The sex chromosomes of dioecious white campion, Silene latifolia (Caryophyllaceae), are of relatively recent origin (10–20 million years), providing a unique opportunity to trace the origin and evolution of sex chromosomes in this genus by comparing closely related Silene species with and without sex chromosomes. Here I demonstrate that four genes that are X-linked in S. latifolia are also linked in nondioecious S. vulgaris, which is consistent with Ohno's (1967) hypothesis that sex chromosomes evolve from a single pair of autosomes. I also report a genetic map for four S. latifolia X-linked genes, SlX1, DD44X, SlX4, and a new X-linked gene SlssX, which encodes spermidine synthase. The order of the genes on the S. latifolia X chromosome and divergence between the homologous X- and Y-linked copies of these genes supports the “evolutionary strata” model, with at least three consecutive expansions of the nonrecombining region on the Y chromosome (NRY) in this plant species.
IT has been suggested that sex chromosomes evolve from a pair of homologous autosomes (proto sex chromosomes), which stop recombining with each other and gradually diverge into mostly nonhomologous X and Y chromosomes (Ohno 1967; Bull 1983). Alternatively, sex chromosomes could have evolved in several stages, through consecutive translocations of autosomal fragments to the (proto-)sex chromosomes and gradual expansion of the nonrecombining region. Y-linked genes advantageous in males and detrimental in females may possess a selective advantage, promoting translocation of autosomal regions containing sexually antagonistic genes to the Y chromosome, as well as further expansion of the nonrecombining region on the Y chromosome (Charlesworth and Charlesworth 1980; Bull 1983; Rice 1987). Indeed, the distal regions of the human sex chromosomes are autosomal in marsupials (Spencer et al. 1991), suggesting that modern mammalian sex chromosomes have formed due to translocations of fragments of autosomes to the sex chromosomes (Watson et al. 1991). The Drosophila Y chromosome also appears to contain a patchwork of genes translocated from autosomes (Carvalho 2002), casting doubts on the generality of Ohno's (1967) hypothesis that sex chromosomes evolve from a single pair of autosomes.
Genes with homologous X- and Y-linked copies on the human sex chromosomes fall into four groups, depending on silent X/Y divergence (Ks), suggesting that different genes spent different amounts of time in the nonrecombining region of the Y chromosome, the NRY (Lahn and Page 1999). On the basis of this observation, it was suggested that the cessation of recombination on the mammalian Y was not a single event, but occurred in several stages spanning a timescale of 240–350 MY and creating four “evolutionary strata” with differing levels of divergence between the X- and Y-linked homologs (Lahn and Page 1999). Mouse (Sandstedt and Tucker 2004) and bird (Handley et al. 2004) sex chromosomes were also reported to have evolutionary strata. It is not clear, however, whether the stepwise formation of the sex chromosomes is a general process or is confined to mammals and birds.
Studying the origins of sex chromosomes in model organisms, such as Drosophila, mouse, human, and chicken is extremely difficult, as they arose hundreds of millions of years ago. The Y (or W in chicken) chromosomes in these species have diverged from the X (or Z) and lost almost all functional genes due to such processes as Muller's ratchet and genetic hitchhiking, leading to the accumulation of deleterious mutations and gene loss in the nonrecombining regions (Charlesworth and Charlesworth 2000). Many organisms, however, have much younger sex chromosomes [e.g., plants Silene latifolia (Filatov et al. 2000) and Carica papaya (Liu et al. 2004) and the stickleback fish (Peichel et al. 2004)]. In particular, the plant genus Silene is very convenient for studying the early stages of sex chromosome evolution because sex chromosomes in this genus have been found only in a small cluster of dioecious Silene species (section Elisanthe: S. latifolia, S. dioica, S. diclinis, S. heuffelii, and S. marizii), where males are the heterogametic sex (XY) and females are homogametic (XX) (Westergaard 1959). The rest of the Silene genus is nondioecious (exept S. otites, which apparently evolved dioecy independently from Elisanthe), allowing a comparison of closely related species with and without sex chromosomes and a tracing of the origin of the sex chromosomes.
