SHOT GLASS, an R2R3‐MYB transcription factor, promotes gemma cup and gametangiophore development in Marchantia polymorpha

Summary

• Many plants reproduce asexually by generating clonal progeny from vegetative tissues, a process known as vegetative reproduction. This reproduction mode contrasts with sexual reproduction, which enhances genetic diversity.

• The bryophyte Marchantia polymorpha L. adjusts its reproductive strategy in response to seasonal environmental cues, transitioning between vegetative and sexual reproduction. In this study, we identified a gene encoding the R2R3‐MYB transcription factor SHOT GLASS (MpSTG) as a critical regulator of gemma cup development. MpSTG was predominantly expressed in the gemma cup, apical notch, and sexual reproductive organs (gametangiophores). MpSTG mutation resulted in the formation of abnormal shot‐glass‐shaped structures lacking gemmae, which replaced functional gemma cups. Additionally, MpSTG‐disrupted plants failed to develop sexual reproductive organs, even under inductive conditions.

• In Arabidopsis thaliana, the MpSTG ortholog LATERAL ORGAN FUSION1 (AtLOF1) plays a pivotal role in lateral bud formation. We demonstrated that MpSTG can partially compensate for AtLOF1's function in lateral bud formation in A. thaliana.

• Our findings suggest that MpSTG is a key regulator of vegetative and sexual reproduction in M. polymorpha, and illustrate that evolutionarily conserved developmental mechanisms may function in both the gametophyte generation of bryophytes and the sporophyte generation of angiosperms.

Introduction

Plants reproduce to ensure the transmission of genetic material to future generations via sexual or asexual means. Sexual reproduction involves gamete fusion to produce offspring genetically distinct from the parents. Conversely, asexual reproduction bypasses gamete fusion, resulting in clonal offspring genetically identical to the parent plant. Vegetative reproduction is a prevalent form of asexual reproduction in plants, which occurs when new individuals develop from plant fragments, cuttings, or specialized reproductive structures (de Kroon & van Groenendael, ¹⁹⁹⁷; Klimešová & Klimeš, ²⁰⁰⁷). Notably, c. 70% of the plant species in temperate regions are capable of vegetative reproduction (de Kroon & van Groenendael, ¹⁹⁹⁷).

However, the molecular mechanisms underlying vegetative reproduction remain largely unexplored and are thought to differ from those governing sexual reproduction. Many plant species exhibit both sexual and vegetative reproductive strategies, which require a tightly regulated switch between these modes in response to environmental cues. These regulations are crucial for balancing the distinct advantages and disadvantages of each reproductive strategy.

The bryophyte Marchantia polymorpha reproduces through both sexual and vegetative modes. Vegetative reproduction in M. polymorpha involves the production of clonal progeny, known as gemmae. These gemmae form within goblet‐shaped vegetative reproductive structures called gemma cups, which develop on the dorsal surface of the flattened plant body (the thallus). Each gemma originates from a single initial cell located in the basal floor epidermis of the gemma cup.

The dorsal surface of M. polymorpha thallus is covered with assimilatory structures called air chambers. During vegetative growth, periodic gemma cup formation occurs and replaces air chambers. Air chamber formation is initiated by periclinal cell division in dorsal epidermal cells of the apical meristematic region (Ishizaki et al., ^2013b). Once these epidermal cells are designated as gemma initial cells, periclinal cell division is suppressed and anticlinal division promoted. Consequently, the gemma cup basal floor epidermis forms (Barnes & Land, 1908; Suzuki et al., ²⁰²⁰). As the basal floor epidermis of the gemma cup expands, surrounding dorsal tissues rise to form a rim and give the gemma cup its characteristic goblet cup shape (Barnes & Land, 1908).

Several key factors involved in gemma cup development have been identified in M. polymorpha (Kato et al., ²⁰²⁰). GEMMA CUP‐ASSOCIATED MYB 1 (MpGCAM1), an R2R3‐MYB transcription factor, acts as a critical regulator of gemma cup initiation (Yasui et al., ²⁰¹⁹). MpGCAM1 confers stem cell‐like properties to the floor epidermal cells within the gemma cup (Yasui et al., ²⁰¹⁹). The plant hormone cytokinin promotes gemma cup development via the cytokinin signaling pathway, which is a prerequisite for gemma formation (Aki et al., ²⁰¹⁹). In this pathway, the type‐B response regulator (MpRRB) positively regulates MpGCAM1 expression (Aki et al., ²⁰²²).

In addition, the KAI2‐ligand (KL) signaling, mediated by KARRIKIN INSENSITIVE2 (KAI2; Varshney & Gutjahr, ²⁰²³), promotes gemma and gemma cup development in M. polymorpha by regulating the expression of MpLONLY GUY (MpLOG), which encodes a cytokinin biosynthesis enzyme, ultimately inducing the expression of MpGCAM1 (Mizuno et al., ²⁰²¹; Komatsu et al., ^{2023, 2025}). Another regulator of MpGCAM1 expression is the Golden2, ARR‐B, and Psr1 (GARP) transcription factor MpKANADI (MpKAN), which acts independently of the cytokinin pathway (Briginshaw et al., ²⁰²²). Despite recent advances in our understanding of M. polymorpha as a model system, the molecular mechanisms governing gemma cup development remain poorly understood.

The gemma cup and gemmae are specific to species in the order Marchantiales. However, phylogenetic analyses indicate that MpGCAM1 belongs to the same subfamily as the REGULATOR OF AXILLARY MERISTEMS (RAX) genes in Arabidopsis thaliana (AtRAX) and SlBlind in tomato (Solanum lycopersicum), both of which regulate lateral bud formation in angiosperms (Yasui et al., ²⁰¹⁹). Phenotypic complementation of the MpGCAM1 gene in the Arabidopsis AtRAX1‐3 triple mutant suggests partial conservation of molecular functions between M. polymorpha GCAM1 and A. thaliana RAX genes (Yasui et al., ²⁰¹⁹). Collectively, these findings suggest a shared molecular mechanism underlying the formation of specific lateral organs in both angiosperms and Marchantiales, such as lateral shoots and gemma cups, respectively.

Marchantia polymorpha transitions from the vegetative phase to the sexual reproductive phase under long‐day and far‐red (FR)‐enriched conditions, resulting in male and female gametangia formation (Durand, 1908; Kubota et al., ²⁰¹⁴; Yamaoka et al., ²⁰²¹; Cui et al., ²⁰²³). This phase transition is an FR high‐irradiance response regulated by the red/FR photoreceptor phytochrome and its downstream transcription factor PHYTOCHROME INTERACTING FACTOR (MpPIF) (Inoue et al., ²⁰¹⁹). Conversely, MpDELLA inhibits gametangia formation through physical interactions with MpPIF (Hernández‐García et al., ²⁰²¹). Additionally, the basic helix–loop–helix transcription factor BONOBO (MpBNB) – a member of subfamily VIIIa – functions as a master regulator of gametangia formation. For example, MpBNB overexpression induces gametangia development independent of environmental cues (Yamaoka et al., ²⁰¹⁸; Saito et al., ²⁰²³). Under FR light, MpKAN can regulate gametangia formation by suppressing MpDELLA and promoting MpBNB expression (Briginshaw et al., ²⁰²²).

In this study, we identified the R2R3‐MYB transcription factor MpSHOT GLASS (MpSTG) as a key regulator of both sexual and vegetative reproduction in M. polymorpha. We investigated MpSTG functions during gemma cup and gametangiophore development and focused on genetic interactions between MpSTG and established regulators. Our findings provide valuable insights into the molecular mechanisms underlying gemma cup and gametangiophore formation in M. polymorpha. Furthermore, Marchantia STG partially rescued the loss‐of‐function phenotype of its ortholog in Arabidopsis, which supports the hypothesis that the molecular mechanisms underlying lateral shoot formation in angiosperms and gemma cup formation in M. polymorpha are at least partially conserved.

Materials and Methods

Plant materials and growth conditions

Male and female wild‐type accessions of Marchantia polymorpha L., Takaragaike‐1 (Tak‐1) and Takaragaike‐2 (Tak‐2), respectively (Ishizaki et al., ²⁰⁰⁸), were grown vegetatively. Plants were cultured on half‐strength Gamborg's B5 medium (Gamborg et al., ¹⁹⁶⁸) supplemented with 1% agar and adjusted to pH 5.5 with 2‐(N‐morpholino)ethanesulfonic acid (MES). Cultures were maintained at 22°C under continuous white Light‐Emitting Diode (LED) lighting (50–60 μmol photons m⁻² s⁻¹; VGL‐1200W; SYNERGYTEC, Tokushima, Japan). When dexamethasone (DEX)/cycloheximide (CHX) treatment was performed following mock treatment, the plants were grown on medium overlaid with cellophane to facilitate transfer (Yasui et al., ²⁰¹⁹). Sexual organ formation was induced using far red (FR) light irradiation as described previously (Chiyoda et al., ²⁰⁰⁷). Mature sporangia were collected 3–4 wk after crossing, air‐dried for 1 wk, and stored at −80°C until use.

