View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector , Vol. 81, 467-470, May 19, 1995, Copyright 0 1995 by Cell Press Axis Formation Minireview in Plant Embryogenesis: Cues and Clues

Gerd Jiirgens focus on pattern formation in the that establishes Lehrstuhl fur Entwicklungsgenetik the basic body organization of flowering plants. Universitat Ttibingen The Shoot Me&tern: Linking Up the Embryo D-72076 Tiibingen with the Flower Federal Republic of Germany Pattern formation in animals is largely confined to em- bryogenesis such that the future adult form is represented in the body organization of the mature embryo. By con- Imagine a textbook entitled Developmental that trast, plant embryogenesis produces a juvenile form, the focuses entirely on plants, mentioning animals only for seedling, that lacks most structures of the adult plant. Em- their peculiar way of making germ cells by setting aside bryogenesis in essence organizes two groups of stem cells a group of precursor cells early in the embryo. The con- at the opposite ends of the body axis, the primary meri- verse has been, and still is, common practice. It is true stems of the shoot and the root. These then that the regenerative potential of plants, which is indeed add new structures to the seedling, thus generating the impressive, sets them apart from the more familiar animal species-specific adult form during postembryonic devel- models: individual cells can give rise to in culture; opment (Steeves and Sussex, 1989). Regardless of the localized groups of stem cells called meristems make the appearance of the adult plant, the shoot is orga- adult plant in a seemingly autonomous fashion not only nized essentially the same way in different plant species. during normal development, but also by regeneration from Two functional units can be distinguished within the meri- lumps of undifferentiated cells in culture. These special stem: a central zone, which is required for self-renewal features notwithstanding, plants do develop from a fertil- and integrity of the meristem, and a peripheral zone, which ized egg cell, the zygote, during the normal course of their makes primordia of lateral organs and their associated life cycle and, like animals, have to establish thecharacter- secondary shoot meristems, such as leaves and flowers istic body organization of the multicellular adult form. Are (Steeves and Sussex, 1989). Whether leaves or flowers the underlying mechanisms the same or different? In the are produced depends on the physiological state of the past few years, the genetic and molecular analysis of meristem. In embryonic flower (emf) mutant seedlings, for flower development in two plant model species, Arabi- example, the primary shoot meristem skips the vegetative dopsis and Antirrhinum (snapdragon), has established a phase of making leaves altogether, producing flowers di- network of MADS domain transcription factors and others rectly (Sung et al., 1992). Shoot meristem identity seems that regulate this best-characterized process in plant de- to be conferred by the continuous expression of genes like velopment (Weigel and Meyerowitz, 1994). However, the maize homeobox gene Knotted7 (Knl), whose ectopic flower development is like putting the finishing touches expression can cause the formation of shoot meristems on the adult plant and may thus not give clues to mecha- on leaves (Sinha et al., 1993; Jackson et al., 1994). A nisms that underlie earlier processes such as axis forma- putative Arabidopsis homolog of Knl, the SHOOT MEW tion and the generation of the overall body organization. By STEM-LESS (STM) gene, is required for shoot meristem drawing largely on recent genetic studies in Arabidopsis, I formation both in the embryo and during regeneration from will briefly discuss postembryonic development that even- tissue culture (Barton and Poethig, 1993). How the activity tually culminates in the formation of flowers, but mainly of such shoot meristem-specific genes is established and

Figure 1. Formation of the Apical-Basal Axis in the Arabidopsis Embryo (A) Asymmetric division of the zygote, giving a a proE small apical (a) and a large basal (b) cell. (B)The8-cellstage.Theproembryo(proE)con- sists of two tiers each of four cells (regions marked A for apical and and C for central) and b sus is connected to the extraembryonic suspensor (sus) via the founder cell of the basal region (B) of the embryo. I- (C) Embryo at heart stage. Approximate loca- tions of cell groups that give rise to the primor- dia of seedling structures are indicated. (D) Embryo at Torpedo stage. Clonal bound- aries are marked by thibk lines. The broken line indicates the upper end of the embryonic root derived from the root meristem initials (RMI). A B C D Below the quiescent center (QC) of the root meristem are the initials of the central root cap (CRC). Primordia of seedling structures: COT, ; HY, hypocotyl; ER, embryonic root; RM, root meristem; SM, shoot meristem. Cdl 466

