A fern WUSCHEL-RELATED HOMEOBOX gene functions in both gametophyte and sporophyte generations

Background Post-embryonic growth of land plants originates from meristems. Genetic networks in meristems maintain the stem cells and direct acquisition of cell fates. WUSCHEL-RELATED HOMEOBOX (WOX) transcription factors involved in meristem networks have only been functionally characterized in two evolutionarily distant taxa, mosses and seed plants. This report characterizes a WOX gene in a fern, which is located phylogenetically between the two taxa. Results CrWOXB transcripts were detected in proliferating tissues, including gametophyte and sporophyte meristems of Ceratopteris richardii. In addition, CrWOXB is expressed in archegonia but not the antheridia of gametophytes. Suppression of CrWOXB expression in wild-type RN3 plants by RNAi produced abnormal morphologies of gametophytes and sporophytes. The gametophytes of RNAi lines produced fewer cells, and fewer female gametes compared to wild-type. In the sporophyte generation, RNAi lines produced fewer leaves, pinnae, roots and lateral roots compared to wild-type sporophytes. Conclusions Our results suggest that CrWOXB functions to promote cell divisions and organ development in the gametophyte and sporophyte generations, respectively. CrWOXB is the first intermediate-clade WOX gene shown to function in both generations in land plants. Electronic supplementary material The online version of this article (10.1186/s12870-019-1991-8) contains supplementary material, which is available to authorized users.


