Evolutionary history of sex and accessory chromosomes in hornworts

Abstract

Hornworts represent one of the six major lineages of land plants already present in the Devonian (Villarreal et al., 2010; Frangedakis et al., 2021, 2023). As one of the three lineages of bryophytes, the hornwort life cycle is gametophyte-dominant, with the sporophyte dependent on the maternal gametophyte. As with the other lineages of bryophytes, the mosses and liverworts, dioicy (separate sexes) is prevalent, with c. 40% of hornwort species being dioicous, with the other 60% monoicous (Villarreal & Renner, 2013). In cases that have been examined, sex is often correlated with the presence of sex chromosomes, with females harboring a U chromosome and males a V chromosome, and the sporophyte having a UV karyotype. The phylogenetic distribution of sexual systems within hornworts suggests multiple independent transitions between sexual systems, with the ancestral condition being equivocal, although monoicy has been suggested (Villarreal & Renner, 2013). However, similar statistical analyses of sexual systems across extant moss species predicted patterns of transition that were not supported by the subsequent phylogenetic analysis of genes residing on the sex chromosomes (McDaniel et al., 2013; Carey et al., 2021). Thus, even statistically supported reconstructions based on extant taxa may be suspicious. In Micheli's original description of Anthoceros major (now Phaeoceros laevis (L.) Prosk.), separate female and male plants are depicted (Micheli, 1729). Subsequent descriptions confirmed the dioicy of this and several other hornwort species (Rink, 1935; Proskauer, 1948, 1957a). Also similar to sex determination in liverworts and mosses, the presence of specific chromosomes is correlated with sex in hornworts. The hornwort sex chromosomes are usually small in relation to the autosomes, with the typical karyotype of dioicous hornworts being four or five large autosomes and one small sex chromosome (Heitz, 1927; Tatuno, 1934, 1941; Rink, 1935; Proskauer, 1948, 1957b; Segawa, 1956). While little is known about the genetic basis of sex determination in hornworts, Rink reported that gametophytes derived from diploid spores (harboring both U and V chromosomes) of dioicous Anthoceros sampalocensis (Burgeff) Prosk. only developed archegonia, suggesting the existence of a U-chromosome-dominant ‘feminizer’ in this species (Rink, 1935). Monoicous hornworts have a similar karyotype to the dioicous species, with five or six chromosomes and one or two of the chromosomes being small compared with the others. In monoicous species, the small chromosomes have been referred to as accessory chromosomes, and Proskauer suggested that these accessory chromosomes may have descended from ancestral sex chromosomes in a transition from dioicy to monoicy (Proskauer, 1948). Recent chromosome-level, telomere-to-telomere assemblies of several hornwort species have allowed a molecular examination of sex and accessory chromosomes (Li et al., 2020; Zhang et al., 2020; Dong et al., 2025; Schafran et al., 2025). A comparative phylogenomic analysis across extant hornwort diversity found little evidence for synteny between either sex chromosomes or sex chromosomes and accessory chromosomes of six hornwort species, concluding that the chromosomes do not have a shared evolutionary history (Schafran et al., 2025). Here, we re-examined these chromosomes and suggest that sex chromosomes in several dioicous hornwort lineages are homologous with each other and also with at least some accessory chromosomes in monoicous hornworts. Furthermore, FEMALE GAMETOPHYTE MYB (FGMYB) is proposed to be a candidate ancestral sex-determining gene in hornworts. For hornwort species for which a sequenced genome is available (Table 1), the respective genome, coding sequence (CDS) and protein sequences were downloaded from either www.hornwortbase.org (Schafran et al., 2025) or bryogenomes.org (Dong et al., 2025). For hornwort species for which a transcriptome is available (Leebens-Mack et al., 2019; Shen et al., 2025), the raw reads were downloaded from the databases listed in Table 1. The transcriptome sequences were assembled using TRINITY (Grabherr et al., 2011) and the CDS extracted (Haas et al., 2013) on the Galaxy server (usegalaxy.org.au). BLASTp and tBLASTn (Camacho et al., 2009; Cock et al., 2015) were employed to identify homologs of genes encoded on the putative