Developmental Cell 4
inductive activity under normal embryonic circumstances. Traditionally, the Spemann’s organizer has been regarded as the primary source of signaling activity that specifies the neural ectoderm and the somitic mesoderm (Harland and Gerhart, 1997; De Robertis et al., 2000). The findings of Mariani et al. show that the primitive neural ectoderm may be an additional source of the mesoderm inducing activity in the Xenopus gastrula (see figure). An intriguing outcome of the study by Mariani et al. is that, although the tissues induced by the expanded neural plate initially express the MyoD gene (and therefore are bona fide paraxial mesoderm), they fail to differentiate into muscles, suggesting that myotomal differentiation may be defective. In the avian embryo, the delineation and differentiation of somitic tissues (dermis, myotome, and sclerotome) is regulated by SHH, WNT, and BMP signaling (Hirsinger et al., 2000). Specifically, signaling activities from the dorsal neural tube, the surface ectoderm, and the lateral mesoderm are essential for the formation and the differentiation of the dermomyotome and the lateral myotome (see figure). In the experimental embryos, the expanded neural plate is mostly made up of neural tissues characteristic of the intermediate region of the spinal cord, and the dorsal cell types are either displaced laterally or lost from the neural plate produced by higher XBF-2 activity. Marker expression also suggests that there is a reduction or loss of intermediate and lateral plate mesoderm. The defects in myogenic differentiation may therefore result from the loss or diminution of the signaling activities from these adjacent tissues. Besides the defect in muscle differentiation, the expanded somitic mesoderm display abnormal segmentation particularly in the caudal region of the embryo. A more severe disruption of segmentation is encountered when the prospective neural ectoderm is replaced by ectodermal tissue that lacks the ability to respond to neuralizing factors. It would be interesting to examine more closely whether in the Xenopus the neural plate may impart any morphogenetic effect on the segmentation of the paraxial mesoderm.
In a classical study on size regulation, chimaeras are made by combining embryos of mice that were selected for differences in adult body size. These chimaeras grew up in sizes that were intermediate of that of the original large and small mice. The organs of these mice displayed a wide variation in the relative abundance of cells from the large and small mice, but none of the organs showed a consistent correlation of the extent of chimaerism with the body size of the whole animal. It was then concluded that there is not a “master” organ that regulates the adult body size or the relative proportion of the body parts (Falconer et al., 1978). The discovery that the early neural plate of the Xenopus embryo may influence the size of the paraxial mesoderm highlights the interaction between tissues as one of the many mechanisms by which tissues coordinate their size with one another during development.
Unraveling the Mystery of Double Flowers
numerous descriptions of double flowers occur in the literature (Meyerowitz et al., 1989). In spite of the longstanding interest in this flower abnormality, the underlying mechanisms that control its formation have remained mysterious. This mystery has finally been solved, as independent studies from the Weigel and Laux laboratories have led to a molecular model that accounts for the formation of double flowers (Lohmann et al., 2001; Lenhard et al., 2001). The observation that flowers are simply modified shoots is just one of many remarkable insights provided by Goethe in his treatise on plant development (1790). Goethe would probably not be surprised to find that that the basic patterning mechanisms used in shoots are also used to pattern flowers (Parcy et al., 1998). However, one of the major differences between shoots and flowers is that shoots are indeterminate structures that continuously elaborate new primordia, whereas flowers are determinate structures and therefore stop after producing
Two recent papers in Cell have shown that a regulatory loop involving the WUSCHEL, AGAMOUS, and LEAFY genes controls the switch from continuous meristem growth to flower development in Arabidopsis.
