Do grasses have meristemoids?

Disclaimer

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.

Self‐renewal is the process by which stem cells divide to produce at least one daughter cell with the same capacity for self‐renewal and differentiation as the parent cell (He et al., ²⁰⁰⁹; Greb & Lohmann, ²⁰¹⁶). This daughter cell eventually differentiates to build up new tissues and/or cells. Stem cells exist in plants and animals, sharing, despite their substantial divergence at the molecular level, surprising similarities (Sablowski, 2004; Heidstra & Sabatini, ²⁰¹⁴). Their enormous potential to produce new cells to grow or replace specialized tissues is having a strong impact on regenerative medicine (Hoang et al., ²⁰²²) and in agriculture for crop improvement (Lindsay et al., ²⁰²⁴). Stem cells are maintained in specialized microenvironments, which are known as stem cell niches (Sablowski, 2004; Mitsiadis et al., ²⁰⁰⁷; Heidstra & Sabatini, ²⁰¹⁴). In plants, these niches are mainly located within the meristems and they produce most postembryonic development (Sablowski, 2004; Aichinger et al., ²⁰¹²; Heidstra & Sabatini, ²⁰¹⁴; Hong & Fletcher, ²⁰²³). Plant stem cells are also dispersed in the epidermis of the developing leaf and cotyledon of many plant species. These stem cells, named meristemoids (Ms), are responsible for triggering stomatal development after they cross a critical cell size threshold (Gong et al., ²⁰²³).

Meristemoids, therefore, are self‐renewing cells, in which one of the daughter cells can exhibit stem properties. In the model plant Arabidopsis, they arise in a basipetal direction from asymmetric cell divisions (entry divisions) of protodermal cells called meristemoid mother cells (MMCs; Serna & Fenoll, ²⁰⁰⁰; Bergmann & Sack, ²⁰⁰⁷; Peterson et al., ²⁰¹⁰; Vaten & Bergmann, ²⁰¹²; Fig. 1a). These Ms, which are the smaller daughter cells, can either directly assume guard mother cell (GMC) fate or undergo several asymmetric divisions in an inward spiral (amplifying divisions), yielding a larger cell and a smaller M that maintains its self‐renewal activity. Finally, the M assumes GMC identity. The GMC undergoes an equal and symmetric cell division that generates the two kidney‐shaped guard cells (GCs). The larger cells that result from the asymmetric divisions of MMCs or Ms, named larger stomatal lineage ground cells (SLGC), can directly differentiate into pavement cells or execute a division (spacing division) to establish an additional M. This developmental process is responsible for generating almost all the epidermal cells in the Arabidopsis leaves (Geisler et al., ²⁰⁰⁰).

Fig. 1.

AtMUTE and grass MUTE orthologs promote a pre‐GMC stage in Arabidopsis and grasses. (a) In Arabidopsis, a MMC cell divides asymmetrically to produce a M and a larger SLGC. Meristemoids are self‐renewing cells, in which one of the daughter cells retains the properties of the parent cell undergoing asymmetric divisions until it assumes GMC identity. The GMC then divides symmetrically to produce the two kidney‐shaped GCs. AtMUTE promotes the transition from the M to the GMC. Therefore, the timing of AtMUTE expression determines the number of amplifying divisions that Ms undergo. AtMUTE also triggers paired GCs formation from the GMC. (b) Stomatal development in grasses starts with an asymmetric division from a SPC, which produces a pre‐GMC and a SLGC. The pre‐GMC assumes GMC identity. Before GMC division, cells from files on either side of the GMC are fated to become SMCs. SMCs then divide asymmetrically to produce SCs that make direct contact with the GMC. When the GMC is flanked by the two SCs, they symmetrically divide giving rise to the two dumbbell‐shaped GCs. Early expression of grass MUTE genes, among other functions, likely represses the stem cell character of the pre‐GMC, promoting its transition to GMC fate without the need to undergo amplifying divisions. (c) Drawings of stomatal phenotypes of atmute, bzu2/zmmute, c‐osmute and sid/bdmute, showing that AtMUTE and grass MUTE genes repress stem cell character of the SPCs. GC, guard cell; GMC, guard mother cell; M, meristemoid; MMC, meristemoid mother cell; SC, subsidiary cell; SLGC, stomatal lineage ground cell; SMC, subsidiary mother cell; SPC, stomatal precursor cell.

