seminars in CELL & DEVELOPMENTAL BIOLOGY, Vol. 12, 2001: pp. 387–394 doi:10.1006/scdb.2001.0268, available online at http://www.idealibrary.com on Cell fate specification in the cereal endosperm Philip W. Becraft In double fertilization, two different cells of the megagametophyte are fertilized. One pollen sperm fertilizes the egg cell to form the zygote and the other sperm nucleus fuses with the two polar nuclei in the central cell to yield the triploid endosperm. The evolutionary origin of the endosperm is an interesting problem with two competing century-old hypotheses to explain its derivation (reviewed by Friedman 1 ). One suggests that the endosperm is derived from an altruistic second embryo produced by double fertilization, while the alternative hypothesis proposes that the endosperm is derived from an extension of megagametophyte development. This issue is pertinent to understanding endosperm development because the developmental program is built on the genetic platform inherited from the ancestral predecessor of the endosperm. Thus, understanding evolution will help us understand development and vice versa. The endosperm is a transitory tissue in many species, particularly dicots such as Arabidopsis where the cotyledons represent the major site of food storage. In cereals, the endosperm is persistent, contains several specialized cell types and is a prominent feature of the mature seed. Despite the very different fates of the endosperm in species such as Arabidopsis and maize, early endosperm development shows remarkable conservation. The cereal endosperm contains three major cell types. The bulk of the endosperm is composed of the starchy endosperm or internal endosperm cells, which accumulate starch and storage proteins. Adjacent to the pedicel is a transfer layer, specialized for nutrient uptake and transport from the maternal tissues to the developing endosperm. An epidermal layer, called the aleurone, covers the endosperm and during seed germination, secretes hydrolytic enzymes to digest the storage products in the starchy endosperm. Understanding how the identity of these different cell types is specified during development has a number of potential practical applications. For While superficially simple, endosperm development is a complex, dynamic process. Cereal endosperms contain three major cell types: starchy endosperm, transfer cells and aleurone. The localized accumulation of the END1 transcript in the syncitial endosperm suggests that signals from the maternal placental tissue specify transfer cell type early. Aleurone fate is plastic and requires the continual input of positional cues to maintain cell identity. Starchy endosperm appears to be the default cell type. Mutant patterns suggest that a regulatory hierarchy integrates endosperm development. Requirements for gametic imprinting, maternal : paternal genome ratios and putative chromatin modeling factors indicate the importance of genomic control. Key words: aleurone / development / endosperm / positional cues / transfer cells c 2001 Academic Press Introduction The endosperm is well recognized for its importance as food and feed, yet the many fascinating features of endosperm biology are often overlooked. The perception of endosperm as uninteresting is perhaps in part because of its familiarity as a foodstuff and part because of the ‘dead end’, often transitory, nature of the endosperm. However, these impressions belie the dynamism and many unique features of endosperm development. Positional cues, cell cycle modifications, genomic imprinting and programmed cell death figure prominently in endosperm development. From the Zoology and Genetics Department and Agronomy Department, Iowa State University, Ames, IA 50011, USA. E-mail: becraft@iastate.edu c 2001 Academic Press 1084–9521 / 01 / 050387+ 08 / $35.00 / 0 387 P. W. Becraft into tube-like alveoli that remain open at their inner face. The next cell division occurs periclinally and is accompanied by cytokinesis to yield a fully cellular peripheral layer and an alveolar internal layer. This process repeats until the entire volume of the endosperm is cellularized. With minor differences, this early development is remarkably conserved in plants as divergent as Arabidopsis, with a transitory endosperm, and cereals. Shortly after cellularization, the Arabidopsis endosperm begins degenerating as the growing embryo crushes surrounding cells. Only an epidermal layer of endosperm, reminiscent of the cereal