Introduction

Plants have a wide range of leaf shapes. In many cases, morphological diversity is thought to reflect adaptations to various environmental factors, such as temperature, light, and water availability (Tsukaya 2018). In addition to the leaf form diversity observed in nature, diversity is also found in many cultivars of horticultural plants and crops (Nakayama et al. 2022). In extreme cases, morphological diversity is observed in a single plant; heteroblasty and heterophylly are phenomena of leaf morphological alterations within a single individual, during the different developmental stages, or in response to the surrounding environment, respectively (Nakayama et al. 2017; Poethig 2010; Zotz et al. 2011). Therefore, leaf form diversity is observed at different levels, such as a single individual, inter-species, among cultivars, and intraspecies. Understanding the molecular mechanisms underlying leaf morphological diversity is of great significance in the field of plant biology, particularly in terms of evolutionary, developmental, ecological, and physiological biology. Numerous studies have disclosed the genetic factors and molecular mechanisms influencing leaf form diversity since the 2000s (Bharathan et al. 2002; Blein et al. 2008; Garcês et al. 2007; Richardson et al. 2021; Vlad et al. 2014; Whitewoods et al. 2020) after the basic molecular mechanisms of leaf development were revealed using a few model plant species such as Arabidopsis thaliana (Maugarny-Calès and Laufs 2018). However, although significant progress has been made in the field of leaf morphological diversity, particularly regarding the basic molecular framework for the regulation of leaf complexity, many other questions remain unanswered.

This review aims to outline the existing knowledge of the molecular mechanisms underlying leaf morphological diversity at different levels and reveal any unsolved questions and challenges in the field. Fortunately, there have been numerous review articles published in recent years regarding leaf development and molecular mechanisms (Conklin et al. 2019; Maugarny-Calès and Laufs 2018; Tsukaya 2021), therefore, I will not delve into details about the molecular mechanisms of basic leaf development. Instead, this review focuses on summarizing what has not yet been clarified. This may help highlight areas where research is required to deepen our comprehensive understanding of leaf form diversity and I believe that this review will help guide further development in this field.

Leaf morphology between species and cultivars

A key transcription factor in leaf morphological diversification is the class I KNOTTED-like homeobox (KNOX1). Maize knotted1 (kn1), a KNOX1 gene, was the first homeobox gene found in plants (Hake et al. 1993), and detailed studies using several model plants such as Arabidopsis thaliana (Arabidopsis) and Zea mays (maize) have revealed that it is involved in the formation and maintenance of the shoot apical meristem (SAM) (Jackson et al. 1994; Long et al. 1996). Recent studies have revealed that SHOOT MERISTEMLESS (STM), a mobile KNOX1 protein, trafficking among cells, is indispensable not only for the maintenance and differentiation of shoot stem cells but also for axillary meristem formation, meristem size, and organ boundaries (Balkunde et al. 2017; Kitagawa et al. 2022; Liu et al. 2018). Moreover, the kn1 gene also has important functions in leaf morphology, as shown by its original isolation as a maize leaf mutant (Hake et al. 1993). Subsequent studies on tomatoes, a well-studied model plant for leaf development, showed that overexpression of the KNOX1 gene in the leaf primordium increased leaf complexity (Chen et al. 1997). Indeed, there are several examples of KNOX1 contributing to differences in the leaf morphology in wild and heirloom tomatoes. A Galapagean tomato, Solanum galapagense, exhibits higher leaf complexity. It has a single nucleotide deletion in the promoter of PETROSELINUM (PTS), a short KNOX1 gene that lacks a DNA-binding domain, leading to PTS ectopic expression in young leaves. PTS inhibits the KNOX1 protein interaction with BIPINNATA (BIP) and the KNOX1–BIP complex regulates leaf complexity by modulating KNOX1 expression. Higher expression of KNOX1 in the leaf primordia due to a mutation in the PTS promoter leads to increased leaf complexity in this Galapagean tomato (Kimura et al. 2008). Another study demonstrated that Silvery Fir Tree (SiFT), a Russian heirloom tomato showing increased leaf complexity, had a single nucleotide deletion in the BIP gene, leading to a higher expression of KNOX1 in the leaves (Fig. 1). Comparative genomics indicated that the bip mutation in SiFT likely arose de novo, is unique to SiFT, and is not introgressed from other tomato genomes (Nakayama et al. 2021). In addition, several studies have shown that KNOX1 is involved in leaf complexity in different plant species (Wang et al. 2022).

