The widely used Marchantia polymorpha Tak-1 (MpTak-1) ecotype was cultivated via propagation of vegetative propagules (gemmae) on solid Johnson’s medium (Ishizaki et al., 2008) supplemented with 0.8% micro agar under axenic conditions. Cultivation petri dishes were sealed using Micropore tape to ensure gas exchange while preventing microbe contamination. Gemmae were grown under long day condition (16 h light/8 h darkness cycle) and white light irradiation (60 μmol m^–1 s^–1) at 21°C and 60% humidity. After 2.5–3 weeks, a few thallus fragments of approximately 0.5–1 cm² were transferred onto a small petri dish (6–9 cm in diameter) containing fresh solid Johnson’s medium on the day of transformation (Figure 1A).

The Arabidopsis thaliana Col-0 ecotype used for DAPI staining was cultivated on soil and grown under long day conditions at 21°C and 120 μmol m^–1 s^–1 light intensity.

All constructs used in this study are summarized in Supplementary Table S1, including their origin, promoter, fluorescent tag and oligonucleotide sequences used for PCR-based amplification of new constructs from Marchantia whole-thallus cDNA. The 35S promoter was used for all expression experiments (except for expression of AtSYP32, AtGot1p and LifeAct) to guarantee comparability of subsequent analyses. The coding sequences of interest were cloned into Gateway (GW)-compatible entry vectors, pDONR201 and pDONR207 (Invitrogen), and then remobilized to be integrated in the respective GW destination vectors (Supplementary Table S1). The cloning procedure was as described before (Westermann et al., 2019).

For a single shot, 300 ng of vector DNA were mixed with gold, serving as micro-carriers (30 mg/ml, 1 μm), CaCl₂ (2.5 M), spermidine (0.1 M) and _ddH₂O under thorough shaking. Subsequently, micro-carriers were washed with 70% EtOH and 100% EtOH. The DNA-coated gold particles were suspended in 100% EtOH and placed onto macro-carriers. The EtOH was allowed to vaporize and the prepared macro-carriers were then used for biolistic transformation.

Marchantia thallus fragments were placed into a PDS-1000/He Biolistic^® Particle Delivery System (Bio-Rad). A vacuum of 25 in Hg vac was applied and the DNA-coated gold particles were shot at 900 psi from a distance of 10 cm. Finally, the bombarded plant material was allowed to recover for 24 h in darkness while remaining in its humid environment, i.e., on the media in the closed petri dish (Figure 1C). Biolistic transformation generally yielded n > 50 transformed cells per sample shot. A representative example for transformation efficiency is shown in Figure 1B. Moderate to strong expression levels in each individual cell could be observed irrespectively of the protein construct or promoter used (Figure 1B and Supplementary Table S1). The use of strong promoters such as pro35S, proAtUBQ10 or proMpEF1α can sometimes lead to overexpression artifacts that may impede drawing secured conclusions. However, yielding a wide range of expression level in the same experimental round and plant sample allows for identification of biologically meaningful protein localization patterns and to distinguish them from unwanted artifacts, such as protein over-accumulation. In order to reliably assess the potential of transformed constructs as single cell fluorescent markers, we co-bombarded all described vectors with either the nuclear marker AtKRP1 or the plasma membrane markers AtNPSN12 or MpSYP13a fused to fluorescent tags and subsequently created a collection of functional and useful Arabidopsis- and Marchantia-derived fluorescent protein fusions (Supplementary Table S1). In order to determine the efficiency of co-transformation, we counted cells expressing both markers in relation to the total number of transformed cells in nine independent co-transformations of protein fusions used in this study. Successful biolistic co-transformation reached on average 74% (see Supplementary Table S3).

For fluorescein diacetate (FDA) staining, young (2- to 5-day-old) gemmae were placed onto depression slides and covered with an FDA solution (5 mg/L FDA in _ddH₂O, diluted from a stock solution of 5 mg/ml FDA in acetone) for 5–10 min. Afterward, samples were rinsed in _ddH₂O.

