Evolution of Sex and Breeding Systems:
Revision Notes

Compiled as a final-year zoology student at the University of Edinburgh, based on the information given in lectures.

Evolution of haploidy and diploidy

(Mable & Otto 1998, in BioEssays)

Advantages of haploidy - haploids are smaller and have a larger relative surface area for absorbing nutrients so may grow faster; haploids have a lower mutational load at equilibrium.
Advantages of diploidy - somatic mutations are masked, DNA damage can be repaired using the homologous copy as a template, diploid populations might evolve faster.
In haploids, mean fitness is W=~1-µ (where µ is mutation rate), in diploids W=~1-2µ (double the mutational load), assuming mutations are deleterious and have at least a small effect on heterozygotes.
Adaptive mutations are fixed more rapidly in diploid than in haploid yeast populations (Paquin & Adams 1983).
Haploid and diploid stages may coexist because they exploit different niches (Hughes & Otto 1999).
Distinct mating types (e.g. in yeast, Chlamydomonas) can evolve by loss of one or other component of a cell surface recognition (Hoekstra 1982), and help to avoid inbreeding.
Anisogamy (two sizes of gametes) can be maintained if there is an accelerating relationship between gamete size and fitness (k>1 where viability of gametes is proportional to (m₁+m₂)^k and m is gamete size), and numbers of gametes produced are inversely proportional to their size (Bell 1978).

Inbreeding vs. outbreeding

Selfing in hermaphrodites may be prevented by spatial or temporal separation of male/female gamete release, or by self-incompatibility (due to an S-locus with many different alleles).
Self-incompatibility in plants may be sporophytic (incompatibility type of pollen controlled by S-locus of plant that produces it), or gametophytic (incompatibility type controlled by the haploid gene of pollen itself). It may be heteromorphic (heterostyly), in which flowers of different incompatibility types are morphologically different.
Selfing rates can be estimated from genotypes of parents and progeny (assuming selfing has no effect on offspring survival and that pollinators are indiscriminate), or from heterozygosity of population (assuming equilibrium).
Selfing rate, S = 2F / (1+F), where F is 'fixation index', and 1 - F = observed heterozygote frequency / 2pq
Advantages of selfing - reproductive assurance, saves resources, allows locally-adapted genotypes to persist.
Selfish B-chromosomes are more prevalent in outcrossing plant species - the fitness cost they impose upon the host leads to them being purged from inbred lineages (Burt & Trivers 1998).
Outcrossing mechanisms are commoner in long-lived species than annuals, due to annuals' need for reproductive assurance (Barrett et al 1996). Distribution of selfing rates is highly binomial, with extremes over-represented.
Cost of outcrossing - an individual only transmits one copy of its genome to next generation, rather than two. Selfing should thus be favoured if inbreeding depression is less than 0.5. A rare gene for selfing can spread rapidly, e.g. spread of homostyled primroses in heterostyled populations (Crosby 1949).
Inbreeding depression is lowered by the fact that inbreeding facilitates purging of recessive deleterious alleles.
Advantage of selfing is lowered if it uses up pollen that could otherwise be exported (pollen discounting).
Why intermediate selfing rates might persist: increasing disadvantage of selfing with successive generations of inbreeding, local adaptation, inbreeding between relatives (which lowers advantage of selfing relative to outcrossing), trade-offs between male and female function (which cause fertility to be frequency-dependent).
Inbreeding depression of at least 2/3 is necessary for initial spread of self-incompatibility genes. Once self-incompatibility has arisen, rare S-alleles have an advantage, so large numbers should evolve.
Selfing leads to increased homozygosity, and in the long term reduced diversity within a population. In plants it leads to reduced investment in flower production, high fruit/flower ratios, and increased allocation of resources to female functions.

Intragenomic conflict

(Maynard Smith & Szathmáry 1995, in The major transitions in evolution)