Background
Stem cells are self-renewing pluripotent cells. In vascular plants, they are located in the shoot apical meristem (SAM) and root apical meristem (RAM). Stem cells divide at a low frequency to produce daughter cells that will either maintain the stem cell pool or actively divide and take on new identities to form new organs [1,2]. The size of a stem cell population varies among different species and is strictly maintained as a part of the meristem [3,4]. Failure to coordinate multiple inter-and intra-cellular signals disrupts development and results in altered plant body architecture [4,5]. In addition to hormonal signals, inter-cellular signaling is mediated by small peptide ligands and their cognate receptors. These signals converge to regulate specific transcription factors to achieve a balance among the populations of stem cells, the faster dividing cells, and the differentiating cells of the meristem [6,7]. In Arabidopsis thaliana, the homeobox transcription factor WUSCHEL (WUS) is a key player in shoot meristem maintenance; WUS expression is transcriptionally regulated and the protein acts non-cell-autonomously by moving from the organizing center (OC) to the central zone (CZ) of the SAM to both activate and repress gene transcription in order to maintain meristem cells in a pluripotent state [8].
WUS belongs to the family of WUSCHEL-RELATED HOMEOBOX (WOX) transcription factors, which are characterized by the presence of a conserved homeodomain [9,10]. Phylogenetic analyses of land plant WOX genes group members into three clades: ancient, intermediate, and modern ( [11] Additional file 1: Figure S1). The progenitor of WOX genes existed in the last common ancestor of land plants and green algae and, through successive gene duplication and functional diversification, gave rise to the three clades of WOX genes [9,12]. All land plants that have been examined, nonvascular and vascular alike, possess WOX genes of the ancient clade, while the intermediate clade only exists in vascular plants, and the modern clade is found in seed plants and ferns, but has not been found in lycophytes [13,14]. Based on the presence of two subgroups of the intermediate clade in the lycophytes and sequence relatedness of only one subgroup to the modern clade WOX genes, it has been proposed that the intermediate subgroup shared a progenitor with the modern clade [11,14]. The modern clade, or the WUS clade, has experienced further expansion in seed plants as Picea abies possesses five and A. thaliana possesses eight WUS clade members [9,15], compared to the single member found in the fern, Ceratopteris richardii [11]. Modern clade WUS proteins contain, in addition to the canonical homeobox, the WUS box (TL-LFPMILV) [9]. Both conserved domains are required for meristem maintenance in the A. thaliana SAM [16]. Maintenance of the SAM [17] and RAM [18,19] is under the control of AtWUS and AtWOX5, respectively, and in addition, AtWOX4 functions in the vascular cambium stem cells [20,21] where continually dividing cells produce phloem and xylem during secondary growth. All other AtWOX genes of the three clades play roles in early embryo development or in organ development, including leaf, root, and floral organs [12,[21][22][23][24][25][26][27][28].
The WOX gene family has been widely studied in land plants, including C. richardii, but functional studies are limited to seed plants such as A. thaliana (e.g. [9,16],), Oryza sativa (e.g. [29],), P. abies [30,31] and moss Physcomitrella patens [32]. Five WOX genes, CrWOX13A and CrWOX13B of the ancient clade, CrWOXA and CrWOXB of the intermediate, and CrWUL of the modern clades have been identified in C. richardii [11]. RT-PCR results showed that CrWOX13A and CrWOXB are equally expressed in all tissue examined, including root tip, gametophyte, and young sporophyte; whereas CrWOXA was expressed more strongly in the root tip and CrWUL in the root tip and gametophyte [11]. In situ hybridization analyses of the latter two genes showed localized expression. CrWOXA is expressed in the root apical cell and in the lateral root apical cell. In addition to expression in the vascular bundle of the leaves [33], CrWUL mRNA is localized to the cutting edge of the lateral root apical cell which divides asymmetrically, proximal to the main root axis [11]. CrWOXB shows a broad expression pattern in the root tip, consistent with the high levels of expression detected by RT-PCR [11].
The unbiased expression of CrWOX13A and CrWOXB in both gametophyte and sporophyte generations presents an opportunity to understand the ancestral functions of WOX proteins. In the moss P. patens, only the ancient clade of WOX genes exists, and, in contrast to A. thaliana ancient WOX genes, P. patens ancient WOX genes function in both generations [32]. Only two AtWOX genes, AtWOX2 and AtWOX8, are expressed in both sporophyte and gametophyte generations [25]; all other AtWOX genes seem to function only in the sporophytes [12,[21][22][23][24][25][26][27][28]. Interestingly, in Nicotiana tabacum, transcripts of two ancient and one intermediate WOX genes are found in both the gametophyte and sporophyte tissues [34].
Sister clade to seed plants, ferns have sporophytic SAMs that are composed of multiple zones resembling that of the seed plants [35,36]. How these zones are involved in stem cell maintenance and organ initiation is unclear. Moreover, how the fern gametophyte notch meristem is maintained is completely unknown. Thus, WOX genes provide an entry point for understanding the meristem of ferns at both the development and the evolution levels. So far, the combination of WOX gene family evolution and their developmental functions has only been studied in detail in the moss P. patens and angiosperms. Similar investigation in the fern will bridge the gap in our knowledge of meristem evolution. Furthermore, comparison between the gametophyte and the sporophyte meristems within the fern will provide insight into the co-option of gene network between the meristem of the two generations. This understanding can only be fully realized with the expression and functional analyses of all five WOX fern genes. Here, we present the completed study of one of the five WOX genes found in C. richardii, CrWOXB, which is expressed in both sporophyte and gametophyte generations [11], to examine its expression in the meristems of sporophyte shoot and gametophyte using sectioned and whole mount in situ hybridization, respectively. The possible function of CrWOXB in both generations was examined by RNAi suppression of CrWOXB expression in transgenic C. richardii plants. These results show that CrWOXB, an intermediate-clade WOX gene, is expressed in regions of cell proliferation in both the gametophyte and sporophyte. The phenotypes of the RNAi suppression lines were consistent with meristem defects, providing the first demonstration of WOX gene function in a fern.

