitness. On the other hand,
selection acting on nuclear genes that affect
the sex ratio will tend to establish an equal
investment in both sexes. If the frequency of
male steriles in a population rises because
selection favours a mitochondrial gene causing
male sterility, there will be increasing selection
pressure favouring any nuclear mutation that
suppresses the cytoplasmic male sterility and
restores a more equal sex ratio. Such nuclear
restorer genes are indeed found in gynodioecious species. This is a clear example of
genomic conflict: selection at the level of
mitochondria favouring mitochondrial male
sterility in conflict with selection at the individual level favouring nuclear restorers of
male fertility.
The long-term outcome of such genetic
conflict between mitochondrial male sterility
genes and nuclear restorer genes is not easy
to predict: it depends on the precise relations
between the fitness of the different genotypes.
There is indirect evidence that at least in some
cases the nuclear genome ultimately 'wins'
the conflict. It has been observed that malesterile individuals can result from crosses
between related species, while within these
species no male sterility occurs. This can be
interpreted as evidence for complete withinspecies restoration of male fertility, but
involving different mitochondrial mutations
and nuclear restorers in the two species. In
the hybrids, a nuclear restorer from one species
can combine with a mitochondrial malesterility gene from the other species against
which it is ineffective. The complexity of the
evolutionary dynamics of male sterility is
illustrated by the finding that natural populations often can simultaneously harbour several
different mitochondrial male sterility mutations together with the corresponding sets of
nuclear restorer genes.
There are some suggestive findings in
humans that allow an interpretation along
these lines. Leber's hereditary optic neuropathy (LHON) is associated with mitochondrial mutations and appears to affect males
more severely than females. Of individuals
affected by LHON, -85% are male (Wallace,
1992). A few other examples of mitochondrial
mutations that seem to cause more severe
disease effects in males than in females have
been mentioned (Frank and Hurst, 1996).
These authors also point out that reduced
sperm motility and poor male fertility have
been associated with mitochondrial mutations.
Sperm dysfunction could perhaps be caused
by mitochondrial mutations that affect female
fitness only weakly or not at all, so that
such mutations might reach relatively high
frequencies at mutation-selection equilibrium.
105
R.F.Hoekstra
Uniparental transmission produces
asexual mitochondrial lineages
Under exclusive uniparental inheritance, mitochondria remain captured in separate female
lines of descent and are excluded from recombination with mitochondria derived from a
different line of descent. Thus mitochondria
form extremely ancient asexual lineages (the
animal mitochondrial lineage must be > 1
billion years old). An intriguing question is
how non-recombining mitochondria have survived for so long and apparently managed to
escape too severe an accumulation of
mutations. Asexual reproduction is basically
a copying process in which newly copied
genomes cannot contain fewer mutations than
were present in the parental genome, except
in the rare case of back-mutations. When
in an asexual population the best class of
individuals (containing the lowest number of
deleterious mutations) either produces no surviving offspring or only produces offspring
that acquired a new deleterious mutation, the
population suffers an irreversible increase in
the mean number of deleterious mutations per
genome. The previously second-best class of
individuals will eventually undergo the same
fate, and so on. Muller was the first to consider
this mutation accumulation consequence of
asexual reproduction formally (Muller, 1964),
and the process is now known as Muller's
ratchet (Felsenstein, 1974).
Quantitative description of Muller's
ratchet
The process of mutation accumulation in asexual lineages (Muller's ratchet) has been analysed mathematically (Haigh, 1978). If, in a
population, the number of individuals in the
optimal class (containing genomes with the
smallest number of deleterious mutations)
equals n0, then the following relationship will
hold: n0 = N e~u/s, where N is the total
106
population size, U is the expected number of
novel deleterious mutations per genome per
generation, and s is the average selective
disadvantage per mutation. The model (Figure
1) assumes that different mutations in the
same genome interact multiplicatively on
organismal fitness. Clearly, there is only a real
danger that the optimal class will get lost (and
the ratchet will click) if « 0 is small. The
formula suggests that mutations with very
weakly deleterious effects will have the greatest chance of accumulating. When s is small,
g-u/s c a n b e v e r y s m a n anc j UQ W[\I stiii be small
even in large populations. Strongly deleterious
mutations will not accumulate because they
are likely to be eliminated quickly by natural
selection. The slighter the deleterious nature
of asexual lineage mutations, the more likely
it is that such mutations become fixed and
accumulate.
