# Source paper archive

- Source title: Robust planar morphogenesis in leaf-like organs requires callose-dependent coordination of division and growth
- Author(s): Lin, W.; Hallet, S.; Cifrova, P.; Deshpande, M.; Gregoire, D.; Lacroix, A.; Pleskot, R.; Smith, R. S.; Kierzkowski, D.; Coudert, Y.
- Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.05.764497v1
- PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.05.764497v1.full.pdf
- PDF SHA-256: a6e6dfafb13ac1d0c33d9ed17050d609a0b2fef74d8d15d23abc13510797deec
- Converter: pymupdf4llm 0.0.27
- Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version.

---

[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)

# **Robust planar morphogenesis in leaf-like organs requires** **callose-dependent coordination of division and growth**


Wenye Lin [1,5], Stéphanie Hallet [2,#], Petra Cifrova [3,#], Mandar Deshpande [1], Daniel Grégoire [1], Alexis
Lacroix [2], Roman Pleskot [3], Richard S. Smith [4], Daniel Kierzkowski [1,@] and Yoan Coudert [2,@]


1 Institut de Recherche en Biologie Végétale, Département de Sciences Biologiques, Université de Montréal,
Montréal, QC, H1X 2B2 Canada

2 Laboratoire Reproduction et Développement des Plantes, Université de Lyon, ENS de Lyon, UCB Lyon 1,
CNRS, INRAE, INRIA, Lyon 69007, France

3 Institute of Experimental Botany, Czech Academy of Sciences, Prague 16500, Czech Republic

4 Computational and Systems Biology, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK

5 Current address: Laboratory of Biochemistry, Wageningen University and Research, Stippeneng 4, 6708
WE Wageningen, Netherlands

# denotes equal contribution

@ [Correspondence: daniel.kierzkowski@umontreal.ca,](mailto:daniel.kierzkowski@umontreal.ca) [yoan.coudert@cnrs.fr](mailto:yoan.coudert@cnrs.fr)


**Summary**


Planar organs, such as leaves, are key developmental innovations underlying the evolutionary success of


plants. In vascular plants, building a planar organ requires coordinating cell division and growth both within


the organ plane and across multiple, mechanically distinct cell layers. Phyllids, the leaf-like organs of


bryophytes, typically consist of a single cell layer and offer a minimal system to decouple in-plane growth


coordination from inter-layer mechanics. Here, we investigate how robust planar morphogenesis is


achieved in phyllids, using the moss _Physcomitrium patens_ as a bryophyte model. We show that impaired


cytokinesis, caused by the loss of a callose synthase gene, disrupts growth gradients, ultimately yielding


smaller, non-planar organs of variable shape. Combining quantitative live imaging with computational


modeling, we show that the severity of shape defects in mutants depends on the spatial distribution and


developmental timing of local growth arrests. These findings reveal how callose-dependent coordination


between cell division and in-plane growth prevents mechanical conflicts and maintains planarity in a


monolayered plant organ.


**Keywords**


_Physcomitrium_, moss, cytokinesis, callose synthase, leaf, phyllid, growth, mechanical conflicts


1


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


**Introduction**


Plant organs acquire their diverse three-dimensional shapes through differential growth, _i.e.,_ spatial and


temporal variations in expansion rate, anisotropy, and direction across cells and tissues (1, 2). Because


plant cells are encased within a rigid cell wall and cannot migrate, growth differences between cells and


tissues can generate mechanical conflicts (3, 4). These growth-derived conflicts can be critical drivers of


organ shape development (5–7), but they also need to be locally buffered to maintain planar geometries


(8–10). Developing and maintaining a flat organ requires strict coordination of growth at multiple scales:


between the adaxial and abaxial sides of the organ, between distinct anatomical regions such as the midrib,


petiole, lamina (blade) and margin, and within the plane of the organ itself, where gradients in growth rate


or anisotropy can generate compressive stresses sufficient to cause buckling or wrinkling (11–13). Our


understanding of how this coordination is achieved has advanced considerably through the study of


_Arabidopsis thaliana_ leaves and sepals as model systems (9, 14–19). However, these organs are


multilayered structures comprising distinct tissue types, each with different mechanical properties, growth


rates, and cellular architectures, which makes it difficult to isolate the contribution of individual cellular


processes to overall organ planarity.


Phylids, the leaf-like organs of bryophytes, evolved planar shapes independently of vascular plant leaves.


In the moss _Physcomitrium patens_, the phyllid lamina is a one-cell-thick layer, which makes cell division


and growth directly observable and offers a minimal framework for dissecting planar morphogenesis (20).


Despite this major structural difference from vascular plant leaves, recent work in _P. patens_ has uncovered


strikingly similar morphogenetic principles (21). Phyllids undergo an early proliferative phase driven by a


basal zone of cell division, followed by cell expansion as cells exit the division zone. Cells of the developing


midrib cease anticlinal divisions and elongation before lamina cells, while spatial and temporal growth


gradients across the lamina collectively determine the final organ shape (21). Time-lapse imaging combined


with quantitative cellular analyses has established the phytohormone auxin as a major positional cue


promoting cell expansion and differentiation (21–24). Despite the identification of various molecular


regulators of phyllid development (20, 23, 25–29), the mechanism that coordinates cell division with growth


gradients remains unknown.


A direct way to test how cell division contributes to planar morphogenesis is to perturb it and examine the


consequences for organ shape. Cytokinesis – the final stage of cell division – ends with the centrifugal


expansion of the cell plate until it fuses with the parent wall (30, 31). Callose, a β-1,3-glucan polysaccharide


synthesized by callose synthase (CALS) proteins (also called glucan synthase-like, GSL), accumulates


transiently at the expanding cell plate and scaffolds the deposition of other cell wall components, so


reducing callose synthesis impairs wall completion (32–37). In _A. thaliana_, _cals10/gsl8_ mutants fail to


complete wall closure and show severe alterations of leaf shape and flatness, but the multilayered structure


of the leaf makes it difficult to relate defective cytokinesis in individual cells to altered planarity (38–40).


_CALS_ genes are conserved across land plants, with 12 putative members in the moss _P. patens_, and their


2


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


role in cytokinesis likely predates land plants, as suggested in the streptophyte alga _Penium margaritaceum_


(41–44). This evolutionary conservation makes CALS-dependent cytokinesis a tractable entry point to


interrogate how cell division shapes planar organs. Here, taking advantage of this conserved feature, we


investigate how the coordination between cell division and growth drives robust morphogenesis in the moss


phyllid.


**Results**


_**Physcomitrium patens CALS6**_ **regulates phyllid shape robustness**


To explore the role of callose in cell division during phyllid development, we stained dissected _P. patens_


gametophore (leafy shoot) apices with aniline blue and imaged the fluorescence in emerging phyllids by


confocal microscopy. At the 2-2.5-day stage, the signal coincided with cell walls across the whole organ,


whereas at the 3-day stage it was mostly restricted to the proximal region, which corresponds to the zone


of active cell division (21). This dynamic pattern indicates that callose is actively deposited and then


removed as phyllids grow (Figure 1A-1B).


