An Evolutionarily Conserved Coreceptor Gene Is Essential for CLAVATA Signaling in Marchantia polymorpha

Growth and development of land plants are controlled by CLAVATA3/EMBRYO SURROUNDING REGION-related (CLE) family of peptide hormones. In contrast to the genetic diversity of CLE family in flowering plants, the liverwort Marchantia polymorpha possesses a minimal set of CLE, MpCLE1(TDIF homolog), and MpCLE2 (CLV3 homolog). MpCLE1 and MpCLE2 peptides exert distinct function at the apical meristem of M. polymorpha gametophyte via specific receptors, MpTDIF RECEPTOR (MpTDR) and MpCLAVATA1 (MpCLV1), respectively, both belonging to the subclass XI of leucine-rich repeat receptor-like kinases (LRR-RLKs). Biochemical and genetic studies in Arabidopsis have shown that TDR/PXY family and CLV1/BAM family recognize the CLE peptide ligand in a heterodimeric complex with a member of subclass-II coreceptors. Here we show that three LRR-RLK genes of M. polymorpha are classified into subclass II, representing three distinct subgroups evolutionarily conserved in land plants. To address the involvement of subclass-II coreceptors in M. polymorpha CLE signaling, we performed molecular genetic analysis on one of them, MpCLAVATA3 INSENSITIVE RECEPTOR KINASE (MpCIK). Two knockout alleles for MpCIK formed narrow apical meristems marked by promMpYUC2:GUS marker, which were not expanded by MpCLE2 peptide treatment, phenocopying Mpclv1. Loss of sensitivity to MpCLE2 peptide was also observed in gemma cup formation in both Mpclv1 and Mpcik. Biochemical analysis using a Nicotiana benthamiana transient expression system revealed weak association between MpCIK and MpCLV1, as well as MpCIK and MpTDR. While MpCIK may also participate in MpCLE1 signaling, our data show that the conserved CLV3-CLV1-CIK module functions in M. polymorpha, controlling meristem activity for development and organ formation for asexual reproduction.