Silent divergence between dioecious S. latifolia and nondioecious S. vulgaris is ∼20%, suggesting that the S. latifolia sex chromosomes are probably no older than 10–20 million years (Filatov and Charlesworth 2002). Degeneration has probably started on the S. latifolia Y chromosome, as one of the first isolated X-linked genes, MROS3X, has a degenerate Y-linked copy (Guttman and Charlesworth 1998). However, four further genes isolated from the X [SlX1/Y1 (Delichère et al. 1999; Filatov et al. 2000), SlX4/Y4 (Atanassov et al. 2001), DD44X/Y (Moore et al. 2003), and SlssX/Y (Filatov 2005)] have intact X- and Y-linked homologs, suggesting that genetic deterioration of the S. latifolia Y-linked genes is at a very early stage. The silent divergence between the homologous S. latifolia X- and Y-linked genes varies from as little as 1.7% for the SlX1/SlY1 to ∼16% for the SlX4/SlY4 genes, suggesting that the cessation of recombination between the proto-X and proto-Y chromosomes has been gradual, rather than the result of a single event (Atanassov et al. 2001; Filatov and Charlesworth 2002), and resembles the evolutionary strata discovered on the human sex chromosomes (Lahn and Page 1999). It is not known, however, whether the order of the genes on the X supports the evolutionary strata model. The difference in X/Y divergence of the S. latifolia sex-linked genes may also be due to consecutive translocations of fragments of autosomes to the sex chromosomes. Such translocations are probably quite common because they were reported for many insect (Bachtrog and Charlesworth 2002; Jacobs 2004), mammalian (Watson et al. 1991; Pack et al. 1993), and plant (Ainsworth 1999) species. Thus, without knowing the order of the genes on the Silene X chromosome, and without comparisons with the closely related nondioecious Silene species, it is difficult to infer the sequence of events that has led to the formation of the Silene sex chromosomes.
To shed light on the evolutionary history of the S. latifolia sex chromosomes, I constructed and compared genetic maps for X-linked genes in S. latifolia and their homologs in nondioecious S. vulgaris. For this study I used three published X-linked genes, the SlX1 (Delichère et al. 1999; Filatov et al. 2000), SlX4 (Atanassov et al. 2001), and DD44X (Moore et al. 2003) as well as a new X-linked gene, the SlssX encoding spermidine synthase (Filatov 2005). The MROS3X (Guttman and Charlesworth 1998) has not been used because this gene is a member of a multicopy gene family with multiple autosomal copies (Kejnovsky et al. 2001), which would complicate genetic mapping and interpretation of the results.
To establish the relative positions of the SlX1, SlX4, DD44X, and SlssX genes on the S. latifolia X chromosome, I tested cosegregation of these genes in the F1 progeny of the genetic cross described previously (“family 5” in Filatov 2005). The fragments of the four X-linked genes were PCR amplified using primers shown in Tables 1 and 2. Conveniently, the female parent used in the cross was heterozygous for the PCR fragment size differences in three of four genes, SlX1, DD44X, and SlX4. Thus, the segregation of the alleles in these genes was inferred from the electrophoretic mobility of PCR products on 1.5% agarose gels. For the SlssX, however, no size difference of the PCR products was detected. Direct sequencing of the SlssX PCR product from the maternal and paternal individuals of the family 5 (Filatov 2005) revealed two nucleotide positions heterozygous in the female and hemizygous in the male. As none of these polymorphisms were located at restriction sites, the segregation of the SlssX gene was tested by direct sequencing of the PCR products of all the F1 progeny.
The four X-linked genes were genotyped in the parents and in 136 F1 progeny (40 males and 96 females). No recombinants were found between the SlssX and DD44X genes, demonstrating that the two genes are tightly linked (<1 cM). The SlX1 and SlX4 genes are the most distant from each other (24.3 cM), while the distances from the SlssX/DD44X cluster to the SlX1 and the SlX4 genes are shorter: 7.4 and 17.9 cM, respectively (see Figure 1 and Table 3). Hence, the most likely order of the four genes on the X chromosome is SlX1, SlssX/DD44X, and SlX4. This order corresponds to what would be expected from the evolutionary strata model: the youngest (SlX1/Y1, Ks = 1.7%) and the oldest (SlX4/Y4, Ks = 16%) sex-linked genes are at the two ends of the map, while the two genes with intermediate X/Y silent divergence (SlssX/Y, Ks = 8% and DD44X/Y, Ks = 7%) are located in the middle (Figure 1).
The order of the three S. latifolia genes, SlX1, DD44X, and SlX4, mapped in an independent genetic cross (Nicolas et al. 2005) is consistent with the results reported in the current study. However, the genetic distances between these genes reported here are are somewhat shorter, compared to those observed by Nicolas et al. (2005). The reasons for this discrepancy are unclear. It may well reflect the variation in recombination rate among the S. latifolia individuals. If the maternal individual in the S. latifolia cross reported here is heterozygous for an inversion located between the SlX1 and SlX4 genes, this could explain the shorter genetic distances reported in the current study. Interestingly, according to the map reported by Nicolas et al. (2005), the SlX1 gene is the closest to the pseudoautosomal region, which is also consistent with the evolutionary strata model.
