Seed

General Aspects

Plants are sessile organisms and use development as a strategy to adapt to their environment. In contrast, animals are motile and can exploit their environment by moving around it. Therefore, animals have a rather rigid developmental program, whereas plant development is much more flexible in order to take environmental variation into account. Higher plants have a relatively simple and flexible basic body organization with a highly developed capacity for indeterminate growth. This means that the final shape and size of an adult plant can often be significantly modulated by the environment. A major factor behind the indeterminate growth habit of plants is their sessile nature, which means that an individual plant is unable to move once it begins to develop. In order to exploit their local habitats most efficiently, plants require great flexibility in their growth form and their responses to the many biotic and abiotic opportunities and threats they face during their lifetimes.

Plants have highly flexible growth patterns. Meristematic regions in a plant contain rapidly dividing, undifferentiated cells that remain active throughout its lifetime. Although they tend to have characteristic adult forms, plants do not have anything like the preprogrammed body plan of animals. Differentiation in animals is almost always irreversible and is specified very early in development. This means that adult body shape and size is more or less fixed, with the result that mature members of the same animal species tend to look very similar to each other. In plants, some features might be relatively constant, such as leaf shape and branching patterns (opposite, alternate, etc.) but the exact shape of the mature organism can be highly unpredictable. Thus, a plant like a beech can vary from being a small contorted shrub less than 1 m high if it grows on a rocky, windswept hillside to a magnificent 50 m tree growing in a more fertile and sheltered habitat.

One reason for this difference is that plants continue to grow throughout their life via multiple meristematic regions. This capacity for indeterminate growth in plants is widely exploited in practices, such as pruning, coppicing, pollarding, and the use of cuttings in commercial and domestic plant husbandry. In contrast, animal growth tends to be allometric whereby all parts of the body grow such that its overall proportions are maintained. Hence, highly active apical meristems located at the root or shoot apex tend to give rise, respectively, to relatively tall and deeply rooted plants, whereas more active lateral meristems will lead to a shorter, bushier type of plant. Nowadays biotechnologists can take even greater advantage of the plasticity of plant growth by producing novel forms that are both higher yielding and more amenable to management and harvesting. A recent example is the development of new commercial apple varieties that are no more than 2 m high and grow like vines, but produce far more fruits than traditional apple trees.

The plant embryo provides a basic body plan for the seedling, but postembryonic growth generates an entirely new adult body plan based on the continuous proliferation of shoot and root meristem populations. The contrasting lifestyles of animals and plants are also reflected in their reproductive strategies. Most animals set aside a germ line early in embryonic development, and this is the sole source of gametes. In contrast, there is probably no germ line in the plant embryo. The somatic cells of plants can switch from vegetative to reproductive growth during postembryonic development.

Cell Fate and Potency

The plasticity of plant development is most clearly demonstrated by the ability of differentiated plant cells to regenerate into whole plants when isolated and placed in the appropriate culture medium. This shows that many plant cells remain totipotent, i.e., they are able to recapitulate the entire developmental pathway and produce all the normal cell types of a complete plant. Both embryonic and postembryonic development can be recapitulated (somatic embryogenesis, organogenesis), although the capacity for these processes differs from species to species and shows strong genotype dependence. In contrast, the potency of animal cells is progressively restricted during development. No matter how many embryonic signaling proteins and regulatory factors are added to the culture medium, human liver cells will never give rise to a human embryo.

The molecular basis of this distinction between animals and plants reflects the mechanisms used to establish cell fate. The fate of a cell is the sum of cell types that cell and its progeny will produce during normal development (development unperturbed by experimental manipulation). In both animals and plants, cell fate is restricted throughout development by a series of hierarchical decisions. In plants, however, these decisions are reversible because they depend on the position of the cell and signals from surrounding cells. Epidermal cells normally arise from the L1 layer of the embryo, and normally remain as epidermal cells throughout the lifetime of the plant. However, an unusual periclinal cell division that causes an epidermal cell to invade the underlying cortex results in the respecification of the epidermal cell as a cortical cell. The fact that this cell comes from a different lineage is not important; it is the new position that counts. In animals, position is also important in development, but there comes a time when each cell becomes determined in its developmental pathway, which means it becomes irreversibly committed to its fate. Under these circumstances, moving the cell to a new site has no effect on that cell – it will carry on in its determined path. The fate of the cell then depends on its lineage (i.e., developmental decisions that were made in its past) rather than on its interactions with its neighbors. Determination and differentiation each results from changing patterns of gene expression. Put simply, these patterns of gene expression become ‘locked’ in animal cells, through a variety of genetic and epigenetic mechanisms, while in most plant cells they are labile.

