7.1 - Axial growth: root, shoot and leaf development

A vascular plant begins its existence as a single cell, the zygote. The early embryo derived from growth of a zygote is globular whereas a mature embryo has a defined apical–basal growth axis (Figure 7.1). In other words, it has become a polar structure.


Figure 7.1 (a,b) Typical dicotyledonous embryo (Arabidopsis thaliana) showing suspensor (S) at globular and heart stages.  EP = embryo proper, Hs = hypophysis, C = cotyledons.  Photographed with Nomarski optics. (c) Mature monocotyledonous embryo within maize grain.  The shoot apex (with coleoptile and pre-formed leaves, together with scutellum), root apex and an adventitious root (arrowed) are all visible. (Based on Yadegari et al. 1994; Raven et al. 1992)

During longitudinal axis formation, two distinct zones that subsequently retain the capacity for continuous growth are set apart at opposite poles. These regions are the apical meristems, one producing the shoot system, the other producing the root system (Figure 7.1c). These are ‘open-ended’ indeterminate growth systems from which the same kinds of organs and/or tissues are produced continuously and which result in the primary plant body. Often in response to environmental cues such as photoperiod and low temperature, the shoot apical meristem may undergo transition to a floral state. In this case, the meristem has become determinate and ceases to produce new organs. In contrast, most root meristems remain indeterminate, although lateral roots which branch off a primary axis can become determinate. Shoot buds containing meristematic cells give rise both to terminal apices and to lateral branches, for example the crown of a eucalypt and its side branches, respectively. Roots also branch profusely, but from meristematic tissue deep within the root axis, so generating extensive root systems typical of most land plants. In monocots, an intercalary meristem located at each node of the stem provides the facility for continued longitudinal growth if the shoot tip is destroyed, for example by a grazing animal or a lawnmower. Patterns of plant development contrast sharply with those of higher animals where the fundamental body plan, complete with rudimentary organs, is laid down in the embryo. In the case of animals the organ number is finite, unlike the plant body in which an indefinite number of organs such as leaves are produced from indeterminate apical meristems. Within the organs of an animal, further cell divisions replace degenerating cells whereas plant cell division primarily provides new organs to replace those lost through senescence.

The so-called primary plant body, described above, may constitute a whole plant, for example annuals like pea, cereals and Arabidopsis. However, plants with extended lifespans have additional meristem layers called cambium which develop within roots and stems, and lead to an increase in girth along the plant’s longitudinal axis (see Section 7.2). Vascular cambium generates extra conducting tissue; cork cambium produces protective tissue, replacing the functions of epidermis in stems, and cortex and epidermis in roots. Cambial meristems and their derivative tissues are referred to as the secondary plant body. Although no new organs are produced by these lateral meristems, the secondary plant body may constitute the bulk of the plant, for example a tree’s trunk, branches and roots.


Figure 7.2 (a) The first few leaves of many Acacia seedlings often have pinnate leaves (arrow), but phyllodes (flattened petioles) take over photosynthetic functions at later nodes. (b) Juvenile seedling (opposite leaf pairs) and adult shoot (spiral phyllotaxis) forms of Eucalyptus. (Photographs courtesy C.G.N. Turnbull)

Localised meristems, whether axial or lateral, have profound implications for morphogenesis. Changes in the fate of cells emerging from meristems will be evident in the resultant tissues and organs. For example, the abrupt transition from juvenile to mature leaves in eucalypts and acacias (Figure 7.2) reflects this change of fate.
Meristems are restricted to localised regions in higher plants, but in algae cell divisions are not always organised this way. Unicellular organisms undergo divisions to produce a new biological entity capable of further cell divisions and multicellular algae often have diffuse meristems. The latter could in part reflect the less exacting demands of their homogeneous aquatic environment. Producing new cells throughout a developing thallus may be feasible simply because it is well supported under water.

7.1.1 - Root apical meristems

Although roots are sometimes neglected by researchers and called the ‘forgotten half’ or ‘hidden half’ of plants, root apical meristems have been studied extensively for two reasons. First, roots are viewed as a simpler system than shoot meristems – the root meristem is much more accessible than the shoot meristem which is ensheathed by developing leaves. Second, complicating lateral structures arise from the terminal shoot meristem (leaf and bud initials) but not from the terminal root meristem, which produces cells solely for the primary axis. This is where the simplicity ends. A primary root meristem generates two tissues simultaneously, the main root axis extending proximally towards the shoot, and the root cap pushing relentlessly forward into the soil, succumbing to sloughing and hence rapid turnover. The detailed organisation of root meristems, which we consider here primarily from the view of the cell biologist, reveals deeper complexities and questions of cell determination.