For PI staining, young gemmae were placed onto depression slides and directly covered with a PI solution for 10 min (10 mg/L in _ddH₂O). Subsequently, samples were rinsed with _ddH₂O.

For FM4-64 staining, young gemmae were mounted onto depression slides in 2 μM FM4-64 diluted in liquid Johnson’s growth medium (Ishizaki et al., 2008) and allowed to incubate for 10 min prior to imaging. For FM4-64 and FDA co-staining, gemmae were first stained in a FDA solution and then mounted in a FM4-64 solution, both as described above. Marchantia thallus fragments, transiently transformed with eYFP-MpRAB5 or MpARA6-eYFP, were stained the day after particle bombardment prior to imaging as described above.

For DAPI staining, several methods were used. Experiments were done using 0-, 4-, and 7-day-old gemmae. The DAPI staining solutions were composed of 10–100 mg/L DAPI in either 1xPBS-T (phosphate buffered saline + 0.1% Tween-20) and 5% DMSO or _ddH₂O with 0.1 or 1% Tween-20 and 5% DMSO. Different staining incubation times of 10, 30, or 60 min were tested. The staining was tested with and without preceding or subsequent shaking of the samples in 70% EtOH at 80°C. To enhance permeability of membranes, 10 or 50 mg/L digitonin was added to the aforementioned staining solutions. As all attempts for staining living cells failed, the following fixation methods were tested. Gemmae were fixed in a 3:1 EtOH:acetic acid mixture on ice for 1 h, washed three times in 100% EtOH and stained in aforementioned DAPI solutions for 1 h. In another attempt, gemmae were fixed in 3% glutaraldehyde in 1× PBS-T (phosphate buffered saline + 0.1% Tween-20) overnight, subsequently washed in 1× PBS-T, and incubated in aforementioned DAPI solutions in darkness overnight. Furthermore, a modified version of a DAPI staining protocol published for gametophore leaflets and protonemata of Physcomitrella patens (Sato et al., 2017) was used. Gemmae were placed in 3.7% formaldehyde in 1× PBS for 30 min. Subsequently gemmae were immersed in 100% MeOH on ice for 10 min. Afterward, gemmae were soaked in 1% Triton X-100 and then stained with the aforementioned DAPI solutions for 30 min. Unfortunately, none of the experimental procedures described here led to a reliable staining of nuclei by DAPI in viable or fixed epidermal cells of M. polymorpha gemmae.

The transformed or stained plant material was transferred onto a depression slide supplemented with 300 μL _ddH₂O and covered with a 18 × 18 mm cover slip. Rhizoid growth experiments were performed using young gemmae mounted with Johnson’s growth medium instead of _ddH₂O. Microscopic analysis was performed using a Leica SP8 CLSM with an argon gas laser intensity set at 20%. Fluorophore excitation and fluorescence caption were performed at the wavelength spectra shown in Supplementary Table S2. Images were taken using a digital gain of 100% at a resolution of 1024 × 1024 pixels, a pinhole size of 1 AU, and a scan speed of 400–700 Hz using bidirectional confocal scanning and hybrid detectors (HyD). For the caption of multiple fluorophore types sequential or, if suitable, simultaneous scanning was performed. Usage of a laser scanning confocal system is strongly recommended for image capture, as it allows for scanning on multiple focal planes to perform maximum projection, while reducing unspecific background noises (as compared to epifluorescence microscopy).

Analysis of all microscopic captions was performed using ImageJ/FIJI (Schindelin et al., 2012), software version 1.51n. Data manipulation included maximum projections from Z-stacks (≤20 frames, 1 μm slice intervals) for some of the markers (as individually mentioned in the figure captions), as well as generation of composite images from separate individual channels.