Pattern Formation and Morphogenesis

In contrast to pattern formation, there are some major differences between plants and animals in terms of morphogenesis. In animals, the formation of cell and tissue structures is strongly dependent on cell–cell interactions, particularly the gain and loss of cell adhesion. This allows cells to move relative to each other either as whole tissues or as individual migrating cells. One of the landmark events in animal development – gastrulation – is driven almost entirely by the relative movement of cells. In contrast, the rigid plant cell wall and the way cells are tightly cemented together means that such mechanisms cannot be used in plants. There is no equivalent to gastrulation in plant development. Changes of shape and structure in the developing plant are, therefore, mediated by alternative mechanisms, such as controlling the plane and symmetry of cell division, and by increasing and decreasing the size of cells by varying the amount of water in the vacuole.

The structure and organization of plant cells also means that cell–cell communication in development is different to that seen in animals. Animal cells lack a cellulose wall and bristle with receptors for all manner of molecules. Furthermore, certain signaling molecules can diffuse through the plasma membrane and interact directly with intracellular receptors. The diversity of signaling pathways is somewhat more limited in plants, and cell–cell communication often occurs through plasmodesmata, the cytoplasmic threads that connect plant cells into a continuous symplasm.

The Use of Signal Gradients for Axial Patterning in the Plant Embryo

The two major developmental processes occurring in plant embryonic development are the specification of regions along the apical–basal axis (axial patterning) and the specification of the three radial cell layers (radial patterning). Before discussing these processes in detail, however, a brief overview of embryonic development in flowering plants is necessary. Fertilization is followed by elongation of the zygote along the future apical–basal axis of the embryo, which corresponds to the shoot–root axis of the seedling. The elongated zygote divides asymmetrically to produce a small apical cell (which gives rise to most of the embryo) and a larger basal cell (which forms part of the root as well as the suspensor that attaches the embryo to the ovule). In order to form the suspensor, the basal cell divides many times in the same plane, while the apical cell undergoes a stereotyped sequence of horizontal and vertical divisions to generate a ball of cells called the proembryo. As cell divisions continue, the embryo is organized into regions representing the layout of the seedling – the SAM, the cotyledon(s), the epicotyl, hypocotyl, radicle (embryonic root), and root apical meristem (RAM). The embryo also becomes radially organized into three fundamental layers that pervade the adult plant – L1, L2, and L3. The identification of mutations in Arabidopsis thaliana and other plants that perturb these processes has shown that they are largely independent.

In animals, axial patterning of the embryo is often dependent on graded signals. Mutations in the genes that provide these signals or respond to them generate patterning defects in which, characteristically, parts of the axis are missing. Very similar mechanisms appear to be involved in the patterning of the plant embryo. Polarized auxin transport has been shown to play an important role in establishing the apical–basal axis, and mutations in the genes that are required either to set up this gradient or respond to it generate phenotypes in which the apical, central, or basal regions of the embryo are affected. For example, the MONOPTEROS (MP) gene is usually expressed in the bottom tier of cells of the proembryo, which gives rise to the hypocotyl, radicle, and initial cells of the RAM. In mp mutants, these structures are absent. The product of the MP gene is a transcription factor that binds to auxin-responsive elements, so the mp mutant is likely to be insensitive to the basal-to-apical gradient of auxin. The product of the GNOM (GN) gene is expressed just after fertilization and appears to be important for the localization of auxin efflux pumps, such as PIN1, specifically in the basal membranes. In gn mutants, the zygote divides symmetrically rather than forming apical and basal cells of different sizes. There is no progressive localization of the PIN1 auxin transporter to basal membranes and no accumulation of auxin in the basal region of the embryo. At the heart stage, gn mutants are more rounded than wild-type embryos, and the RAM is not specified. Note that not all axial patterning mutants affect auxin transport or perception. For example, the defect in the gurke (gk) mutant, in which the apical region of the embryo is absent, appears to reflect abnormal control of cell division.

Balancing Proliferation and Differentiation in the SAM

Meristems are small populations of rapidly proliferating cells that produce all the adult organs of a flowering plant. Two meristem populations are established in the embryo, the SAM (which gives rise to the aerial parts of the plant) and the RAM (which gives rise to the root system). The SAM comprises a central zone of slowly dividing mother cells surrounded by a peripheral zone of rapidly dividing apical initial cells. As cells leave the meristem proper, they enter the rib meristem, where files of cells are laid down to extend the shoot. Adjacent to the rib meristem are the lateral morphogenesis zones, where organ primordia begin to differentiate. The spatial arrangement and timing of lateral organ initiation is discussed later.

