The evolution of sex chromosomes similarities and differences

The evolution of sex chromosomes: similarities and differences between plants and animals Deborah Charlesworth Institute of Evolutionary Biology, University of Edinburgh Papaya female Silene dioica male Silene latifolia female

• 1. WHAT are sex chromosomes? – and what are NOT sex chromosomes • 2. WHY do sex chromosomes evolve loss of recombination? • 3. WHEN did sex chromosomes of some important species evolve? – and when did recombination stop? • and 4. HOW did recombination stop? • 5. WHERE are the sex-determining loci in relation to the regions where recombination is absent? • 6. WHAT are the consequences for sex chromosomes of stopping recombination?

Sex chromosomes have been known to geneticists for a long time • Muller (1914): reviewed evidence for X-Y pairing (indicating their homology) and Y genetic degeneration (suggested by C. W. Metz) and discussed recessive loss of function mutations as the cause of degeneration • Haldane (1922, p. 107): “If sex were determined by a single factor, it is very difficult to see what advantage there could be in its being linked with other factors)” • An excellent review of the classical work is JJ Bull’s 1983 book “Evolution of Sex Determining Mechanisms” but many important things have only become clear very recently, and great progress is occurring • Lahn and Page (1999): human genome sequence reveals sequences of genes shared between X and Y, often highly diverged. Carvalho (2001): Drosophila Y-linked genes • Recent data are starting to help us understand why and how recombination gets stopped between the X and Y (and what the consequences are) • Further evolutionary changes of sex chromosomes can now be studied in detail

Classical sex chromosomes Humans Y is ~ 1/3 of the size of the X X ~ 1, 098 genes Y 24 genes Y X Male-specific Y region Pseudo. MSY autosomal region without recombination PAR In Drosophila, the Y is about the same size as the X, but the X has several thousand genes, while the Y has around 20. No X gene has a Y homologue.

Some species, including many plants have small sexdetermining regions Several plant ‘sex chromosomes’ have the sexdetermining genes located within a small region (blue; only 10% of chromosome 1 of papaya) where recombination does not occur (Liu et al. 2004) – Some fish sex chromosomes may be similar – Does small size mean young, or primitive? Other plants have heteromorphic sex chromosomes like those of humans and Drosophila, or neosex chromosomes neo-sex chromosomes also occur in plants Liu et al. 2004. Nature 427: 348 -352

Some highly heteromorphic plant sex chromosomes Y Silene latifolia X PAR MSY region The liverwort, Marchantia polymorpha (haploid) X Y Yamato et al. (2007) Proc. Natl. Acad. Sci. USA 104, 6472 -6477

Haploid sex chromosomes in a bryophyte with separate sexes (Ceratodon purpureus) Meiosis 2 n → n Diploid sporophyte XY AA NOTE No XX Haploid Fertilization n → 2 n YA XA Male Female gametophyte Mc. Daniel et al. 2007 Genetics 176: 2489 -2500 Genetic map showing 15 linkage groups, including the XY chromosome pair (121 AFLP, 3 genic markers )

Neo-sex chromosomes due to Fusions/translocations X Y Autosome Y-autosome fusion X-autosome fusion neo-X Y 1 X 1 neo-Y or Y 2 neo-X or X 2 neo-Y Chromosome(s) transmitted to female progeny Chromosome(s) transmitted to male progeny In Drosophila, there is no recombination in males Thus, both kinds of fusion create non-recombining neo-sex chromosomes Chromosome fusions can lead to heteromorphism Fusion can also occur in X 0 systems

Neo sex chromosomes in the genus Drosophila Complete degeneration of the ancestral Y Carvalho, A. B. , and A. G. Clark. 2005. Y chromosome of D. pseudoobscura is not homologous to the ancestral Drosophila Y. Science 307: 108 -110.

Neo-sex chromosomes occur in many species Rowell, D. (1985). Complex sex-linked fusion heterozygosity in the Australian huntsman spider Delena cancerides (Araneae: Sparassidae). Chromosoma 93, 169 -176. RENS et al. , 2004 PNAS 101: 16257 -16261 GRÜZNER et al. , 2004 Nature 432: 913 -917.

• The mammalian sex chromosomes evolved via a fusion since the split from marsupials X Y marsupial X autosome in marsupials • The Y can be lost entirely if genes required for male fertility can move to a different chromosome – Some species have X/0 male genotype, but males are still fertile, e. g. Drosophila affinis • In D. pseudoobscura, the X chromosome has been fused to an autosome, and the Y has lost all male fertility genes – So, even if the Y chromosome degenerates, we do not need to worry about a future without males • Clearly, the Y cannot be lost unless the sex-determination function is replaced by a new gene (or the Y gene moves to another chromosome) – Such changes are theoretically possible: e. g. DOORN and KIRKPATRICK, 2007 Turnover of sex chromosomes induced by sexual conflict. Nature 449: 909 -912, KOZIELSKA et al. , 2010 Segregation distortion and the evolution of sex-determining mechanisms. Heredity 104, 100 -112.