Columbia (Col‐0) was used as the wild‐type accession for Arabidopsis  thaliana. The lof1‐1 T‐DNA insertion line (SALK_025235; Lee et al., ²⁰⁰⁹) was obtained from the Arabidopsis BioResource Center (https://abrc.osu.edu). Seeds were germinated under sterile conditions on Murashige & Skoog (MS) medium supplemented with 1% sucrose and 0.5% gellan gum adjusted to pH 5.7 with MES. Arabidopsis seedlings were cultured at 23°C under continuous light (40–50 μmol photons m⁻² s⁻¹). After 2 wk, the seedlings were transferred to potted soil and grown at 23°C under continuous light (40–50 μmol photons m⁻² s⁻¹) provided by a fluorescent lamp. Marchantia polymorpha and A. thaliana plants used in this study are listed in Supporting Information Table S1.

Reverse transcription quantitative Polymerase Chain Reaction (RT‐qPCR)

Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen), according to the manufacturer's protocol. Thereafter, total RNA (100 ng) was reverse‐transcribed in a 10 μl reaction mixture to generate cDNA using ReverTra Ace quantitative polymerase chain reaction RT Master Mix with gDNA remover (Toyobo, Osaka, Japan). Subsequently, the mixture was diluted with 40 μl of distilled water, and 5 μl aliquots were used for RT‐qPCR. For RT‐qPCR, the cDNA samples were further diluted with 220 μl of distilled water, and 2 μl aliquots were amplified on the Light Cycler Nano Real‐time PCR Detection System (Roche Applied Science, Penzberg, Germany) using KOD SYBR^TM quantitative polymerase chain reaction Mix (Toyobo). Two‐step PCR cycling was performed according to the manufacturer instructions. Primers used in the experiments are listed in Table S2. MpEF1α (Mp3g23400) and MpACT7 (Mp6g11010) were used as internal controls.

Vector construction and transformation

All vectors used in this study are listed in Table S3.

Mpstg ^ko

To generate the MpSHOT GLASS (MpSTG; Mp8g11870) gene‐targeting vector, 5′ and 3′ homologous arms (c. 4.5 kb each) were amplified from Tak‐1 genomic DNA by PCR using KOD FX Neo (Toyobo) with the primer pairs STG_5IF_L/STG_5IF_R and STG_3IF_L/STG_3IF_R (Table S2), respectively. Thereafter, the PCR‐amplified 5′ and 3′ homologous arms were cloned into the PacI and AscI sites of pJHY‐TMp1 (Ishizaki et al., ^2013a), respectively, using an In‐Fusion HD Cloning Kit (Clontech, San Jose, CA, USA). The MpSTG gene‐targeting vector (Fig. S1a) was transformed into F1 sporelings derived from sexual crosses between Tak‐1 and Tak‐2 following the method described by Ishizaki et al. (2008). The transformants were selected using 10 μg ml⁻¹ hygromycin B and 100 μg ml⁻¹ cefotaxime. Gene‐targeted lines were screened using genomic PCR, as previously described (Fig. S1a,b; Ishizaki et al., ^2013b). Primers used for screening are listed in Table S2. Male and female MpSTG knockout lines were identified by sex genotyping. MpSTG expression levels were confirmed using reverse transcription PCR (Fig. S1c).

proMpSTG‐GUS

The regulatory region of MpSTG, including a 5070‐bp sequence upstream of the start codon and a 246‐bp coding sequence downstream up to and including the codon for the third methionine (Met81) (Fig. S1a), was amplified from Tak‐1 genomic DNA via PCR using KOD‐Plus‐Neo (Toyobo) with the primer set STG_pro_F/STG_pro_R (Table S2). Thereafter, the amplified fragment was cloned into the pENTR/D‐TOPO^TM vector (Thermo Fisher Scientific, Waltham, MA, USA). This entry vector was used in an LR reaction with the gateway binary vector pMpGWB104 (Ishizaki et al., ²⁰¹⁵). The resulting proMpSTG:GUS vector (Fig. S1d) was introduced into the regenerating thalli of Tak‐1 as previously described (Kubota et al., ²⁰¹³). Finally, the transformants were selected using 10 μg ml⁻¹ of hygromycin B (Hyg) and 100 μg ml⁻¹ of cefotaxime.

gMpSTG‐Citrine/tdTomato

The genomic region of MpSTG, including a 5070‐bp sequence upstream of the start codon but excluding the stop‐codon and 3′ UTR, was amplified from Tak‐1 genomic DNA via PCR using KOD‐Plus‐Neo (Toyobo) with the primer pair STG_pro_F/STG_cds_nsR (Table S2). Thereafter, the genomic fragment was subcloned into the pENTR/D‐TOPO^TM vector (Thermo Fisher Scientific) and subsequently into the gateway binary vectors pMpGWB307 and pMpGWB329 via an LR reaction (Fig. S1d; Ishizaki et al., ²⁰¹⁵). The resulting gMpSTG‐Citrine/tdTomato vector (Fig. S1d) was introduced into the regenerating thalli of male and female MpSTG knockout lines (Mpstg ^ko #1, Mpstg ^ko #3). The transformants were selected using 0.5 μM of chlorsulfuron (CS) and 100 μg ml⁻¹ of cefotaxime.

Mpstg ^ge

Loss‐of‐function mutants of MpSTG were generated using the CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats)/Cas9 (CRISPR‐associated protein 9) genome editing system, following the protocol described by Sugano et al. (2014, 2018). A target sequence within the 5′ end of the R2R3‐MYB domain of MpSTG (Fig. S2a) was selected based on CRISPRdirect guidelines (https://crispr.dbcls.jp). Synthetic DNA oligonucleotides corresponding to the target site (STG_ge_oligo F and STG_ge_oligo R; Table S2) were annealed, inserted into the entry vector pMpGE_En03, and cloned into the destination vector pMpGE011 (Sugano et al., ^{2014, 2018}). The resulting MpSTG genome‐editing vector was introduced into regenerating thalli of transgenic plants (proMpHA9:Citrine‐GUS (Hiwatashi et al., ²⁰¹⁹), proMpRSL1:tdTomato‐NLS (Sakamoto et al., ²⁰²²), Mpgcam1 ^ko , proMpEF1α:MpGCAM1‐GR, MpGCAM1‐Citrine ^KI (Yasui et al., ²⁰¹⁹), MpBNB‐GR male/female, and MpBNB‐Citrine male/female (Yamaoka et al., ²⁰¹⁸)) via Agrobacterium tumefaciens GV2260 (Kubota et al., ²⁰¹³). Transformants were selected using 0.5 μM of CS and 100 μg ml⁻¹ of cefotaxime. Genomic DNA sequences around the target site were analyzed using an Applied Biosystems^TM 3130xl Genetic Analyzer (Thermo Fisher Scientific). Genome‐edited lines were named following the nomenclature rules for M. polymorpha (Bowman et al., ²⁰¹⁶), as shown in Fig. S2(b) and Table S1.

MpSTG‐GR

Briefly, the MpSTG coding sequence was amplified from the Tak‐1 cDNA pool generated from the total RNA using ReverTra Ace (Toyobo). PCR amplification was performed using KOD‐Plus‐Neo (Toyobo) with the STG_cds_F/STG_cds_nsR primer set (Table S2). Thereafter, the amplified sequence was cloned into a pENTR/D‐TOPO^TM vector (Thermo Fisher Scientific). The entry vector was subsequently used in a Gateway LR reaction (Thermo Fisher Scientific) with gateway binary vectors pMpGWB113 (Hyg) and pMpGWB313 (CS), respectively (Fig. S1d; Ishizaki et al., ²⁰¹⁵). The proMpEF:STG‐GR (Hyg) vector was introduced into the regenerating thalli of Tak‐1 cells, as previously described (Kubota et al., ²⁰¹³). Transformants were selected using 10 μg ml⁻¹ of hygromycin B and 100 μg ml⁻¹ of cefotaxime. Similarly, the proMpEF:STG‐GR (CS) vector was introduced into the regenerating thalli of Mpgcam1 ^ko plants (Yasui et al., ²⁰¹⁹). Transformants were selected using 0.5 μM of CS and 100 μg ml⁻¹ of cefotaxime. At least five transformants from each vector were analyzed to examine the effects of DEX treatment.

pro35S:AtLOF1 and pro35S:MpSTG

Briefly, the coding sequences of AtLOF1 (At1g26780.2) and MpSTG were amplified using KOD‐Plus‐Neo (Toyobo) with the primer sets AtLOF1_F/AtLOF1_sR and STG_cds_F/STG_cds_sR, respectively. Thereafter, the amplified sequences were cloned into a pENTR/D‐TOPO^TM vector (Thermo Fisher Scientific). These coding sequences were subsequently transferred to the Gateway binary vector pGWB2 (Nakagawa et al., ²⁰⁰⁷) using the Gateway LR reaction (Thermo Fisher Scientific). The resulting binary vectors were introduced into Arabidopsis lof1‐1 (SALK_025235) plants using A. tumefaciens (strain C58MP90), as previously described (Goto et al., ²⁰²³). Transgenic lines were selected on the MS medium containing 10 μg ml⁻¹ of hygromycin B. T3 homozygous plants obtained through self‐pollination were used for analysis.