The patterning mutants discussed above, such as mp and gk, produce plants in which either the SAM or the RAM is never established, and there is consequently a complete failure of postembryonic development. In other mutants, the meristems are established correctly, but there is a defect that prevents cell division. This occurs in the shoot meristemless (stm) mutant of Arabidopsis and the knotted-1 (kn1) mutant of corn (Zea mays; maize). In Arabidopsis, STM gene expression begins in the embryo at the early globular stage. Null mutations cause the arrest of development after rudimentary cotyledons have formed, indicating that the meristem cells have differentiated prematurely. In postembryonic development, STM is expressed strongly in the apical initial cells but is downregulated in the rib meristem and morphogenesis zones, where differentiation takes place. Weaker stm mutants produce shoots that give rise to a few leaves and flowers with reduced numbers of floral organs. This indicates that the meristem is ‘consumed’ during the formation of leaves so that a smaller number of cells are available for floral organ formation.

Other mutations cause the opposite phenotype, i.e., a larger meristem. For example, the PRIMORDIA TIMING (PT) gene influences meristem size in the embryo, and strong mutants show progressively increasing meristem size from the heart stage until germination. Loss-of-function mutants for the CLAVATA1 (CLV1) gene show increased meristem size during postembryonic development. It appears that the function of this gene is to recruit cells leaving the meristem into organogenesis. CLV1 encodes a receptor kinase and functions in a signaling pathway that also includes the genes CLV2 and CLV3. Mutations in two other genes, MGOUN 1 (MGO1) and MGO2 cause cells leaving the central zone of the SAM to accumulate in the peripheral zone of initial cells rather than forming organ primordia. All these data suggest a model in which STM maintains the meristem by blocking differentiation, which is facilitated by the CLAVATA and MGOUN genes acting sequentially (Figure 1). The WUSCHEL gene, which is expressed as early as the 16-cell stage, may be required to maintain meristem identity in the postembryonic meristem since in wus mutants, the presumptive meristem cells differentiate but are not recruited into lateral organs. The function of WUS in the embryo prior to the formation of the SAM is unknown. Another gene, ZLL, is required to maintain the meristem population in the embryo and probably acts upstream of STM because in zll mutants, STM expression is not maintained.

Figure 1. Genes involved in the maintenance of the shoot apical meristem (SAM). WUS is required to maintain the population of stem cells. STM inhibits the differentiation of stem cells, and its activity is opposed by genes of the CLAVATA (CLV) and MGOUN (MGO) families.

Lineage versus Position in the Specification of Cell Fates in the Shoot

The specification of cell fates in the SAM has been investigated using cell chimeras, i.e., plants in which the meristem contains two genetically and phenotypically distinct cell populations. In periclinal chimeras, the distinct cells are found in different radial layers. Most of the cells in the outer two layers (L1 and L2) undergo anticlinal divisions, and remain within the same layer, but there are occasional periclinal divisions, which allow cells from one layer to invade the other. Under these circumstances, the fate of the invading cell is respecified, i.e., it begins to behave in the same manner as its neighbors rather than its forebears. As discussed earlier, this indicates that the radial fate of cells in the stem is maintained by cell–cell contacts rather than being dependent on lineage.

In mericlinal chimeras, segments of the meristem (comprising all three radial layers) are genetically marked, and these can be used to investigate the specification and determination of organ primordia. Although the details differ between different species, a common factor is that the fate map of the meristem is organized roughly as a series of concentric circles, with the outer circles representing nodes near the base of the stem and the inner circles representing nodes nearer the growing tip. There are always more cells around the edge of a circle than in its center, so marked cell clones of equivalent size at the periphery and in the center of the meristem fate map behave differently. At any stage of development, a small clone of marked cells from the periphery of the meristem will generate a marked sector of tissue in a basal leaf or stem, whereas a similar clone of marked cells from the center will probably give rise to several apical leaves derived entirely from marked cells (Figure 2).

Figure 2. Cell fate in the shoot apical meristem (SAM). Marked cell clones from the periphery of the SAM fate map give rise to older, lower nodes, while those from the center give rise to younger, more apical nodes. Peripheral cells may contribute to one leaf or a sector within one leaf, whereas central clones may cover several phytomers.

Meristem fate maps have been produced in corn and Arabidopsis. The growth of corn plants is determinate, i.e., there is a predetermined number of nodes (18 plus a terminal spike and tassel, representing the male reproductive organs). Conversely, the growth of Arabidopsis is indeterminate, i.e., the number of leaves is not fixed. However, the determinacy of the corn SAM is reversible. The SAM removed from a corn plant after the formation of six or seven leaves can be deprogrammed in culture and will regenerate into an entire plant, not just the terminal portion of the plant from which it was removed. Therefore, it is not only radial cell types that are maintained by signaling, but also the positional identity of cell types within the meristem itself.

