Auxin signal dynamics during Ceratopteris hermaphrodite and male development and male-to-hermaphrodite conversion

To visualize auxin signals in Ceratopteris gametophytes, we used the DR5 synthetic promoter, which contains tandem repeats of auxin response elements that marks active auxin signaling sites [40]. Specifically, we stably transformed Ceratopteris with the well-established DR5v2::ntdTomato reporter construct [41] and isolated multiple independent transgenic lines that exhibited comparable expression patterns in Ceratopteris gametophytes. To test whether the DR5 reporter was responsive to auxin in Ceratopteris gametophytes, we performed live-imaging experiments with IAA treatment. At 3 days after germination (DAG), individual males and hermaphrodites from the DR5v2::ntdTomato reporter line were transferred to medium supplemented with either mock treatment or 10 µM IAA (S1A–S1P Fig). After 24 h of IAA treatment, we found that, in contrast to the localized DR5 expression patterns observed in mock-treated controls (S1A–S1H Fig), DR5 signals became broadly distributed across nearly all cell in both male and hermaphrodite gametophytes (S1I–S1P Fig). Heat maps clearly illustrated significantly elevated DR5 signals in response to exogenous IAA compared to mock controls (S1A–S1D and S1I–S1L Fig). These results confirm that the DR5 reporter is functional and responsive to auxin signaling in Ceratopteris gametophytes.

We then performed non-invasive, time-lapse imaging to analyze auxin signaling dynamics during Ceratopteris hermaphrodite and male gametophyte development. Spores of the DR5v2::ntdTomato transgenic reporter line were plated on fern medium (FM) or conditioned FM (CFM, containing antheridiogen) to observe hermaphrodite and male development, respectively. Once germinated, hermaphrodite samples were live-imaged every day on the original FM plate from 1 to 6 DAG. At 1 DAG, DR5 signals were first detected in the developing prothallus (white arrowhead, Fig 1A), as well as in the rhizoid (yellow arrowhead). By 2 DAG, the DR5 reporter expression was localized to the basal part of the hermaphrodite (Fig 1B). From 3 to 6 DAG, corresponding to meristem initiation and proliferation, DR5 expression gradually formed an intensity gradient, with peak expression at the center of the notch meristem and lower expression further away from the notch (Magenta arrowheads, Fig 1C, 1D, and 1G–1L). DR5 signals were also observed in developing archegonia (Fig 1I–1L). In contrast, in males germinated and grown on CFM, DR5 signals were generally weaker compared to hermaphrodites on FM, suggesting that antheridiogen, the male differentiation inducer, negatively influences auxin signaling in males (Fig 2A–2L). Males and hermaphrodites exhibited differences in auxin signal distribution beginning at 2 DAG (Figs 1B, 1F, 2B, and 2E). We observed new auxin signals in the cells that eventually differentiated into the antheridium (cyan arrowheads, Fig 2). As the antheridia matured, DR5 signal intensity decreased, exhibiting a temporary increase followed by a reduction in signal intensity in the differentiating antheridium cells (Fig 2B, 2C, and 2G–2I). These dynamic expression patterns, especially the DR5 signal peak at the center of actively proliferating multicellular meristems, suggest potential roles of auxin in regulating Ceratopteris gametophyte development.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Auxin signaling dynamics during Ceratopteris hermaphrodite development.

(A–L) Confocal time-lapse images showing auxin signaling dynamics in Ceratopteris hermaphrodite gametophytes expressing the DR5v2::ntdTomato transgenic reporter. A hermaphrodite (A–L) was live-imaged every day from 1 to 6 days after germination (DAG). (A–D, I–J) Z-projection views of tdTomato signals (Fire LUT). (E–H, K–L) Merged channels of tdTomato (magenta) and DIC (gray, showing cell outlines). The white arrowhead (A) indicates the initial DR5 signals in the early developing prothallus, while yellow arrowheads (A–D) indicate DR5 signals in developing rhizoids. Magenta arrowheads (D, H, I–L) highlight the auxin signaling peak during the initiation and establishment of a multicellular meristem in the hermaphrodite. Color bars (A–D, I–J) represent the relative intensity of DR5 signals in Fire LUT images, with low intensity shown in black, intermediate intensity in blue to red, and high intensity in yellow to white. The color scale applies to the entire prothallus but excludes the spore coat. Scale bars: 50 µm. Three independent samples were analyzed, showing comparable results.

https://doi.org/10.1371/journal.pbio.3003592.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Auxin signaling dynamics during Ceratopteris male development in response to antheridiogen.

(A–L) Confocal time-lapse images showing auxin signaling dynamics in Ceratopteris male gametophytes expressing the DR5v2::ntdTomato transgenic reporter, grown on CFM. A male gametophyte was live-imaged every day from 1 to 6 DAG. (A–C, G–I) Z-projection views of tdTomato signals (Fire LUT). (D–F, J–L) Merged channels of tdTomato (magenta) and DIC (gray, showing cell outlines). The white arrowhead (A) indicates the initial DR5 signal in the early developing prothallus, while yellow arrowheads (A–C) indicate DR5 signals in developing rhizoids. Cyan arrowheads and white dashed circles (B–C, G–I) highlight DR5 signaling dynamics during antheridium initiation and maturation. Color bars (A–C, G–I) represent the relative intensity of DR5 signals in Fire LUT images, with low intensity shown in black, intermediate intensity in blue to red, and high intensity in yellow to white. The color scale applies to the entire prothallus but excludes the spore coat. Scale bars: 50 µm. Three independent samples were analyzed, showing comparable results.

https://doi.org/10.1371/journal.pbio.3003592.g002

Next, we performed time-lapse imaging to investigate the DR5 reporter expression patterns during male-to-hermaphrodite conversion (Fig 3A–3P), comparing them with patterns observed during normal male and hermaphrodite development (Figs 1A–1L and 2A–2L). DR5v2::ntdTomato spores were surface-sterilized and plated on CFM to induce male differentiation, as described in previous studies [8,9,42]. Individual male gametophytes at 3 DAG were transferred to antheridiogen-free FM, with one gametophyte per plate to initiate de novo meristem formation without potential interference from antheridiogen released by neighboring hermaphrodites. Confocal live-imaging was performed immediately after transfer (0 h after transfer, 0 h; Fig 3A) and continued at multiple time points from 48 to 216 h after transfer, capturing the initiation and establishment of new hermaphrodites (Fig 3B–3P). At 0 h, weak DR5 signals were observed in developing antheridia (Fig 3A), similar to those observed in males continuously grown on CFM (Fig 2B, 2C, 2E, and 2F). These signals gradually diminished as each antheridium matured (Fig 3B–3F; white dashed circles in Fig 3B and 3C). At subsequent time points, no new antheridium-differentiated cells appeared, and no further DR5 signals were detected within existing antheridia (Fig 3B–3F), consistent with the previous observation that continuous exposure to antheridiogen is necessary for new antheridium initiation [5]. More importantly, time-lapse imaging revealed that the emergence of new DR5 signals at 48 h after transfer (yellow circles, Fig 3A and 3B), which persisted (yellow circle, Fig 3C) and continued to expand, correlating with the initiation and growth of the new hermaphrodite body (magenta arrowheads, Fig 3B–3F). By 108 h, DR5 signals were highly expressed at the basal region of the newly formed hermaphrodite (Magenta arrows, Fig 3F and 3G), mimicking to expression patterns observed during normal hermaphrodite development (Fig 1B and 1C). DR5 signals maintained highly expressed in the base region of the newly formed hermaphrodite throughout its expansion and gradually formed a distinct intensity gradient as the de novo meristem was initiated and became established (Fig 3H–3P). Heat maps revealed that, once a concave meristem notch formed, DR5 signals peaked at the center of the notch (arrowheads, Fig 3K–3P). In summary, auxin signaling dynamics are directly associated with the initiation of male-to-hermaphrodite conversion (Fig 3B–3F) and the subsequent meristem initiation and notch formation (Fig 3K–3P).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Auxin signaling dynamics during male-to-hermaphrodite conversion and de novo meristem formation in the absence of antheridiogen.