INTRODUCTION
CLAVATA3/EMBRYO SURROUNDING REGION-related (CLE) peptides are a family of peptide hormones in land plants, mediating cell-to-cell communication in the plant body (Murphy et al., 2012;Hirakawa and Sawa, 2019;Fletcher, 2020). CLE peptides are genetically encoded as a precursor protein possessing a conserved CLE domain(s) at or near the C-terminus. Biosynthesis of CLE peptide hormone from the CLE domain involves post-translational events including proteolytic cleavage, posttranslation modifications and secretion to the apoplast (Ito et al., 2006;Kondo et al., 2006;Ohyama et al., 2009;Tamaki et al., 2013: Matsubayashi, 2014. In flowering plants, a large number of CLE genes are encoded in the genome, which have been extensively studied for the past two decades (Cock and McCormick, 2001;Oelkers et al., 2008;Jun et al., 2010;Fletcher, 2020). The function of CLE genes cover a wide range of physiological processes including stem cell homeostasis in meristems, vascular cell differentiation, stomata differentiation and responses to various environmental cues (Fletcher et al., 1999;Suzaki et al., 2008;Okamoto et al., 2009;Stahl et al., 2009;Etchells and Turner, 2010;Hirakawa et al., 2010;Mortier et al., 2010;Kondo et al., 2011;Fiume and Fletcher, 2012;Depuydt et al., 2013;Endo et al., 2013;Araya et al., 2014;Czyzewicz et al., 2015;Gutiérrez-Alanís et al., 2017;Rodríguez-Leal et al., 2017;Qian et al., 2018;Takahashi et al., 2018;Ma et al., 2020). In bryophytes, which are distantly related to flowering plants in the land plant lineage Puttick et al., 2018), relatively low number of CLE genes are encoded in the genome, providing simplified models to study the function of CLE genes (Bowman et al., 2017;Whitewoods et al., 2018). The minimal set of CLE genes, MpCLE1 (Mp6g07050) and MpCLE2 (Mp5g18050), are encoded in the genome of the liverwort Marchantia polymorpha (Bowman et al., 2017;Montgomery et al., 2020; Figure 1A). MpCLE1 and MpCLE2 are the orthologs of TDIF (tracheary element differentiation inhibitor factor) and CLV3 (CLAVATA3) of Arabidopsis thaliana, respectively, representing the two distinct subgroups of CLE peptide family. In Arabidopsis, specific bioactivities of TDIF and CLV3 are attributed to the difference in a few amino acids between them, which are mediated by two distinct groups of receptors, TDIF RECEPTOR/PHLOEM INTERCALTED WITH XYLEM (TDR/PXY) and CLAVATA1/BARELY ANY MERISTEMs (CLV1/BAMs), respectively (Fletcher et al., 1999;DeYoung et al., 2006;Fisher and Turner, 2007;Hirakawa et al., 2008Hirakawa et al., , 2017Ogawa et al., 2008;Rodriguez-Villalon et al., 2014;Shimizu et al., 2015;Shinohara and Matsubayashi, 2015;Crook et al., 2020). Since the ligand-receptor pairs are conserved among flowering plants and bryophytes and no CLE homologs were found in sister streptophyte algae, the specific CLE peptidereceptor pairs may have originated in the common ancestor of land plants (Whitewoods et al., 2018;Hirakawa et al., , 2020. In M. polymorpha, CLE genes regulate the activity of the apical meristem located at the apical notch of the thalloid gametophyte body. MpCLE1-MpTDR signaling acts as a negative regulator of cell proliferation at the apical notch, while MpCLE2-MpCLV1 signaling functions as a positive regulator of stem cell activity in the apical notch (Hirakawa et al., , 2020. Both TDR/PXY and CLV1/BAM belong to the subclass XI of leucine-rich repeat receptor-like kinase (LRR-RLK) family. In addition to CLE peptides, a number of peptide ligands have been shown to bind to specific members of subclass-XI receptors, which possess a long extracellular domain (ECD) composed of more than 20 LRRs (Shiu and Bleecker, 2001;Yamaguchi et al., 2006;Hou et al., 2014;Tabata et al., 2014;Ou et al., 2016;Shinohara et al., 2016;Song et al., 2016;Doblas et al., 2017;Nakayama et al., 2017;Toyokura et al., 2019;Doll et al., 2020). Accumulating evidence indicates that subclass-II receptors, such as SOMATIC EMBRYOGENESIS RECEPTOR KINASE/BRASSINOSTEROID INSENSITIVE1-ASSOCIATED KINASE1 (SERK/BAK1) family, participate in the peptide hormone perception by forming a heterodimeric complex with subclass-XI receptors (Hohmann et al., 2017;Gou and Li, 2020). Structural studies have revealed that SERK coreceptors have a short ECD containing five LRRs. The ECD of subclass-II receptors do not interact strongly or at all to the peptide ligand by themselves and rather recognize the ligand-receptor complex (Santiago et al., 2013(Santiago et al., , 2016Sun et al., 2013;Wang et al., 2015;Okuda et al., 2020). In line with this scheme, PXY/TDR and SERK2 are reported to form a heterodimeric complex for TDIF recognition, and multiple knockout mutants for Arabidopsis SERK genes show reduced TDIF sensitivity in vascular development (Morita et al., 2016;Zhang et al., 2016a,b).
In contrast to TDIF, involvement of SERK family has not been observed in CLV3-type CLEs. Instead, another group of subclass-II receptors, CLAVATA3 INSENSITIVE RECEPTOR KINASEs (CIKs), have been implicated in CLV3 peptide perception. CIK proteins can form protein complexes with CLV1/BAM receptors Hu et al., 2018). Quadruple mutants for Arabidopsis CIK1-4 genes develop enlarged shoot apical meristems, which is similar to those of clv mutants. The growth from the enlarged meristems is not arrested by treatment with CLV3 peptide, a negative regulator of stem cells in Arabidopsis . Furthermore, full activity of Arabidopsis CLE26/CLE45 peptides in root phloem cell differentiation requires CLE-RESISTANT RECEPTOR KINASE (CLERK)/CIK2 although biochemical interaction is not detected between the ECDs of CLERK and the subclass-XI receptor BAM3 (Anne et al., 2018). In this study, we searched for the homologs of CIK genes in M. polymorpha and analyzed their involvement in CLE peptide signaling by molecular genetic approach.