Lineage versus Position in the Specification of Cell Fates in the Root

The root system of the plant derives from the RAM, which is established at the basal pole of the embryo, and derives partly from the embryo proper and partly from the terminal cell of the suspensor (known as the hypophysis). While the SAM and RAM have similar functions, there are also some important differences between them. The SAM not only elongates the stem, but also gives rise to primordia that form branches and lateral organs (leaves, thorns, etc.). Conversely, lateral roots arise from the differentiated primary root well behind the growing tip. Also, the RAM gives rise not only to the differentiated cells of the root (which lie behind the meristem), but also to a distal root cap that is continually replenished as the cells are sloughed off by abrasion.

The RAM is organized in a similar way to the SAM with a quiescent center surrounded by proliferative initial cells that give rise to the file meristem and the distal root cap. Together, the quiescent center and the initial cells comprise the promeristem. The Arabidopsis promeristem has three layers. The lower layer comprises 12 central cells that produce the root cap and 16 surrounding cells that generate the lateral root cap and the epidermis. The middle layer comprises the four cells of the quiescent center and eight surrounding cells that give rise to the endodermis and cortex. Finally, the upper layer of stele initials gives rise to the pericycle and vascular bundles (Figure 3). Unlike the situation in the shoot, the division and differentiation of cells in the root meristem is invariant. For example, the cortical/endodermal (CE) initial cells divide to produce one initial cell and one daughter cell. The daughter cell then goes on dividing to produce one cortical cell and one endodermal cell, which enter the file meristem. The process is repeated throughout root growth (Figure 4(a)).

Figure 3. Organization and activity of the root apical meristem. The different cell types in the file meristem and mature root are shown in different shades and are labeled as follows: Ep, epidermis; C, cortex; End, endodermis; P, pericycle; V, vasculature. The cells of the root cap are labeled as follows: LRC, lateral root cap; RCC, root cap columella. They are derived from the initial cells of the promeristem, which are identified as follows: SI, stele initials; PI, pericycle initials; CEI, cortical/endodermal initials; ELI, epidermal/lateral root cap initials; QC, quiescent center; RCI, root cap columella initials.

Figure 4. Laser ablation studies in root development. Left panel shows schematic representation of the promeristem (CEI, cortical/endodermal initials; QC, quiescent center; PI, pericycle initial; VI, stele (vascular) initials). Cell layers in the file meristem are shown above the corresponding initial cells and are identified at the top of the figure. (a) In normal development, each CEI gives rise to a daughter cell (D), which divides to form cortical and endodermis cells (C, E) adding to existing files. (b) If the CEI cell is ablated, an unexpected division of the PI cell replaces it. (c) If the daughter cell is ablated, the CEI undergoes an extra division to replace it. Wavy lines indicate squashed remains of ablated cells.

Reproduced with permission from Twyman, R.M., 2000. Instant Notes in Developmental Biology. BIOS Scientific Publishers, Oxford.

Ablation experiments have shown that although the cell divisions are stereotyped, cell fates are still dependent predominantly on position rather than lineage. For example, the ablation of one CE initial cell is followed by the unexpected division of an adjacent pericycle initial cell, so that the CE initial is replaced by a pericycle daughter cell (Figure 4(b)). The replacement cell then behaves in the same manner as a CE initial. If root development were lineage dependent, the pericycle initial would define a separate lineage (i.e., it would only be capable of producing pericycle cells), and such a replacement would not occur. If the daughter cell of a CE initial is ablated, the underlying CE initial cell divides again to replace it (Figure 4(c)). However, if three or more daughter cells lying above the CE initial are ablated, the CE initial becomes unable to divide asymmetrically. This shows that cell behavior is determined by signals from adjacent cells in the same file. Recently, it has been shown that these signals may be constrained within layers of cells by channeling through plasmodesmata.

Lateral Inhibition to Produce Regular Spacing Patterns

The periodic initiation of lateral organs by the SAM defines a developmental unit known as a plastochron, which gives the shoot time to extend before the next organ is initiated. Shoots are thus composed of nodes and internodes, the nodes bearing the lateral organs (usually leaves). A simple model for the regular initiation of lateral organs is that the formation of one primordium inhibits the formation of the next until a certain time has passed. This can be demonstrated by removing organ primordia as soon as they have formed. For example, if a new leaf primordium is ablated in the lupin (Lupinus albus) shoot apex, the next primordium emerges as expected, but the position of the one after that is altered. It is shifted toward the site of excision. This strongly suggests that lateral inhibition is used to delay the emergence of organ primordia to ensure the correct spacing of leaves. This also confirms that cell fates in the SAM are labile until just before the emergence of each primordium and are maintained by signaling between cells.