To establish whether the S. latifolia sex chromosomes originated from a single pair of autosomes or were assembled from fragments of several autosomes, I conducted a segregation analysis of four S. vulgaris genes—SvXY1, Svss, SvDD44, and SvXY4, which are homologous to the S. latifolia sex-linked genes SlX1/Y1 (Delichère et al. 1999; Filatov et al. 2000), SlssX/Y (Filatov 2005), DD44X/Y (Moore et al. 2003), and SlX4/Y4 (Atanassov et al. 2001), respectively. If S. latifolia sex chromosomes have evolved from a single pair of autosomes, these genes would be expected to be linked in nondioecious S. vulgaris. S. vulgaris is gynodioecious (females and hermaphrodites) and lacks sex chromosomes. The genetic control of gynodioecy in this species is nuclear cytoplasmic (Charlesworth and Laporte 1998) and apparently originated independently from that of dioecy in S. latifolia.
For the segregation analysis in S. vulgaris I used 58 F1 progeny of the cross between two S. vulgaris individuals grown from seeds kindly provided by D. Charlesworth (University of Edinburgh). Two parental S. vulgaris individuals and 58 F1 progeny were genotyped for the SvXY1, Svss, SvDD44, and SvXY4 genes. In the S. vulgaris cross, only the maternal parent was heterozygous for the size of the PCR product of the SvXY1 gene, which allowed me to study segregation of this variant on agarose gels. Segregations of all the other genes were studied by direct sequencing of the PCR products of the parents and F1 offspring. The primers used for PCR amplification and sequencing are listed in Tables 1 and 2. The segregation analysis in the S. vulgaris cross demonstrated that all four genes are linked in S. vulgaris, supporting the hypothesis that S. latifolia sex chromosomes evolved from a single pair of autosomes (Figure 1 and Table 3). Interestingly, the distance between SvDD44 and SvXY4 reduced to only 3.5 cM, while the distance between the SvDD44 to Svss increased to 27.6 cM, compared to the distances on the X chromosome of S. latifolia. This demonstrates that the order of the genes in S. vulgaris differs from that in S. latifolia, probably due to an inversion, including Svss and SvXY4 genes.
Mapping of the four genes on the S. latifolia X chromosome and in the nondioecious S. vulgaris has demonstrated that, despite very different divergence between the homologous X- and Y-linked copies of these genes, all four genes are linked in both species, suggesting that S. latifolia sex chromosomes have evolved from a single pair of autosomes via formation and expansion of the NRY. Although only four genes were used in this study, they fall into at least three evolutionary strata and might represent a significant proportion of the S. latifolia sex chromosomes (∼25 cM according to the map reported here and >50 cM according to Nicolas et al. 2005). Nevertheless, it is still possible that parts of the S. latifolia sex chromosomes were translocated from the autosomes. Indeed, the autosomal S. latifolia Slap3 gene was reported to have a functional Y-linked homolog in this species, which probably originated due to a translocation from the autosome to the Y chromosome (Matsunaga et al. 2003). Thus, translocation of genetic material to the sex chromosomes observed in many species (Watson et al. 1991; Pack et al. 1993; Ainsworth 1999; Bachtrog and Charlesworth 2002; Jacobs 2004) may also be common in Silene. With only four genes it may be difficult to detect such translocations. Isolation and mapping of further sex-linked genes will lead to a much more detailed picture of the evolution of S. latifolia sex chromosomes.
Originally, cessation of recombination between the proto X and Y chromosomes in the S. latifolia lineage probably occurred in the region including the SlX4/Y4 gene. The formation of the original NRY (stratum I, containing SlX4/Y4) in S. latifolia was followed by at least two further expansions of the NRY region, creating stratum II, including the SlssX/Y and the DD44X/Y genes and, finally, stratum III, which includes the SlX1/Y1 gene. Because a similar sequence of NRY expansions apparently resulted in formation of the human sex chromosomes (Lahn et al. 2001), this might be a general evolutionary pathway, which may be followed in most groups of animal and plant species, which evolved sex chromosomes independently.
I thank Deborah Charlesworth for providing S. vulgaris seeds and Joe Ironside, Dave Gerrard, and Graham Muir for critical reading of the manuscript. This work was funded by the Biotechnology and Biological Sciences Research Council.
Communicating editor: O. Savolainen
- Received October 1, 2004.
- Accepted March 14, 2005.
- Genetics Society of America