maintained properly is not known. However, there are can- mainder of the root meristem, comprising the quiescent didate genes in Arabidopsis for performing these func- center and the initials of the central root cap. Data from tions. For example, the ZWLLE (ZLL) gene is specifically clonal analysis support the view that cell ancestry does involved in establishing the primary shoot meristem in the not play a role in generating apical-basal pattern ele- embryo but not in any other (nonembryogenic) context ments: clone boundaries are variable and, moreover, can (Jtirgens et al., 1994). Mutations in another gene, CLA- run across specificseedling structures, such as thecotyle- VATAl (CLW), cause overgrowth of primary and second- dons or the root meristem (Scheres et al., 1994; Dolan et ary shoot meristems that also results in the formation of al., 1994). Furthermore, the regularity of cell divisions in one or more supernumerary whorls in the center of the the early embryo is deceptive since mutations in the FASS flower (Clark et al., 1993). This phenotype suggests that (FS) gene totally alter the pattern of , without the organization of the meristem is altered, owing to an affecting pattern formation (Torres Ruiz and Jiirgens, enlargement of the central zone required for self-renewal. 1994). Thus, the apical-basal pattern elements appear to Although more genes need to be identified to clarify this be established by cellular interactions in a position- point, it is conceivable that the primary shoot meristem dependent manner. may acquire a specific organization in the embryo that is What is the significance of the early regions along the subsequently maintained by cellular interactions within apical-basal axis? Evidence comes from the embryonic the meristem. phenotypes of mutations in three genes, MONOPTEROS Pattern Formation in the Arabidopsis Embryo (MP), FACKEL (FK), and GURKE (GK), which delete spe- The primary meristems of the shoot and the root originate cific seedling structures (Mayer et al., 1991). In mp em- at distinct positions in the embryo as part of the overall bryos, the cells in the central and basal regions divide body organization of the seedling (Barton and Poethig, abnormally, resulting in the absence of hypocotyl, root, 1993; Dolan et al., 1993). The latter may be viewed as the and root meristem (Berleth and Ji.irgens, 1993). In fk em- superimposition of an apical-basal pattern along the main bryos, the central region is affected while the basal region body axis and a radial pattern perpendicular to this axis. is not; in the seedling, the hypocotyl is missing such that The apical-basal pattern consists of a top-to-bottom array the cotyledons are directly attached to the root (Mayer et of elements: shoot meristem, cotyledons (embryonic al., 1991; Jiirgens et al., 1994). Since the root meristem leaves), hypocotyl (embryonic stem), radicle (embryonic is composed of cells from both the basal and the central root), and root meristem. The radial pattern is made up of regions, the basal region appears to induce the adjacent concentric rings of tissue layers: epidermis, cells of the central region to become root meristem initials (cortex and endodermis), and (pericycle, that in turn produce the meristemderived embryonic root xylem, and phloem). The origins of the apical-basal and (see Figures 1 C and 1 D). A similar argument can be made radial pattern elements have been traced back to cell for the cotyledons, which are derived from both the apical groups of the early embryo in Arabidopsis where the very and the central regions. The phenotype of gk mutants is regular patterns of cell division facilitated these analyses essentially complementary to mp: the apical region of the (Mansfield and Briarty, 1991; Jiirgens and Mayer, 1994). embryo is altered, and, later on, gk seedlings lack shoot Formation of the Apical-Basal Axis meristem and cotyledons (Mayer et al., 1991). This sug- Apical-basal pattern formation starts with the asymmetric gests that the cotyledons are initiated within the apical division of the zygote that gives two