Results
CrWOXB is expressed in regions of cell division in both the gametophyte and sporophyte generations The expression of CrWOXB was observed in developing gametophytes and in sporophyte leaves. CrWOXB was expressed at low but discernable levels before (d8 and d10) sexual maturation and increased in gametophytes at sexual maturity (d13) (Fig. 1a). Whole-mount in situ hybridization revealed that CrWOXB mRNA was expressed in recently germinated gametophytes (Fig. 1b, c) and then in the notch meristem region in gametophytes before (d8, Fig. 1d, e; day 10, Fig. 1f, g) and after (d13, Fig. 1h, i) sexual maturation. Consistent with increased expression at d13 detected by RT-PCR analysis (Fig. 1a), d13 gametophytes showed a broader area of CrWOXB expression than d8 and d10, that was also further away from the notch region. In addition, CrWOXB expression was also detected in developing archegonia (Fig. 1d, f red arrows) suggesting a role in organ specification in hermaphrodites. At d8, similar to hermaphrodites, male gametophytes expressed CrWOXB in cells prior to antheridia development (Fig. 1j). The expression declined in d13, when most cells have developed into antheridia ( Fig. 1l). Thus, the expression of CrWOXB in gametophytes is in actively dividing region in both hermapthrodites and males and the notch meristem and the archegonium of hermaphrodites. The first set of leaves (total of 16-21 leaves) formed by C. richardii sporophytes are vegetative (do not produce sporangia), followed by an indeterminate number of sporophylls (produce sporangia). Pinnae refer to the leaflets of both vegetative and sporophyll leaves. The expression of CrWOXB in sporophyte tissues detected by RT-PCR showed the highest expression in sporophytes with one fully expanded leaf and in the later vegetative leaves but was barely detectable in the sporophylls, the pinnabase and tip of the vegetative leaves (Fig. 2a). CrWOXB mRNA was detected by in situ in young vegetative leaves (Fig. 2b), the shoot tip (Fig. 2d), leaf vascular bundles (Fig. 2f), and in root primordia ( Fig. 2h) of young sporophytes bearing 10-11 vegetative leaves, when the first visible fiddlehead was observed. Notably CrWOXB expression is seen within the apical cells of leaf primordia and more mature leaf ( Fig. 2b, d, green arrowhead) but not in the apical cell of the SAM (Fig. 2d, purple arrowhead). Thus, CrWOXB expression is localized to regions of actively dividing cells in the sporophyte generation.

CrWOXB is required for proper growth of gametophytes
To carry out functional analyses, we created CrWOXB RNAi suppression, crwoxb, lines, by using Agrobacterium-mediated transformation of gametophytes. Suppression of CrWOXB transcripts were quantified in young sporophytes with 6-7 fully expanded vegetative leaves by RT-qPCR. CrWOXB expression in T 2 sporophyte of crwoxb lines were found to express a range of levels from a high average of~40% in crwoxb10 to~16% in crwoxb1 compared to wild-type plants (Fig. 3a). The variation is likely due to position effect of the transgene [37]. No plants were recovered with undetectable CrWOXB expression.
Consistent with expression in the meristem and other regions of cell division ( Fig. 1b- Figure S2a-d inset) between the two lobes of the thallus. The wider notch appeared to be the result of a localized combination of fewer trapezoidal meristem cells and altered cell division planes, which prevented the lobes of the gametophyte from growing together. To quantify the gametophyte size, gametophyte nuclei were stained and counted. Wild-type and crwoxb lines produced similar numbers of gametophyte cells prior to d8. After d8, development of crwoxb lines is delayed by 1 day (Fig. 3f; Additional file 5: Table S2). The gametophyte notch meristem is generally formed from d7 to d8. The average numbers of cells produced by crwoxb gametophytes were less than wild-type gametophytes and the difference increased with time ( Fig. 3f; Additional file 5: Table S2).
Archegonia house the eggs and are direct derivatives of the notch meristem in C. richardii [38]. Because the gametophytes of crwoxb lines had fewer cells, we hypothesized that they would also develop fewer archegonia. To test this, we compared numbers of archegonia in wild-type and the crwoxb lines and found that indeed the crwoxb lines produced fewer archegonia than wildtype plants ( Fig. 3g; Additional file 6: Table S3). The archegonia of the crwoxb lines were functional as they produced sporophytes. The reduction in archegonia numbers could be due to fewer cells of crwoxb gametophytes or to involvement of CrWOXB in the specification of archegonia progenitor cells. To distinguish between these two possibilities, we compared the number of archegonia to number of cells in the entire gametophyte of wild-type and crwoxb lines (Additional file 6: Table S3). Thirteen-day-old crwoxb gametophytes, although having fewer cells, had on average 55 more, not fewer, cells for each archegonium than wild-type gametophytes. This result rules out the first but not the second scenario.