aleurone, remains in the mature seed. 5 The function of this cell layer is unknown. In cereals, the persistent endosperm grows to a substantial size. Cell division initially occurs throughout the cellular endosperm and then becomes restricted to a cambial zone at the periphery. More interior cells continue to grow by cell enlargement as they accumulate starch and protein bodies. This growth is accompanied by an increase in genomic DNA content through endoreduplication. As growth continues, programmed cell death (PCD) initiates and progresses throughout the starchy endosperm until only the aleurone cells remain alive in the mature endosperm (reviewed by Young 7 ). Castor bean contains a living endosperm that remains attached to the cotyledons and serves as a storage site for lipids. This endosperm undergoes PCD after germination, 8 similar to the aleurone of cereals. 9 example, manipulating the transfer layer might allow enhanced solute uptake efficiency, leading to increased yield. The aleurone is relatively lipid rich and so altering aleurone development could allow seed composition to be manipulated. In addition, the aleurone layer is rich in phytic acid, which is detrimental in uncooked corn products because of its ability to bind dietary Ca2+ and Zn2+ . It is also a leading source of phosphorus in livestock wastes, posing a serious water pollution problem. Finally, the aleurone is the major source of amylase in the malting process and is therefore of great interest to the malt beverage industry. This paper will highlight recent advances in our understanding of how cell fates are decided during cereal endosperm development. Overview of endosperm development There are three main types of endosperm development, nuclear, cellular and helobial. Only nuclear endosperm development will be considered here as it occurs in most familiar plants including cereals and Arabidopsis. Following fertilization of the central cell, the primary endosperm nucleus enters a period of free nuclear division. Although the early endosperm is syncitial, the daughter nuclei of early divisions migrate to relatively invariant positions. Clonal analysis in maize showed that following the first division, daughter nuclei occupy lateral positions and give rise to the left and right halves of the endosperm. 2 Products of subsequent divisions give rise to predictable quarters and eighths and so forth. These results demonstrate that positional information is present within the early syncitial endosperm, possibly preexisting within the megagametophyte. After a period of free nuclear division, the nuclei occupy the peripheral cytoplasm surrounding a large central vacuole and cytoskeletal arrays define nucleocytoplasmic domains (NCDs). 3–5 In cereals, syncitial divisions appear synchronous 3 while in Arabidopsis there appear to be mitotic domains. 6 Cellularization ensues through a process of free cell wall formation whereby unusual microtubule structures called adventitious phragmoplasts form at the boundaries between NCDs. The phragmoplasts direct membrane vesicles containing cell wall constituents to the growing points of the cell walls. Beginning at the periphery and growing centripetally, cell walls grow between non-daughter nuclei, separating the nuclei Cell fate acquisition in cereal endosperm As mentioned, the cereal endosperm has three major cell types. In addition, there are regions or subtypes of cells within the starchy endosperm that are histologically distinct or show specialized patterns of gene expression. The organization of these cells in the maize endosperm is illustrated in Figure 1. Transfer layer The transfer layer, also called ‘modified aleurone’, is the first cell type to become histologically differentiated during endosperm development, becoming visible at approximately 6 days after pollination (DAP) in maize. 10 These cells develop extensive inward cell wall projections that increase the area of plasma membrane by approximately 20fold [Figure 1(c); see Thompson in Reference 11 for review]. Molecular evidence suggests that positional 388 Cell fate specification in the cereal endosperm is also unknown. BETL1 was originally proposed to have a structural function because of its tight association with the cell wall but the small cysteinerich BETL1 and BETL3 peptides also show similarity with defensins, suggesting possible antimicrobial functions. However, SCR, the pollen determinant of sporophytic self-incompatibility in Brassica, and likely ligand for the S receptor kinase (SRK), shows a similarity to defensins. 