daughter cells of un- region, which then signals to adjacent cells of the central equal sizes and different fates (Figure 1). The small apical region to participate in formation. Thus, the cell generates, by cleavage divisions, an 8-celled proem- early regions may define genetically distinct groups of cells bryo that will give rise to most of the embryo. The large (compartments?) that generate the apical-basal pattern basal cell produces a file of 7-9 cells, of which all but the by cellular interactions. uppermost one will form the extraembryonic suspensor; How are the early regions established along the apical- the uppermost cell (hypophysis) joins the proembryo later basal axis? The boundary separating the central from the to give rise to part of the root meristem. Within 1 day of basal region originates with the first division of the zygote fertilization, three embryonic regions are thus established (Figures 1A and 1B). The basal daughter cell produces a along the axis: apical and central, which correspond to file of cells of which the uppermost one becomes the the upper and lower tiers within the proembryo, and basal, founder cell of the basal region of the embryo. This file of which is represented by the adjacent hypophysis (Figures cells is initially normal in mp embryos when the earliest 1A and 1 B). The three embryonic regions differ in their cell defect becomes apparent in the proembryo, whereas later division patterns: apical cells divide without preferential on both the central and the basal regions develop abnor- orientation; central cells produce cell files, thus expanding mally (Berleth and JOrgens, 1993). Thus, in normal devel- the axis; and the founder cell of the basal region undergoes opment, the central region may induce the uppermost de- a stereotyped program of division. However, the early re- rivative of the basal daughter cell of the zygote to become gions (apical, central, and basal) do not correspond to the founder cell of the basal region. By contrast, the apical primordia of seedling structures (Figures 1C and 1 D). The and the central regions are established by the transverse apical region gives rise to the shoot meristem and most cell divisions within the proembryo that is derived from the of the cotyledons; the central region also contributes to apical daughter cell of the zygote. the cotyledons, but mainly produces hypocotyl, root, and Before dividing asymmetrically, the zygote elongates root meristem initials; and the basal region gives the re- about 3-fold in the future axis, and concomitantly the corti- Minireview 469

Figure 2. Formation of the Radial Pattern in the Arabidopsis Embryo Schematic cross sections through the central region (see Figure 1) of embryo (A and B) and through the root primordium (C and D). (A) Dermatogen stage. Periclinal (radial) divi- sions within the prcembryo(Figure lB)give the outer epidermis layer (EP) and an inner cell mass (ICM). (B) Globular embryo. The inner cell mass has A D split into the ground tissue(G) and the centrally located vasculated primordium 01). (C) Embryo at heart stage (Figure IC). The pericycle layer (PC) surrounds the primorctium of the conductive tissue (CT). The basic organization of the radial pattern is complete. (D) Embryo at Torpedo stage (Figure 1D). Periclinal divisions in the ground tissue generate an outer cortex (CO) and an inner endodermis layer (EN).

cal microtubules, which were previously oriented at ran- (phloem and xylem). This basic organization of the radial dom, become aligned perpendicular to the axis (Webb and pattern is complete before the heart stage of embryogene- Gunning, 1991). This reorganization may reflect polariza- sis (Figure 2C). One further subdivision occurs within the tion of the zygote in response to some as yet unknown ground tissue, generating an outer cortex layer and an signal(s). Its significance for axis formation is suggested inner endodermis layer (Figure 2D). by mutations in the GNOM (GN) gene that affect the zygote By the time the root meristem has become active, clonal and can completely abolish regional differentiation of the boundaries reflect the basic organization of the radial pat- seedling axis (Mayer et al., 1991). The gn zygote expands tern (Dolan et al., 1994; Scheres et al., 1994). Are tissue- but does not elongate, producing an enlarged apical cell specific cell fates fixed early and then transmitted clonally? at the expense of the basal cell (Mayer et al., 1993). Subse- Mutant phenotypes suggest that this may be the case. For quent