CrWOXB promotes leaf development in the sporophyte generation
In situ hybridization revealed localized expression of CrWOXB in leaf primordia and developing leaves in the sporophyte (Fig. 2b, d), consistent with a role of leaf initiation. These results prompted an examination of leaf initiation and development in the crwoxb lines. In the wild type, 16-21 vegetative leaves are formed, followed by sporophylls (Fig. 4e). In contrast, crwoxb lines initially produced fewer vegetative leaves (Fig. 4e) before producing sporophylls. The number of sporophylls produced seemed to be comparable to the wild-type as both continue producing sporophylls indefinitely; however, the number of pinnae of each sporophyll was greatly reduced compared to that of wild-type plants (Fig. 4f).
These results indicate a role of CrWOXB in leaf morphogenesis during both the vegetative and reproductive phases of the sporophyte generation.
CrWOXB promotes root and lateral root initiation during sporophyte development The expression of CrWOXB in root primordia (Fig. 2h) and during lateral root formation suggested a role for CrWOXB in root initiation and/or development. Sporophytes with 6-7 fully expanded leaves were grown in liquid media for 14 days before root observation. Wildtype sporophytes (Fig. 5a) exhibited more roots and were more branched than crwoxb lines (Fig. 5b-d). When quantified, the average number of roots (Fig. 5e) and lateral roots per plant (Fig. 5f) were significantly reduced in crwoxb lines compared with wild-type plants. Similar to leaf initiation and development, the significant reduction of root and lateral root numbers further confirmed the role of CrWOXB in organ initiation.

Discussion
The WOX genes, especially the modern clade member WUS, are well studied in angiosperms. Considering their important roles in meristem maintenance, how these genes function in the ferns will help the understanding of fern meristems and their maintenance. Here we presented the first functional analysis of a fern WOX gene, CrWOXB, and showed its role in both gametophyte and sporophyte generations.

CrWOXB functions in both the gametophyte and sporophyte generations
Reduced cell divisions in hermaphroditic gametophytes of crwoxb lines suggested that CrWOXB promotes cell division, mirroring intermediate WOX proteins in A. thaliana and P. abies where these proteins activate cyclin genes, which regulate the progression of the cell cycle [30,39,40]. In addition to CrWOXB's function in cell division, in the hermaphrodites CrWOXB also seemed to play a role in specifying cells to become archegonia ( Fig. 1g; Additional file 6: Table S3) in the region where CrWOXB is expressed highly. The reduced number of archegonia in crwoxb lines could be explained by noncell autonomous action of CrWOXB, where decreased expression in crwoxb lines would need more cells to produce some threshold concentration for specification. Once specified, the egg-cell develop normally: its maturation and embryo development were unaffected in crwoxb lines, based on the observation that sporophytes formed after 5 days post-fertilization in both wild-type and crwoxb gametophytes. In the male, cell proliferation is followed closely by differentiation into antheridia [41,42]. We detected CrWOXB in cells before, but not after  N ≥ 11 fronds). Scale bar = 9.3 cm. One-Way ANOVA (*, p < 0.05; **, p < 0.01; ***, p < 0.001 df = 24) (a-d). Scale bar = 4.65 cm (a-d inset). Error bars represent standard deviation differentiation into antheridia in d8 gametophytes (Fig. 1j, l).
In the sporophyte generation, abnormal phenotypes were observed in both the shoot and the root of crwoxb lines. In the shoot, the number of sterile fronds and pinna of the fertile fronds was diminished and similarly in the root, both root and lateral-root numbers were decreased. Therefore, we conclude that CrWOXB functions to promote cell division and possibly to specify organ formation in both generations of C. richardii.
CrWOXB functions in both generations, whereas its ortholog in A. thaliana, AtWOX9, has only been shown to function in sporophytes [27,43]. Prior to this work, only the ancient clade of WOX genes have been shown to function in both the gametophyte and the sporophyte generations of P. patens [32]. The trend of diminishing WOX gene function in the gametophytes during evolution is consistent with the comparative transcriptome profiling between the moss Furnaria hygrometrica and A. thaliana, in which an enrichment of bryophyte gametophyte-biased transcription factors are found in sporophyte-biased (and sporophyte-specific) A. thaliana orthologs [44].

