Patterns of Molecular Evolution of an Avian Neo-sex Chromosome

Abstract

Newer parts of sex chromosomes, neo-sex chromosomes, offer unique possibilities for studying gene degeneration and sequence evolution in response to loss of recombination and population size decrease. We have recently described a neo-sex chromosome system in Sylvioidea passerines that has resulted from a fusion between the first half (10 Mb) of chromosome 4a and the ancestral sex chromosomes. In this study, we report the results of molecular analyses of neo-Z and neo-W gametologs and intronic parts of neo-Z and autosomal genes on the second half of chromosome 4a in three species within different Sylvioidea lineages (Acrocephalidea, Timaliidae, and Alaudidae). In line with hypotheses of neo-sex chromosome evolution, we observe 1) lower genetic diversity of neo-Z genes compared with autosomal genes, 2) moderate synonymous and weak nonsynonymous sequence divergence between neo-Z and neo-W gametologs, and 3) lower GC content on neo-W than neo-Z gametologs. Phylogenetic reconstruction of eight neo-Z and neo-W gametologs suggests that recombination continued after the split of Alaudidae from the rest of the Sylvioidea lineages (i.e., after ∼42.2 Ma) and with some exceptions also after the split of Acrocephalidea and Timaliidae (i.e., after ∼39.4 Ma). The Sylvioidea neo-sex chromosome shares classical evolutionary features with the ancestral sex chromosomes but, as expected from its more recent origin, shows weaker divergence between gametologs.

Introduction

Degeneration and Genetic Variation of Sex-Linked Genes

Sex chromosomes have evolved independently in many different lineages and some common key features in the evolution from an ancestral homologous pair have been recognized. These include reduced recombination between the members of the sex chromosome and a progressive degeneration of the sex-limited chromosome (i.e., Y in mammals and W in birds) (Bergero and Charlesworth 2009). According to the sexual antagonistic model, recombination cessation is triggered by the presence of sexual antagonistic alleles (beneficial to one sex and harmful to the other) associated to a sex-determining factor at one of the sex chromosomes (Rice 1987). Once multiple loci are involved in the establishment and functionality of one of the sexes, reduced recombination between the two chromosomes will be selected for. As a consequence, the nonrecombining sex chromosome is predicted to accumulate deleterious mutations, start degenerating, and progressively diverge relative to the other member of the sex chromosome pair (Lahn and Page 1999; Bachtrog 2005; Wilson and Makova 2009). The degeneration of the nonrecombining sex chromosome is expected to be driven by multiple processes, including Muller’s ratchet, background selection, and the Hill–Robertson effect with weak selection (Lahn and Page 1999; Charlesworth et al. 2005; Bergero and Charlesworth 2009; Wilson and Makova 2009).

The timing of recombination cessation may differ along the sex chromosome, from the initiation points to the regions where recombination still occurs, and this regional variation will be evident in the degree of sequence divergence of gametologous genes (i.e., homologous genes on the opposite sex chromosomes; sensu Garcia-Moreno and Mindell 2000). By analyzing divergence between gametologs, regions that have ceased to recombine at different times, corresponding to clusters of distinct divergence, evolutionary strata, were initially identified in the human X chromosome (Lahn and Page 1999; Ross et al. 2005). More recently, such evolutionary strata, although not always in a linear fashion over the chromosome, have also been reported in other mammals (Sandstedt and Tucker 2004; but see Murtagh et al. 2012), plants (Nicolas et al. 2004; Bergero et al. 2007), and birds (Handley et al. 2004; Nam and Ellegren 2008), that is, in both male and female heterogametic systems. This suggests that similar processes of chromosome evolution occur in both types of sex chromosome systems (Charlesworth and Mank 2010). The presence of evolutionary strata has been related to two main events in chromosome evolution: the suppression of recombination as a means to resolve sexual antagonism between clusters of genes located on sex chromosomes (van Doorn and Kirkpatrick 2007) and the occurrence of chromosomal rearrangements that can affect the physical linearity of the blocks of reduced recombination (Lemaitre et al. 2009).

In birds, the evolution of gametologous genes and the occurrence of evolutionary strata on sex chromosomes have been almost exclusively addressed in the chicken (Gallus gallus) (Ceplitis and Ellegren 2004; Handley et al. 2004; Nam and Ellegren 2008). In other bird groups, mainly in passerines (perching birds or song birds), studies have focused on the evolution of genes located on the Z chromosome (Borge et al. 2005), the putative role of this chromosome in reproductive isolation (Storchová et al. 2010), and prominently in the comparative analysis of the organization of genes located on sex chromosomes between different species (Itoh et al. 2006; Nanda et al. 2008; Völker et al. 2010). The general scarcity of information about gametologs in birds results mostly from a highly heterochromatic W chromosome, in which considerable divergence in gene content and sequence variation makes it difficult to isolate Z and W gene pairs within nonrecombining areas. To our knowledge, even in model species such as the zebra finch (Taeniopygia guttata), structure and expression patterns of Z/W gametologs have only been investigated for the CHD1 gene pair (Fridolfsson and Ellegren 2000; Agate et al. 2004). In fact, the high divergence between the sex chromosomes or more specifically the degeneration of the sex-limited chromosome constitutes a major drawback of using chromosomes with very deep divergence times for assessing patterns of sex chromosome evolution. Very often in ancestral sex chromosome pairs, such as the mammalian XY (approximately 150–170 Ma; Veyrunes et al. 2008; Livernois et al. 2012) or the avian ZW (an at least 150-My-old sex chromosome pair; Nam and Ellegren 2008; Livernois et al. 2012), direct comparisons of gametologous genes are no longer possible. Even the 80–130-My-old X- and Y-added region in eutherian mammals has started to deteriorate (Waters et al. 2001; Livernois et al. 2012).