Next, we used a genetic approach to alter callose biosynthesis. Among the five _CALS_ genes expressed in


phyllids and gametophores, _CALS6_ had the highest expression levels in all conditions tested (Figure 1C).


To examine its function, we disrupted its coding sequence by CRISPR-Cas9 genome editing. Using four


sgRNAs, we deleted a chromosome fragment of more than 18 kb spanning nearly the entire _CALS6_ coding


sequence. The deletion introduced a frameshift at codon 20 and was predicted to yield a truncated protein


of 29 amino acids instead of the 2,020 amino acids of the wild-type protein (Figure 1D).


To assess the developmental consequences of loss of _CALS6_ function, we compared the gross morphology


of _cals6_ mutants with that of wild-type controls (Figure 1E, 1H). Gametophores of _cals6_ mutants were


stunted, with an average length more than 50% shorter than in wild type (Figure 1F, 1I; Figure S1). In wild


type, phyllids are heteroblastic: their length increases gradually during the juvenile-to-adult transition and


stabilizes around the 10 [th] position from the base, and adult phyllids have a lanceolate shape (Figure 1G,


1K) (22, 45). In _cals6_ mutants, phyllids were shorter at all positions, and the heteroblastic gradient was less


pronounced, with length reaching a plateau around the 4 [th] position (Figure 1K). Mutant phyllids were also


more irregular in shape, with crinkled margins, as reflected by a higher excess length ratio, defined as the


phyllid perimeter relative to its length (Figure 1J, 1L) (46). Independent lines generated with the same


sgRNAs, but carrying different mutations showed similar phenotypes, supporting the conclusion that loss


of _CALS6_ function causes these defects (Figure S1). Together, these results identify _CALS6_ as an essential


regulator of phyllid morphogenesis in _P. patens_ .


_**CALS6**_ **regulates tissue planarity in phyllids**


To elucidate the cellular basis of the morphological defects of _cals6_ mutants, we segmented individual cells


in mature adult phyllids using MorphoGraphX and quantified their area and curvature. The proximal–distal


3


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


gradient of cell area typically observed in wild type was disrupted in _cals6_ phyllids, which showed frequent


patches of juxtaposed small and large cells (Figure 2A). In addition, whereas wild-type cells were


predominantly elongated and aligned along the proximal–distal axis, _cals6_ mutant cells were frequently


irregular in shape and smaller overall, suggesting impaired growth (Figure 2I). In wild type, cell curvature


followed a consistent medio-lateral gradient, with negative values along the midrib, near-zero values across


the blade, and positive values along the margins (Figure 2B, 2J). Cross-sections confirmed that the two


halves of the blade flanking the midrib were flat at the tissue level, consistent with the distribution of cell


curvature values along the proximal-distal axis. In contrast, the medio-lateral curvature gradient was


severely impaired in _cals6_ mutants (Figure 2B). Although margin cells retained an overall positive curvature


similar to wild type, cell curvature across the blade and midrib was irregular, resulting in greater variability


in curvature values than in wild type (Figure 2J). Consistent with this, cross-sections of mutant blades


revealed a loss of planarity (see insets in Figure 2B).


**Cellular architecture is altered in** _**cals6**_ **phyllids**


To determine whether the cellular defects were associated with impaired cytokinesis, we performed


confocal imaging of mature phyllid laminae. Two-dimensional surface views revealed frequent cell wall


discontinuities in the _cals6_ mutant, whereas wild-type cells displayed intact walls throughout (Figure 2C,


2E). Three-dimensional (3D) reconstructions showed that these discontinuities corresponded to gaps of


diverse sizes and geometries within the cell walls (Figure 2D, 2F). Quantification of gap number per cell


wall in 3D confocal images showed that longitudinal walls contained more gaps than transverse walls,


probably because of their larger surface area (Figure 2K). Of all gaps, 73% were located at cell edges, 14%


occurred within the wall faces away from the edges, and 13% spanned the entire wall from one edge to the


other, creating large openings between adjacent cells. Most gaps were smaller than 5 µm, but some


extended beyond 30 µm (Figure 2L). Cell wall thickness was more variable and, on average, greater in


_cals6_ mutants than in wild type, suggesting that loss of _CALS6_ function likely alters cell wall assembly


(Figure 2M).


To test whether these defects were accompanied by changes in the plasma membrane, we knocked out


_CALS6_ by CRISPR/Cas9 in a transgenic line constitutively expressing a YFP fused to an N-terminal


myristoylation signal (21). We stained cell walls of the resulting _cals6_ mutant with propidium iodide (PI),


which also labels dead cells as it penetrates compromised membranes. This dual labelling revealed groups


of cells with an altered appearance in the mutant ( _e.g._, loss of YFP signal and PI-positive cytoplasm), which


was not seen in wild type and suggested a local loss of membrane integrity (Figure 2G, 2H). Together,


these results indicate that _CALS6_ is required to maintain cellular architecture, including cell wall and plasma


membrane integrity.


**Altered cell division and growth gradients underpin developmental defects in** _**cals6**_ **phyllids**


4


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


To investigate how morphological perturbations arise during _cals6_ phyllid development, we performed time

lapse imaging of upper phyllids at 0.5-day intervals and compared their growth patterns with those of wild

type, using the protocol we previously established (21). The number and arrangement of merophytes ( _i.e._,


groups of cells each derived from a single initial cell; delimited by white lines in Figure 3A-3D) were similar


in _cals6_ mutants and wild type, suggesting that phyllid initiation, marked by the oblique division of the initial


cell, proceeds normally in mutants. Two days after initiation, _cals6_ phyllid primordia had a shape and size


comparable to wild type (Figure 3A-D). The total duration of phyllid development, approximately 5.5 days


in wild type, was also unchanged in _cals6_ mutants (Figure 3A-3D). A basipetal growth gradient, in which


cells progressively cease dividing and expanding from the phyllid tip toward the base, was detected in both


genotypes (Figure 3A-3D). In the mutant, however, the actively growing region was smaller and confined


to more proximal positions (Figure 3A-3D). Within this region, fewer cells were dividing or expanding, and


the rates of cell division and area expansion were lower than in wild type (Figure 3E-3F), consistent with


the smaller cells and reduced cell number in mature _cals6_ phyllids (Figure 2A-2I, 3E). The cellular growth


gradient within the actively growing zone was also less pronounced in _cals6_, with a more heterogeneous


distribution of expanding cells (Figure 3A-3D, 3G-3J). In addition, growth at the base of wild-type phyllids,


where they attach to the stem, was strongly inhibited, giving the base a characteristic narrowing (Figure


3C, 3I). This pattern was absent in _cals6_, which probably explains its broader phyllid base (Figure 3D, 3J).


Thus, loss of _CALS6_ function lowers cell division and growth rates and makes growth more heterogeneous.