Changes can occur from XY to ZW system (even in different populations of the same species) An example of the use of genes to demonstrate that the XY pair of chromosomes changed into a ZW pair Uno et al. 2008. Comparative chromosome mapping of sex-linked genes and identification of sex chromosomal rearrangements in the Japanese wrinkled frog (Rana rugosa, Ranidae) with ZW and XY sex chromosome systems. Chromosome Research: 1217 -7

Summary of diversity of sex-determining chromosomes • CLASSICAL – Non-recombining over a large genome region, with small “pseudo-autosomal region(s)”, e. g. mammal and Drosophila X and Y, bird and Lepidopteran Z and W – Genetically degenerated • loss of genes relative to the X (or lower function — see later) • The Y can sometimes be totally lost (X 0 systems) – Y is enriched in male-function genes, and is rearranged relative to the X • “LOCAL SEX-DETERMINING REGION” – Chromosome is largely pseudo-autosomal – The same properties as classical sex chromosomes, in a restricted region of genome – We know much less about these, and modern molecular approaches are helping get information • HAPLOID – Haploid male genotype is Y and female is X • NOT SEX CHROMOSOMES – single-gene systems, e. g. sex-determination factor that replaced a previous one, honeybee complementary sex-determiner – plant, fungal and algal incompatibility regions (but have some similar properties)

Monoecious Hermaphrodite Environmental sex-determination Mutation to loss of ♂ or ♀specific developmental pathway ♂-sterility mutation Genetically + environmentally determined unisexuals Gynodioecious (♀and hermaphrodite) Silene vulgaris ♀-sterility mutation GSD dioecy (♂ and ♀) X Proto-sex Y chromosomes Hermaphrodite Female Non-recombining region Replacement by new sexdetermining gene Mutation to sexually antagonistic gene New non-recombining region evolves Translocation onto another chromosome

Possible origins of sex-determining chromosomes

2. WHY does recombination stop on sex chromosomes? • Haldane (1922, p. 107): “If sex were determined by a single factor, it is very difficult to see what advantage there could be in its being linked with other factors)” (Sex ratio and unisexual sterility in hybrid animals. J. Genetics 12: 101 -109) – Nei (1969, 1970): models for lack of recombination and consequent accumulation of detrimental mutations leading to degeneration – Nei, M. 1969. Linkage modification and sex difference in recombination. Genetics 63: 681 -699; 1970. Accumulation of nonfunctional genes on sheltered chromosomes. American Naturalist 104: 311 -322. • But many modern authors are much less clear e. g. “heteromorphic sex chromosomes have evolved …. when one autosome develops a dominant sex-determining mutation” – Itoh et al. 2007. Molecular cloning of zebra finch W chromosome repetitive sequences: evolution of the avian W chromosome. Chromosoma 117: 111 -121.

Why are 2 genes involved? In many plants, males and females are simply hermaphrodites with parts missing. In S. latifolia, sex-determination is genetically simple. Mutants support the hypothesis that at least 2 genes are involved 1. Loss of stamen promoting factor (SPF or M) creates females 2. Gynoecium suppressing factor (GSF or Su. Female) reduces female functions Hermaphrodite Picture from Shigeyuki Kawano Neuter

The simple 2 gene evolutionary model suggests that (1) sex determining loci must initially be linked for separate sexes to evolve and (2) once females are present, hermaphrodites are selected to re-allocate more to male and less to female functions 1 2 M Su. Female “proto-Y” “proto-X” m f Selection should then act to reduce recombination between the initial 2 genes slightly older Female M 2 M Su proto-Y, with MSY region

Summary of question 2: There is no selection to reduce recombination unless at least 2 genes interact • Qvarnstrom & Bailey. 2009. Heredity 102: 4 -15 are completely wrong! – • The evolution of identifiable heteromorphic sex chromosomes is initiated by the spread of a sex-determining gene. This occurs when a new mutation at a locus leads all its carriers to become the same (subsequently heterogametic) sex, with the chromosome carrying this mutation becoming the Y/W chromosome. In eutherian mammals, for example, the development of males is controlled by the SRY gene found only on the Y chromosome Here is another misunderstanding (Sekido & Lovell-Badge. 2009. Tr. Genet. 25: 1929) – In eutheria, Sox 3 is X linked and involved in the development of the CNS, the pituitary, pharyngeal region and is perhaps involved in male fertility, but it has no demonstrable role in sex determination. The original mutation that led to the origin of Sry, therefore, seems to have involved the acquisition of a novel function (neomorph) and it could have been the primary drive for the separation of the two sex chromosomes. • These authors are confused between the evolution of sex chromosomes and the evolution of modified sex-determining systems • To understand why the sex chromosomes don’t recombine, we need to understand WHY interacting genes are involved, which requires understanding HOW separate sexes evolved, and what kinds of genes were involved – BC, DC (1978, Amer. Nat. 112: 975 -997) : evolution of sex-determining region with two loci (driving selection for less recombination) • This question is separate from: how is recombination lost? – single step when the sex chromosomes orginated or a gradual process, with several successive steps, and whether inversions were involved?

How can one get sex-linked genes to map sex-determining chromosomes? • To estimate ages of sex chromosomes, and to study degeneration, we need to find genes and study alignable X and Y alleles • There are few known mutant phenotypes (as Muller realised) • Molecular methods are needed (Muller realised this too, in 1922) • Even with “complete” genome sequences of important “model organisms” there are still great difficulties – The gene content of the Y chromosomes of important “model organisms” have only recently been determined • Drosophila: Carvalho et al. 2001 PNAS 98: 13225 -13230 • Humans: SKALETSKYet al. , 2003 Nature 423: 825 - 837, BHOWMICK et al, 2007 Genome Res. 17: 441 -450; Chimpanzee: HUGHES et al. 2010 Nature (13 January 2010). The mouse Y is still not well characterized • and sex chromosomes of non-model organisms” are only now starting to be studied — EST sequences can be very helpful – e. g. Dreyer et al. 2007. ESTs and EST-linked polymorphisms for genetic mapping and phylogenetic reconstruction in the guppy, Poecilia reticulata. BMC GENOMICS 8: 269, Tripathi et al. 2009 Genetic linkage map of the guppy, Poecilia reticulata, and quantitative trait loci analysis of male size and colour variation. Proceedings of the Royal Society B 276, 2195 -2208. How can one get sex-linked genes to map sex-determining chromosomes?