New sex chromosome elements, neo-sex chromosomes, can be formed either by translocations of autosomal parts to the sex chromosomes (Bachtrog 2003; Zhou et al. 2008; Howell et al. 2009; Kitano et al. 2009) or through the translocation of a master sex-determination locus (Tanaka et al. 2007). Neo-sex chromosomes offer unique possibilities to studying the evolution of sex-linked genes because degeneration of the sex-limited chromosome is still limited. If, as hypothesized, reduced recombination and degeneration of the sex-limited chromosome are key factors in the evolution of most sex chromosome systems, then these features should be evident on genes located on neo-sex chromosomes. Neo-sex chromosome evolution has been intensively studied in some model systems, for example, in Drosophila miranda where the relatively recently established sex-limited neo-Y chromosome shows clear signs of degradation in the form of an excess of nonsynonymous mutations on neo-Y-linked genes (Bartolomé and Charlesworth 2006). One of the rare examples of neo-sex chromosomes in mammals is found in the black muntjac (Muntiacus crinifrons) where a 0.5-My-old neo-Y alleles have accumulated mutations indicating recombination cessation and ongoing neo-Y degeneration. In contrast, fish seem particularly prone to chromosome rearrangements involving sex chromosomes (Volff 2005; Mank and Avise 2009), and several cases of neo-sex chromosomes are described (Kitano et al. 2009; Ross et al. 2009). An interesting case is the neo-sex chromosome in the threespine stickleback (Gasterosteus aculeatus) where newly sex-linked genes on the neo-sex chromosome are involved in mate recognition, reproductive isolation, and speciation (Kitano et al. 2009, 2010). However, in birds, a very prominent group in the context of female heterogamety, the opportunity of using neo-sex chromosomes as basis for the study of sex chromosome evolution has not yet been available.

Neo-sex Chromosome in Sylvioidea Passerines

We have recently detected a neo-sex chromosome in Sylvioidea passerine birds (Pala et al. 2011). A significant part (ca. 10 Mb) of chromosome 4a (in total 20.7 Mb), an autosome in the zebra finch genome (Warren et al. 2010), has been translocated to the sex chromosomes in the Sylvioidea lineage (including, e.g., warblers and larks; fig. 1). The translocated material is present in all Sylvioidea assayed so far but not in non-Sylvioidea species, which suggest that the translocation occurred at the base of the Sylvioidea branch of the avian phylogeny, at approximately 42.2 Ma (Pala et al. 2011). Our previous results indicated the presence of both Z- and W-linked copies of the neo-sex chromosome in Sylvioidea and show no indications of ongoing recombination (Pala et al. 2011), which offers the possibility of isolating a number of these more recently formed gametologous genes.

The Sylvioidea neo-sex chromosome opens the unprecedented possibility of studying the evolutionary trajectories of sex-linked genes and addressing the question of what initial factors affect sex chromosome evolution. In vertebrates, this opportunity has seldom arisen (but see, e.g., Zhou et al. 2008; Ross et al. 2009) and never before in a female heterogametic system. Although with obvious differences, the two main sex chromosome systems (XY and ZW) have evolved from autosomes, with specific evolutionary processes modulating the progressive structural and functional divergence of each member of the pair (Fridolfsson et al. 1998; Graves 1998; Bergero and Charlesworth 2009). The Sylvioidea neo-sex chromosomes, with their autosomal origin and verified linkage to ancestral Z, can be looked upon as an emergent sex chromosome pair, in which divergence might still be ongoing (Pala et al. 2011). Interestingly, chromosome 4a is orthologous to the X chromosome in mammals (International Chicken Genome Sequencing Consortium 2004; Warren et al. 2010) and holds genes for which sex linkage may be favorable (e.g., the androgen receptor gene; Pala et al. 2011), a notion that may help to explain the establishment of the neo-sex chromosome.

In this work, we evaluate the initial signatures of sex chromosome evolution by analyzing silent and functional sequence divergence of gametologs on the neo-sex chromosome and by searching for patterns that would hint to a progressive loss of recombination over the neo-sex chromosome. We sequenced genes located on the neo-sex chromosome and on the autosomal part of chromosome 4a in three Sylvioidea lineages: Acrocephalidea (great reed warbler, Acrocephalus arundinaceus); Timaliidae (common whitethroat, Sylvia communis), and Alaudidae (skylark, Alauda arvensis). To test hypotheses of gene degeneration and sequence evolution in response to loss of recombination and decrease in population size on neo-sex chromosomes, we evaluated 1) the degree of silent and functional sequence divergence between neo-Z and the neo-W gametologs, 2) the difference in GC content between the neo-Z and neo-W gametologs (reflecting recombination, gene degradation, and thus gene density; Meunier and Duret 2004), and 3) the amount of genetic diversity of neo-Z and autosomal genes (reflecting effective population sizes; Pool and Nielsen 2007). The results from the neo-sex chromosome were compared with those of the ancestral sex chromosome. Furthermore, we used a phylogenetic approach to evaluate whether or not recombination cessation between neo-Z and neo-W chromosomes preceded speciation using data from eight neo-sex gametologs in the three Sylvioidea species.