Fig. 1
figure 1

Leaf form variation of tomatoes. Left: Solanum lycopersicum cv. M82; right: an heirloom tomato “Silvery Fir Tree.” Bar = 1 cm

Full size image

Simple and compound leaves, the most basic classifications of plant leaves, can also be regarded as having differences in complexity. Simple and compound leaves are thought to share common gene regulatory networks and compound leaves undergo repeated leaflet organogenesis (Tsukaya 2014). In this context, KNOX1 remains a key player in facilitating the formation of different leaf shapes between simple and compound leaves. A previous comparative study of Cardamine hirsuta (Brassicaceae) and Arabidopsis demonstrated that STM is necessary for leaflet morphogenesis and that the cis-regulatory divergence in two KNOX1 genes, STM and BREVIPEDICELLUS (BP), is associated with differences in leaf shape between them (Hay and Tsiantis 2006). Additionally, a more recent study using loss- and gain-of-function mutants and transgenics suggested that stm mutants are more pleiotropic than bp mutants. Yet, the BP genomic locus from C. hirsuta is sufficient to modify the Arabidopsis leaf shape to a much greater extent than the C. hirsuta STM locus. Therefore, the authors concluded that although BP has a lesser pleiotropic effect on plant development, cis-regulatory differences in the BP locus between C. hirsuta and Arabidopsis are also sufficient to change leaf shape (Rast-Somssich et al. 2015). Regarding the Arabidopsis leaf shape, a previous study on several species in the genus Arabidopsis demonstrated that the evolution of its unlobed leaf form involved the loss of STM expression in the leaf primordia, and cis-regulatory divergence contributed to this process (Piazza et al. 2010). However, considering broader taxa, factors other than KNOX1 are also known to be involved in the evolution of leaf complexity. For instance, it appears that some simple-leafed species (e.g., Lepidium oleraceum and Pimpinella anisum) have a cryptic compound developmental program in the early stages of initiation, which is then secondarily modified to generate a simple morphology (Bharathan et al. 2002; Champagne and Sinha 2004; Champagne et al. 2007). This indicates that KNOX1 expression in the leaf primordia is not necessarily associated with compound leaves. In Fabaceae, compound-leafed species belonging to the inverted repeat–lacking clade use UNIFOLIATA (UNI), an ortholog of the floral regulator LEAFY/FLORICAULA (LFY/FLO), instead of KNOX1 to regulate compound leaf development (Hofer et al. 1997). In Medicago truncatula (Fabaceae), PINNATE-LIKE PENTAFOLIATA1, a BIP ortholog, negatively regulates the transcription of the LFY ortholog SINGLE LEAFLET1 through direct protein–DNA interactions (He et al. 2019; Whitewoods et al. 2020). In this study, a model was developed that could generate diverse leaf forms and account for the evolutionary origins of cup-shaped leaves through simple alterations in gene expression (Whitewoods et al. 2020). Similar processes may be happening in the bullate leaves of Dischidia plants (Fig. 3g). Similarly, a study focusing on the domatium, a cavity within the structure of plants inhabited by ants and mites, suggested that excessive cell proliferation at the basal part of the leaf blade warps the shape of the blade and disturbs the development of the proximodistal axis, leading to the generation of domatia in Callicarpa saccata (Lamiaceae; Sarath et al. 2020), which is a mymecophytic species (Nakashima et al. 2016; Fig. 3h). These studies indicate that relatively parsimonious cues can generate diverse leaf forms via changes in the orientation of cell division and/or gene expression involved in the establishment of polarity fields. Additionally, recent studies have revealed the importance of cell proliferation position: a study on the relationship between cell division patterns and the final organ shape of sepals and petals, which are the same lateral organs as leaves, revealed that the difference in the location of cell division associated with the spatial AN3 expression pattern mainly contributes to the final organ shape (Kinoshita et al. 2022). This difference may indicate that cell division patterns and the area in which cell division occurs contribute differently to the 2D and 3D morphologies of leaves. However, the detailed developmental patterns, impact of cell division patterns, and position of cell division area in leaf form diversification still need to be clarified in further studies to evaluate each contribution.

