KANADI promotes thallus differentiation and FR-induced gametangiophore formation in the liverwort Marchantia.

In angiosperms, KANADI transcription factors have roles in the sporophyte generation regulating tissue polarity, organogenesis and shade avoidance responses, but are not required during the gametophyte generation. Whether these roles are conserved in the gametophyte-dominant bryophyte lineages is unknown, which we examined by characterising the sole KANADI ortholog, MpKAN, in the liverwort Marchantia polymorpha. In contrast to angiosperm orthologs, MpKAN functions in the gametophyte generation in Marchantia, where it regulates apical branching and tissue differentiation, but does not influence tissue polarity in either generation. MpKAN can partially rescue the sporophyte polarity defects of kanadi mutants in Arabidopsis, indicating that MpKAN has conserved biochemical activity to its angiosperm counterparts. Mpkan loss-of-function plants display defects in far-red (FR) light responses. Mpkan plants have reduced FR-induced growth tropisms, have a delayed transition to sexual reproduction and fail to correctly form gametangiophores. Our results indicate that MpKAN is a modulator of FR responses, which may reflect a conserved role for KANADI across land plants. Under FR, MpKAN negatively regulates MpDELLA expression, suggesting that MpKAN and MpDELLA act in a pathway regulating FR responses, placing MpKAN in a gene regulatory network exhibiting similarities with those of angiosperms.

Introduction

The evolution of land plants from an ancestral alga more than 450 million years ago was facilitated by changes in life history, body plan and responses to new environmental cues on land. Recent analyses of streptophyte genomes have indicated key genetic innovations, which helped facilitate the transition to land, with a small number of transcription factor (TF) families and phytohormone pathways either originating or displaying increased diversity in the ancestral land plant (Hori et al., 2014; Ju et al., 2015; Catarino et al., 2016; Bowman et al., 2017; Cheng et al., 2019; Li et al., 2020). The elucidation of ancestral vs derived functions provides insight into the nature of the ancestral land plant and how genes direct development of plants with radically different body plans.

Among the land plant‐specific TFs are the KANADI (KAN) TFs, a subfamily of GARP (GOLDEN2, Arabidopsis response regulator (ARR) and Phosphorus Stress Response1 (PSR1)) transcriptional repressors, which possess conserved GARP DNA‐binding and EAR motif repressor‐like domains (PDLSL and LEFTL), with the latter domains likely to facilitate interactions with TOPLESS co‐repressors (Hosoda et al., 2002; Causier et al., 2012; Bowman et al., 2017; Safi et al., 2017). In Arabidopsis, KAN is not required for gametophyte development (Izhaki & Bowman, 2007), but functions in the sporophyte generation, where it regulates tissue polarity by specifying abaxial–peripheral tissue types in leaves and vasculature, as well as organogenesis by co‐ordinating lateral organ initiation sites (Eshed et al., 2001, 2004; Kerstetter et al., 2001; Emery et al., 2003; Heisler et al., 2005; Caggiano et al., 2017; Ram et al., 2020). Additionally, KAN TFs have been shown to regulate hypocotyl elongation and vasculature responses to shade (Brandt et al., 2012; Xie et al., 2015; Merelo et al., 2017; Botterweg‐Paredes et al., 2020). A model has been proposed whereby KAN and Class III HOMEODOMAIN‐LEUCINE ZIPPER (C3HDZ) TFs regulate both development and shade processes by the antagonistic regulation of shared target genes, including genes encoding auxin biosynthesis enzymes (i.e. TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS (TAA) and YUCCA (YUC)) and Class II HD‐ZIP (C2HDZ) TFs (Brandt et al., 2012; Merelo et al., 2013, 2017; Reinhart et al., 2013; Huang et al., 2014; Xie et al., 2015). Thus, distinct processes (developmental patterning and shade avoidance) are regulated by a shared KAN‐C3HDZ regulatory module (Merelo et al., 2017). During shade avoidance responses, auxin biosynthesis genes are also downstream targets of the PHYTOCHROME‐INTERACTING FACTOR (PIF) TFs, and shade‐induced elongation growth is associated with high auxin levels, thus providing a potential link between KAN‐C3HDZ regulatory activity and phytochrome‐mediated regulation of shade avoidance responses (Tao et al., 2008; Hornitschek et al., 2012; Li et al., 2012; Merelo et al., 2017).

The role of KAN TFs in nonangiosperm lineages, however, is not well understood. Expression analysis of vascular plant KAN orthologs suggests that a role in regulating tissue polarity may be conserved in other euphyllophytes (ferns and seed plants) (Zumajo‐Cardona et al., 2019). The function of KAN in the bryophyte lineages (i.e. liverworts, hornworts and mosses) remains unknown. In contrast to the four KANADI homologs present in Arabidopsis, Marchantia has a single ortholog – MpKAN – which by parsimony reflects the ancestral copy number (Bowman et al., 2017). To better understand the ancestral role of KAN in land plants, we characterised the sole Marchantia ortholog – MpKAN – in both generations of the life cycle.

Materials and Methods

MpKAN promoter cloning

Marchantia gene nomenclature follows the guidelines designed by Bowman et al. (2016). Gene and coding nucleotide sequences of MpKAN (v.6.1: Mp3g04970.1) were collected from the MarpolBase genome portal (Supporting Information Fig. S1). All primers are listed in Table S1. Primer sets #1/#2 and #3/#4 were used to amplify 6.1kb of MpKAN 5′ regulatory region in two fragments from Marchantia genomic DNA (Fig. S1). Gibson Assembly was performed using the two MpKAN promoter fragments as inserts and BsaI/HindIII‐cut pMpGWB401 vector (Ishizaki et al., 2015) to clone MpKAN:GW:pMpGWB401. GUS:pENTR/D‐TOPO was recombined with MpKAN:GW:pMpGWB401 to clone MpKAN:GUS:pMpGWB401.

gRNA design and cloning

Two gRNAs targeting MpKAN intronic sequence (Fig. S1; Table S1) were first cloned into pMpGE_EN02 plasmid vectors, and next recombined with pMpGE010 and pMpGWB301 destination vectors to generate MpKAN‐gRNA1‐pMpGE010 and MpKAN‐gRNA2‐pMpGWB301 transformation vectors (Sugano et al., 2014, 2018; Ishizaki et al., 2015).

MpKAN coding sequence cloning

Wild‐type cDNA from thallus tissue was used to amplify a 2.1‐kb MpKAN coding sequence (CDS) with and without a stop codon (primers #9–#11), and cloned into pENTR/D‐TOPO (Life Technologies). To generate a T‐DNA for the complementation of Mpkan loss‐of‐function lines, MpKAN‐CDS‐Stp‐pENTR/D‐TOPO and MpKAN:GW:pMpGWB401 were recombined to generate a MpKAN:MpKAN‐CDS:pMpGWB401 plasmid. The MpKAN:MpKAN‐CDS T‐DNA is Cas9 resistant, as MpKAN‐gRNAs target intronic sites. To create an inducible system for the overexpression of MpKAN, we translationally fused MpKAN‐CDS with the glucocorticoid receptor (GR) and expressed the MpKAN‐CDS‐GR with the MpEF1 promoter. To achieve this, MpKAN‐CDS‐StpL‐pENTR/D‐TOPO and pMpGWB113 vectors were recombined to create a MpKAN‐CDS‐StpL‐pMpGWB113 plasmid.

