Chun-MIng Liu's Plant Embryogenesis Site

Plant Embryogenesis

Based on a morphological comparison of serial sections of different stage embryos, substantial information on the pattern of cell division has been accumulated (Maheshwari, 1950, Johri, 1984). Our understanding of plant ontogeny has progressed step by step, in parallel with the technical break throughs like the invention and improvement of the light and electron microscope. The first systematic observation of embryo development in dicotyledons (dicots) and monocotyledons (monocots) was carried out more than 100 years ago by Hanstein (1870), who, with the aid of a light microscope and histological methods, followed the formation of quadrants, octants and even the establishment of the three germinal layers and hypophysis. Such kind of histological analysis gives a general description of fertilization and embryo pattern formation.

The life cycle of flowering plants is divided into two phases, a dominant diploid sporophyte phase and a transient haploid gametophyte phase. During the complex process of plant sexual reproduction, the male gametophytes or pollen grains , which contain two sperm cells and one vegetative nucleus when mature, are formed in the anther. The female gametophyte, or embryo sac, is formed in the ovule, and consists of seven cells: one egg, two synergids, one central cell and three antipodal cells. The egg cell and synergid cells locate at the micropylar pole of the embryo sac. The polarity of the egg cell is evident from the facts that the nucleus and the most of the cytoplasm situate at the chalazal end, whereas the micropylar end is highly vacuolate (Reiser and Fischer, 1993). Dual fertilisation, which is typical in angiosperms, initiates the development of both the diploid embryo (via the fusion of the sperm and the egg) and triploid endosperm (the fusion of the sperm and the central cell).

Based on the first two divisions of the zygote and the contributions of apical and basal cells to the formation of the embryo proper (yellow in the figure) and the suspensor (pink in the figure), plant embryogeny has been divided into six types: Solanad, Chenopodiad, Caryophyllad, Onagrad, Asterad and Piperad (Johri, 1984). In the first 5 types, the first division of zygote is transverse to the long axis of the cell, and often asymmetrical to produce a large basal cell and a small apical cell. When the first division of the apical cell is transverse, if most organogenetic part of the embryo is derived from the apical cell, it is called Solanad type; if the basal cell has been involved considerably in embryonic organogenesis, it is called Chenopodiad type; and if absence of division in the basal cell and the formation of the suspensor mainly from the apical cell, it is called Caryophyllad type. When the apical cell divides longitudinally, if the basal cell contribute very little or nothing to the formation of embryo organs, it is called Onagrad type (or Crucifer type); if the basal cell contributes greatly to the embryonic organogenesis beside the formation of a suspensor, it is called Asterad type. The remaining type is called Piperad to which very few species adapt, with the first division of the zygote is longitudinal to the axis of the cell (Natesh and Rau, 1984; Maheshwari, 1950).

This classification is applicable to all flowering plants, but not directly related to their taxonomy. Angiosperms (Anthophyta) are divided into two subdivisions, Dicotyledoneae (dicots) and Monocotyledoneae (monocots), representing two prototypes of body organization with either one or two cotyledons which can be seen in their seeds and seedlings. The dicots are regarded as primitive type of flowering plants, in which two cotyledons are symmetrically positioned across the axis, and are composed of a shoot apex, a hypocotyl and an embryonic root. The monocots differ from dicots in three aspects: firstly, there is only one cotyledon (called scutellum) on the axis in which the shoot apex situates laterally; secondly, the embryonic root does not bear a functional root meristem, so the roots are derived from the lower portion of the hypocotyl (Jurgens et al., 1994); Thirdly, the shoot apex is normally well-developed in the mature seeds.

Embryogenesis in dicots and monocots is similar up to the octant stage by following one of the six types of embryogeny mentioned above and dissimilar there after (Raghavan, 1986). Typical embryo development of a dicot embryo is seem in several successive stages: globular, heart-shape, torpedo-shape and cotyledonary stages, followed by an extended period of physiological changes and storage products accumulation to prepare the seed for desiccation and dormancy.

Experimental Embryology

Due to the obstacle that plant embryo development occurs within the ovule and is therefore inaccessible to experimental manipulation, the study of plant embryogeny has remained at a descriptive level for a long time. Experimental embryology, as a mean of understanding plant cell differentiation and pattern formation, has been limited and its manipulation heavily dependent on the availability of in vitro systems.

Totipotency is a feature of plant cells. In many species, differentiated cells retain the ability to form a whole plant through organogenesis or somatic embryogenesis (Nomura and Kumamine, 1985). Several recent reports show that embryogenesis, at least in vitro appears to be rely on certain proteins (Lordenener et al, 1991; De Jong et al, 1992; Kreuger and Van Holst, 1993; Gavsh et al, 1992). These proteins are most likely to represent cell wall components, some of which may have a function in the control of wall expansion (Van Engelen and de Vries, 1992; 1993). Other experiments indicate that nodulation factors, lipooligosaccharides released from Rhizobium , are necessary for the transition from the globular to the heart-shape stage embryos (De Jong et al, 1993). Embryo manipulation by microsurgery has also been carried out in carrot somatic embryo to determine how pattern regulation occurs following bisection (Schiavone and Cooke, 1985; Schiavone and Racusen, 1990; 1991). These studies suggest that the extra-embryonic tissues of maternal or non-maternal origin surrounding a zygotic embryo may not be vital for the embryo pattern formation. The role of these tissues may mainly be nourishing the developing embryo.