The role of CrWOXB in the meristems of gametophytes and sporophyte
In the gametophyte, CrWOXB was expressed in both the male and hermaphrodite. Expression in the male persisted briefly, during thallus growth, before cells differentiate into antheridia. Similarly, in the hermaphrodite, CrWOXB was expressed shortly after spore germination; however, its expression appeared in the notch region, during and after the emergence of the lateral meristem. The expression pattern of CrWOXB in the male and hermaphrodite is in agreement with cell proliferation regions delineated by [42]. Our results suggest that CrWOXB function is required soon, if not immediately, after spore germination.
All growth in C. richardii sporophytes, as in other ferns, can be traced back to single, apical cells [45]. The shoot apical cell of C. richardii sits atop a slender stalk of meristem cells in a region defined as the proliferation  [36]. Leaf initiation commences with the specification of one of the peripheral shoot meristem cells as a leaf apical cell, which persists throughout leaf development [46,47]. Roots similarly are formed by the persistent action of a root apical cell [48]. Expression of CrWOXB is homogenous in the shoot, root, leaf primordia, and the vascular tissues signifying a more general role of CrWOXB in each region of cell proliferation. The homogenous expression pattern of CrWOXB in the primordia is similar to what was observed of AtWOX9 in Arabidopsis shoot apical meristem [9,40]. The expression pattern of CrWOXB in the root primordia is similar to that in the mature root tip reported by Nardmann et al. [11], and is expressed in tissues outside of where CrWOXA, the other intermediate CrWOX gene, is expressed, suggesting some functional divergence between these two paralogs. Interestingly, CrWOXB expression was not detected in apical cell of the shoot and root but was observed in the leaf primordia (Fig. 1h). We have consistently observed this difference, but the significance is unclear. The CrWOXB may be regulated differently in the shoot and root apices and in the leaf primordia.
It may not be surprising to find that CrWOXB played a role in both gametophytes and sporophylls. As observed by Hagemann [60], the pinnae of sporophyll and the gametophytes of the ferns are of structural similarity as both are dorsiventral with marginal meristematic growth and produce abaxial reproductive organs.

The relationship between CrWOXB and the intermediate WOX transcription factor family
In A. thaliana, the homeodomain of AtWOX8 and AtWOX9 can partially rescue meristem function in a wus-1 background, which establishes the intermediate WOX homeodomain as a key motif for meristem function [16]. Outside of the homeodomain, seed plant intermediate WOX members contain conserved N-terminal and C-terminal motifs that are not shared with intermediate clade proteins in C. richardii, CrWOXA and CrWOXB (Additional file 3: Figure S3). Despite the divergence outside the homeodomain, phenotypes of crwoxb in the shoot and root are reminiscent of AtWOX9 null-mutant seedlings which fail to form leaves, secondary shoots and lateral roots [27]. The presence of the C-terminal domain and N-terminal motifs may be required for embryo patterning and development of the suspensor in A. thaliana because, in a complementary experiment, AtWUS, which does not contain the N-and C-terminal motifs of the intermediate clade, cannot rescue embryo arrest in Atwox8 Atwox9 double mutants [16]. Therefore, in C. richardii, the homeodomain is the most likely motif involved in cell proliferation and organ specification in the gametophyte and sporophyte, while the divergent N-and C-terminal sequences may contain yet-to-be recognized motifs that have additional function during embryogenesis.