Our results indicate a modified evolutionary trajectory of loci upon their linkage to the sex chromosomes, and we detect chromosome regions of higher divergence, which potentially correspond areas on the neo-sex chromosome where recombination cessation started. Different levels of silent divergence are suggestive of the occurrence of a nonlinear process of recombination arrest along the neo-sex chromosomes, resembling the well-described evolutionary strata of ancestral chromosome systems (Handley et al. 2004; Nam and Ellegren 2008). The potential role of sexual antagonism in modulating the process of recombination cessation (Rice 1987) is also addressed, as the neo-sex chromosome harbors the androgen receptor gene, a locus involved in sex determination and with potential selective advantage for males (cf. Pala et al. 2011). The sex-limited neo-sex chromosome was overall only moderately degenerated, and many of the gametologs were still showing high amino acid conservation.

We discuss our results in light of central questions regarding the evolution of sex chromosomes and the fate of sex-linked genes: Will recombination cessation start at specific points in the neo-sex chromosome (e.g., from the fusion area or around genes with sexual antagonistic value) (Rice 1987; Lahn and Page 1999; Charlesworth and Mank 2010)? Will autosomal loci put into a sex chromosomal context evolve according to the neutral expectations of their new inheritance mode or will they progressively diverge due to actions of specific selective pressures, distinct from those that affect autosomes (Lahn and Page 1999; Bachtrog 2005; Charlesworth et al. 2005; Bergero and Charlesworth 2009; Wilson and Makova 2009)?

Materials and Methods

Samples and Marker Isolation

DNA from individuals of great reed warbler and common whitethroat were obtained from the Molecular Ecology and Evolution Lab, Department of Biology, Lund University, Sweden. Skylark samples were provided by P. Zehtindjiev, Institute of Zoology, Bulgarian Academy of Sciences, Sofia, Bulgaria. Samples were sexed as described in Pala et al. (2011).

Three coding sequence (exon) markers (ZNF711; P2RY4; and ZC3H12B) were selected from the subset of sex-linked loci reported in Pala et al. (2011). Additionally, we designed primers to amplify exons at another five loci along the neo-sex chromosome (table 1; and supplementary table S1, Supplementary Material online). This was done in the Primer3 design module in Geneious (Rozen and Skaletsky 2000; Drummond 2010) based on sequences of genes located on chromosome 4a in the zebra finch genome assembly (build taeGut3.2.4 at Ensembl, www.ensembl.org; Warren et al. 2010).

For the three exon markers isolated in Pala et al. (2011), reference polymorphisms for Z- and W-linked copies had already been assessed and the presence of both gametolog copies confirmed in females. For these loci, the amplified polymerase chain reaction (PCR) product of one female of each species was cloned into PCR 2.1 TOPO vector (Invitrogen). Colonies were screened by blue/white ampicillin selection, and inserts were amplified with M13F and M13R universal primers. We sequenced a minimum of eight clones of the expected size per individual and per locus using BigDye Terminator Sequencing Kit (Applied Biosystems) according to manufacturer’s recommendations.

For the remaining five exon markers, each locus was initially amplified from three females and two males of each species. The identity of the sequences was confirmed through Basic Local Alignment Search Tool search. Variation among sequences was identified using Geneious (Drummond 2010). We assessed the presence of sex-specific polymorphism and occurrence of heteromorphic sequence patterns unique to the females (suggesting the presence of both Z- and W-linked alleles) through comparative analysis of male and female sequences, as described in detail in Pala et al. (2011). Briefly, if the amplified markers are linked to the Z chromosome, males should be either homozygous or heterozygous (corresponding to the amplification of the two Z-linked alleles). At male-specific variable positions, females should present a single sequence (corresponding to the single Z-linked allele), while being heterozygous at positions in which no variability should be observed in males (corresponding to the female-specific W gametolog). Cloning of the PCR product of one female per species and locus was conducted as described earlier.

Ancestral Gametologs

Available primers to amplify exons of ancestral gametologs were tested, and additional ones were designed (table 2 and supplementary table S1, Supplementary Material online). Whenever possible, we used zebra finch sequences as basis for primer design for the ancestral gametologs. In most cases (with the exception of the CHD1 gene) only Z sequences were available in zebra finch, thus Z and W sequences of these genes, extracted from the chicken genome (build WASHUC2.1), were additionally used as reference.

Sequence Analysis

The cloned sequences obtained for each locus/species combination were aligned in Geneious (Drummond 2010), where we also performed the analysis of sequence variation. The cloned sequences were compared with the direct sequencing results for the same female (to assess the representation of the two gametologous sequences among clones). Additionally, the two sequence types were classified as Z-linked and W-linked based on the comparison with male and female sequences for the same locus: Z sequences should present a similar polymorphism landscape as the one observed in males, whereas all the female-specific polymorphisms observed through direct sequencing should be evident in the sequence clones classified as W.

To estimate the Z-W divergence in coding sequences, we calculated synonymous (Ks) and nonsynonymous (Ka) substitution rates according to the Nei–Gojobori method (Nei and Gojobori 1986) using DnaSP (Librado and Rozas 2009). GC content at the third positions of all loci was computed using DnaSP (Librado and Rozas 2009).

To evaluate whether or not recombination cessation between neo-Z and neo-W chromosomes preceded speciation, we used a phylogenetic approach. Phylogenetic analysis of both ancestral (three loci) and neo-sex gametologs (eight loci) in the three Sylvioidea species was performed using the neighbor-joining and maximum likelihood methods as implemented in MEGA version 5 (Tamura et al. 2011) If cessation of recombination continued after the speciation events, we expected the gametologs to group according to species (i.e., the Z and W gametologs of species A would be more similar to each other than to the gametologs of species B). We tested different nucleotide substitution models, including the Jukes–Cantor model (Jukes and Cantor 1969) (assuming equilibrium of base frequencies and equal nucleotide substitution rates) and the Tamura–Nei (Tamura and Nei 1993) and Hasegawa–Kishino–Yano models (Hasegawa et al. 1985) (assuming variable transition and transversion rates). We used zebra finch sequences for each locus as outgroups for phylogenetic reconstruction. Support for the observed topologies was generated by performing 1,000 bootstrap replicates.