We first picked the Arabidopsis thaliana INHIBITOR OF CYCLIN-DEPENDENT KINASE 1 (AtICK1)/KIP-RELATED PROTEIN 1 (AtKRP1), which localizes to the nucleus and functions in cell growth, differentiation, and cell cycle progression (Wang et al., 1998; De Veylder et al., 2001; Schnittger et al., 2003; Weinl et al., 2005; Jakoby et al., 2006). Upon biolistic transformation of Marchantia thalli, we observed AtKRP1-eCFP protein localization to the nucleus of epidermal cells (Figure 2A). We therefore co-transformed AtKRP1 as a nuclear marker and indicator of successful cell transformation in subsequent experiments.

As a second potential marker, we chose the Arabidopsis thaliana NOVEL PLANT SNARE 12 (AtNPSN12), which represents a non-polar plasma membrane-localized protein commonly used as plasma membrane marker (Alassimone et al., 2012; Kirchhelle et al., 2016). Biolistically transformed Marchantia thallus epidermal cells showed AtNPSN12-mCherry fluorescence at the cell periphery consistent with plasma membrane localization (Figure 2B). To confirm this localization, we co-transformed AtNPSN12-mCherry with the known Marchantia plasma membrane marker mCitrine-MpSYP13a (Kanazawa et al., 2016) (Figure 2C). As single and co-bombardments with AtNPSN12 showed (co)localization to the plasma membrane, we conclude that AtNPSN12-mCherry and mCitrine-MpSYP13a are both suitable plasma membrane markers for Marchantia epidermal cells (Figure 2D).

Receptor-like kinases of the Malectin-like receptor (MLR) subfamily have been the subject of intensive research in the past years given their multitude of functions in plant development and immunity signaling (Franck et al., 2018a). The plasma membrane localized MLRs ANXUR1 and 2 (AtANX1/2) control cell wall integrity during pollen tube growth (Boisson-Dernier et al., 2009; Miyazaki et al., 2009) and negatively regulate plant immune responses in Arabidopsis (Mang et al., 2017). During pollen tube growth control, AtANX1/2 act genetically upstream of the cytosolic and plasma membrane-attached receptor-like cytoplasmic kinase of the PTI1-like family, AtMRI, while the AtANX1 homolog AtFERONIA (AtFER) acts upstream of AtMRI during root hair growth control (Boisson-Dernier et al., 2015). We showed recently that tip-growth control in Marchantia rhizoids relies on an evolutionarily conserved signaling module comprised of the unique Marchantia MLR MpFER and its downstream component and unique Marchantia PTI1-like MpMRI (Westermann et al., 2019). We transiently co-expressed the fluorescent protein fusions AtMRI-YFP, MpFER-YFP, and MpMRI-YFP with AtNPSN12-mCherry. While MpFER-YFP showed signal exclusive to the plasma membrane, AtMRI and MpMRI displayed plasma membrane localization with traces in the cytoplasm as reported before (Figure 3) (Boisson-Dernier et al., 2015; Westermann et al., 2019).

Noteworthily, we also wanted to test expressing the plasma membrane localized Arabidopsis MLRs in Marchantia and thus co-transformed AtANX1-RFP with mCitrine-MpSYP13a and AtFER-Citrine with AtNPSN12-mCherry. Intriguingly, while many cells expressed the plasma membrane markers mCitrine-MpSYP13a and AtNPSN12-mCherry, a great majority of them did not show expression of either AtANX1 or AtFER (Supplementary Figures S1A,B). This suggests that, unlike MpFER, the Arabidopsis MLRs fused to single fluorescent tag cannot be expressed in Marchantia epidermal cells. Thus, we next tried to express AtFER with a triple Citrine tag instead of a single one. It resulted in many Citrine-expressing cells but mostly in the cytoplasm, with no hints of plasma membrane localization (Supplementary Figure S1C). These results indicate that fusion of long protein tags may prevent transmembrane receptor kinases such as MLRs to be correctly integrated into cellular membranes. To check if this was specifically due to Arabidopsis proteins or to certain protein families, we co-expressed MpFER-3xCitrine with MpFER-TdTomato and MpMRI-3xCitrine with MpMRI-RFP (Supplementary Figures S1D,E). Interestingly, the 3xCitrine tag did not perturbate the cytosolic and plasma membrane localization of MpMRI, as MpMRI-3xCitrine co-localized with MpMRI-RFP at the cell periphery. However, while MpFER-TdTomato exhibited PM localization, MpFER-3xCitrine-derived signal was clearly present in the cytoplasm. Therefore, for some plasma membrane-localized protein families, fusion with a triple tag can lead to localization artifacts, and the use of single tag is thus recommended by default. Why MpFER but neither AtFER nor AtANX1 can be expressed in Marchantia thallus epidermis remains puzzling.