A Single CIK Ortholog in Marchantia polymorpha
In the M. polymorpha genome, three LRR-RLK genes (Mp7g09160/Mapoly0068s0069, Mp7g14210/Mapoly0009s0106, and Mp7g15980/Mapoly0560s0001) have been classified into subclass II (Sasaki et al., 2007;Bowman et al., 2017;Montgomery et al., 2020). To better understand the evolutionary relationships, we performed phylogenetic analysis of the subclass-II genes from land plants (A. thaliana, Amborella trichopoda, Picea abies, Selaginella moellendorffii, Physcomitrium patens, Sphagnum fallax, M. polymorpha) and charophycean algae (Spirogyra pratensis and Coleochaete orbicularis) based on the amino acid sequence of the kinase domain using a Bayesian method ( Figure 1B). The tree inferred three subgroups diverged in the land plant lineage, each of which contains a single M. polymorpha gene. Mp7g14210, designated as MpCIK, was grouped into a single subgroup with all CIK genes from Arabidopsis. Likewise, Mp7g09160/MpSERK was grouped into the SERK subgroup with all Arabidopsis SERK genes. Thus, M. polymorpha genome may lack redundancy in CLE ligand/receptor/coreceptor orthlogs ( Figure 1A). In transcriptome of M. polymorpha, MpCIK was expressed in thalli, gametangiophores and sporophytes while its expression was relatively low in sporelings. MpCLV1 expression showed a similar trend (Bowman et al., 2017).

CRISPR-Cas9 Editing of MpCIK Does Not Affect Overall Growth of Gametophyte
To analyze the physiological function of the CIK coreceptor gene in M. polymorpha, we generated loss-of-function alleles for MpCIK using CRISPR-Cas9 editing (Sugano et al., 2018). Sanger sequencing revealed that two independent transgenic lines, Mpcik-1 ge and Mpcik-2 ge , possess different mutations at the CRISPR/Cas9 target site, both predicted to result in gene knockout due to premature termination of translation ( Figure 1C). We could not find significant differences in the overall morphology of thalli in 10-day-old Mpcik-1 ge and Mpcik-2 ge plants grown from gemmae, compared to any of wildtype (Tak-1), Mpcle2-2 ge and Mpclv1-3 ge genotypes ( Figure 1D).

MpCIK Is Necessary for MpCLE2 Signaling to Control Apical Notch Expansion
To analyze the involvement of MpCIK in MpCLE2 peptide signaling, we examined the apical notch morphology in 4-day-old gemmalings grown on liquid M51C medium supplemented with or without 3 µM MpCLE2 peptide. In wild-type gemmalings, apical notches were expanded by treatment with MpCLE2 peptide, as reported previously (Figures 2A,B; Hirakawa et al., 2020). By contrast, apical notches in both Mpcik-1 and Mpcik-2 were insensitive to MpCLE2 peptide, which is similar to those in Mpclv1-3 (Figures 2A,B). Importantly, apical notches of the Mpcik and Mpclv1 alleles were narrower than those of wild type in the growth without MpCLE2 peptide. Consistently, Mpcle2-2 ge developed narrow apical notches but it was sensitive to the Frontiers in Plant Science | www.frontiersin.org treatment with the MpCLE2 peptide as reported previously ( Figure 2B; Hirakawa et al., 2020). pro MpYUC2(YUCCA2):GUS is a marker for the tip of apical notch, and pro MpYUC2:GUSpositive (MpYUC2 + ) region is affected by MpCLE2-MpCLV1 signaling Hirakawa et al., 2020). Compared to wild-type (Tak-1) background, MpYUC2 + region was reduced in Mpcik backgrounds, phenocopying Mpclv1-3 ( Figure 2C). Confocal imaging in 2-day-old gemmalings showed the number of apical and subapical cells, was reduced in Mpcik and Mpclv1 alleles (Figures 2D,E). These data support that MpCIK is an essential component of the MpCLE2 peptide perception.