In Arabidopsis, leaf meristematic tissues can be divided into two types, long-lived plate meristems (PM) and short-lived marginal meristems (MM), which are controlled by two different pathways (Tsukaya 2021). Although leaf morphology is thought to depend on two meristematic tissues, even in Arabidopsis, the contributions of PM and MM to leaf morphogenesis are unknown. For example, although it has been reported that the transient cell proliferation activity of MM is due to the biased expression of the CIN-TCP-NGA pathway components along the proximodistal axis and that at least part of this effect is mediated by the direct activation of transcription of the KIP-related proteins which are cell cycle inhibitor genes (Rath et al. 2022), little is known about the factors that determine the length of the meristematic activity period. Additionally, the factors that regulate the position of the meristematic regions within a leaf are unknown. Moreover, studies on non-model plants have indicated that other meristematic tissues contribute to the diversification of leaf morphology in some plants. For instance, studies on Juncus (Juncaceae) unifacial leaves that lack the adaxial fate of leaf lamina have shown that thickening meristem also plays an important role in leaf morphology, and an ortholog of DROOPING LEAF (DL), a CRABS CRAW–type YABBY gene, contributes to the thickening growth along the axis perpendicular to the direction of the expansion of bifacial leaves (Yamaguchi et al. 2010). This was sustained by active cell divisions distributed diffusely throughout the cross-sectional area of the leaf blade (Yin and Tsukaya 2019). A subsequent study revealed that the proper placement of auxin induces DL ortholog expression, leading to initial leaf flattening (Nukazuka et al. 2021). A similar mechanism was reported in other independently acquired unifacial leaves, and the importance of thickening meristem is worth emphasizing (Golenberg et al. 2023). Therefore, understanding each leaf meristem in detail and recapturing the complexity and specialization of leaf morphology through changes in these will help us understand the process of morphological change and the molecular mechanisms behind them more precisely. In Arabidopsis petals, cell proliferation mainly occurs in the distal region (Kinoshita et al. 2022), and is considered a good tool for studying the relationship between organ morphology and mitotic regions. The mechanisms underlying the morphological diversification of not only leaves but also petals, which have high morphological diversity and are the same lateral organ as leaves, may be useful in revealing the fundamental mechanism of morphological diversification in lateral organs, including leaves.

Leaf-like organs

Foliage leaves in the genus Asparagus are scale leaves that are reduced in size and do not appear to be fully functional as leaves. Instead, Asparagus plants have leaf-like organs called cladodes. Cladodes develop in the axil where a lateral shoot generally arises; however, they are flattened and leaf-like in shape (Fig. 4a b). Due to the unusual morphology and position of emergence, cladodes have received much attention from several aspects and their origin has been debated for a long time; Arber (1924) concluded that cladodes are prophylls of abortive lateral shoots in the axils based on their leaf-like morphology and anatomical features. However, since it is known that organ arrangement in plants is quite rigid, previous studies have concluded that cladodes are modified lateral branches based on their axillary position (Cooney-Sovetts and Sattler 1987; Kubitzki and Rudall 1998). A study using A. asparagoides with leaf-like cladodes showed that the anatomical features of the cladodes were similar to those of the leaf, except that the positions of the xylem and phloem were inverted in cladodes compared to foliage leaves (Nakayama et al. 2012a). Although KNOX1 expression was limited to the peripheral region of the cladode primordia and finally decreased, an ortholog of ASYMMETRIC LEAVES1 (AS1) was also expressed in the central part of the primordia, suggesting that cladodes differ from true foliage leaves. Additionally, the orthologs involved in the establishment of the ad/ab polarities were expressed in a leaf-like manner. These data indicate that cladodes are modified axillary shoots that co-opt the leaf gene regulatory network, and this co-option allows the axillary shoots to establish ad/ab polarity and to be flattened and indeterminate (Nakayama et al. 2012a, b).