An allele with 'meiotic drive' is one that can increase its probability of being transmitted to the next generation at meiosis. Such alleles are rare in nature because meiotic drive is hard to achieve and there will be selection among the other genes in an organism to resist it.
The segregation distorter (SD) gene in Drosophila produces a toxin inactivating any sperm with a particular allele (Rsp⁺) at a 'responder' locus. The two loci are tightly linked. A similar system (the t-locus) occurs in mice.
B chromosomes are useless, parasitic chromosomes, found in many groups, that impose a fitness cost upon their host but persist because of their ability to increase in number within a lineage (Burt & Trivers 1998). In some plant tissues, B chromosomes multiply as a result of non-disjunction. In the grasshopper Myrmeleotettix maculatus, they move preferentially to the egg nucleus rather than the polar bodies.
Transposable elements of DNA can replicate selfishly within a genome. Their spread is counteracted by lowered fertility resulting from transposons' tendency to cause mutations or to lead to the production of aneuploid gametes (as a result of 'ectopic recombination' between elements at different chromosome locations).
Uniparental inheritance of organelles may have evolved to avoid intragenomic conflict - it occurs even in isogamous organisms (such as Chlamydomonas).
Selfish genetic elements may distort the sex ratio. In the wasp Nasonia vitripennis, maternal sex ratio (msr) is a maternally-transmitted factor that causes females to produce mostly daughters; son-killer (sk) is a maternally transmitted bacterium that causes male eggs to die, and paternal sex ratio (psr) is a B chromosome that destroys other paternal chromosomes to produce a haploid male (males are better than females at transmitting the B chromosome to their descendants). (Werren et al 1988)

Sex determination systems

Sex determination may be environmental (e.g. due to temperature in crocodilians) or genetic.
In male heterogamety, females are XX and males XY; occurs in mammals, where maleness is due to Sry gene on Y chromosome (Graves 2002), and in Drosophila, where ratio of X chromosomes to autosomes determines sex.
In female heterogamety, males are ZZ and females ZW; occurs in birds, Lepidoptera and various lower vertebrates.
In some species there is a dominant gene M that causes development as a male regardless of karyotype, e.g. in house fly Musca domestica (reviewed by Shearman 2002).
In some lemmings, there is an X* variant of X chromosome and X*Y individuals are female. In the wood lemming Myopus schistocolor, non-disjunction means that X*Y females only produce X* eggs; in varying lemming Dicrostonyx torquatus there is no non-disjunction and a quarter of zygotes are non-viable YY, but this isn't a big problem because more embryos are implanted than survive to birth. In both cases, X* has transmission advantage over X.
The fixation of such genes may lead to the evolution of new sex chromosomes. There is evidence of transitions between male and female heterogamety in evolutionary history.
Haplodiploidy (arrhenotoky), in which unfertilised eggs develop as males, occurs in all Hymenoptera. Sex is determined by a locus with multiple alleles at which heterozygotes are female, and haploids (or homozygous diploids, though these are inviable or infertile) are male. In some other groups (mites, scale insects, sciarid flies), all eggs are fertilised but paternal genome is eliminated in males.
In some crustaceans (e.g. woodlouse Armadillidium), maternally-transmitted endosymbionts such as Wolbachia override sex determination systems causing feminisation of carriers (reviewed by Rigaud 1997).
When dioecy has evolved from cosexuality, theory predicts a higher incidence of male heterogamety, which is what is observed in plants (exception: strawberry), but this is not necessarily expected when genetic sex determination has evolved from environmental sex determination.
Environmental sex determination will be favoured when relative fitness of sexes depends on environmental conditions (e.g. in marine worm Bonellia viridis, where larvae that settle next to females become male). A bias towards the sex with the lowest average fitness across environments is expected (reviewed by Shine 1999).
Why should genetic sex determination evolve? A gene conferring maleness will spread if sex ratio is variable since it gains an advantage when males are rare; this creates a male-biased sex ratio that favours evolution of 'female' genes too. Same applies if a femaleness factor appears first.
Haplodiploidy may arise due to a gene carried by the mother, to which haplodiploidy gives a transmission advantage, provided the fitness of haploid males is at least half that of diploid males.

Sex ratios

Fisher (1930): equal resources should be invested in male and female offspring, because an investment ratio biased in favour of one sex would give a fitness advantage to the other (it would be rarer and get more matings).
The hyperparasitoid wasp Encarsia tricolor, in which male eggs are laid within female parasites, will lay 50:50 male/female eggs if eggs are limiting, but if hosts are limiting, the ratio will depend on proportion of already-parasitized hosts encountered (Hunter & Godfray 1995).
In polygynous species such as red deer (Sheldon & West 2004), females in good condition should have more sons since a male's fitness is more dependent on quality than female's fitness. In blue tits, females with a more attractive mate lay more male eggs for similar reasons (Sheldon et al 1999).
In parasitoid wasps, more female eggs are laid in good (large) hosts since female quality is important for fitness (fecundity depends on body size) but male quality is less important (West & Sheldon 2002).
Sex-changing fish where big males control harems will be female early in life and male later (Shapiro 1980).
Local Resource Enhancement - if one sex of offspring stays behind as 'helpers' (e.g. male African wild dogs), more of this sex should be produced. In Seychelles warblers, where helpers (female) are only helpful in good territories, more female offspring are produced under these conditions (Schwarz 1988).
Local Resource Competition (LRC) - where one sex disperses, more of this sex should be produced since they are less likely to compete with siblings. Suggested in primates, but evidence is not clear (Brown & Silk 2002).
Most obvious example of LRC is in social insects, where offspring of swarming species are strongly male-biased.
Local Mate Competition is a specific form of LRC. Female-biased sex ratios increase fitness of sons (more partners, less competition between brothers) and increase relatedness of mothers to daughters in haplodiploid species. Higher inbreeding should lead to more female-biased sex ratios - observed in fig wasps, where there is less inbreeding bigger figs containing more individuals (Herre 1987).
In Plasmodium and related apicomplexan parasites (where male gametocytes can produce eight gametes but female gametocytes produce only one), inbreeding favours female-biased sex ratios. Sex ratio is less biased where transmission intensity (and thus number of clones per host) is higher (West et al 2001).