Many of the most beautiful flowers, including hybrid tea roses, double camellias, and carnations, have layer upon layer of petals. Highly prized for their beauty, these so-called double flowers were selected from their plain relatives that have only a single layer of petals. Theophrastus first described double roses more than 2000 years ago, and in the centuries that have followed,
Patrick Tam Embryology Unit Children’s Medical Research Institute University of Sydney Locked Bag 23, Wentworthville NSW 2145 Australia Selected Reading Cooke, J. (1975). Nature 254, 192–199. De Robertis, E.M., Larrain, J., Oelgeschlager, M., and Wessely, O. (2000). Nat. Rev. Genet. 1, 171–181. Falconer, D.S., Gould, I.K., and Robert, R.C. (1978). In Genetic Mosaics and Chimaeras in Mammals, L.B. Russell, ed. (New York: Plenum), pp 39–50. Harland, R., and Gerhart, J. (1997). Annu. Rev. Cell Dev. Biol. 13, 611–667. Hirsinger, E., Jouve, C., Dubrulle, J., and Pourquie, O. (2000). Int. Rev. Cytol. 198, 1–65. Itasaki, N., Sharpe, J., Morrison, A., and Krunlauf, R. (1996). Neuron 16, 487–500. Liem, K.F., Jessell, T.M., and Briscoe, J. (2000). Development 127, 4855–4866. Mariani, F.V., Choi, G.B., and Harland, R.M. (2001). Dev. Cell 1, this issue, 115–126.
Previews 5
AG/WUS Feedback Loop in the Flower Meristem (A) In the SAM, WUS acts in a feedback loop with CLV3 to maintain undifferentiated stem cells in the center of the meristem. (B) In floral meristems, the same feedback loop between WUS and CLV3 acts to maintain stem cells. LFY, which is expressed throughout the meristem, acts with WUS to activate transcription of AG in the central dome, which then feeds back to turn off WUS.
a defined number of organs. Double flowers resemble shoots in that they remain indeterminate. In order to understand how determinacy is achieved in flowers, it is helpful to review recent insights that explain how shoots remain indeterminate. Essentially all of the above-ground portion of plants is derived from the shoot apical meristem (SAM), a group of cells at the growing tip of plants. The center of the SAM consists of a small population of undifferentiated stem cells, which must be maintained throughout the life of a plant, whether it is a short-lived weed such as Arabidopsis or a gargantuan tree that lives for thousands of years such as the Giant Sequoia. To generate new primordia indefinitely, the SAM must balance the demands of new organs for cells with the necessity of maintaining undifferentiated stem cells for future growth. A regulatory circuit involving CLAVATA3 (CLV3), a secreted ligand that is expressed in undifferentiated stem cells, and WUSCHEL (WUS), a homeodomain transcription factor expressed in a few cells in the organizing center of the meristem, creates this balance (see figure, panel A). WUS acts to increase the number of stem cells and consequently increases the number of cells expressing CLV3. CLV3 in turn binds to the CLAVATA1 receptor leading to downregulation of WUS in the center of the meristem creating a negative feedback loop (Waites and Simon, 2000). This same regulatory loop occurs in the floral meristem to maintain stem cells until differentiation of the central carpels, at which point WUS expression stops (Lenhard et al., 2001). What causes WUS expression to cease in the developing flower? The answer comes from examining the interactions between WUS and the MADS-box gene AGAMOUS (AG) (Lenhard et al., 2001; Lohmann et al., 2001). AG is required for floral determinacy since mutations in this gene cause the indeterminate double flower phenotype. Molecular studies now show that WUS expression fails to be switched off in the center of ag mutant flowers. Determinacy is restored in wus ag double mutants, indicating that it is the ectopic activity of WUS in ag mutants that causes indeterminacy. These studies demonstrate that AG normally functions to negatively regulate WUS in order to prevent the indeterminate growth of the flower meristem. If AG activity in the center of flowers is necessary to turn off WUS, then what activates AG in the first place? In wild-type plants, AG is activated in the center of young flower primordia in cells that will later give rise to stamens and carpels, the organs that are missing in ag mutants. The LEAFY (LFY) transcription factor is known to bind to the AG regulatory elements and activate AG (Lohmann et al., 2001 and references therein). However, since LFY is expressed throughout young flowers, how does it lead to AG activation only in the center?