Lateral root meristems enable generation of massive networks of fine roots. The evolutionary processes which led to root systems of a very branched nature (e.g. grasses) through to coarse unbranched root systems (e.g. orchids) are a fascinating basis for further research into control of root branching. In addition, molecular intervention is giving us new plant forms which can be used to unravel the controls on root development and branching.

Root meristem anatomy

Primary roots arise through controlled cell divisions in the apical meristem and subsequent expansion and differentiation of these cells. Root formation (rhizogenesis) is usually extremely rapid: daughter cells exiting the meristem may be found 24 h later in a fully differentiated structure (e.g. phloem), even though further modifications to cell function are still possible (e.g. formation of an exodermis).


Figure 7.3 Longitudinal section through a primary root tip of radish (Raphanus sativus). Files of cells extend forward from the centre of the apex to form the root cap, and backwards to form the main root tissues. (Based on Raven et al. 1992)

Critical steps in setting up dimensions and thickness of the root axis, and supply of cells to the zone of elongation, are the rate and position of cell divisions in the meristem. This can be appreciated from the two-dimensional view of a developing radish root in Figure 7.3. Divisions can be in any of three planes, either anticlinal (normal to the root axis), periclinal (tangential to the root axis) or radial to the axis. These divisions will give rise, respectively, to increased root length, increased root thickness (more layers of cells through the root), or increased root circumference. The apical meristem supplies all the cells for the primary root axis and the consequences of the planes of cell division are evident long after meristematic activity ceases.

Separate cell divisions at the leading edge of the root meristem generate a root cap which extends forward as a protective structure. The central cells of the root cap are often oriented in longitudinal arrays (columella) and are destined for rapid attrition as the ‘advancing’ soil particles slough off the surface layers. These cells also fulfil a vital chemico-physical role by secreting a glycoprotein-rich mucigel which reduces friction between root and soil matrix. Root caps advance at a dramatic speed: a root might elongate by 5 cm per day and new root cap cells can be pushed in advance of the apex of the primary axis at about the same rate. This means that the leading tip of a primary root, supplied with new root cap cells, advances through the soil at up to 60 µm min–1. The sloughing off of roughly one cell layer per hour may explain our observation that the root cap does not increase in size over time. Intriguingly, root caps are still conspicuous in roots grown in nutrient solutions but still never dominate the primary root axis, so we deduce that sloughing off may induce a feedback mechanism that upregulates root cap meristem activity.


Figure 7.4 Most root apices contain a quiescent centre of very slowly dividing cells. (a) Diagram of longitudinal section through a maize (Zea mays) root tip. The quiescent centre is shaded dark green. (b) Autoradiograph of transverse section through root apex of Vicia faba (broad bean), fed for 24 h with radioactive [3H]thymidine which specifically labels DNA in nuclei of dividing cells. The quiescent centre has significantly fewer labelled cells (dark silver grains). (Based on Clowes 1959; Waisel et al. 1996)

Another remarkable feature of root apices is the quiescent centre, a paradox at the heart of the meristem (Figure 7.4). The quiescent centre is a zone of relatively inactive, slowly dividing cells, numbering about 500–600 in a mature maize root. Its discovery (Clowes 1959) involved studies on mitotic frequencies, thymidine incorporation into nuclear DNA (Figure 7.4b), and ploidy after colchicine treatment. These led to a radical change in view of plant roots. The quiescent centre functions as a reserve of cells which can survive stresses and provide cells to a regenerating meristem. Recovery from surgical removal of parts of the meristem and irradiation to destroy dividing cells supports this concept. Likewise, short determinate lateral roots often lack a quiescent centre, suggesting it is closely tied to sustained indeterminate development.

Passage of cells from meristem to differentiated structures has been studied in simple roots such as ferns in which a single apical cell can be the progenitor of all root cells. Higher plants such as maize or beans have more complex roots, but the whole root can still be traced back to as few as 12 cells in the middle of the quiescent centre (Lyndon 1990). Cell destiny appears to follow predictable patterns, suggesting the notion of clonal development in which cell fate is fixed from the first divisions in the meristem. This view is under challenge from experiments using laser ablation. Individual cells, or groups of cells, can be eliminated by laser treatment, then the behaviour of adjacent cells is followed to see how the meristem is organised. This demonstrates that cells have considerable scope for taking over the meristematic role of their nearest neighbours. However, the process depends on physical contact between dividing cells and their daughter cells, which suggests a local transfer of information. The implication is that cell fate and the asymmetric divisions which give rise to various cell lines are regulated at a tissue level, but we do not yet know the nature of the mobile signals which might program cells in the meristem. We next turn to the fate of cells in their temporal journey from division to differentiation.