(A–P) Confocal time-lapse images of a Ceratopteris gametophyte expressing the DR5v2::ntdTomato reporter. A 3-DAG male gametophyte was first imaged at 0 h (0 h) after being transferred to an antheridiogen-free medium and subsequently live-imaged at multiple time points from 48 to 216 h after the transfer. (A–P) Z-projection views of tdTomato signals (Fire LUT). White dashed circles (B–C) mark regions with reduced DR5 signals corresponding to antheridium differentiation. Yellow dashed circles (A–C) highlight the region with newly merged auxin signals at the site of de novo hermaphrodite formation. Magenta arrows (B–F) indicate the emergence (B) and persistence (C–F) of DR5 signals at the initiation site of the newly forming hermaphrodite. Magenta arrowheads (K–P) indicate dynamic DR5 signal patterns associated with de novo multicellular meristem formation. A white curved arrow (G) indicates the adjusted orientation of the sample at the same point (108 h, G) for clear visualization of the newly formed hermaphrodite body. Color bars (A–P) represent the relative intensity of DR5 signals in Fire LUT images, with low intensity shown in black, intermediate intensity in blue to red, and high intensity in yellow to white. The color scale applies to the entire prothallus but excludes the spore coat. Scale bars: 50 µm. Three independent samples were analyzed, showing comparable results.

https://doi.org/10.1371/journal.pbio.3003592.g003

Auxin biosynthesis is essential for male-to-hermaphrodite conversion

To explore the role of auxin, particularly the establishment of new peak of auxin signaling, in controlling male-to-hermaphrodite conversion and associated de novo meristem formation, we first employed chemical perturbation on new auxin production. L-Kynurenine (Kyn) is a specific inhibitor of TAA1-mediated auxin biosynthesis, as previously demonstrated in the flowering plant model Arabidopsis and the seed-free plant model Marchantia [31,43]. Kyn therefore served as a valuable tool to assess how disrupted auxin biosynthesis affects sex-type conversion and meristem initiation in Ceratopteris gametophytes. For this test, individual 3-day-old WT male gametophytes initially germinated on CFM were transferred onto antheridiogen-free FM plates (one male per plate) supplemented with either mock or Kyn at concentrations of 50 or 150 µM (Fig 4A–4I). This experimental design allowed us to directly evaluate the effects of Kyn treatment on male-to-hermaphrodite conversion relative to mock controls (Fig 4A–4I). Light microscopy images of these samples were captured immediately after transfer (0 days after transfer [DAT]) and subsequently at 7 and 10 DAT. Our results demonstrated a clear, dosage-dependent inhibitory effect of Kyn on male-to-hermaphrodite conversion (Fig 4J and S1 Data). Mock-treated controls exhibited no inhibition, achieving over 93% successful conversions of sex types (Fig 4A–4C, 4J, and S1 Data). In contrast, treatment with 50 µM Kyn reduced the conversion rate to approximately 73% (Fig 4J and S1 Data), and the resulting hermaphrodites were notably smaller compared to mock-treated controls (Fig 4C and 4F). At the higher concentration (150 µM), Kyn strongly inhibited conversion and prevented new meristem formation (Fig 4G–4I), with only 10% of males successfully transitioning to hermaphrodites (Fig 4J and S1 Data). Notably, at this concentration, no meristem structures were observed at 7 DAT, and most males exhibited minimal growth between 7 DAT and 10 DAT (Fig 4H–4I).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Auxin biosynthesis inhibition blocks male-to-hermaphrodite conversion in the absence of antheridiogen.

Light micrographs of representative male gametophytes at 0, 7, and 10 days after transfer (DAT) to antheridiogen-free FM, treated with either mock (A–C), 50 µM L-kynurenine (D–F), or 150 µM L-kynurenine (G–I). In the mock-treated sample, the yellow arrow (A) indicates the site where a meristem and a hermaphrodite body begin to form de novo. White dashed rectangles (B–C) indicate the original male body observed in (A). Blue arrowheads (A, D, G–I) indicate several antheridia, while magenta arrowheads (B–C, F) indicate the newly formed meristem. (J) The percentage of the successfully converted hermaphrodites out of the total samples examined was calculated from three groups of experiments, each containing 20 independent gametophytes (60 in total). The data underlying panel (J) can be found in S1 Data. (K–V) Confocal time-lapse images of Ceratopteris gametophytes expressing the DR5v2::ntdTomato reporter. 3-DAG male gametophytes were first imaged at 0 DAT to antheridiogen-free FM (K, S), treated with either mock (K) or 150 µM L-kynurenine (S), and then live-imaged at multiple time points from 2 to 8 DAT (K–R, S–V). (K–V) Z-projection views of tdTomato signals (Fire LUT). White arrowheads (K–L, S–T) indicate DR5 signals in the antheridia of male gametophytes. Magenta arrows indicate the site of the de novo meristem and hermaphrodite body formation in the mock-treated sample (M–O). The magenta arrowhead (R) indicates DR5 signals in the newly formed multicellular meristem. The white curved arrow (O) indicates the adjusted orientation of the mock-treated sample from 4 DAT to 5 DAT for clear visualization of the developing hermaphrodite. Color bars (K–V) represent the relative intensity of DR5 signals in Fire LUT images (K–V), with low intensity shown in black, intermediate intensity in blue to red, and high intensity in yellow to white. The color scale applies to the entire prothallus but excludes the spore coat. Scale bars: 100 µm (A–I) and 50 µm (K–V). Three independent samples from each treatment were analyzed, showing comparable results.

https://doi.org/10.1371/journal.pbio.3003592.g004

We then investigated how Kyn treatment affects DR5 expression patterns following the transfer of males to antheridiogen-free medium. 3-DAG DR5v2::ntdTomato male gametophytes were transferred to antheridiogen-free FM plates supplemented with either mock or 150 µM Kyn. In mock-treated controls (Fig 4K–4R), DR5 expression patterns closely resembled those during male-to-hermaphrodite conversion described above (Fig 3), with initial new DR5 peaks overlapping with the initiation site of newly formed hermaphrodites, followed by a clear intensity gradient established in the de novo meristem during sex-type conversion (Fig 4L–4R). In contrast, Kyn treatment substantially diminished auxin signaling dynamics (Fig 4S–4V). Despite transfer to an antheridiogen-free environment, no new proliferation sites or associated DR5 signal were observed in Kyn-treated samples (Fig 4S–4V). These results suggest that reducing auxin biosynthesis by Kyn-mediated inhibition of CrTAA1 diminishes auxin signaling essential for male-to-hermaphrodite conversion.