Why is it difficult to sequence Y chromosomes? • Low gene density makes finding genes very difficult. • Rearrangements: one homolog cannot used to help align the other, unlike the autosomes – Y can be sequenced from a single individual • Their intergenic regions and introns contain large amounts of repetitive sequence, so it is difficult to find the different parts of the same gene • Assembly of highly repetitive genomes is very difficult – it requires large sequenced regions, such as BAC clones, but these may be difficult to sequence if they contain repetitive sequences • These are sometimes unstable when cloned, and so cannot be sequenced • They may compete in PCR reactions, so that some copies fail to amplify • If the repetitive sequences are AT-rich, poor strand separation may impede sequencing reactions • In humans and Drosophila, Y-linked genes have been found, and, in humans, some have X-linked alleles

The human MSY region genes Many Human Y genes have male functions Genes on Y only Y genes with male functions can be kept on the Y because the sex chromosomes don’t recombine across much of their length. These genes are probably prevented from degenerating Genes on X and Y Heterochromatin There a few X-Y gene pairs (X homologous genes), even in the nonrecombining regions (NRY)

Maybe one should sequence the genome? Overview of the Marchantia polymorpha YR 2 region — so far in this species mainly Y chromosome data, not X and Y. This species is expected to have an old Y chromosome 50 Y-linked housekeeping genes are also found in females (presumably nondegenerated genes, with autosomal or X-linked copies) 14 Y-linked genes are unique to males, and expressed only in reproductive organs G = genes (indicated by arrows ) P = pseudogenes O= organelle sequence T = transposable element

I emphasized how helpful it is The genus to identify genes, not just Silene anonymous markers or sequences, and that plants are interesting for studying de novo evolution of sex chromosomes (because sex chromosomes have evolved recently in several taxa) Plants BUT finding X and Y genes in non-model species is difficult, and the S. latifolia genome is big! Humans Dioecious (independent evolution) Gynodioecious Dioecious Hermaphrodite Estimated from ITS sequences by Desfeux, C. , et al. 1996. Proc. Roy. Soc. Lond. B. 263: 409 -414 Recent work with more nuclear genes supports these phylogenetic relationships

How else can one find sex-linked loci? • Testing linkage of known genes involved in flower development, using families – MROS 3 -X and -Y (Dave Guttman, 1998) – Sl. Ap 3 (Sachi Matsunaga 2003) • c. DNA probing of micro-dissected Y chromosomes – Sl. X/Y 1 (Delichère et al. , 1999) – Sl. X/Y 4 (Atanassov et al. , 2001) – Sl. X/Y 3 (Nicolas et al. 2005 – Genes can now be discovered from c. DNA libraries and EST sequences of any species of interest – – – Sl. Ss-X/Y Sl. Cyp-X/Y Sl 8 -Y only Sl 6 a and b X/Y Sl 7 X/Y – RB 11 and RB 18 Dmitry Filatov Roberta Bergero Isomerase, cyclophilin type Roberta Bergero Mono-oxygenase/haem binding protein Roberta Bergero Unknown protein (2 Y and X copies) Roberta Bergero Unknown protein Roberta and Vera Kaiser • Differential display – DD 44 (Moore et al. , 2003) • It is interesting to combine sequence divergence estimates with genetic maps

EST sequences were used to obtain sequences Intron positions of genes at low copy number were determined from the Arabidopsis thaliana and rice genomes PCR primers were designed to cross introns to find length variants to do genetics Parents F 1 progeny Y-linked 1830 bp 730 bp maternal X 2072 bp 700 bp 510 bp maternal X 590 bp paternal X X- and Y-linkage for locus Sl 6

Roberta’s ISVS method Forward primer Intron region Exon A Exon B Reverse primer FAM Incorporation of labeled universal primer after the first PCR cycles FAM For product sizes > 450 -500 bp, digest FAM with restriction enzyme Mbo. I FAM Mbo. I Hae. III Analysis by capillary electrophoresis

Evidence for X/Y linkage of the Sl. Cyp gene Intron 3 variants, showing Y-linkage of 438 bp band Intron 2 variants , showing X-linkage of 259 and 260 bp bands 2 male and 2 female F 1 plants Parents 260 Xm 257 Y 260 Xm 259 Xp 257 Y 447 bp 438 bp in males only 447 bp 259 Xp 260 Xm

3. WHEN did sex chromosome systems evolve, and when did recombination stop? • Some classical sex chromosomes are probably old – We don’t yet know how long it takes for the full set of features to evolve • It is often assumed that all other systems are young • but we need data. It is now possible to get evidence, using DNA sequences, estimating divergence between homologous X and Y sequences, and assuming a molecular clock – heteromorphism can evolve rapidly, e. g. by chromosome fusions • For most species, it is difficult to get the genes for such studies