Cells divide in planes which are identifiably targeted to become the various root tissues even before all cell divisions are complete. For example, cells giving rise to the stele are generally clustered around the axis of the root, proximal to the quiescent centre (Figure 7.4a), while those giving rise to outer tissues (endodermis, cortex and epidermis) are peripheral to the pre-stelar cells. In roots of Arabidopsis, which has become a favoured plant for this work, the numbers of cells which generate individual tissues (e.g. eight meristematic cells generate eight cortical files) are known and the order of divisions giving rise to tissues such as pericycle, cortex and endodermis have been defined.

The changes which cells undergo in root meristems are profound; mitotic activity is most rapid in the distal regions (with a mitotic index of up to 23% in some files) but the mitotic cycle slows dramatically within 0.5–1.0 mm from a wheat apex. Surprisingly, mitotic frequency in adjacent cell files can vary widely (Figure 7.5; Table 7.1).

The rate and plane of cell division and subsequent rate of cell elongation determine the rate of delivery of new cells to mature root tissues. The coordination of cell flux is presumably under tight control, achieving the final anatomical outcomes recognisable as mature roots – single layers of pericycle and endodermal cells, long conducting vessels and epidermal cells are some examples. The role of growth in cells exiting the meristem and the direction of expansion are major factors in rhizogenesis, with a 30- to 150-fold volume expansion required to generate the primary axis.


Figure 7.5 Cell division rates vary along different zones of the root apical meristem. Frequency of observed cell division is represented by different densities of stippling (onion root tip). (Based on Jensen and Kavaljian 1958)

Lateral roots

Lateral roots are important in nutrient and water uptake (Chapter 4.1). The cellular reorganisation which leads to lateral root primordia forming and developing into new axes starts with a latent meristematic activity in the root pericycle being de-repressed and cell divisions resume. Occasionally, endodermal cells are also recruited. Periclinal divisions underlie the out-growth of cells and disruption of the outer tissues of the root. However, before the cortex and epidermis have been penetrated by the young lateral root, it has formed its own terminal meristem and root cap. The new organ is thus prepared for growth in the external matrix. Lateral roots formed from the pericycle must breach the endodermis of the parent root. How is this achieved without rupturing of the Casparian strip and preventing outflow of concentrated nutrients to the cortex? Dyes which penetrate only the apoplasm have shown that endodermal disruption is a transient feature of lateral root growth, but the consequences are not well understood.

Lateral roots generally do not form within 1 cm from the terminal apex, and almost never in the zone of elongation. This makes sense as laterals in the growing zone would act as barbs impeding growth of the primary axis through the soil. An exception which supports this view comes from Eichhornia (water hyacinth) which does produce laterals in the elongation zone very near the root tip, but because of its aquatic environment this does not interfere with growth.

7.1.2 - Shoot apical meristems

Shoot apical meristems are minute yet complex structures that are ensheathed within new developing leaves or bracts. A vegetative meristem gives rise to leaves or other organs, for example thorns and tendrils. Axillary buds are themselves complete shoot meristems from which branches are produced (cf. lateral roots described above). In angiosperms, when a plant shifts from vegetative to reproductive growth some meristems undergo a transition to the reproductive state and give rise either to multiple flowers in an inflorescence, as in mango (Figure 7.6a), or to a single terminal flower, for example a poppy or waterlily (Figure 7.6b). All axial growth from meristems, be they vegetative or floral, is continuous or indeterminate until topped by the formation of a flower. When this occurs, floral organ primordia arise in whorls from the shoot meristem and differentiate into the familiar sepals, petals, stamens and carpels.


Figure 7.6 Shoot meristems change morphology and function throughout the life cycle, leading to different mature structures. (a) Indeterminate inflorescence of mango; (b) determinate single flower of a waterlily. (Photographs courtesy C.G.N. Turnbull)

Sometimes, however, the indeterminate inflorescence meristem reverts back instead to a vegetative status. Think of a bottlebrush or a pineapple with a leafy axis extending beyond the flower or fruit (Figure 7.7). What determines whether a meristem is vegetative or floral? In many species, environmental signals cause the switching between vegetative and floral states and a picture is now emerging of the genes and molecular mechanisms responsible for defining structures that are generated by meristems.