Growth conditions

Plant growth conditions were presented as outlined in previous studies (Gamborg et al., 1968; Ishizaki et al., 2008; Flores‐Sandoval et al., 2015). For controls, Australia/Melbourne accession wild‐type male and female lines were used. Plants were grown under either constant white light (WL, R : FR = 6) or FR‐enriched white light (FR, R : FR = 0.5), with spectral compositions identical for both light treatments apart from FR light. For experiments using the auxin analog 2,4‐dichlorophenoxyacetic acid (2,4‐D), plants were grown on ½ B5 media supplemented with mock (equivalent vol. 100% EtOH) or 0.5 μM 2,4‐D. For MpEF1:MpKAN‐GR experiments, plants were grown on ½ B5 media supplemented with either mock (equivalent vol. 100% EtOH), 1 μM DEX or 10 μM DEX.

Plant transformation and generation of loss‐of‐function alleles

All transgenic lines were generated by the transformation of Australia/Melbourne accession wild‐type sporelings, unless otherwise specified. Marchantia sporelings were transformed following Ishizaki et al. (2008). To generate CRISPR/Cas9‐mediated Mpkan mutants, sporelings were cotransformed with MpKAN‐gRNA1‐pMpGE010 and MpKAN‐gRNA2‐pMpGWB301. Lines harbouring loss‐of‐function alleles were genotyped by sequencing PCR amplicons using primers #12/#13. Two independent lines – Mpkan‐5ge (a male line) and Mpkan‐7ge (a female line) – were isolated (Fig. S2).

Complementation of Mpkan lines

Complementation of Mpkan plants was attained by first crossing Mpkan‐5ge male and wild‐type female lines to generate Mpkan‐5ge/+ sporangia. Sporelings produced from Mpkan‐5ge/+ sporangia were transformed with MpKAN:MpKAN‐CDS‐pMpGWB401 T‐DNA. In this manner, a 1 : 1 Mpkan‐5ge/ MpKAN:MpKAN‐CDS:+/ MpKAN:MpKAN‐CDS population was produced, and complemented Mpkan‐5ge × MpKAN:MpKAN‐CDS lines were identified by genotyping using primer sets #12/#13 and #14/#15.

GUS staining

Tissues from three independent MpKAN:GUS lines were incubated in GUS staining solution (1mM potassium ferrocyanide, 1mM potassium ferricyanide and 1 mM X‐Gluc) for 2–4 h at 37°C and cleared with 70% ethanol.

Tissue sectioning, in situ hybridisation and scanning electron microscopy

Tissue sectioning and in situ hybridisation were performed as described previously (Zachgo, 2002; Retamales & Scharaschkin, 2014). Histological sections were stained with either toluidine blue (1 min) for thalli sections or toluidine blue (1 min) followed by ruthenium red (1 min) for sporophyte sections. The sections were set to 7 and 8 μm thicknesses for histological and in situ sections, respectively. In situ hybridisation was performed using hydrolysed (200 bp) digoxygenin‐labelled antisense MpKAN mRNA as probe. Tissue fixation and SEM were performed as described previously (Flores‐Sandoval et al., 2015).

Phenotype quantification

Air pore counts, thallus area measurements and quantification of dorsal differentiation indices were measured with 7‐d gemmalings. Thallus surface area was measured using the ImageJ software (Schneider et al., 2012). To quantify dorsal differentiation, the surface area of the differentiation zone was measured, and calculated as a fraction of total dorsal surface area to calculate the ‘differentiation area index’ (see Methods S1).

Calculation of θ TS

For the methodology of measuring the angle between the thallus and substrate (θ TS), see Methods S1.

RNAseq experiments

To identify transcriptional targets of MpKAN, RNA sequencing (RNA‐seq) was used to compare the transcript abundance between wild‐type, Mpkan and MpEF1:MpKAN‐GR plants. Transcriptome data were collected from wild‐type and Mpkan lines grown under continuous white light (WL = 15 d WL) and FR‐enriched light (FR = 10 d WL, then 5 d FR) (Methods S1). For transcriptomes generated from wild‐type male and MpEF1:MpKAN‐GR L1 male plants, RNA was extracted from 15‐d gemmalings subsequently treated with either (1) mock (0 h), (2) 10 μM dexamethasone (DEX) (3 h) or (3) 10 μM DEX (24 h) (Methods S1).

Whole thallus material was collected from three replicates for each independent line. Total RNA was isolated (RNeasy Plant Kit, Qiagen, Germany) and assessed for quality using a Nanodrop spectrophotometer and a 2100 Bioanalyser (Agilent, Waldbronn, Germany). Library preparation used polyA mRNA selection and MGIEasy stranded mRNA chemistry. Sequencing used MGI Tech MGISEQ‐2000RS hardware (400 million raw reads per lane, 100‐pb paired‐end reads). Raw reads were mapped onto the Marchantia polymorpha v.5.1 assembly using the Hisat2 alignment software (Kim et al., 2015). Read counting was performed using FeatureCounts to generate matrices of raw reads per gene locus (Liao et al., 2014). Raw counts and normalised transcripts per million (TPM) values for each loci (Table S2).

To corroborate the similarity between replicates for each genotype and condition, principal component analysis (PCA) was performed using R v.3.6.1 and the prcomp function (Fig. S3), with raw reads normalised as log‐transformed TPM values. Pairwise differential gene expression analysis was performed using the edgeR software package (Robinson et al., 2010). Raw outputs of each edgeR pairwise comparison are included in Table S2. Only genes with associated adjusted (adj.) P values of < 0.001 were considered differentially expressed.

Volcano plots were generated using the ggplot2 function in R v.3.6.1 and were annotated for genes encoding transcription regulators, phytohormone‐related proteins, peptide/receptors or light signalling proteins (Table S3). Genes were considered downregulated in DEX‐treated MpEF1:MpKAN‐GR plants when the following conditions were met: (1) logFC < −1, MpKAN‐GR 3h DEX/WT 3 h DEX; (2) logFC < −1, MpKAN‐GR 24 h DEX/WT 24 h DEX; and (3) logFC ≥ 0, WT 3 h DEX/WT 0 h mock. Genes were considered up‐ or downregulated in Mpkan plants relative to wild type when logFC values were > 1 or < −1, respectively, for pairwise comparisons between genotypes and independent of light treatment. For the analysis of differentially expressed genes (DEGs) in Mpkan and wild‐type plants under FR compared with WL, genes were considered FR‐upregulated when logFC > 1 and FR‐downregulated when logFC < −1 for pairwise comparisons between light treatments and within genotypes. To generate a robust list of gene candidates negatively regulated by MpKAN, genes with logFC > 0.5 in all Mpkan/WT pairwise comparisons and logFC < −0.5 in MpKAN‐GR 3 h DEX/WT 3 h DEX and MpKAN‐GR 24 h DEX/WT 24 h DEX pairwise comparisons were selected. To remove genes that were additionally downregulated in DEX‐treated wild‐type plants, genes were also required to have logFC ≥ 0 for the WT 3 h DEX/WT 0 h mock pairwise comparison. GO enrichment analysis was performed on selected gene lists using Plant Transcriptional Regulatory Map (http://plantregmap.cbi.pu.edu.cn/go.php; Peking University) and using Fisher’s exact test to identify statistically significantly enriched GO terms. TPM plots for selected gene loci of interest were generated using the PlotsOfData web application (Postma & Goedhart, 2019).