(1) Mostly old part of X (all but 2 genes present in marsupial X chromosomes) (2) Few genes on the X are still detectable on the Y Xp (3) In contrast with Xq, many Xp genes still have detectable homologues on the Y Stratum 3 Stratum 2 LAHN & PAGE, 1999 Four evolutionary strata on the human X chromosome. Science 286: 964 -967. SKALETSKY et al. 2003. Nature 423: 825 - 837. Stratum 1 Strata 4 & 5 X-Y divergence, Ks X-Y divergence in humans Autosomal in marsupials (added to X and Y by transposition. Xq recent transposition PAR 1 2 genes transposed very recently to the Y

Strata are found in organisms other than humans Human Y) (X versus (Z versus W) (X versus Y) NOTE the different y axis scale (X versus Y) PAR 1 transposition Autosomal in Mostly old part of X marsupials (all but 2 genes (added to X present in and Y by marsupial X transposition) chromosomes) PAR inversion Lawson-Handley et al. , 2004 Genetics 167: 367 -376 Nam & Ellegren. 2008. Genetics 180: 1131 - 1136 PAR Bergero et al. , 2007 Genetics 175: 1945 -1954

Phylogenetic analysis of bird Z and W chromosomes also suggests that recombination between them stopped at different times Pseudoautosomal end Chicken Z Genes that stopped recombining after split of taxa Genes in region where ZW recombination stopped before split of major bird taxa Some bird taxa probably have small sex-determining regions from LAWSON-HANDLEY et al. , 2004 Genetics 167: 367 -376

Gradual evolution of bird sex chromosomes is also evident when different taxa are compared — some taxa have not undergone all the steps that others have taken Giemsa staining C-bands G-bands by Brd. U Painting with Locations of chicken Z markers probe Z W Non-recombining region has probably remained small Ostrich Large nonrecombining region Chicken Markers: Z chromosomes of both taxa share several markers Thus they probably had the same ancestral sex chromosome Recombination has been suppressed only in the chicken lineage (including other neognathae), and not in palaeognathous birds Nishida-Umehara et al. 2007 Chromosome Research 15: 721 -734 Nanda, I et al. . 2008. Cytogenet Genome Res 122: 150 -156.

Gradual evolution of snake sex chromosomes P. molurus (Pythonidae) Females are WZ E. quadrivirgata (Colubridae) Matsubara et al. (2006) PNAS 103: 18190 Many (11/11) genes shared between Z and W (small sexdetermining region) 3/11 genes No genes shared between Z and W (W has lost most genes) T. flavoviridis (Viperidae)

Ks values in 6 X and Y Marchantia polymorpha genes suggest that this sex chromosome system is old (Ks is uncorrected synonymous or silent site divergence) Ks • If we had a good molecular clock, we could translate Ks values into times when X-Y recombination stopped • It is not yet possible to tell whethere are strata in this plant, or if the Y and/or X is degenerated

Is papaya (with a small MSY) a young system? Divergence is low between papaya X and Y gene sequences X and Y from hermaphrodite (Yh, YU et al, 2007 Plant Journal 53: 124 -132) Ks values in 4 papaya genes from a BAC clone X and Y from male (YU et al, 2008 Tropical Plant Biology 1: 49 -57) It is not yet possible to tell whethere are strata in this plant, because only 2 BACs were sequenced (< 150 kb, whereas the size of the MSY is ~ 10 Mb)

Summary of question 3 • Some classical sex chromosomes are old • Even in such old sex chromosomes, recombination in most of the chromosometimes continued for long after the sex chromosome first evolved, and, in some species (but not all) later stopped in some regions – in mammals, birds and Silene latifolia • It is not yet clear whether species with small nonrecombining regions of their sex-determining chromosomes are always young sex chromosomes • Some of them could be sex-determining chromosomes with single gene control of gender, and therefore without selection for reducing recombination

WHY are there strata? Why doesn’t recombination just stop across the entire sex chromosome? 2 The simplest 2 gene evolutionary model above suggested that sex determining loci must initially be linked for separate sexes to evolve 1 3 M Su. Female “proto-Y” “proto-X” m f Other genes may be added to the system in a 3 rd step (and so on) “proto-Y” sexually antagonistic male-enhancer “Y” M 2 M Su. Female Reduced recombination between initial 2 genes

• The hypothesis of sexually antagonistic male-enhancers is plausible, but all evidence to date is indirect, and the only such genes yet identified are in the guppy – indeed many are wholly or partially sex-linked – without sexual antagonism, there should be no selective pressure converting hermaphrodites into males (the female functions of hermaphrodites could be maintained unchanged while male functions improve). • There may be molecular ways to test for antagonistic genes – in chicken and mouse, Mank et al. (2008 American Naturalist 171: 35 -43) searched for genes with different male and female expression patterns – many of these will NOT have antagonistic effects (they could just have sex-specific expression), but the set of such genes should include genes with antagonistic effects – They found that this set of are less likely to be expressed in multiple tissues (with the potential for conflicting selection pressures) than the genome average, even after excluding sexlinked genes; however, a difference in tissue-specificity could be explained without sexually antagonistic effects

• Drosophila experiments that allowed selection in males only show that female fitness indeed declines • This is consistent with a tradeoff between the sex functions, but it could be due just to stopping selection in females • Reversal in quality of progeny, depending on whether they had high or low fertility parents, is clear evidence for trade-offs, but it does not prove intralocus sexual conflict Fertility of female offspring High fertility female parents Low fertility female parents Fertility of male offspring Low fertility female parents High fertility female parents Low High Fertility of male parents Pischedda & Chippindale (2006, PLo. S Biology 4: e 356)