Although meristems function as generic sources of cells for differentiation into organs, each type of meristem is programmed to produce only certain structures. Across all species, there is a small, finite range of these structures, yet we observe an amazingly diverse array of final vegetative and floral morphologies. A leaf is always recognisable as a leaf but consider the vast structural differences between a pine needle, a waterlily pad and a tree fern frond. The generation and spatial patterning of plant organs are determined by early events within the vegetative meristem. This precise positioning of organs around the shoot meristem is called phyllotaxis. Later in development, a dramatic meristematic switch will give rise to a terminal inflorescence, often with an abrupt change in patterning of organs. Phyllotaxis also applies to floral structures, for example spiral patterns of scales of a pine cone or bottlebrush flowers (Figure 7.7).


Figure 7.7 Alternating phases of vegetative and reproductive growth in the bottlebrush Callistemon ‘Pink Perth’). Spiral phyllotaxis is visible in the young inflorescence. (Photograph courtesy M. Fagg, ANBG)

Organ spacing is a final determinant of shoot appearance. For example, leaves forming a rosette as in Arabidopsis are separated by short internodes compared with longer internodes intervening between whorls of leaves of a blue gum seedling. The resulting morphologies are strikingly different. The question of what determines phyllotaxis and internode length is discussed later.

Shoot meristem anatomy


Figure 7.8 Three-dimensional reconstruction of vegetative shoot apex of lupin showing central dome and spirally arranged leaf primordia on flanks. (Based on Williams 1974)

Shoot meristems are small, with a dome typically of 100–300 µm in diameter consisting of no more than a few hundred cells. In the 1970s, Williams (1974) published elegant reconstructions of shoot meristems derived from serial sections (Figure 7.8) which revealed the extent of variability in the shape and dimensions of shoot apices. However, the overriding organisation is of a central dome with groups of cells partitioned off from its periphery to form either determinate organ primordia or secondary meristems (axillary buds). Some cells in between are not destined to become primordia and will instead later become the internodes of the axis. Different models have been proposed to describe the regions of shoot meristems. In a functional sense, vegetative meristems have three main components: the central zone, peripheral zone and the file meristem zone, all of which tend to disappear or become indistinct in infloresence meristems (Figure 7.9).


Figure 7.9 The structure and zonation of shoot apices can be described in two main ways. Zonation in vegetative apices (top) can be based on relative cell division rates and differential staining; central zone (cz, slow division rate), peripheral zone (pz) and file meristem zone (fmz). Layers of cells are usually also visible. L1 and L2 are often referred to as the tunica and L3 as the corpus. The layering remains visible during early floral development (bottom), but the central zone disappears in determinate inflorescences. (Based on Huala and Sussex 1993)

Superimposed on this functional zonation are usually three distinct cell layers which give rise to separate cell lineages (Kerstetter and Hake 1997). These cell layers, designated L1, L2 and L3, are distinguishable by their positions in the meristem and their pattern of cell divisions, and are evident in both vegetative and reproductive meristems. Surface cells of L1 divide anticlinally while within the meristem and during sub-sequent differentiation of the organs. Not surprisingly, they form the epidermis. Within the apical dome, the plane of cell division within L2 is also purely anticlinal, but later on during organ formation divisions occur in other planes. In contrast, cells of the deepest layer, L3, divide in all planes. The two inner layers, L2 and L3, contribute cells to form the body of the plant with the proportion of cells derived from each layer varying in different organ types. Although the cell lineages produced by each layer usually contribute to distinct regions within each organ, invasion of cell derivatives of one layer into another has been observed.

Invading cells differentiate in accordance with their new position, which we interpret to mean that developmental fate of cells appears to be governed more by position than by cell lineage. However, meristematic cells may already be functionally distinct as evidenced by patterns of gene expression which reflect the layered cellular organisation (Figure 7.10).


Figure 7.10 Complexities of meristem functioning are revealed by specific staining (in situ mRNA hybridisation and antibody techniques) for expression of different genes. Dark shading represents protein patterns, light shading represents mRNA expression. Many of these patterns match the zonations in Figure 7.9. (Based on Meeks-Wagner 1993)

In order to maintain the very precise organisation of vegetative meristems over long periods, or to accommodate rapid changes during flower formation, some signalling process must exist to coordinate division between the cell layers. Evidence for such signalling has been established through the development of chimeras, where genetically different cell types exist together in a single apex, yet still achieve normal developmental patterns.

