A XY system on chromosome 15 of Salix arbutifolia and Salix triandra, and possible turnovers in willows
The two species studied in detail here, Salix arbutifolia and S. triandra, both have SDSs on chromosome 15 (Fig. 2, Fig. 3) as Gulyaev et al. (2022) hypothesized for species of the Vetrix clade. Our heterozygosity analyses suggest that both have male heterogamety. Li et al. (2020) suggested S. triandra has a ZW SDS on chromosome 15, based on linkage mapping, but they used the female S. purpurea v1.0 reference genome sequence, whose Z and W were not phased, and represent a chimera (https://phytozome-next.jgi.doe.gov/info/Spurpurea_v1_0; Li et al. 2020; Zhou et al. 2020), which can cause false inferences. The chromosome 15 of S. purpurea v1.0 is 22.6 Mb in length, and after phasing (S. purpurea v5.1), its 15W and 15Z sequences are 15.7 Mb and 13.3 Mb, respectively (Zhou et al. 2020). It is however worth considering that the SDSs are quite flexible in Salix, it be possible that there is a change within S. triandra. It is a very widespread species and different evolution could happen in different areas.
Our ancestral state reconstruction suggests that Salix arbutifolia and S. triandra are in a transitional stage between clades 1 and 4, whereas a Y-linked region on chromosome 7 appears to be the ancestral state for clade 4 (Salix) (Fig.1; 4a). Since our ancestral state analysis yielded similar probabilities for XY on chromosome 7 (0.395) or chromosome 15 (0.356) as the ancestral state for the genus Salix, we cannot rule out the possibility that XY on chromosome 15 is the ancestral state of the genus Salix. Two scenarios are thus possible, both involving at least two turnover events (Fig. 4a): 1) if Y-linked region on chromosome 7 is the ancestral state for the genus Salix, at least two turnover events may have occurred in the Vetrix clade (XY on chromosome 7 to XY on chromosome 15, and to ZW on chromosome 15), while the entire Salix clade maintained the chromosome 7 ancestral state (Fig. 4a, blue arrow), or 2) if an XY system on chromosome 15 is the ancestral state for genus Salix, at least two turnover events must have occurred within the genus Salix (XY to ZW on chromosome 15 in the Vetrix clade, and XY on chromosome 15 to XY on chromosome 7 in the Salix clade, Fig. 4a, gray arrow). As the sex-linked regions of Populus, the sister genus of willows, are mostly on chromosome 19, the two groups of willows with sex-linked regions on different chromosome probably evolved independently from chromosome 19 (He et al. 2021a) (Table 1), perhaps involving different sex determination gene(s). This remains to be tested, as willow sex determination gene(s) have not yet been identified.
Our phylogenomic analyses and the sex-linked region identified in S. arbutifolia and S. triandra supports their assignment to the Vetrix clade (Fig. 1), consistent with previously studies (Lauron-Moreau et al. 2015; Wu et al. 2015; Gulyaev et al. 2022).
Possible reasons for turnovers in the genus Salix
Bull (1983) and Vicoso (2019) proposed that some clades have extremely conserved sex chromosomes because they have become genetically degenerated, while others have not degenerated and can undergo sex chromosome turnovers. The physical sizes of sex-linked regions of willows are relatively large (Table 1), but their S-LRs are, at most, only slightly degenerated (Almeida et al. 2020; He et al. 2021a). Hence, turnover events should be possible in Salix and Populus (He et al. 2021a; Gulyaev et al. 2022). It has, however, been proposed that genetic degeneration can promote turnovers, replacing degenerated Y with new, non-degenerated sex chromosomes (Blaser et al. 2014). Accumulation of deleterious mutations on the ancestral sex chromosomes seems unlikely to explain the XY to ZW change on chromosome 15 in the Vetrix clade, because such “homologous” turnovers are predicted to maintain the ancestral heterozygous sex (Bull, 1983; Blaser et al. 2014; van Doorn and Kirkpatrick 2010; Jeffries et al. 2018; Scott, Osmond and Otto 2018). Drift-induced turnovers are also generally expected to maintain the ancestral heterogamety (Veller et al. 2017).
The sex-ratio selection and sexually antagonistic selection models are more likely to lead to changes in the heterogametic sex (Jeffries et al. 2018) like those inferred in willows. To explain the XY to ZW transition in the Vetrix clade, the W-linked factor must be dominant and cause femaleness in the presence of the Y. Sexual selection in plants, such as pollinator attraction, is very similar to some forms of sexual selection in many animals (Charlesworth et al. 1987). In S. dunnii, Apis cerana's significant preference for visiting female flowers suggests that sexual selection pressure can be placed on male flowers (Zeng et al. 2022). Likewise, evidence of sexual dimorphism was also found in S. purpurea (Gouker et al. 2021). Hence, sexually antagonistic selection may play a role in the turnover of the genus Salix.