• The best evidence so far for sexually antagonistic male-enhancers is in the guppy fish, Poecilia reticulata – Guppy males are highly polymorphic for color patterns and their genetics has been studied analysis since 1927 – This fish has 23 pairs of chromosomes — 22 autosomal and one sexdetermining. Males are heterogametic (the sex determination mechanism is “XX/XY”, and the “YY” genotype is viable) – Almost all the genes determining guppy colour patterns (except for body color) are fully or partially sex-linked or sex limited (unlike what is found in other teleosts) • Winge, O. 1927. The location of eighteen genes in Lebistes reticulatus. Journal of Genetics 18, . A peculiar mode of inheritance and its cytological explanation. Journal of Genetics 12: 137. • With the possibility of using naturally occurring polymorphic sequence variants as genetic markers, it is now possible to make a more detailed genetic linkage map and find out if the Y has an excess of male attractiveness factors • Molecular markers have now been found on the Y chromosome, closely (but none fully) linked to the sex-determining region. – Shen et al (2007, Aquaculture 271: 178 -187), TRIPATHI et al. (2009, Proc. Royal Society B 276: 2195 -2208) – Overall, the results suggest that the non-recombining MSY region may not be very large, and that the colour variants may be controlled by polymorphic genes in the PAR

4. HOW did recombination stop, how do MSY regions expand? BUT the known inversions on the Y occurred relatively recently, and these cannot be involved in stopping recombination (since, in most of the Y, it stopped long ago) SRY/ SOX 3 Translocation • ‘. . there is little evidence demonstrating the importance of [chromosome rearrangements versus genes modifying recombination] in the evolution of X-Y crossover suppression’ (Bull 1983) • There are many inversions on the mammalian Y Lahn & Page’s suggested evolutionary history of the mammalian sex-chromosomes

Heterochromatic region X chromosomes Human Mammalian X chromosome gene arrangements are stable, while Y chromosomes are highly rearranged PAR 1 Yp BUT inversions occurred since humans (or split from chimpanzees PSA) and modifier genes can also change recombination rates during evolution NRY regions recently transposed from the X degenerated copies of X genes “Ampliconic” (duplications) degenerated X genes inverted PSA 2 Yq inverted Chimpanzee

• The papaya MSY (malespecific non-recombining) regions have been rearranged, even in just the two BACs so far studied • It is not known whether this region contains the sex-determining genes, or whether these inversions caused recombination to stop, but this region is only a small part of the MSY • The rearrangement in this region is shared by the Y and Yh, and differentiates both of them from the X • Thus it either dates from, or pre-dates, the evolution of hermaphrodites, suggesting that several events suppressing recombination have occurred X from hermaphrodite Y from hermaphrodite (Yh) 6. 5 Mb Y from male Y from hermaphrodite (Yh) 6. 5 Mb X from male X from hermaphrodite X-Y divergence (%) 13 - 19 2 -4 10 - 13 Yu et al. 2008. Tropical Plant Biology 1: 49 -57

The most recent strata in the human MSY already have several inversions • Stratum 5 may involve an inversion, but stratum 4 includes several inversions ( ) • Rearrangements may thus occur as a consequence of lack of recombination, as well as causing recombination to cease (Lemaitre et al. 2009 Footprints of Inversions at Present and Past Pseudoautosomal Boundaries in Human Sex Chromosomes. Gen Biol Evol. 1, 56 - 66) Stratum 5 PAR 1 Ross et al. 2005. Nature 434: 325 -337 Stratum 4 Stratum 3

5. Where are the sex-determining genes? Did X-Y recombination stop in S. latifolia due to inversions? Y chromosome deletion map of the Y, based on 3 parental plants A C B In one parent plant, the Y chromosomes gene Sl. Y 1 is in a different location In two parent plants the Y chromosomes gene Sl. Y 6 b is absent

Possible rearrangements in the Y, relative to the X Pseudo autosomal p arm Present X gene order Proto-Y 1 Present Y Alternative Y gene order RB 11 (no Y copy) RB 18 X M M m Suf q arm RB 18 Y Su. Female Paracentric inversion Pericentric inversion These results show that inversions happened after X-Y recombination first stopped in the region containing genes Sl. XY 3, 4, 7 and 6 a

• Accumulation of transposable element on Y chromosomes may promote rearrangements • Loss of a gene in primates: Nakayama & Ishida. 2006 Genome Res. 16: 485 -490. Cytogenetic maps of the threespine stickleback X and Y chromosomes, based on FISH with genes X Y • Rearrangements can obscure X-Y heteromorphism • Gene conversion between paralogs in the human Y: Bosch et al. 2004. Genome Res. 14: 835 -844. inversion X Y Heteromorphic X-Y pair deletion of part of Y inversion on Y and/or insertion, making heteromorphism hard to detect deletion Ross and Peichel. Genetics 2008; 179: 2173 -2182