Leiosporoceros dussii V chromosome (Schafran et al., 2025). Chromosome numbers are derived from Proskauer (1957b) and Fritsch (1991). De novo whole-genome assembly of putative opposite sexes for the three hornwort species was performed using Abyss using Illumina short reads (150-bp paired) on the trace archive sequences listed in Table 1. Supporting Information Table S1 outlines the assembly statistics of the draft genomes for Leiosporoceros dussii, Phaeomegaceros chiloensis and two isolates of Phymatoceros phymatodes. N50 suggests that Ldussi and Pphyma1 assemblies are poor and fragmented, while Pchilo and Pphyma2 assemblies are more contiguous. To better enrich hornwort sex chromosome sequences, sequences likely representing the sex chromosomes were also assembled de novo. Briefly, the entirety of the Illumina short reads was first aligned to their respective reference genomes (minus the putative sex chromosomes), outputting read pairs where both reads were unaligned. These unaligned pairs representing sequences not found in the reference gene autosomes were then assembled de novo to obtain a putative sex chromosome assembly for the opposite sexes. Assembled chromosome sizes in Table S1 are likely an overestimation due to the presence of reads that have organellar origins and ambiguous repetitive regions. Ldussi, Pchilo and Pphyma2 contigs longer than one kilobase are provided in Datasets S1– S3. Primary protein transcripts on accessory chromosomes of Anthoceros agrestis ‘Oxford’, Notothylas orbicularis and Paraphymatoceros pearsonii were mapped to their corresponding Leiosporoceros dussii homolog via reciprocal Blastp (v.2.15.0) searches (-evalue 1e-20; Camacho et al., 2009). Chromosomal identities of these accessory chromosome gene homologs within the male L. dussi genome were parsed in R Studio (R Core Team, 2024), and the observed gene distributions across the six L. dussi chromosomes were tested for overrepresentation via a binomial test (alternative = ‘greater’). Complete or partial coding nucleotide sequences were manually aligned as amino acid translations using AliView (https://ormbunkar.se/aliview/). Ambiguously aligned sequences were removed, and alignments of nucleotides were employed in the subsequent phylogenetic analysis. The graphical representation of the trees was generated using the FigTree (v.1.4.0) software (http://tree.bio.ed.ac.uk/software/figtree/). Sequence alignments can be provided upon request. The best substitution models for each of the alignments were obtained using the ‘optimize’ function in raxmlGui v.2.0 (Edler et al., 2020). A maximum likelihood phylogeny was constructed, with nodal support calculated after 1000 replications, using raxmlGui v.2.0 (Edler et al., 2020). The phylogenetic relationships of hornwort species for which there exists genome sequence, some telomere-to-telomere (T2T) or transcriptiome sequence are shown in Fig. 1. These species span the known extant phylogenetic diversity of hornworts and include three dioicous species (Leiosporoceros dussii, Phymatoceros phymatodes and Phaeomegaceros chiloensis) for which T2T genome sequences have been assembled. While L. dussi has been previously placed as the sister to all other extant hornworts (Duff et al., 2007; Villarreal et al., 2015), gene tree incongruences, possibly due to gene flow or incomplete lineage sorting at the base of extant hornworts, suggests that L. dussi could instead be sister to the Anthocerotaceae (Peñaloza-Bojaca et al., 2025). Thus, the tree in Fig. 1 is drawn as a polytomy and in our analysis Leiosporoceros dussi as sister to all other dioicous liverworts analyzed. We first performed Blastp with the protein set of Leiosporoceros dussii Chromosome s6, the putative V chromosome (Schafran et al., 2025), against the Marchantia polymorpha (L.) v.7.1 protein complement (Tanizawa et al., 2025) and identified 84 genes with homologs in both hornworts and liverworts (Table S2). We performed this first step as a filter, since in liverworts many genes on the sex chromosomes appear to be recent immigrants (Bowman et al., 2017). We examined these 84 genes and eliminated those that encode genes of repetitive elements and those that appear to be recent immigrants into the sex chromosome, as described previously for the M. polymorpha sex chromosomes (Yamato et al., 2007; Bowman et al., 2017; Montgomery et al., 2020; Iwasaki et al., 2021), leaving 29 conserved genes (Table S3). This