Evidence of mutation accumulation in
mitochondrial genomes
Several studies, on both mammalian and
Drosophila mitochondrial genomes, have
reported an excess of non-synonymous nucleotide polymorphisms (i.e. base changes
resulting in amino acid changes) between
related species compared with the number
expected under the assumption of strictly neutral mutations (Ballard and Kreitman, 1994;
Nachman et al, 1996; Rand and Kann, 1996).
In other words, mitochondrial genomes seem
to contain more mutations with (presumably
negative) fitness consequences than expected,
probably because, with genetic drift, weakly
deleterious mutations can reach appreciable
frequencies.
Lynch has found molecular evidence for
the action of Muller's ratchet in animal mitochondrial genomes (Lynch, 1996). He compared transfer RNA genes in mitochondrial
and nuclear genomes, and observed that the
mitochondrial genome accumulates nucleotide
10
10
8
8
0
0
Figure 1. Muller's ratchet. The population histograms show the relative number of subpopulations with 0, 1,2, 3... etc. deleterious mutations. In the upper histogram,
there is a small proportion of the population still with no mutations. If the population with no mutations (n0) is small in absolute numbers, then there is in each generation
a chance that, despite high fitness, all n0 individuals die without leaving offspring. If so, the optimal class is lost, the ratchet will have clicked (middle diagram), and the
class can be reconstituted only by back mutation (whereas in a sexual population it can be reconstituted by recombination). In the lower diagram the ratchet has clicked a
second time (after Maynard Smith, 1989).
1
10
8
0
G
i
65
"3"
R.F.Hoekstra
substitutions -25 times faster than the nuclear
genome does. The average binding energy
between complementary strands in the stems
of mitochondrial tRNA is less than half that
of nuclear tRNA, and tRNA loop sizes are 50
times more variable in the mitochondrial than
in the nuclear genome (Lynch, 1996): an
observation interpreted as meaning that there
has been a loss of fitness of mitochondrial
tRNAs compared with nuclear tRNAs. He
further argues (Lynch, 1996, 1997) that these
findings support the idea that animal mitochondrial lineages are subject to slow and very
long-term fitness decline due to the accumulation of slightly deleterious mutations not seen
in the nuclear genome. The nuclear genome is
less likely to accumulate deleterious mutations
because of the purging effect of meiotic
recombination. This raises the question for
how long on an evolutionary time scale animal
mitochondrial lineages can survive the inevitable decline in fitness. It has been estimated
that the process is not likely to imperil many
species on time scales of less than a million
years (Lynch and Blanchard, 1998), but
nonetheless it is conceivable that mitochondrial mutations accumulation will have contributed to extinction of some animal species.
Features of mitochondrial genetics
related to the risk of mutation
accumulation
Deficient DNA protection and repair
Unlike the nuclear chromosomes, animal mitochondrial genomes are not protected by histones, spend significant time as single DNA
chains during replication (Clayton, 1991;
2000), and are deficient in several DNA repair
enzymes, such as those involved in excision
repair, photoreactivation and recombination
repair (Lansman and Clayton, 1975). Both
features will contribute to an elevated mitochondrial mutation rate.