Because the duration of development is unchanged, this alters the developmental gradients and results in


smaller phyllids.


**Timing and location of local growth arrests account for** _**cals6**_ **phyllid shape defects**


Although altered patterns of cell division and growth were observed in all _cals6_ mutant phyllids analyzed,


their morphological consequences differed substantially between organs, including within the same


gametophore, reflecting a general loss of shape reproducibility (Figure 1J). This variability makes it difficult


to dissect the individual contributions of the defects by live imaging alone. It suggests that the defects arise


stochastically during development, with both their timing and their location likely influencing their impact on


final organ shape. To test whether such defects could account for the abnormal shapes, we used a


previously developed computational model, which qualitatively recapitulates the growth dynamics of


developing phyllids by using positional cues to control growth and cell division (21). We asked whether the


model could reproduce the morphological abnormalities of mutant phyllids by introducing local growth


arrests that mimic the cell expansion defects observed in the mutant.


First, we simulated growth arrest in multiple cells randomly but evenly dispersed throughout the lamina


(Figure 4A-4B). These perturbations altered local cell organization and reduced organ size, but overall


phyllid shape remained largely similar to that of control simulations. The model also predicted that early


growth arrests are more detrimental than equivalent arrests occurring later, consistent with cells formed


5


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


early contributing more to final organ size. These simulations qualitatively reproduced the effects of isolated


single-cell growth arrests observed in _cals6_ mutants (Figure 4C-4D).


Next, we simulated growth arrest in a group of cells in the central region of the lamina (Figure 4E). This


produced pronounced tissue-level defects, including constrictions that became more severe as


development progressed, showing that, in the model, coordinated arrest across larger groups of cells has


much stronger morphological consequences than dispersed arrests. These simulations resembled the


effects of arrested growth in central cell patches observed in _cals6_ mutants (Figure 4F-4G).


Finally, when multiple cells located at the margin were prevented from growing, the model produced


pronounced indentations (Figure 4H). Marginal growth arrest in _cals6_ phyllids also altered shape, but such


indentations were not systematically observed. Instead, neighboring cells appeared to compensate partially


for the arrest by expanding toward the region resistant to growth, as shown by the orientation of growth


tensors at 3.5 d (Figure 4I-4J). This suggests intercellular feedback, possibly mechanical, that buffers local


growth defects during tissue growth. Together, these results indicate that local growth arrests can account


for the abnormal morphology of _cals6_ phyllids, with consequences that depend strongly on their timing and


spatial distribution.


**Discussion**


In vascular plants, organ planarity depends on a strict coordination of growth rate and orientation both within


the organ plane and between cell layers (11). In contrast, moss phyllids offer a minimal system to isolate


the in-plane component of this problem, independently of the inter-layer mechanics that dominate most


angiosperm models. Here, through a detailed quantitative analysis of phyllid morphogenesis in the _P._


_patens cals6_ mutant, we reveal how the coordination between cell division and growth contributes to the


robust shaping of these organs.


Our data reveal that _P. patens_ _CALS6_ disruption strongly impairs both cell division and growth processes,


producing organs with variable, non-planar shapes. The similarity of these phenotypes to those of _A._


_thaliana_ _cals10/gsl8_ mutants supports the idea that _CALS_ -dependent regulation of cell division is


evolutionarily conserved across distantly related land plant lineages (38, 40, 41). This result is also


consistent with the broader observation that moss phyllids and angiosperm leaves share a common


developmental logic. In both systems, tightly controlled spatio-temporal growth heterogeneities across the


lamina underlie robust planar shape establishment (11, 20, 21). Using live imaging and computational


modeling of the stochastic defects induced by _cals6_ loss-of-function, we found that local growth


perturbations can produce minor or severe shape defects depending on their timing and spatial location.


Because the phyllid lamina consists of a single cell layer, these defects cannot result from a mismatch


between mechanically distinct cell layers, as proposed in _A. thaliana_ (9, 18, 47). Instead, localized within

layer growth arrests are sufficient to generate growth conflicts that compromise shape robustness. Recent


work in _A. thaliana_ showed that _jaw-D_ mutant leaves curve because proximal–distal growth is uneven


6


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


across the width of the blade, not because of differential growth between the adaxial and abaxial sides,


indicating that in-plane growth coordination is also required for flatness in a multilayered organ (16).


Whereas the _A. thaliana_ _jaw-D_ mutant reflects a systematic imbalance between midrib and blade, the _P._


_patens_ _cals6_ mutant shows that cell-scale growth arrests are sufficient to produce a comparable outcome


in a monolayered organ.


Our observations indicate that some growth defects in _cals6_ phyllids are at least partially buffered by the


surrounding tissue, possibly through mechanical feedback between neighboring cells. This buffering


capacity parallels findings in _A. thaliana_, where robust patterning is achieved despite substantial cell

autonomous stochasticity, and growth conflicts are mechanically compensated to avoid buckling (48, 49).


Tolerance to local, random variations in cellular growth may thus be a general requirement for achieving


reproducible planar organ shapes across land plants. However, the out-of-plane tissue deformations


observed in _P. patens_ _cals6_ phyllids reveal that some growth conflicts are resolved at the expense of


planarity. Notably, marginal growth arrest was not fully compensated by neighboring cells, consistent with


the distinct morphological identity of margin cells relative to lamina cells (21, 45), and the notion that margin


domains may act as a structural frame constraining the expanding blade (11, 12). This suggests that this


frame-like role of the margin might be a deeply conserved feature of laminar growth, present even in a


monolayered organ type.


Together, our findings position callose-dependent cytokinesis as a determinant of in-plane growth


coordination that, beyond its role in cell division, safeguards against growth conflicts to maintain robust


planarity in a monolayered tissue. By isolating this within-layer contribution from the inter-layer mechanics


at play in angiosperm models, our study further establishes the moss phyllid as a minimal system for


dissecting the general principles of planar organ morphogenesis in land plants (20).


**Material and methods**


**Plant material and growth conditions**


_Physcomitrium patens_ Gransden ecotype was used as the genetic background for all experiments. Plants


were grown _in vitro_ under sterile conditions on BCD medium supplemented with dibasic ammonium tartrate


(BCDAT medium) and solidified with 0.7% agar, following (50). Plants were maintained in a growth chamber


at 23°C under a 16 h light/8 h dark photoperiod or continuous light for live-imaging experiments, with a light


intensity of 50–80 μmol m⁻² s⁻¹. For live imaging, a transgenic line expressing _PpEF1α::acyl-YFP_ (21) was


used as the wild-type (WT) background for generating the callose synthase mutant _Ppcals6_ .