15 Thus it is also possible that the BETL1 and 3 peptides function as signaling molecules. BETL2 has recently been shown to be a member of a small family of related peptides that possess antifungal activity. 16 These peptides are called BAP1-3 (basal layer antifungal proteins) and BETL2 has been renamed BAP2. Interestingly, a proper 2 : 1 maternal to paternal genomic ratio is critical for transfer cell differentiation in maize. 17 The basal cells of tetraploid (2m : 2p) endosperms fail to initiate cell wall ingrowths. In addition, they show cytoplasmic disorganization and abnormal inclusions in the cell wall matrix. Defects in other parts of the endosperm, including necrosis and failure of aleurone cell differentiation were considered potential secondary effects of impaired nutrient uptake. These effects of paternal genomic excess suggest that a regulatory factor, critical for transfer cell fate or differentiation, is differentially imprinted in the pollen and the egg. Consistent with the role of the basal transfer cells in solute uptake, a cell wall acid invertase gene, INCW1, is specifically expressed in these cells. 18 In fact, sugar sensing, and possibly flux, appear to influence the development of maternal and endosperm tissues. Endosperms mutant in mn1, which codes for INCW2, cause degeneration of the maternal placental tissues. 19,20 Since hydrolysis of sucrose to hexose is a key step in sugar import to the endosperm, the placental defects might result from reduced flux. The reduced grain filling1 (rgf1) mutant has a similar phenotype to mn1 but has apparently normal sugar levels, suggesting the lesion may be impaired sugar sensing. Both mn1 and rgf1 showed morphologically normal basal transfer cells but in the rgf1 mutant the levels of BETL1 and BETL2 (BAP2) proteins were decreased. 21 This did not occur in mn1 or several starch biosynthetic mutants. The basal transfer cell-specific INCW1 transcript shows sugar dependent post-transcriptional regulation. 22 Thus, sugars influence gene expression in the basal transfer cells. The differentiation of transfer cells in other systems has been shown to be dependent on solute transport (reviewed by Thompson 11 ), hence it would B A p a sa se C wp D em embryo pericarp BETL ESR starchy endosperm conducting tissue aleurone subaleurone BETL1 Figure 1. Cell types of the cereal endosperm. (a) A diagrammatic representation of a developing maize kernel in longitudinal section. The pericarp is maternally derived tissue that covers the kernel and the embryo is the other product of double fertilization. Everything else is endosperm. (b) A histological section through the peripheral crown region of a mature maize kernel, showing starchy endosperm (se), subaleurone (sa), aleurone (a) and the pericarp (p). (c) A histological section of the basal transfer layer cells showing the extensive cell wall projections (wp) associated with this cell type. The cells to the lower right part of the figure belong to the maternal placental tissue, which unloads the solutes taken up into the endosperm (upper left) by the transfer cells. (d) In situ hybridization showing the expression of the BETL1 transcript specifically in the basal transfer layer. cues demarcate the presumptive transfer cell region even prior to cellularization. In barley, the END1 transcript accumulates in the syncitial endosperm over the placental crease, the region destined to form transfer cells. 12 It was hypothesized that a signal from the placenta induces the expression of END1 and that this transcript may be anchored to the cytoskeleton in this region of the syncitial endosperm. The function of this gene is currently not known. In maize, transcripts for four genes, BETL1-4, are transiently expressed in the transfer cells during early- to mid-stage endosperm development. 13,14 Whether these genes have a developmental function 389 P. W. Becraft peripheral cell layer adapts a starchy endosperm fate. Therefore the wild-type DEK1 gene product is required for the perception or response to the positional cues that specify aleurone cell fate. In a transposon-induced dek1 mutant, revertant sectors as small as a single cell allowed aleurone cell development. Because the revertant cell’s ability to perceive or respond to the positional cues was not restored until near the end of active cell division, the positional cues that specify aleurone identity must still be present late in development. Furthermore, starchy endosperm cells in a peripheral position remain competent to respond to those signals. In a reciprocal experiment, Ds induced chromosome breakage was used to cause the loss of the wild-type DEK1 gene in a dek1/DEK1 heterozygote, uncovering the mutant allele in sectors. 