Auxin biosynthesis is required for MPC establishment and cell proliferation during de novo meristem formation

Our previous studies demonstrated that upon removal from antheridiogen, male gametophytes transition into hermaphrodites, establishing a newly formed and actively proliferating meristem. This multicellular meristem originates from a single MPC in the original male [9]. To visualize cellular proliferation dynamics at single-cell resolution under conditions where auxin biosynthesis was disrupted, we used the transgenic reporter line pCrUBQ10::H2B-GFP::3′CrUBQ10, which ubiquitously expresses a nuclear-localized GFP in all cells, enabling dynamic analysis of individual cells over time [9]. Spores from the reporter line were surface-sterilized and germinated on CFM to induce male differentiation. Two DAG, male gametophytes were transferred onto antheridiogen-free FM supplemented with either mock treatment or with 150 µM of the auxin biosynthesis inhibitor, Kyn (Figs 5A–5F and S2–S7).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. L-kynurenine inhibits the formation of MPC during male-to-hermaphrodite conversion.

(A–F) Confocal images of 2-DAG Ceratopteris male gametophytes expressing CrUBQ10p::H2B-GFP::3′CrUBQ10, taken at 0 (A, D), 60 (B, E), and 120 (C, F) hours after treatment with either mock (sample 1, A–C) or 150 µM Kyn (sample 4, D–F). Green: GFP. (C, F) Merged channels of GFP (green) and propidium iodide (PI) (magenta); PI staining was performed at 120 h to visualize cell outlines. Time-lapse imaging was performed every 6 h, and all time points from 0 to 120 h for sample 1 and sample 4 are shown in S2A–S2V and S5A–S5V Figs, respectively. Panels (A–F) correspond to zoomed-in regions from the supplementary figures: (A) from S2A Fig, (B) from S2K Fig, (C) from S2V Fig; (D) from S5A Fig, (E) from S5K Fig, and (F) from S5V Fig. Three independent biological replicates per treatment were imaged under the same conditions, yielding comparable results for each treatment. Time-lapse imaging results from the other samples are shown in S3 and S4 Figs (mock-treated samples 2 and 3) and S6 and S7 Figs (Kyn-treated samples 5 and 6). (G–L) Cell lineage dynamics of live-imaged sample 1 (G–I) and sample 4 (J–L) at 0, 60, and 120 h, corresponding to panels (A–F). Each solid circle represents either a single nucleus or a group of nuclei within one antheridium. At 0 h, adjacent nuclei were labeled with different colors as a reference for visualizing distinct lineages at subsequent time points. Cells and their progeny were labeled with the same color across all time points to represent the same lineage. When an antheridium developed into a 3D complex structure, all nuclei within the same antheridium were represented as a single solid circle to simplify visualization. Magenta arrows highlight either the MPC (yellow) lineage (A, G) contributing to de novo meristem formation in mock-treated sample 1 or the lineage (yellow) (D, J) with the most division events in Kyn-treated sample 4. Full lineage dynamics from 0 to 120 h for samples 1 and 4 are shown in S8A–S8U and S11A–S11U Figs, respectively. Panels (G–I) correspond to zoomed-in regions from the supplementary figures: (G) from S8A Fig, (H) from S8K Fig, and (I) from S8U Fig; (J) from S11A Fig, (K) from S11K Fig, and (L) from S11U Fig. Lineages from three independent samples per treatment were analyzed, showing comparable results for each treatment. Lineage dynamics of the other samples are shown in S9 and S10 Figs (mock-treated samples 2 and 3) and S12 and S13 Figs (Kyn-treated samples 5 and 6). (M–R) Quantification of total division events for each cell lineage across all analyzed samples (from S2 to S7 Figs). Each solid circle represents an individual nucleus at 0 h. Numbers indicate the total division events over 120 h within the MPC lineages from mock-treated samples 1-3 (M–O) and the lineages with the most division events from Kyn-treated samples 4-6 (P–R). Total division events are color-coded, ranging from blue (minimum, 0) to red (maximum, > 53). Scale bars (A–R): 50 µm.

https://doi.org/10.1371/journal.pbio.3003592.g005

Confocal live imaging was performed directly on growing Petri dishes every 6 h, beginning immediately after transfer and continuing for a total of 120 h (Figs 5A–5F and S2–S7). Between imaging intervals, samples were returned to growth chambers and maintained under standard growth conditions (29 °C, continuous light). Consistent with previous observations in WT [9], non-antheridium cells in mock-treated samples actively proliferated after transfer to antheridiogen-free medium, successfully transitioning from males to hermaphrodites by 120 h (Figs 5A–5C and S2–S4). This transition was characterized by the formation of a clear, notch-shaped de novo meristem (Figs 5C, S2U, S2V, S3U, S3V, and S4U). In contrast, Kyn-treated samples, though placed on antheridiogen-free medium, exhibited low proliferation activity and remained male throughout the entire 120-h imaging period (Figs 5D–5F and S5–S7), consistent with results from light microscopy (Fig 4G–4J). These findings demonstrate that Kyn treatment efficiently inhibits the male-to-hermaphrodite transition triggered by antheridiogen-free conditions.