Arabidopsis complementation experiments

For AtKAN1:MpKAN kan1‐2 kan2‐1 rescue experiments in Arabidopsis, AtKAN1:MpKAN was created by cloning the MpKAN CDS downstream of the AtKAN1 (At5g16560) promoter that consists of a 884‐bp fragment of the conserved second intron fused to a 5.3‐kb fragment upstream of AtKAN1, which has been previously described (Efroni et al., 2008). The AtKAN1:MpKAN construct was then subcloned into pMLBART. Transgenic lines were generated by Agrobacterium‐mediated transformation into wild‐type Columbia and kan1‐2 kan2‐1/+ plants. Plants were grown under long‐day conditions (16‐h day), and transformants were selected on soil for BASTA resistance. Presumptive AtKAN1:MpKAN kan1‐2 kan2‐1 T1 plants were sequenced to determine their genotype. The region of the first exon corresponding to the respective kan1‐2 and kan2‐1 mutations was PCR‐amplified and PCR‐sequenced to confirm the homozygosity of mutant alleles (Eshed et al., 2001) (primers #16–#19).

Results

Expression of MpKAN in gametophyte vegetative and reproductive tissues

To ascertain the gametophyte spatiotemporal expression pattern of MpKAN in Marchantia, MpKAN:GUS transcriptional reporter lines including 6.1 kb upstream of the MpKAN translation start site were stained each day from gemma germination to form a time series of expression. After removal from the gemmae cups (i.e. 0 d), gemmae exhibited staining localised to the apical notches, where the apical meristems are located, with reporter gene expression maintained at the apical notches following germination (Fig. S4a–e). In situ hybridisation on longitudinal thallus sections also indicated MpKAN expression in the region of the apical notch, in both dorsal and ventral epidermal and subepidermal cell files (Fig. 1a). At the commencement of the first bifurcation event after 4 d, reporter gene expression was detected at both apical notches, as well as in photosynthetic filaments of developing air chambers (Fig. 1b). In addition to staining in the apical notches, adult thalli exhibited staining along the midrib of the thallus, and at the bases of primordial gemmae cups, and in situ hybridisation additionally indicated the expression of MpKAN in developing gemmae within gemmae cups (Fig. S4f–h). To determine whether MpKAN expression differed in plants grown under FR, as experienced in shade, gemmalings were also grown under WL for 18 d, followed by FR for 7 d. The staining pattern in vegetative tissues of MpKAN:GUS plants was largely unchanged under FR (Fig. S4i). After prolonged FR exposure, thalli transition to sexual reproduction via the production of gametangia in specialised gametangium‐bearing structures (i.e. gametangiophores). GUS staining was additionally observed in gametangiophores during sexual reproduction, primarily in the receptacles which harbour gametangia (Figs 1c,d, S4j,k). In the sporophyte generation, in situ hybridisation indicated that MpKAN is expressed broadly in developing sporophytes, with no localisation of expression along the axes of symmetry as would be expected for genes that regulate tissue polarity in the sporophyte generation (Fig. S4l,m).

Fig. 1

Expression analysis of MpKAN and generation of Mpkan loss‐of‐function plants in Marchantia. (a) MpKAN transcripts were detected via in situ hybridisation in the immediate dorsal and ventral derivatives of apical cell divisions in longitudinal sections of thalli apical notches. Asterisk indicates putative site of apical cell. d., dorsal; v, ventral; ap, air pore. (b–d) Expression patterns of MpKAN as assessed by MpKAN:GUS reporter gene expression in (b) 4‐d‐old gemmalings, (c) archegoniophores and (d) antheridiophores. (b) In 4‐d‐old gemmalings, when apical notches first bifurcated during plastochron 1, expression was detected in the apical notches (asterisks) and the filaments (arrow) of air chambers closest to the apical notch. (c, d) Reporter gene expression was observed in archegoniophores (c) and antheridiophores (d), predominantly in receptacle tissues. (e) Schematic diagram depicting Mpkan gene structure, domain‐encoding sites and gRNA sites for mRNA sequence of wild‐type and loss‐of‐function alleles. Mpkan‐5ge and Mpkan‐7ge harbour deletions at gRNA1 and gRNA2 targeting sites. (f) Raw RNA‐seq density reads (Sashimi plots) of wild‐type and Mpkan lines at the MpKAN locus. Wild‐type MpKAN mRNA consists of six exons, whereas Mpkan alleles resulted in loss of expression of exons 2–5. WT (g–h), Mpkan (i–j) and Mpkan‐5ge × MpKAN:MpKAN‐CDS (complemented) (k, l) plants grown for 7 d from gemmae. Asterisks indicate apical notches. (m) Wild‐type, Mpkan and complemented lines had largely similar thallus areas. Complemented lines showed restored branching and air pore numbers (n, o), but not gemmae cup numbers (p). (m–p) Data are mean ± SD. Statistical differences among genotypes were determined using one‐way ANOVA and Tukey’s multiple comparisons tests, with letters indicating statistically significant groups (P < 0.01). Where present, numbers after genotype indicate independent lines. Bars: (a) 125 μm; (b–d) 100 μm; (e) 500 bp; (g–l) 2 mm.

MpKAN influences apical growth and differentiation in vegetative tissues

To investigate MpKAN function, loss‐of‐function alleles were created using CRISPR‐Cas9‐mediated mutagenesis, using two gRNAs constitutively coexpressed in transformants to generate deletion alleles (Fig. 1e). Two independent transgenic lines – Mpkan‐5ge (a male) and Mpkan‐7ge (a female) – with 3.3‐kb deletions of genomic sequence spanning the region between gRNA target sites were identified (Fig. S5a–c). The deletion in both lines included 385 bp of exonic sequence encoding 57% of the GARP DNA‐binding domain and caused a frameshift putatively disrupting the LEFTL domain (Fig. 1e). Transcriptome reads demonstrated that Mpkan lines expressed a transcript lacking exons 2–5 (Fig. 1f). To confirm that Mpkan phenotypes described below can be attributed to Mpkan loss‐of‐function alleles, mutant alleles were complemented. Two independently complemented Mpkan‐5ge lines expressing MpKAN:MpKAN‐CDS were corroborated by genotyping (Fig. S5d,e). Complemented Mpkan‐5ge lines showed restored wild‐type phenotypes (Figs 1, 3, 3c–h) with the exception that gemmae cups were not restored (Fig. 1p). The lack of complete complementation suggests that not all endogenous regulatory elements are present in the complementation construct.

Fig. 3

Mpkan mutants fail to complete far red (FR)‐induced developmental transitions in Marchantia. (a) θ TS values of 5‐d‐old gemmalings grown under constant WL (solid bars) and FR (striped bars). Data are mean θ TS (columns) ± SD. Statistical differences obtained using two‐way ANOVA and Tukey’s multiple comparisons tests. Letters denote statistically significant different groups (P < 0.05). (b) Ratio of receptacles : apical notches in wild‐type and Mpkan FR‐induced plants. Data are mean ± SD. (c–h) Comparison between the sexual tissues of wild‐type, Mpkan and complemented lines induced under FR‐enriched light conditions. Plants were grown under FR from gemmae for 35 d (c–e, g, h) and 45 d (f). (i–q) SEM imaging of dorsal (i, j, m–o) and ventral (k–l, p–q) reproductive tissues (*, other positions where archegonia are evident). Wild‐type male (i, k) and female (m, p) and Mpkan male (j, l) and female (n, o, q) thalli were grown for 55 d under FR. Magnified insets depict (o) archegonia and (p) involucres. Panels (i–n) and (p) are composites of SEM images. Bars: (c–h, i–n, p) 1 mm; (o, q) 500 μm.