28 Cells within mutant sectors assumed starchy endosperm identity, even in late sectors as small as two cells. Because aleurone cells were clearly recognizable at earlier stages of development, the late loss of DEK1 gene function caused peripheral cells to lose aleurone identity and switch fate to starchy endosperm. This demonstrates that the positional cues that specify aleurone identity are also required to maintain it. Such transdifferentiation events demonstrate that cell fate specification is a continual process and highlight the developmental plasticity of the endosperm. The maize crinkly4 mutant causes mosaic aleurone development. In regions that lack aleurone, the peripheral cells develop as starchy endosperm. This fate switch of peripheral cells indicates that CR4 is involved in the aleurone cell fate decision. The identification of CR4 as a receptor-like kinase 27 suggested the exciting possibility that CR4 may be the receptor for the positional cues that specify aleurone identity. 24,27 CR4 is not specific to the aleurone but rather regulates a wide variety of cellular developmental processes in the plant and endosperm, suggesting that the function is analogous to a growth factor receptor, where a widespread signal(s) regulates various responses in a contextdependent manner. 30 A region in the extracellular domain shows similarity to the ligand-binding domain of the mammalian tumor necrosis factor receptor (TNFR) family suggesting that the ligand for CR4 (i.e. the putative positional cue for aleurone identity) may be related to TNF. 27 Unfortunately, the TNF family shows poor primary sequence conservation, 31 creating difficulty for sequence based strategies for identifying TNF-related molecules as potential CR4 not be surprising if source–sink relationships were important for inducing the formation of endosperm transfer cells. Aleurone During cellularization, the periclinal division that first sets aside a cellular peripheral cell layer gives rise to two daughter cell lineages with different fates; the internal daughter has the invariant fate of producing only starchy endosperm cells while the peripheral cell layer will produce aleurone cell initials. Because of the different fates of the daughter cells, this early periclinal division has been called a formative division 4,23,24 and it was commonly believed that the aleurone formed a separate cell lineage from the starchy endosperm. 10,25–27 However it is clear, at least in maize, that this initial cell division is not a determinative event but the daughter cell fates remain plastic through late stages of endosperm development. 28 This conclusion is based on endosperm cell lineage analysis and developmental genetic analysis of the DEK1 gene function, required for aleurone cell fate. Cell lineage analysis was conducted in maize endosperm using a technique similar to that of Barbara McClintock. 2 A chromosome breaking Ds transposon was used to simultaneously mark starchy endosperm and aleurone by uncovering the wx and C1 markers, respectively. 28 The size of a sector indicates the relative time during development when an event occurred; small sectors occurred late in development because there were few cell divisions following the event. Small sectors containing both aleurone and starchy endosperm were observed, indicating that both cell types could be derived from a common precursor at late stages of development. These results show that the aleurone cell lineage is not distinct but that the peripheral cell layer contributes cells internally to differentiate as starchy endosperm throughout development. This is consistent with a report from wheat that describes the redifferentiation of aleurone cells to starchy endosperm following an aleurone cell division that contributed a daughter cell internally. 29 That starchy endosperm cells can be derived from the aleurone cell lineage implies that position, not lineage, determines endosperm cell fate. The question of aleurone cell fate determination was further addressed by examining unstable dek1 alleles. 