We then applied the previously established pipeline [9,44] to performed 2D imaging analyses on samples treated with either mock or Kyn (Figs 5G–5L and S8–S13). We segmented each GFP-tagged nucleus at various time points and assigned unique IDs to each individual nucleus. This enabled tracing of each nucleus and its descendants over time, generating dynamic lineage maps of all segmented nuclei from 0 to 120 h, corresponding to the period of initiation and formation of the new multicellular meristems in mock-treated samples (Figs 5G–5L, S8, and S10). At 0 h, each nucleus represented an independent lineage, with their clonal descendants sharing with the same color in all subsequent time points (Figs 5G–5L, S8, and S10). Consistent with previous observations in WT samples [9], in mock-treated samples, a single progenitor cell at 0 h, identified as the MPC (magenta arrows, Figs 5G and S8A), gave rise to the entire de novo meristem (Figs 5G–5I and S8A–S8U). The MPC lineage (yellow) progressively expanded, becoming the dominant cell sector (Figs 5G–5I and S8A–S8U) that eventually contributing to all cells within the de novo meristem (Figs 5I, S2U, S2V, and S8U). The MPC lineage represents the most actively dividing cell population, surpassing all other lineages. Conversely, in the Kyn-treated sample, the actively dividing lineages (such as the one in yellow, Figs 5J–5L and S11A–S11U) exhibited a significantly reduced number of progenies, with a much smaller sector compared to the MPC lineage in the mock control at the same time points (Figs 5G–5I and S8A–S8U). Due to the failure in de novo meristem establishment under Kyn treatment, no MPC lineage could be defined (Figs 5J–5L and S11A–S11U). These results were consistent across multiple samples and confirmed through lineage analysis (mock-treated samples 1–3 in S8–S10 Figs versus Kyn-treated samples 4–6 in S11–S13 Figs).

Aligning with lineage maps (Figs 5G–5L and S8–S13), quantitative analyses demonstrated that, over the 120-h period, the total number of division events in MPC lineages from mock-treated samples 1–3 (Fig 5M–5O) consistently exceeded those observed in any lineage from three Kyn-treated samples 4–6, including the most actively dividing lineage (Fig 5P–5R). These results suggest that de novo auxin biosynthesis is crucial for meristem cell proliferation. Detailed 12-h division maps (S14–S18 Figs) further illustrated the inhibitory effects of Kyn treatment on cell proliferation during the male-to-hermaphrodite switch and associated de novo meristem formation, providing high spatial and temporal resolution. In these maps, segmented cells undergoing division within each 12-h interval were highlighted in magenta, while those remaining undivided were shown in green (S14–S18 Figs). Cell division occurring within mature antheridia was not included in this analysis. Specifically, during the first 60 h after transfer (S14–S18 Figs), cell division events were randomly distributed across the apical region in both mock- and Kyn-treated samples, and the MPC lineage in the mock-treated samples (S14A–S14D, S14I, S15A–S15E, and S16A–S16E Figs) did not yet exhibit a clear difference in division activity, which is consistent with the previously defined Phase I of the WT male-to-hermaphrodite conversion [9]. During this phase, Kyn-treated samples showed fewer overall dividing cells (S14E–S14H, S14L, S17A–S17E, and S18A–S18E Figs) compared to mock samples (S14A–S14D, S14I, S15A–S15E, and S16A–S16E Figs), suggesting that auxin may facilitate cells re-entering the cell cycle. During the subsequent 60 h, corresponding to Phase II as previously defined [9], division events in mock-treated samples became gradually localized and eventually restricted to the MPC lineages (S14J, S14K, S14O–S14Q, S15F–S15J, and S16F–S16J Figs). In contrast, the number and pattern of division events in Kyn-treated samples remained unchanged compared to Phase I (S14M, S14N, S14R–S14T, S17F–S17J, and S18F–S18J Figs), and no actively proliferating MPC lineages could be identified. For direct comparison, the most actively dividing non-antheridium cell lineages from each Kyn-treated sample were outlined (white dashed outlines, S14, S17, and S18 Figs), clearly showing fewer divisions compared to the MPC lineage (yellow dashed outlines, S14–S16 Figs) in mock-treated samples. In summary, disrupting auxin biosynthesis efficiently blocks the male-to-hermaphrodite transition by suppressing the initiation and subsequent proliferation of the MPC lineage.

Dynamic expression patterns of CrTAA1 during male-to-hermaphrodite conversion

We next set out to identify and characterize TAA1 homologs in Ceratopteris. Protein sequences of Arabidopsis AtTAA1 and AtTAR homologs were used as references for BLASTp searches against the genomes of Ceratopteris richardii, Azolla filiculoides, and Salvinia cucullata. Previously identified TAA1 and TAR homologs from several representative land plant species [45] were obtained from the Phytozome database. Multiple sequence alignments were performed, and a maximum-likelihood phylogenetic tree was constructed (S19 Fig). Candidate Ceratopteris orthologs were further analyzed with InterProScan and compared to Arabidopsis TAA1 family members (S2 Data). Our phylogenetic analysis revealed that only one C. richardii gene, CrTAA1 (Ceric.10G039800), grouped within the AtTAA1/AtTAR1/AtTAR2 lineage (S19 Fig and S2 and S3 Data). Furthermore, CrTAA1 is the only Ceratopteris gene classified within the TAA clade, which is characterized by the exclusive presence of an Alliinase-C domain (IPR006948) (S2 Data). In contrast, all other candidate Ceratopteris orthologs fell within a different (AtTAR3/AtTAR4) clade and possessed both an N-terminal Alliinase-EGF (Epidermal Growth Factor, IPR006947) domain and the Alliinase-C domain (S2 Data). Together, these results demonstrate that CrTAA1 (Ceric.10G039800) is the sole TAA1 homolog in the Ceratopteris genome.

We initially characterized the expression dynamics of CrTAA1 using quantitative PCR (qPCR) (Fig 6A and S4 Data). Wild-type (WT) spores were plated on FM, and RNA was separately extracted from male and hermaphroditic gametophytes harvested from the same Petri dishes at 4, 5, 8, or 10 DAG. This allowed for a direct comparison of CrTAA1 expression levels in males and hermaphrodites over time. At each time point, males or hermaphrodites were pooled as a single biological replicate, with at least three biological replicates included to assess CrTAA1 expression at each stage. CrACTIN1 was used as the reference gene, as described previously [42]. CrTAA1 transcripts were consistently detected at all time points, with expression observed at 4, 5, 8, and 10 DAG during the gametophyte stage (Fig 6A and S4 Data). Notably, starting at 5 DAG, when meristems are established in hermaphrodites, CrTAA1 expression was significantly higher in hermaphrodites compared to males at the same DAG. This suggests that CrTAA1 expression correlates to meristem development in hermaphrodites.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. CrTAA1 expression patterns in Ceratopteris gametophytes.