Intron Analysis

Five males (ZZ) of each species were used in the amplification of 14 intron markers (one locus failed to amplify in great reed warblers and common whitethroat, respectively; see later), distributed along zebra finch chromosome 4a (covering both the neo-sex and the autosomal part of 4a) (supplementary table S2, Supplementary Material online). Primers listed in Pala et al. (2011) were used for amplification, and locus FMR1 (FMR1-F, 5’-RCATGAAGATTCAATAACAGTTKCAT-3’ and FMR1-R, 5′-ATTTGTCTCTCTGGTTGCCA-3’) was additionally included. We exclusively targeted the neo-Z chromosome; thus, sequenced males only. The PCR products were subjected to direct sequencing, sequences were aligned per locus, and only the regions common to all species were considered in the analysis. Diploid sequences were haplotyped using the algorithms provided by PHASE as implemented in DnaSP (Librado and Rozas 2009). Intraspecific estimates of divergence, including nucleotide diversity (π), were calculated using DnaSP (Librado and Rozas 2009). Alignment gaps were excluded from the analysis.

To directly compare the polymorphism levels between the sex-linked part and the autosomal region of chromosome 4a, we performed a two-locus Hudson–Kreitman–Aguadé (HKA) direct test (Begun and Aquadro 1991) in DnaSP (Librado and Rozas 2009). According to the predictions of this test, regions of the genome that evolve at a higher rate should present a higher level of polymorphism within species. Genes located on different chromosomes are expected to differ in genetic diversity, because of the difference in population size between sex chromosomes relative to autosomes: the Z chromosome will have a population size of three-fourths of that of autosomes. The direct mode of the HKA test will take this population size difference into account and consider the estimation of the nucleotide diversity level as 4 N_e (N_e = effective population size) for autosomal loci and 3 N_e for Z-linked genes.

In each of the three species, we also tested the degree of interspecific divergence in relation to inheritance (autosomal vs. sex linkage) by comparing six introns within the sex-linked area (comprising approximately 2,500–2,900 bp) and six introns (five in the skylark) within the autosomal part of 4a (comprising approximately 1,350–1,970 bp). Intra- and interspecific divergence levels were calculated independently for the two regions, through a HKA test (Hudson et al. 1987) and subsequently used as a basis for the direct test calculations. A single sequence of each species was used for the estimation of interspecific polymorphism levels relative to the other two species. The individual contributions of each locus were assessed through a multilocus HKA test (http://genfaculty.rutgers.edu/hey/home).

Results

Amplification Success of Gametologs on the Neo-sex and the Ancestral Chromosomes and of Introns on Neo-sex and Autosomal Loci on Chromosome 4a

A total of eight pairs of gametologous genes located within the sex-linked area of 4a were sequenced in the great reed warbler, the common whitethroat, and the skylark (table 1 and supplementary table S3, Supplementary Material online). We identified the Z- and W-specific copies by cloning and sequencing the female PCR products and by comparing those sequences with the Z- and W-specific polymorphisms identified in male and female sequences obtained through direct sequencing. We managed to isolate both gametologs in all species at five loci but had difficulties in doing so at the three loci located within the first (i.e., telomeric) 3 Mb of the sex-linked part of 4a. For one locus each in the great reed warbler (DIAPH2) and the skylark (PCDH11), and for three loci in the common whitethroat (PCDH19, DIAPH2 and PCDH11), the W-linked copy was not amplified (table 1 and supplementary table S3, Supplementary Material online).

We additionally attempted to amplify and sequence exons of four gametologs on the ancestral sex chromosomes and successfully isolated both gametologs at four loci in the common whitethroat and at three loci in the great reed warbler and in the skylark (table 2 and supplementary table S4, Supplementary Material online).

Finally, we amplified 14 intron markers along chromosome 4a, 7 loci on the neo-sex chromosome, and 7 on the autosomal part (supplementary table S2, Supplementary Material online; Pala et al. 2011). Here, we were targeting the neo-Z gametolog, and therefore, we sequenced males only. Most loci were successfully amplified and sequenced, with the exception of one locus each in the great reed warbler and the common whitethroat, respectively (supplementary table S2, Supplementary Material online).

Nucleotide Diversity and Interspecific Divergence at Neo-Z and Autosomal Introns

The comparison between neo-Z and autosomal introns on chromosome 4a (fig. 2) revealed that the nucleotide diversity was on average lower among the neo-Z sequences in all three species (mean ± standard error [SE]: 0.002 ± 0.001 in the great reed warbler, 0.014 ± 0.003 in the common whitethroat, and 0.010 ± 0.005 in the skylark) than over the autosomal part of 4a (mean ± SE: 0.013 ± 0.004 in the great reed warbler, 0.017 ± 0.004 in the common whitethroat, and 0.024 ± 0.009 in the skylark). The difference was statistically significant in the great reed warbler (P = 0.04, two-tailed t-test) and nearly significant in the skylark (P = 0.06). The lower level of diversity among Z-linked intronic loci is in line with the hypothesis of loss of variation due to the reduced effective population size (three-fourths of autosomal genes; Pool and Nielsen 2007). The observed neo-sex:autosome nucleotide diversity ratio based on the values above is 0.15 in the great reed warbler, 0.82 in the common whitethroat, and 0.42 in skylark.