6. Why does stopping recombination lead to sex chromosome degeneration, and how fast is it? • It has been known since 1918 that classical Y chromosomes are degenerated chromosomes – Muller, H. J. 1918. Genetic variability, twin hybrids and constant hybrids, in a case of balanced lethal factors. Genetics 3: 422 -499. • “It is probably needless to point out that the W and especially the Y chromosome …. . show the expected evidences of …. degeneration and differentiation from their homologues, both genetically and cytologically. The evidences are now as follows: – X-linked mutations affecting visible phenotypes are manifested in XY males • therefore the Y does not carry alleles that can cover up mutations – Infrequent dominant Y (and W) linked mutations” – “Great variations in their own size and shape even in closely related species” – Synaptic attraction between them and their homologues • “but the sex chromosomes in the heterozygous sex tend to remain condensed during the growth period, while the autosomes are spinning out for intimate conjugation, and there is frequently delayed synapsis” • “also lack of crossing over between them and their homologues, even …. where other chromosomes are undergoing crossing over”

Y chromosome degeneration • Loss of genes – Well illustrated by classical sex chromosomes • for example, the human X region that has been non-recombining longest has the lowest proportion of intact genes on the Y (at most, 5), whereas the probable number carried on the X chromosome is 734 (based on a count done by Gabriel Marais, using Ensembl version 47) • Worse gene function – amino acid substitutions that reduce functioning – less use of optimal codons – expression levels changed relative to X (presumably wrong levels) • Transposable element insertion is often included as an aspect of Y degeneration, and degeneration may indeed be caused partly by transposable element insertions, but we don’t actually know this – it is possible that these insertions are neutral – they could insert after genes or the sequences controlling their expression have degenerated

• NOTE that most of the genes present on the human Y are found in the youngest stratum of the X (strata 3 and 4 in the initial paper on strata) • This indicates that the older strata are genetically degenerated and have lost most of the genes that were once on the Y • Notice how helpful it is to have identified genes, not just anonymous markers or sequences LAHN & PAGE, 1999 Four evolutionary strata on the human X chromosome. Science 286: 964 -967

G C T A T Leaf X/Y 4 Flower Leaf X/Y 3 X/Y 7 Leaf Cyp. X/Y DD 44 Gene X/Y 1 Leaf Differences between X and Y homologues, estimated using PCR with primers recognising the same sequence in X and Y alleles Pyrosequencing Flower Little is known about degeneration in plants. Expression studies in S. latifolia give some direct evidence of low Y function C A pyrosequencing primer Sl. Cyp. Y AATTTGCACACCAACAAAGCATCACG Sl. Cyp. X AATTTGCACACCAACAAAGTATCACG Work of Michael Nicolas and Roberta Bergero

6. WHY are Y chromosomes degenerated? I. The ‘sheltering hypothesis’ (the Y is always heterozygous with an X) • It is a challenge to evolutionary biologists that a common observation such as degenerate Y chromosomes is still so far from being understood’ (Bull 1983, p. 258) • “The reason for this rapid decay of things Y-chromosomal is thought to be quite simple: once the Y chromosome became sex-determining, its presence was limited to the heterogametic sex (in our case, males). Because the Y chromosome was never found in the absence of an X chromosome, there was presumably little selection against the mutational inactivation of those genes on the Y chromosome that were also present on the X chromosome. Thus, over evolutionary time, the Y chromosome gradually lost most of its functional genes by the accumulation of deleterious mutations, resulting in that little dab of male-determining chromatin that we have today. ” – HAWLEY, R. , 2003 The human Y chromosome: Rumors of its death exaggerated. Cell 113: 825 -828. • have been greatly This is wrong — it ignores the central importance of the lack of recombination

Models of non-recombining genomes have largely answered Bull’s challenge There are now several other theories for degeneration when there is no recombination

Selection for advantageous mutations causes fixation of deleterious mutations , which reduces the effective population size Many different sequences, one carrying the deleterious mutation Only a single sequence, i. e. all carry the deleterious mutation several generations Deleterious mutations prevent spread of advantageous mutations unless their selective advantage is large several generations

Muller’s ratchet (probably less important) 10 different sequences Genotype number 1 2 3 4 5 6 7 8 9 Selection against mutations reduces the effective size After many generations with stable numbers of deleterious mutations 4 ancestral sequences Ancestral genotype number 1 4 7 Loss of mutant-free class 8

Summary of question 6: WHY are Y chromosomes degenerated? • Hitch-hiking processes in non-recombining regions lower the effectiveness of natural selection • There are several molecular evolutionary approaches that allow us to detect this • I will next describe transposable element accumulation and some sequence analyses

A sign of low effectiveness of selection is accumulation of repetitive DNA on Y chromosomes Autosomes Accumulation of retrotransposons on the Drosophila miranda neo-Y chromosome BACHTROG, D. , 2003. Mol. Biol. Evol. 20: 173 -181.

Drosophila miranda Transposable elements Genes The neo-Y is turning into heterochromatin

TE insertions do NOT necessarily cause loss of function Two Silene latifolia Y genes DD 44 Y S. latifolia X Work of Gabriel Marais Blastn Y 3 / Y 3 Blastn Genbank Repeat. Masker (Repbase) Blastx prot TEs Arabidopsis S. vulgaris (not sex-linked) Sl. XY 3 Y X Introns LTR retrotransposon Exons Non-LTR retrotransposon DNA transposon Inverted repeats Direct repeats

Active MITE insertions were detected by searching for polymorphic inserts in introns of Silene latifolia genes EITRI: 11 -bp terminal inverted repeats (5'-CTAGGTAGCAC-3') and 8 -bp target site duplications (TSDs, like h. AT or P elemenst) A Tourist-like element M Roberta Bergero & DC, in press in Genetics GAAATTCTTT//Sl-To 1//TAGTTTC GAAATT-------TAGTTTC GAACTTCTTC-----------AGTTTC