The A. thaliana type-one protein phosphatases (TOPP) AtATUNIS1/2 (AtAUN1/2) have recently been reported as negative regulators of cell wall integrity maintenance during Arabidopsis tip-growth (Franck et al., 2018b). The nucleocytoplasmic localization of AtAUN1-YFP and AtAUN2-YFP was demonstrated in Arabidopsis pollen tubes and leaf epidermal cells (Franck et al., 2018b). In Marchantia epidermal cells, expression of AtAUN1/2-YFP led to a comparable nucleocytoplasmic localization, as opposed to the co-expressed plasma membrane localized AtNPSN12-mCherry fusion (Figure 4), therefore qualifying these phosphatases as reliable Marchantia nucleocytoplasmic markers.

As for endosomal compartments, we chose two Ras-related in brain (RAB) GTPases, the canonical MpRAB5 and the plant-unique MpARA6, that were recently described in M. polymorpha. Both proteins were successfully expressed in stably transformed lines and co-localized to endosomal punctate structures stained by FM1-43 (Minamino et al., 2018). Upon biolistic co-transformation of the protein fusions mCherry-MpRAB5 and MpARA6-eYFP with the nuclear marker AtKRP1-eCFP (Figures 5A,B), we found a comparable localization in punctate structures for both markers. Moreover, both GTPases strongly co-localized with each other (Figure 5C) showing that MpRAB5 and MpARA6 are suitable endosomal markers also for transient transformation studies.

The carboxyl-terminal amino acid sequence serine–lysine–leucine (SKL) is well known as the consensus peroxisomal targeting sequence 1 (PTS1) and is sufficient to induce protein targeting and import to peroxisomes. SKL was first shown to be able to signal protein import into peroxisomes of mammalian cells (Gould et al., 1989) but later was also found to be functional in yeast and plants (Keller et al., 1991). In Arabidopsis, SKL motif fused to fluorescent tags is frequently used as a peroxisomal marker (Mathur et al., 2002; Kim et al., 2013; Rodríguez-Serrano et al., 2016). In M. polymorpha, SKL targeting was utilized for evaluation of CRISPR-Cas9 modules (Konno et al., 2018). In transiently transformed M. polymorpha cells, we also found a clear and distinct localization of mCherry-SKL in punctate structures, likely representing peroxisomes (Figure 6A).

Both the LifeAct peptide - a short peptide of 17 amino acids – and the C-terminal 197 amino acids of mouse talin are known to bind to filamentous actin (Kost et al., 1998; Riedl et al., 2008). Therefore, to visualize the actin filaments in Marchantia epidermal cells, we used the Citrine-mTalin and LifeAct-Citrine reporters described previously (Kimura and Kodama, 2016). As in stably transformed Marchantia lines (Kimura and Kodama, 2016), both markers successfully revealed the actin filament networks around chloroplasts in epidermal cells (Figures 6B,C).