In grasses, among them the model Brachypodium (Brachypodium distachyon) and the cereal crops maize (Zea mays) and rice (Oryza sativa), the stomatal precursor cells (SPCs) form from the base to the tip of stomatal‐forming cell files, and they undergo a single transverse division (entry division) producing a distal cell that gives rise, without undergoing amplifying divisions and after a symmetric division, to two dumbbell‐shaped GCs (Stebbins & Shah, 1960; Serna, ²⁰¹¹; Hepworth et al., ²⁰¹⁸; Nunes et al., ²⁰²⁰; Fig. 1b). Some authors have called this immediate precursor of the stoma GMC, which has led to the speculation that grasses do not form Ms (Peterson et al., ²⁰¹⁰; Hepworth et al., ²⁰¹⁸; Serna, ²⁰²⁰). This stomatal precursor divides to produce the two GCs only after cells from ontogenetically unrelated files acquire subsidiary mother cell (SMC) identity and undergo a longitudinal division producing the lateral subsidiary cells (SCs). Like SMCs, the stomatal precursor divides with its cell division plane being parallel to the main axis of the leaf. The proximal cell arising from the entry division (SLGC) could be a way to adjust the stomatal pattern, spacing stomata within a row and perhaps achieving an adequate distance with stomata located in neighboring files.

In addition to regulating the transition from GMC to GCs (Han et al., ²⁰¹⁸), one of the key factors that represses the stem cell character of the Ms in Arabidopsis is the basic helix–loop–helix (bHLH) transcriptional factor AtMUTE (MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷). AtMUTE expression and the localization of the protein encoded by this gene are restricted, in addition to the GMC, to a subset of Ms (MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷). These Ms will presumably undergo GMC transition (Serna, 2013). In these Ms, AtMUTE integrates with cell cycle machinery upregulating the cyclin‐dependent kinase inhibitor SIAMESE‐RELATED4 (SMR4), which represses stem cell divisions by inhibiting the G1‐specific D‐Type Cyclin CYCD3;2 (Han et al., ²⁰²²). SMR4 also permits MUTE‐induced G1 cyclin CYCD5;1 to proceed to a terminal GMC division. In grasses, the primary role of MUTE proteins, which laterally move from the immediate SPCs where their genes are expressed to the neighboring epidermal cell files, is promoting SC recruitment (Raissig et al., ²⁰¹⁷; Wang et al., ²⁰¹⁹). Grass MUTE genes, in addition to regulating other steps of stomatal development (Serna, 2025), also inhibit cell divisions of the pre‐GMCs and perhaps the expression of a possible stem cell character of these cells. The integration of grass MUTE genes with the cellular machinery is not known. Here, I focus on what underlies the absence of manifestation of a SPC self‐renewal stage in grasses.

AtMUTE represses the stem cell character of Ms and promotes GMC fate

In Arabidopsis, stomata develop from Ms, which usually after a few divisions transition to the GMC fate. Studies of cell divisions through time using serial imprints, clonal analysis and long‐term confocal time‐lapse imaging revealed that the number of self‐renewal divisions that undergo the Ms before producing the paired GCs is not fixed but varies from zero to three within a given ecotype and plant organ (Berger & Altmann, 2000; Geisler et al., ²⁰⁰⁰; Serna et al., ²⁰⁰²; Robinson et al., ²⁰¹¹; Serna, ²⁰¹³).

But what are the molecular bases of this plasticity? Why do some Ms divide, while others are immediately fated to GMCs? In the absence of AtMUTE function, stomata are replaced by Ms, which arrest after having undergone an excess of self‐renewing cell divisions that occur in an inward‐spiral pattern (Fig. 1c; MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷). In addition, AtMUTE overexpression converts all epidermal cells into stomata (MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷). This is telling us that AtMUTE terminates the self‐renewing state of the Ms and promotes GMC fate (MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷). This gene also positively regulates the transition from GMC to GCs (Han et al., ²⁰¹⁸). Its expression and the localization of the protein encoded by AtMUTE are restricted to a subset of Ms and to the GMCs (MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷). These Ms presumably will not divide, and they will undergo GMC transition (Serna, 2013). Thus, it is the variability in the timing of AtMUTE expression during the development of stomatal complexes that appears to determine the plasticity in this process, with premature expression preventing the manifestation of the stem cell divisions of Ms, and late expression allowing these cells to undergo amplifying divisions before the transition to GMC.