**Generation and genotyping of** _**Ppcals6**_ **mutants**


To generate the _Ppcals6_ knockout mutants, four protospacers were designed to target the exons 1, 3, 21


and 50 of the _PpCALS6_ gene (Pp3c4_1790v3.1) using CRISPOR software (51). For enhanced readability,


7


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


_PpCALS6_ is referred to as _CALS6_ throughout the manuscript. The corresponding sgRNAs (sgR-PpCALS6

EX01-59-F, CCATCCTGGGAACCGCACGCCTGG; sgR-PpCALS6-EX01-59-R,


AAACCCAGGCGTGCGGTTCCCAGG; sgR-PpCALS6-EX03-33-F, CCATCCGTGAGAAGTCTGACATCT;


sgR-PpCALS6-EX03-33-R, AAACAGATGTCAGACTTCTCACGG; sgR-PpCALS6-EX21-47-F,


CCATGTACACTGTAGCATCGGCAT; sgR-PpCALS6-EX21-47-R, AAACATGCCGATGCTACAGTGTAC;


sgR-PpCALS6-EX50-53-F, CCATTGCCGCTTGAAACAGGCGCA; sgR-PpCALS6-EX50-53-R,


AAACTGCGCCTGTTTCAAGCGGCA) were synthesized, annealed and cloned into the destination vector


pZeo-Cas9-gate published in (52), yielding the DM12-PpCALS6 plasmid. This construct was transformed


into moss protoplasts of either wild type or the _PpEF1α::acyl-YFP_ line using PEG- mediated method


previously described (53). The transgenic lines were transiently selected based on antibiotic resistance and


genotyped by PCR using the primers PD40_F6 (GGTGATTTCTCGAGGAGGGC) and PD40_R6


(CATACGGGCAATCTCTCGCA). Mutant lines with complete gene deletions were retained (lines #2 and


#7 in the wild-type background, and lines #15, #19, #21 and #24 in the _PpEF1α::acyl-YFP_ background),


and all presented similar morphological defects. The lines #2 and #15 were used for experiments shown in


Figure 1 and Figures 2-4, respectively. Predicted off-target sites suggested by CRISPOR software were


screened by PCR, confirming the absence of non-specific mutations.


**Gene identifiers and expression data**


Expression profiles of the _CALLOSE SYNTHASE_ ( _CALS_ ) gene family members were obtained using the


expression dataset “ _P. patens_ vegetative development” of the PEATmoss database (54, 55)


[(https://peatmoss.plantcode.cup.uni-freiburg.de/easy_gdb/tools/expression/expression_input.php).](https://peatmoss.plantcode.cup.uni-freiburg.de/easy_gdb/tools/expression/expression_input.php)


_PpCALS4_ corresponds to Pp3c10_19330V3.1 gene model, _PpCALS5_ corresponds to Pp3c9_4140V3.1


gene model, _PpCALS6_ corresponds to Pp3c4_1790V3.1 gene model, _PpCALS11_ corresponds to


Pp3c15_22960V3.1 gene model and _PpCALS12_ corresponds to Pp3c4_15060V3.1 gene model.


Throughout the manuscript, and unless otherwise stated, gene names are abbreviated for readability, _e.g._


_CALS6_ refers to the _P. patens_ gene _PpCALS6_ .


**Microscopy**


_Gross phenotyping_


Colony morphology, gametophore shoots, and individual phyllid outlines were imaged using a Keyence


VHX-1000E digital microscope equipped with a 20–200× objective.


_Live and fixed imaging of phyllids_


Confocal images of individual phyllids were acquired using a Zeiss LSM 800 microscope equipped with a


40× water-immersion Apochromat objective (NA 1.0). Z-stacks were acquired at 0.5-μm intervals along the


z-axis, with an image resolution of 1024 × 1024 pixels. YFP fluorescence from the membrane marker was


8


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


excited using a 488-nm diode laser, and emission was collected over the 400–614-nm range. To image


phyllids stained with propidium iodide (PI), the plant tissue was first immersed in a 50 μg/mL PI solution


and subsequently scanned using 488 and 561 nm diode lasers. YFP and PI emissions were collected at


400–550 nm and 588–617 nm, respectively. For time-lapse imaging of upper phyllid development,


gametophores bearing newly emerged 12th phyllids were selected from 18-day-old plates. Neighboring


phyllids were carefully removed using fine tweezers to expose the targeted phyllid. The same phyllid was


imaged at 12-h intervals over approximately six days, until maturity. When necessary, gametophores were


tilted to expose the abaxial surface of the target phyllid. For samples extending beyond the objective's field


of view, multiple overlapping z-stacks were acquired to cover the entire phyllid. During imaging, plant


samples were immersed in sterile water. After each imaging session, samples were rinsed three times with


sterile water, gently blotted with tissue paper, and returned to their original growth conditions in the growth


chamber. At least three independent time-lapse series were acquired for each experiment. Fine Z-stacks


were acquired using DAPI imaging settings (excitation at 358 nm; emission at 461 nm) to visualize cell wall


autofluorescence. Raw image data were processed using ZEN Black software (Carl Zeiss, Jena, Germany).


Three-dimensional reconstructions in ZEN Blue were used to determine the number and spatial distribution


of gaps.


_Callose imaging_


Aniline blue fluorochrome (Biosupplies) was prepared at 0.1 mg/mL in sterile deionized water and stored


at 4 °C in the dark, according to (56). Before use, the stock was diluted 1:3 (33 µg/mL final) in 0.1 M


phosphate buffer, pH 8.5 (97 mM K₂HPO₄, 3 mM KH₂PO₄). Gametophore apices/filaments were harvested


from the edge of well-grown colonies, avoiding agar, and placed in 100 µL of staining solution per well in a


24-well plate. The plate was wrapped in aluminum foil and incubated for 30 min at room temperature.


Samples were then mounted in distilled water between a slide and a coverslip. Imaging was performed on


a Leica SP8 confocal microscope equipped with a 405 nm diode laser (10% power), a PMT spectral


detector, and 10× dry and 40× water-immersion objectives. Aniline blue fluorescence was collected with a


440–520 nm emission band. The signal at the cell wall between the apical and subapical cells was first


located under UV illumination with the 10× objective, then imaged with the 40× objective. Acquisition


parameters were: frame size 1024 × 1024, 12-bit depth, zoom 1.5×, pinhole 1 A.U., 8-frame averaging,


8000 Hz scan speed. Z-stacks comprised 30 steps of 0.7 µm. Images were processed in Fiji.


**Image processing and quantitative analyses**


_Phyllid reconstruction_


Confocal time-lapse images were processed using MorphoGraphX (57, 58). Images were subjected to two


successive Gaussian blur operations, followed by Edge Detect using a threshold of 6,000–13,000 and Edge


Detect Angle using a threshold of 4,000–8,000. Two-and-a-half-dimensional (2.5D) meshes were


9


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


generated using a cube size of 5 μm, then subdivided and smoothed two to three times. Fluorescence


signals were projected onto the meshes from a distance of 4–6 μm from the surface. The meshes were


manually segmented, and parent cells were identified and labeled across consecutive imaging sessions.


Cell lineages were reconstructed across multiple days based on parent-cell labels, and cell clone identities


were assigned manually.