108
Reduction of genome size
Some features that reduce mitochondrial mutation accumulation are related to the reduction
of the mitochondrial genome size. The
ancestors of mitochondria are thought to derive
from the alpha-division of the purple photosynthetic bacteria (Yang et al., 1985). Present
day animal mitochondria have much smaller
genomes than their bacterial relatives. This is
mostly the result over evolutionary time of
extensive gene transfer (in most taxa some
90% of the genes) from the mitochondrial
genome to the nuclear genome. Additionally,
the mitochondrial genome is streamlined in
several respects (Kurland, 1992). Mammalian
mitochondria do not contain introns and have
only a single site for the initiation of transcription. Small genomes will carry fewer
mutations than larger genomes and the advantage of a single initiation site (apart from saving
the space required for separate regulation of
each gene) is that a mutation at this site
renders the complete mitochondrial genome
(instead of just one gene) defective. In this way
such a mutation has a more severe deleterious
effect and is more easily eliminated by
selection.
Multiplicity of mitochondrial genomes
The multiplicity of mitochondrial genomes in
a cell also contributes to a slowing down
of mutation accumulation. Because of the
multiplicity of mitochondrial genomes per cell
and the 'relaxed control' (Birky, 1983; see
also Clayton, 2000) of the partitioning of
mitochondria among daughter cells, a heteroplasmic cell with a given number of mitochondrial mutations can give rise to progeny cells
that carry fewer (or more) mutations. In this
way, variation in mutational load among cells
would be produced, enabling natural selection
to favour cells with a relatively small mutation
load. When an individual starts its develop-
Uniparental mitochondrial inheritance
ment from a homoplasmic condition, this
argument only applies to novel mitochondrial
mutations arising during development of an
individual organism. The effect of intracellular
random genetic drift on reducing the rate at
which Muller's ratchet clicks has been analysed (Takahata and Slatkin, 1983). They
suggest that this effect of random genetic
drift could even halt the ratchet, but their
simulations are based on very strong selective
effects (10% reduction in fitness of individuals
fixed for a deleterious mitochondrial mutation). The effect of random drift is expected
to be much smaller when mutational effects on
fitness are small, which seems more realistic
(Lynch, 1997); it is precisely for very weakly
deleterious mutations that the danger of mutation accumulation is highest (as discussed
above).
Bottleneck in mitochondrial genome copy
number
Genetic drift in heteroplasmic cell lineages is
very much enhanced by the occurrence of
bottlenecks in the number of mitochondrial
genomes in the cell. If in every generation
germline mitochondrial genomes are at one
point severely reduced to only a few per cell,
drift would be maximal and natural selection
could potentially be very effective in eliminating inferior mitochondrial genotypes. This
requires that at this stage, the mitochondrial
genome can be put to the test, i.e. that all
mitochondrial functions are activated and that
cells containing suboptimal mitochondria are
in some way competitively inferior to cells
with optimal mitochondria.
Bottlenecking could also make natural
selection for advantageous mitochondrial
mutations more effective. Kearsey et al. have
studied the fate of mutations for mitochondrial
chloramphenicol resistance in a mouse cell
line. The mutations are selectively advantageous in the presence of this antibiotic but
disadvantageous in its absence. Despite the
continuous presence of chloramphenicol,
mutants did not increase in frequency sufficiently to reach homoplasmy (Kearsey et al,
1985). They suggested that a bottleneck in the
oocyte could favour evolution towards fixation
of advantageous mitochondrial mutations.
Paternal leakage
A small amount of biparental mitochondrial
transmission (with recombination between the
maternal and the paternal mitochondrial genomes) could be sufficient to stop Muller's
ratchet. Pamilo et al. concluded from a theoretical analysis that the difficulties created by
Muller's ratchet can be alleviated with a very
low rate of sex-induced recombination (Pamilo
et al, 1987). Recent analyses of human mitochondrial sequences suggest that transmission
of paternal mitochondrial genomes and subsequent recombination may occasionally (but
probably very rarely) occur (Eyre-Walker
et al, 1999; Hagelberg et al, 1999). Here we
encounter
an
interesting
evolutionary
dilemma: selection for a small percentage of
biparental transmission would prevent mitochondrial mutation accumulation, but would
on the other hand allow an increase in population frequency of selfish mitochondrial
mutations, as discussed above.