But how can AtMUTE inhibit the stem character of Ms? A recent study has shown that AtMUTE cooperates with epigenetic machinery to regulate chromatin remodeling and therefore, cell fate specification through transcriptional regulation (Liu et al., ²⁰²⁴). In addition, the regulation of the stem cell character of Ms depends on the heterodimerization of AtMUTE with the functionally redundant bHLH proteins ICE1 (also known as SCREAM) and SCREAM2 (Kanaoka et al., ²⁰⁰⁸). What is not known is whether this physical interaction between these bHLH proteins is necessary to regulate chromatin remodeling and/or gene transcription.

ZmMUTE and OsMUTE define a pre‐GMC stage

In grasses, including maize and rice, the SPC that arises from the asymmetric division of a protodermal cell undergoes a single and symmetric cell division producing a cell that gives rise directly to the paired GCs (Stebbins & Shah, 1960; Serna, ²⁰¹¹; Hepworth et al., ²⁰¹⁸; Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹; Nunes et al., ²⁰²⁰). Because Ms have been defined as stem cells that undergo several amplifying divisions before transition to GMC fate (Bergmann & Sack, 2007; Pillitteri & Dong, ²⁰¹³; Han & Torii, ²⁰¹⁶), does this mean that grasses lack Ms? If the answer is yes, considering that AtMUTE represses the stem cell character of Ms, mutations in their orthologs, BZU2/ZmMUTE and OsMUTE, should not trigger phenotypes like those of atmute. On the contrary, if the answer is negative, the repression of MUTE expression or activity in grasses may resume the self‐renewal capacity of their pre‐GMCs.

The loss‐of‐function bzu2/zmmute mutants have a higher leaf temperature than wild‐type (WT) plants, reflecting decreased transpiration (Wang et al., ²⁰¹⁹). In addition, their photosynthetic activity is also decreased (Wang et al., ²⁰¹⁹). These lethal physiological changes are a direct consequence of the defects in stomatal development. Indeed, the stomatal precursors of the bzu2/zmmute mutant, instead of symmetrically dividing to produce a pair of GCs, undergo excessive, randomly oriented and/or asymmetric divisions that give rise to short columns of two, three or even four elongated and undifferentiated cells (Fig. 1c; Wang et al., ²⁰¹⁹). Unlike AtMUTE expression, BZU2/ZmMUTE is expressed in all young SPCs from the timeframe of SMC establishment (Wang et al., ²⁰¹⁹). Therefore, BZU2/ZmMUTE expression is defining a stage which could be equivalent to the stage in which the last Ms of Arabidopsis are formed.

Like the bzu2/zmmute mutant, the lethal loss‐of‐function c‐osmute mutant exhibits also short columns of two cells, produced by misoriented and/or asymmetric cell divisions, instead of two dumbbell‐shaped GCs (Fig. 1c; Wu et al., ²⁰¹⁹). Consistent with this phenotype, OsSPCH1 and OsSPCH2, which control early steps of stomatal development (Liu et al., ²⁰⁰⁹; Wu et al., ²⁰¹⁹), were expressed in c‐osmute (Wu et al., ²⁰¹⁹), and OsFAMA, which is required for GC differentiation (Liu et al., ²⁰⁰⁹; Wu et al., ²⁰¹⁹), was downregulated (Wu et al., ²⁰¹⁹). Both transpiration and photosynthetic activity of the c‐osmute mutant, like those of the bzu2/zmmute mutant (Wang et al., ²⁰¹⁹), must also be dramatically affected. In contrast to AtMUTE and like BZU2/ZmMUTE, OsMUTE expression was observed in all the stomatal precursors from the moment SMC divides (Wu et al., ²⁰¹⁹). Therefore, OsMUTE expression, like BZU2/ZmMUTE, defines also a pre‐GMC stage.