The direct mode HKA tests used to assess the role of selection for the degree of divergence of neo-Z and autosomal sequences performed over the pooled loci were nonsignificant in all sex-linked/autosomal comparisons (P > 0.05). The multilocus HKA test over individual loci also produced nonsignificant results (P > 0.05) for all species, with exception of the common whitethroat data set when using the great reed warbler for the interspecific comparisons (P = 0.02).

Silent and Functional Divergence between Gametologs, and GC Content at the Neo-sex and Ancestral Chromosomes

The synonymous substitution rate (Ks) of gametologs at the neo-sex chromosome ranged between 0.03 and 0.34 in great reed warbler, 0.01 and 0.12 in common whitethroat, and 0.04 and 0.25 in the skylark (fig. 3A). This was in general lower than the Ks values observed at the ancestral gametologs (0.24–0.38 in the great reed warbler; 0.20–0.34 in the common whitethroat, and 0.20–0.55 in the skylark) (fig. 3A); significantly so for the great reed warbler and the common whitethroat (P = 0.03 and P = 0.01, respectively; two-tailed t-test) but not for the skylark (P = 0.15). In the skylark, there was a weak increase in Ks over the neo-sex chromosome, but still overall low Ks values, and a similar pattern was seen for loci located between 4.4 and 9.4 Mb in the great reed warbler and the common whitethroat (fig. 3A). In the first 3-Mb region of the neo-sex chromosome, the divergence patterns differed more between species, but the lack of data for some loci (see earlier) makes comparisons difficult. The great reed warbler had comparatively high Ks for the two loci in the first 3-Mb region (PCDH19, Ks = 0.22 and PCDH11, Ks = 0.34) (fig. 3A), higher than for the skylark in this region and similar to the levels observed at the ancestral sex chromosomes. However, even though the skylark showed considerably lower substitution levels at the loci in this region (with alignment gaps excluded from calculations) (fig. 3A), these loci exhibited the highest insertion/deletion polymorphism levels for the Z-and W-linked copy comparisons: a 2 bp indel in PCDH19 and a more extended indel region (comprising a total of 105 bp) in DIAPH2.

The nonsynonymous substitution rate (Ka) was generally low among the neo-sex gametologs (0–0.03 in great reed warbler, 0–0.08 in common whitethroat, and 0.01–0.12 in the skylark) (fig. 3B). Low Ka values were also observed at the ancestral sex chromosomes in all species, ranging from 0 (ATP5A1 and UBE2R2) to 0.05–0.09 (CHD1) over the ancestral sex chromosome (fig. 3B). No statistical difference in the nonsynonymous substitution rates was found between the neo-sex and the ancestral sex chromosome (P ≥ 0.7).

Regarding the GC content at third codon positions, the expectation was that it should be lowest in nonrecombining W gametolog. In line with this, all neo-W gametologs, with the exception of NKRF in the great reed warbler and DIAPH2 in the skylark, had lower GC content at third codon positions than their Z-linked gametologs (fig. 4A). However, the difference in GC content between neo-Z and neo-W was statistically nonsignificant (P = 0.10 in the great reed warbler; P = 0.12 in the common whitethroat and the skylark). The difference in GC content was even more pronounced when comparing the gametologs located on the ancestral sex chromosomes; the W-linked gametolog exhibited lower GC content for all genes analyzed (fig. 4B).

Phylogenetic Analysis of the Neo-sex and Ancestral Gametologs

The analysis of the phylogenetic relationships between gametologs located on the ancestral sex chromosomes revealed a consistent grouping of loci according to chromosomal class (Z or W), irrespective of species (fig. 5A). This pattern is in agreement with the old age (150 Ma) of the ancestral Z and W chromosomes and that recombination between the sex gametologs ceased before the Sylvioidea radiation.

Loci located on the neo-sex chromosome did not exhibit such a clear phylogenetic signal, with some markers clustering according to chromosomal class and others according to species (fig. 5B). Among the clades with strong support (i.e., bootstrap values >70%), we found that the Z and W gametologs in skylarks group together in four of the eight genes (PCDH19, DIAPH2, ZNF711, and ZC3H12B; fig. 5B) suggesting that recombination was still ongoing when the Alaudidae diverged from the rest of the Sylvioidea (i.e., ca. 42.2 Ma). Similarly, the neo-Z and W gametologs grouped together (with >70% bootstrap support) also in the great reed warbler at locus ZNF711 and in the common whitethroat at locus KLHL13 (fig. 5B), indicating that recombination continued at least in some regions of the neo-sex chromosome also after the split of Acrocephalidae and Timaliidae (i.e., ca. 39.4 Ma). However, the gametologs of the great reed warbler and the common whitethroat of loci located on each end of the neo-sex chromosome, respectively, clustered (with > 70% bootstrap support) according to chromosomal origin (the Z gametolog at PCDH19 and the W gametolog at NKRF; fig. 5B). This pattern indicates that recombination had already ceased in these regions before the split of Acrocephalidae and Timaliidae, suggesting that nonuniform recombination suppression events have occurred during neo-sex chromosome evolution.