Mpkan mutants were of similar size to wild‐type plants, and dorsal air pores and air chambers had a morphology comparable to those of wild type (Figs 1m, S6a,b,e). However, significantly fewer air pores were formed in the mutant background (Figs 1n, S6f). Additionally, the first bifurcation event was further progressed in Mpkan plants, with apical notches more laterally displaced from each other (Fig. 2a–f), indicating that the reduction in air pore numbers was not caused by reduced or delayed apical activity. Mpkan plants also developed abnormal, wing‐like outgrowths from the central zone of the gemmaling, a region that is not developmentally active in wild type, and produced more dorsal rhizoids than wild type from the central region (Fig. 2a–f). Ventrally, Mpkan plants developed scales and smooth rhizoids of similar morphology and positioning to those of wild type (Fig. S6c,d). No evidence was found of polarity abnormalities in Mpkan plants, with dorsal and ventral tissue types correctly positioned along the dorsiventral axis (Figs  2a,b, S6c,d), and longitudinal sectioning of 5‐d gemmalings corroborated that the loss of MpKAN function did not impair the establishment of a dorsiventral thallus (Fig. S6g,h).

Fig. 2

Loss of MpKAN activity affects growth and differentiation in the vegetative gametophyte in Marchantia. (a, b) Scanning electron micrographs of dorsal tissues in wild‐type (a) and Mpkan‐5ge (b) gemmalings. (c–f) Light microscopy images of wild‐type (c, d) and Mpkan (e, f) gemmalings. (a–f) Asterisks indicate apical notches, and arrows, outgrowths from central zone. (g) Differentiation area index values of wild‐type and Mpkan lines. Statistical differences among genotypes were determined using one‐way ANOVA and Tukey’s multiple comparisons tests, with letters indicating statistically significant groups (P < 0.001). (h) Timing of completion of the first three branching plastochrons in wild‐type and Mpkan plants. Statistical differences were calculated using one‐way ANOVA and Tukey’s multiple comparisons tests. P‐values: **, <0.01; *, <0.05; ns, nonsignificant. (g, h) Data are mean ± SD. (i–l) Morphology of 24‐d‐old wild‐type and Mpkan thalli. (i, j) Wild‐type plants have entered plastochron 4 and have produced two series of gemmae cups originating during plastochrons 2 and 3. (k, l) Mpkan plants are further progressed into plastochron 4; however, gemmae cups have only emerged during plastochron 3. Bars: (a, b) 500 μm; (c–f) 1 mm; (i–l) 5 mm.

We further characterised the apparent differentiation defect in Mpkan dorsal tissue types. After 8‐d growth from gemmae, a distinct boundary is discernible between differentiated dorsal tissue originating from the apical cell, herein referred to as the differentiation zone (dz), and a region derived from the central region of the gemma where cells do not divide and remain developmentally inactive, referred to as the central zone (cz) (Methods S1). To quantitate the extent of dorsal differentiation, the dz surface area was measured and calculated as a fraction of total dorsal surface area to give a value called the ‘differentiation area index’. Differentiation area indices were significantly reduced in Mpkan mutants (Fig. 2g), indicating that they are compromised in their ability to differentiate dorsal tissue types.

Mpkan mutants produced more apical notches and fewer gemmae cups when compared with wild type (Figs 1o,p, 2i–l, S6i,j). Since apical notch numbers were increased in Mpkan plants, we assessed whether the phenotype reflected faster apical branching during specific plastochrons. To test this, plastochron duration was measured for each of the first three branching events. Plastochron 1 (the time taken to bifurcate from 2 to 4 notches) was completed in significantly fewer days in Mpkan lines compared with wild type, whereas plastochrons 2 and 3 took a comparable number of days to be completed (Fig. 2h). In agreement with this, SEM images of gemmalings showed that Mpkan mutants were already markedly bifurcating at this timepoint, whereas in wild type, bifurcation was either yet to commence or in its early stages (Fig. S6k–n). Altogether, the increase in apical notch number in Mpkan plants was caused primarily by an increased rate of bifurcation during plastochron 1.

Since KAN orthologs in angiosperms are active in the sporophyte generation, and in situ hybridisation indicated expression of MpKAN in the sporophyte, we next examined the phenotypes of Mpkan homozygous sporophytes. Homozygous Mpkan sporophytes were not impaired in polarity along either radial or apical–basal axis of symmetry, with distinct foot, seta and capsule tissues forming in correct positions along the apical–basal axis, and amphithecial (i.e. capsule wall cell types) and endothecial (i.e. spores and elaters) layers correctly positioned along the radial axis of the capsule (Fig. S6o–r). Thus, similarly to gametophyte tissues, we found no evidence, suggesting MpKAN regulates polarity in the sporophyte.

Given that KAN represses auxin biosynthesis gene expression in angiosperms (Brandt et al., 2012; Merelo et al., 2013; Xie et al., 2015), we looked for the presence of auxin‐related phenotypes in Mpkan plants. Wild‐type gemmalings grown on low concentrations of the auxin analog 2,4‐D display two developmental abnormalities. First, numerous rhizoids are produced from the dorsal cz (Fig. S7a,b) (Tarén, 1958; Flores‐Sandoval et al., 2015). Second, secondary thalli are occasionally initiated at the cz/dz boundary (Fig. S7a,b), a process thought to be caused by growth from a second gemma apical cell that under normal conditions is repressed via the eventual establishment of a dorsiventral auxin gradient (Halbsguth & Kohlenbach, 1953). Similar to 2,4‐D‐treated wild‐type plants, mock‐treated Mpkan plants produce profuse dorsal rhizoids from the cz and occasionally produce secondary thalli from the cz/dz boundary (Fig. S7c–e). Together, these results suggest that Mpkan lines phenocopy pharmacological treatment with low auxin concentrations.

MpKAN as a modulator of growth responses to FR‐enriched light

Given that KAN TFs regulate FR light responses in angiosperms (Merelo et al., 2017; Botterweg‐Paredes et al., 2020), we next examined whether the loss of MpKAN activity affected FR responses in Marchantia. The growth under FR causes two developmental changes compared with WL, namely a transition from plagiotropic to orthotropic growth, with the thallus tips increasingly elevated from the substrate, and the transition to sexual development via the formation of gametangiophores (Fig. S8a–f) (Fredericq & de Greef, 1966; Inoue et al., 2019; Yamaoka et al., 2021). In contrast to wild type, Mpkan mutants continued to grow plagiotropically under FR‐enriched light and exhibited delayed gametangiophore initiation (Fig. S8a–l). The angle formed between the thallus and substrate (θ TS) was compared between Mpkan and wild type under WL and FR. In wild type, θ TS significantly increased under FR when compared with WL (Fig. 3a). By contrast, Mpkan plants had significantly lower θ TS under both light conditions than wild type, and θ TS did not significantly increase under FR (Fig. 3a). In the case of Mpkan‐7ge, mean θ TS values were negative, with the thallus margins growing into the substrate. Thus, Mpkan plants displayed reduced orthotropic growth compared with wild type under both light conditions, and failed to grow orthotropically in response to FR.

We next examined gametangiophore initiation in Mpkan lines. In wild type, the first gametangiophores were produced after 18–20 d in males and 20–22 d in females, with a marked increase in the number of gametangiophores per notch from 22 to 26 d (Fig. 3b). In Mpkan plants, gametangiophore‐like tissues were produced, and the appearance of the first reproductive tissues was delayed by 8 d in males and 14 d in females (Fig. 3b). Thus, FR‐induced gametangiophore initiation was delayed in the Mpkan mutant background.