The early expression of BZU2/ZmMUTE and OsMUTE strongly suggests that these genes prevent the manifestation of self‐renewal behavior of the pre‐GMCs, accelerating their immediate transition to the GMC fate. However, because both bzu2/zmmute and c‐osmute not only replace their stomata by short columns of cells but also lack SCs (Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹), the observed defects of the SPC division could simply be a consequence of the absence of SCs. Interestingly, although all stomata in c‐osmute are replaced by small cells that lack SCs (Wu et al., ²⁰¹⁹), some complexes of bzu2/zmmute (c. 5%) developed one single SC (Wang et al., ²⁰¹⁹). Considering that all stomata in bzu2/zmmute, including those that have one SC, were replaced by columns of undifferentiated cells (Wang et al., ²⁰¹⁹), it seems that the defect of the SPC division is not a consequence of the absence of SCs, but of the absence of BZU2/ZmMUTE activity. Thus, the expression of BZU2/ZmMUTE and probably OsMUTE not only defines but also promotes a stage that may be equivalent to that in which the last M are formed in Arabidopsis.

BdMUTE or BdFAMA are required to advance to a pre‐GMC stage

The subsidiary cell identity defective (sid/bdmute) mutant, with alterations in the Arabidopsis MUTE orthologue BdMUTE, in addition to having impaired other steps of stomatal development (Serna, 2025), undergoes misoriented divisions in c. 30% of the SPCs, aborting the formation of GCs (Fig. 1c; Raissig et al., ²⁰¹⁷). However, in contrast to both bzu2/zmmute and c‐osmute mutants, which exhibit a complete lack of stomata (Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹), the remaining 70% of the stomatal precursors of the sid mutant produce dicot‐like stomata. Because of this, it is fully viable and fertile, although its gas exchange capacity and stomatal diffusion capacity are also affected (Raissig et al., ²⁰¹⁷). The BdMUTE promoter is also induced early from the timeframe of SMC establishment (Raissig et al., ²⁰¹⁷). Therefore, BdMUTE expression also defines a stage that could be equivalent to the stage in which the last Ms of Arabidopsis are formed.

Because the sid/bdmute mutant also lacks SCs (Raissig et al., ²⁰¹⁷), the observed defects in SPCs could simply reflect a consequence of the absence of SCs. Interestingly, a line expressing 3×GFP‐BdMUTE (the weakest expressing one) in the sid/bdmute background almost completely rescued paired GCs formation but failed to recruit SCs (Spiegelhalder et al., ²⁰²⁴). This indicates that the misoriented divisions of the stomatal precursors of the sid/bdmute mutant are not due to the absence of SCs, but to the sid/bdmute mutation. In agreement with this, although another of the lines expressing 3×GFP‐BdMUTE (the highest expressing one) almost completely rescued the sid/bdmute phenotype, with the mutant exhibiting 98% of complexes identical to those of WT plants, a considerable number of GMCs, at the time of their division, lacked adjacent SCs because the recruitment of SCs was delayed (Spiegelhalder et al., ²⁰²⁴). Therefore, the lack of SCs does not trigger the defects of the SPC division in the sid/bdmute mutant, and such defects are a consequence of the absence of BdMUTE activity. It is then likely that premature expression of BdMUTE, like those of ZmMUTE and OsMUTE, represses cell divisions and perhaps the expression of a possible stem cell character of the pre‐GMC, accelerating its transition to GMC fate without the need to undergo amplifying divisions. Because grass MUTE genes function by correctly orienting the division plane of their stomatal precursors (Raissig et al., ²⁰¹⁷; Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹; Spiegelhalder et al., ²⁰²⁴), it cannot be ruled out that the primary role of these genes is to change the orientation of cell division, from transverse to longitudinal, which in turn could regulate the cell division capacity.

Interestingly, a severe double bdmute; bdfama fails to produce any stomata and instead develops to short columns of undifferentiated cells like those of bzu2/zmmute or c‐osmute mutants (McKown et al., ²⁰²³). This is telling us that BdFAMA, in addition to positively regulating the transition from GMC to GC, also represses the stem cell character of the pre‐GMC. Consistently, BdFAMA expression is not restricted to the last stage of stomatal development, but its expression, starting during SC recruitment, overlaps with those of BdMUTE (Raissig et al., ²⁰¹⁷; McKown et al., ²⁰²³). Assuming that grass FAMA function is conserved, the fully penetrant phenotype of both bzu2/zmmute and c‐osmute mutants suggests that in maize and rice, FAMA expression may depend on MUTE activity. In agreement with this, both bzu2/zmmute and c‐osmute barely express FAMA (Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹). Therefore, this MUTE‐dependent activation of FAMA could explain the complete abortion of stomatal development in the bzu2/zmmute and c‐osmute mutants. BdMUTE not only represses but also promotes stem cell divisions, which depend on the cell type. Indeed, overexpression of BdMUTE induces polarized SMC‐like divisions, giving rise to stomata with multiple layers of SCs (Raissig et al., ²⁰¹⁷). This indicates that BdMUTE, in SMCs, promotes rather than ends stem cellness. The formation of similar cells throughout the epidermis when OsMUTE is overexpressed (Wu et al., ²⁰¹⁹) could also point in the same direction.