protein dataset was then used to interrogate the genomes of Phymatoceros phymatodes and Phaeomegaceros chiloensis, both of which have sex chromosomes (Schafran et al., 2025), identifying Genes 9 and 8 (of the 29 genes), respectively, that are conserved on the sex chromosomes (Table 2). We then examined whether these nine genes were present on the accessory chromosomes of Paraphymatoceros pearsonii, Notothylas orbicularis and Anthoceros agrestis, finding that Genes 9, 5 and 3, respectively, are found on the accessory chromosomes of these species (Table 2). While we identified nine gametologs that are shared across all three species for which sex chromosomes were assembled (Leiosporoceros dussii, Phaeomegaceros chiloensis and Phymatoceros phymatodes), an additional five potential gametolog pairs were identified in comparisons between the V and putative U of L. dussii (Table S4). Based on all vs all BLAST comparisons, one of these five may also be located on the sex chromosome in Phaeomegaceros chiloensis (Table S5). Finally, based on all vs all BLAST comparisons, eight potential gametolog pairs were identified as being shared between Phaeomegaceros chiloensis and Phymatoceros phymatodes, but these were not on the UV of Leiosporoceros dussii (Table S6). Further investigation of the evolution of these potential gametolog pairs could illuminate additional aspects of hornwort sex chromosome evolution. To examine whether the other accessory chromosomes also retained genes form the putative ancestral sex chromosome, genes located on the accessory chromosomes of Anthoceros agrestis ‘Oxford’, Notothylas orbicularis and Paraphymatoceros pearsonii were mapped to their corresponding homologs in the male Leiosporoceros dussii reference genome to visualize their distributions across Leiosporoceros dussii chromosomes. A binomial test (Fig. S1) showed that across all three species, homologs of accessory chromosomes genes are significantly overrepresented on the male sex chromosome (Chromosome 6) of Leiosoroceros dussi. A significant proportion of homologs from genes on the accessory chromosomes of Notothylas orbicularis and Paraphymatoceros pearsonii also mapped to Chromosome 4 of Leiosporoceros dussii, suggesting a translocation from ancestral Chromosome 4 with the accessory chromosome early in the Notothyladaceae lineage. The low intersection of Chromosome 4 homologs on the accessory chromosomes between N. orbicularis and P. pearsonii (18/84 unique homologs) suggests either multiple independent chromosomal rearrangement events or, alternatively, an accompanying duplication followed by degeneration. While other explanations, such as a fission of ancestral Chromosome 4 and subsequent migration of genes from ancestral sex chromosomes to this new chromosome, are possible, we consider them unlikely given the relative stability of hornwort karyotypes. To identify gametologs from the opposite sex of the three dioicous species (Leiosporoceros dussii, Phymatoceros phymatodes and Phaeomegaceros chiloensis), we assembled putative sex chromosome sequences from the opposite sex genomes for each of the species. These assemblies are not T2T, but are highly fragmented as expected for sex chromosomes containing mostly repetitive DNA. Despite this, putative gametolog sequences were readily identified as most genes were found in the small number of contigs larger than 10 kb. Phylogenetic trees of each of the gametologs were constructed using the above sequences as well as orthologs obtained from other available hornwort transcriptomes (Leebens-Mack et al., 2019; Shen et al., 2025) and hornwort genome sequences (Dong et al., 2025). The phylogenetic patterns of the gametolog pairs suggest that at least six, and possibly eight, gametologs have been in the nonrecombining region of the sex chromosomes since the origin of extant hornwort diversity (Dataset S4). While Leiosporoceros dussii and Phymatoceros phymatodes presumed V chromosomes share nine genes, only eight were likely in nonrecombining regions, with Leiosporoceros dussii S6G002100 in a pseudoautosomal region (Fig. 2). Furthermore, based on the phylogenetic distribution of gametologs, we surmise that the genome assembly of Phaeomegaceros chiloensis was derived from a female. Finally, there is little synteny of gametologs between the V chromosomes of Leiosporoceros dussii and Phymatoceros phymatodes or between their V chromosome and the accessory chromosome of