MpCLE2 Signaling Affects Thallus Branching and Gemma Cup Formation via MpCLV1 and MpCIK
To further address the involvement of MpCIK in MpCLE2 signaling, we used pro MpYUC2:MpCLE2, a gain of function allele of MpCLE2, which stably develops supernumerary branching from the expanded apical meristems (Hirakawa et al., 2020). We generated Mpcik and Mpclv1 knockout alleles in pro MpYUC2:MpCLE2-2 background by CRISPR/Cas9-mediated genome editing (Supplementary Figure 1). Both Mpcik-3 ge and Mpclv1-4 ge suppressed the supernumerary branching phenotype in 16-day-old plants ( Figure 3A), which is consistent with the results in peptide treatment assay ( Figure 2B). Furthermore, the number of gemma cup formed on thalli was reduced in pro MpYUC2:MpCLE2 and it was suppressed in both Mpcik-3 ge and Mpclv1-4 ge ( Figure 3A). Time-course analysis showed that the gemma cup formation was significantly reduced and delayed in pro MpYUC2:MpCLE2 plants compared to wild type ( Figure 3B). Meanwhile, all Mpcik and Mpclv1 alleles showed minor increase in gemma cup formation compared to wild type ( Figure 3B).

Biochemical Interaction of MpCIK and MpCLV1 Proteins
Since the genetic analysis suggests that MpCIK may function as a coreceptor for MpCLV1 for the perception of MpCLE2 peptide, we examined the biochemical interaction between MpCIK and MpCLV1 proteins expressed in a Nicotiana benthamiana transient expression system, which has been utilized to analyze the interaction of CLV and CIK receptors of Arabidopsis (Kinoshita et al., 2010;Betsuyaku et al., 2011;Hu et al., 2018). MpCIK, MpCLV1 and MpTDR were expressed under the control of 35S promoter in N. benthamiana as proteins C-terminally fused to 3 × HAs-single StrepII or 3 × FLAG (MpCIK-3HS, MpCLV1-3FLAG, MpTDR-3FLAG), respectively ( Figure 4A). MpTDR, a receptor for MpCLE1, was also included in this interaction assay . In co-immunoprecipitation experiments using an anti-HA affinity matrix, MpCLV1-3FLAG was detected not strongly but reproducibly in the immunoprecipitates containing MpCIK-3HS, suggesting a weak or transient interaction between MpCIK-3HS and MpCLV1-3FLAG ( Figure 4B). Similarly, MpTDR-3FLAG was also shown to associate weakly with  MpCIK-3HS ( Figure 4B). Thus, MpCLV1 and MpTDR are capable of interacting with MpCIK.

Mpcik Is Sensitive to MpCLE1 and TDIF
Our biochemical data suggests that MpCIK could also function as a coreceptor for MpTDR upon the perception of MpCLE1, a TDIF-type CLE peptide. Previously, we showed that synthetic TDIF-type peptides cause slight reduction of overall growth and twisted lobes in M. polymorpha thalli . In order to address the possible involvement of MpCIK in MpCLE1 signaling, we first examined the effects of TDIF, the strongest analog among known TDIF-type CLE peptides including MpCLE1 peptide. In 14-day-old plants grown from gemmae on growth medium supplemented with 3 µM TDIF, both Tak-1 and Mpcik-1 showed slight reduction of overall growth and twist in the thallus lobes (Figure 5A), indicating that MpCIK is not necessary for the perception of TDIF. We further analyzed the effects of MpCLE1 overexpression using pro MpYUC2:MpCLE1 transformants. In the wild-type (Tak-1) background, pro MpYUC2:MpCLE1 resulted in small, twisted thalli in 14-day-old plants. This phenotype was not observed in pro MpYUC2:MpCLE1 1−417 , a truncated version of MpCLE1 lacking an essential asparagine residue in the CLE peptide motif (Figure 5B; . In the Mpcik-1 ge background, pro MpYUC2:MpCLE1 also resulted in small twisted thalli although the effects on growth was mild compared to the Tak-1 background, as judged from the ground cover area in 14-day-old plants (Figures 5B,C). These data suggests that MpCIK could be partially involved in MpCLE1 perception.