Using mathematical analysis and computer
simulation, Hastings has analysed the effects
of a low level of paternal leakage on the spread
of selfish cytoplasmic mutations (Hastings,
1999). He concluded that even a small amount
of paternal leakage can have devastating
effects on host population fitness unless it is
either exceedingly rare or tight bottlenecking
through the female germline reduces the
effective population size to nearly unity (a
combination of the two mechanisms is
allowable). This result emphasizes the essential role of a mitochondrial bottleneck in
rapidly exposing a mutation deleterious to
109
R.F.Hoekstra
individual host fitness and allowing natural
selection to remove affected genomes.
Hastings also investigated the effect of
paternal leakage in life cycles with a number
of asexual divisions between sexual reproduction, as is common among many protists
(protozoa, including algae) and fungi
(Hastings, 1999). The threat posed by selfish
cytoplasmic elements seems to be greatly
diminished if sexual reproduction is a rare
event in an otherwise asexual life cycle,
because during asexual reproduction lineages
containing selfish mutations can be eliminated
by selection. This could explain the fact that
some degree of biparental cytoplasmic transmission is not uncommon among these groups
of organisms (yeast being a well-known
example).
In conclusion, it seems that the effect of
natural selection on mammalian mitochondrial
transmission operates in a balance, between
the Scylla of mutation accumulation and the
Charybdis of invasion by selfish mutations,
by allowing just a little bit of paternal genomic
leakage from a successful fertilizing spermatozoon (with competitive mitochondrially driven
motility and hence, by implication, mitochondrial genomic integrity): too little to allow
selfish elements to spread, but enough to
contribute significantly to slowing down
Muller's ratchet.
References
Ballard, J.W.O. and Kreitman, M. (1994) Unraveling
selection in the mitochondrial genome of Drosophila.
Genetics, 138, 757-772.
Birky, C.W., Jr. (1983) Relaxed cellular controls and
organelle heredity. Science, 222, 468^75.
Birky, C.W., Jr. (1995) Uniparental inheritance of
mitochondrial and chloroplast genes: Mechanisms and
evolution. Proc. Natl Acad. Sci. USA, 92, 11331—
11338.
Clayton, D.A. (1991) Replication and transcription of
vertebrate mitochondrial DNA. Annu. Rev. Cell Biol,
7, 453-478.
Clayton, D.A. (2000) Transcription and regulation of
110
the mitochondrial genome. Hum Reprod., 15 (Suppl.
2), 11-17.
Cosmides, L.M. and Tooby, J. (1981) Cytoplasmic
inheritance and intragenomic conflict. /. Theor. Biol.,
89, 83-129.
Eberhard, W.G. (1980) Evolutionary consequences of
intracellular organelle competition. Q. Rev. Biol., 55,
231-249.
Eyre-Walker, A., Smith, N.H. and Maynard Smith, J.
(1999) How clonal are human mitochondria? Proc.
R. Soc. Lond. B, 266, 477-483.
Felsenstein, J. (1974) The evolutionary advantage of
recombination. Genetics, 78, 737-756.
Frank, S.A. and Hurst, L.D. (1996) Mitochondria and
male disease. Nature, 383, 224.
Gillham, N.W. (1994) Organelle Genes and Genomes.
Oxford University Press, New York, USA.
Grun, P. (1976) Cytoplasmic Genetics and Evolution.
Columbia University Press, New York, USA.
Hagelberg, E., Goldman, N., Lio, P. et al. (1999)
Evidence for mitochondrial DNA recombination in a
human population of island Melanesia. Proc. R. Soc.
Lond. B, 266, 485^92.
Haigh, J. (1978) The accumulation of deleterious genes
in a population - Muller's ratchet. Theor. Popul. Biol.,
14, 251-267.