Paraphymatoceros pearsonii that harbors orthologs of all nine gametologs (Fig. 2). Thus, over the time frame of extant hornwort evolution, chromosomal rearrangements have scrambled the ancestral gene order. That sex/accessory chromosomes spanning extant hornwort diversity share these nine homologs (or subsets thereof) strongly suggests these sex and accessory chromosomes are homologs. The accessory chromosomes retain vestiges of their evolutionary history as old sex chromosomes, similar to what has been observed in derived monoicous liverworts where the micro-chromosome of monoicous Ricciocarpos natans (L.) Corda retains characteristics of the sex chromosome from which it was derived (Singh et al., 2023). Also similar to what has been observed in liverworts (Singh et al., 2023; Levins et al., 2025), monoicous hornworts in different lineages (e.g. Anthocerotaceae, Notothyladaceae, Phymatocerotaceae and Dendrocerotaceae) have retained different patterns, that is derived from either the ancestral female or the ancestral male, of gametologs (Table 2). These patterns suggest independent gene losses in the different lineages and imply independent transitions from ancestral dioicy to derived monoicy. Some gametologs have been retained in all three dioicous species (Leiosporoceros dussii, Phymatoceros phymatodes and Phaeomegaceros chiloensis) for which there exist genome sequences, for example orthologs of TCP1, C4HDZ. By contrast, others appear to have been in the nonrecombining region of the common ancestor of extant hornworts but may have been subsequently moved to an autosome in some lineages. Finally, one gametolog ARGONAUTE1 (AGO1) appears to be in the pseudoautosomal region in Leiosporoceros dussii, but subsequently incorporated into the nonrecombining region in the ancestor of Phymatoceros phymatodes and Phaeomegaceros chiloensis. Many of the gametologs encode regulatory proteins and could thus be involved in hornwort sex determination. The most promising candidate CHASE-domain containing histidine kinase (CHK) encoding a protein related to cytokinin receptors, based on the observation that cytokinin signaling promotes aspects of female gametophyte development in other land plants (Yuan et al., 2016; Liu et al., 2017; Bao et al., 2024), does not exhibit a phylogenetic pattern consistent with it being the ancestral sex determination gene. One of the strongest candidates for a hornwort sex-determining gene on the U chromosome is FGMYB, which promotes female gametophyte development in both the liverwort Marchantia polymorpha and the angiosperm Arabidopsis thaliana (Kasahara et al., 2005; Rabiger & Drews, 2013; Hisanaga et al., 2019). While this gene is autosomal in Marchantia, it is a direct target of the U chromosome–encoded feminizer, BASIC PENTACYSTEINE ON THE U CHROMOSOME (MpBPCU) (Iwasaki et al., 2021). Furthermore, in the liverwort genus Riccia, it has been hypothesized that FGMYB acts as the female sex determination gene following the loss of BPCU in this lineage (Levins et al., 2025). It was previously reported that FGMYB was missing from the genomes of two of the three dioicous hornwort species, Leiosporoceros dussii and Phymatoceros phymatodes, but was present in Phaeomegaceros chiloensis (Schafran et al., 2025). However, since both these assemblies lacking FGMYB were derived from males and our gametolog analysis suggested that Phaeomegaceros chiloensis was derived from a female, we examined our female assemblies for the former two species for the presence of FGMYB. FGMYB orthologs can be found on the presumed U chromosome of Phaeomegaceros chiloensis and in the sequences presumably from the U chromosomes of Leiosporoceros dussii and Phymatoceros phymatodes (Fig. 1). In Notothylas orbicularis, FGMYB is found on the accessory chromosome, but in most monoicous hornwort species, FGMYB resides on a chromosome that was ancestrally autosomal (Fig. 1). A phylogenetic analysis of hornwort FGMYB sequences is consistent with accepted hornwort relationships (Fig. S2; Dataset S5). These data suggest that in an ancestral hornwort, FGMYB resided in the U chromosome and is thus a prime candidate for the sex-determining gene in this lineage. If FGMYB was the ancestral hornwort sex-determining gene and was retained in Leiosporoceros dussii, Phymatoceros phymatodes and Phaeomegaceros chiloensis, it implies multiple independent transitions from dioicy to monoicy, as proposed for the Marchantiopsida