Histology and light microscopy

Images of whole plant bodies, gemma cups, and air chambers were captured using a digital camera (Tough TG‐5; Evident (formerly Olympus), Tokyo, Japan), digital stereoscopic microscope (DMS1000; Leica Microsystems, Heerbrugg, Switzerland), digital microscope (VHX‐5000; Keyence, Osaka, Japan), and M205 FA stereoscopic microscope (Leica Microsystems) equipped with a Charge‐Coupled Device (CCD) camera (DFC7000 T; Leica Microsystems).

Morphology of paraclade junctions in A. thaliana was observed as follows. Plants were grown according to the method described above (see ‘Plant materials and growth conditions’ in the Materials and Methods section). After cultivation on the soil for 10–20 d, the floral buds were removed from the bolting primary shoot (c. 10 cm in length). Ten days after floral bud removal, images were taken at the paraclade junctions using a digital stereoscopic microscope (DMS1000; Leica Microsystems).

For plastic‐embedded sectioning, plant samples were fixed in a solution containing 3.7% formaldehyde, 50% ethanol, 0.5% acetic acid, and 0.025% glutaraldehyde, evacuated using a water aspirator until they sank, followed by incubation at 4°C for 2 d. Thereafter, the samples were dehydrated using a graded ethanol series, embedded in Technovit 7100 plastic resin (Kulzer, Wertheim, Germany), and cut into 8‐μm thick sections using a microtome (HM335E; Leica Microsystems) equipped with a tungsten carbide microtome blade (SH35W; Feather, Osaka, Japan). After staining with 0.2% Toluidine Blue O, the sections were observed under a light microscope (BX51; Evident). For scanning electron microscopy, plant samples were frozen in liquid nitrogen and directly observed using a VHX‐D500 microscope (Keyence).

β‐Glucuronidase (GUS) staining and imaging

For histochemical GUS staining, transgenic plants expressing GUS reporters (proMpSTG:GUS, proMpHA9:Citrine‐GUS, and Mpstg ^ge /proMpHA9:Citrine‐GUS) were grown on half‐strength B5 medium supplemented with 1% agar under continuous white light for various durations as described above (see ‘Plant materials and growth conditions’ in the Materials and Methods section). Thereafter, the plants were incubated in GUS staining solution (0.5 mM potassium ferrocyanide, 0.5 mM potassium ferricyanide, 10 mM EDTA, 50 mM sodium phosphate buffer (pH 7.2), 0.01% (v/v) Triton X‐100, and 1 mM X‐Gluc) at 37°C for 24 h and made transparent with 70% ethanol. For plastic‐embedded sectioning, GUS‐stained samples preserved in 70% ethanol were embedded in plastic resin and sectioned into semi‐thin sections (8‐μm thick) as described in the ‘Histology and light microscopy’ in the Materials and Methods section. Finally, the sections were stained with 0.01% Safranin and observed under a light microscope (BX51; Evident).

Subcellular localization analysis

To analyze the subcellular localization of MpSTG‐Citrine, MpBNB‐Citrine, and MpSTG‐tdTomato, fluorescence from Citrine/tdTomato and chloroplasts was observed using confocal laser scanning microscopes FV1000 (Evident) or STELLARIS5 (Leica Microsystems). The excitation and detection wavelength ranges were set as follows: Citrine fluorescence, 515 and 527 nm, respectively; tdTomato fluorescence, 559 and 591–630 nm, respectively; and chloroplast autofluorescence, 559 and 650–750 nm, respectively.

Functional induction of transcription factors using DEX

For phenotypic analysis of transgenic plants overexpressing glucocorticoid receptor (GR)‐conjugated transcription factors, plants were grown on a cellophane sheet placed over half‐strength Gamborg's B5 medium supplemented with 10 μM DEX or 0.1% dimethyl sulfide (mock) under vegetative conditions. The gemmae of proMpEF1α:MpSTG‐GR were examined for 14 d under DEX and mock treatment or for 7 d under mock treatment followed by 7 d under DEX. Apical explants of proMpEF1α:GCAM1‐GR and Mpstg ^ge /proMpEF1α:MpGCAM1‐GR plants were examined for 14 d in the presence or absence of DEX. Apical part‐excluded explants from proMpEF1α:MpSTG‐GR and proMpEF:MpSTG‐GR/Mpgcam1 ^ko plants were analyzed for 9 d. The apical explants of male and female MpBNB‐GR and Mpstg ^ge /MpBNB‐GR plants were examined for 3 wk.

RNA‐seq

Gemmae of proMpEF1α:MpSTG‐GR plants were grown on cellophane sheets placed over half‐strength Gamborg's B5 medium under vegetative conditions for 7 d. Thereafter, the cellophane sheets were transferred to fresh half‐strength Gamborg's B5 medium supplemented with 10 μM DEX or 0.1% dimethyl sulfoxide (mock treatment). After 0, 2, and 8‐h treatments with DEX or mock, 100 mg of fresh tissue (five to seven plants) was collected, immediately frozen in liquid nitrogen, and stored for subsequent analysis. Biological replicates (n = 3) were prepared for each condition.

Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen) and treated with DNase I (RNase‐Free DNase Set; Qiagen). RNA was further purified using the RNeasy MinElute Cleanup Kit (Qiagen) following the manufacturer's protocols. Sequence libraries were constructed using the TruSeq Stranded mRNA Library Prep Kit (Illumina, San Diego, CA, USA) according to the manufacturer's instructions. RNA sequencing was performed on a NovaSeq6000 platform (Illumina) to generate 101‐base paired‐end data. All sequencing data were deposited in the DDBJ Sequence Read Archive, and are available under the specified accession nos. DRR639648–DRR639659.

The sequence data were assessed for quality using FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), processed using the Cutadapt program (Martin, 2011), and aligned to the Marchantia genome v.6.1, which was obtained from the public database MarpolBase (http://marchantia.info/; Bowman et al., ²⁰¹⁷), using Hisat2 (Kim et al., ²⁰¹⁹). The read counts were generated using feature counts (Liao et al., ²⁰¹⁴). Thereafter, the transcripts per million (TPM) for each gene were calculated (Table S4). To identify differentially expressed genes (DEGs), mapped reads from different samples were analyzed using edgeR (Robinson et al., ²⁰¹⁰), and the results are presented in Table S5.

Phylogenetic analysis of MpSTG

For the phylogenetic analysis of MpSTG and its homologous R2R3‐MYBs, sequence information was retrieved from Phytozome and NCBI using the accession numbers reported in previous studies (Table S8; Naz et al., ²⁰¹³; Jiang & Rao, ²⁰²⁰). Multiple sequence alignments of the amino acid sequences of MpSTG and its homologous R2R3‐MYBs were performed using the Muscle program (Edgar, 2004) implemented in SnapGene v.6.2 with default parameters. For phylogenetic analysis, we extracted and aligned the conserved R2R3‐MYB domain region from each protein sequence and constructed the tree based on this alignment. A maximum likelihood phylogenetic tree was constructed using IQ‐Tree v2.2.6 (Minh et al., ²⁰²⁰). ModelFinder, implemented in IQ‐Tree, was used to select the best‐fit substitution model according to the Bayesian information criterion. Branch support was estimated using 1000 ultrafast bootstrap replicates. The resulting phylogenetic tree was visualized using Molecular Evolutionary Genetics Analysis (Mega) v.11.0.13 (Kumar et al., ²⁰¹⁸; Tamura et al., ²⁰²¹).

Results

MpR2R3‐MYB5 is highly expressed in gemma cups and reproductive organs

MpR2R3‐MYB5 (gene ID: Mp8g11870) has been previously identified as one of the genes highly expressed in gemma cups (Yasui et al., ²⁰¹⁹). RT‐qPCR confirmed that MpR2R3‐MYB5 expression was higher in the gemma cups on 2‐wk‐old thalli than that in the apical region or 1‐wk‐old whole thalli, before gemma cup formation, and MpR2R3‐MYB5 was highly expressed in the female reproductive organs (Figs 1a, S3).

Fig. 1.