Silene latifolia Not Y-linked (none fixed) Genetic mapping allows us to find ones that are Y-linked As predicted from population genetics theory, MITE insertions are generally at low frequencies, but on the Y chromosome they reach high frequencies, perhaps in part because there are few genes Y-linked (3/25 fixed, others at high frequencies)

Data from non-degenerated genes can provide indirect evidence that degeneration is occurring • Degeneration is thought to be caused by lack of recombination – This changes evolutionary processes in several ways – The different processes all cause lower “effective population size” and thus they lead to low genetic variation • Diversity studies can thus detect these processes – If degeneration is happening, we should find lower diversity of Ylinked than X-linked genes – We must take into account that the population size of X-linked genes is 3 time higher than for Y-linked genes

• S. latifolia Y diversity is low compared with homologous X-linked genes – significant difference in diversity compared with the autosomal genes (by an HKA test) • Diversity of autosomal and X-linked genes is very variable, and can be very high – No significant difference by an HKA test Y-linked X-linked or autosomal Nucleotide diversity values (silent sites) To estimate subdivision for Y and X genes within S. latifolia, we sampled plants from 23 European populations, and sequenced Y and X alleles

• What do we expect for the S. latifolia Y chromosome? – X chromosome = 1/7 of genome – Assume that the S. latifolia genome contains 21, 000 genes (a conservative estimate for a plant), we get ~ 3, 000 genes for the Y if it has lost no genes, or perhaps 2, 000 is the PAR has 1/3 of the genes (it probably has many fewer than this), i. e. 25 times more than the dot chromosome Diversity relative to the neutral value • This predicts a much greater reduction in diversity that the estimated value, and therefore suggests that the Y may have lost many genes 1. 0 0. 1 0. 001 D. melanogaster 4 th (dot) chromosome (80 genes) 0. 0001 Predicted with moderately deleterious mutations Predicted with moderately and slightly deleterious mutations and no recombination

Molecular evolutionary comparisons of X, Y and outgroup sequences allows one to test whether the X or Y has changed, again using non-degenerated genes X chromosome. Y chromosome G A Outgroup species A A G substitution Ancestral sequence

Inferring the causes of degeneration One can infer maladaptation from divergence • If the neo-Y genes are acquiring harmful mutations that impair their functionality, we expect to find more changes in functionally significant sequences in the ancestry of the neo-Y copies than the neo-X copies (e. g. more non-synonymous substitutions). The opposite is true if the neo-X genes experience adapt more than the neo-Y We can test this by sequence comparisons X-Y or along the 2 lineages Y X outgroup species

The neo-sex chromosomes in D. miranda are a model system for studying sequence divergence across many genes A consistent pattern of higher Ka than Ks for Y than X sequence divergence across many genes in D. miranda is difficult to explain by molecular adaptation of the Y (it would be very strange if all Y genes were adapting) Bachtrog, D. 2005. Genome Research 15: 1393 -1401

Similar results have been found in birds Chicken Z Turkey Z Chicken W Turkey W The W chromosome (restricted to females) has relatively more non-synonymous substitutions (higher Ka/Ks) than the Z (but a caveat is that the outgroup is ~VERY distant) Berlin & Ellegren. 2006. Journal of Molecular Evolution 62: 66 -72

In Silene latifolia genes, Gabriel Marais sees both failure of selection to prevent deleterious substitutions, and favorable changes switching to neutral Gene (Numbers of codons analyzed) Site model indicates purifying selection at all loci Branch model X versus Y d. N/d. SX d. N/d. SY Significance Branch-site analysis suggests weak efficacy of selection on Y % of codons suggesting degeneration, and changes in Y versus X and outgroup (to neutral evolution or positive selection) Sl. X 1/Y 1 (458) 100% <<1 10 -4 0. 11 X < Y*** 6% No significant switching Sl. Cyp. X/Y (519) 94% <<1 6% w~1 0. 14 X=Y 10% No significant switching Slss. X/Y (259) 97% <<1 3% >1 0. 18 0. 23 X < Y ns 0 No significant switching 63% <<1 36% ~1, 1% > 1 0. 13 0. 90 d. NX=d. NY=d. SY (all low values) 5. 5% under positive selection ( =14. 8) Significant switching to positive selection Sl. X 3/Y 3 (318) 95% <<1 5% ~1 0. 04 0. 13 X < Y* 4% under positive selection ( =3. 5) Significant switching to positive selection Sl. X 7/Y 7 (246) 88% <<1 12% ~1 0. 08 0. 11 X < Y ns 4% No significant switching Sl. X 4/Y 4 (362) 92% <<1 6% ~ 1, 2% > 1 0. 11 0. 25 X < Y* 14% Significant switching to neutrality DD 44 X/Y (217) Significance of differences in LR tests: ns = non-significant, * p < 0. 05, *** p < 0. 0005

The Chimpanzee Sequencing and Analysis Consortium 2005 Nature 437: 69 -87. Number of 1 MB windows Human-chimpanzee divergence Over all sites, genes on the mammalian X evolve more slowly than autosomal genes (indicating selective constraints, which are more important on the X because of hemizygosity in males) and the Y evolves unusually fast (because there is a higher mutation rate in males, due to multiple cell divisions in spermatogenesis) Autosomes X Human-chimpanzee divergence Y X