Additionally, a number of morphological abnormalities were observed in Mpkan reproductive tissues (Fig. 3c–q). Wild‐type gametangiophores develop via modifications of the thallus, giving rise to the two specialised regions of the gametangiophore: the stalk and the gametangium‐bearing receptacle. In males, three successive rounds of dichotomous branching produce an 8‐lobed receptacle, with apical notches located at lobe tips (Fig. 3c,i,k). Antheridia are produced from dorsal derivatives of the apical cell and from within dorsal cavities of antheridiophore receptacles. The dorsal surface of the antheridiophores receptacle produces both air pores and antheridial pores, with the latter exuding sperm when antheridia are mature (Fig. S9a). In wild‐type females, three rounds of dichotomous branching, with occasional extra bifurcations, give rise to archegoniophores possessing 9–11 digitate rays, with apical notches residing between the rays (Fig. 3d,m,p). Archegonia form from dorsal epidermal derivatives of apical cells and are located between the bases of the rays. Archegonia are protected by involucres (Figs 3p, S9d). In both sexes, the receptacles are elevated on stalks that form by inward rolling of the thallus, where the dorsal surface of the thallus envelopes an internalised ventral side producing scales and bundles of rhizoids (Figs 3c,d, S9g,j). No air pores form on the surface of the stalk, and dorsal epidermal cells are columnar in shape (Fig. S9g,j).

In Mpkan lines grown under FR, mature antheridia and archegonia form in male and female lines, respectively, but without characteristic thallus modifications. In both sexes, gametangiophore morphology resembled the wild‐type vegetative thallus morphology and lacked the clear distinction between stalk and receptacle tissues. Stalk regions were stunted and not inward‐rolled, and columnar‐shaped air pores were present (Figs 3e,f,j,n, S9h,i,k,l). Gametangiophore receptacle morphology was also intermediate between wild‐type vegetative and gametangiophore morphology (Fig. 3e,f,j,n,o). In Mpkan‐5ge males, antheridia formed internally in receptacles that were less lobed than wild‐type antheridiophore receptacles (Fig. 3e,j,l). Dorsally, most antheridial pores and air pores formed normally (Fig. S9b), but occasionally, antheridium‐like structures were observed protruding from large pore‐like structures (Fig. S9c). In Mpkan‐7ge females, archegoniophores were similarly less modified from the vegetative thallus morphology (Fig. 3f,n,q). Archegonia were produced, but lacked involucres (Figs 3o, S9e,f). Mature Mpkan‐5ge and Mpkan‐7ge gametangia had fertile sperm and egg cells, respectively, as crosses between Mpkan‐7ge and Mpkan‐5ge produced viable Mpkan‐7ge/Mpkan‐5ge sporophytes (Fig. S8n,o). Complemented Mpkan‐5ge lines of both sexes grown under FR displayed normal gametangiophore development (Fig. 3g,h). Altogether, these results suggest after gametangium initiation, MpKAN regulates the transition from vegetative to sexual tissue types (i.e. gametangiophore formation) under FR.

Effect of inducible MpKAN overexpression on development and FR growth responses

To examine the effects of MpKAN overexpression, DEX‐inducible MpEF1:MpKAN‐GR transgenic lines were generated. Mock‐treated MpEF1:MpKAN‐GR lines resembled wild‐type plants, whereas DEX‐treated MpEF1:MpKAN‐GR lines displayed a range of weak‐to‐strong phenotypic defects (Figs 4a, S10a). DEX‐treated MpEF1:MpKAN‐GR gemmalings were smaller and produced fewer air pores and no gemmae cups, and stronger lines displayed severely reduced apical growth and loss of dorsal differentiation (Figs 4a, S10a–c). MpEF1:MpKAN‐GR plants grown on mock for 20 d and then transferred to DEX‐produced gemmae cups with significantly fewer gemmae, with plastochron 4‐derived gemmae cups containing almost no gemmae (Fig. S10d–f). DEX‐treated MpEF1:MpKAN‐GR lines also displayed abnormal growth tropisms in response to light treatments. Under WL, DEX‐treated MpEF1:MpKAN‐GR plants grew more orthotropically than wild type, and comparably with wild‐type plants grown under FR (Fig. 4b), indicating that MpKAN overexpression can induce growth tropisms under WL similar to those normally observed under FR. Under FR, DEX‐treated MpEF1:MpKAN‐GR plants failed to produce gametangiophores (Fig. S11). Altogether, constitutive overexpression of MpKAN causes a reduction in growth, the number of air pores, gemmae cups and gemmae, and promotes orthotropic growth under both WL and FR.

Fig. 4

DEX‐inducible overexpression of MpKAN and analysis of transcriptional targets of MpKAN in Marchantia. (a) Phenotypes of wild type (left) and MpEF1:MpKAN‐GR (right) transformants treated with mock (top) and 10 μM DEX (bottom). Plants were grown for 9 d. (b) Comparison of angle between thallus and substrate (θ TS) of mock and DEX‐treated plants. Data are mean (columns) ± SD. Statistical differences were obtained using two‐way ANOVA and Tukey’s multiple comparisons tests. Letters denote statistical significant differences (P < 0.05). (c, d) TPM plots of (c) MpTAA and (d) MpDELLA loci from transcriptome profiles of white light (WL)‐ and far red (FR)‐treated wild‐type and Mpkan plants. Values associated with brackets are logFC values (adj. P < 0.001). (e‐h) Volcano plots showing FR‐upregulated and FR‐downregulated genes in wild‐type (e, f) and Mpkan (g, h) transcriptomes. LogFC values obtained using the edgeR software comparing wild‐type FR with wild‐type WL (e, males; f, females) and Mpkan FR with Mpkan WL (g, males; h, females). Yellow loci, nonsignificant (adj. P < 0.001); grey loci, −1 < logFC < 1; blue loci, FR‐upregulated (logFC > 1) or FR‐downregulated (logFC < −1) genes. Transcriptional regulators, peptides/receptors, and phytohormone‐ and light‐related genes (Bowman et al., 2017) were annotated on volcano plots. Genes of interest with logFC > 1 or logFC < −1 are labelled in black. Genes of interest in teal showed same FR response in all sexes and genotypes. Genes of interest in pink showed same FR response in both sexes for one genotype, but not in the other genotype (and are shown in red text in volcano plots of other genotype). Bars, (a) 1 mm.

MpKAN negatively regulates expression of MpTAA and MpDELLA, and has reduced FR‐induced changes in gene expression

To investigate downstream targets of MpKAN, we looked at transcriptome changes in whole thallus tissue collected from Mpkan, MpEF1:MpKAN‐GR and wild‐type lines. We first identified gene expression upregulated or downregulated in MpEF1:MpKAN‐GR lines compared with wild type after DEX treatment (Table S2). DEG counts in DEX‐treated MpEF1:MpKAN‐GR plants revealed a higher number of downregulated than upregulated genes after 3 h of DEX treatment, suggesting that, as expected, MpKAN acts primarily as a transcriptional repressor (Figs S12, S13a,b). The GO analysis of differentially expressed genes revealed that a significantly enriched number of phenylpropanoid/flavonoid biosynthesis genes were downregulated in DEX‐treated MpEF1:MpKAN‐GR plant transcriptomes, suggesting MpKAN regulates flavonoid biosynthesis (Tables S2, S4).