28 dek1 is a recessive mutation in which aleurone differentiation fails to initiate and the 390 Cell fate specification in the cereal endosperm starchy endosperm fate instead of aleurone. In a sense then, the question of how starchy endosperm cell fate is specified comes down to the question of how the endosperm, with the same genetic makeup as the embryo except for the 2 : 1 maternal : paternal genomic ratio, is specified. This, in turn returns to the question of the evolutionary origins of the endosperm. More work is required before these questions can be answered. ligands. Genetic analysis suggests that a hierarchy of genetic functions is involved in the specification and differentiation of aleurone cells. 28 Genetics may represent the most viable approach to identifying the CR4 ligand and will help identify other components of the CR4 signal transduction system. Maternal : paternal genomic ratio is also important for aleurone development because tetraploid endosperm fails to differentiate an aleurone layer. 17 It was not reported whether tetraploidy caused the peripheral endosperm cells to switch fate to starchy endosperm or simply to differentiate incompletely. Whether this particular effect was due to a regulatory factor that controls aleurone development being differentially imprinted in male and female gametes or was a secondary consequence of disrupted transfer cell differentiation cannot be determined at present, but there is ample evidence that imprinting is critical for aleurone development. Imprinting of anthocyanin factors in the aleurone is a well-known phenomenon 32,33 and Dappled mutants that disrupt aleurone development only show phenotypes when inherited through the female, suggesting they are imprinted. 34 Among the molecular markers that have been identified for the aleurone are lipid transporters, 35–37 Myo-inositol-1-phosphate synthase, 38 a defenserelated gene, 39 and in certain genotypes of maize, genes of the anthocyanin biosynthetic pathway. 34,40 Specialized regions within the starchy endosperm Within the starchy endosperm are several specialized regions of note. These include the subaleurone, the ‘embryo surrounding region’ (ESR) and conducting zone (Figure 1). The subaleurone is a cambial zone of active cell division. Cells are small and generally contain small amyloplasts in the early stages of starch accumulation. 29,45 The specific expression of the basic leucine zipper transcriptional factor, RISBZ1, in the aleurone and subaleurone 46 suggests that the subaleurone might in fact possess a distinct identity. However, this might also merely reflect a phase in starchy endosperm cell development because there is a progression of cells passing from the subaleurone to mature regions of the endosperm as new cells are added to the periphery during endosperm growth. Cells surrounding the embryo have a distinct morphology, with compact size and dense cytoplasm containing extensive rough endoplasmic reticulum. 47 This morphology suggests a synthetic function, possibly producing substances to be taken up by the suspensor of the developing embryo. Three genes called ESR1-3 are specifically expressed in this region. 48,49 The function of these genes is unknown but the promoters of all three drive GUS expression in the ESR indicating that regulators function specifically within this region. 48 The pattern is fairly static in the endosperm, initially surrounding the whole embryo but as the embryo proper grows up and out of this region, the ESR is left surrounding just the suspensor. It has been reported that developing maize endosperm contains a column of elongated conducting cells extending through the central core of the endosperm from the transfer layer to near the apex. 17 These cells are called prismatic cells in barley 50 and are believed to transport solutes throughout the kernel. Like the transfer cells and aleurone, this conducting tissue failed to differentiate in tetraploid endosperms, suggesting either that a determinant is differentially imprinted in male and Starchy endosperm Starchy endosperm cells constitute the bulk of the cereal endosperm and represent the major storage site for starch and protein. Despite intensive study, it is surprising that no genetically defined regulatory factors for starchy endosperm development have been isolated. Even among the various metabolic pathways to storage product accumulation, surprisingly few regulators have been isolated. Mutants are known in many of the enzymes of the starch biosynthetic pathway but no genetic regulators have been identified. OPAQUE2 is the only genetically defined regulator of storage proteins that has been isolated. 