Expression profile of MpR2R3‐MYB5 in Marchantia polymorpha. (a) RT‐qPCR to detect MpR2R3‐MYB5 (Mp8g11870) expression in vegetative and reproductive tissues; 1‐wk thallus (Th), apical notches (AN) and gemma cup (GC) in 3‐wk thallus, and male (mRO) and female (fRO) reproductive organs. MpELONGATION FACTOR1α (MpEF1α) and MpACTIN8 (MpACT8) genes were used as internal references. White dots indicate each data point. The box plot illustrates the interquartile range (IQR), spanning from the first quartile (Q1) to the third quartile (Q3). The horizontal line within the box marks the median (second quartile, Q2). The whiskers extend to the minimum and maximum values within 1.5 times the IQR from Q1 and Q3, respectively. n = 4. Alphabets indicate significant differences (one‐way ANOVA followed by Tukey's HSD test, P < 0.05). (b–j) Histochemical GUS assay of a representative proMpSTG:GUS lines. (b) Three‐week old thallus. (c) Apical region with a gemma cup on the 3‐wk‐old thallus. (d, e) Transverse sections of developing and mature gemma cups on a 3‐wk‐old thallus. (f) Magnified image of dashed square in (e). Male reproductive organ (antheridiophore) in early stage (g) and developing stage (h). Female reproductive organ (archegoniophore) in early stage (i) and developing stage (j). dGe, developing gemma; Fl, floor cell; Mu, mucilage cell. Bars: (b) 5 mm; (c, h, j) 1 mm; (d, f) 100 μm; (e, g, i) 500 μm.

To further characterize the expression profile of MpR2R3‐MYB5, we analyzed transgenic M. polymorpha lines expressing a GUS reporter under the control of the Mp8g11870 promoter. The GUS reporter construct was generated by translationally fusing a 5‐kb genomic region upstream of Mp8g11870, including its 5′ UTR (Fig. S1a). GUS staining patterns were examined across the five transformants and a generally consistent pattern was observed. Among the five independent transgenic lines, we selected a line for detailed analysis based on its consistent and representative GUS expression patterns.

Obvious promoter activity was not observed in the gemmae and 1‐wk‐old gemmalings (Fig. S4a,b). In mature thalli containing gemma cups, significant promoter activity was observed in the apical notches and the midribs and smaller gemma cups near the apical notches (Figs 1b,c, S4c–g). Gemma cups in M. polymorpha originate from the dorsal epidermal cells at the meristem notch. The gemma cups that were located closer to the notch at earlier developmental stages exhibited strong GUS signals in floor cells, mucilage cells, and developing gemmae (Figs 1d, S4d). However, GUS signals were restricted to the peripheral region of the floor, in addition to mucilage cells and developing gemmae located at the margins, in larger mature gemma cups located farther from the apical notches (Figs 1e,f, S4e). Further, MpR2R3‐MYB5 expression became spatially restricted and declined in intensity as gemma cups matured (Fig. S4f).

After transition to the sexual reproductive phase, GUS expression was additionally observed in the male and female reproductive organs (antheridium and archegonium) (Figs 1g–j, S4h–k). Collectively, these results suggest that MpR2R3‐MYB5 is involved in the early developmental stages of the gemma cups during the vegetative phase and gametangia formation during the sexual reproductive phase.

MpR2R3‐MYB5/SHOT GLASS is essential for gemma cup formation

To investigate the functions of MpR2R3‐MYB5, we generated knockout mutants using a homologous recombination‐based gene‐targeting technique (Fig. S1a; Ishizaki et al., ^2013b). Four independent knockout lines were generated (Fig. S1b,c). These knockout lines were derived from F1 plants (spores) obtained by crossing Tak‐1 and Tak‐2, which led to some natural variation in thallus morphology; however, all lines exhibited identical phenotypes during vegetative thallus growth. In 3‐wk‐old wild‐type thalli derived from apex‐bearing explants, > 30 gemma cups filled with gemmae were observed along the midrib on the dorsal surface (Fig. 2a,d,g). Conversely, MpR2R3‐MYB5 knockout plants (Mpstg ^ko ) never developed normal gemma cups capable of producing gemmae (Fig. 2b,g). When the apical region of the thallus was cultured, no significant defects were observed in thallus growth or in the development of other vegetative structures, such as air chambers, rhizoids, and ventral scales despite the absence of gemma cups (Fig. S5a,b).

Fig. 2.

Shot glass‐shaped structures on MpR2R3‐MYB5/MpSHOT GLASS gene‐disrupted plants. (a–c) Two‐week‐cultivated apical explants of wild‐type (Tak‐1; a), MpR2R3‐MYB5/SHOT GLASS (MpSTG) gene‐disrupted mutant (Mpstg ^ko ; b), and complementation line (gMpSTG‐Citrine/Mpstg ^ko ; c). Bar, 5 mm. (d–f) Representative gemma cups or shot‐glass‐shaped structures on each plant; Tak‐1 (d), Mpstg ^ko (e), and gMpSTG‐Citrine/Mpstg ^ko (f). Bar, 500 μm. (g) The number of normal and abnormal gemma cups on the 3‐wk‐cultivated apical explants of wild‐type, Mpstg ^ko , and gMpSTG‐Citrine/Mpstg ^ko plants. Data represent counts from six individual plants for each line. P‐values were calculated on the number of normal gemma cups using a Students' t‐test following an F‐test. Asterisk indicates statistically significance (P < 0.05) between normal gemma‐cup numbers in WT and those in gMpSTG‐Citrine/Mpstg^ko . (h, i) Transverse section of the gemma cup in wild‐type plant (h) and the shot glass‐shaped structure formed in Mpstg ^ko plant (i). Bar, 100 μm. (j) Expression and intracellular localization of MpSTG‐tdTomato at the basal region of gemma cup in gMpSTG‐tdTomato/Mpstg ^ko plants. Bar, 100 μm. (k–n) MpHA9 expression profile in gemma cup and the shot glass‐shaped structure. GUS staining in the gemma cup of pMpHA9:GUS plants (k) and the shot glass‐shaped structure of Mpstg ^ge /pMpHA9:GUS plants (l). Transverse section of GUS‐stained gemma cup in pMpHA9:GUS plants (m) and the shot glass‐shaped structure of Mpstg ^ge /pMpHA9:GUS plants (n). Bar: (k–n) 500 μm.

In rare cases, protrusions or shot‐glass‐shaped structures, presumed to be abortive gemma cups, were observed along the midrib of the Mpstg ^ko line (Fig. 2e,g). These gemma cup defects were successfully rescued by introducing the genomic sequence of MpR2R3‐MYB5, including a 5‐kbp promoter region (Figs 2c,f,g, S1a,d). To compare normal gemma cup formation, we performed a Student's t‐test. The number of normal gemma cups was reduced in complemented lines compared to wild‐type (mean ± SD: 22.00 ± 5.63 vs 30.67 ± 4.58; t = 2.89, df = 10, P = 0.016; Fig. 2g). In complemented lines, MpR2R3‐MYB5 fused to the fluorescent protein tdTomato at its C‐terminus driven by the MpR2R3‐MYB5 promoter exhibited distinct nuclear fluorescent signals in gemma‐cup basal floor cells, mucilage cells, and the nuclei of developing gemmae during early stages of gemmae development (Figs 2j, S1d). Based on the abortive gemma cup‐like structures observed in the knockout mutants, we designated MpR2R3‐MYB5 as MpSHOT GLASS (MpSTG).

In wild‐type plants, several developing gemmae were consistently observed in the floor cells of the gemma cups (Fig. 2h). Conversely, the shot‐glass‐shaped structures in Mpstg ^ko plants had narrow bases and lacked gemma‐like structures (Fig. 2i). To investigate whether the floor cells of these structures acquired a similar cell identity to gemma cup floor cells, we analyzed the promoter activities of two marker genes, MpHA9 (H ⁺ ‐ATPase9; Hiwatashi et al., ²⁰¹⁹) and MpRSL1 (RHD6‐LIKE1; Proust et al., ²⁰¹⁶).

MpHA9 is specifically expressed at the base of the gemma cup (Hiwatashi et al., ²⁰¹⁹). In proMpHA9:Citrine‐GUS transgenic plants, GUS signals were exclusively observed at the base of the gemma cups (Figs 2k,m, S2c). By contrast, no GUS signals were detected in the shot‐glass‐shaped structures (Fig. 2l,n) in Mpstg ^ge /proMpHA9:Citrine‐GUS plants carrying an MpSTG loss‐of‐function mutation (Fig. S2b). In proMpRLS1:tdTomato‐NLS transgenic plants, tdTomato signals were observed in the developing gemmae at the base of the gemma cups (Fig. S2d). No proMpRSL1‐positive cells were detected at the base of the shot‐glass‐shaped structures in the Mpstg ^ge /proMpRLS1:tdTomato‐NLS line (Fig. S2d). Collectively, these results indicate that the shot‐glass‐shaped structures in MpSTG mutants fail to acquire the cell identity of gemma cup floor cells and are unable to form gemmae, confirming the essential role of MpSTG in gemma cup development and clonal propagation.