(A) Quantitative PCR analysis showing the relative expression levels of CrTAA1 (normalized to the internal control CrACTIN1) in hermaphroditic gametophytes (HG) and male gametophytes (MG) at the indicated DAG. Bar: mean ± sem (n = 4 biological replicates for 4-DAG MG, 5-DAG HG or MG, 8-DAG HG or MG, and 10-DAG HG or MG; n = 3 biological replicates for 4-DAG HG). ns, not significant (p > 0.05); ** p < 0.01; *** p < 0.001. Each biological replicate included pooled hermaphrodites or males collected at the indicated DAG. The data underlying panel (A) can be found in S4 Data. (B–H) Confocal images of pCrTAA1::H2B-GFP::3′CrTAA1 reporter expression in different hermaphrodites at the indicated DAG. (C–E) One hermaphrodite at 6 DAG. (F–H) One hermaphrodite at 7 DAG. (D, G) Zoomed-in views of the meristems of the hermaphrodites shown in (C) and (F), respectively. (I–M) Confocal images of different gametophytes during male-to-hermaphrodite conversion, expressing the pCrTAA1::H2B-GFP::3′CrTAA1 reporter, taken at indicated days after transfer (DAT) to antheridiogen-free medium. (I, J) Gametophytes at 3 DAT. (K) Zoomed-in view of the initiation site of hermaphrodite formation shown in (J). (L, M) A newly formed hermaphrodite at 9 DAT. (M) Zoomed-in view of the de novo meristem shown in (L). Magenta arrowheads (B–D, F–G, L–M) indicate CrTAA1 expression specifically in the notch region of multicellular meristems during normal hermaphrodite development (B–D, F–G) and in de novo formed meristems during male-to-hermaphrodite conversion (L–M). Magenta arrows (J, K) indicate GFP signals at the initiation site of de novo meristem formation. (B–D, F–G, I–M) Z-projection views of GFP signals (Fire LUT). (E, H) Merged channels of GFP (green) and DIC (gray, showing cell outlines). Scale bars: 50 µm. Color bars (B–D, F–G, I–M) represent the relative intensity of GFP signals in Fire LUT images, with low intensity shown in black, intermediate intensity in blue to red, and high intensity in yellow to white. The color scale applies to the entire prothallus but excludes the spore coat.

https://doi.org/10.1371/journal.pbio.3003592.g006

Next, we generated a transgenic fluorescent transcriptional reporter for CrTAA1 (pCrTAA1::H2B-GFP::3′CrTAA1) to investigate its expression dynamics with higher spatial resolution (Fig 6B–6M). First, we focused specifically on CrTAA1 expression patterns during hermaphrodite development, performing confocal live imaging of hermaphrodite samples at various DAG (Fig 6B–6H). Quantitative imaging results revealed that the CrTAA1 reporter exhibited relatively weak, yet highly specific, expression patterns (Fig 6B–6H). At 3 DAG, CrTAA1 expression localized specifically to one side of the hermaphrodite, overlapping with the initiation site of the multicellular meristem (Fig 6B). As the meristem further developed, CrTAA1 expression became exclusively confined to the central region of the meristem, particularly concentrated at the meristem notch region (Fig 6C-6H). Heat maps showed an expression gradient, with peak CrTAA1 expression at the notch center (Fig 6D and 6G). These findings suggest that during meristem establishment and notch formation in hermaphrodites, auxin biosynthesis is locally upregulated at the meristem center through at least the TAA1 pathway. Then, we examined CrTAA1 reporter expression during the male-to-hermaphrodite transition and de novo meristem formation. Males of the CrTAA1 reporter line, initially germinated on CFM, were transferred onto antheridiogen-free FM and imaged at different DAT. By 3 DAT, when males began initiating a new proliferation site, this reporter signal became specifically localized at this site (Fig 6I–6K), aligning with the DR5 activity pattern observed during de novo meristem initiation (Fig 3B–3E). At 9 DAT, after the male-to-hermaphrodite transition had been completed and the de novo meristem was fully established, CrTAA1 reporter expression was explicitly localized at the center of the newly formed meristem (Fig 6L and 6M), similar to the expression pattern observed during normal hermaphrodite development (Fig 6D and 6G).

Stable gene-editing demonstrates the crucial role of CrTAA1 in male-to-hermaphrodite conversion and de novo meristem formation

Given that CrTAA1 expression is dynamically and specifically localized to the initiation site of newly formed hermaphrodites and subsequently at the meristem notch region (Fig 6), we hypothesized that CrTAA1 is essential for the de novo formation of meristems and conversion from males to hermaphrodites. This hypothesis was tested and supported by results from chemical perturbation experiments using the TAA1-specific inhibitor Kyn (Figs 4, 5, S2–S7), prompting us to further test the hypothesis through genetic perturbation using CRISPR-Cas9-mediated gene editing.

To genetically test the requirement of CrTAA1, we generated stable Ceratopteris knockout lines using CRISPR-Cas9. The CRISPR construct was designed with two gRNAs targeting exon 1 of CrTAA1, each driven by an independent CrU6 promoter and terminated by a single CrU6 terminator. Following the established procedure for stable transformation and regeneration of transgenic lines, we identified multiple lines with different mutations in CrTAA1. Following self-fertilization, confirmed knockout lines were obtained, and their spores were harvested. Two independent CrTAA1 CRISPR mutants, crtaa1-1 with a single nucleotide insertion (A) and crtaa1-2 with a single nucleotide deletion (C) in exon 1 (Fig 7A–7C), both introduced frameshifts that led to premature termination, generating null alleles (S20 Fig). These mutant lines, together with WT controls, were selected for detailed phenotypic analysis. To evaluate their ability to undergo the male-to-hermaphrodite transition, individual 3-DAG male gametophytes from WT or CrTAA1 knockout lines (crtaa1-1 and crtaa1-2) were transferred onto antheridiogen-free FM plates, and their development was monitored by light microscopy immediately after transfer and at 10 DAT. At 10 DAT, all WT gametophytes (20 out of 20) had successfully transitioned into hermaphrodites (Fig 7D and 7E). In contrast, at least 90% of gametophytes from both CrTAA1 knockout lines remained male at 10 DAT (Fig 7F–7I). The genetic knockout of CrTAA1 phenocopied the inhibitory effects observed with Kyn treatment, providing robust genetic evidence that TAA1-mediated auxin biosynthesis is essential for male-to-hermaphrodite conversion and the associated de novo meristem formation in Ceratopteris.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. CrTAA1 is crucial for male-to-hermaphrodite conversion in the absence of antheridiogen.

(A–C) Sanger sequencing results showing representative WT and edited CrTAA1 sequences. (D–I) Representative micrographs showing male gametophytes (3 DAG) from WT (highlighted in blue) and CrTAA1-CRISPR lines (crtaa1-1 and crtaa1-2, highlighted in magenta) transferred from CFM (containing antheridiogen) to antheridiogen-free FM. Images were taken immediately after transfer (0 DAT) (D, F, H) and 10 days after transfer (DAT) (E, G, I), with 20 out of 20 for WT, 18 out of 20 for crtaa1-1 (line 99), and 20 out of 20 for crtaa1-2 (line 103). (E) Magenta arrowhead: de novo meristem; yellow arrowheads: newly formed archegonia. Scale bars (D–I): 100 µm.

https://doi.org/10.1371/journal.pbio.3003592.g007

Auxin peak and male-to-hermaphrodite conversion

Our previous work demonstrated that transferring Ceratopteris males from an antheridiogen-containing to an antheridiogen-free environment triggers a profound developmental switch: cell proliferation is re-initiated at a new site, establishing the MPC lineage and subsequently leading to hermaphrodite formation [9]. However, the molecular basis by which the absence of antheridiogen induces this reprogramming of cell fate remains unclear. This study begins to address this knowledge gap by identifying a key signaling event, localized auxin biosynthesis, that is dynamically activated under antheridiogen-free conditions and is essential for driving male-to-hermaphrodite conversion.