Hastings, I.M. (1992) Population genetic aspects of
deleterious cytoplasmic genomes and their effects on
the evolution of sexual reproduction. Genet. Res., 59,
215-225.
Hastings, I.M. (1999) The cost of sex due to deleterious
intracellular parasites. J. Evol. Biol., 12, 177-183.
Hoekstra, R.F. (1987) The evolution of sexes. In Stearns,
S.C. (ed.) The Evolution of Sex and its Consequences.
Birkhaeuser Verlag, Basel, Switzerland.
Hoekstra, R.F. (1990) Evolution of uniparental
inheritance of cytoplasmic DNA. In Maynard Smith,
J. and Vida, J. (eds.) Organisational Constraints on
the Dynamics of Evolution. Manchester University
Press, Manchester, UK.
Hurst, L.D. and Hamilton, W.D. (1992) Cytoplasmic
fusion and the nature of sexes. Proc. R. Soc. Lond.
B, 247, 189-194.
Kaneda, H., Hayashi, J.I., Takahama, S. et al. (1995)
Elimination of paternal mitochondrial DNA in
intraspecific
crosses
during
early
mouse
embryogenesis. Proc. Natl Acad. Sci. USA, 92,
4542^546.
Kearsey, S.E., Munro, E. and Craig, I.W. (1985) Studies
of heterogeneous mitochondrial populations in a
mouse cell line: The effects of selection for or against
mitochondrial genomes that confer chloramphenicol
resistance. Proc. Roy. Soc. Lond. B., 224, 315-324.
Kurland, C.G. (1992) Evolution of mitochondrial
genomes and the genetic code. Bioassays, 14, 709714.
Uniparental mitochondrial inheritance
Lansman, R.A. and Clayton, D.A. (1975) Selective
nicking of mammalian mitochondrial DNA in vivo:
Photosensitization by incorporation of 5-bromodeoxyuredine. J. Mol. Biol, 99, 761-776.
Law, R. and Hutson, V. (1992) Intracellular symbionts
and the evolution of uniparental inheritance. Proc. R.
Soc. Lond. B, 248, 69-77.
Lynch, M. (1996) Mutation accumulation in transfer
RNAs: molecular evidence for Muller's ratchet in
mitochondrial genomes. Mol. Biol. Evoi, 13,209-220.
Lynch, M. (1997) Mutation accumulations in nuclear,
organelle, and prokaryotic transfer RNA genes. Mol.
Biol. EvoL, 14, 914-925.
Lynch, M. and Blanchard, J.L. (1998) Deleterious
mutation accumulation in organelle genomes.
Genetica 102/103, 29-39.
Maynard Smith, J. (1989) Evolutionary Genetics. Oxford
University Press, Oxford, UK, pp. 241-243.
Muller, H.J. (1964) The relation of recombination to
mutational advance. Mutat. Res., 1, 2-9.
Nachman, M.W., Brown, W.M., Stoneking, M. and
Aquadro, C.F. (1996) Nonneutral mitochondrial DNA
variation in humans and chimpanzees. Genetics, 142,
953-963.
Pamilo, P., Nei, M. and Li, W.-H. (1987) Accumulation
of mutations in sexual and asexual populations. Genet.
Res., 49, 135-146.
Rand, D.M., and Kann, L.M. (1996) Excess amino
acid polymorphism in mitochondrial DNA: Contrasts
among genes from Drosophila, mice and humans.
Mol. Biol. EvoL, 13, 735-748.
Takahata, N. and Slatkin, M. (1983) Evolutionary
dynamics of extranuclear genes. Genet. Res., 42,
257-265.
Wallace, D.C. (1992) Mitochondrial genetics: a paradigm
for aging and degenerative disease? Science, 256,
628-632.
Yang, D., Oyaizu, Y., Oyaizu, H. et al. (1985)
Mitochondrial origins. Proc. Natl Acad. Sci. USA, 82,
4443^447.
Ill