Phyllotaxis and internode length

We previously raised the question of what determines phyllotaxis and internode length. Organs derived from the shoot meristem can arise in whorls (two or more organs simultaneously at one node), alternately (two files displaced by 180° with a single organ at each node) or in spirals (each organ displaced from the previous one by approximately 137° with a single organ at each node). These organs may be separated by very short or long internodes. Phyllotaxis patterns are usually stable, but often change abruptly with floral induction or when seedlings undergo transition to their mature morphology. In many species of Eucalyptus, this ‘phase change’ from juvenile to adult is very striking and is accompanied by a change from whorled to alternate or spiral phyllotaxis. How is the change in phyllotaxis effected? We gain some insight from experiments on chrysanthemum meristems where application of the inhibitor of polar auxin transport TIBA (tri-iodo benzoic acid) induced changes in internode length and displacement angle between leaf primordia. The data are consistent with a change from 137° spiral (control) to alternate (50 ppm TIBA) phyllotaxis and are presumed to result from increased concentrations of auxin in the meristem resulting from inhibition of transport away from the existing primordia. Meristems are deduced to be sites of auxin synthesis (Schwabe and Clewer 1984). One interpretation is that each primordium acts as a field with a defined radius of inhibition preventing other primordia from initiating too close to it. Another option is mechanical control by pressure and tension gradients within the meristem (Green et al. 1996). Supporting evidence comes from experiments on cells in tissue culture in which applied pressure altered their morphogenetic patterns. At a whole-plant scale, tension and compression wood in trees are further examples of specific developmental responses to physical forces.

Meristems as templates for morphogenesis

The location and activity of individual meristems give rise to the diverse morphologies we recognise within the Plant Kingdom. Palms and grass trees have a distinctive morphogenesis with the entire shoot canopy produced from the activity of a single apical meristem. Removal of the crown of a coconut palm inevitably kills the whole plant. The roots of palms and grass trees are also extraordinary in that they grow and senesce in a seasonal pattern which confers tolerance to poor soils and fire.

In contrast, woody trees produce complex shoot morphologies through combined activity of terminal and lateral apices. We see the product in the height and diverse branching pattern of large trees. Australian eucalypts show a diversity of shoot forms, ranging from the single slender trunk of a mountain ash or karri, topped by a branched canopy, to the multiple trunks of mallee eucalypts. The branching form of mallee species is determined by simultaneous activity of many apical meristems. Similarly, excavation of roots of large trees has often revealed complex branching patterns which enable effective exploration of large volumes of soil and extraction of water and nutrients. In the case of Eucalyptus marginata, a root system arises from strong meristematic activity in the surface levels of the root system as well as proliferation of deep sinker roots. Some species within the Proteaceae can form proteoid (cluster) roots that are adapted to nutrient-poor soils.

Grasses have a distinctive morphology which arises from the local activity of intercalary meristems. These meristems give rise to semi-autonomous plants called tillers which comprise leaves, stems and reproductive parts and are subtended by nodal roots. Cell divisions within the intercalary meristem are developmentally responsible for the characteristic morphology of grasses, a family that is well adapted to herbivory.

7.1.3 - Leaf development

Mary Byrne, School of Biological Sciences, University of Sydney

Leaves are the main photosynthetic organs of vascular plants, and their optimisation for conversion of light to chemical energy production has resulted in a striking array of shapes. There is enormous variation in the shape of leaves of different plant species, of leaves within species, and for some species there is variation between juvenile and mature leaves on an individual plant. In the most general type of dicot leaf, there is a basal stalk or petiole and a distal flat blade or lamina. The leaf forms this structure through development in three axes; the proximal-distal (tip to base), the adaxial-abaxial (top to bottom) and the medial-lateral (middle to margin) (Figure 7.11). The adaxial-abaxial axis is evident at early stages of development of the leaf primordium. The adaxial side of the leaf is the side nearest to the shoot meristem and the abaxial side of the leaf is the side further from the meristem (Figure 7.11). These will eventually become the upper and lower sides of the leaf.


Figure 7.11 Diagram of a plant apex and mature leaf of a dicot such as Arabidopsis. (a) The shoot meristem comprises a population of undifferentiated dividing cells. Cells on the flanks of the meristem contribute to production of leaves. The adaxial-abaxial axis is marked. (b) Mature leaf showing the proximal-distal and medial-lateral axes.

The resulting basic leaf shape is a flat, planar structure but modifications during development result in variations on the basic shape. For example, other shapes include leaves that are flattened along a lateral dimension or are radial, such as tendrils of Pisum sativum (pea) leaves or spines in many cacti. Varying degrees of serrations and lobing at the leaf margins also contribute to differences in leaf shape. For instance, within Australian Banksia, B. grandis leaves have deeply indented margins, whereas B. integrifolia leaves have smooth margins (Figure 7.12).