development is highly abnormal, and the region- example, the epidermis cells are abnormally enlarged in specific MP gene was shown to be ineffective in gn em- early keole (keu) embryos (Mayer et al., 1991) and the bryos. The asymmetric division of the zygote seems to be endodermis layer is absent in short root (shr) embryos involved in fixing the apical-basal axis of the embryo, but (Benfey et al., 1993; Scheres et al., 1995). Furthermore, it is not known how the GN gene brings about its specific the epidermis is marked by the tissue-specific expression effect at the molecular level. Although the predicted 160 of the LTPgene both in the embryo and during postembry- kDa GN (also called EMB30) protein showssequence simi- onic development (Thoma et al., 1994). The genetic dis- larity to several other proteins over a stretch of 200 amino tinction of different tissues might involve blocking of plas- acids that is referred to as the Sec7 domain, the signifi- modesmata-mediated cell communication as has been cance of this domain is not known, nor does the remainder observed between the epidermis and subepidermal tis- of the GN protein readily suggest a specific function within sues in the seedling (Duckett et al., 1994). While the avail- the cell (Shevell et al., 1994). Since GN is the only Arabi- able evidence seems to favor a model of genetically deter- dopsis gene known to be specifically required for apical- mined tissue types, oriented cell divisions do not appear basal pattern formation from the zygote on, determining to be required for radial pattern formation: cell divisions the primary function of the GN protein will be an important occur at random in early fs embryos such that the charac- step in further analysis of axis formation. teristic radial organization of tissues is not apparent, but Radial Pattern Formation later on the mutant seedlings display all tissues found in Radial pattern formation, which commences in the &celled wild type (Torres Ruiz and Jlirgens, 1994). This flexibility proembryo, involves two choices of oriented cell division: suggests that cellular interactions may play an important periclinal (radial), which gives a new tissue layer, and anti- role in initiating and (possibly) maintaining tissue-specific clinal (circumferential), which increases the number of gene expression in a position-dependent manner. cells in a given layer. New tissue layers are successively How is the radial pattern initiated in the early embryo? formed from the periphery toward the center (Figure 2). So far, only one gene has been identified that may be Initially, an outer layer of epidermal precursor cells is sepa- involved in this process: in knolle (kn) embryos, the epider- rated from an inner cell mass. The epidermis layer then mis is not separated from the inner cell mass by periclinal expands by anticlinal cell divisions, thus maintaining its divisions within the proembryo, and the mutant seedlings integrity in the growing embryo. Further development of appear to lack the characteristic epidermis layer (Mayer the radial pattern is confined to the central region of the et al., 1991). How the KN gene relates, at the molecular axis that gives rise to hypocotyl, root, and root meristem level, to the periclinal divisions in the early embryo remains initials; neither the apical nor the basal region is involved to be determined. (Jiirgens and Mayer, 1994; Scheres et al., 1995). The inner What Mechanisms May Underlie Pattern Formation cell mass splits into an outer layer of ground tissue and in the Plant Embryo? the centrally located vascular precursor cells. The latter Pattern formation in the Arabidopsis embryo appears to cells again divide periclinally to give a layer of pericycle depend largely on cell-cell communication, both in the cells surrounding the precursors of conductive tissue apical-basal axis and in the radial dimension. This is for- Cell 470