We next looked at changes in gene expression in Mpkan plants relative to wild type under both WL and FR. Genes whose expression is upregulated in the Mpkan mutant background included the sole TAA ortholog MpTAA (Eklund et al., 2015), and the R2R3‐MYB TF MpR2R3‐MYB14, which has been characterised as a master regulator of flavonoid biosynthesis in Marchantia (Albert et al., 2018; Kubo et al., 2018) (Fig. S13c–f). By combining analyses of both RNAseq experiments, a subset of 29 genes were identified, which showed opposite transcript abundance changes in DEX‐treated MpEF1:MpKAN‐GR and Mpkan plants compared with relevant controls (Table 1; Fig. S13g). Of particular note, MpTAA and MpDELLA were upregulated in Mpkan plants under both light conditions and downregulated in DEX‐treated MpEF1:MpKAN‐GR plants, suggesting that both are targets, either directly or indirectly, of negative regulation by MpKAN (Table 1; Figs 4c,d, S14a,b). The overexpression of MpDELLA has been shown to cause complete suppression of gametangiophore initiation under FR, and promotes expression of flavonoid biosynthesis genes (Hernández‐García et al., 2021b), both phenotypes that are shared with Mpkan plants. To investigate the link between MpKAN and MpDELLA further, we compared genes upregulated in MpDELLA overexpression lines (data from Hernández‐García et al., 2021) with those upregulated in FR‐treated Mpkan lines. There was substantial overlap between gene lists (Fig. S13h), with phenylpropanoid and flavonoid biosynthesis‐related overrepresented in the subset of genes upregulated in both MpDELLA‐ox and FR‐treated Mpkan lines (Fig. S15; Table S4), suggesting that flavonoid biosynthesis genes are shared downstream targets of MpKAN and MpDELLA.

Table 1

Genes upregulated in Mpkan plants and downregulated in DEX‐treated MpEF1:MpKAN‐GR plants in Marchantia.

			Fold change response (log2FC)
			Mpkan‐5ge ♂/WT ♂		Mpkan‐7ge ♀/WT♀		pMpEF1:MpKAN‐GR/WT
v.5.1 ID	Name	Description	WL	FR	WL	FR	3 h DEX	24 h DEX
Mp1g04680		Cytochrome P450	1.89	2.37	2.02	2.44	−0.85	−0.89
Mp1g21930	MpLOX5	LOX2S; lipoxygenase	1.24	1.16	1.53	1.83	−1.53	−1.76
Mp1g28830		Unannotated	4.42	6.24	3.01	5.61	−2.59	−3.39
Mp2g02350		Protein of unknown function (DUF3455)	2.34	3.2	1.47	3.19	−1.02	−1.92
Mp2g03050	MpCHK1	Cytokinin receptor	0.93	0.53	0.84	0.56	−1	−0.69
Mp2g03280	MpCCD7	9‐cis‐beta‐carotene 9′,10′ cleaving dioxygenase	1.47	3.77	1.37	3.6	−1.56	−3
Mp2g13230		Unannotated	3.02	3.47	3.07	2.46	−1.62	−2.43
Mp3g10530		Serine/threonine protein kinase	2.82	2.2	2.64	1.38	−1.43	−2
Mp3g15790		Peroxidase	7.27	8.01	6.17	6.31	−3.99	−4.49
Mp3g17960		Unannotated	1.07	1.15	1.12	1.09	−1.05	−1.66
Mp3g21740		Peroxidase/oxygenase	2.2	1.88	1.16	1.1	−1.1	−1.62
Mp4g03600		Sulphite exporter TauE/SafE	2.15	2.64	2.84	3.32	−1.76	−1.46
Mp4g12310		Serine/threonine protein kinase	0.84	0.63	0.68	0.53	−0.72	−0.91
Mp4g15440		Indole‐3‐acetate O‐methyltransferase	0.86	1.49	0.53	1.2	−0.59	−0.77
Mp4g16270		Unannotated	1.17	1.13	1.58	0.75	−0.62	−0.74
Mp5g00560		Methyltransferase	3.12	2.7	3.02	1.84	−1.52	−2.93
Mp5g06850		O‐methyltransferase	0.67	1.3	0.71	1.47	−1.1	−0.69
Mp5g14320	MpTAA	Aminotransferase	1.94	2.48	1.55	1.82	−1.15	−0.64
Mp5g14990		Unannotated	1.59	0.89	1.55	1.4	−0.88	−1.03
Mp5g16240		EamA‐like transporter	3.68	3.83	3.23	2.59	−2.78	−2.67
Mp5g20660	MpDELLA	Transcription factor, GRAS family	0.58	1.14	0.6	1.12	−1.45	−0.9
Mp7g01790		6‐methylsalicylate decarboxylase	1.26	1.54	1.25	1.44	−0.92	−0.75
Mp7g08280		Unannotated	0.53	1.24	0.63	1.41	−0.6	−1.45
Mp7g09830		Alpha/beta hydrolase	1.44	1.02	1.68	1.29	−0.83	−0.87
Mp7g11540		3‐phosphoglycerate kinase	1.91	1.11	2.31	1.13	−0.55	−1.19
Mp7g14650		3‐methylcrotonyl‐CoA carboxylase	1.08	0.76	1.55	1.16	−1.05	−0.65
Mp7g16380		rnc, DROSHA, RNT1; ribonuclease III	1.21	1.32	1.72	1.82	−4.36	−2.49
Mp8g06340		Dirigent‐like protein	1.05	0.95	1.49	1.4	−0.83	−1.54
Mp8g11680		Polyketide cyclase/dehydrase	0.93	1.31	1.11	1.36	−0.68	−1.46

Gene list of loci with opposite changes in transcript abundance of RNA samples from Mpkan, wild‐type and MpEF1:MpKAN‐GR lines; fold changes are measured using EdgeR analysis (adj. P < 0.001). Genes in list meet the following requirements: logFC > 0.5, Mpkan/WT; logFC < −0.5, MpEF1:MpKAN‐GR/WT; and logFC ≥ 0 WT 3 h DEX/WT 0 h mock.

To identify genes downstream of MpKAN activity that may cause reduced FR responses in Mpkan plants, we examined the differences in gene expression between wild‐type and Mpkan lines after 5 d in FR light conditions, targeting genes involved in early gametangiophore specification (Fig. 4e–h). FR‐upregulated genes in wild‐type males and females included MpBONOBO (MpBNB), which encodes a TF necessary for gametangium specification in Marchantia (Yamaoka et al., 2018), as well as the TALE‐homeodomain TF‐encoding gene MpBELL1, consistent with a previous study (Inoue et al., 2016) (Figs 4e,f, S14i,j). In wild‐type females, the regulator of female sexual differentiation, MpFGMYB (Hisanaga et al., 2019), was FR‐upregulated (Fig. 4f). Candidate gibberellic acid biosynthesis genes orthologous to those encoding enzymes that catalyse GA12 from GGDP in angiosperms (Fig. S14c) (Hedden, 2020), as well as cytokinin biosynthesis genes (MpIPT2 and MpLOG) (Bowman et al., 2017), were also FR‐upregulated in wild‐type males and females (Figs 4e,f, S14d–g). In Mpkan lines, transcript levels of MpBNB and most candidate GA12 biosynthesis genes were no longer FR‐upregulated (Figs 4g,h, S14d–i; Table S3), indicating that FR‐induced upregulation of these genes is dependent on MpKAN activity. Gene ontology (GO) analysis indicated that FR‐upregulated genes in Mpkan plants were not enriched for gibberellin biosynthesis and metabolism terms, as observed in FR‐upregulated genes in wild‐type plants (Table S4). In Mpkan lines, however, MpBELL1 and the cytokinin biosynthesis genes were still FR‐upregulated (Figs 4e–h, S14j), implying independence from MpKAN activity.