Figure 7.12 Contrasting leaf shape of (left) Banksia grandis and (right) Banksi integrifolia. (Photographs courtesy M. Fagg and D. Greig, ANBG)

In addition to these types of elaborations, the leaf lamina may be a simple undivided shape or may be complex and subdivided into leaflets. Complex or compound leaves are found in many species including tomato (Solanum lycopersicum) and pea (Pisum sativum) (Figure 7.13).


Figure 7.13 Leaf shape variation. Banksia serrata leaf has serrated margins (A) and B. integrifolia leaf has smooth margins (B). The leaf of Arabidopsis thaliana is simple with small marginal serrations (C). The leaf of tomato is compound and a single leaf has primary and secondary leaflets (D). The leaf of pea is compound and includes central leaflets, proximal stipules and distal radial tendrils (E).

Different leaf shapes may be produced during the life of the plant, and may be influenced by environmental conditions. For instance many Eucalyptus species have short, broad juvenile leaves and narrow, elongated adult leaves. Leaves in Acacia species change with plant maturity where juvenile leaves have a dissected leaf lamina and adult leaves (which are a modified petiole, or phyllode), are simple in shape (Figure 7.14).


Figure 7.14 Branch of an Acacia species showing two leaf shapes on the one plant. The juvenile leaves are compound leaves and are subdivided into leaflets. The adult leaves (phyllodes) are simple in shape.

Leaf initiation

Leaves initiate in peripheral regions of the shoot meristem (Figure 7.15). Sites of initiation occur where the hormone auxin forms a localised maximum. Local high levels of auxin are achieved through the action of a membrane-localised auxin efflux carrier PIN-FORMED1 (PIN1). The distribution of PIN1 to the plasma membrane on just one side of cells leads to directed flux of auxin away from sites where organs have already initiated and toward the site where organs will next form (Figure 7.15). Coincident with high levels of auxin, class I homeodomain transcription factor KNOX genes are down-regulated in initiating organs. Class I KNOX genes are expressed in the shoot meristem and are essential to maintain growth of the shoot meristem (Figure 7.15).


Figure 7.15 Leaf initiation and shape is determined by auxin and KNOX genes. (a) Diagram of the shoot apex with central meristem, incipient leaves I1-I3, and initiated leaves P1-P5. Arrows indicate the flow of auxin to the site where a new leaf will initiate. (b) in situ hybridisation showing expression of the class I KNOX gene SHOOT MERISTEMLESS in the shoot meristem and down-regulation in initiating and developing leaves. (c) Young Arabidopsis wild type seedling showing two cotyledons and first leaves. (d) The class I KNOX gene mutant shoot meristemless (stm) has 2 cotyledons and fails to produce leaves due to lack of a shoot meristem.

Each initiating leaf primordium is flanked by a boundary of slowly dividing cells and this boundary serves to delineate the leaf from the meristem and from adjacent leaf primordia. Several members of a transcription factor gene family known as CUP-SHAPED COTYLEDON (CUC) are expressed at this boundary and downregulation of CUC genes results in fusion of adjacent organs. This reflects the importance of the meristem-organ boundary in leaf development. As we see below, auxin, as well as the KNOX and CUC genes are also involved in determining leaf shape.

Proximal-distal axis

Following initiation leaf primordia go on to develop as discrete organs through cell division, cell expansion and cell differentiation. Leaf primordia become defined morphologically through cell divisions that lead to outgrowth from the meristem and hence formation of a proximal-distal axis. In the model species Arabidopsis thaliana (Arabidopsis), as the leaf grows a gradient of maturation occurs along this axis so that cells cease division and undergo terminal differentiation firstly in the distal region and lastly in the proximal region of the leaf. The leaf proximal-distal axis may have distinguishing features of lamina and petiole as in Arabidopsis and many dicotyledonous plants. One variation of this shape occurs in monocotyledons, which includes the grasses such as wheat, barley, rice and maize. Strap-like leaves of many monocotyledons have a distal blade and a proximal sheath, which wraps around the stem. The blade and sheath are separated by specialised cells that form a hinge, and this hinge allows the blade to bend away from the sheath and stem to be exposed to light.

Adaxial-abaxial axis

The adaxial-abaxial axis is evident at early stages of development of the leaf primordium. The adaxial side of the leaf is the side nearest to the shoot meristem and the abaxial side is furthest from the meristem (see Figure 7.11 above). Experimental separation of initiating leaf primordia from the shoot meristem results in loss of adaxial fate. Amazingly this loss of adaxial fate leads to development of an abaxial, radial leaf. This indicates that adaxial fate is determined either by contact with the meristem, or by signalling from the meristem.