Regular spacing patterns are common in development, and in plants one example is the regular spacing of trichomes on the dorsal surface of the leaf. Trichomes are derived from epidermal cells, and all epidermal cells have the capacity to form these structures. In animals, it has been shown that regular spacing patterns can arise spontaneously in fields of equivalent cells through the following mechanism. Initially, all the cells are undifferentiated and send out signals to their neighbors to suppress differentiation. However, if one cell begins to differentiate, its inhibitory signals become stronger, and its ability to respond to inhibitory signals from adjacent cells begins to decline. The system becomes unstable due to random fluctuations that transiently increase or reduce the inhibitory signals produced from each cell. Eventually, individual cells become dominant over their neighbors, and the spacing of these cells depends on the range of the signal (Figure 5).

Figure 5. How spacing patterns are produced by lateral inhibition. (a) Initially all cells are undifferentiated (white) and inhibit each other's differentiation equally (T-bars). (b) Random fluctuations in the level of inhibition then cause some cells to assert dominance and cause the surrounding cells to lose their inhibitory activity. (c) Eventually, the dominant cells differentiate and completely inhibit the surrounding cells.

Reproduced with permission from Twyman, R.M., 2000. Instant Notes in Developmental Biology. BIOS Scientific Publishers, Oxford.

The Arabidopsis leaf is a useful model to study the genetic basis of trichome spacing because the trichomes are very distinct, and they are dispensable when plants are grown in the laboratory. Genes have been identified that regulate trichome differentiation at various stages, from initiation to final morphogenesis. Those acting early in the pathway are the most interesting from a spacing perspective, and they include TRYPTICHON (TRY); GLABROUS1 (GL1), GL2, GL3; and TRANSPARENT TESTA GLABRA (TTG). Mutations in TRY cause the formation of clusters in which four or five adjacent epidermal cells can form trichomes. TRY therefore appears to be involved in the lateral inhibition process that normally causes the trichomes to be interspersed by four or so epidermal cells. Mutations in GL1, GL2, GL3, and TTG reduce the number of trichomes, suggesting that these genes promote trichome development. However, weak ttg and gl1 alleles also produce trichome clusters, so TTG and GL1 also appear to play a role in trichome spacing. A recent model suggests that GL1 and TTG form a trimeric complex that contains either GL3 or TRY. The balance of the availability of GL3 and TRY would therefore determine whether the complex promoted or inhibited trichome development through the production of a diffusible inhibitory signal (Figure 6). The model is attractive in its simplicity but is complicated by the existence of a number of other regulators, such as the MYB-like protein AtMYB23 that functionally overlaps with GL1 and the putative negative regulators CAPRICE (CPC) and COTYLEDON TRICHOME1 (COT1).

Figure 6. A model for the equilibrium reactions that establish trichome fate in the developing leaf. The upper box symbolizes a cell that has exceeded a threshold of trichome-promoting activity and is inhibiting an adjacent, lower cell symbolized by the lower box. Arrows are shown between components that can disturb the equilibrium, either favoring trichome formation or inhibiting it. The major complexes are boxed and both contain the GL1 and TTG gene products. The trichome-promoting complex also contains GL3 while this is replaced by TRY in the trichome-inhibiting complex.

Reproduced from Szymanski, D.B., Lloyd, A.M., Marks, M.D., 2000. Progress in the molecular genetic analysis of trichome initiation and morphogenesis in Arabidopsis. Trends Plant Sci. 5, 214–219, with permission of Elsevier Ltd.

Integration of Multiple Environmental Signals into One Developmental Pathway

The transition from vegetative growth to flowering has to be timed correctly in order to maximize reproductive success, especially in outcrossing species where flowering must be synchronized. The initiation of flowering is controlled by both an autonomous pathway that acts as a developmental clock and exogenous pathways that respond to different cues in the environment. Many of the components of these pathways have been identified in Arabidopsis, a facultative long-day plant (i.e., a plant where flowering can occur at any time, but is promoted on long days). The transition to flowering represents a paradigm in signal integration. It appears that all the signals, internal and external, converge on three so-called floral integrator genes that act as master switches in reproductive development. These are LEAFY (LFY), SUPPRESSOR OF OVEREXPRESSION OF CO 1 (SOC1), and FLOWERING LOCUS T (FT).

An autonomous pathway is required to prevent flowering before the plant has undergone sufficient vegetative growth to enjoy reproductive success. Mutations that show late-flowering phenotypes under both long- and short-day conditions identify proponents of the autonomous pathway. These include genes such as FLOWERING LOCUS CA (FCA), FLOWERING LOCUS PA (FPA), and LUMINIDEPENDENS (LD). The common molecular feature of such mutants is the accumulation of the mRNA for another gene, FLOWERING LOCUS C (FLC), whose product is an MADS box transcription factor that inhibits (probably indirectly) the three floral integration genes listed above. The inhibition of FLC is, therefore, a central step in the transition to flowering.