rnally similar to pattern formation in the familiar animal Duckett, C. M., Oparka, K. J.. Prior, D. A. M., Dolan, L., and Roberts, models (e.g., vulva development in Caenorhabditis ele- K. (1994). Development 720, 3247-3255. gans or postblastoderm embryogenesis and imaginal disc Evans, T. C., Crittenden, S. L., Kodoyianni, V., and Kimble, J. (1994). development in Drosophila). However, the underlying Cell 77, 163-194. Jackson, D., Veit, B., and Hake, S. (1994). Development 720, 405- mechanisms might be different in plants. For example, 413. plant cells are cytoplasmically interconnected by plasmo- Jt’irgens, G., and Mayer, U. (1994). In Embryos: Colour Atlas of Devel- desmata, allowing various kinds of molecules to pass opment. J. B. L. Bard, ed. (London: Wolfe Publishing), pp. 7-21. freely from cell to cell. Unfortunately, it is not known Jtirgens, G., Torres Ruiz, R. A., Laux, T., Mayer, U., and Berleth, T. whether this potential plant-specific means of intercellular (1994). In Plant Molecular Biology: Molecular-Genetic Analysis of communication is actually used and how it is regulated and Metabolism, G. Coruzzi and P. PuigdomBn- during development. Another serious problem in the analy- ech, eds.(Berlin: Springer-Verlag), pp. 95-103. sis of plant pattern formation is the shortage of markers, Mansfield, S. G., and Briarty, L. G. (1991). Can. J. Botan. 69, 461- 476. both morphological (which cannot be changed) and molec- Mayer, U., Torres Ruiz, R. A., Berleth, T., MisBra, S., and JOrgens, ular. As more genes with tissue-specific or region-specific G. (1991). Nature 353, 402-407. expression patterns in the embryo will become available, Mayer, U., Biittner, G., and Jiirgens, G. (1993). Development 717, some of the ideas presented here can be tested. 149-162. The initial events of pattern formation in the Arabidopsis Quatrano, R. S., Brian, L., Aldridge, J., andSchultz,T. (1991). Develop embryo are different from the later phases since they can- ment 773 (Suppl. I), 11-16. not involve cell-cell communication: the asymmetric divi- Scheres, B., Wolkenfelt, H., Willemsen, V., Terlouw, M., Lawson, E., sion of the zygote as well as the periclinal divisions in the Dean, C., and Weisbeek, P. (1994). Development 720, 2475-2487. 8-celled proembryo appear to segregate fates within cells. Scheres, B., Di Laurenzio, L., Willemsen, V., Hauser, M.-T., Jaanmat, How is position translated into cell fate in these cases? K., Weisbeek, P., and Benfey, P. N. (1995). Development 721, 53- 62. One possible mechanism is suggested by the observation Shevell, D. E., Leu, W.-M., Gilimor, C. S., Xia, G., Feldmann, K. A., that in C. elegans, the unequal division of the zygote is and Chua, N.-H. (1994). Cell 77, 1051-1062. associated with the differential degradation of a specific Sinha, N., Williams, R., and Hake, S. (1993). Genes Dev. 7,787-795. uniformly distributed maternal mRNA, thus establishing Sleeves, T. A., and Sussex, I. M. (1989). Patterns in Plant Development territories of differential gene expression along the main (Cambridge: Cambridge University Press). body axis of the early embryo (Evans et al., 1994). A poten- Sung, 2. R., Belachew, A., Shunong, B., and Bertrand-Garcia, R. tially plant-specific alternative for embryonic axis fixation (1992). Science 258, 1645-l 647. is suggested by recent results from the brown alga, Fucus. Thoma, S., Hecht, U., Kippers, A., Botella, J., de Vries, S. C., and The Fucus zygote also divides asymmetrically, giving an Somerville, C. (1994). Plant Physiol. 705, 35-45. upper thallus cell and a lower rhizoid cell. If the cell wall Torres Ruiz, R. A., and Jiirgens, G. (1994). Development 720, 2967- 2978. of the zygote is removed, the axis can be formed, but not fixed (Quatrano et al., 1991). Moreover, cell ablation Webb, M. C., and Gunning, 8. E. S. (1991). Planta 784, 167-195. Weigel, D., and Meyerowitz, E. M. (1994). Cell 78, 203-209. studies on early embryos suggest that the different cell fates of thallus versus rhizoid may be imprinted into the cell wall (Berger et al., 1994). Which mode, if either, of cell-fate segregation applies in the early embryo of Arabi- dopsis remains to be determined. The two proteins thus far identified do not appear to regulate gene expression directly, unlike the transcription factors that play decisive roles in flower development. Does axis formation in higher- plant embryos involve an extracellular detour on its way to the nucleus?

References

Barton, M. K., and Poethig, R. S. (1993). Development 119, 823-831. Benfey, P. N., Linstead, P. J., Roberts, K., Schiefelbein, J. W., Hauser, M-T., and Aeschbacher, R. A. (1993). Development 719, 57-70. Berger, F., Taylor, A., and Brownlee, C. (1994). Science 263, 1421- 1423. Berleth, T., and Jijrgens, G. (1993). Development 778, 575-587. Clark, S. E., Running, M. P., and Meyerowitz, E. M. (1993). Develop- ment 119, 397-418. Dolan, L.. Janmaat, K., Willemsen, V., Linstead, P., Poethig, S., Rob- erts, K., and Scheres, B. (1993). Development 179, 71-64. Dolan, L., Duckett, C. M., Grierson, C., Linstead, P., Schneider, K., Lawson, E., Dean, C., Poethig, S., and Roberts, K. (1994). Develop merit 720, 2465-2474.