Fig. 4
figure 4

Flattened stems and roots. (a) Asparagus myriocladus. Note that the needle-like structures that look like leaves are all cladodes. (b) Asparagus cladodes. Left: A. asparagoides. Right: A. cochinchinensis. (c) Muehlenbeckia platyclados. Note that the flattened structures that look like leaves are phylloclades. (d) Branching lobed thalli of Hydrobryum japonicum. Photograph courtesy of Dr. Natsu Katayama. (e). Korthalsella japonica shoots grown on Eurya emarginata. Bars = 1 cm

Full size image

In addition to Asparagus, many other plants exhibit flattened stems. For instance, Muehlenbeckia platyclados (Polygonaceae) has phylloclades, a type of flattened stem that consists of multiple internodes (Fig. 4c). The cacti of the genus Opuntia (Cactaceae) also have phylloclades. Therefore, it is worth investigating whether these flattened stems are the result of convergent or parallel evolution. A previous study using Ruscus aculeatus (Asparagaceae) suggested that both RaSTM which is expressed in vegetative and reproductive shoot apices, and RaYABBY2 which is highly expressed in leaves, were co-expressed in the phylloclade primordia (Hirayama et al. 2007), indicating that the Ruscus leaf developmental pathways may have been co-opted into the phylloclade developmental pathway, such as in A. asparagoides and that the acquisition of leaf-like structures in these species seems to be a result of parallel evolution.

Additionally, there are many other examples of leaf-like organs that are not originally leaves, such as Hydrobryum japonicum, in which the root becomes leaf-like (Fig. 4d), and Korthalsella japonica (Fig. 4e), in which the stem becomes leaf-like. Moreover, there are many other organs that are not originally leaves but have planarity and function like leaves. In H. japonicum, the expression of genes involved in the maintenance of SAM and stem cells, such as STM and WUSCHEL, are reduced at the shoot apex, and the AS1 ortholog is expressed instead, resulting in the shoot apex being converted to a leaf and the shoot becoming determinate. The acquisition of determinate shoot formation is thought to have enabled Podostemaceae to expand and diversify further into torrential environments (Katayama et al. 2010). However, the details of root flattening remain unclear in H. japonicum. Although little research has been done on how leaf-like organs develop and evolve, the understanding of the “original” leaf developmental mechanism is progressing, and its application to these leaf-like organs is expected to make rapid progress and unveil the molecular mechanisms behind their evolutionary process in the future.

Succulent leaf

The examples of leaf diversification that I have provided so far are just part of the diversity that plants exhibit. Although the detailed developmental processes and molecular mechanisms are not clear, many plants exhibit interesting leaf morphologies. For instance, the succulent leaf, a type of leaf characterized by the presence of water-storage tissues that allow it to be temporarily independent of external water supply (Males 2017), presents an unusual leaf shape. In general leaves, morphological diversity is generated by deformation or alteration of the whole or a part of a flat leaf blade along the proximo-distal and mediolateral axes, and succulent leaves differ from other leaves in that they generate diversity through changes in the direction of leaf thickness (Fig. 5). Previous studies have suggested that cell enlargement begins with the loosening of the cell wall (Cosgrove 2016, 2023). This initial step induces the subsequent relaxation of wall stress, water uptake, and cell enlargement. Plant cell walls consist of cellulose microfibrils embedded in hydrophilic pectins and cellulose-binding hemicellulose (Cosgrove 2016, 2023). Recent studies have revealed that the wall strength, plasticity, and elasticity depend mostly on the stretching, straightening, and sliding of cellulose microfibril networks. In addition to the cell wall, vacuoles represent another component. In root cells, vacuolar occupancy increases from approximately 40% in meristematic cells to more than 85% in cells of the late elongation zone (Dünser et al. 2019). A recent study revealed that vacuolar enlargement is important for rapid cell expansion rates (Dünser et al. 2019). Although this change did not cause dramatic leaf succulence in Arabidopsis, further elucidation of this mechanism in terms of an increase in cell number in the direction of thickness is important.