Evolution of separate sexes

In both plants and animals, dioecy evolved separately many, many times
Dioecy in plant lineages is probably recent - this is inferred from its distribution and the fact that unisexual plants often have opposite-sex rudiments. Dioecious lineages may have short evolutionary lives.
Models for evolution of dioecy assume trade-offs between male and female functions.
Advantages of dioecy vs. hermaphroditism depend on gain curves - concave curves (intermediates have higher fitness than pure males or females) favour hermaphroditism; convex curves favour dioecy.
Evolution of two sexes requires a minimum of two genetic changes, producing male sterility and female sterility.
Cytoplasmic male sterility (e.g. in maize) will invade a population if it gives females even a slight fertility advantage (leading to gynodioecy). Nuclear genomes may have genes restoring hermaphroditism.
Nuclear male sterility genes will invade if they double female fertility, or if they help avoid inbreeding depression - both advantages are probably important. Females producing 1+k times as many seeds as hermaphrodites will be fitter if k > 1-2Sd where S is selfing rate and d is inbreeding depression.
Females are most likely to invade hermaphrodite populations that allocate resources highly to male functions.
Experiments on gynodioecious populations fit the models well. In the gourd Cucurbita foetidissima there is intense inbreeding (Kohn 1988). In New Zealand Umbelliferae, female frequencies increase with difference in fertility between females and hermaphrodites (Webb 1979).
Female flowers may appear hermaphroditic to fool pollen-seeking insects into visiting them (Charlesworth 1984).
Invasion of males into hermaphrodite populations is not helped by inbreeding, since this simply reduces availability of potential partners. Males can only invade if their fitness is more than twice that of hermaphrodites. Androdioecy is therefore rarer than gynodioecy. In the plant Datisca glomerata, androdioecy has evolved from dioecy due to occasional need for self-fertilisation (Fritsch & Rieseberg 1992). In Mercurialis annua, cosexuals devote a high proportion of resources to female function and unisexual males have much higher pollen production (Pannell 1997). In both species, denser populations (with higher outcrossing rates) have a higher proportion of males.
In androdioecious nematode Caenorhabditis elegans, males persist because they can mate with hermaphrodites but other hermaphrodites cannot.
In a newly-dioecious species, linkage of male-sterility and female-sterility genes would be favoured to avoid the production of sterile offspring carrying both. This may start the evolution of sex chromosomes.

Why sex and recombination?