During this transition, an auxin signal maximum appears to define an early step in cell fate re-specification. Upon transfer to antheridiogen-free conditions, DR5 reporter activity in males is strongly and specifically induced at a newly initiated site of cell proliferation (Fig 3B and 3C). This DR5 signal marks the re-entry of cells into the cell cycle (Fig 5A–5C, 5G–5I, and 5M–5O) and persists throughout the formation of the new hermaphrodite (Fig 3B–3P). This new DR5 signal peak (Fig 3B and 3C) coincides spatially and temporally with the localized activation of CrTAA1 expression shortly after antheridiogen removal (Fig 6I–6K). Furthermore, this early auxin activation event not only serves as a molecular marker for hermaphrodite initiation but also plays a functional role in enabling sex-type conversion. Chemical inhibition of TAA1-dependent auxin biosynthesis by Kyn abolished the formation of the DR5 signal maximum (Fig 4S–4V) and suppressed male-to-hermaphrodite conversion in a concentration-dependent manner (Fig 4A–4J and S1 Data). Lineage-tracing and quantitative cell division analyses revealed that Kyn treatment severely disrupted MPC lineage establishment and subsequent cell proliferation (Figs 5 and S8-S18). Consistently, CRISPR-Cas9-mediated knockout of CrTAA1 phenocopied these effects (Fig 4A–4J), effectively blocking male-to-hermaphrodite conversion (Fig 7). Together, these results reveal an essential molecular pathway underlying sex-type conversion: the absence of antheridiogen induces a local auxin biosynthesis hotspot that generates a new auxin signal maximum. This auxin maximum is required to initiate the MPC lineage, sustain cell proliferation, and drive de novo meristem formation, which are essential for successful sex-type conversion. Given that DR5 signals are also dynamically present in female reproductive organs (egg-bearing archegonia) during both normal hermaphrodite development and male-to-hermaphrodite conversion (Figs 1 and 3), future studies should determine whether auxin directly promotes female organ formation and defines hermaphrodite identity, or acts indirectly by regulating meristem activity, which in turn triggers archegonium formation in Ceratopteris [2,44].

Conserved auxin biosynthesis plays context-specific roles in fern gametophytes

Previous genetic studies have shown that local auxin biosynthesis, mediated by the TAA1-YUC pathway, regulates stem cell maintenance and differentiation in both seed plants and seed-free nonvascular plants [23,25,31,46,47]. In seed plants, disruptions of this pathway impairs stem cell activity and identity in both root and shoot meristems [23,25]. For instance, in Arabidopsis SAMs, auxin biosynthesis genes are expressed in the epidermal layer, and the taa1 tar2 double mutant exhibits reduced shoot meristem size compared to WT [25]. In the liverwort Marchantia, auxin biosynthesis is localized to the apical meristem region, as indicated by MpTAA transcript and IAA accumulation [31]. Loss-of-function mptaa mutants display reduced thallus growth and impaired tissue differentiation [31]. In this study, we uncover a previously undefined role for auxin biosynthesis in regulating sex type and de novo meristem formation in fern gametophytes. Our results (S19 Fig and S2 Data), together with previous analysis [39], demonstrate that CrTAA1 (Ceric.10G039800) is the closest homolog of Arabidopsis TAA1, whereas other related Ceratopteris proteins group into a distinct clade corresponding to the functionally uncharacterized Arabidopsis TAR3/4 (S19 Fig). CrTAA1 is specifically activated at the onset of the male-to-hermaphrodite transition, and its expression becomes confined to the center of emerging de novo meristems (Fig 6L and 6M), overlapping with the developing concave meristem notch. These patterns align with the crucial role of CrTAA1 in initiating and establishing new meristems during the haploid gametophyte phase. Consistent with these findings, previous transcriptomic and phylogenetic analyses identified multiple YUC homologs in the Ceratopteris genome [42]. One of these homologs is activated during early hermaphrodite development and meristem establishment, dependent on CrHAM, a meristem-specific transcription factor that promotes meristem cell division and prevents male differentiation [42,48]. Moreover, treatment with yucasin [49], a specific inhibitor of YUC family enzymes in both flowering plants (such as Arabidopsis) and bryophytes (such as Physcomitrium and Marchantia) [49,50], effectively blocked de novo meristem formation and male-to-hermaphrodite conversion in Ceratopteris (S21 Fig and S5 Data), similar to the defects caused by chemical or genetic perturbation of CrTAA1. Together, these results highlight that the TAA1-YUC-mediated auxin biosynthesis pathway is evolutionarily conserved across land plants, including bryophytes, seed-free vascular plants, and seed plants, yet plays distinct, context-specific roles in fern gametophytes, where it governs de novo meristem formation and defines the early, essential steps of sex-type conversion. Furthermore, because Ceratopteris hermaphrodites and Marchantia gametophytes share comparable notch meristems and similar localized expression of their TAA1 homologs, comparative studies of auxin function in these systems will further illuminate the evolution of meristem regulation in land plants.

Beyond the meristem initiation sites and newly formed meristems observed during sex-type conversion (Fig 3), auxin signals revealed by the DR5 reporter are also active during normal meristem development in hermaphrodites and in differentiated organs, including rhizoids, sperm-producing antheridia, and egg-bearing archegonia (Figs 1 and 2). This suggests multiple and potentially distinct roles of auxin in regulating Ceratopteris gametophyte development across different stages. Future genetic studies using the established CrTAA1 mutants and other CRISPR-generated auxin-related mutants, focusing on developmental processes such as meristem notch formation in hermaphrodites, as well as antheridium differentiation and spermatogenesis in males, will help establish a more comprehensive understanding of auxin function during the fern gametophyte phase. Furthermore, the DR5 reporter examined in this study reflects the transcriptional output of auxin signaling, based on ARF gene activity [41]. Although no previous studies have applied it in ferns, the ratiometric auxin sensor R2D2 has been shown across multiple plant lineages to serve as a direct readout for auxin input or accumulation [41]. Future studies using R2D2 as an independent reporter system to determine auxin dynamics in Ceratopteris gametophytes, in direct comparison with DR5 expression patterns, will provide more insights into auxin function during fern development.