Most flat leaves have adaxial-abaxial polarity where the two sides of a leaf have distinct morphological features. On the leaf surface, adaxial and abaxial epidermal cell size and shape, stomata density, or the presence of elaborated cell types such as hairs may be different. Vasculature often has a discrete adaxial-abaxial polar arrangement with adaxial water-conducting xylem and abaxial nutrient-conducting phloem (Figure 7.16a). Internal mesophyll cells of the leaf may also display differential arrangement along the adaxial-abaxial axis. In some species, the adaxial domain has closely aligned and chloroplast-rich palisade mesophyll cells whereas the abaxial domain has dispersed spongy mesophyll cells. Mature leaves are typically oriented with the adaxial side facing the light and the abaxial side away from the light. As such, distinct cell types along the adaxial-abaxial axis serve to optimize light capture and gas exchange for photosynthesis.

On top of this layer of regulation, adaxial and abaxial fates are each determined by specific genes. A reduction in expression of genes required for either adaxial or abaxial fate results in development of leaves that comprise only the opposite fate. Furthermore these leaves are radial. Together such observations have led to a fundamental tenet of leaf development, whereby adjacent adaxial and abaxial tissue types are essential to development of an adaxial-abaxial axis and hence to formation of a flattened leaf.


Figure 7.16 Leaf adaxial-abaxial fate. (a) Cross section through the mid-vein region of an Arabidopsis leaf. Tissues of the leaf including vasculature show distinct adaxial (upper surface) and abaxial (lower) differences. (b) Diagram showing adaxial and abaxial gene interactions. Solid lines indicate direct repression, dashed lines indicate indirect repression.

Key genetic determinants of adaxial fate include class III HD-ZIP transcription factors (Figure 7.16b). These proteins have a homeodomain, which is involved in binding DNA, a leucine zipper domain, which is involved in protein-protein interaction, and a sterol-binding domain that is predicted to bind small lipid molecules. In Arabidopsis, there are five genes encoding class III HD-ZIP transcription factors. Three of these genes, PHABULOSA (PHB), PHAVOLUTA (PHV) and REVOLUTA (REV), have a significant role in leaf development. Transcripts of all three genes are confined to the adaxial domain through transcriptional and post-transcriptional repression in the abaxial domain. For example, PHB, PHV and REV transcripts are all targets of the microRNAs miR165 and miR166, which are expressed on the abaxial side of the leaf. In mutants with reduced class III HD-ZIP activity adaxial fate diminishes. Conversely, gain of class III HD-ZIP expression in the abaxial domain results in conversion of the abaxial side of the leaf to adaxial fate.

Two other transcription factors, ASYMMETRIC LEAVES1 (AS1) and ASYMMETRIC LEAVES2 (AS2), form heterodimers and promote adaxial fate, in part through repression of miR165/166. Abaxial fate is promoted by additional factors, including KANADI (KAN) proteins, which are part of a plant-specific group of transcription factors. In Arabidopsis, four KAN genes have overlapping roles in leaf development. KANADI proteins are expressed on the abaxial side of the leaf and are necessary and sufficient for specifying the abaxial fate. It is also known that KAN1 directly represses AS2. Other abaxial factors are AUXIN RESPONSE FACTOR (ARF) proteins, which bind to promoter sequence elements in auxin responsive genes and regulate plant response to auxin. The two ARF genes, ARF3 and ARF4, are repressed on the adaxial side to the leaf by a set of small RNAs known as tasiR-ARF. The AS1-AS2 module directly represses ARF3. Thus the emerging picture is that repressive interactions between adaxial and abaxial genes helps to set up the adaxial-abaxial axis of the developing leaf (Figure 7.16).

Mediolateral axis

The mediolateral axis of leaves extends from the margin to the mid-region of the leaf, which is characterised by the presence of the main vein or midrib. Most leaves are bilaterally symmetrical and the two halves of the leaf, from margin to midrib, form mirror images. Leaf width is determined by the degree of lateral expansion along this axis. However, the relationship between the leaf margin and the midrib depends on cell recruitment from the shoot meristem during the process of initiation. Leaves of dicots initiate from a limited region of the shoot meristem so that the margins arise from cells close to the mid-region and lateral expansion of the leaf occurs post-initiation. On the other hand, leaves of monocots such as grasses, recruit cells from around the entire circumference of the shoot meristem. In this case, the leaf margins are derived from cells that are distant from the mid-region and consequently this early lateral expansion results in the leaf primordium wrapping around the entire shoot meristem.