Conclusion
We have functionally characterized an intermediate clade WOX protein CrWOXB throughout gametophyte and sporophyte development in the fern model C. richardii and found that CrWOXB is expressed in proliferating tissues of both generations. Knockdown crwoxb lines produce fewer gametophyte cells, and smaller sporophytes with fewer sporophyte organs, suggesting a conserved function in gametophytes and sporophytes despite their different architecture. Methods and results presented here serve as model for the analysis of the remaining WOX genes in C. richardii in order to understand how this gene family has diversified its functions in proliferative regions of the gametophyte and sporophyte generations.

Plant growth conditions
Spores of C. richardii strain Rn3 (wild-type) were originally obtained from Carolina Biological Supply Company (Burlington, NC). Wild-type and CrWOXB RNAi suppression lines (crwoxb) were surface sterilized in 4% sodium hypochlorite and 0.5% Tween-20 for 5 min, rinsed 4-5 times with sterile water and incubated at room temperature in the dark for 3-5 days to synchronize germination. Spores were then plated on basal media (1/2 MS, pH 6.0) supplemented with 100 μg ml − 1 ampicillin and maintained in humidity domes at 26°C with a light/ dark cycle of 16/8 under light intensity of 100 μM m − 2 s − 1 for gametophyte development. Plates were inverted after 10 days of growth (d10) to deter fertilization. Sporophytes were grown in BLP germination soil #1 (Beautiful Land Products, West Branch, IA) under humidity domes in the same light and temperature regime as gametophytes.

Transformation of C. richardii gametophytes
A 302-bp fragment (See Additional file 4: Table S1 for primer sequences) of CrWOXB was cloned into vectors pK7GWIWG2 and pH7GWIWG2 to generate CrWOXB RNAi constructs using the Gateway technology as described by Curtis and Grossniklaus [49] and Bui et al. [50]. Each construct was introduced into Agrobacterium tumefaciens strain GV3101 from Escherichia coli with an E.coli helper strain containing the pRK 2013 plasmid [51]. Stable transformation of young gametophyte tissue was conducted as described previously [52]. Successfully transformed gametophytes (T 0 ) were selected on media containing 50 μg ml − 1 kanamycin or 5 μg ml − 1 hygromycin. Resistant gametophytes were isolated and allowed to self-fertilize to produce sporophytes (T 1 ). Sporophytes were moved to liquid basal media and allowed to root before transplanting to soil. From the more than 20 independent transgenic lines isolated, 10 were chosen for qPCR analysis and characterization. Detailed phenotyping of three lines are presented here.

Whole-mount and sectioned in situ hybridization
Antisense and sense RNA probes used for in situ hybridization experiments were synthesized from 1 μg of PCR products amplified using primers containing T7 promoter sequences (Additional file 4: Table S1) with T7 RNA polymerases (Agilent, Santa Clara, CA), and DIG RNA labeling mix (Roche Diagnostics, Indianapolis, IN). DIG-labeled RNA probes were precipitated in 2.25 M LiCl overnight at − 20°C, before resuspension in nuclease-free water. RNA concentration was measured with a Nanodrop One (Thermo-Scientific, Waltham, MA) and then diluted 1:1 with deionized formamide and stored at − 20°C.
The SAM from young sporophytes with 10-11 vegetative leaves, the youngest a visible fiddlehead, was dissected and vacuum infiltrated with fixing solution (4% paraformaldehyde in 1x PBS) for 45 min and then incubated in fixing solution overnight at 4°C. Dehydration, embedding, pre-hybridization, hybridization and posthybridization washes were based on Jackson [59], except that acetic anhydride washes were omitted from prehybridization. Embedded tissues were sectioned at 8 μm thickness with a rotary microtome. Probe detection and color development protocols were based on Ambrose et al. [58]. Whole-mount in situ was adapted from the protocol of Ambrose et al. [36,58], with the following modifications. Gametophytes were fixed in FAA (formaldehyde: ethanol: acetic acid, 3.7%:50%:5% v/v respectively) at room temperature for 1 h, then stored in 70% ethanol at − 20°C. Fixed gametophytes were processed without Histoclear II. Color development of wholemount in situ tissues was stopped in ddH 2 O and mounted in 50% glycerol. Whole-mount samples were viewed with a Zeiss compound light microscope and imaged with the Zeiss Axiocam ERc 5 s digital camera (Carl Zeiss Microscopy LLC, Thornwood, NY). DIC images of sectioned samples were viewed with a Nikon Eclipse E800 (Nikon Instruments Inc., Melville, NY) and captured with a Photometrics CoolSNAP cf. (Photometrics, Tucson, AZ). To confirm gene expression patterns, each in-situ experiment was repeated at least two times using different biological samples.