MpSTG restricts dorsal vegetative tissue formation

To investigate the developmental function of MpSTG, we generated transgenic plants overexpressing MpSTG fused to a glucocorticoid receptor (GR) domain at the C‐terminus, driven by the MpELONGATION FACTOR 1α promoter (MpSTG‐GR; Fig. S1d). In these transgenic plants, MpSTG activity as a transcriptional regulator was induced in a DEX‐dependent manner (Schena et al., ¹⁹⁹¹). Under mock treatment (Fig. 3a), the thalli of MpSTG‐GR plants grew flat and resembled wild‐type plants (Fig. 3b). On the ventral side, rhizoids oriented towards the ground (Fig. 3c). On the dorsal surface, regularly distributed air pores corresponding to air chambers and gemma cups formed (Fig. 3b,d). However, DEX treatment drastically suppressed thallus growth in the MpSTG‐GR line (Fig. 3a,f). DEX‐treated thalli failed to develop a flat morphology and exhibited numerous rhizoids growing upwards (Fig. 3g). Mature air pores were absent, although intercellular spaces resembling air pores were initially observed on the dorsal surface via scanning electron microscopy (Fig. 3h). Transverse sections showed that the air chambers with assimilatory filaments did not form in DEX‐treated transverse sections (Fig. 3e,i). However, intercellular spaces in the epidermal cell layers, which are typically observed during the initial stages of air chamber development, were recognized (Fig. 3i′).

Fig. 3.

MpSTG overexpression suppresses air chamber development. (a–i) Two‐week‐old proMpEF1α:MpSTG‐GR transgenic plants treated with mock (b–e) or 10 μM o dexamethasone (DEX; f–i). (a) Schematic representation of treatment. (b, f) Top view of the thallus. (c, g) Side views of the thallus. Bar, 1 mm. (d, h) High‐magnification images within the dotted rectangles in (b, f) were captured using a scanning electron microscope. Bar, 100 μm. (e, i) Transverse sections of the thalli. Bar, 200 μm. Enlarged views of the red dotted rectangles are shown in the lower panels (i′). Arrowheads indicate intercellular gaps or spaces corresponding to air chamber initiation. Bar: (i′) 100 μm. (j–n) Two‐week‐old wild‐type (k, l) and proMpEF1α:MpSTG‐GR transgenic plants (m, n) treated with mock for 1 wk, followed by an additional 1‐wk treatment with 10 μM of DEX. (j) Schematic of treatment. (k, m) Top view of the thallus. (l, n) Side views of the thallus. Bar, 1 mm. (o–r) Top view (o, q) and black‐field images of transverse section (q, r) of gemma‐developing areas in proMpEF1α:MpSTG‐GR transgenic plants treated with mock (o, p) or 10 μM of DEX (q, r) for 3 wk. Bar, 500 μm.

Treatment of 1‐wk gemmalings of MpSTG‐GR plants with DEX (Fig. 3j) caused the thalli to grow vertically instead of flat (Fig. 3k–n).

However, MpSTG overexpression did not induce ectopic gemma cup or gemmae formation across the thallus. In DEX‐treated MpSTG‐GR plants, gemmae were limitedly formed in narrow, spot‐like regions along the midrib of the dorsal surface after 3 wk. This corresponds to the typical sites where gemmae and gemma cups are normally formed in mock‐treated MpSTG‐GR plants (Fig. 3o,q). Similar to wild‐type plants, mock‐treated plants formed gemmae on the peripheral area of the gemma cup basal floor (Fig. 3p; Komatsu et al., ²⁰²³). However, gemmae formed at a high density within these narrow spots in DEX‐treated MpSTG‐GR plants (Fig. 3r).

Transcriptomic analysis reveals downstream targets of MpSTG

To identify the factors downstream of MpSTG that regulate gemma cup development, we analyzed transcriptomic changes in 1‐wk gemmalings of MpSTG‐GR plants using RNA sequencing (Table S4). Compared with those in the mock‐treated group, differential gene expression analysis revealed that MpSTG overexpression induced by 2‐ and 8‐h DEX treatments upregulated 82 and 297 genes, respectively, and downregulated 36 and 164 genes, respectively (Fig. 4a; Table S5). We further listed DEGs that showed more than a twofold change in expression (|log₂FC| > 1) (Tables S6, S7). The gene showing the greatest upregulation after the 2‐h DEX treatment was MpWRKY14 (log₂FC = 1.61), which, however, was not included in the DEG list following the 8‐h treatment.

Fig. 4.

Gene expression changes by MpSTG overexpression. (a) Venn diagram showing differences between differentially expressed genes (DEGs) in mock treatment plants and those in MpSTG‐GR plants after 2 and 8 h of treatment with 10 μM dexamethasone (DEX). DEGs were identified based on RNA‐seq data, with expression levels quantified as transcripts per million (TPM). (b) Heatmap of expression profiles of transcription factor genes commonly regulated at both 2 and 8 h after DEX treatment in MpSTG‐GR plants. RNA‐seq expression values (TPM) were transformed into Z‐scores for each gene. Columns represent differentially expressed transcription factor genes, and rows represent treatment conditions (mock or 10 μM DEX) at each time point. Red and blue indicate higher and lower expression, respectively, relative to the mean value for each gene. (c, d) Relative expression levels of MpDELLA and Mp1R‐MYB17 (c) MpCYCB;1 (d) in 1‐wk‐old gemmalings of MpSTG‐GR plants treated with mock or 10 μM DEX for 4 and 24 h, and 10 μM cycloheximide (CHX) or 10 μM DEX + 10 μM CHX for 4 h. MpACT7 and MpEF1a were used as reference genes in RT‐qPCR. Bars represent the mean ± SE. The dots represent individual data points. Asterisks indicate statistically significant differences (*, P < 0.05; **,P < 0.01) and ns indicates not significant (P > 0.05) (Student's t‐test following an F‐test); n = 4.

Among the DEGs at both 2 and 8 h after DEX treatment, 22 genes were consistently upregulated and 14 were consistently downregulated (Figs 4a, S6a; Table S4). From this subset, we focused on transcription factors; MpWRKY11 (Mp1g24950) and MpDELLA (Mp5g20660) were upregulated, while MpGARP2 (Mp4g08700), MpTRIHELIX12 (Mp2g25660), and Mp1R‐MYB17 (Mp4g21300) were downregulated (Figs 4b, S6a). We focused on these transcription factors because they showed reproducible expression; therefore, they were strong candidates for direct or indirect downstream targets of MpSTG. RT‐qPCR confirmed that MpSTG overexpression significantly upregulated MpDELLA expression and downregulated Mp1R‐MYB17 expression (Fig. 4c). Notably, DEX‐induced expression changes in MpDELLA and Mp1R‐MYB17 were observed under cycloheximide treatment, which inhibits de novo protein synthesis, thus implicating these genes as likely direct transcriptional targets of MpSTG (Fig. 4c). Additionally, MpWRKY11, MpGARP2, and MpTRIHELIX12 showed consistent changes in RNA‐seq analysis; however, these genes were not analyzed further in this study because the results were not reproducible by RT‐qPCR (Fig. S6b).

Our gene ontology (GO) term enrichment analysis of genes differentially expressed after 8 h of DEX treatment in MpSTG‐GR plants revealed enrichment in categories related to cell cycle and cell division (downregulated genes) (Fig. S7). We observed significant downregulation of MpCYCB;1 expression in MpSTG‐GR plants following 8‐h (Table S4) and 24‐h DEX treatment (Fig. 4d).

MpSTG affects the MpGCAM1 pathway involved in gemma cup development

Although not statistically significant, MpGCAM1 expression was slightly lower in the Mpstg ^ko mutant that in the wild‐type plant (Fig. 5a). To investigate whether MpSTG regulates MpGCAM1 expression, we used a transgenic line, in which Citrine was knocked in before the stop codon of the MpGCAM1 gene (MpGCAM1‐Citrine ^KI ; Yasui et al., ²⁰¹⁹); MpGCAM1‐Citrine signals were detected in developing gemmae and gemma cup floor cells (Fig. 5e). In this line, although the number of gemma cups were similar to that in wild‐type plants (Fig. 5b), it had been reported that the number of gemmae per gemma cup was three times higher than that in the wild‐type plant (Fig. 5c), possibly because of the enhancements of MpGCAM1 expression in gemma cup (Komatsu et al., ²⁰²³). Nevertheless, following CRISPR/Cas9‐mediated MpSTG disruption in the MpGCAM1‐Citrine ^KI background (Mpstg ^ge /MpGCAM1‐Citrine ^KI ; Fig. S2b), only shot‐glass‐like structures lacking gemmae were formed instead of the normal gemma cups (Fig. 5b,d). MpGCAM1‐Citrine fluorescence was never detected in the shot‐glass‐like structures (Fig. 5f).

Fig. 5.