Discussion

This study constitutes the first analysis of the evolutionary patterns of genes located on the recently detected neo-sex chromosome in Sylvioidea passerines (Pala et al. 2011). In the following sections, we discuss the fate of these relatively newly sex-linked genes and whether they evolve according to the neutral expectations of their new inheritance mode and/or progressively diverge due to specific selective pressures, distinct from those that affect autosomes. We approach questions that are of relevance to both ancestral and incipient chromosome systems and can be more effectively tackled when, as in the case of the present neo-sex chromosome system, the sex-limited gametolog (neo-W) is not yet too degraded and an extensive chromosomal region can still be directly compared.

Low Nucleotide Diversity on the Neo-Z

The first aspect that we wanted to explore was the impact of the transition from an autosomal to a sex chromosome on the overall variability of the loci involved. A decrease in nucleotide diversity is expected for sex chromosomes due to drift effects following from the reduction in effective population size compared to autosomes (Montell et al. 2001; Pool and Nielsen 2007), and this is supported by previous analyses in both male and female heterogametic systems (Sachidanandam et al. 2001; Berlin and Ellegren 2004; Sundström et al. 2004).

We have studied birds, a female heterogametic system, and expected a reduction in the effective population size (N_e) of three-fourths for neo-Z genes compared to autosomes. Under neutrality, the genetic diversity is expected to be proportional to N_e times the mutation rate. In line with this expectation, the neo-Z chromosome intron markers showed generally lower nucleotide diversity (π) when compared with loci mapped to the autosomal part of chromosome 4a in all three Sylvioidea species (fig. 2). The observed neo-sex:autosome nucleotide diversity ratio was 0.15 in the great reed warbler, 0.82 in the common whitethroat, and 0.42 in skylark and thus differed to some extent from the expected three-fourths. Such deviations have also been found in other studies and reflect that diversity values often do not proportionally translate to N_e of each chromosome type. A possible explanation for this is that the selection regimes differ between sex chromosomes and autosomes, e.g., because sexually antagonistic selection may intensify the Hill–Robertson effect and hitchhiking on sex-linked loci.

However, the amount of diversity observed at the neo-Z did not depart significantly from neutral expectation when formally assessed through a HKA test. The lack of significance could have different reasons: demographic events could have violated the assumptions of a strict neutral model of evolution (such as constant population sizes; Singh et al. 2007), ongoing recombination after the constitution of the neo-sex chromosome could have affected the diversity of neo-Z-linked loci (discussed later), and the male-driven evolution effect with slightly higher mutation rate on Z compared to autosomes in birds (Axelsson et al. 2004) could have obscured the effects of selection. Thus, it is perhaps premature to exclude selection on the basis of nonsignificant HKA tests. The difficulty of significantly detecting departures from neutrality with the HKA test does not seem restricted to recently established and slower evolving neo-sex systems such as the present one but has been observed both in old ancestral (Storchová et al. 2010) and fast evolving neo-sex chromosome systems (Bartolomé and Charlesworth 2006).

Divergence between Gametologs

The moderate loss in diversity of neo-Z-linked sequences that we observed is in line with the predictions from sex chromosome evolutionary theory but does not shed light upon the patterns of differentiation of the neo-sex gametologs. To address this question, we evaluated coding (exon) sequence divergence over eight gametologous gene pairs on the Sylvioidea neo-sex chromosome. The use of coding sequences allowed us to assess two important levels of sex chromosome evolution: the silent divergence, which is expected to correlate with the time since recombination between gametologs was reduced, and the functional divergence, which can be looked upon as a diagnostic feature of a progressive degeneration of one of the members of the pair.

Measures of silent divergence per site (Ks) of neo-sex chromosome loci in the three species were generally lower when compared with the Ks values of gametologs on the ancestral sex chromosome (fig. 3), as expected from the difference in age of the ancestral and the neo-sex chromosomes.

However, we observed some potentially interesting variation in divergence between different regions of the neo-sex chromosome and between species. In the skylark, there was a weak increase in Ks over the neo-sex chromosome, and a similar pattern was found for loci located between 4.4 and 9.4 Mb in the great reed warbler and the common whitethroat (fig. 2). According to the linkage analysis performed in the great reed warbler (Pala et al. 2011), the new enlarged sex chromosome present in Sylvioidea species resulted from the fusion between the distal end of ancestral chromosome Z and the distal (centromeric) end of the first 10-Mb part of chromosome 4a (fig. 1). Markers toward the 10-Mb position of the neo-sex chromosome would therefore be located near the fusion point between the neo-sex chromosome and the ancestral Z. Because silent divergence is predicted to correlate to the time since recombination ceased (Lahn and Page 1999; Handley et al. 2004; Bergero et al. 2007), the weak increase in Ks toward the fusion area provide some support for the suggestion that recombination cessation of the neo-sex chromosome initially started in the area with closest proximity to the ancestral sex chromosomes. Such a spread of reduced recombination from the fusion point is in line with hypothesis of recombination suppression for neo-sex arisen through translocation of autosomal material to the sex chromosome.