MpKAN can rescue the sporophyte polarity defects of Arabidopsis kanadi mutants

In angiosperms KAN, TFs promote a sporophyte abaxial identity (Eshed et al., 2001), whereas MpKAN does not regulate tissue polarity in either generation of Marchantia. To assess whether MpKAN can fulfil this role in angiosperms, and whether KAN TFs are conserved at the protein level across land plants, we expressed MpKAN with the Arabidopsis AtKAN1 promoter in kanadi1‐2 kanadi2‐1 (kan1‐2 kan2‐1) plants. Remarkably, the phenotype of AtKAN1:MpKAN kan1‐2 kan2‐1 plants approaches that of wild type in leaf, inflorescence and flower morphology (Figs 5, S16), indicating that MpKAN can substitute for reduced AtKAN abaxial identity activity in kan1‐2 kan2‐1 mutants, and demonstrating that MpKAN maintains an AtKAN biochemical function.

Fig. 5

MpKAN can rescue the polarity defects of kanadi mutants in Arabidopsis. Comparison between leaves (a) and bolting plants (b) of representative wild‐type, kan1‐2 kan2‐1/+, kan1‐2 kan2‐1 and pAtKAN1:MpKAN kan1‐2 kan2‐1 long‐day grown plants. (a) Detached 6th leaf of each of the four genotypes with the adaxial surface facing upwards, with the exception of the kan1‐2 kan2‐1 leaf, which is positioned on its side with abaxial (ab), and adaxial (ad) surfaces indicted. kan1‐2 kan2‐1/+ and pAtKAN1:MpKAN kan1‐2 kan2‐1 leaves can be seen to curl inwards relative to wild type but lack the pronounced abaxial outgrowths and reduced expansion of kan1‐2 kan2‐1 leaves. (b) 7‐week‐old bolting plants with the exception of kan1‐2 kan2‐1, which bolt later because of growth disruptions to the inflorescence stem. Some perturbance in kan1‐2 kan2‐1/+ and pKAN1:MpKAN kan1‐2 kan2‐1 inflorescence growth is apparent as twisted stem in axillary shoots (arrowheads). Bars: (a) 5 mm; (b) 1 cm.

Discussion

KANADI TFs are biochemically conserved across land plants

Since the loss of KAN activity in Arabidopsis can be partially complemented by transgenic expression of MpKAN (Figs 5, S16), we conclude that KAN protein structure, and its ability to act with partners, is largely conserved across land plants. However, as discussed below, our results suggest distinct biological roles of KANADI in liverworts and angiosperms, indicating the evolution of either (1) changes in expression domains between lineages or (2) the divergence of downstream transcriptional targets via changes in cis‐regulatory binding sites.

MpKAN influences apical growth and dorsal differentiation

In contrast to vascular plants, MpKAN – while expressed in the sporophyte generation – did not have a conspicuous developmental role in the sporophyte of Marchantia (Figs 1, S6). Again in contrast to angiosperms, MpKAN plays a functional role in the gametophyte generation of Marchantia (Fig. 1). MpKAN expression was observed throughout the apical notches in the epidermal and subepidermal files close to the meristematic apical cell (Fig. 1) and is thus expressed where most developmental decisions are occurring. The loss of MpKAN activity resulted in more rapid bifurcations during the first plastochron, as well as reduced tissue differentiation (Figs 2, S6). Constitutive expression of MpKAN caused strong suppression of apical growth and the loss of gemmae production in developing gemmae cups, the latter process involving meristematic gemma initials (Figs 4, S10). Taken together, these results suggest that MpKAN is either (1) a negative regulator of meristem activity in the apical notch; (2) a positive regulator of differentiation of subepidermal and epidermal cells as they are displaced from the apical cell; or (3) a combination of both roles. Given that two key meristem‐associated genes – MpTAA and MpCLE1 (Eklund et al., 2015; Hirakawa et al., 2019) – were downregulated in DEX‐treated MpEF1:MpKAN‐GR plants (Tables 1, S2), one might presume that MpKAN has at least the first of these roles.

In angiosperm sporophytes, KAN members regulate radial patterning by promoting abaxial and peripheral tissue types in the lateral organs and shoot (Eshed et al., 2001, 2004; Kerstetter et al., 2001; Emery et al., 2003; Izhaki & Bowman, 2007). In the gametophyte generation of Marchantia, MpKAN is not expressed in a polarised manner, and Mpkan plants do not display defects in patterning along spatial axes (Figs 1, 2, S6). Additionally, Mpkan‐5ge/Mpkan‐7ge homozygous sporophytes did not display tissue polarity defects (Fig. S6). Thus, in contrast to KAN orthologs in angiosperms, MpKAN is not required for co‐ordinating tissue polarity in either generation in Marchantia, suggesting that a role in patterning along spatial axes may have evolved de novo within vascular plant sporophytes, or was lost in bryophytes – a scenario that is equally parsimonious if bryophytes are monophyletic (Puttick et al., 2018).

The function of KAN members in Arabidopsis embryo development goes beyond tissue polarity, as kan triple mutants show ectopic organ formation, where KAN TFs suppress organogenesis via the repression of PIN1‐FORMED (PIN1) auxin transporters (Izhaki & Bowman, 2007). This bears resemblance to the role of MpKAN in the Marchantia gametophyte generation, with MpKAN‐mediated thallus differentiation potentially connected with its capacity to downregulate MpTAA and cause auxin depletion in Marchantia (Figs 4, S7). Indeed, the loss of MpKAN function in Marchantia phenocopies auxin‐treated wild‐type plants possibly due to the regulation of MpTAA. Further functional analyses in bryophytes and seedless tracheophytes could corroborate the deep conservation of this function across land plants.

MpKAN modulates developmental FR responses, including the transition to sexual reproduction

MpKAN functions in multiple aspects of development in response to FR in Marchantia. First, Mpkan thalli displayed reduced orthotropic growth compared with wild type under both WL and FR (Fig. 3). Conversely, MpKAN overexpression caused orthotropic growth under both WL and FR (Fig. 4), suggesting MpKAN may be sufficient to initiate FR‐induced orthotropic growth. Similarly, Mpphy mutants grow orthotropically under both WL and FR, suggesting that MpPHY inhibits orthotropic growth (Inoue et al., 2019). Thus, the plagiotropic growth form of Mpkan mutants under both light treatments may reflect reduced phytochrome‐mediated growth tropisms. While not homologous structures, in Arabidopsis leaves phytochrome (PHYB) and KAN activities also affect leaf flatness, that is plagiotropic growth, in an opposite manner (Kozuka et al., 2013; Johansson & Hughes, 2014). Second, Mpkan mutants exhibited a delayed transition to sexual reproduction following FR treatment. In wild type, FR‐upregulated genes included MpBNB, a master regulator of gametangium initiation in Marchantia that is required for subsequent gametangiophore formation (Yamaoka et al., 2018). MpBNB transcripts were not FR‐upregulated in Mpkan plants (Fig. 4), suggesting that the loss of MpKAN activity causes a delay in the transition to sexual reproduction via delayed activation of MpBNB. Third, Mpkan mutants produced vegetative‐like gametangium‐bearing structures, with the Mpkan reproductive thallus morphologically intermediate between wild‐type vegetative thallus tissue and wild‐type gametangiophore tissues (Fig. 3). Given that gametangiophores form only after MpBNB‐mediated gametangium initiation (Yamaoka et al., 2018, 2021), this would suggest that MpBNB activity is not entirely lost in Mpkan plants and that after gametangium specification, MpKAN is additionally required for gametangiophore formation under FR. One might predict that MpKAN overactivity would cause early or ectopic formation of gametangiophores; however, MpKAN overexpression caused complete suppression of gametangiophore formation (Fig. S11), likely due to the previously mentioned role as a negative regulator of meristem activity.