Functions of MpSTG and MpGCAM1 in gemma‐cup formation. (a) Relative expression levels of MpSTG in the apical part of regenerated thalli 2 wk after cutting thalli of wild‐type, Mpstg ^ko , and gMpSTG‐Citrine/Mpstg ^ko plants. MpACT7 and MpEF1a were used as reference genes in RT‐qPCR. Dots represent each data point. The box plot illustrates the interquartile range (IQR), spanning from the first quartile (Q1) to the third quartile (Q3). The horizontal line within the box marks the median (second quartile, Q2). The whiskers extend to the minimum and maximum values within 1.5 times the IQR from Q1 and Q3, respectively. n = 12. (b) The number of normal and abnormal gemma cups on 3‐wk‐cultivated apical explants of wild‐type (Tak‐1), MpGCAM1‐Citrine ^KI (GCAM1‐Cit), and Mpstg ^ge /MpGCAM1‐Cit (stg ^ge /GCAM1‐Cit) plants. Data represent counts from five individual plants for each line. No statistically significant difference (ns, P > 0.05) was detected in gemma‐cup numbers between Tak‐1 and GCAM1‐Cit, based on P‐value calculated using a Students' t‐test following an F‐test. (c, d) Representative gemma cups in MpGCAM1‐Citrine ^KI (c) and Mpstg ^ge /GEMMA CUP‐ASSOCIATED MYB 1 (GCAM1)‐Cit plants (d) captured with a scanning electron microscope. Bar, 1 mm. (e, f) MpGCAM1‐Citrine signal in the basal region of a gemma cup in MpGCAM1‐Citrine ^KI (e) and a shot glass‐like structure in Mpstg ^ge /MpGCAM1‐Citrine ^KI (f) plants. Magnified image of the gemma cup basal floor cells are shown in the insets. Bar, 500 μm. (g, h) Two‐week‐cultivated apical explants of Mpgcam1 ^ko (g) and Mpstg ^ge /Mpgcam1 ^ko double mutant (h). Bar, 5 mm. (i–l) Effects of ectopic MpSTG overexpression on the growth of regenerating thallus. Cut thalli of MpSTG‐GR (i, j) and MpSTG‐GR/Mpgcam1 ^ko (k, l) were cultured for 9 d in the absence (i, k) or presence (j, l) of 10 μM of dexamethasone (DEX). Bar, 1 mm. (m–p) Effects of ectopic MpGCAM1 overexpression. Apical explants of MpGCAM1‐GR (m, n) and Mpstg ^ge /MpGCAM1‐GR (o, p) were treated with mock (m, o) or 10 μM of DEX (n, p) for 14 d. Scanning electron microscopy images of cell clamps generated following DEX treatment are presented in the insets (n, p). Bar, 2 mm; white bars in insets, 500 μm.

To further explore the relationship between MpSTG and MpGCAM1, we generated double mutants of MpSTG and MpGCAM1 (Mpstg ^ge Mpgcam1 ^ko ; Fig. S2b). The Mpgcam1 ^ko mutant exhibited complete loss of the gemma cup, while the Mpstg ^ge /Mpgcam1 ^ko double mutant exhibited no additional phenotype to gemma‐cup impairment (Fig. 5g,h).

Next, we examined transgenic plants overexpressing MpSTG‐GR on an MpGCAM1 loss‐of‐function background (MpSTG‐GR/Mpgcam1 ^ko ). DEX treatment of MpSTG‐GR plants suppressed the dorsal tissue growth, resulting in the orthotropic growth of thallus and the exposure of ventral tissues to air (Fig. 3f,g,m,n). Considering that Mpgcam1 ^ko plants lack gemma formation, we examined the effects of DEX treatment on explants regenerated from thalli excision. After 9 d of mock treatment, flat thalli with clear dorsoventral polarity were regenerated from the MpSTG‐GR plant fragments (Fig. 5i). Conversely, DEX‐treated explants exhibited non‐flat growth, with numerous rhizoids extending upwards (Fig. 5j). Similar phenotypes were observed in the MpSTG‐GR/Mpgcam1 ^ko line (Fig. 5k,l). Moreover, transcriptomic analysis demonstrated that 2‐ and 8‐h DEX treatment did not alter MpGCAM1 expression in MpSTG‐GR plants (Table S4), suggesting that MpGCAM1 is not a direct target of MpSTG.

Furthermore, we generated MpSTG loss‐of‐function mutants in the MpGCAM1‐GR background (Mpstg ^ge /MpGCAM1‐GR) via genome editing (Fig. S2b). Mock‐treated MpGCAM1‐GR plants showed a similar phenotype with wild‐type plants (Fig. 5m). DEX treatment resulted in undifferentiated cell clump formation in the MpGCAM1‐GR (Fig. 5n) and Mpstg ^ge /MpGCAM1‐GR (Fig. 5o,p) lines. Thus, MpSTG does not act downstream of MpGCAM1 during undifferentiated cell clump formation.

MpSTG is essential for sexual reproduction

RT‐qPCR and GUS reporter assay showed that MpSTG was expressed not only in vegetative organs, but also in sexual reproductive organs, gametangiophores (Figs 1a,g–j, S4h–k), thus indicating MpSTG as part of the sexual reproductive phase. To further investigate this role, we analyzed the phenotypes of MpSTG knockout mutants (Mpstg ^ko ) during sexual reproduction. Wild‐type male and female plants developed gametangiophores within 3 wk under FR light supplemented with white light; antheridiophores in males and archegoniophores in females (Fig. 6a,b). By contrast, Mpstg ^ko plants failed to form any reproductive organs, even after > 60 d of FR irradiation (Fig. 6a,b). These defects in gametangiophore formation were largely rescued in the complementation line (gMpSTG‐Citrine/Mpstg ^ko ), although gametangiophore formation occurred more slowly than in wild‐type plants (Fig. 6a,b). Overall, these results indicate that MpSTG plays an indispensable role in gametangiophore formation.

Fig. 6.

MpSTG functions in gametangiophore formation. (a) Representative apical regions of male and female thalli in wild‐type, Mpstg ^ko , and gMpSTG‐Citrine/Mpstg ^ko plants after 3 wk of supplemental irradiation with far‐red light (FR). Bar, 5 mm. (b) Days until the first gametangiophore recognition after the start of FR irradiation. Numbers above the graph indicate the number of plants that formed the gametangiophore among eight biological replicates. nd indicates that gametangiophore formation was not detected. Each dot represents an individual data point. The box plot illustrates the interquartile range (IQR), which spans from the first quartile (Q1) to the third quartile (Q3). The horizontal line within the box marks the median (second quartile, Q2). The whiskers extend to the minimum and maximum values within 1.5 times the IQR from Q1 and Q3, respectively. (c) Representative apical regions of male and female explants of MpBNB‐GR and Mpstg ^ge /MpBNB‐GR plants treated with mock solution or 10 μM of dexamethasone (DEX) for 3 wks. Bar, 2 mm. (d) Sexual reproductive development in male and female MpBNB‐Citrine and Mpstg ^ge /MpBNB‐Citrine plants. Representative apical meristematic regions after 3 wk of culture under the inductive conditions are shown (upper panels). Arrowheads indicate sexual reproductive organs in the apical meristematic regions. Bar, 5 mm. Thallus epidermal cells and the MpBNB‐Citrine fluorescence at the apical notch of the male and female plants are shown (lower panels). Images in white‐dashed rectangles are magnified on the right. Asterisks indicate the apical notches. Bar: (black) 50 μm; (white) 10 μm.

MpSTG facilitates MpBNB accumulation and reproductive development

To determine whether MpSTG functions downstream of MpBNB in gametangiophore formation, we generated MpSTG loss‐of‐function mutants in the MpBNB‐GR background (Mpstg ^ge /MpBNB‐GR) using CRISPR/Cas9 genome editing system (Fig. S2b). Similar to MpBNB‐GR plants, Mpstg ^ge /MpBNB‐GR plants formed gametangiophores following DEX treatment (Fig. 6c), suggesting that MpSTG is not involved in MpBNB‐induced gametangiophore formation.

To determine whether MpSTG influences the spatiotemporal expression of MpBNB, we generated MpSTG loss‐of‐function mutants in MpBNB‐Citrine plants (Mpstg ^ge /MpBNB‐Citrine) using CRISPR/Cas9 genome‐editing system (Fig. S2b). After 12 d of culture under inductive conditions, MpBNB‐Citrine accumulation was clearly observed in the nuclei of gametangia initial cells in the developing gametangiophores of both male and female MpBNB‐Citrine plants. By contrast, no initial cells showing MpBNB‐Citrine fluorescence were observed in the apical notches of Mpstg ^ge /MpBNB‐Citrine plants (Fig. 6d).

Phylogenetic relationship of MpSTG

Phylogenetic analysis confirmed that MpSTG belongs to subfamily V, which is distinctly separated from the closely related subfamily IV (Fig. 7a). Members of R2R3‐MYB subfamily V, including MpSTG, share highly conserved exon‐intron structures and a C‐terminal motif, in addition to the characteristic R2R3‐MYB domain (Figs 7b, S8a,b). However, the full‐length amino acid sequence of MpSTG was approximately twice as long as that of the other members of subfamily V (Fig. S8a,b).

Fig. 7.