Again, WUS is as a good candidate for activating AG expression because WUS is already present in the center of the meristem, and the rare flowers that form in wus mutants lack stamens and carpels, organs that are also missing in ag mutants. Lohmann et al. found that in plants with a weak wus phenotype, AG expression is reduced, hinting that WUS may be an activator of AG. When both groups ectopically expressed WUS, they found that ectopic stamens and carpels were produced. In all of these transgenic lines, AG was ectopically expressed in a pattern that matched that of WUS, demonstrating that within flowers, expression of WUS is sufficient to activate AG. Lohmann et al. further show that WUS protein directly binds to DNA sequences within the AG regulatory region and that deletion of these WUS binding sites prevents AG activation. In lfy mutants, AG is not expressed in the majority of flowers, indicating that the endogenous level of WUS protein is not sufficient for activation of AG in the absence of LFY. However, when WUS is overexpressed in lfy mutants, AG is ectopically expressed, suggesting that high levels of WUS can bypass the need for LFY (Lohmann et al., 2001). These results indicate that under normal circumstances, LFY provides the flower specificity and WUS provides regional specificity, so that AG is only activated in flowers and not the SAM. Collectively, these results suggest a simple model in which AG and WUS interact in a feedback loop where WUS activates transcription of AG which then feeds back to turn off WUS (see figure, panel B). It has previously been proposed that factors used to pattern shoot meristems may have been coopted to pattern the expression of the floral homeotic genes (Parcy et al., 1998). The cooption of WUS in the activation of AG is the first direct evidence of this interaction. Is the use of meristem factors such as WUS as spatial transcriptional activators an exception applying only to AG or a rule repeatedly used throughout plant development? And how does WUS, which is expressed in only a few cells, activate AG expression in a much larger domain? One possibility is that additional factors act together with WUS to activate AG. Alternatively, WUS protein could move throughout the meristem, similar to other transcription factors, including LFY and the homeodomain protein KNOTTED, which are known to move from cell to cell. (Sessions et al., 2000). In animals, developmental patterns are commonly established by morphogen gradients where different genes are activated at different threshold concentrations (Teleman et al., 2001). Could WUS be establishing a morphogen gradient in the meristem? To test this hypothesis, it would be useful to know whether WUS is also involved in the activation of other genes. The SEPALLATA1/2/3 (SEP) MADS-box genes
Developmental Cell 6
which control petal, stamen, and carpel development are good candidates for regulation by a gradient of WUS because they are expressed in the floral meristem in a pattern that is slightly broader than that of AG. Furthermore, it is likely that the SEP genes, like AG, are involved in repression of WUS because the sep1/2/3 triple mutant flowers are also indeterminate (Pelaz et al., 2000). Another good candidate for activation by WUS is the FRUITFULL (FUL) MADS-box gene, which is expressed in a domain that is slightly smaller than that of AG and whose expression persists in the meristem of ag mutant flowers (Mandel and Yanofsky, 1995). It will be interesting to see how many genes achieve their spatially restricted expression domains through activation by WUS. Adrienne H.K. Roeder and Martin F. Yanofsky Section of Cell and Developmental Biology University of California at San Diego La Jolla, California 92093-0116
Self-Destruction in the Line of Duty Three cellular processes, microautophagy, macroautophagy, and the cytoplasm-to-vacuole (Cvt) pathway, are involved in the cargo delivery from the cytosol to the vacuole or lysosome. Recent findings have identified Cvt19 at the receptor for specific cargo binding in the Cvt pathway.