Types of sex: automixis (selfing), amphimixis (outcrossing sex).
The union of haploid cells (syngamy) can be explained by the advantage of masking different deleterious alleles, but recombination requires a different explanation.
Recombination (mixing up of genomes) occurs by crossing-over or by segregation of chromosomes.
Recombination in bacteria can occur by transformation (uptake of DNA, which may be done primarily for nutrition), transduction (transfer by a phage), or conjugation (transfer of plasmids). It may be accidental. Bacterial sex may have evolved to benefit selfish genetic elements by allowing them to spread (Maynard Smith & Szathmáry 1999).
Recombination in eukaryotes probably evolved from DNA repair mechanisms; many enzymes involved have homologues in prokaryotes (such as recA in E. coli, which unwinds DNA so that homologous strands can pair up). It cannot simply be accidental side-effect of DNA repair, however - recombination rates differ among organisms (male Drosophila lack recombination, and artificial selection experiments can alter recombination rates), and double-stranded DNA breaks are actually induced during meiosis.
Weissman argued that sex evolved 'for the good of the species'. However, there must also be short-term advantages.
Selection among species does help to maintain sex - asexual taxa go extinct rapidly (with some exceptions).
In insects, arrhenotoky (unfertilised eggs becoming males) arose 8 times, including in some fairly large taxonomic groups; thelytoky (unfertilised eggs becoming females) arose ~1000 times and is sporadically distributed.
In some groups parthenogenesis occurs cyclically (e.g. aphids, gall wasps) or facultatively (e.g. nematodes, snails).
The short-term advantage to sex must be strong to overcome the disadvantages: meiosis is slow, courtship and mating are costly, organisms risk not finding a mate, sexual selection may lower survival, recombination breaks up favourable gene combinations (comparison of chromosomes in male and female Drosophila suggest that recombination reduces viability by ~1% and fecundity by ~7%).
In many cases selection has reduced recombination rates: 'supergenes' (such as the clusters of genes that determine appearance of mimics), Y chromosomes
The composition of the population is only changed if there is non-random association between genes (linkage disequilibrium).
Sibling competition may explain the advantage of recombination and variety in some situations, although this cannot be a general explanation. Experiments on Drosophila show that more variable families (those with different fathers) compete less. Survivors of intra-family selection may compete better with other families in later life.
Recombination may offer an advantage in unpredictable environments. It will be favoured only if selection favours positive associations of alleles in some generations and negative associations in others (fluctuating epistasis).
Recombination facilitates adaptation, although only if there tend to be negative associations between favourable alleles. Artificial selection experiments tend to raise recombination rates (Otto & Lenormand).
Sex and recombination may help maintain resistance against rapidly-evolving parasites (Hamilton 1980).
Many mutualists have lost the ability to reproduce sexually (due to less need for rapid adaptation?).
Kondrashov suggested that synergistic epistasis (the ability of organisms to tolerate moderate mutational loads with little loss of fitness) could favour recombination, but this isn't well supported by evidence, and would require an unrealistically-high mutation rate.
Experiments with Chlamydomonas show that adaptation is faster in bigger populations, but in asexual populations there are diminishing returns with increasing population size, because in the absence of recombination favourable mutations have to be acquired one after the other (Colegrave 2002).

Evolution of sex chromosomes

(Charlesworth 2002, in The Genetics and Biology of Sex Determination)

The X and Y chromosomes pair at meiosis but fail to undergo recombination along some or part of their length.
Sex chromosomes evolved from autosomes that acquired sex-determining genes. The mammalian X and Y have ancient homologies (Lahn & Page 1999).
Sexually antagonistic genes - those that favour one sex at the other's expense, such as guppy colour genes (Fisher 1931) - become linked to the sex-determining locus, favouring non-recombination between sex chromosomes.
In 'young' sex chromosomes, non-recombination may be restricted to a small region, e.g. in papaya (Liu et al 2004).
In bird sex chromosomes, phylogenetic relationships between genes show that some sets stopped recombining earlier than others (Lawson-Handley et al 2004).
Y chromosomes tend to accumulate deleterious mutations because they never recombine, they exist at lower numbers than other chromosomes in the population, and their genes are never homozygous.
Muller's ratchet is the inexorable accumulation of deleterious mutations in the absence of recombination, as the least mutation-loaded types are lost from the population (Muller 1964). It is faster in smaller populations.
Background selection involves the rapid loss of strongly-selected deleterious mutations from a non-recombining population, carrying any new variants with them and decreasing the effective population size.
When a highly beneficial allele arises in a non-recombining population and goes to fixation (in a 'selective sweep'), mildly deleterious mutations associated with it will also be fixed.
'Dosage compensation' ensures that the activity of genes on the X chromosome is similar in males and females. In Drosophila, X in males is up-regulated; in C. elegans (where hermaphrodites are XX and males X0), X is down-regulated in hermaphrodites; in mammals, one X is inactivated in female cells. (Marín et al 2000)
Many groups have 'neo-X' and 'neo-Y' chromosomes, with a pair of autosomes attached to the sex chromosomes.
In different Drosophila species, different stages of degeneration of the neo-Y can be seen (there is no recombination in male Drosophila so the neo-Y totally stops recombining). The neo-X shows signs of faster change in two adaptive genes (the neo-X adapts faster than the neo-Y), the neo-Y shows faster change in other genes (slightly deleterious mutations accumulate faster), and some neo-Y genes have mutations that destroy their function. There is an excess of rare variants of the neo-Y, perhaps due to 'selective sweeps'. (Bachtrog 2003/4)

More notes and essays

Evolution of Sex and Breeding Systems: Revision Notes