As potential markers for the Golgi apparatus, we selected the Arabidopsis proteins SYNTAXIN OF PLANTS 3 (AtSYP3) and the GOLGI TRANSPORT 1 p homolog (AtGot1p). Both proteins have been shown to localize to the Golgi apparatus (Conchon et al., 1999; Uemura et al., 2004) and are reliable Golgi markers for Arabidopsis, as being part of the Wave line multicolor marker set for membrane compartments (WAVE22 and WAVE18, respectively; Geldner et al., 2009). Upon transient biolistic expression of eYFP-AtSYP3 and eYFP-AtGot1p in M. polymorpha, a distinct and comparable localization pattern of both proteins, likely representing the Golgi apparatus, was visible (Figures 7A,B). Furthermore, upon co-expression of eCFP-AtSYP3 and eYFP-AtGOT1p, we also found perfect co-localization (Figure 7C) confirming that both markers are reliable to illuminate the Golgi in Marchantia.

mRNA processing bodies (p-bodies), have been found to play a crucial role in mRNA processing comprising deadenylation, decapping, degradation, mRNA storage and mRNA quality control (thoroughly reviewed for A. thaliana in Maldonado-Bonilla, 2014). As p-bodies markers, we chose the Arabidopsis thaliana DECAPPING PROTEIN 1 (AtDCP1) and AtDCP2, whose function has been well studied in the past years (Xu et al., 2006; Xu and Chua, 2009; Steffens et al., 2015; Bhasin and Hülskamp, 2017). Upon transformation of the protein fusion AtDCP1-mCherry we found a comparable expression in dot-like structures, likely representing p-bodies (Figure 8A). In contrast, transformation of mCherry-AtDCP2 revealed a diffused expression throughout the cytoplasm and in the nucleus (Figure 8B), as reported in Arabidopsis in the absence of stress (Motomura et al., 2015). Therefore, we assume that AtDCP2 is also generally localized in the cytoplasm and nucleus in M. polymorpha and is only recruited to p-bodies upon stress conditions (Motomura et al., 2015).

The similar localization of AtDCP1/2 in Arabidopsis (Iwasaki et al., 2007; Motomura et al., 2015) and Marchantia suggests that the function of DCPs in mRNA processing has been evolutionarily conserved. To assess whether the two Marchantia DCP-homologs MpDCP1/2 localize similarly as their Arabidopsis counterparts, we transformed different combinations of fluorescent fusions (MpDCP1-mCherry, MpDCP1-eYFP, MpDCP2-mCherry, MpDCP2-eYFP). As anticipated, MpDCP1 displayed a dot-like localization pattern similar to AtDCP1, while MpDCP2 exhibited an AtDCP2-like nucleocytoplasmic localization (Figures 8C,D).

Fluorescein diacetate is a cell-permeable, per se non-fluorescent esterase substrate. As soon as it passes the plasma membrane, it is hydrolyzed by esterases in the cytoplasm of viable cells (Rotman and Papermaster, 1966). Thereby, FDA is converted to a negatively charged, green-fluorescent fluorescein unable to either cross back the plasma membrane or pass the tonoplast and thus it accumulates in the cytoplasm. Owing to these properties, FDA is suitable for cell viability assays and can be used as a negative stain for vacuoles. FDA staining has been reliably used for testing Arabidopsis root hair and guard cell viability (Schapire et al., 2008; Hao et al., 2012), to visualize vacuoles in root hairs (Saedler et al., 2009) and trichomes (Mathur et al., 2003), and to study pathogen response (Jones et al., 2016), as well as to assess Marchantia protoplast viability (Sugawara and Fukukawa, 1995).

Here, we successfully utilized FDA to stain the cytoplasm of rhizoids and epidermal cells in young gemmae (Figure 10). FDA showed a strong, green fluorescence already after a short incubation time of 10 min, demonstrating the viability of rhizoids and epidermal cells. We here present FDA as a tool to be readily used for visualization of the cytoplasm in M. polymorpha. As it is not able to pass the tonoplast, it can also be used to detect vacuolar architecture, especially in rhizoids, where vacuolar volume was clearly visible after staining with FDA (Figure 10D).

Propidium iodide (PI) is an intercalating, red-fluorescent cell dye. It penetrates damaged cell membranes and visualizes nuclei of dead cells by intercalating DNA with low base preference. However, PI cannot pass intact cell membranes and thus is excluded from viable cells, while remaining fluorescent. Therefore, PI can also readily be used to visualize cell wall of living cells. In Arabidopsis, PI is regularly utilized for counterstaining of cell walls (Takano et al., 2002; Ubeda-Tomás et al., 2009), such as for viability assays, frequently combined with FDA (Shahriari et al., 2010; Kong et al., 2018).

