Abstract
The aleurone layer of cereal grains is important biologically as well as nutritionally and economically. Here, current knowledge on the regulation of aleurone development is reviewed. Recent reports suggest that the control of aleurone development is more complex than earlier models portrayed. Multiple levels of genetic regulation control aleurone cell fate, differentiation, and organization. The hormones auxin and cytokinin can also influence aleurone development. New technical advances promise to facilitate future progress.
Introduction
Cereal grains represent one of the most important plant products for humankind. As food and animal feed, they are the major source of human caloric intake. They are also important for leisure activities as the raw material for recreational beverages, and in recent times they have become increasingly important for biofuels and other industrial purposes. Several key cell types of the endosperm are important for conferring properties that are biologically important and make endosperm useful to people. The predominant cell type is starchy endosperm, the major storage cell type that accumulates starch grains and storage protein bodies. This forms the central bulk of the cereal endosperm. Transfer cells occur in a layer at the interface with the maternal pedicel tissue and are specialized to take up solutes from the maternal tissues into the developing grain. The epidermal-like aleurone layer forms on the surface of the endosperm and is important for digesting the endosperm storage products during germination. This review focuses on recent research into the developmental basis of how aleurone cell differentiation is regulated.
In most cereals, including typical maize (Zea mays L.) lines, the aleurone is a single cell layer (Fig. 1), although in barley (Hordeum vulgare L.) it comprises about three cell layers and is variable in rice (Oryza sativa L.). Aleurone cells are distinguished from starchy endosperm cells by their morphology, biochemical composition, and gene expression profiles. Starchy endosperm cells tend to be large and irregularly shaped, packed with starch grains and storage protein bodies, and to express genes associated with starch and storage protein deposition. As shown in Fig. 1, aleurone cells have regular cuboidal shapes, contain densely granular cytoplasms due to the accumulation of vacuolar inclusion bodies called aleurone grains (Buttrose, 1963), lack starch grains, and specifically express a variety of genes such as vp1 that can be used as molecular markers (Fig. 1; Kalla et al., 1994; Costa et al., 2003; Wisniewski and Rogowsky, 2004; Furtado and Henry, 2005; Cao et al., 2007; Gomez et al., 2009). In particular maize genotypes, the aleurone cells specifically accumulate anthocyanin pigment, which serves as a convenient marker (Becraft and Asuncion-Crabb, 2000; Cone, 2007).
Microscopic sections of maize endosperms. (A, C) The wild type and (B, D) the dek1 mutant. (A, B) Histological staining. Starchy endosperm cells are filled with starch grains, which stain pink with periodic acid–Schiff (PAS) reagent. The yellow arrow in A highlights the aleurone layer with dense granular cytoplasm and cuboidal cells. The aleurone is absent in a dek1 mutant and surface cells have starchy endosperm identity. (C and D) A VP1:GUS transgene marks aleurone cells (Cao et al., 2007). Wild-type endosperm shows aleurone-specific β-glucuronidase (GUS) activity, while the marker is not expressed in dek1. Size bars=50μM.
As mentioned, the most well known function of the aleurone is as a digestive tissue. During seed maturation, abscisic acid (ABA) induces aleurone cells to acquire desiccation tolerance and survive seed drying, while starchy endosperm cells undergo programmed cell death (Young et al., 1997; Young and Gallie, 2000). Upon imbibition, the embryo produces gibberellin (GA), which induces the aleurone to secrete amylases and proteases that break down the stored starch and proteins in the dead starchy endosperm, making free sugars and amino acids available to the growing seedling (reviewed in Fath et al., 2000). The aleurone is also a major site of mineral storage (Stewart et al., 1988) and serves to protect the nutrient-rich endosperm by expressing an array of stress- and pathogen-protective proteins such as PR-4 (Jerkovic et al., 2010). In Arabidopsis thaliana, the aleurone controls seed dormancy (Bethke et al., 2007).
Besides its biological importance, the aleurone is also of practical importance. Approximately half of cereal bran is composed of aleurone, which is the most dietarily beneficial fraction of the bran (Cheng et al., 1987; Harris et al., 2005; Graham et al., 2009; Stewart and Slavin, 2009; Okarter and Liu, 2010). Compositional analysis shows that many bran constituents considered healthful are highly concentrated in the aleurone (Buri et al., 2004).