A major disadvantage of using somatic embryos to study pattern formation is that a high proportion of abnormal embryos occur quite often in tissue culture. For example, carrot somatic embryos often possess fused or multiple cotyledons. In most cases, only 14% of them retain the ability to become plantlets, which might be a result of the failure to establish a functional shoot meristem (Nickle and Yeung, 1993). The recent establishment of an in vitro culture method for zygotic proembryos (Liu et al, 1993), makes it possible to manipulate embryos in an attempt to understand the mechanism of embryo pattern formation. Using this technique, it has been observed that auxin polar transport is essential for the establishment of bilateral symmetry during early embryogenesis in Brassica (Liu et al, 1993). Such kind of experiments will provides important clues to the mechanism of embryo pattern formation.

Genetic Dissection of Embryo Pattern Formation

Genetic dissection has been applied to study plant embryo pattern formation in the last decade (Jurgens et al, 1994; Meinke, 1991). It is well-known that many genes are switched on and off to control cell differentiation and pattern formation during plant embryogenesis (Goldberg et al, 1989). One problem that has handicapped the search for regulatory genes is the complexity of the pathways they control. A practical way to fulfill this task is to relate growth and differentiation problems to the gene expression or gene regulation errors. Such "errors" can be induced by mutagenesis. This has been carried out mainly in Arabidopsis and maize.

Arabidopsis as a material to dissect embryo pattern formation

In the last twenty years, Arabidopsis has become a model plant for the study of embryo pattern formation, mainly because of its small genome size, short life cycle, and the possibility to carry out whole-mount analysis of developing embryos (Meinke, 1979; Mayer et al.,1993; Pyke, 1994). Its small genome size (~100mb) facilitates gene isolation by chromosome walking (Leyser et al, 1993). In Arabidopsis, T-DNA from Agrobacterium (Feldmann, 1991) and Ac/Ds transposon systems (Bancroft et al., 1993) from maize have been used to generate insertional mutations. Once an interesting mutant has been identified, the inserted element can be used as a probe to isolate the developmental gene, such as the homeotic gene AGAMOUS (Yanofsky et al, 1990) and apical-basal pattern gene GNOM

The main disadvantage of insertional mutagenesis has been a relatively low mutation rate which makes it difficult to tag a specific gene (Lindsey et al, 1995). In contrast, chemical mutagenesis like ethylmethane sulphate (EMS) treatment gives rise a high mutation frequency without apparent preferences for specific genomic regions, and it can also generate many alleles which enables one to get null phenotypes. The frequencies of producing allelic mutations during mutagenesis could give rise an indication of the number of genes performing essential functions at the screening stage. At the present time, two screening stages are generally chosen to identify embryo mutants: one is at the seed development stage to isolate embryo-lethals (Meinke, 1986; Johnson et al, 1994); another is at the seed germination stage to identify pattern mutants (Mayer et al, 1991). According to the estimation of Jurgens et al (1991), there are approximately 4,000 genes essential for embryogenesis in Arabidopsis, and 40 of them are zygotically involved in pattern formation. Existing collections of embryonic mutants, therefore, may have saturated putative pattern genes, but not the whole embryogenesis (Jurgens et al., 1991; Meinke, 1995).

Single-gene mutations can alter the following processes without causing embryo-lethality: (1) production and proper distribution of pigments in the embryo, endosperm and seed coat; (2) production of storage material in the cotyledons; (3) establishment of embryo pattern and cotyledon morphology; (4) the size and the shape of the seed.

Recessive embryo lethals are the most common mutants recovered following EMS mutagenesis in Arabidopsis (Meinke, 1986). Since such mutant embryos will die before they reach maturity, embryo lethals must be maintained as heterozygotes. Fruits produced from self-pollinated heterozygous plants contain approximately 25% aborted seeds. The remaining 75% phenotypically normal seeds contain 1/3 homozygous wild-types and 2/3 heterozygous wild- types. Therefore, to sustain such mutant lines, screening for heterozygous plants is required in every generation.

The isolation of non-lethal embryo mutants has been carried out mainly by Jurgens and his co-workers (Mayer et al., 1991; Jurgens, 1992), most of the genes identified are involved in pattern formation, especially in the establishment of the apical-basal pattern and radial pattern.

The primary body along the apical-basal axis of an Arabidopsis seedling can be described as four distinct elements, from top to bottom: the shoot apex, two cotyledons, the hypocotyl and the root apex. Embryo pattern mutations could cause deletion, duplication or transformation of these elements. Pattern mutant can affect one or more of the elements without changing the development of other elements.