Plant material and growth condition

Plant material and growth condition in this study followed the procedures described in Geng and colleagues (2022) [44]. Ceratopteris richardii (Ceratopteris) strain Hn-n [3] was used as the WT for generating transgenic plants. Ceratopteris gametophytes were grown on either fern media (FM) containing 0.5 × MS salts and vitamins (PhytoTechnology Laboratories) and 0.7% (w/v) agar (Sigma-Aldrich) or conditioned FM (CFM, supplemented with antheridiogen) as previously described [5]. Hermaphrodites were individually transferred to 48-well plates for self-fertilization. Ceratopteris calli were induced from young sporophytes (shoot tips or fronds) on callus induction medium (pH 5.8) containing 1 × MS salts and vitamins (PhytoTechnology Laboratories), 2% (w/v) sucrose, 1 mg/L benzylaminopurine, and 0.7% agar (Sigma–Aldrich). Gametophytes, calli, and sporophytes were cultured under continuous light at 29 °C. Fertilized sporophytes or regenerated sporophytic shoots were transferred to soil. Adult sporophytes were subsequently transferred to large soil pots and cultivated in the Purdue LILY greenhouse for spore harvesting.

Plasmid construct and transformation

The DR5v2::ntdTomato reporter was generated by PCR amplifying the DR5v2::ntdTomato fragment from the pGreen-DR5v1::3xGFP-DR5v2::tdtomato [41], followed by cloning into the pCAMBIA1300 vector using PstI and SpeI restriction sites. The ubiquitous nuclear marker, pCrUBQ10::H2B-GFP::3′CrUBQ10 (line 10), used in this study, was previously described [9]. To generate the CrTAA1 transcriptional reporter (pCrTAA1::H2B-GFP::3′CrTAA1), a 2,496 base-pair (bp) promoter and 1,466 bp 3′ terminator of CrTAA1 were introduced into the 5′ and 3′ ends of the H2B-GFP fragment, respectively, in the pENTR vector using enzyme restriction and ligation. The CrTAA1 promoter was amplified from the Ceratopteris genome using the primers: 5′-ACAAGCGGCCGCTTCGAGGACCTAGACCCGAGACTTC-3′ and 5′-ACAAGCGGCCGCAGCTACTACAGCAAAACGCTTAACC-3′. The 3′ terminator of CrTAA1 was amplified using the primers: 5′-TTGGCGCGCCTCGACAGATAAGTTTTGTTCATTCG-3′ and 5′-TTGGCGCGCCTCAATACATGCTCGCTGGGCATTAAG-3′. The entire expression cassette (pCrTAA1::H2B-GFP::3′CrTAA1) was subsequently cloned into the pMOA34 vector for stable transformation.

The CRISPR construct for Ceratopteris was generated using the template from a previously published system for Arabidopsis [51]. Specifically, the modified system contains the CrUBQ10 promoter [9] to drive Cas9 expression and two distinct Ceratopteris Pol III promoters to independently drive gRNAs. For editing the CrTAA1 genomic sequence, two gRNAs specific to CrTAA1 were designed using the website tool (chopchop.cbu.uib.no), with the Ceratopteris richardii genome sequence [46] as the reference. The gRNA1 (5′-AUGGGCUUCAUCAGACCUUC-3′) and gRNA2 (5′-AUAUACUAACUGUUCGGGUC-3′) were positioned 78 bp apart, targeting exon 1 of CrTAA1. The expression cassettes for gRNA1 and gRNA2 were independently generated through Q5 PCR mutagenesis (New England Biolabs) in two separate pCR4 plasmids, CRP01 and CRP02, using the following primers: 5′- tcagaccttcGTTTTAGAGCTAGAAATAGCAAGTTAAAATAAGGCTAG-3′ and 5′- tgaagcccatCGAGTGCAAGGACGCGCG-3′ for gRNA1; 5′- tgttcgggtcGTTTTAGAGCTAGAAATAGCAAGTTAAAATAAG-3′ and 5′-gttagtatatCAACAGCAGCCAATCGCT-3′ for gRNA2. Subsequently, the two gRNA expression cassettes were integrated into a new pCR4 plasmid (CRP03) by enzyme restriction and ligation. The complete gRNA expression cassettes (shown in S22 Fig), along with the pCrUBQ10::Cas9 expression fragment, were introduced into pCAMBIA1300 for stable transformation.

All constructs were transformed into Ceratopteris Hn-n calli using microparticle bombardment, following the previously described procedure [48] Plasmid DNA was coated onto tungsten microparticles and delivered into Ceratopteris calli using the BioRad Biolistic PDS-1000/He particle delivery system at 1,100 psi. Transformed calli were cultured on selection media containing Hygromycin (40 µg/mL) for shoot regeneration. Spores (T1) were harvested from each regenerated sporophyte (T0) and tested again on selection media containing 18 µg/ml Hygromycin B. For fluorescent reporters, multiple independent transgenic lines were generated for each construct, showing comparable expression patterns. All live-imaging experiments were performed on FM or CFM without any selection pressure. For initially screening CrTAA1-CRISPR mutants, leaves from regenerated sporophytes derived from transformed calli were collected for genomic DNA isolation and subsequent genotyping of CrTAA1. Spores from edited sporophytes were then harvested, surface-fertilized, and germinated to produce gametophytes for further genotyping. Individual gametophytes were self-fertilized in 48-well plates, and the resulting sporophytes were genotyped again. CrTAA1 CRISPR mutants confirmed by Sanger sequencing were used for phenotypic characterization. Chromatographs of Sanger sequencing results were visualized using SnapGene Viewer.

Effects of chemical and genetic perturbations of auxin biosynthesis on male-to-hermaphrodite conversion

Ceratopteris spores were surface-sterilized and spread on CFM to induce male differentiation. At 3 DAG, individual male gametophytes were transferred to FM plates (one gametophyte per plate) lacking antheridiogen to induce male-to-hermaphrodite conversion. Live light micrographs of gametophytes were captured at 0, 7, and 10 days after transfer (DAT) using an Olympus CKX53 microscope equipped with a MIchrome 5 Pro camera under a 10x objective lens. To assess the effects of L-kynurenine, a TAA1 inhibitor, on male-to-hermaphrodite conversion, 3-DAG WT male gametophytes were transferred onto fresh FM plates (without antheridiogen), supplemented with either 0.15% Dimethylsulfoxide (DMSO, as a mock control), 50 µM L-kynurenine (with 0.15% DMSO), or 150 µM L-kynurenine (with 0.15% DMSO). The sex type of each gametophyte was determined at 10 DAT, and the sex ratio for each treatment group was calculated as the average of three independent groups, each consisting of 20 independent biological replicates (for a total of 60 individual gametophytes). To assess the genetic role of CrTAA1 in this process, gametophytes from two CrTAA1-CRISPR mutant lines (Line 99 and Line 103, corresponding to crtaa1-1 and crtaa1-2, respectively) and the WT control were imaged and compared under the same conditions. At 10 DAT, the sex type of each gametophyte was determined, with 20 independent gametophytes analyzed per genotype. The genotypes of the CrTAA1-CRISPR lines were confirmed by PCR using primers 5′-AGACAACTTCAGATGCAAAATGA-3′ and 5′-TACGGCAGGCAACTCTGTAGG-3′ and followed by Sanger sequencing. Additionally, to assess the effects of Yucasin (Sigma), an inhibitor of YUC family enzymes [49], on male-to-hermaphrodite conversion, 3-DAG WT male gametophytes were transferred onto fresh FM plates (without antheridiogen) supplemented with either 0.3% DMSO (mock control) or 150 µM Yucasin in 0.3% DMSO. The sex type of each gametophyte was determined at 10 DAT, with 20 individuals analyzed per treatment group.