41,42 The paucity of identified genetic regulators in these systems suggests that there is either a high degree of redundancy or that such mutants are lethal, perhaps present in the many non-descript defective kernel (dek) mutants. 43,44 Available evidence suggests that starchy endosperm represents the default cell type because loss-offunction mutants cause peripheral cells to adapt a 391 P. W. Becraft female gametes or that the activity of the transfer layer is required for the differentiation of this tissue as well. No molecular markers have yet been reported for this tissue but two barley mutants specifically eliminate this tissue from the endosperm. 50 Pattern formation in the endosperm From the array of specialized cell types and molecular markers, each with their characteristic spatial patterns within the endosperm, it is apparent that pattern formation is an essential component of endosperm development. Analysis of mosaic aleurone mutants of maize (Figure 2) suggested that endosperm development is integrated by at least two systems. 28 Mosaic patterns of mutants that affect early steps of the aleurone development have a strong tendency for showing defective aleurone differentiation preferentially in the abgerminal crown region, reflecting a gradient from the silk scar region of the kernel. The source of the positional information for this gradient is not the embryo because a mutant that altered embryo position did not alter the distribution of the mosaic pattern. This pattern is reminiscent of the anterior–posterior pattern shown by the MEA, FIS2 and FIE genes of Arabidopsis. 51,52 These putative chromatin-modeling factors are required in the megagametophyte to inhibit endosperm development until fertilization and are involved in establishing axial pattern in the developing endosperm (see accompanying article by Chaudhury 53 ). It will be interesting to see whether the homologs of these genes are involved in establishing the germinal–abgerminal pattern in maize. Mutants that affect later stages of aleurone differentiation show blocky patterns of mosaicism, first in large patterns, then in smaller patterns, suggesting that development is somehow integrated in progressively smaller fields. What these fields represent, whether the limits of a diffusible factor, symplastic fields, or something else, and the functional meaning of these fields in endosperm development are questions that will be answered as the genes that show these mutant patterns are isolated. (a) (b) (c) Figure 2. Representative mutants showing the characteristic patterns for each class of mosaic aleurone phenotype. 28 The dark regions of the kernels contain anthocyanin specifically in the aleurone. The light regions show areas that fail to accumulate anthocyanin because aleurone differentiation is perturbed. The highly characteristic patterns of mutants in multiple genes suggest they reflect an underlying developmental pattern. (a) The bareback* mutant showing the germinal–abgerminal pattern also observed in crinkly4 and dek1 mutants. (b) A collapsed2-o12 kernel showing a large blocky pattern of mosaicism, also typical of several Dap and Msc mutants. (c) The white2-dek21 pattern consisting of small patches of aleurone. Summary cycle is highly modified, first to allow nuclear division without cytokinesis, then to allow an unusual cytokinesis involving free cell wall growth in the absence of nuclear division. Later the cell cycle is modified again to allow endoreduplication in the starchy endosperm The picture that emerges of endosperm development is of a dynamic system, containing several novel features. In early stages of development, the cell 392 Cell fate specification in the cereal endosperm cells before they undergo programmed cell death. The establishment of pattern and the specification of cell fates are also very complex, involving positional cues, signal transduction, chromatin modeling factors, genomic imprinting, probable interactions with maternal tissues and metabolic signaling. Understanding how all these processes are integrated to allow the development of a functional endosperm will require much more research but the potential rewards of being able to improve seed composition, not to mention the intellectual gratification, are ample. 