The processing or turnover of cellular macromolecules in yeast vacuoles or their mammalian counterpart, the lysosomes, fulfills biosynthetic and recycling functions that are important for cell growth, survival, and development. These processing events depend on the activity of vacuolar or lysosomal hydrolases and require mechanisms for the delivery of cargoes (specific proteins, unnecessary organelles, or bulk cytosol) from the cytosol and the extracellular medium to the lumen of the lysosomes. Multiple pathways exist for the delivery of specific and nonspecific cargoes from the cytosol to the vacuole or lysosome (Klionsky and Ohsumi, 1999). Three of these routes (microautophagy, macroautophagy, and the Cvt pathway; see figure) exhibit substantial overlap and share many proteins whose functions are still being elucidated. Two recent papers (Leber et al., 2001; Scott et al., 2001) describe the characterization of an unusual membrane-bound receptor required for the delivery of proteins specifically to the Cvt pathway. Delivery of Cargo to the Vacuole or Lysosome by the Autophagy and Cvt Pathways There are two forms of autophagy, termed microautophagy and macroautophagy (Dunn, 1994). In the former, nonspecific bulk cytosol or specific but dispensable organelles are engulfed by invagination of the vacuolar membrane. The engulfing arms then fuse, resulting in the vacuolar delivery of the cargo wrapped in a single
Selected Reading von Goethe, J.W. (1790). Versuch die Metamorphose der Pflanzen zu erkla¨ren (Gotha, Germany: C.W. Ettinger). Lenhard, M., Bohnert, A., Ju¨rgens, G., and Laux, T. (2001). Cell 105, 805–814. Lohmann, J.U., Hong, R., Hobe, M., Busch, M.A., Parcy, F., Simon, R., and Weigel, D. (2001). Cell 105, 793–803. Mandel, M.A., and Yanofsky, M.F. (1995). Plant Cell 7, 1763–1771. Meyerowitz, E.M., Smyth, D.R., and Bowman, J.L. (1989). Development 106, 209–217. Parcy, F., Nilsson, O., Busch, M.A., Lee, I., and Weigel, D. (1998). Nature 395, 561–566. Pelaz, S., Ditta, G.S., Baumann, E., Wisman, E., and Yanofksy, M.F. (2000). Nature 405, 200–203. Teleman, A.A., Strigini, M., and Cohen, S.M. (2001). Cell 105, 559–562. Sessions, A., Yanofsky, M.F., and Weigel, D. (2000). Science 289, 779–781. Waites, R., and Simon, R. (2000). Cell 103, 835–838.
membrane derived from the invaginated vacuolar membrane (Sakai et al., 1998). In macroautophagy and the Cvt pathway, a double membrane of unknown origin wraps around the proteins, cytosol, or organelles to be degraded and fuses at its edges to create a dual-membrane vesicle surrounding the cargo. The outer membrane of these double-membrane vesicles, called autophagosomes or Cvt vesicles, respectively, fuses with the vacuolar membrane to deliver cargo surrounded by a single membrane into the vacuole (Klionsky and Ohsumi, 1999). In all three processes, the single membrane surrounding the cargo is then destroyed by vacuolar hydrolases and the enclosed contents are released into the vacuolar lumen for delivery, processing, or turnover (see figure). In contrast to the autophagy modes, the Cvt pathway is biosynthetic rather than degradative and appears to be a mechanism for the delivery of oligomerized proteins into the vacuole lumen, where their functions are necessary. In Saccharomyces cerevisiae, the Cvt pathway selectively delivers enzymes, such as a precursor of aminopeptidase I (prAPI) and ␣-mannosidase (Ams1) to the vacuole. The Cvt pathway is distinct from macroautophagy in terms of its cargo selectivity, saturability, faster kinetics, constitutive nature, and the smaller size of the Cvt vesicles. However, many genes involved in this process (CVT genes) are shared with those necessary for macroautophagy (APG/AUT genes) and to a lesser extent with those needed for glucose-stimulated microautophagy (GSA genes) in Pichia pastoris (Yuan et al., 1999). Furthermore, proteins such as API that are normally routed via the Cvt pathway can find their way to the vacuole via macroautophagy under starvation conditions, which induce this process (Klionsky and Ohsumi, 1999). The specific degradation of organelles such as peroxisomes (called pexophagy) in S. cerevisiae also requires many of the same CVT/APG/AUT genes and may be a form of autophagy that targets specific cargo (Hutchins et al., 1999). This view is supported by the finding that specific peroxisome degradation in methylotrophic yeasts can occur both by macroautophagy and