Imaging auxin signaling dynamics

Confocal time-lapse live imaging was performed to examine the dynamic expression patterns of the DR5 reporter in Ceratopteris gametophytes under different growth conditions. Spores from DR5v2::ntdTomato transgenic lines were surface-sterilized and spread on either FM or CFM. Live imaging was performed every day from 1 to 6 DAG. Hermaphrodites germinated on FM were imaged directly on their original plates, while male gametophytes germinated on CFM were also imaged directly on CFM plates. To examine DR5 expression during male-to-hermaphrodite conversion, spores were surface-sterilized and spread on CFM to induce male differentiation. At 3 DAG, male gametophytes were transferred to FM (one gametophyte per plate), and time-lapse confocal imaging was performed at 0 h after transfer (HAT) and every 12 h from 48 to 216 HAT. tdTomato signals were excited with a 561-nm laser line, and emissions was collected within 561–650 nm range. To assess the effects of L-kynurenine on DR5 expression dynamics during male-to-hermaphrodite conversion, 3-DAG male gametophytes were transferred to FM lacking antheridiogen but supplemented with either 0.15% DMSO (mock control) or 150 µM L-kynurenine containing 0.15% DMSO. Confocal images were taken at 1 DAT and at 24-h intervals from 3 to 9 DAT. To examine the dynamic response of this DR5 reporter to auxin, 3-DAG gametophytes of the DR5v2::ntdTomato reporter line were transferred from FM to FM supplemented with either 0.1% ethanol (mock control) or 10 µM IAA containing 0.1% ethanol. Confocal live images were taken at 0 and 24 h after treatment using identical settings. To analyze CrTAA1 expression patterns during hermaphrodite development, spores from Ceratopteris pCrTAA1::H2B-GFP::3′CrTAA1 transgenic lines were surface-sterilized and cultured on FM plates. Reporter signals were imaged at various DAG in different hermaphrodite samples. To examine CrTAA1 expression patterns during male-to-hermaphrodite conversion, the pCrTAA1::H2B-GFP::3′CrTAA1 spores were spread on CFM to induce male differentiation. At 3 DAG, individual male gametophyte were transferred to an antheridiogen-free FM plate (one sample per plate) to initiate the conversion process. Reporter signals were examined at 3 and 9 DAT in different samples. GFP was excited using a 488-nm laser line, with the detection wavelength set to 490–535 nm.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis

The expression pattern of CrTAA1 (Ceric.10G039800) in WT male gametophytes (MG) and hermaphroditic gametophytes (HG) was determined by qPCR (Fig 6A and S4 Data). Ceratopteris WT spores were surface-sterilized and spread on FM plates. HG and MG were separately collected from the same WT populations at 4, 5, 8, or 10 DAG. Each biological replicate consisted of pooled HG or MG samples, and at least three biological replicates per stage were used for RNA isolation and qPCR analysis. The procedure for culturing, harvesting gametophytes, RNA isolation, and qPCR followed the previously described methods [42]. All gametophytes used for qPCR were cultured in growth chambers (Percival) under continuous light at 29 °C. Total RNA was extracted using the RNeasy Mini Kit (Qiagen), and cDNA was synthesized using the SuperScript Reverse Transcription system (Invitrogen). qPCR was performed using SYBR Green qPCR Master Mix (Selleck Chemicals) on a CFX duet Real-time PCR system (Bio-Rad). Relative expression levels of CrTAA1 were normalized to the internal control CrACTIN1. The qPCR primers for CrACTIN1 were 5′-GAGAGAGGCTACTCTTTCACAACC-3′ and 5′-AGGAAGTTCGTAACTCTTCTCCAA-3′ [42]. The qPCR primers for CrTAA1 were 5′-ACAGCTGAGACATGCAATCGG-3′ and 5′-GACAGCATGACAGTCTTCATCC-3′.

Confocal time-lapse imaging during male-to-hermaphrodite conversion with auxin biosynthesis inhibitor treatment

Spores of pCrUBQ10::H2B-GFP::3′CrUBQ10 were surface-sterilized and spread on CFM to induce male differentiation. At 2 DAG, male gametophytes were transferred to FM lacking antheridiogen but supplemented with either 0.15% DMSO (mock control) or 150 µM L-kynurenine containing 0.15% DMSO. To avoid potential effects of newly released antheridiogen, only one gametophyte was placed per plate. Time-lapse confocal live imaging was performed every 6 h from 0 to 120 HAT. GFP signals were excited with a 488-nm laser line, and emissions were collected within a 490–535 nm detection range. At the final time point (120 h), gametophytes were stained with propidium iodide (PI) to visualize cell outlines, as described previously [9,44,52]. PI was excited using a 561 nm laser line, and emissions were collected in the 596–650 nm range. Confocal image stacks were processed using Fiji/ Image J software to generate maximum intensity z-projection views.

All confocal imaging of Ceratopteris gametophytes was performed using a Zeiss LSM 880 upright confocal microscope with a Plan Apochromat 10×/0.45 objective lens. When necessary, the brightness of each entire confocal image was uniformly adjusted in Fiji/ImageJ to improve visualization.

Lineage tracking and cell division quantification

Nucleus segmentation, lineage tracking, and cell division quantification were performed following the detailed procedure described previously [9,44].

Identification and phylogenetic analysis of TAA1 homologs in Ceratopteris

The identification and phylogenetic analysis of TAA1/TAR1/TAR2 homologs in Ceratopteris richardii were performed as follows: Protein sequences of Arabidopsis AtTAA1 and AtTAR homologs were used as queries for BLASTp searches against the genomes of Ceratopteris and two additional fern species, Azolla filiculoides and Salvinia cucullata. BLASTp searches were performed using Phytozome 13 (phytozome-next.jgi.doe.gov) and Fernbase (https://fernbase.org/), with an E-value threshold set at 10⁻¹⁶. Previously published TAA1 and TAR homologs from selected species (listed in S3 Data) were obtained from Carrillo-Carrasco and colleagues [45] via the Phytozome 13 database. Multiple sequence alignments were carried out using MAFFT (https://www.ebi.ac.uk/jdispatcher/msa/mafft), and maximum-likelihood phylogenetic trees were constructed using IQ-TREE v2.4 (http://www.iqtree.org/) with 100,000 ultrafast bootstrap replicates for statistical support. The phylogenetic tree was visualized using Figtree (http://tree.bio.ed.ac.uk/software/figtree/). Protein domain identification for each clade was performed using INTERPROSCAN [53], which was also used to confirm the identities of protein domains in shorter BLASTp hits (less than 200 amino acids) that passed the E-value threshold.