11. Thompson RD, Hueros G, Becker H, Maitz M (2001) Development and functions of seed transfer cells. Plant Sci 160:775–783 12. Doan DNP, Linnestad C, Olsen O-A (1996) Isolation of molecular markers from the barley endosperm coenocyte and the surrounding nucellus cell layers. Plant Mol Biol 31:877–886 13. Hueros G, Varotto S, Salamini F, Thompson RD (1995) Molecular characterization of BET1, a gene expressed in the endosperm transfer cells of maize. Plant Cell 7:747–757 14. Hueros G, Royo J, Maitz M, Salamini F, Thompson RD (1999) Evidence for factors regulating transfer cell-specific expression in maize endosperm. Plant Mol Biol 41:403–414 15. Schopfer CR, Nasrallah ME, Nasrallah JB (1999) The male determinant of self-incompatibility in Brassica. Science 286:1697–1700 16. Serna A et al. (2001) Maize endosperm secretes a novel antifungal protein into adjacent maternal tissue. Plant J 25:687–698 17. Charlton WL, Keen CL, Merriman C, Lynch P, Greenland AJ, Dickinson HG (1995) Endosperm development in Zea mays; implication of gametic imprinting and paternal excess in regulation of transfer layer development. Development 121:3089–3097 18. Taliercio EW, Kim JY, Mahe A, Shanker S, Choi J, Cheng WH, Prioul JL, Chourey PS (1999) Isolation, characterization and expression analyses of two cell wall invertase genes in maize. J Plant Physiol 155:197–204 19. Miller ME, Chourey PS (1992) The maize invertase-deficient miniature-1 seed mutation is associated with aberrant pedicel and endosperm development. Plant Cell 4:297–305 20. Cheng WH, Taliercio EW, Chourey PS (1996) The Miniature1 seed locus of maize encodes a cell wall invertase required for normal development of endosperm and maternal cells in the pedicel. Plant Cell 8:971–983 21. Maitz M, Santandrea G, Zhang Z, Lal S, Hannah LC, Salamini F, Thompson RD (2000) rgf1, a mutation reducing grain filling in maize through effects on basal endosperm and pedicel development. Plant J 23:29–42 22. Cheng WH, Taliercio EW, Chourey PS (1999) Sugars modulate an unusual mode of control of the cell-wall invertase gene (Incw1) through its 30 untranslated region in a cell suspension culture of maize. Proc Natl Acad Sci USA 96:10512–10517 23. Olsen O-A, Linnestad C, Nichols SE (1999) Developmental biology of the cereal endosperm. Trends Plant Sci 4:253–257 24. Olsen O-A, Lemmon B, Brown R (1998) A model for aleurone development. Trends Plant Sci 3:168–169 25. Randolph LF (1936) Developmental morphology of the caryopsis in maize. J Agric Res 53:881–916 26. Walbot V (1994) Overview of key steps in aleurone development, in The Maize Handbook (Freeling M, Walbot V, eds) pp. 78–80. Springer-Verlag, New York 27. Becraft PW, Stinard PS, McCarty DR (1996) CRINKLY4: a TNFR-like receptor kinase involved in maize epidermal differentiation. Science 273:1406–1409 28. Becraft PW, Asuncion-Crabb YT (2000) Positional cues specify and maintain aleurone cell fate during maize endosperm development. Development 127:4039–4048 29. Morrison IN, Kuo J, O’Brien TP (1975) Histochemistry and fine structure of developing wheat aleurone cells. Planta 123:105–116 30. Jin P, Guo T, Becraft PW (2000) The maize CR4 receptor-like kinase mediates a growth factor-like differentiation response. Genesis 27:104–116 Acknowledgements The author is grateful to members of the Becraft lab for helpful discussions and critical reading of the manuscript. Figures 1(c) and (d) were kindly provided by Richard D. Thompson, Max Plank Institute, Koln, Germany. Work in the author’s lab is supported by grants from the National Science Foundation (IBN-9604426) and the Department of Energy, Energy Biosciences (DE-FG02-98ER20303). References 1. Friedman WE (1998) The evolution of double fertilization and endosperm: an ‘historical’ perspective. Sex Plant Reprod 11:6–16 2. McClintock B (1978) Development of the maize endosperm as revealed by clones, in The Clonal Basis of Development (Subtelny S, Sussex IM, eds) pp. 217–237. Academic Press Inc., New York 3. Brown RC, Lemmon BE, Olsen O-A (1994) Endosperm development in barley: microtubule involvement in the morphogenetic pathway. Plant Cell 6:1241–1252 4. Brown RC, Lemmon BE, Olsen O-A (1996) Polarization predicts the pattern of cellularization in cereal endosperm. Protoplasma 192:168–177 5. Brown RC, Lemmon BE, Nguyen H, Olsen O-A (1999) Development of endosperm in Arabidopsis thaliana. Sex Plant Reprod 12:32–42 6. Boisnard-Lorig C, Colon-Carmona A, Bauch M, Hodge S, Doerner P, Bancharel E, Dumas C, Haseloff J, Berger F (2001) Dynamic analyses of the expression of the histone::yfp fusion protein in arabidopsis show that syncytial endosperm is divided in mitotic domains. Plant Cell 13:495–509 7. Young TE, Gallie