Evolution of sexual systems and regressive evolution in Riccia

Abstract

Introduction

Organismal complexity is correlated with transcription factor (TF) diversity, and the origin of new TFs is often associated with developmental or physiological innovations (Lang et al., 2010; Weirauch & Hughes, 2011; Mendoza et al., 2013; Catarino et al., 2016; Wilhelmsson et al., 2017). Conversely, the loss of transcription factor (TF) genes or cis-regulatory sequences through which TFs act has been correlated with the loss of developmental or physiological traits, a process commonly referred to as regressive evolution. For example, parallel reductions in vision have occurred among cave-dwelling animals (e.g. Jeffery, 2009; Langille et al., 2022), hindlimbs have been lost independently in snakes and cetaceans (e.g. Cohn & Tickle, 1999; Thewissen et al., 2006), and similar parallel reductions are observed in the ability of land plants to support mycorrhizal fungal interactions (e.g. Delaux et al., 2014; Favre et al., 2014; Bravo et al., 2016). Many cases of regressive evolution have been correlated with the loss of regulatory sequences or losses of genes. For example, a characteristic suite of genes, including key TFs controlling the process, have been lost in species that no longer have the ability to form mycorrhizal fungal interactions (Delaux et al., 2014; Favre et al., 2014; Bravo et al., 2016).

Among the Marchantiopsida, a liverwort clade includes the complex thalloid liverworts (Crandall-Stotler et al., 2009), by far the most speciose genus is Riccia, with well over 200 species (Schuster, 1992; Cargill et al., 2016). Phylogenetically, Riccia is nested within the other Marchantiopsida (Forrest et al., 2006; He-Nygrén et al., 2006; Villarreal et al., 2016), lending support to Goebel's idea that the Riccia are highly reduced from a more complex Marchantiopsida ancestor (Goebel, 1930). For example, oil bodies, a synapomorphy of liverworts, are lacking in Riccia gametophytes (Müller, 1939; Jovet-Ast, 1986). In contrast to the ancestral and majority of extant liverworts that are dioicous, most Riccia species are monoicous. The Riccia sporophyte is dramatically reduced, lacking an elaborate foot, seta elongation, any specialized mechanism for capsule rupture and elaters, all of which are ancestral characters in liverworts and the Marchantiopsida, with the latter three characters all linked to spore dispersal. Thus, in Riccia, the mature sporophyte is embedded in the gametophytic thallus (cleistocarpy), and spores are released only upon disintegration of the sporophyte capsule wall and subsequent degeneration of the maternal gametophyte. Consistent with a lack of dispersal mechanism, spores of most Riccia are large and can remain dormant for an extended period, often years. Sister to the Riccia genus is the monotypic Ricciocarpos natans, whose sporophyte is also morphologically reduced. Whether the reductions occurred independently in the two genera is unknown.

The Riccia genus is most diverse in Mediterranean-type climates, where growth is often ephemeral or especially with seasonal precipitation (Schuster, 1992; Perold, 1995; Jovet-Ast, 2000; Cargill et al., 2016). Many species of Riccia colonize disturbed habitats and act as pioneer species, while others are specialized for colonization of harsh ephemeral dry habitats, perhaps the least likely places where one expects to find plants that rely on aquatic fertilization. To cope with such extremes, two divergent life cycle strategies have evolved. Some species inhabiting regions wherein the growth period is ephemeral and unpredictable have evolved desiccation tolerance and the ability to lie dormant for time periods spanning years. By contrast, other species are rapid annuals wherein the entire spore-to-spore life cycle occurs in as little as 2 months. While the majority of Riccia species are monoicous, dioicous Riccia species exist, indicating that there have been shifts between the two sexual systems within the genus.

The ancestral liverwort was likely dioicous, with females and males harbouring U and V sex chromosomes, respectively (Berrie, 1963; Bowman, 2016; Iwasaki et al., 2021). In M. polymorpha, a U-linked feminizer, MpBPCU, promotes female development via suppression of an antisense RNA transcript, encoded by MpSUF, which in turn represses the autosomal female-promoting gene, MpFGMYB (Hisanaga et al., 2019; Iwasaki et al., 2021). Thus, in the presence of MpBPCU, MpFGMYB is expressed and development as a female occurs, and in the absence of MpBPCU, MpSUF represses MpFGMYB and male differentiation ensues. The V chromosome gametolog of Basic Pentacysteine on the U (BPCU), called Basic Pentacysteine on the V (BPCV), promotes the transition to reproduction (as does BPCU), but lacks any sex-determining function (Iwasaki et al., 2021). While U-linked feminizers have been identified in phylogenetically diverse liverworts, for example in Marchantia, Sphaerocarpos and Pellia genera (Haupt, 1932; Knapp, 1935; Lorbeer, 1936, 1938; Burgeff, 1937; Heitz, 1949), it is not known whether feminizers in these genera are orthologous (Bowman, 2016). The evolution of monoicy, where both female and male sex organs develop on a single individual, from dioicy could also be considered a form of regressive evolution. Monoicy has evolved multiple independent times within liverworts, possibly in response to the occupying of ephemeral habitats (Bischler, 1998). While there may be multiple mechanisms to evolve monoicy, for example polyploidization whereby a species inherits both U and V chromosomes, in the one case examined to date, Ricciocarpos natans, monoicy evolved via aneuploidy, where both the ancestral U and V chromosomes were retained, albeit with the ancestral V largely intact and fragments of the ancestral U fused to previously autosomal chromosomes (Singh et al., 2023). In Ricciocarpos natans, both BPCU and BPCV, as well as SUF, were retained, suggesting that developmental regulation of BPCU allows switching between female and male development (Singh et al., 2023).

To investigate the evolution of sexual systems within Riccia and to examine whether genomic signatures of regressive evolution exist in Riccia, we analysed the genomes of Riccia sorocarpa (Krawczyk et al., 2025) and Riccia fluitans, along with the transcriptomes of Riccia cavernosa (Dong et al., 2019b), Riccia huebeneriana (Shen et al., 2024) and Riccia anguillarum. The former four Riccia species have a karyotype of eight chromosomes, one of which is quite small and is referred to as an ‘m’ chromosome (Siler, 1934; Krawczyk et al., 2025). This karyotype is widespread in the genus (Müller, 1941; Tatuno, 1956; Berrie, 1964; Jovet-Ast, 1970; Na-Thalang, 1980; Bornefeld, 1984; Fritsch, 1991; Bischler-Causse et al., 1995; Akiyama & Odrzykoski, 2020). Both R. sorocarpa and R. cavernosa are widespread, monoicous, short-lived annuals occupying temporary habitats throughout much of the northern and southern hemispheres (Schuster, 1992). R. anguillarum is a recently described dioicous species endemic to Western Australia (Cargill et al., 2024), and R. huebeneriana is a related monoicous species widespread in the northern hemisphere (Schuster, 1992). R. fluitans, a common aquarium plant available world-wide, is sterile but is reported to be dioicous because when reproductive structures have been induced, only archegonia were observed (Selkirk, 1979). These Riccia species span most of the phylogenetic breadth of the genus, allowing general conclusions to be formulated.