DISCUSSION
In this study, we performed functional analysis of MpCIK, the sole M. polymorpha ortholog of Arabidopsis CIK genes, and showed that MpCIK is essential for MpCLE2 peptide signaling to regulate the apical meristem activity in gametophyte. Biochemical analysis in N. benthamiana supports the idea that the MpCLV1-MpCIK is a receptor-coreceptor pair for MpCLE2 peptide perception. Since the same ligand-receptor-coreceptor relationship of CLV3-CLV1-CIK is consistently observed for multiple paralogs in Arabidopsis (Anne et al., 2018;Cui et al., 2018;Hu et al., 2018), this system is likely an evolutionarily conserved mechanism for the perception of CLV3-type CLE peptide in land plants. In addition, we suggested partial involvement of MpCIK in signaling of TDIF-type CLE peptide, MpCLE1, although MpCIK is not necessary for MpCLE1 signaling. Studies in Arabidopsis have suggested that subclass-II receptors in the SERK subgroup function as coreceptors for TDIF-PXY/TDR signaling (Zhang et al., 2016b). Further studies on MpSERK would clarify the contribution of different coreceptors. The phylogenetic analysis inferred the divergence of CIK and SERK subgroups in the common ancestor of land plants, which coincides with the appearance of subclass-XI genes. Recent studies have also shown that SERK-interacting receptors from subclass X, such as BRI1/BRL and EMS1, are also encoded in bryophytes (Ferreira-Guerra et al., 2020;Furumizu and Sawa, 2021). Interestingly, the two sequences from S. pratensis ( Figure 1B) showed high similarity to the subclass-II genes from land plants. Studies on Zygnematales algae would provide a clue to understand the evolution of these receptors.
Gemma cups are specialized structures for vegetative propagation, found in certain species of Marchantiopsida (Yasui et al., 2019;Kato et al., 2020). Gemma cup formation initiates at the cells in the dorsal epidermis close to the apical cells (Suzuki et al., 2020). We show that gain-of-function of MpCLE2 results in the delay of gemma cup formation. Loss-of-function phenotypes support that the intrinsic MpCLE2-MpCLV1-MpCIK signaling module functions as a negative regulator of gemma cup formation. Although it is still unclear if the phenotypes in gemma cup formation can be uncoupled from the defects in the apical/subapical cells, MpCLE2 likely regulates cell fates in both the lateral derivatives of the apical cells and the dorsal derivatives of apical/subapical cells. It is known that hormonal and environmental cues affect the formation of gemma cup Aki et al., 2019;Li et al., 2020;Rico-Reséndiz et al., 2020). A possible role for MpCLE2 peptide signaling would be to mediate certain environmental cues to control the timing of gemma cup formation and thus clonal propagation, cooperatively with other hormonal inputs. In addition, involvement of CIK subgroup members into antiviral responses has been suggested in Arabidopsis (Fontes et al., 2004). Further studies of Mpcik knockout plants under various environmental conditions would provide a new insight into signals that allowed plants to survive on land.
Our biochemical data reveals that MpCIK is capable of associating with MpCLV1 or MpTDR in an ectopic and transient expression system of N. benthamiana. Furthermore, weak associations observed for MpCIK-MpCLV1 as well as MpCIK-MpTDR indicates possible requirement of other components in MpCIK-containing complex formation. For instance, in a ligand-induced dimerization model, receptorcoreceptor interaction can be induced by the perception of ligand at their ectodomains, which in turn allows for their kinase domains to transphosporylate and activate signaling (Jaillais et al., 2011;Hohmann et al., 2018;Perraki et al., 2018). Thus, ligand and/or other membrane receptors could be required for strong MpCIK-MpCLV1/MpTDR association. With its genetic simplicity in CLE signaling, M. polymorpha will be a nice experimental system to address this point in future studies.