liverworts (Singh et al., 2023; Fu et al., 2025; Levins et al., 2025; Potente et al., 2025). In liverworts, during the transition from dioicy to monoicy there is a bias toward retaining the V chromosome, with essential U-chromosome genes translocated to autosomes (Singh et al., 2023; Fu et al., 2025; Levins et al., 2025; Potente et al., 2025); whether this is also the case in hornworts remains to be investigated. In the Notothyladaceae, FGMYB is missing in monoicous Paraphymatoceros pearsonii (Schafran et al., 2025) and we were unable to find FGMYB in dioicous Phaeoceros laevis, but it was found in the other species within the Notothyladaceae (Fig. 1). As dioicous Phaeoceros laevis is nested within a clade of monoicous species (Villarreal & Renner, 2013), a derived dioicous condition for this species is plausible. Thus, for at least some monoicous and dioicous species in the Notothyladaceae, FGMYB may be dispensable, suggesting the evolution of a new sex determination gene (and possibly new sex chromosomes). Likewise, in the Anthocerotaceae, that both Folioceros and Anthoceros monoicous species exhibit similar phylogenetic patterns of gametologs (Table 2) might suggest a transition to monoicy early in this lineage with a subsequent reversion(s) to dioicy in some lineages. An extension of this scenario is that the dominant ‘feminizer’ observed in Anthoceros sampalocensis (Rink, 1935) may be of derived origin. These speculative hypotheses can now be investigated once T2T genome sequences of the relevant taxa become available. The hypothesized transitions from ancestral dioicy to monoicy and then back to dioicy suggested here for hornworts are reminiscent of a similar scenario described in the liverwort genus Riccia, wherein an ancestral dioicous species transitioned to monoicy and then re-evolved dioicy, likely by replacing the ancestral sex determination gene (BPCU) with its downstream target (FGMYB), and could provide a conceptual model for these hornwort lineages (Levins et al., 2025). In summary, we provide evidence for the evolutionary relationship between at least some sex and accessory chromosomes in hornworts and suggest that FGMYB was the ancestral sex-determining gene in this land plant lineage, while in some hornwort lineages, a new sex determination mechanism appears to have evolved. These data lend support to the idea that FGMYB was the female sex-determining gene in the ancestral land plant, which has been speculated to have been dioicous (Bowman, 2022). We thank Tom Yamato and Yang Liu for their comments on this manuscript. This work was supported by funding from the Australian Research Council Centre of Excellence for Plant Success in Nature and Agriculture (CE200100015). Open access publishing facilitated by Monash University, as part of the Wiley - Monash University agreement via the Council of Australian University Librarians. Open access publishing facilitated by Monash University, as part of the Wiley - Monash University agreement via the Council of Australian University Librarians. JLB designed the research. JLB and JL analyzed the data and wrote the manuscript. The data that support the findings of this study are available as Datasets S1–S5. Dataset S1 Putative Leiosporoceros dussii U-chromosome scaffolds longer than 1 kb. Dataset S2 Putative Phymatoceros phymatodes U-chromosome scaffolds longer than 1 kb. Dataset S3 Putative Phaeomegaceros chiloensis V chromosome scaffolds longer than 1 kb. Dataset S4 Hornwort gametolog trees. Dataset S5 Hornwort FGMYB sequences. Fig. S1 Distribution of genes of accessory chromosomes of Anthoceros agrestis ‘Oxford’, Notothylas orbicularis, and Paraphymatoceros pearsonii across chromosomes of Leiosporoceros dussii. Fig. S2 Phylogram of hornwort FGMYB genes. Table S1 Reference and de novo whole-genome assembly stats, and sex chromosome reference and de novo assembly stats. Table S2 BLAST results of L. dussi s6 against M. polymorpha v7.1 proteome. Table S3 Edited BLAST results from Table S2. Table S4 BLAST results of Leiosporoceros conserved proteins from Table S3 vs presumed Leiosporoceros U-chromosome sequences. Table S5 Possible gametologs shared between Phaeomegaceros chiloensis and Phymatoceros phymatodes sex chromosomes. Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.

Abstract from the paper’s OpenAlex record; open copy licensed cc-by-nc. Read at the source ↗

The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.