Sex ratio selection is, however, likely in plants, and could be involved in the turnover events of the genus Salix. Sex ratio imbalances are common in nature (Scott et al. 2018), and can favour a new sex determining locus that restores an equal sex ratio (Beukeboom and Perrin 2014; Bull 1983; Mank, Hosken and Wedell 2014). In willows, female biased sex ratios are associated with ZW systems (Table S8), possibly reflecting incompatibility between maternally and paternally inherited Z-linked alleles (Pucholt et al. 2017a). Willows with XY SDS may generally have male-biased or balanced sex ratios (Table S8). If the ancestor of the Vetrix clade had male heterogamety and a male biased sex ratio, a feminizing gene might be able to invade and cause a transition to ZW SDS (Kozielska et al. 2010). On the other hand, as many willows with female biased sex ratios have ZW systems. Male or female bias can be also introduced during polyploidization, and many willows are polyploids (He and Hörandl 2022). Our ancestral state reconstruction supports changes in both the heterogamety and the stamen number, with a reduction to two stamens accompanying the change from male to female heterogamety. However, it is worth noting that in Salix group there are some polyploid species with just 2 stamens, eg. S. chienii (tetraploid) and S. matsudana (tetraploid) (Gulyaev et al. 2022), whether such a pattern exists in polyploidy requires further investigation. Hence, the relationship between turnover events and biased sex ratios remains unclear.
Can turnovers explain homomorphic sex chromosomes in Salicaceae?
The sex chromosomes in the Salicaceae stricto sensu (Populus and Salix) show no signs of major genetic degeneration, and most have remained homomorphic (Pucholt et al. 2017b), despite almost all species being dioecious, and dioecy having evolved more than 45 million years ago in the genus Salix. The rate at which genetic degeneration occurs after loss of recombination between sex-linked regions is not yet understood (reviewed by Charlesworth 2021). However, a turnover event will evidently create a new non-degenerated sex-linked region (Beukeboom and Perrin 2014; Gamble et al. 2015; Bachtrog et al. 2014; Jeffries et al. 2018; Kuhl et al., 2021; Vicoso 2019). Turnovers via chromosomal mutations (translocations) are in Salix probably feasible because the chromosomes are morphologically very similar to each other (Pucholt et al. 2017b). Hence, turnovers and also the return of previous SDS-bearing chromosomes to autosomes would be probably easier than in genera with more differentiated karyotypes. Turnovers, in turn, would inhibit degeneration and loss of size of sex chromosomes. Turnovers, including transitions between XY and ZW systems, may be associated with a lack of sex chromosome heteromorphism (Bachtrog et al. 2014, Zhou et al. 2018). Poplars have more known turnover events than willows (Table 1). The estimated sizes of Salix sex-linked regions are generally significantly larger than in Populus, possibly due to more frequent turnover events in Populus. Another possibility is that the S-LRs of Salix species are within pericentromeric regions that recombine very rarely (Chen et al. 2016; Zhou et al. 2018; He et al. 2021a; Wikerson et al. 2022), creating larger fully sex-linked regions than in Populus.
The S-LRs in Salix arbutifolia and S. triandra were estimated above smaller than those of other willows of the Vetrix clade (see Table 3). This might seem to be evidence against the turnover events inferred here, since, if the ZW on chromosome 15 is the derived state, it might be expected to be smaller than that in species with a male-determining region on chromosome 15. However, our estimates used the S. purpurea reference genome, and may reflect the actual sizes in S. arbutifolia and S. triandra.
We only identified two sex-link genes in S-LRs, among which Sapur.15ZG052500 (best hit in the A. thaliana genes: ATGUS2) is involved cell elongation and beta-glucuronidase activity, and Sapur.15ZG056900 (reciprocal best hits found the A. thaliana genes ATPPRT3) in positive regulation of the abscisic acid-activated signaling pathway. The homolog of Sapur.15ZG052500 also shows female-biased expression in shoot tip of S. purpurea (Carlson et al. 2017).
In poplars, turnovers have also been inferred, and different partial duplications of ARR17-like genes have been shown to be involved in different species (Yang et al. 2020; Li et al. 2022; Xue et al. 2020). We searched for ARR17-like genes in the S-LRs of S. arbutifolia and S. triandra, but did not find any sex-linked copies in either species. We speculate that willows probably do not share the same sex-determining gene as the one found in poplars (He et al. 2021a), or duplicated copy of ARR17-like gene could be found in their own genomes (S. arbutifolia and S. triandra). However, turnovers that involve a change from male to female heterogamety, or vice versa, also seem unlikely to involve the same gene (for example, a dominant male-determining factor in a population with male heterogamety cannot cause a turnover by simply being duplicated to a different location, as a dominant factor is required, that can convert individuals to females even if they carry the ancestral male-determining factor). Thus such turnovers seem unlikely to involve the duplication mechanism, and this is consistent with different genes being involved in Populus and Salix. The results in are also consistent with the expectation that, when heterogamety changes, female heterogamety is generally the derived state. This is because, if the first step in de novo evolution of separate sexes in plants is probably a loss of function sterility mutation, which is likely to be recessive. A recessive male sterility mutation, a common kind of mutation, can initiate evolution of an XY system, but a ZW system requires an unusual type of mutation causing dominant male sterility. A turnover can, however, readily produce a ZW system through a mutation that dominantly suppresses male fertility, for example a duplication that leads to production of an RNA sequence that suppresses a function essential for anther or pollen development.