Phylogenetic relationship between MpSTG and its homologs in land plants. (a) Phylogenetic analysis of MpSTG and its homologs across the land plants. An unrooted maximum‐likelihood tree was generated using amino acid sequences for the R2R3‐MYB DNA‐binding domains from MpSTG and related R2R3‐MYB proteins representing diverse land plant lineages; Arabidopsis thaliana, Solanum lycopersicum, Oryza sativa, Selaginella moellendorffii, Physcomitrium patens, and Marchantia polymorpha. Bootstrap values (%) from 1000 replicates are shown at the nodes. The scale bar represents evolutionary distance as the rate of amino acid substitutions. (b) Multiple alignment of the R2R3‐MYB domains of MpSTG, AtLOF1, AtLOF2, AtBRAVO, and SlTrifoliate. Black asterisks indicate conserved tryptophan (Trp) residues typical of plant R2R3‐MYB proteins. Amino acids are color‐coded according to the ClustalW convention based on their physicochemical properties. (c–f) Morphology of paraclade junctions in A. thaliana Col‐0 wild‐type (c), lof1‐1 (d), AtLOF1 ^OX /lof1‐1 (e), and MpSTG ^OX /lof1‐1 (f) plants. White arrowheads indicate accessory buds. ax, axillary stem; c, cauline leaf; ps, primary stem. Bar, 2 mm.

Previous studies demonstrated that M. polymorpha GCAM1 can complement the axillary bud formation defects of the A. thaliana rax1 rax2 rax3 triple mutant, suggesting a functional analogy between gemma cup formation in M. polymorpha and axillary bud development in angiosperms (Yasui et al., ²⁰¹⁹). Given that the Mpstg mutant displays defects in gemma cup formation, we investigated whether MpSTG might functionally resemble members of the R2R3‐MYB subfamily V in angiosperms, such as AtLOF1, AtLOF2, and SlTrifoliate, which have been implicated in the regulation of accessory (or axillary) bud formation (Lee et al., ²⁰⁰⁹; Naz et al., ²⁰¹³). Furthermore, in addition to the characteristic R2R3‐MYB domain, MpSTG shared more conserved motifs with AtLOF1 (motifs 15, 16, and 18; Fig. S8a) than with AtLOF2 (motifs 16 and 18; Fig. S8a). Therefore, to examine whether M. polymorpha STG shares functional similarities with angiosperm homologs, we introduced MpSTG into the Arabidopsis lof1‐1 mutant, which exhibits defects in accessory bud formation at the junctions between axillary stems and cauline leaves (Fig. 7c,d; Lee et al., ²⁰⁰⁹). Similar to A. thaliana LOF1 overexpression (pro35S:AtLOF1), MpSTG overexpression under the control of the CaMV35S promoter (pro35S:MpSTG) partially rescued accessory bud formation defects in the mutant (Fig. 7e,f). These results demonstrate that MpSTG from M. polymorpha functions within the genetic machinery responsible for accessory bud formation in the angiosperm A. thaliana.

Discussion

MpSTG is a novel regulator of gemma cup development

In the present study, we focused on MpSTG – an R2R3‐MYB transcription factor that is highly expressed in the gemma cup. Loss‐of‐function mutants of MpSTG failed to form functional gemma cups capable of producing gemmae (Fig. 2), indicating that MpSTG is essential for both gemma cup formation and gemma production. Instead of normal gemma cups, Mpstg mutants occasionally develop shot‐glass‐like structures with a narrower floor at the base that lack the ability to generate gemmae initials (Fig. 2). Thus, MpSTG seems to play a critical role in the development and/or maintenance of the basal floor during gemma cup formation. Notably, ectopic MpSTG overexpression did not increase the number of gemma cups or induce the generation of ectopic gemmae, suggesting that the MpSTG is not involved in the initiation or spatial determination of gemma cup formation.

Alternatively, ectopic MpSTG overexpression suppressed air chamber development on the dorsal surface of the thallus (Fig. 3). Quantitative reverse transcription polymerase chain reaction confirmed that MpSTG expression was higher in the gemma cups than that in the thallus or apical region, and MpR2R3‐MYB5 was highly expressed in the female reproductive organs (Fig. 1a). Overall, these results were consistent with the expression profiles available in the open‐source database MBEX (MarpolBase Expression; https://mbex.marchantia.info; Kawamura et al., ²⁰²²; Fig. S3). Transcriptome analysis revealed that MpSTG upregulated MpDELLA expression and downregulated MpCYCB;1 expression (Fig. 4b–d; Hernández‐García et al., ²⁰²¹). These results suggest that MpSTG regulates cell division via the MpDELLA pathway and inhibits air chamber formation, which may be a prerequisite for basal floor expansion during gemma cup formation. However, as our data did not directly demonstrate a regulatory relationship between MpSTG and MpCYCB;1 mediated by MpDELLA, additional experiments will be required to validate this regulatory pathway. Although MpSTG was also expressed on the ventral side of the thallus midrib (Fig. S4g), neither the loss of function nor overexpression of MpSTG affected ventral organogenesis in the thallus. Thus, the involvement of other regulators that spatially restrict MpSTG function to the dorsal side may be involved.

Furthermore, we investigated the genetic relationship between MpSTG and MpGCAM1, a key regulator of gemma cup development (Yasui et al., ²⁰¹⁹; Kato et al., ²⁰²⁰). MpSTG was not required by MpGCAM1 to confer stemness to epidermal cells (Fig. 5m–p). Conversely, MpGCAM1 did not play a role in MpSTG‐mediated suppression of air chamber development (Fig. 5i–l), suggesting that MpSTG and MpGCAM1 functions in parallel or in partially independent pathways. This aligns with previous findings that MpSTG is not a downstream target of MpRRB in the cytokinin signaling pathway, whereas MpGCAM1 is directly upregulated by MpRRB (Aki et al., ²⁰²²). However, low expression of MpGCAM1 in the Mpstg ^ko mutant and the loss of function of MpSTG in MpGCAM1‐Citrine ^KI plants (Fig. 4a–f) suggest that MpSTG is necessary for maintaining MpGCAM1 expression to facilitate the expansion of the gemma cup's basal floor and promote gemma formation. MpSTG indirectly contributes to the maintenance of MpGCAM1 expression by enabling the expansion of the MpGCAM1‐positive domain, even if MpGCAM1 initially functions upstream of MpSTG. Thus, the regulatory direction may sift depending on developmental timing and spatial context.

Based on these results, we propose that the gemma cup development comprises several sequential processes, including basal floor establishment, basal floor maintenance and expansion, and rim formation (Fig. 8). The gemma cup is not derived from a single initial cell but instead formed by a group of cells originating from lateral and dorsal merophytes, which acquire gemma‐forming ability in response to developmental and environmental cues (Suzuki et al., ²⁰²⁰). The essential regulator, MpGCAM1, has been proposed to maintain dorsal cells in an undifferentiated state while initiating and establishing the gemma cup basal floor (Yasui et al., ²⁰¹⁹). The KL signaling promote gemma cup formation by upregulating cytokinin biosynthesis and signaling pathway, ultimately leading to induction of MpGCAM1 expression (Fig. 8; Aki et al., ²⁰²²; Komatsu et al., ²⁰²⁵). MpSTG is necessary for basal floor expansion, possibly by maintaining and promoting MpGCAM1 expression in the basal floor region (Fig. 8). Simultaneously, MpSTG negatively regulates air chamber development, potentially through the MpDELLA pathway. This creates space in the dorsal epidermis for basal floor expansion (Fig. 8). Gemma cup development is completed by the three‐dimensional outgrowth of the gemma cup rim, which is regulated by an RWPRK domain‐containing transcription factor, MpRKD (Rövekamp et al., ²⁰¹⁶). This factor is a target of miRNAs derived from MpMIR319 (Lin et al., ²⁰¹⁶; Tsuzuki et al., ²⁰¹⁶). MpMIR319 has been suggested to influence the frequency of gemma cup formation, although its specific targets and underlying mechanisms remain unknown (Futagami et al., ²⁰²⁵).

Fig. 8.

Scheme of developmental processes of gemma cup in Marchantia polymorpha. Solid arrows, known positive regulatory pathways; blant‐ended arrow, known negative regulatory pathway; dashed arrows, positive regulatory pathways suggested in this study; thick arrows, positive regulatory pathways with unknown processes; thick blant‐ended arrows, negative regulatory pathways with unknown processes.

MpSTG is involved in the regulation of sexual reproductive development

Our study showed that MpSTG plays a critical role in the formation of sexual reproductive organs. Loss‐of‐function mutations in MpSTG inhibited the formation of reproductive organs under inductive conditions (Fig. 6). Sexual reproductive organ formation is triggered by FR light, which is mediated by the R/FR photoreceptor Mpphy and its downstream MpPIF (Inoue et al., ²⁰¹⁹). This likely promotes the activation of MpBNB (Yamaoka et al., ²⁰¹⁸). The antheridium and archegonium primordia develop within the structural contexts of the male and female sexual reproductive organs, respectively.