_Cellular growth and division_


Cellular growth and division were quantified using cell lineage information and the topology of neighboring


cells. Area expansion was calculated from the total area of the daughter cells at time point (T_{i+1}) relative


to the area of their parent cell at time point (T_i). Cell division was quantified as the number of daughter


cells identified at the next time point minus one. The principal directions of growth between consecutive


time points were characterized using the mutually perpendicular Stretch Max and Stretch Min


measurements in MorphoGraphX.


_Cell position_


To quantify cell position along the phyllid, a group of cells at the phyllid base was manually labeled to define


the basal boundary, which was positioned perpendicular to the main phyllid axis. The distance of each cell


from this boundary was calculated using the Cell Distance function in MorphoGraphX, with Wall Weights


set to Euclidean to obtain absolute distances. Relative distances from the phyllid base were calculated by


normalizing absolute distances to the maximum distance measured at the corresponding time point.


Distances from the phyllid tip were calculated using the same approach, with the single apical cell used as


the reference.


_Growth defects_


To investigate the local effects of growth defects in _cals6_ phyllids, growth-arrested or slow-growing cells


were identified from heat maps of cell area expansion calculated over consecutive 12-h intervals. Direct


neighboring cells sharing a common cell wall with the affected cells were also identified, excluding cells


belonging to the midrib lineage. Cumulative area expansion was calculated for the affected cells and their


neighbors from the onset of the growth defect until the phyllid first reached its mature shape. A


corresponding group of cells in WT phyllids was analyzed for comparison.


_Curvature_


Cellular curvature was quantified using segmented meshes in MorphoGraphX. Mean cell curvature was


calculated using the Project Mesh Curvature function and visualized as a heat map using the Heat Map –


Measures – Signal Total function with default parameters.


_Gametophore and phyllid length_


10


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


The length of gametophores and the length, width, and perimeter of individual phyllids were measured


using Fiji software (59). Shoot length was measured by tracing the gametophore stem using the segmented

line tool. Phyllid length was measured along the midrib, width was measured at the widest point, and


perimeter was determined by tracing the phyllid contour using the segmented-line tool.


_Cell wall gap size and thickness_


Cell wall autofluorescence was used to examine the number, size, and distribution of cell wall gaps in _cals6_


phyllids and to measure cell wall thickness in both wild-type (WT) and _cals6_ phyllids. Phyllids were cleared


for 14 days in a solution containing 120 g chloral hydrate, 7.5 mL glycerol, and 150 mL water. Cleared


samples were embedded in the same solution and imaged using a Zeiss LSM 900 confocal microscope


equipped with an Airyscan 2 detector and a C-Apochromat 40×/1.20 W objective, using multiplex Airyscan


mode. Cell wall thickness and gap size were measured from two-dimensional images using ZEN Blue, with


both transverse and longitudinal cell walls included in the analysis. Cell wall thickness was measured using


the cell wall autofluorescence between two adjacent cell walls in at least 200 cells. Gap size in _cals6_ was


quantified from 350 measurements, and gap numbers were determined in more than 100 cells.


**Computational modeling**


The model was implemented in the MorphoMechanX framework (60, 61), following the approach previously


described in Lin et al. (21), as a 2D representation of the phyllid at two scales: cells, each refined into


triangular finite elements that never cross cell boundaries. Growth parameters were assigned to each cell


according to its type and position, and inherited by its elements. When a cell divided, its elements were


discarded and the daughter cells were re-triangulated. Because residual stresses are released at each


growth step, this remeshing did not alter the force distribution in the tissue. Growth was simulated with a


growing finite element method (FEM), using 2D three-node triangular membrane elements instead of 3D


six-node wedge elements, and a linear St. Venant isotropic material (Young’s modulus 100 MPa, Poisson’s


ratio 0.3). At each step, the reference configuration of each element was enlarged, with independent control


of growth parallel (K_par) and perpendicular (K_per) to a polarity direction, itself derived from the gradient


of a distance field computed from the apical cell. Mechanical equilibrium was then solved, vertex positions


updated and residual stresses released. To define the initial template, a representative day-1 mesh was


flattened into 2D and its cell outlines extracted as polygons to serve as the template. The triangulation was


updated upon division, with edge points regularized to maintain suitable triangle aspect ratios. The apical


cell was given a distinct identity, and the most basal vertices were fixed along the longitudinal axis (Dirichlet


conditions). Concerning tissue zonation and cell division, the phyllid was divided into three zones: a slow

growing basal attachment zone; above it, a proliferative zone defined by distance from the base (in cells),


where cells divided on reaching a threshold area along the shorter of two candidate walls, oriented parallel


or perpendicular to their polarity; and a differentiation zone, entered as cells were displaced out of the


11


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


proliferative zone, where division stopped. The apical cell instead divided at 60° in an alternating left–right


pattern. Growth rates could be set as gradients within zones, and zone sizes varied to match developmental


phases and/or the influence of auxin. Zones and parameters were updated after each growth step. Growth


arrest was simulated by specifying “wooden cells” that stop growing and dividing, and are 100-fold stiffer


than other cells; these are shown in grey in the model output visualizations. At a defined time point, a midrib


was specified from cells close to the base and to the central longitudinal axis. These cells received distinct


growth parameters, notably a strong reduction of medio-lateral growth, and their identity was passed on to


daughter cells.


**Acknowledgments**


This work was supported by Czech Science Foundation grant 22-35680M (to R.P.) and 26-23380N (to


P.C.), the Centre SÈVE, the Fonds de Recherche du Québec Nature et Technologies New Researcher


grant 2020-NC-267497, and the Natural Sciences and Engineering Research Council of Canada Discovery


grant RGPIN-2025-04418 (to D.K.), and a Biotechnological and Biological Sciences Research Council


Institute Strategic Program Grant to the John Innes Centre BB/X01102X/1 (to R.S.S). We acknowledge the


Imaging Facility of the IEB CAS, supported by MEYS CR LM2023050 ‘Czech-BioImaging’ and IEB CAS.


Confocal imaging was mainly conducted using the instrument supported by Canada Foundation for


Innovation grant (37805) (to D.K.).


**Author contributions**


Conceptualization: W.L., D.K., and Y.C. ; Methodology: W.L., P.C., R.S.S., D.K., and Y.C. ; Software: M.D.,


D.G., R.S.S. ; Formal analysis: W.L., M.D., D.G., R.S.S., D.K., and Y.C. ; Investigation: W.L., S.H., P.C.,


M.D., D.G., A.L., and Y.C. ; Writing – original draft: W.L., D.K., and Y.C. ; Writing – review & editing: W.L.,


P.C., R.P., R.S.S., D.K., and Y.C. ; Visualization: W.L., D.K., and Y.C. ; Supervision: R.P., R.S.S., D.K.,


and Y.C. ; Project administration: D.K., and Y.C. ; Funding acquisition: R.P., D.K., and Y.C.


**References**


1. E. Coen, A.-G. Rolland-Lagan, M. Matthews, J. A. Bangham, P. Prusinkiewicz, The genetics of
geometry. _Proc. Natl. Acad. Sci. U.S.A._ **101**, 4728–4735 (2004).