A long period of cold temperature (below 10 °C) promotes flowering in many species including Arabidopsis and is a useful reproductive strategy because it delays flowering until the spring. This vernalization response is epistatic to the autonomous pathway, and this reflects the fact that vernalization response genes such as FRIGIDA (FRI) also regulate the expression of FLC. The expression of FLC is downregulated (indirectly) by FRI at low temperatures and then maintained in a repressed state when the weather gets warmer (and when FRI activity is reduced) by epigenetic mechanisms dependent on genes such as VERNALIZATION RESPONSE 2 (VRN2), which encode chromatin remodeling proteins.

Flowering time in Arabidopsis is promoted under long-day conditions, which means that genes for light perception and circadian rhythm must be involved. Information converges on the CONSTANS (CO) gene, whose product is active only under long-day conditions. The CO protein is a transcription factor that directly regulates two of the floral integrator genes (FT and SOC1). In co mutants, flowering is delayed under long-day conditions, but occurs normally under short-day conditions, showing that the autonomous pathway is not affected. Other genes have been identified that act as repressors in the light response pathway, e.g., EARLY FLOWERING 3 (ELF3), ELF4, and EARLY FLOWERING IN SHORT DAYS (EFS). Other environmental signals also affect flowering, including overcrowding, heat, drought, nutrient depletion, and changes in light quality. It is unclear how many of these signals are integrated although some may result in the increased synthesis of plant hormones. All plant hormones that have been identified thus far have been shown to influence flowering time, although only the gibberellin pathway, which promotes flowering through gibberellin response elements in the LFY promoter has been investigated in any detail.

Positional Information in the Flower as a Paradigm of Pattern Formation

Floral integration, described above, results in the conversion of the SAM into a floral meristem. The latter is a determinate structure that no longer gives rise to periodic lateral organs, but to concentric whorls of floral organs (sepals, petals, stamens, and carpels) that are required for reproduction. Despite their very different structures and functions, all these organs begin as undifferentiated cell clumps on the surface of the floral meristem. Therefore, an important aspect of flower development is pattern formation, i.e., the regional specification of cells in the meristem so that each forms the correct structure in the correct place.

Many mutants that show floral patterning defects have been identified, but the most interesting class is the homeotic mutants, in which one whorl of floral organs develops with the likeness of another. This is analogous to the situation in Drosophila, where certain mutations cause body segments to take on incorrect regional identities, resulting in mutants with misplaced appendages (e.g., legs growing out of the head instead of eyes). The similarities go further in that in both Drosophila and plants, the genes responsible for homeotic mutations encode transcription factors that appear to provide each cell with a ‘positional code’ to confer its regional identity. In plants, these are known as floral organ identity genes and many (but not all) of them encode transcription factors of the MADS box family.

The control of floral organ identity is explained by the ABC model, in which each whorl of floral organs is defined by a combination of three classes of genes: A, B, and C. Expression of class A genes alone specifies sepals (whorl 1), class A and B genes in combination specify petals (whorl 2), class B and C genes in combination specify stamens (whorl 3), and class C genes alone specify carpels (whorl 4). Class A genes are therefore expressed in whorls 1 and 2, class B genes in whorls 2 and 3, and class C genes in whorls 3 and 4 (see REPRODUCTION AND BIODIVERSITY | Flower Development). The last component of the model is that the class A and C genes negatively regulate each other so that their expression patterns never overlap. Examples of all three classes of genes have been identified in Arabidopsis and all the homeotic mutant phenotypes can be explained in terms of this model. Additional complexities are generated by so-called class D genes, which are required for ovule development, and genes of the SEPALATA class, which are expressed in whorls 2, 3, and 4.

7.1. The Zygote Can Set Up Its Own Polarity

Several light and electron microscopic studies have shown that the cytoplasmic organization of the zygote is asymmetric before the first transverse division, with more endoplasmic reticulum (ER) and organelles, such as mitochondria and plastids, concentrated toward the apical (or chalazal) end and one or two large vacuoles toward the basal (micropylar) end. The unfertilized egg cell shows a similar cytoplasmic asymmetry (Fig. 3-6), and it is likely that in situ (in place, within the plant) the egg cell, or maternal influences, determines the plane of the first asymmetric division of the zygote.

FIGURE 3-6. Schematic drawing of a female germ unit in an angiosperm. The female germ unit consists of the egg cell (E), central cell (CC) with two polar nuclei, two synergids (S), and a filiform apparatus (FA), the site of entry of the pollen tube. The egg cell shows a polar organization with a large vacuole toward the micropylar end and nucleus and other organelles toward the chalazal end. The egg cell will unite with a sperm cell to form the zygote and embryo, whereas polar nuclei will unite with the other sperm nucleus to form the primary endosperm nucleus and endosperm. Synergids are attached to the embryo sac wall at points indicated by small arrows.

From Russell (1993).