• • The PAR recombines at a much higher frequency than the rest of the X, increasingly so as its size is restricted by evolution of new strata High recombination may also cause a high mutation rate (if recombination causes mutations) The PSA is smaller in humans than other mammals, i. e. the boundary has moved towards the tip, compare with its location in bovine X chromosomes and in mice it is small due to movement of genes off the X Van Laere A. et. al. Genome Res. 2008; 18: 1884 -1895

Male sterility, m Female fertility Summary of steps in the evolution of the Y Maleness factor, M Female suppressor, Su. F Proto-X and Y 1 Addition of male function genes, further recombination suppression, rearrangements Loss of parts of Y M 2 small MSY region MSY Evolution of a sex. Su. F determining region in M Suppressed an ancestral recombination on chromosome pair, part of proto-Y forming a non. Transposable element heteromorphic pair 2 The simplest evolutionary model suggests that males and females evolved from hermaphrodites by loss of functions Newest Su. F stratum M Oldest stratum PAR accumulation and Some plants, expanded MSY fish, snakes region 3 4 m M 2 Y X Classical (humans and Drosophila) 5

Adaptations during sex chromosome evolution • (1) In species with XY systems – The Y chromosome may acquire genes with male functions – Genes with male functions may also evolve more readily on the X than the autosomes, because the X spends a higher proportion of its evolutionary history in males than females • VICOSO & CHARLESWORTH, 2006 Evolution on the X chromosome: unusual patterns and processes. Nature Reviews Genetics 7: 645 -653; POTRZEBOWSKI et al. 2009 The emergence of new genes on the young therian X. Trends in Genetics 26: 1 -4.

• (2) If degeneration has occurred, then, at most X-linked loci XY males have only one gene copy, compared with XX females’ two copies – except for a few housekeeping genes on both X and Y – These changes must involve sexually antagonistic effects – The simplest solution is over-expression of the X in males, as in Drosophila 2 X 2 A Females Mammals low expression of both Xs • • Some species have evolved control of levels of expression so that levels of X-linked gene products are correct, relative to expression of other genes: dosage compensation The X must therefore undergo considerable evolutionary change The compensation mechanisms are different in different organisms inactivation of one X • 2 X 2 A XO 2 A Females Males Drosophila “Females” Males C. elegans

7. What about other non-recombining systems? • WHY did recombination stop, and what are the consequences, e. g. do these genome regions degenerate? • Some fungal incompatibility systems – Different loci (pheromones and receptors) are involved at some of these “loci”, and are sometimes present in inverted regions, but the selective reason for lack of recombination is unknown • Lee et al. 1999. The mating-type and pathogenicity locus of the fungus Ustilago hordei spans a 500 -kb region. PNAS 96: 15026 -15031. – Some seem to have undergone degeneration. Allelic forms of the genome region each lack genes found in the other orthologous region – TEs are sometimes abundant (50% in U. hordei, maybe accumulated, but comparisons with other genome regions should be made) • Bakkeren et al. 2006. Fungal Genetics and Biology 43: 655 -666 • Angiosperm self-incompatibility systems (SI, S-allele systems) – Alleles at two different loci (encoding pollen and pistil function proteins) must be present in the correct combinations, which would lead to selection against recombination, but it is not yet certain whether these regions do have unusually low recombination – Studies testing whether homozygotes for the same allelic form at these “loci” are disfavoured (suggesting “linked load”) are not yet conclusive

Primrose Distyly Primrose Short -styled Long-styled • The different flower characters are controlled by at east 3 closely linked genes • This is called a ‘supergene” • Style length • Anther position • Pollen compatibility type Buckwheat

• Even if the S-locus region has low recombination, it may be too small (contain too few genes to drive the processes described earlier) to undergo genetic degeneration • The possibilities were reviewed by Uyenoyama, M. K. 2005. Evolution under tight linkage to mating type. New Phytologist 165: 63 -70. • Some empirical tests have suggested that individuals identical by descent for S alleles may have low survival • BUT it is difficult to test more than a few alleles – one has to compare homozygotes and heterozygotes, matching their inbreeding coefficients, I. e. ensuring that both sets are noninbred (otherwise inbreeding depression might be the cause of lower survival of homozygotes) – it is also hard to rule out an early effect of the incompatibility (which might slow the growth of the pollen that would generate homozygotes, and might lead the maternal plant to abort those zygotes)

• Fungal and algal mating type loci are NOT sex chromosomes, but they show some very similar evolutionary behaviour • In Neurospora tetrasperma, there is a non-recombining region whose function is to link the incompatibility gene region to the centromere, guaranteeing first division segregation and thus compatibility among pairs of meiotic products (a mechanism for selffertilisation) • This region has expanded – Divergence between genes in the two haplotypes varies in a pattern like the sex chromosome strata • Menkis et al. 2008. The mating-type chromosome in the filamentous ascomycete Neurospora tetrasperma represents a model for early evolution of sex chromosomes. PLo. S Genetics 4 – The expansion is due to an inversion and the effects of modifier genes • Jacobson 2005. Blocked recombination along the mating-type chromosomes of Neurospora tetrasperma involves both structural heterozygosity and autosomal genes. Genetics 171: 839 -843. • The mating-type loci of Cryptococcus and Microbotryum fungi (also self-fertilising) may be similar – FRASER et al. , 2004 PLo. S Biology 2: 2243 -2255, Votintseva & Filatov, 2009 Genetics May 2009.

This is like a neosex chromosome system — a part was added recently NOTE: incorrect use of the word ‘rate’ (they mean ‘divergence’)