Fig. 5
figure 5

Various succulent plants. (a) Crassula perforate. (b) Caputia tomentosa, known as wooly senecio. (c) Senecio antandroi, a species of the genus Senecio endemic to Madagascar. (d) Haworthia truncata. The truncated tip has a leaf window. (e) Haworthia cooperi. A leaf window can be seen at the tip of the leaf. (f) Curio herreanus. The white arrowheads indicate leaf windows on the adaxial side. Bars = 1 cm

Full size image

Another clue to consider regarding the association with succulent leaf development might be endoreduplication, which is a specialized cell cycle in which mitosis is bypassed, thereby producing cells with doubled genome ploidy. Endoreduplication is involved in cell enlargement, particularly in the epidermal pavement cells of Arabidopsis (Galbraith et al. 1991; Gendreau et al. 1997; Hülskamp et al. 1999; Katagiri et al. 2016; Kawade and Tsukaya 2017; Melaragno et al. 1993). Additionally, recent studies have revealed that endoreduplication is linked to cell expansion in the tomato pericarp, and the level of GA is a limiting factor in cell size and ploidy (Renaudin et al. 2023). Interestingly, previous studies have suggested that some succulent plants exhibit endoreduplication (De Rocher et al. 1990; Powell et al. 2020), and certain cells are known to expand in succulent leaves. Therefore, endoreduplication may be involved in the expansion of succulent leaves. However, even in Arabidopsis, the relationship between cell enlargement and endoreduplication and its molecular mechanisms are not fully understood, although a recent study emphasized that endoreduplication is a phenotype (result) of differentiated pavement cells and not a trigger of pavement differentiation (Dubois et al. 2023). Additionally, a recent study on tomato pericarps suggested that not only endoreduplication but also the position of the cell within the tissue is important for cell expansion (Tourdot et al. 2024). Therefore, further studies are required to determine the relationship between cell expansion and endoreduplication.

Finally, some succulents have leaf windows, also known as epidermal windows, and fenestrations in their leaves to allow light exposure (Fig. 5d–f). Leaf windows are seen at the tip or top part of the adaxial surface of leaves, allowing light to be captured even when the plant body is entirely below the soil surface in harsh environments, such as deserts, minimizing the exposure of the leaf surface area to desiccation due to intense heat (Egbert et al. 2008). There are different types of leaf windows, and it is interesting to understand how these windows have developed and evolved.

Origin of leaves

Thus far, we have discussed leaf morphological diversity and its evolution; however, the detailed process of leaf evolution and the underlying molecular mechanisms remain unknown. The fossil record suggests that the ancestors of land plants had a bifurcated branching (equal branching) system with no leaves and is thought to be similar to that of Horneophyton spp. (Fig. 6a). Thus, flattened leaves were acquired during the evolution of land plants (Romanova et al. 2023). The most famous hypothesis for the evolution of leaves from branching forms is the ‘telome theory’. This theory proposed that the transformative evolutionary processes of unequal branching “overtop**,” rearrangement of lateral branches into a single plane “planation”, and infilling of gaps between branches with laminar tissue “webbing” generated leaves (Zimmermann 1952). Fossil records that suggest distinct “webbing” have not been found to date, although “overtop**” and “planation” have been found in Trimerophyton. Additionally, there is no known process for the secondary fusion of multiple developed branches. Therefore, there is still no conclusive evidence supporting the telome theory. However, each process highlighted in the theory is thought to suggest a fundamental process in leaf development and evolution, as attempts are still being made to understand the evolutionary process of leaves by modifying the telome theory to take into account new findings (Harrison and Morris 2018; Vasco et al. 2016). One of the reasons why the process of leaf evolution is difficult to understand is that it has occurred multiple times during land plant evolution (Harrison and Morris 2018; Vasco et al. 2016). In addition to moss leaves, vascular plant leaves (lycophyte and horsetail leaves, and fern and seed plant fronds and leaves; Fig. 6b–d) have evolved multiple times independent of the branching shoot systems. Therefore, previously used terms such as “microphyll” and “megaphyll” are no longer used to emphasize that these leaves have been acquired in each lineage. In addition, our understanding of evolutionary origins of leaves is improving owing to the increased accuracy of phylogenetic relationships among existing and pre-existing land plants, and several studies have suggested that a specific set of genes (toolkit genes) tends to be used even in the development of leaves that have different origins (Harrison and Morris 2018). Moreover, recent genome analyses have suggested that many of the genes involved in plant development exist earlier on in the evolutionary process (Bowman 2022). This suggests that common toolkit genes (or part of the genes) may have been used in independently acquired leaves and raises new questions about what those genes were and how they changed and were used in each lineage. This question is perhaps one of the greatest mysteries related to leaves. The detailed evolutionary process of leaves has not yet been clarified, and further analyses on this subject are required.