Nonetheless, several instances indicate that the zygote can set up its own polarity and that new asymmetries can be established in an otherwise radially symmetrical cell. This section reviews some evidence from the in vitro fertilization of male and female gametes, somatic embryogenesis, and apomixis.

i. In vitro fertilization. A determination of the causal factors for the establishment of polarity in embryos of seed plants in planta is not possible because embryo development occurs while it is surrounded by other tissues, e.g., female gametophyte in gymnosperms or endosperm in angiosperms, and, in both taxa, by still another maternal tissue, i.e., the seed coat. However, recent improvements in microdissection and micromanipulation techniques have made it possible to isolate the embryo sac and egg cell from ovules and also male gametes, perform in vitro fertilization, and generate whole plants from fertilized zygotes (Fig. 3-7). These experiments have been performed so far with maize and are being extended to other plants (e.g., wheat, barley, ryegrass, tobacco, and rape seed). In maize, under the right conditions, the fusion of protoplasts of a male gamete and egg cell occurs very quickly (∼1 s) and karyogamy follows generally within 60 min, with the first division within ∼42-46h after fusion. The egg cell may retain some of its polar organization or may lose it before fertilization. Asymmetry of cytoplasm and organelles is reestablished in the zygote before the first division, which is also asymmetric. In vitro embryo development copies in vivo embryo development; divisions may occur irregularly for a while, but within 12-13 days an oblong embryo with an attached suspensor-like structure is formed. Later, the embryo can be transferred to solid medium and whole plants obtained.

FIGURE 3-7. Technique of in vitro fertilization and isolated male and female cells in maize. (A) Protocol of isolation of gametes and fertilization. Ovules at a receptive stage (i.e., when they are ready to receive the pollen tube and male gametes) are harvested and nucellar tissue is softened by cell wall hydrolyzing enzymes. This is followed by a manual dissection of the egg cell, synergids, and the central cell. Male gametes are obtained by bursting pollen grains or pollen tubes in a hypotonic solution. Protoplasts of egg cells and male gametes are brought in contact using glass microcapillaries, and the fusion of membranes is facilitated by short DC electric pulses or by calcium ions. The whole operation is performed on the stage of an inverted microscope. Embryo development from the zygote does not require endosperm or maternal tissues, but is facilitated by “feeder” systems of sporophytically derived nurse cell cultures. (B) An isolated unit of central cell (cc), synergids (s) and egg cell (e). (C) An isolated sperm and egg protoplast justbef ore fusion.

Modified from Kranz and Dresselhous (1996). Courtesy of Erhard Kranz, University of Hamberg.

The successful isolation of egg cell and male gametes, and in vitro fertilization and zygotic development, allows a dissection of intrinsic and environmental factors responsible for the polarization and patterning of zygotic embryos in a higher plant. It would also allow an isolation of embryo-specific genes and their precise patterns of expression and roles in development.

Zygotes produced in planta can also be isolated and cultured in vitro. In this connection, work on egg cells and zygotes of Pelvetia and Fucus (both members of Fucales among brown algae) is important. Careful and detailed experiments have established that the egg and the newly fertilized zygote are apolar and have a radially symmetrical organization, but environmental cues, such as light and pH, set up gradients that lead to axis formation, which is labile at first but then is stabilized, followed by cytoskeletal reorganization and partitioning of the cell into two cytologically dissimilar halves by the first cell division (see Box 3-1).

BOX 3-1

AXIS FORMATION IN ZYGOTES OF FUCALES

Fucales are common brown seaweeds abundant in the upper intertidal shores of temperate oceans. Fucks and Pelvetia produce relatively large egg cells, up to 1 mm in diameter, in large numbers, which are released into the seawater along with the sperm, and fertilization and embryo development occur in seawater free from maternal influences. The zygotes divide asymmetrically to produce a small rhizoidal cell and a large apical cell. The rhizoidal cell gives rise to the holdfast, which anchors the plant to the rocky substratum, whereas the large cell forms the stipe and fronds. Electron microscopy has revealed that the egg itself and the freshly fertilized zygote are both symmetrical and apolar (Fig. 3-8).

External signals, such as unidirectional light or a pH gradient, provide the cue for the formation of the axis of polarity; this axis is labile for some time (in zygotes of Fucales, up to 10 h postfertilization); then it becomes stabilized or fixed (10-14 h). Among the early signs of axis formation are the flow of an electric current from outside, through the presumptive rhizoid pole, up through the cell, and exiting through the thallus pole; a directed exocytosis of vesicles—“cortical clearing”—at the presumptive rhizoid pole; and a possible involvement of actin filaments in axis formation (cytochalasin B is a drug that disrupts the polymerization of globular actin monomers into actin filaments; a short treatment with cytochalasin B disrupts the axis formation). For axis stabilization, not only are actin filaments necessary, but also Golgi-derived vesicles carrying cell wall materials need to accumulate at the rhizoid pole. Stabilization of the axis is followed by the asymmetric distribution of zygotic cytoplasm, organelles, and cytoskeleton (14-18 h), which in turn is followed by the first cell division and partitioning of the cell into two cytologically dissimilar halves. Thus, the apical basal polarity is set before the first asymmetric division.