Notably, while sexual reproductive organs failed to form in MpSTG loss‐of‐function mutants (Fig. 6a,b), the induction of MpBNB activity via an MpBNB‐GR inducible line in the Mpstg mutant background successfully restored gametangiophore formation upon DEX treatment (Fig. 6c). This result suggests that although MpSTG is required for initiating gametangiophore development under natural inductive conditions, its function can be bypassed by ectopic MpBNB activation. Therefore, MpSTG is not strictly required downstream of MpBNB. Indeed, MpBNB protein accumulation was clearly reduced in the apical meristem notch of MpSTG loss‐of‐function mutants under inductive conditions (Fig. 6d), indicating that MpSTG likely promotes sexual reproductive organ formation by modulating MpBNB expression or accumulation. Supporting this idea, the MpBNB expression pattern was not altered by MpSTG during the vegetative phase (Table S4), thus implying that MpSTG likely affected MpBNB expression specifically during the sexual reproductive phase. The molecular mechanisms by which MpSTG modulates MpBNB accumulation remain unclear; however, its role in gemma cup formation suggests that MpSTG contributes to the generation of spatiotemporal gaps and spaces around the apical notch, enabling MpBNB to accumulate in appropriate cells.

Notably, MpSTG expression can also be observed in the later developmental stage of gametangiophore, implying potential functions beyond the initial phase of reproductive induction. Future studies that temporally manipulate MpSTG activity – such as the application of inducible gMpSTG‐GR systems after reproductive fate has been established – could provide valuable insight into its stage‐specific roles, including potential contributions to gametangia development, gamete maturation, or male fertility. Elucidating these later functions of MpSTG will be key to understanding the full scope of its role in coordinating reproductive development in Marchantia.

Dual function modulators are involved in both vegetative and sexual reproduction

This study demonstrated that the R2R3‐MYB transcription factor MpSTG regulates both vegetative and sexual reproduction. Similarly, MpKANADI (MpKAN) – another transcription factor – functions in both vegetative and sexual reproductive phases (Briginshaw et al., ²⁰²²). Loss of function mutations in MpKAN result in defects in both vegetative and sexual reproduction, suggesting that MpKAN positively influences gemma cup and sexual reproductive organ formation. Transcriptome analysis during the vegetative and reproductive phases revealed that MpKAN modulated the expression of multiple genes involved in various developmental processes, including meristem‐, gemma cup‐, and gametangia‐associated genes (Briginshaw et al., ²⁰²²). MpKAN appears to coordinate the harmonized development of the thallus by acting as a negative regulator of meristem activity in the apical notch while functioning as a positive regulator of dorsal tissue differentiation.

MpSTG is one of the genes modulated by MpKAN during both the vegetative and reproductive phases (Briginshaw et al., ²⁰²²). During the vegetative phase, MpSTG functions as a positive regulator of gemma cup formation, as well as MpKAN. On the other hand, ectopic and excessive activation of MpSTG suppressed air chamber formation, resulting in severe dorsal developmental defects (Fig. 3). Therefore, MpSTG activity appeares to be regulated spatiotemporally to ensure propper gemma cup formation. Genetic evidence suggests that MpKAN downregulates MpSTG expression (Briginshaw et al., ²⁰²²), which may restrict MpSTG function spatiotemporally and maintain dorsal tissue development.

In the reproductive phase, MpSTG promoted FR‐induced accumulation of MpBNB in gametangia initial cells (Fig. 6d) and facilitated the formation of both male and female gametangiophores. A previous study showed that FR irradiation significantly increased MpSTG expression in females but not in males (Briginshaw et al., ²⁰²²). Similarly, our results showed that MpSTG expression levels in the gametangia differed between males and females (Fig. 1a). Additionally, FR irradiation upregulated MpSTG expression in male Mpkan mutants. These findings suggest that MpSTG downregulation by MpKAN is higher in males than that in females, and that MpKAN appears to fine‐tune MpSTG expression in a sex‐dependent manner.

Furthermore, MpDELLA has been proposed to negatively regulate MpBNB induction under FR (Hernández‐García et al., ²⁰²¹). MpKAN appeared to simultaneously downregulate both MpSTG and MpDELLA, which suggested a complex interplay between positive and negative feedback in the regulation of gametangia initiation. Further studies are required to elucidate the precise relationships between these regulators.

MpSTG is orthologous to regulators of lateral bud formation in angiosperms

Consistent with this study and previous phylogenetic analyses, classified MpSTGs into subfamily V (subgroup 21) of the R2R3‐MYB transcription factors (Fig. 7a; Lee et al., ²⁰⁰⁹; Du et al., ²⁰¹⁵; Jiang & Rao, ²⁰²⁰). In A. thaliana, this subgroup includes multiple homologs such as AtLOF1, AtLOF2, AtBRAVO, AtMYB10, and AtMYB89 – phylogenetically close to MpSTG (Jiang & Rao, 2020). These genes have diversified to acquire distinct functions: for example, AtMYB89 is involved in seed oil accumulation (Li et al., ²⁰¹⁷), AtMYB10 contributes to vascular development, and AtBRAVO plays critical roles in root apical meristem maintenance and stem cell identity (Vilarrasa‐Blasi et al., ²⁰¹⁴). In this study, we focused our functional analyses on AtLOF1, given its reported involvement in axillary meristem development and boundary specification (Lee et al., ²⁰⁰⁹). Loss‐of‐function mutants of AtLOF1 (lof1‐1) exhibit defects in accessory bud formation at the junction between the axillary branches and cauline leaves (Lee et al., ²⁰⁰⁹). Similar to AtLOF1 expression, ectopic MpSTG expression partially rescued accessory bud defects in the lof1‐1 mutant in this study. This suggests that the biochemical functions of MpSTG and AtLOF1 are partially conserved and MpSTG can activate target genes in A. thaliana.

In addition to MpSTG, MpGCAM1 (an ortholog of the regulators of axillary bud formation in angiosperms) is essential for gemma cup formation in the liverwort, M. polymorpha (Yasui et al., ²⁰¹⁹). In the present study, we demonstrated that MpSTG, is necessary for gemma cup formation. Gemma cups in Marchantia are lateral organs with a secondary meristem (stem cell pools) distant from the apical meristem and analogous to the axillary or accessory buds in angiosperms. Our findings support the idea that the fundamental mechanisms for generating secondary meristems are conserved across land plants. Specifically, establishing special boundaries to maintain meristematic regions via processes such as boundary formation and organ separation is a crucial step in the development of secondary shoots.

On the other hand, members of subfamily V of the R2R3‐MYB transcription factors include BRASSINOSTEROIDS AT VASCULAR AND ORGANIZING CENTER in A. thaliana (AtBRAVO; Fig. 7a). AtBRAVO plays a central role in maintaining the quiescent center and regulating stem cell identity in the root meristem, thereby contributing to the establishment of the global body plan (Vilarrasa‐Blasi et al., ²⁰¹⁴). Although our study did not examine the function of AtBRAVO in detail, the evolutionary proximity of MpSTG to multiple Arabidopsis class V MYBs – including AtBRAVO – raises the possibility that diverse meristem‐regulatory functions have ancient origins and were subsequently partitioned among paralogs in the angiosperm lineage. Functional explortion of additional Arabidopsis orthologs, including AtBRAVO, will be important for further elucidating the degree of functional conservation and divergence within this subfamily.

Competing interests

None declared.

Author contributions

YS and KI planned and designed experiments. YS, HT, SY and HF performed the experiments. YS, HT, SY, HF, HK and KI analyzed the data. YS and KI wrote the manuscript with input from all authors. HF and TK supervised this study. All the authors have read and approved the final version of this manuscript.

Disclaimer

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.

Supporting information

Fig. S1 Generation of MpR2R3‐MYB5/STG (Mp8g11870) knockout mutants and their complementation lines.

Fig. S2 Genome editing at the MpR2R3‐MYB5/STG (Mp8g11870) locus.

Fig. S3 Expression profile of MpR2R3‐MYB5/STG (Mp8g11870), derived from the open‐source MBEX (MarpolBase Expression) database, related to Fig. 1(a).

Fig. S4 Histochemical GUS assay of proMpSTG:GUS lines related to Fig. 1.

Fig. S5 Dorsal and ventral organs in the MpSTG knockout mutant.

Fig. S6 Expression profile in MpSTG‐GR plant.

Fig. S7 Gene Ontology (GO) term enrichment analysis of DEGs in MpSTG‐GR plants.

Fig. S8 Schematic representation of domain structures in the MpSTG and its homologs.

Fig. S9 Arabidopsis thaliana transgenic plants expressing AtLOF1 and MpSTG in the lof1‐1 mutant background.

Table S1 List of plants analyzed in this study.

Table S2 List of primers and oligos used in this study.

Table S3 Vectors used in this study.

Table S4 RNA‐seq data (TPM values).

Table S5 Differentially expressed genes (DEGs) induced by MpSTG overexpression.

Table S6 DEGs induced by MpSTG overexpression for 2 h (|log₂FC| > 1).

Table S7 DEGs induced by MpSTG overexpression for 8 h (|log₂FC| > 1).

Table S8 List of accession numbers of the R2R3‐MYB proteins analyzed in this study.

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.

Data availability

The data that support the findings of this study are available in the article and in Figs S1–S9 and Tables S1–S8. The RNA‐seq data here generated are available on DDBJ Sequence Read Archive (SRA) (DRR639648–DRR639659) (https://www.ddbj.nig.ac.jp).