RNA extraction and RT-PCR analyses
Gametophyte and sporophyte tissue were harvested and flash frozen in liquid nitrogen, then stored at − 70°C. Total RNA was extracted from frozen tissue with the Quick-RNA MiniPrep (Plus) kit (Zymo Research, Irvine, CA) and 750 ng of gametophyte total RNA or 500 ng of sporophyte total RNA was used in reverse transcriptase reaction using MMLV (New England Biolabs, Ipswich, MA) with N9 random primers (IDT Coralville, IA). PCR was conducted with the following cycles: 2 min at 94°C, followed by 37 cycles of 30s at 94°C, 30s at 59°C, and 30s at 72°C, with a 5 min final extension time at 72°C for CrWOXB, and 25 cycles under the same conditions for CrUBQ transcripts.
For RT-qPCR, three biological and two technical replicates were performed for each line. Total RNA from whole young sporophytes with 6-7 fully expanded round leaves was extracted and 200 ng were used in cDNA synthesis as described above. Due to delay in development of crwoxb lines, the age of the sporophytes in both the wild-type and crwoxb lines was determined by numbers of leaves and not days. Primers for qPCR are listed in Additional file 4: Table S1. Detection of amplification was performed using SYBR green chemistry (Roche Diagnostic, Indianapolis, IN) with the Roche LightCycler 480 Real-Time PCR system (Roche Diagnostic). The PCR cycle was as follows: 10 min at 95°C, followed by 45-55 cycles of 10s at 95°C, 10s at 62°C, and 20s at 72°C, with a single fluorescence read at the end of each extension time. A melting curve analysis was also performed and analyzed using the Tm calling software module to verify the absence of primer dimers and nonspecific products. Calibrator normalized relative quantification was performed using the 2nd derivative maximum algorithm with three internal relative standards. CrWOXB expression was measured relative to CrUBQ.

Phenotypic analysis of crwoxb lines
To count cells of the gametophytes, gametophytes were cleared overnight in 100% ethanol at 4°C, then rinsed 3 times for 5 min in water and stained with Hoechst 33342 (40 μg ml − 1 ) (Invitrogen, Carlsbad, CA) for at least 15 min, rinsed in water and mounted on slides with 50% glycerol. Gametophytes were then imaged with a Leica stereomicroscope and a Qicam camera (Qimaging, Surrey, BC, Canada) with a DAPI filter. Nuclei of gametophytes were counted in Photoshop CC (Adobe systems, San Jose, CA). Brightness and contrast were increased slightly to facilitate cell counting.
For root and lateral root counts, spores of wild-type and crwoxb lines were grown on basal media for 13 days, after which individual hermaphroditic gametophytes were isolated for self-fertilization by adding a few drops of water. Resulting sporophytes were transferred to 100 ml of liquid basal media with 100 μg ml − 1 ampicillin and grown for an additional 2 weeks before roots and lateral roots were counted. Vegetative leaves and pinnae were counted on soil-grown sporophytes when each sporophyte had 5-7 sporophylls.

Statistical evaluation of the data
Statistical analyses of CrWOXB levels in crwoxb lines, gametophyte archegonia numbers, and sporophyte phenotypes were conducted with one-way ANOVA, while gametophyte cell numbers were conducted with twoway ANOVA. Both analyses were followed by Dunnett's multiple comparisons test. All calculations were done in GraphPad Prism version 8.0.1 (GraphPad Software, San Diego, CA).