Four classes of apical-basal deletion phenotypes have been isolated in Jurgens's lab: apical, basal, central and terminal, and shoot apex (Mayer, et al., 1991; Barton and Poethig, 1993). The apical deletion mutant alleles of the GURKE gene corresponds to the deletion of shoot apex and cotyledons; while central deletions are represented by mutant alleles of the FACKL gene with the cotyledons being attached directly to the root. In the basal deletion class, the monopteros seedling, which eliminates both the hypocotyl and the root, is almost the complete converse of the gurke seedling. The mutant embryo shows abnormalities as early as the octant stage. Nevertheless, the MONOPTEROS gene is likely to be required for organising the basal elements rather than making a root, as roots can be induced from the mutant seedling grown in tissue culture (Berleth and Jurgens, 1993). In the terminal deletion class, e.g. gnom, shoot apex, cotyledons and root are reduced or eliminated. Mutant alleles of the GNOM gene show either cone- or ball-shaped embryos. Unlike monopteros , the gnom mutant c annot produce a root even in tissue culture. Histological analysis indicated that the gnom allele may fail to promote the asymmetric cell division which is necessary for the formation of both the cotyledon and the root (Mayer et al., 1993).

This is an intriguing model since it makes a plant embryo more like a fruitfly embryo with distinct segments. Whether such model reflect the real cue of plant pattern formation need further convincing evidence, since normally there is no clear boundary between these segments. When a segment, like an organ, was removed, it is possible to regenerate it from its adjacent organ. Cell linage experiment also indicated that cell linage can cover two neighbouring segments (Dolan et al, ). In Jurgens' recent review article (1995), he also recognized that plant pattern formation rely largely on the cell-cell communications. Similarly, Vernon and Meinke (1995) point out recently that many of the embryo mutants may not involve genes directly in morphogenesis but rather than genes involved in basic cellular and physiology process. In support this notion they cite the further study of the GNOM gene. Using T-DNA tagging, the GNOM gene has been cloned recently from Arabidopsis (Shevell, et al, 1994). Unexpectedly, the predicted amino acid sequence of the GNOM is similar to a yeast protein (Sec7p) and is expressed in every tissue tested.

Mutant deletion of only shoot apex has been identified recently, which is called shoot meristemless (stm). (Barton and Poethig, 1993). Mutation of the STM gene completely blocks the initiation of the shoot apical meristem, but has no other obvious effects on embryo development. Failure to regenerate shoots in tissue culture from stm mutant suggests that this gene regulates both embryonic and adventitious shoot formation.

The radial pattern of a mature embryo consists of 3 primary tissues: epidermis, ground tissue and vascular bundles, which are originated at the globular stage and are maintained during postembryonic development (Laux and Jurgens, 1994). To carry out special functions, the primary tissues may undergo further specifications, such as the formation of trichomes and guard cells in the epidermis. Two mutants, keule; and knolle, so far have been identified in Arabidopsis, which are defective in the establishment of epidermis (Mayer et al., 1993). Globular keule embryos show enlarged cells in the outer layer that normally will give rise to the epidermis, while the inner cells appears normal. By contrast, in knolle the enlarged cells are not restricted to the outer cell layer but also take place in the inner tissues (Mayer et al , 1993). There is only one mutants identified as deletion in part of ground tissue, and none, so far, deletion in vascular bundles. A mutant called short root lacks the endodermal layer which is derived from ground tissue (Benfey et al ., 1993). The scarcity of mutants defective in radial pattern could be due to either that such kinds of mutants are difficult to be recognized, or deletion of a primary tissue would cause embryo lethality.

Homeotic mutants involved in embryo development have also been isolated, for example leafy cotyledon , which converts cotyledons to leaves (Meinke, 1992). The cotyledons of this mutant produce trichomes which are characteristic of leaves, lack embryo-specific protein bodies and exhibit a vascular pattern intermediate between that of leaves and cotyledons. In another homeotic mutant, emf , the vegetative growth is replaced by reproductive growth. Flower organs appear immediately after seed germination (Sung et al, 1992).

Maize as a material to dissect embryo pattern formation

Maize, with its well-characterized embryogenesis and endogenous transposons, offers a system for the study of embryo development in monocots (Breton et al., 1995). Extensive screening of lethal mutants defective in embryo development has also been carried out in the last few years (Sheridan and Clark, 1993). Fifty-one embryo-specific mutants of maize were isolated by Clark and Sheridan (1993) from Robertson's mutator transposon stocks. Twenty-seven of them were characterized by examining freshly dissected mature embryos. All mutants result in the arrest of development at different stages: 9 of them are blocked during the proembryo or transition stage, 10 mutants are blocked at the time of embryonic axis establishment and 8 mutants are blocked at the time of embryo maturation and the imposition of dormancy. Such mutants generated by transposon tagging will facilitate molecular cloning of related genes by using the known DNA fragment as a probe (Clark and Sheridan, 1991). The cloning and characterization of the corresponding genes will shed additional light on early embryogenesis.

In addition, other species have also been used to produce embryo mutants and to study embryogenesis, for example, rice (Hong et al, 1995), barley (Bosnes et al, 1987; Felker et al, 1984, 1985), carrot (Lo Schiavo et al, 1990; Schnall et al, 1988), rice (Nagato et al, 1989), and pea (Liu et al, 1996). For more information about seed development in pea, please see Trevor Wang's homepage at John Innes Centre

References