Phylogenetic Analysis
Gene sequences of land plants were retrieved from Phytozome v12.1 database 1 except for those of Arabidopsis thaliana 2 , Picea abies 3 , and Marchantia polymorpha 4 . Sequences of charophycean algae were reported in Bowman et al. (2017), obtained from transcriptome databases for Spirogyra pratensis 5 and Coleochaete orbicularis 6 . Gene IDs and the protein sequences are listed in Supplementary Table 1. Predicted protein sequences were aligned in Clustal W 7 . We excluded ambiguously aligned sequence to produce an alignment of 297 amino acid characters in the conserved cytosolic domain. Bayesian analysis was performed using MrBayes 3.2.7 (Ronquist et al., 2012). Two runs with four chains of Markov chain Monte Carlo (MCMC) iterations were performed for 1,500,000 generations, keeping one tree every 100 generations. The first 25% of the generations were discarded as burn-in and the remaining trees were used to calculate a 50% majority-rule tree. The standard deviation for the two MCMC iteration runs was below 0.01, suggesting that it was sufficient for the convergens of the two runs. Convergence was assessed by visual inspection of the plot of the log likelihood scores of the two runs calculated by MrBayes (Gelman and Rubin, 1992). Character matrix used for the Bayesian phylogenetic analysis is provided in Supplementary Data Sheet 1.

Plant Materials and Growth Conditions
Marchantia polymorpha male Takaragaike-1 (Tak-1) accession was used as wild type in this study. M. polymorpha plants were grown on half-strength Gamborg B5 medium (pH 5.5) solidified with 1.4% agar at 22 • C under continuous white light. N. benthamiana seeds were grown on BM2 soil (Berger) in a growth room at 23 • C under continuous LED light.

Constructs
Primers and plasmids are listed in Supplementary Tables 2, 3. All plant transformation vectors were generated using the Gateway cloning system (Thermo Fisher Scientific, Waltham, MA, United States). Gateway destination vectors are described in Kinoshita et al. (2010); Ishizaki et al. (2015), and Sugano et al. (2018), except for pMpGWB301-YUC2p, which was generated in this study. A 3,032 bp DNA flagment of MpYUC2 promoter sequence franking the translation initiation site was PCR amplified from pENTR-proMpYUC2 vector (Hirakawa et al., 2020) with a primer pair of MpYUC2prom3k_F_InFusion_XbaI and MpYUC2prom_R_InFusion_XbaI, and cloned into the XbaI digestion site of pMpGWB301 using In-Fusion HD Cloning Kit (Takara Bio, Shiga, Japan). For construction of pro MpYUC2:MpCLE1, entry clones, pENTR-MpCLE1 and pENTR-MpCLE(1-417) , were transferred to the pMpGWB301-YUC2p vector using Gateway LR Clonase II Enzyme mix (Thermo Fisher Scientific). For genome editing of MpCIK, a guide RNA was designed at the first exon/intron junction of Mp7g14210 using CRISPRdirect 8 (Naito et al., 2015) and the plasmid for genome editing was constructed according to Sugano et al. (2018). For the expression of epitope-tagged receptors in N. benthamiana, coding sequences of MpCIK, MpCLV1, and MpTDR were PCR amplified from M. polymopha cDNA and cloned into pENTR/D-TOPO vector. Resultant entry clones (pENTR-MpCIK, pENTR-MpCLV1, and pENTR-MpTDR) were transferred to pXCSG-3FLAG or pXCSG-3HS vector using Gateway LR Clonase II Enzyme mix (Thermo Fisher Scientific).

Production of Transgenic Marchantia polymorpha
Transgenic M. polymorpha plants are listed in Supplementary  Table 3.
Agrobacterium-mediated transformation of M. polymorpha was performed using regenerating thalli according to Kubota et al. (2013). CRISPR/Cas9-based genome editing was performed according to Sugano et al. (2018) and mutations in the guide RNA target loci were examined by direct sequencing of PCR product amplified from genome DNA samples with primers listed in Supplementary Table 2. Genome editing of MpCLV1 was performed as described previously (Hirakawa et al., 2020). Nomenclature of genes and mutants are according to Bowman et al. (2016).