Some experiments suggest that cues to axis formation (apical-basal polarity) are also imprinted in the cell wall of the zygote. Subsequently, the two cells in the bicelled embryo carry these wall-specific signals, which determine the fates of their progeny. If the walls of the two-celled embryo are enzymatically hydrolyzed and protoplasts are released, the cells dedifferentiate and revert to the zygotic state. If the two cells are separated but are kept within their walls, they follow their own restricted fate. If a protoplast from one cell type is exposed to wall fragments from the other cell, the fate of the former is switched, suggesting that some cell-specific signal is embedded in the wall. Thus, cell walls can carry information for fate determination

FIGURE 3-8. Axis formation in zygote of Fucus. The fertilized egg has no inherent polarity. By the time the zygote germinates, many cellular components, all labeled structures, are unevenly distributed along the polar axis.

Adapted from Kropf (1994).

FIGURE 3-9. Production of apomictic embryos from nucellus (n) in Citrus. (A) An unfertilized ovule. A nucellar initial cell has divided to produce a two-celled nucellar embryo (ne). Other nucellar initials (ni) can be seen differentiating from the nucellus. (B) A longitudinal section through the micropylar end of a fertilized seed showing the zygotic embryo (ze) and a small nucellar embryo (ne), which is growing into the embryo sac (es).

From Koltunow (1993).

FIGURE 3-11. Structure of a telomere. The telomeric DNA with the 3′ overhang of the G strand is shown at the top. The double-stranded part of telomere DNA loops back on itself, forming a lariat structure. The 3′ G strand extension invades the duplex telomeric repeats and forms a D loop (displacement loop), Duplex telomere-binding proteins bind along the length of the telomere repeats, and a specialized telomere-binding protein binds the D loop at the junction of the lariat.

Modified from Greider (1999) with permission from Elsevier Science.

ii. Somatic embryogenesis, or the production of embryos from vegetative or somatic cells, other than zygote, is common in angiosperms as well as several gymnosperms (see Section 8). The carryover of asymmetry from maternal factors in the development of somatic embryos is not likely.

iii. Apomixis, or the production of embryos from cells in the ovule, other than the zygote, is common in several large families of angiosperms. Several different types of apomixis are known, but the concept of the polarity of the egg determining the asymmetry of zygote is difficult to extend to the apomictic embryos that arise from individual cells of nucellus or inner integument (Fig. 3-9).

7.2. Genes That Regulate Pattern Formation during Embryogenesis Are Not Expressed in Maternal Tissues

All genes that have been identified so far, via mutant analysis, to play a role in plant embryo development are expressed in the zygote or embryo, not in egg cell or ovular tissues, which indicates that major specifiers of the embryo body plan act after fertilization has occurred (see Table 3-1). In addition to the genes mentioned in Section 5, TWIN genes in Arabidopsis (see Chapter 4) are thought to suppress the embryogenic potential of the basal cell and its derivatives, which give rise to the suspensor. GNOM and TWIN genes are among the earliest acting genes identified thus far. mRNAs of both are expressed only in zygotic or embryonic tissues, not in maternal tissues or in the unfertilized egg.

In summary, while asymmetry of the egg cell may be carried to zygote, and maternal influences may play a role in early embryo development in planta, other examples suggest that plant cells can set up their own polarity and give rise to embryos. Moreover, the genes expressed in early embryogenesis are zygotic in nature. Nonetheless, there are reasons to believe that some maternally-inherited alleles keep the embryo and/or endosperm development program suppressed in the female gametophyte up to the time of fertilization. In mutants such as fertilization-independent endosperm (fie) and fertilization-independent seed (fis), this suppression is released, and results in a precocious development of endosperm and embryo (from the central cell nucleus and egg, respectively) without double fertilization; even seed and fruit development (from ovular integuments and ovary, respectively) seems to proceed normally. Usually, however, the embryo and endosperm development is arrested after a certain stage and the seeds atrophy. In the medea mutant also, the central cell nucleus proliferates without fertilization leading to a massive development of endosperm. The FIE and MEDEA genes have been cloned and encode different members of a diverse group of polycomb proteins which in mammals, insects and fungi are known to participate in protein complexes that serve to ensure the stable inheritance of expression patterns through cell division and regulate the control of cell proliferation in developing embryo. It is possible therefore that FIE and MEDEA proteins participate in similar complexes that keep the endosperm development program silenced in the female gametophyte.