MpDELLA physically interacts with PHYTOCHROME‐INTERACTING FACTOR (MpPIF), with both transcription factors acting in a pathway‐regulating gametangium and gametangiophore formation under FR (Inoue et al., 2019; Hernández‐García et al., 2021b). FR‐treated Mpkan plants expressed elevated levels of MpDELLA, and MpDELLA and MpKAN similarly influence gene expression involved in flavonoid biosynthesis (Figs 4, S15). Thus, we postulate that MpKAN and MpDELLA may act in a genetic pathway wherein MpKAN negatively regulates MpDELLA, which in turn inhibits the activity of MpPIF, whose activity promotes both the transition to sexual reproduction and flavonoid biosynthesis (Fig. 6). However, MpPIF and MpKAN possess distinct roles in directing gametangiophore morphogenesis, with the former required only for their initiation. Additionally, as MpTAA is upregulated in Mpkan plants under WL and FR (Fig. 4), and in Mppif plants under FR (Hernández‐García et al., 2021b), MpPIF may negatively regulate auxin biosynthesis gene expression in response to FR, although how auxin biosynthesis is related to FR responses in Marchantia remains unclear. Notably, Arabidopsis KAN orthologs, AtKAN1 and AtKAN4, can physically interact with DELLA TFs, and KAN and DELLA control integument development via a shared regulatory pathway (McAbee et al., 2006; Kelley et al., 2009; Gomez et al., 2016). AtKAN1 also directly negatively regulates DELLA gene expression (Merelo et al., 2013). With these results considered together, we propose that MpKAN modulates the output of phytochrome‐mediated FR responses, including gametangiophore initiation, and that the general role of MpKAN is to promote differentiation, with ancient environmental responses pathways (i.e. FR light responses) recruiting a land plant‐specific module, involving KAN and DELLA, as a potent differentiation factor.

Fig. 6

Theoretical pathways for the regulation of far‐red (FR) responses by KAN transcription factors (TFs) in land plants. Components with yellow background are proposed to be conserved in land plants, ones with white background are not conserved, and those with green background have insufficient information to determine whether pathway position is conserved in land plants. Arrows and blunt‐ended arrows indicate positive and negative regulations, respectively. (a) In Marchantia, MpKAN regulates MpDELLA expression, which in turn negatively regulates MpPIF activity. Under high FR, MpPHY enters the nucleus and positively regulates MpPIF, which promotes the transition to sexual reproduction via gametangium initiation. (b) In Arabidopsis, phytochromeB (phyB) is nuclear localised under high R : FR, where it negatively regulates PIF activity via (1) protein–protein interactions resulting in degradation via the ubiquitin–proteasome system; and (2) reduced GA levels, which frees the DELLA transcription factor AtRGA to repress AtPIF4 and AtPIF5 activity. Under low R : FR (shade), phyB is located in the cytoplasm, resulting in increased PIF activity. PIF and KAN TFs act antagonistically to regulate expression of AtTAA1 expression, which in turn promotes hypocotyl elongation. Theoretical pathways are formulated based on results of this paper and previous work (Casal, 2013; Merelo et al., 2017; Inoue et al., 2019; Hernández‐García et al., 2021b). Question mark indicates partially supported pathway step where only AtKAN1 negatively regulates AtRGA (Merelo et al., 2013).

Loss of KANADI activity disrupts FR‐induced expression of GA12 biosynthesis genes

In wild‐type Marchantia, FR‐upregulated genes included orthologs of cis‐Zeatin and GA12 biosynthesis genes in Arabidopsis, making the activity of both molecules candidates for mediating FR responses. In Mpkan plants, GA12 biosynthesis genes were no longer upregulated in response to FR (Figs 4, S14). AtKAN4 also suppresses GA biosynthesis, including GA12 production (Gomez et al., 2016). Bryophytes are unlikely to produce canonical bioactive GAs found in angiosperms and do not possess GID1 receptor orthologs, which in vascular plants are required for interactions between GA and DELLA proteins (Hernández‐García et al., 2021a). In one plausible scenario, in the ancestral land plant, KAN TFs regulated DELLA activity and GA production independently, and in the vascular plant ancestor, the DELLA pathway became GA‐dependent via acquisition of the GID1 receptor. Alternatively, in the ancestral land plant KAN TFs regulated a GID1‐DELLA‐GA pathway, and in the ancestral bryophyte GID1 orthologs were lost, decoupling DELLA from GA regulation.

KANADI TFs – conserved regulators of developmental FR responses

In flowering plant sporophytes, KAN TFs regulate both tissue polarity and shade avoidance responses via common gene targets (Merelo et al., 2017). Here, we found that MpKAN does not regulate tissue polarity in either generation of the Marchantia life cycle and instead functions in at least two contexts during the gametophyte generation. In the context of vegetative development, MpKAN is active in the derivatives of the apical cell, where it regulates thallus apical growth and tissue differentiation. In the context of FR‐enriched light treatment, MpKAN modulates multiple FR responses including MpBNB‐mediated gametangiophore initiation and positively regulates gametangiophore formation after the gametangium initiation have been specified. While acting in alternating generations, KAN TFs may have conserved roles as modulators of FR responses in both angiosperms and liverworts, a function presumably present in the ancestral land plant.

Author contributions

LNB, EF‐S and JLB designed the research; LNB, JPA, TD and EF‐S performed research and analysed data; LNB and JLB wrote the paper with input from all authors.

Fig. S1 Promoter, gene and protein sequences of MpKAN.

Fig. S2 Sequence alignments of wild‐type and Mpkan loss‐of‐function alleles.

Fig. S3 Principle component analysis of RNA‐seq samples.

Fig. S4 Expression of MpKAN in vegetative and reproductive tissue types.

Fig. S5 Genotyping alleles of Mpkan loss‐of‐function and complemented lines.

Fig. S6 Morphology of wild‐type and Mpkan plants grown under white light.

Fig. S7 Loss of MpKAN activity may cause altered auxin distributions.

Fig. S8 Loss of MpKAN activity causes altered responses to FR light.

Fig. S9 Scanning electron micrography of wild‐type and Mpkan reproductive tissues.

Fig. S10 Effect of constitutive over‐expression of MpKAN.

Fig. S11 Effect of constitutive MpKAN overexpression on sexual reproduction.

Fig. S12 RNA‐seq analysis: Differentially expressed gene counts.

Fig. S13 Volcano plots of gene loci.

Fig. S14 TPM plots of differentially expressed gene loci of interest.

Fig. S15 Heatmap of expression changes in flavonoid biosynthesis genes.

Fig. S16 MpKAN can rescue the flower and silique defects of Arabidopsis kanadi mutants.

Methods S1 Supplementary Materials and Methods.

Click here for additional data file.

Table S1 Primers used in this study.

Click here for additional data file.

Table S2 RNA‐seq data.

Click here for additional data file.

Table S3 RNA‐seq data of FR‐treated wild‐type and Mpkan plants.

Click here for additional data file.

Table S4 GO enrichment analysis.

Please note: Wiley Blackwell 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.

Click here for additional data file.