Plant Imaging and Phenotypic Measurement
Overall morphology of plants was observed under a digital microscope (DMS1000, Leica Microsystems, Wetzlar, Germany) or under a digital camera (TG-6, Olympus, Tokyo, Japan). For the quantification of ground cover area in plant images, blue color was extracted and quantified using ImageJ (Schneider et al., 2012). For the measurement of apical notch width, plants grown on liquid medium were individually transferred onto agar medium and imaged under a digital microscope (DMS 1000, Leica Microsystems). To quantify the apical notch width, distance between the rims of apical notch was measured on the obtained images using ImageJ (Schneider et al., 2012). Confocal imaging of apical notch was performed as described previously (Hirakawa et al., 2020). Briefly, 2day-old gemmalings were fixed by vacuum infiltration in 4% paraformaldehyde in phosphate buffer (Nacalai Tesque, Kyoto, Japan). After fixation, samples were cleared with ClearSee solution (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) and were stained for an hour with SCRI Renaissance 2200 (Renaissance Chemicals, Selby, United Kingdom) in ClearSee solution. Samples were observed under a confocal laser scanning microscopy (Fluoview FV1000, Olympus) using 405-nm excitation. Z-series images were collected at 0.5 µm intervals through the specimens and obtained images were processed using Fiji software to specify apical and subapical cells (Schindelin et al., 2012).

Transient Expression in Nicotiana benthamiana
Agrobacterium tumefaciens strains GV3101 MP90RK carrying expression constructs were grown in YEB medium with appropriate antibiotics, harvested by centrifugation at 4,500 rpm for 10 min, and resuspended in infiltration buffer [10 mM MES (pH 5.7), 10 mM MgCl2, 150 µM acetosyringone]. The cultures were adjusted to an OD 600 of 1.0 and incubated at room temperature for at least 3 h prior to infiltration. Equal volumes of cultures of different constructs were mixed for co-infiltration, and then mixed with agrobacterial cultures (OD 600 of 1.0) carrying the p19 silencing suppressor in a 1:1 ratio (Voinnet et al., 1999). The resulting cultures were infiltrated into leaves of 3-to 4-week-old N. benthamiana. The leaf samples were harvested 3 days after infiltration for subsequent protein extraction (Betsuyaku et al., 2011).

Protein Extraction
Total protein was extracted from the infiltrated N. benthamiana leaves with IP extraction buffer (1:1 w/v, 50 mM Tris-HCl pH 8.0, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 × Proteinase inhibitor cocktail SIGMA P9599 and 1 mM EDTA) and incubate the extract at 4 • C for 30 min. The lysates were centrifuged at 20,000 × g for 20 min at 4 • C and the supernatants were then centrifuged again at 20,000 × g for 5 min at 4 • C. The resultant supernatants were used as total protein lysates.

Co-immunoprecipitation
For immunoprecipitation, 1 ml of the lysates prepared with IP extraction buffer from 0.5 g of leaves was incubated with anti-HA Affinity Matrix (Roche 11815016) for o/n in a rotary shaker at 4 • C. The beads were collected and washed three times with 1 ml of the extraction buffer. Immunoprecipitated proteins were eluted from the beads by boiling in SDS sample buffer at 95 • C and analyzed by Western blot using the corresponding antibodies. We used the following antibodies; Anti-HA-Peroxidase High Affinity (3F10) (Roche 12013819001) and Monoclonal ANTI-FLAG M2-Peroxidase (HRP) (SIGMA A8592).

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
YH conceived the study, designed the work with input from all authors, and prepared the manuscript draft. GT, SB, NO, and YH performed the experiments and analyzed the data. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by grants from the Japan Science and Technology Agency ERATO (JPMJER1502 to SB) and JSPS KAKENHI (JP19K06727 to YH).

ACKNOWLEDGMENTS
We thank colleagues in the peptide/protein center of WPI-ITbM for peptide synthesis, Eriko Betsuyaku and Ikuko Nakanomyo for technical assistance.