An alternative hypothesis suggests that recombination suppression in differentiating sex chromosomes is initiated through a selectively favorable linkage between a sex-determination locus and sexual antagonistic alleles, advantageous to one sex and harmful to the other (Rice 1987; Charlesworth et al. 2005; Bergero and Charlesworth 2009). In the zebra finch, the distal region of the Z chromosome arm to which the neo-sex chromosome is now connected harbors the relevant player in male sex determination, DMRT1 (Smith et al. 2009), so it is possible that this locus has influenced the fate of the neo-sex chromosome that are now brought into its vicinity. Another locus implicated in the sexual differentiation in different species (de Waal et al. 2008; Wang et al. 2009) that could offer a selective advantage when linked to sex determination genes on the sex chromosomes is the androgen receptor gene. Interestingly, this locus is located on the neo-sex chromosome, at approximately 6.4 Mb, close to locus P2RY4 (5.6 Mb) which has relatively high Ks- and Ka values in particular in skylark (however, it is also close to ZC3H12B, which has low substitution rates). Considering the estimated phylogenetic relationships of loci located in the skylark neo-sex chromosome, the region between 4.4 and 6.8 Mb, comprising markers ZNF711, P2RY4, and ZC3H12B, is also the one with the highest genetic distance between Z- and W-linked copies. It is therefore possible to speculate that similarly to other chromosomal systems (Rice 1987; Charlesworth et al. 2005; Bergero and Charlesworth 2009), sexual antagonistic alleles comprised in the neo-sex chromosomes (androgen receptor gene), and in the distal end of the ancestral chromosome (DMRT1), function as initial stepping stones for recombination cessation. This suggestion needs to be followed up by additional analyses (e.g., using genome-wide functional genetics approaches or expression studies).

The diversity estimates for the loci in the first 3-Mb region (PCDH19, DIAPH2, and PCDH11) are much more difficult to interpret. Unlike for the remaining loci, observations were quite divergent in the three species. In the common whitethroat, no W sequences were retrieved for loci in this region, suggesting high gametolog divergence, altering the primer binding regions. Note, however, that we amplified both Z and W of two intronic markers in this region in the common whitethroat in our previous study (see table 1 in Pala et al. 2011). In the great reed warbler, unexpectedly high Ks values (similar to those of the ancestral gametologs) were observed for PCDH19 and PCDH11, and the neo-W gametolog of DIAPH2 failed to amplify. In the skylark neo-sex chromosome, low silent divergence estimates were observed for PCDH19 and DIAPH2. However, other features, more consistent with a divergence scenario were additionally observed, such as the occurrence of deletions affecting protein sequence in the W-linked sequences of these two loci, and amplification failure of the neo-W gametolog of PCDH11. According to the linkage analysis in the great reed warbler, these three loci have a telomeric position on the enlarged sex chromosome (Pala et al. 2011). Because the recombination rate is predicted to be higher in the proximity of telomeric regions—a pattern supported from both chicken (Groenen et al. 2009) and zebra finch (Backström et al. 2010)—the telomeric position of these neo-sex loci cannot easily explain their high divergence estimates in Sylvioidea (which in contrast indicates reduced recombination). However, it is important to point out that although gene order was conserved for the few loci studied between the great reed warbler and the zebra finch (Pala et al. 2011), the possibility that gene order is not conserved cannot be excluded (cf. Warren et al. 2011). Moreover, the terminal end of the neo-sex chromosome could of course hold some unknown sexually antagonistic loci causing recombination reduction and modulating divergence. In fact, mutations in the diaphanous gene in Drosophila cause both male and female infertility (Castrillon and Wasserman 1994), and defects in its X chromosome human homolog (DIAPH2; i.e., one of the loci included here) have been associated with ovarian failure (Simpson and Rajkovic 1999).

However, another hypothesis to explain the variation in divergence between different regions of the neo-sex chromosome and between species is the occurrence of inversions that reduced recombination between the neo-Z and neo-W chromosomes (cf. Zhou et al. 2008). In both ancestral and incipient chromosome systems alteration of structural conformation, the occurrence of selectively favored inversions, the accumulation of repetitive sequences, and the gain of heterochromatin modulate divergence through recombination suppression of gametologs (Bachtrog 2005; Charlesworth et al. 2005). Distinguishing such rearrangements between the gametologs would require detailed cytological studies (e.g., using fluorescence in situ hybridization).

Divergence between neo-Z and neo-W loci was also evident from the analysis of the GC content at the third codon positions (fig. 4). GC content has been shown to be positively correlated with recombination rate in vertebrate genomes (Fullerton et al. 2001; Webster et al. 2006). Third position GC content should therefore be lower for the nonrecombining W chromosome compared with the recombining Z chromosome (Meunier and Duret 2004). Our results are in agreement with this prediction, with all neo-Z-linked loci except NKRF in the great reed warbler and DIAPH2 in the skylark showing a higher third position GC content than their W-linked homologs. Also, the difference in GC content between Z and W gametologs was more pronounced for the ancestral than for the neo-sex chromosome, as expected from their age difference.

Recombination and Gene Degradation of Newly Sex-Linked Loci

The phylogenetic analysis revealed clear differences in the evolutionary trajectories of loci on ancestral and neo-sex chromosome and provided further insight on the evolution of recombination patterns of the newly sex-linked loci in the great reed warbler, the common whitethroat, and the skylark. As expected from the analysis of sex chromosome evolution in other bird species (Handley et al. 2004), the gametologs on the ancestral chromosomes fell into two independent Z and W groups (fig. 5A). Such a pattern indicates that the Z and W chromosomes were already evolving independently and that recombination had ceased on the ancestral chromosomes before the Sylvioidea radiation.

According to our previous estimates, the translocation of part of the 4a autosome to the sex chromosomes should have occurred at approximately 42 Ma, that is, at the time of the split of the Sylvioidea branch of the avian phylogeny (of which the most basal group includes the skylark) (Pala et al. 2011). Furthermore, the presence of Z and W gametologs at chromosome-wide loci in our three study species (based on the analysis of sex-specific polymorphisms at 19 loci; Pala et al. 2011) and the segregation pattern at some loci in the great reed warbler (Pala et al. 2011) suggest that the neo-sex chromosome does not recombine presently. There is no indication of any pseudoautosomal regions on the neo-sex chromosome on the resolution of our markers (Pala et al. 2011; this study). The question is when recombination was suppressed and in which chromosomal regions it was initiated? In this study, we show that most branches with high bootstrap support group neo-sex chromosome gametologs according to species (four loci in skylark and one each in the two warblers; fig. 5B). This indicates that the neo-sex chromosome was still recombining, at least in some regions, when the Alaudidae diverged from the rest of the Sylvioidea (ca. 42.2 Ma), as well as after the split of Acrocephalidae and Timaliidae (ca. 39.4 Ma; divergent times from Pala et al. 2011).