2. E. Echevin, _et al._, Growth and biomechanics of shoot organs. _J Exp Bot_ **70**, 3573–3585 (2019).


3. E. Coen, D. J. Cosgrove, The mechanics of plant morphogenesis. _Science_ **379**, eade8055 (2023).


4. B. P. Lapointe, N. S. Kaur, A.-L. Routier-Kierzkowska, A. Burian, From stress to growth: Mechanical
tissue interactions in developing organs. _Current Opinion in Plant Biology_ **86**, 102759 (2025).


12


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


5. R. Kelly-Bellow, _et al._, Brassinosteroid coordinates cell layer interactions in plants via cell wall and
tissue mechanics. _Science_ **380**, 1275–1281 (2023).


6. S. R. Silveira, _et al._, Mechanical interactions between tissue layers underlie plant morphogenesis.
_Nat. Plants_ **11**, 909–923 (2025).


7. K. Jonsson, Y. Ma, A.-L. Routier-Kierzkowska, R. P. Bhalerao, Multiple mechanisms behind plant
bending. _Nat. Plants_ **9**, 13–21 (2023).


8. N. Hervieux, _et al._, Mechanical Shielding of Rapidly Growing Cells Buffers Growth Heterogeneity and
Contributes to Organ Shape Reproducibility. _Current Biology_ **27**, 3468-3479.e4 (2017).


9. A. Singh Yadav, _et al._, Growth directions and stiffness across cell layers determine whether tissues
stay smooth or buckle. _Current Biology_ (2026). https://doi.org/10.1016/j.cub.2026.08.011.


10. C. Le Gloanec, _et al._, Cell type-specific dynamics underlie cellular growth variability in plants.
_Development_ **149**, dev200783 (2022).


11. A. De Carufel, E. Branchini, M. C. P. Timmermans, D. Kierzkowski, How to grow a leaf: Generating,
maintaining, and modulating flatness. _Current Opinion in Plant Biology_ **92**, 102915 (2026).


12. Z. Peng, M. J. Zimmermann, A. H. K. Roeder, Y. Jiao, How to grow a flat leaf. _Current Biology_ **36**,
R714–R728 (2026).


13. K. Guo, C. Huang, Y. Miao, D. J. Cosgrove, K. J. Hsia, Leaf morphogenesis: The multifaceted roles
of mechanics. _Molecular Plant_ **15**, 1098–1119 (2022).


14. A. Baekelandt, _et al._, Arabidopsis Leaf Flatness Is Regulated by PPD2 and NINJA through
Repression of CYCLIN D3 Genes. _Plant Physiol_ **178**, 217–232 (2018).


15. N. Hervieux, _et al._, A Mechanical Feedback Restricts Sepal Growth and Shape in _Arabidopsis_ .
_Current Biology_ **26**, 1019–1028 (2016).


16. K. Harline, _et al._, Growth-rate coordination across the width of a leaf preserves its flatness. _PLOS_
_Biology_ **24**, e3003993 (2026).


17. R. Liu, _et al._, PRESSED FLOWER works downstream of ASYMMETRIC LEAVES 2 to affect sepal
flatness in Arabidopsis. _J Exp Bot_ **76**, 4101–4113 (2025).


18. S. Xu, _et al._, A 3-component module maintains sepal flatness in _Arabidopsis_ . _Current Biology_ **34**,
4007-4020.e4 (2024).


19. L. Serra, E. T. Smithers, L. Bentall, M. O. Lenz, S. Robinson, Cell geometry and mechanical stress
coordinate stomatal division orientation. _Cell Reports_ **45**, 117366 (2026).


20. W. Lin, Y. Wang, Y. Coudert, D. Kierzkowski, Leaf Morphogenesis: Insights From the Moss
Physcomitrium patens. _Front. Plant Sci._ **12**, 736212 (2021).


21. W. Lin, _et al._, Morphogenesis of moss leaf-like organs through variations in deeply shared
developmental principles. _Science Advances_ **12**, eaee6959 (2026).


13


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


22. R. J. Dennis, C. D. Whitewoods, C. J. Harrison, Quantitative methods in like-for-like comparative
analyses of Aphanorrhegma (Physcomitrella) patens phyllid development. _Journal of Bryology_ **41**,
314–321 (2019).


23. T. A. Bennett, _et al._, Plasma Membrane-Targeted PIN Proteins Drive Shoot Development in a Moss.
_Curr. Biol._ **24**, 2776–2785 (2014).


24. T. Viaene, _et al._, Directional Auxin Transport Mechanisms in Early Diverging Land Plants. _Curr Biol_
**24**, 2786–2791 (2014).


25. M. Ishikawa, _et al._, GRAS transcription factors regulate cell division planes in moss overriding the
default rule. _Proceedings of the National Academy of Sciences_ **120**, e2210632120 (2023).


26. Y. Ge, Y. Gao, Y. Jiao, Y. Wang, A conserved module in the formation of moss midribs and seed
plant axillary meristems. _Science Advances_ **8**, eadd7275 (2022).


27. V. Demko, E. Ako, P.-F. Perroud, R. Quatrano, O.-A. Olsen, The phenotype of the CRINKLY4
deletion mutant of Physcomitrella patens suggests a broad role in developmental regulation in early
land plants. _Planta_ 1–10 (2016). https://doi.org/10.1007/s00425-016-2526-2.


28. H. K. Yip, S. K. Floyd, K. Sakakibara, J. L. Bowman, Class III HD-Zip activity coordinates leaf
development in Physcomitrella patens. _Dev Biol_ 1–14 (2016).
https://doi.org/10.1016/j.ydbio.2016.01.012.


29. C. D. Whitewoods, _et al._, CLAVATA Was a Genetic Novelty for the Morphological Innovation of 3D
Growth in Land Plants. _Current Biology_ 1–18 (2018). https://doi.org/10.1016/j.cub.2018.05.068.


30. A. Smertenko, _et al._, Plant Cytokinesis: Terminology for Structures and Processes. _Trends in Cell_
_Biology_ **27**, 885–894 (2017).


31. P. Livanos, S. Müller, Division Plane Establishment and Cytokinesis. _Annual Review of Plant Biology_
**70**, 239–267 (2019).


32. R. Sinclair, _et al._, Plant cytokinesis and the construction of new cell wall. _FEBS Letters_ **596**, 2243–
2255 (2022).


33. R. Sinclair, _et al._, Four-dimensional quantitative analysis of cell plate development in Arabidopsis
using lattice light sheet microscopy identifies robust transition points between growth phases. _J Exp_
_Bot_ **75**, 2829–2847 (2024).


34. K. Thiele, _et al._, The timely deposition of callose is essential for cytokinesis in Arabidopsis. _The Plant_
_Journal_ **58**, 13–26 (2009).