Relatively simple modifications of the basic leaf shape are achieved through changes in growth dynamics at the leaf margins. In this way, leaves can develop marginal teeth or serrations or may be lobed. Leaf shape is more dramatically modified in compound leaves, where the blade is divided into reiterating units or leaflets. Multiple orders of subdivision occur when leaflets themselves are further subdivided into leaflets, for example in the compound leaves of tomato, which may have primary and secondary leaflets. Surprisingly, studies on the simple, serrated leaf of Arabidopsis, the lobed leaf of the Arabidopsis close relative Cardamine hirsuta, and the compound leaf of tomato have found that genes involved in setting up leaf initiation from the shoot meristem also function to modify the shape of the leaf lamina. The class I KNOX genes are important for modified leaf shape in many species. These genes are not expressed in the simple leaf of Arabidopsis but are expressed in the lobed leaf of Cardamine and the compound tomato leaf. Misexpression of class I KNOX genes in Arabidopsis leaves generates lobed leaves and increasing expression of these genes in Cardamine and tomato results in more lobing or more leaflets. Likewise, decreasing class I KNOX gene levels in Cardamine leads to a more simple leaf shape. Thus the degree of leaf complexity can depend on the expression of growth promoting class I KNOX genes. In addition to KNOX genes, auxin and CUC genes play central role in elaboration of leaves. In compound leaves of tomato and lobed leaves of Cardamine, high levels of auxin occurs at discrete sites along the developing leaf margin, where leaflets or lobes will form. Furthermore CUC gene expression marks the boundaries between initiating leaflets. At these sites CUC genes act to repress growth thereby promoting separation of leaflets. Altering the spatial distribution of auxin or CUC in the young developing leaf leads to reduced leaflet production. Less intricate elaborations of leaf shape, such as leaf serrations, also involve auxin and CUC genes. In developing leaves of Arabidopsis, the margins form periodic auxin maxima flanked by boundaries of CUC gene expression (Figure 7.17).


Figure 7.17 Modifying leaf margins. Serrations in Arabidopsis leaves occur at sites of auxin maxima. The PIN1 protein (straight arrows) promotes formation of an auxin peak. PIN1 is positively regulated by auxin and CUC2 (curved arrows). High levels of auxin in turn represses CUC2.

Together this promotes the formation of outgrowths at the leaf margin and so modulates leaf shape. Differences between the growth potential and intricacy of interactions between transcription factors and hormones likely determine whether the leaf will develop with modest outgrowths in the form of serrations, more prominent lobes or establish a compound leaf with distinct leaflets.

Further reading on leaf development

Bar M, Ori N (2015) Compound leaf development in model plant species. Curr Opin Plant Biol 23: 61-69

Bowman JL, Floyd SK (2008) Patterning and polarity in seed plant shoots. Ann Rev Plant Biol 59: 67-88

Braybrook SA, Kuhlemeier C (2010) How a plant builds leaves. Plant Cell 22: 1006-1018

Byrne ME (2005) Networks in leaf development. Curr Opin Plant Biol 8: 59-66

Byrne ME (2006) Shoot meristem function and leaf polarity: the role of class III HD-ZIP genes. PLoS Genet 2: e89

Chitwood DH, Nogueira FT, Howell MD et al. (2009) Pattern formation via small RNA mobility. Genes Dev 23: 549-554

Efroni I, Eshed Y, Lifschitz E (2010) Morphogenesis of simple and compound leaves: a critical review. Plant Cell 22: 1019-1032

Fukushima K, Hasebe M (2014) Adaxial-abaxial polarity: the developmental basis of leaf shape diversity. Genesis 52: 1-18

Hay A, Tsiantis M (2010) KNOX genes: versatile regulators of plant development and diversity. Develop 137: 3153-3165

Hepworth SR, Pautot VA (2015) Beyond the divide: Boundaries for patterning and stem cell regulation in plants. Front Plant Sci 6: 1052

Machida C, Nakagawa A, Kojima S et al. (2015) The complex of ASYMMETRIC LEAVES (AS) proteins plays a central role in antagonistic interactions of genes for leaf polarity specification in Arabidopsis. WIRES Develop Biol 4: 655-671

Mentink RA, Tsiantis M (2015) From limbs to leaves: common themes in evolutionary diversification of organ form. Front Genet 6: 284

Sluis A, Hake S (2015) Organogenesis in plants: initiation and elaboration of leaves. Trends Genet 31: 300-306.