We here show successful PI staining of cell walls of M. polymorpha (Figure 11), in agreement with former reports (Delmans et al., 2017; Jones and Dolan, 2017; Thamm et al., 2020). Strong fluorescence was observed already after short incubation times of 10 min. PI reliably stained the cell walls of living epidermal cells (Figure 11A) and rhizoids (Figure 11B) and thus can reveal cell shape and size. This staining was non-toxic as stained rhizoids kept elongating, thereby revealing the usefulness of PI staining for studying rhizoid tip-growth (Supplementary Video S1).

DAPI is one of the most common DNA fluorochromes enabling staining and visualization of nuclei of dead but also viable cells, as it is able to pass cell membranes – however, often with weak effectiveness. Upon excitation with ultraviolet light, DAPI emits blue fluorescence at a maximum of 461 nm. DAPI binds stoichiometrically to adenine-thymine rich regions of DNA. DAPI also has a weak binding capacity to RNA, however emission is then shifted to 500 nm. Thus, DAPI is frequently utilized not only to visualize nuclei in trichomes, epidermal pavement cells or root cells (Kirik et al., 2001; Lee et al., 2006; Spitzer et al., 2006), but also to quantify DNA content in Arabidopsis, being a reliable tool to discover endoreduplication (Schnittger and Hülskamp, 2007; Bramsiepe et al., 2010; Bhosale et al., 2018). Kondou et al. (2019) report a functional DAPI staining of nuclei in wholemount samples of fixed epidermal cells of M. polymorpha. In this study, we tested staining of fixed (i.a. after a modified version of the protocol by Kondou et al., 2019) but also of viable thallus epidermal cells of Marchantia. Surprisingly, despite usage of gemmae at different developmental stages, short to long DAPI incubation periods, preceding and subsequent de-staining steps using EtOH, different methods of fixation (for more details see “Materials and Methods” section), we were unable to stain and visualize nuclei of Marchantia with DAPI (Supplementary Figure S3). In our hands, DAPI accumulated on cell walls and to a weaker extent in the cytoplasm but did not enter the nucleus. To demonstrate functionality of the used DAPI solution, we stained Arabidopsis leaves in parallel (Supplementary Figure S3), showing strong and distinctive visualization of nuclei. Staining of DNA by PI after fixation also failed in our hands (data not shown). Additionally, we stained young gemmae with Hoechst33342 for 10 min at 10 mg/L concentration, but it also failed in our hands to consistently stain nuclei. It remains to be elucidated, why nuclei of M. polymorpha seem to be hardly accessible to DNA fluorochromes. Until then, we either suggest to use a protein marker localizing in nuclei (e.g., AtKRP1) and to generate stably expressing Marchantia lines if needed; or to visualize S-phase nuclei with 5-ethynyl-2′-deoxyuridine (EdU) staining, as reports show its functionality in M. polymorpha (Furuya et al., 2018; Busch et al., 2019).

The lipophilic steryl dye FM4-64 ((3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)-phenyl)-hexatrienyl) pyridinium-dibromide) is commonly used as marker for the outer leaf of the cellular plasma membrane. Staining of young gemmae with FM4-64 resulted in a clear fluorescence signal at the cellular boundaries, likely representing the plasma membrane (Figure 12A). Upon co-staining with FDA, the FM4-64-specific plasma membrane signal at the cell periphery was clearly distinct from the cytoplasmic FDA signal (Figure 12B). Additionally, we stained Marchantia thallus fragments, transiently transformed with eYFP-MpRAB5 or MpARA6-eYFP, with the FM4-64 dye. The eYFP- and FM4-64-derived fluorescent signals co-localized at the punctuate endosomal structures (Figures 12C,D). Altogether, these findings support FM4-64 as a reliable marker dye to label the outer cellular membrane and endosomes via single or co-staining in Marchantia.