Fig. 6
figure 6

Leaves in different lineages in land plants. (a) Schematic reconstruction of Horneophyton, an extinct early plant. This figure is created based on Gifford and Foster (1989). (b) Leaves of Phlegmariurus squarrosa. (c) Asplenium nidus. (d) Welwitschia mirabilis. Bars = 1 cm in (b) and 2 cm in (c) and (d)

Full size image

Conclusion

Many of the examples presented in this review are found not only in one specific lineage but also in multiple independent lineages. For example, indeterminate leaf growth, as observed in M. glabra, has also been observed in some Guarea species (Steingraeber and Fisher 1986; Fig. 3d) and Welwitschia mirabilis, a gymnosperm species (Fig. 6d). This indicates that understanding the molecular mechanisms underlying each leaf’s morphological diversification and evolutionary processes will shed light on whether they are the result of convergent or parallel evolution. The effect of KNOX1 on leaf morphology is one of the most interesting cases reported for multiple phenomena at various levels. Although a re-evaluation of its importance and centrality in leaf form diversification is needed, as mentioned above, leaf morphological diversity involving KNOX1 can be considered a “super-hotspot” and not just a hotspot. It is important to understand whether there are other such genes, and if so, what characteristics they have in the leaf developmental regulatory network. Additionally, some highly specialized leaves may be useful for studying complex adaptive traits. This is because it is still unclear how the complex adaptive traits evolve beyond the valley of adaptation, where they are adaptive in their completed state but not in the intermediate stages before completion and become disadvantageous for survival. Therefore, diversity in leaf morphology is not only fascinating in the realm of plant biology but also serves as a valuable research topic to shed light on issues that are relevant to all living organisms in terms of evolutionary aspects that can be adapted to the whole organism.

This review mainly focuses on the leaf lamina. However, the petiole, a stalk-like structure that attaches the leaf lamina to the stem, has not been extensively studied. Many leaves have this structure, and like other organs, petioles exhibit morphological diversity (Bell and Bryan 2008). For instance, in some species (e.g., Acacia spp.), the petiole flattens laterally into a photosynthetic organ, the phyllode (Dong and He 2017). Others have swollen petioles with cavities inhabited by ants, such as in Piper cenocladum (Tepe et al. 2009). Additionally, food bodies, which are structures that may contain proteins, lipids, or carbohydrates in different proportions and serve to maintain mutualistic relationships between the plant and ants, occur at specialized sites called trichilia at the swelling petiole-stem juncture in Ceropia obtusa (Bell and Bryan 2008; Marting et al. 2018). However, the basic molecular mechanisms underlying petiole development have not yet been studied, even in model plants. Therefore, understanding the molecular mechanisms of petiole development is another unpaved research field in leaf shape, and applying what we are about to uncover to understand the mechanisms underlying petiole morphological diversity will become more important in the future.

The latest estimate of the number of existing plants is more than 300,000 and this number is increasing every year (Christenhusz and Byng 2016); this suggests that studying the variety of leaf shapes and structures is an ongoing topic of plant research and this research field is just at its beginning. However, owing to recent advancements in RNA and DNA sequencing technologies and analysis pipelines, we can now study any plant at the molecular level, instead of being limited to model plants or closely related species. This will enable further research to be conducted on plants with interesting leaf morphologies. Therefore, for those interested in the morphology of organisms and their evolutionary processes, it can be argued that this is a never-ending subject of intellectual inquisitiveness and fun.