35. X.-Y. Chen, J.-Y. Kim, Callose synthesis in higher plants. _Plant Signaling & Behavior_ **4**, 489–492
(2009).


36. D. P. S. Verma, Z. Hong, Plant callose synthase complexes. _Plant Mol Biol_ **47**, 693–701 (2001).


37. G. Drakakaki, Polysaccharide deposition during cytokinesis: Challenges and future perspectives.
_Plant Science_ **236**, 177–184 (2015).


14


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


38. X.-Y. Chen, _et al._, The Arabidopsis callose synthase gene GSL8 is required for cytokinesis and cell
patterning. _Plant Physiol_ **150**, 105–113 (2009).


39. A. Töller, L. Brownfield, C. Neu, D. Twell, P. Schulze-Lefert, Dual function of Arabidopsis glucan
synthase-like genes GSL8 and GSL10 in male gametophyte development and plant growth. _The_
_Plant Journal_ **54**, 911–923 (2008).


40. B. Saatian, _et al._, Analysis of a novel mutant allele of GSL8 reveals its key roles in cytokinesis and
symplastic trafficking in Arabidopsis. _BMC Plant Biol_ **18**, 295 (2018).


41. D. Ušák, S. Haluška, R. Pleskot, Callose synthesis at the center point of plant development—An
evolutionary insight. _Plant Physiol_ **193**, 54–69 (2023).


42. R. Panstruga, J. Fürst-Jansen, J. de Vries, Plant terrestrialization: piecing together streptophyte trait
evolution – an introduction to a Virtual Issue. _New Phytologist_ **247**, 1963–1970 (2025).


43. S. Schuette, _et al._, Novel localization of callose in the spores of Physcomitrella patens and
phylogenomics of the callose synthase gene family. _Ann Bot_ (2009).


44. D. J. Davis, _et al._, Callose deposition is essential for the completion of cytokinesis in the unicellular
alga Penium margaritaceum. _J Cell Sci_ **133**, jcs249599 (2020).


45. E. I. Barker, N. W. Ashton, Heteroblasty in the moss, Aphanoregma patens (Physcomitrella patens),
results from progressive modulation of a single fundamental leaf developmental programme.
_Journal of Bryology_ (2013). https://doi.org/10.1179/1743282013Y.0000000058.


46. S. Armon, O. Yanai, N. Ori, E. Sharon, Quantitative phenotyping of leaf margins in three dimensions,
demonstrated on KNOTTED and TCP trangenics in Arabidopsis. _Journal of Experimental Botany_
**65**, 2071–2077 (2014).


47. X. He, _et al._, JAG modulates sepal flatness by regulating cell growth direction, interacts with AS2,
and is antagonized by TCP24. _Cell Reports_ **44** (2025).


48. E. Scacchi, _et al._, A diffusible small-RNA-based Turing system dynamically coordinates organ
polarity. _Nat. Plants_ **10**, 412–422 (2024).


49. F. K. Clark, _et al._, A common pathway controls cell size in the sepal and leaf epidermis leading to a
nonrandom pattern of giant cells. _PLOS Biology_ **23**, e3003469 (2025).


50. D. J. Cove, _et al._, Transformation of the moss Physcomitrella patens using direct DNA uptake by
protoplasts. _Cold Spring Harbor Protocols_ **2009**, pdb.prot5143-pdb.prot5143 (2009).


51. J.-P. Concordet, M. Haeussler, CRISPOR: intuitive guide selection for CRISPR/Cas9 genome editing
experiments and screens. _Nucleic Acids Res_ **46**, W242–W245 (2018).


52. D. R. Mallett, M. Chang, X. Cheng, M. Bezanilla, Efficient and modular CRISPR-Cas9 vector system
for Physcomitrella patens. _Plant Direct_ **3**, e00168 (2019).


53. Y. Coudert, O. Novák, C. J. Harrison, A KNOX-Cytokinin Regulatory Module Predates the Origin of
Indeterminate Vascular Plants. _Current Biology_ **29**, 2743-2750.e5 (2019).


15


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


54. P.-F. Perroud, _et al._, The Physcomitrella patens gene atlas project: large-scale RNA-seq based
expression data. _The Plant Journal_ **95**, 168–182 (2018).


55. N. Fernandez-Pozo, _et al._, PEATmoss (Physcomitrella Expression Atlas Tool): a unified gene
expression atlas for the model plant Physcomitrella patens. _The Plant Journal_ **102**, 165–177 (2020).


56. A. Muller, T. Fujita, Y. Coudert, “Callose Detection and Quantification at Plasmodesmata in
Bryophytes” in _Plasmodesmata_, Methods in Molecular Biology., Y. Benitez-Alfonso, M. Heinlein,
Eds. (Springer US, 2022), pp. 177–187.


57. S. Strauss, _et al._, Using positional information to provide context for biological image analysis with
MorphoGraphX 2.0. _eLife_ **11**, e72601 (2022).


58. P. B. de Reuille, _et al._, MorphoGraphX: A platform for quantifying morphogenesis in 4D. _eLife_
_Sciences_ **4**, e05864 (2015).


59. J. Schindelin, _et al._, Fiji: an open-source platform for biological-image analysis. _Nat Methods_ **9**, 676–
682 (2012).


60. G. W. Bassel, _et al._, Mechanical constraints imposed by 3D cellular geometry and arrangement
modulate growth patterns in the Arabidopsis embryo. _Proceedings of the National Academy of_
_Sciences_ **111**, 8685–8690 (2014).


61. G. Mosca, _et al._, Growth and tension in explosive fruit. _Current Biology_ **34**, 1010-1022.e4 (2024).


16


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


**Figure 1. Robust phyllid morphogenesis in** _**Physcomitrium patens**_ **requires** _**CALS6**_ **.** ( **A-B** ) Callose


deposition patterns revealed by aniline blue fluorescence in emerging phyllids of wild-type (WT)


gametophores. Dotted lines mark phyllid outlines. ( **C** ) Heat map showing that _CALLOSE SYNTHASE6_


_(CALS6)_ is the most highly expressed gene out of five _CALS_ genes detected in _Physcomitrium patens_


phyllids and gametophores, grown either on KNOP or BCDAT medium. ( **D** ) Molecular characterization of


the _cals6_ CRISPR mutant (line #2). Left: diagrams showing the _CALS6_ locus in WT and _cals6_ mutant


backgrounds, with light grey-blue boxes showing exons, yellow arrowheads indicate the positions of


17


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


sgRNAs, and dashed lines indicate the edited region identified in the _cals6_ mutant by Sanger sequencing.