Concluding remarks and future perspectives

Ms possess a stem cell‐like character, reiterating asymmetric amplifying divisions. This has led us to propose that protodermal cells directly produce GMCs, and that grasses therefore lack Ms (Peterson et al., ²⁰¹⁰; Hepworth et al., ²⁰¹⁸; Serna, ²⁰²⁰). However, the phenotype of both bzu2/zmmute and c‐osmute, where all stomata are replaced by short columns of undifferentiated cells (Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹), suggests that maize and rice may have Ms, and that an early expression of both BZU2/ZmMUTE and OsMUTE may repress their stem cell divisions. Consistently, these genes are expressed in all young stomatal precursors from the stage of SMC establishment or division (Wang et al., ²⁰¹⁹; Wu et al., ²⁰¹⁹). Although the sid/bdmute mutant phenotype is not fully penetrant (Raissig et al., ²⁰¹⁷), the same seems to occur in Brachypodium. The early expression of BdMUTE in all SPCs from the timeframe of SMC establishment (Raissig et al., ²⁰¹⁷) suggests that it may repress the manifestation of their stem cell character. Therefore, the early expression of grass MUTE immediately depletes the potential stem cell pool. Delving into what determines this early expression would provide fundamental insights into the cellular biology of self‐renewal and the mechanisms that not only initiate but also regulate stem cell identity.

In Arabidopsis, the bHLH transcription factor SPEECHLESS (SPCH) promotes the transition of protodermal cells to MMC and their subsequent asymmetric divisions (MacAlister et al., ²⁰⁰⁷; Pillitteri et al., ²⁰⁰⁷; Lau et al., ²⁰¹⁴). Grasses have two copies of SPCH (Raissig et al., ²⁰¹⁶; Wu et al., ²⁰¹⁹). The duplication of SPCH in this group of plants was not well understood, mainly because grasses were thought to lack Ms. Assuming that they have Ms, it is likely that grass SPCHs, redundantly and through strategies that differ from those of Arabidopsis SPCH (McKown & Bergmann, 2020), promote the transition of protodermal cells to MMC, and that the early expression of grass MUTE prevents grass SPCHs subsequent function. The function of the signaling peptides EPIDERMAL PATTERNING FACTOR 1 (EPF1) and EPF2 in Arabidopsis is also well known. While EPF2 negatively regulates entry divisions and subsequent M activity (Hara et al., ²⁰⁰⁹; Hunt & Gray, ²⁰⁰⁹), EPF1 predominantly regulates the M‐to‐GMC transition (Hara et al., ²⁰⁰⁷; Hunt & Gray, ²⁰⁰⁹). As it was assumed that grasses lacked Ms, the function of EPF1/2‐like peptides was thought to be also divergent between Arabidopsis and grasses (Chater et al., ²⁰¹⁷). The questioning of the presence of Ms in grasses also questions such divergence. In addition, heterologous and homologous overexpression of EPF1/2 from grasses negatively regulates stomatal development (Hughes et al., ²⁰¹⁷; Lu et al., ²⁰¹⁹; Jangra et al., ²⁰²¹), suggesting that the function of these genes, at least in part, is conserved.

The stomatal complex has been fine‐tuned by several innovations, including incorporation of multiple asymmetric stem cell divisions in stomatal precursors to create a variety of stomatal distributions. For example, amplifying divisions, associated with the development of anisocytic stomatal complexes, are known to occur in some Piperales (Rudall, 2023). Therefore, it is likely that grasses lost their M function due to the early expression of grass MUTE genes, which possibly enhanced gas exchange. Furthermore, the function of MUTE genes appears to have changed over the course of evolution, as evidenced by the fact that the Kalanchoe laxiflora MUTE genes (KlasMUTEs) promote rather than repress Ms divisions (Cheng et al., ²⁰²⁴). Understanding the emergence and evolution of Ms, as well as the molecular basis of these changes in the function of MUTE genes, is an exciting challenge for the future.

Competing interests

None declared.