DR (2000) Programmed cell death during endosperm development. Plant Mol Biol 44:283–301 8. Schmid M, Simpson D, Gietl C (1999) Programmed cell death in castor bean endosperm is associated with the accumulation and release of a cysteine endopeptidase from ricinosomes. Proc Natl Acad Sci USA 96:14159–14164 9. Bethke PC, Lonsdale JE, Fath A, Jones RL (1999) Hormonally regulated programmed cell death in barley aleurone cells. Plant Cell 11:1033–1046 10. Kiesselbach TA (1949) The structure and reproduction of corn. Nebr Agric Exp Stn Res Bull 161:1–96 393 P. W. Becraft Science 238:960–963 42. Schmidt RJ, Burr FA, Aukerman MJ, Burr B (1990) Maize regulatory gene opaque-2 encodes a protein with a ‘leucinezipper’ motif that binds to zein DNA. Proc Natl Acad Sci USA 87:46–50 43. Neuffer MG, Sheridan WF (1980) Defective kernel mutants of maize. I. Genetic and lethality studies. Genetics 95:929–944 44. Scanlon MJ, Stinard PS, James MG, Myers AM, Robertson DS (1994) Genetic analysis of 63 mutations affecting maize kernel development isolated from Mutator stocks. Genetics 136:281–294 45. Kyle DJ, Styles ED (1977) Development of aleurone and subaleurone layers in maize. Planta 137:185–193 46. Onodera Y, Suzuki A, Wu CY, Washida H, Takaiwa F (2001) A rice functional transcriptional activator, RISBZ1, responsible for endosperm-specific expression of storage protein genes through GCN4 motif. J Biol Chem 276:14139–14152 47. Schel JHN, Kieft H, van Lammeren AAM (1984) Interactions between embryo and endosperm during early developmental stages of maize caryopses (Zea mays). Can J Bot 62:2842–2853 48. Bonello JF, Opsahl-Ferstad HG, Perez P, Dumas C, Rogowsky PM (2000) Esr genes show different levels of expression in the same region of maize endosperm. Gene 246:219–227 49. Opsahl-Ferstad H-G, Le Deunff E, Dumas C, Rogowsky PM (1997) ZmEsr, a novel endosperm-specific gene expressed in a restricted region around the maize embryo. Plant J 12:235–246 50. Bosnes M, Weideman F, Olsen O-A (1992) Endosperm differentiation in barley wild-type and sex mutants. Plant J 2:661–674 51. Luo M, Bilodeau P, Dennis ES, Peacock WJ, Chaudhury A (2000) Expression and parent-of-origin effects for FIS2, MEA, and FIE in the endosperm and embryo of developing Arabidopsis seeds. Proc Natl Acad Sci USA 97:10637–10642 52. Sørensen MB, Chaudhury AM, Robert H, Bancharel E, Berger F (2001) Polycomb group genes control pattern formation in plant seed. Curr Biol 11:277–281 53. Chaudhury A, Berger F (2001) Maternal control of seed development. Seminars Cell Dev Biol 12:381–386 31. Locksley RM, Killeen N, Lenardo MJ (2001) The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell 104:487–501 32. Kermicle JL, Alleman M (1990) Gametic imprinting in maize in relation to the angiosperm life cycle. Dev Suppl 1:9–14 33. Alleman M, Doctor J (2000) Genomic imprinting in plants: observations and evolutionary implications. Plant Mol Biol 43:147–161 34. Gavazzi G, Dolfini S, Allegra D, Castiglioni P, Todesco G, Hoxha M (1997) Dap (Defective aleurone pigmentation) mutations affect maize aleurone development. Mol Gen Genet 256:223–230 35. Kalla R, Shimamoto K, Potter R, Nielsen PS, Linnestad C, Olsen O-A (1994) The promoter of the barley aleuronespecific gene encoding a putative 7 kDa lipid transfer protein confers aleurone cell-specific expression in transgenic rice. Plant J 6:849–860 36. Skriver K, Leah R, Muller-Uri F, Olsen FL, Mundy J (1992) Structure and expression of the barley lipid transfer protein gene Ltp1. Plant Mol Biol 18:585–589 37. Aalen RB (1995) The transcripts encoding two oleosin isoforms are both present in the aleurone and in the embryo of barley (Hordeum vulgare L.) seeds. Plant Mol Biol 28:583–588 38. Yoshida KT, Wada T, Koyama H, Mizobuchi-Fukuoka R, Naito S (1999) Temporal and spatial patterns of accumulation of the transcript of Myo-inositol-1-phosphate synthase and phytin-containing particles during seed development in rice. Plant Physiol 119:65–72 39. Skadsen RW, Sathish P, Kaeppler HF (2000) Expression of thaumatin-like permatin PR-5 genes switches from the ovary wall to the aleurone in developing barley and oat seeds. Plant Sci 156:11–22 40. Kao CY, Cocciolone SM, Vasil IK, McCarty DR (1996) Localization and interaction of the cis-acting elements for abscisic acid, VIVIPAROUS1, and light activation of the C1 gene of maize. Plant Cell 8:1171–1179 41. Schmidt RJ, Burr FA, Burr B (1987) Transposon tagging and molecular analysis of the maize regulatory locus opaque-2. 394