Right: schematic representations of the predicted proteins, with WT sequence shown in dark blue and


predicted frameshifted regions in the _cals6_ mutant shown in red. Numbers indicate predicted amino acid


(AA) numbers. ( **E-J** ) Representative colonies (E, H), gametophores (F, I) and phyllid developmental series


(G, J) showing developmental alterations in _cals6_ mutants (H-J) compared with WT (E-G). ( **K** )


Quantification of phyllid length along the gametophore axis, revealing that _cals6_ mutants ( _n_ ≥ 7) have


shorter phyllids than WT ( _n_ ≥ 9). ( **L** ) Phyllid shape complexity (excess length ratio = phyllid perimeter divided


by phyllid length) is greater in _cals6_ mutants ( _n_ = 192) than in WT ( _n_ = 149). For (K) and (L), statistical


significance of a Welch t-test and a Mann-Whitney U one-sided test, respectively, is indicated as follows :


*, (0.01 < p ≤ 0.05) ; **, (0.001 < p ≤ 0.01) ; ***, (p ≤ 0.001). Boxes show the interquartile range, the horizontal


line the median, and dots the mean; whiskers extend to the most extreme values within 1.5× the interquartile


range. Scale bars : A-B, 20 µm ; E-J, 1 mm. For all panels, _CALS6_ refer to the _P. patens_ gene _PpCALS6_ .


18


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


**Figure 2. Cellular defects account for shape variability and planarity loss in** _**cals6**_ **phyllids** _**.**_ ( **A-B** )


Segmented confocal images of isolated phyllids showing the distribution of cell area (A) and curvature (B)


in wild type (WT) and _cals6_ mutants. In (B), insets show cross sections through phyllids at the indicated


positions, yellow arrows indicate protrusions. ( **C-F** ) Confocal images of cell wall (CW) autofluorescence in


WT and _cals6_ mutant phyllid blades, shown as 2D surface views (C, E) and 3D projections of cell


19


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


morphology (D, F). Arrowheads indicate discontinuities and gaps in cell walls. ( **G-H** ) Confocal images of


phyllid blades in WT (G) and _cals6_ mutants (H) expressing a plasma membrane marker (myr-YFP) and


stained with propidium iodide (PI), revealing altered membrane organization in _cals6_ mutants. Dotted lines


delimit putative clonal sectors exhibiting similar alterations. The asterisk marks dead cells. ( **I** ) Boxplots


showing reduced cell area in _cals6_ mutants compared with WT (WT, _n_ = 1599 ; _cals6_, _n_ = 1039). ( **J** ) Plot


showing increased variability in cell curvature in _cals6_ compared with WT. Lines show the median curvature


within distance bins and shaded bands indicate the interquartile range for _n_ independent phyllids (WT, _n_ =


3 ; _cals6_, _n_ = 4). ( **K** ) Boxplots showing the number of gaps in transverse (trans.) (n = 88) and longitudinal


(long.) (n = 89) cell walls in _cals6_ mutants. ( **L** ) Boxplot showing the distribution of gap sizes in _cals6_ mutants


( _n_ = 348). ( **M** ) Boxplots showing increased cell wall thickness in _cals6_ compared with WT (WT, _n_ = 252 ;


_cals6_, _n_ = 208). For (I, K, M), the statistical significance of a Mann-Whitney U one-sided test is denoted as


follows: **, (0.001 < p ≤ 0.01) ; ***, (p ≤ 0.001). Scale bars: A-B, 100 µm ; C-F, 20 µm and G-H, 50 µm. For


all panels, _CALS6_ refer to the _P. patens_ gene _PpCALS6._


20


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


**Figure 3. Growth and division gradients driving phyllid morphogenesis are altered in** _**cals6**_ **mutants.**


( **A-D** ) Heat maps of cell divisions (A-B) and area expansion (C-D) in wild-type (WT) (A,C) and _cals6_ (B,D)


21


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


upper phyllids. Heat values are displayed in the earlier time points. Color scales are similar in (A) and (B),


and (C) and (D). ( **E-F** ) Plots showing that cell division (E) and area expansion (F) are overall reduced in


upper _cals6_ phyllids compared with WT (n = 928 cells). In (F), Boxes show the interquartile range, horizontal


lines the median, and crosses the mean; whiskers extend to the 5th to 95th percentiles. Statistical


significance of a Welch t-test is indicated as follows: *, (0.01 < p ≤ 0.05) ; ***, (p ≤ 0.001). ( **G-J** ) Quantification


of cell division (G-H) and area expansion (I-J) as a function of normalized distance from the organ base,


showing overall reduced cell division and altered growth gradients in _cals6_ mutants (H, J) with respect to


WT (G, I). In (I-J), lines correspond to the median, and shaded areas represent the interquartile range. For


(E-J), data were computed from 3 and 4 independent WT and _cals6_ mutant phyllids, with a maximum


number of 928 and 672 cells, respectively. Scale bars: A-D, 100 µm. For all panels, _CALS6_ refer to the _P._


_patens_ gene _PpCALS6._


22


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


**Figure 4. Stochastic local growth inhibition events contribute to global shape defects in** _**cals6**_


**phyllids.** ( **A-B, E, H** ) Model outputs at successive time points showing phyllid shape in the reference


simulation (A) and in simulations with growth arrests imposed on single cells at random positions across


the blade (B), or on a strip of cells in the middle (E) or at the edge (H) of the blade. Simulations A(i), B(i),


E(i) and H(i) correspond to time frame 120; B(ii), E(ii) and H(ii) to time frame 165; A(ii) and B(iii) to time


frame 275; E(iii) to time frame 245; and H(iii) to time frame 215. ( **C-D, F-G, I-J** ) Heat maps of cumulative


area expansion in various wild-type (WT) and _cals6_ mutant upper phyllids, from 2.5 to 5 days of growth.


23


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


Black lines indicate the direction of maximum growth. The heat maps were computed using the parent


labels from the early time point and displayed on the later time point. Arrowheads in (G) and (J) indicate


protrusions. Scale bars: A-B, E, H = 50 µm; C-D, F-G, I-J = 100 µm. For all panels, _CALS6_ refer to the _P._


_patens_ gene _PpCALS6._


24


[bioRxiv preprint doi: https://doi.org/10.64898/2026.11.05.764497; this version posted November 9, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.05.764497)

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
[available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/)


**Figure S1. Molecular and phenotypic characterization of** _**cals6**_ **CRISPR mutants.** ( **A** ) Molecular


characterization of the _cals6_ CRISPR mutant line #7 (CR#7). Left: diagrams showing the _CALS6_ locus in


wild type (WT) and _cals6_ mutant backgrounds, with light grey-blue boxes showing exons, yellow


arrowheads indicate the positions of sgRNAs, and dashed lines indicate the edited region identified in the


_cals6_ mutant by Sanger sequencing. Right: schematic representations of the predicted proteins, with WT


sequence shown in dark blue and predicted gap in the _cals6_ mutant shown in white. Small numbers indicate


predicted amino acid (AA) position or total number. ( **B-C** ) Representative colony (B) and gametophore (C)


showing developmental alterations in _cals6_ _[CR#7]_ mutants (compare with WT in Figure 1). ( **D** ) Quantification


of gametophore length (n = 20) in _cals6_ mutants (line #2 refers to _cals6_ mutant line shown in Figure 1),


revealing that they are shorter than in WT. Statistical significance of a Welch t-test indicated as ***, (p ≤


0.001). Scale bars : B-C, 1 mm.


25