Exceptions to this more general clustering pattern according to species were observed at two loci. The gametologs of the great reed warbler and the common whitethroat of loci located on each end of the neo-sex chromosome, respectively, clustered (with high bootstrap support) according to chromosomal origin (the Z gametolog for locus PCDH19 and the W gametolog for NKRF; fig. 5B). The phylogenetic grouping according to chromosomal origin at these loci suggests that, although most parts of the neo-sex chromosome were still recombining when the split of the species occurred, recombining cessation had already been initiated in the telomeric region of neo-sex chromosome and in the fusion area to the ancestral sex chromosome.

As mentioned earlier, the loss of recombination has deep effects on the evolutionary fate of the gametologs and will lead to the progressive degeneration of the sex-limited chromosome. Known consequences of recombination cessation are the accumulation of deleterious mutations affecting protein functionality and loss of genes (Bachtrog 2005; Bergero and Charlesworth 2009; Wilson and Makova 2009). Thus the degree of nonsynonymous substitution rate between gametologs can be used as a proxy for the accumulation of deleterious mutations and as an indication of the degree of W chromosome degeneration.

We found relatively low nonsynonymous divergens (Ka in the range of 0–0.12), which suggests that functionality is still highly conserved in neo-sex-linked gametologs. Moreover, our successful isolation of a considerable number of functionally conserved gametologs using standard primer design and PCR approaches (Pala et al. 2011) implies that no widespread gene loss or functional divergence has yet occurred in the Sylvioidea neo-W chromosomes. This contrasts the situation in the old and very diverged ancestral chromosome system in birds, in which the number of identified Z and W homologs is much lower (Montell et al. 2001; Agate et al. 2004; Handley et al. 2004; Nam and Ellegren 2008). Also the 80–130-My-old X- and Y-added region in eutherian mammals has started to deteriorate (Waters et al. 2001; Livernois et al. 2012). Furthermore, the comparatively very recently established neo-Y chromosome in D. miranda (ca. 1 Ma) already shows clear signs of degradation in the form of an excess of nonsynonymous mutations on neo-Y-linked genes (Bartolomé and Charlesworth 2006), implying that, irrespective of other factors, the rate of molecular evolution is also dependent on the generation time and population size of specific systems.

However, features suggestive of incipient degeneration were observed at specific points of the neo-W chromosome in some of our species and more prominently toward the distal region of the chromosome. We failed in amplifying the W gametolog in some species at loci located in the first 3-Mb region of the neo-sex, indicating high gametolog divergence, altering the primer binding regions. Moreover, frame shift mutations and deletions, altering W protein sequence, were observed at PDH19 (0.9 Mb), DIAPH2 (1.6 Mb), and P2RY4 (5.6 Mb) in the skylark. Finally, there was a one base pair deletion at the W-linked copy of P2RY4 in the common whitethroat additionally indicating that the functional divergence of the neo-W chromosome is ongoing in Sylvioidea passerines.

Conclusions

The Sylvioidea neo-sex chromosome shares classical evolutionary features of ancestral sex chromosome systems, such as recombination cessation and progressive gametolog divergence. As expected, the divergence between gametologs is weaker at the younger Sylvioidea neo-sex chromosome than at the much older ancestral avian sex chromosome (Handley et al. 2004). The central question remains of which mechanism triggers the recombination suppression and drives the progressive differentiation of genes brought into a sex chromosome system (Bergero and Charlesworth 2009; Wilson and Makova 2009). In Drosophila, the absence of recombination in males implies that autosomal genes that are translocated to the Y will directly cease to recombine with their gametolog, as the fused chromosome will segregate along with the Y chromosome (Bartolomé and Charlesworth 2006). Recombination is thus quickly suppressed, maximizing divergence levels of the neo-sex-linked genes (Bartolomé and Charlesworth 2006). Similarly, in the black muntjac, a large autosomal inversion coincided with the translocation to the sex chromosome and resulted in recombination suppression (Zhou et al. 2008). Conversely, in the Sylvioidea neo-sex chromosome, recombination was apparently maintained for a considerable period of time after the fusion of 4a autosomal genes to the ancestral Z and W chromosomes. Recombination ceased initially at specific regions along the Sylvioidea neo-sex chromosome, supposedly in the telomeric region and in the fusion area to the ancestral sex chromosome, emphasizing the potential importance of both physical linkage and gene content for the evolution of the neo-sex chromosome. Specific features that remain to be explored include the functional significance of the physical linkage between specific genes on neo-sex and ancestral sex chromosomes. Further investigation might clarify the role of such interactions in modulating the evolutionary trajectories of neo-sex chromosome systems.

Acknowledgments

The authors thank Martin Stervander for his contributions to primer design and molecular dating. This work was supported by the Swedish Research Council (to S.B., D.H., and B.H.), the Crafoord Foundation (to B.H.), the Wenner-Gren Foundation (to I.P.), and Helge Axson Johnsons Stiftelse (to I.P.).

References

Author notes