Overview of endosperm and aleurone development
The endosperm is produced during double fertilization when one of the sperm nuclei undergoes syngamy with the two polar nuclei in the central cell. Thus, the endosperm is triploid, with two copies of the maternal haploid genome and one copy of the paternal haploid genome. The endosperm nuclei then undergo a period of free nuclear division, followed by migration to the periphery of the central cell. Despite the absence of cell walls, the first few divisions appear regularly ordered such that the daughters give rise to predictable sectors of the mature endosperm (McClintock, 1978). With the nuclei arranged around the periphery, cellularization ensues with the formation of anticlinal cell walls to produce alveoli. Each alveolus contains a nucleus and is open centripetally to the common cytoplasm. Periclinal divisions then produce a cellular peripheral layer and an alveolar interior layer. This process is repeated several times in the inner alveolar layers until the whole volume of the endosperm is cellular (Kiesselbach, 1949; Brown et al., 1994, 1996; Olsen, 2001).
Upon cellularization, the peripheral cells and internal cells display distinct behaviours. The peripheral cells show a typical plant cell division cycle with a pre-prophase band of microtubules anticipating the future cell division plane (Brown et al., 1994). In addition, the divisions are highly ordered, occurring almost exclusively in the anticlinal and periclinal planes (Randolph, 1936; Kiesselbach, 1949). In contrast, internal cells show an atypical division cycle lacking a pre-prophase band and showing unordered division planes. Thus, developmental cues distinguish the peripheral cell layer from the internal cells from the onset of the cellular phase of endosperm development.
Despite the different behaviours of these cells, the peripheral layer is capable of contributing daughter cells to the interior. Periclinal divisions can be observed in the peripheral cell layer, and lineage markers demonstrate that the aleurone and underlying starchy endosperm cells remain clonally related throughout endosperm development. When a periclinal division occurs in the aleurone layer, the internally contributed daughter cell initially possesses aleurone characteristics but then redifferentiates to form a starchy endosperm cell (Randolph, 1936; Kiesselbach, 1949; Morrison et al., 1975; Becraft and Asuncion-Crabb, 2000; Gruis et al., 2006).
The plasticity of cell fate was also evident in genetic mosaic studies. Maize defective kernel1 (dek1) loss-of-function mutants have no aleurone layer (Fig. 1), indicating that this gene's function is required for the signalling or perception of the signals that specify the outer cell layer as aleurone (Becraft et al., 2002; Lid et al., 2002). Similar functions have been reported in Arabidopsis and rice (Lid et al., 2005; Hibara et al., 2009). In an unstable transposon-induced maize dek1 mutant, gene reversion in surface cells at late stages of development causes the transdifferentiation of starchy endosperm cells to aleurone. Transdifferentiation was similarly observed in cultured endosperms when fissures developed and starchy endosperm cells newly exposed to a surface then became aleurone (Gruis et al., 2006). Conversely, when dek1 mutant cells were generated late in the development of normal kernels, aleurone cells transdifferentiated to starchy endosperm. These results, summarized in Fig. 2, indicate that developmental fate in the periphery of the endosperm remains plastic throughout development, that the positional cues that specify aleurone identity are present throughout development, and that these cues are required to maintain aleurone identity (Becraft and Asuncion-Crabb, 2000).
Developmental plasticity of endosperm cells. (A) Reversion of a dek1 mutant cell to wild type results in the transdifferentiation of the peripheral cell from starchy endosperm to aleurone. (B) Induction of a dek1 mutant cell in a wild-type background results in the transdifferentiation of aleurone to starchy endosperm. Such transdifferentiation events can occur even late in development, illustrating that the cues that specify aleurone cell fate are present throughout development and required to maintain aleurone identity.
Genetic regulation of aleurone development
Several genes controlling aleurone differentiation have been described. As mentioned, the dek1 gene is required for aleurone cell fate specification. The dek1 gene encodes a large complex integral membrane protein that localizes to the plasma membrane (Lid et al., 2002; Tian et al., 2007). An extracellular loop suggests the DEK1 protein has the potential to interact with extracellular molecules, including signalling ligands. The cytoplasmic domain contains an active calpain protease (Wang et al., 2003).
CRINKLY4 (CR4) is a receptor-like kinase that also functions as a positive regulator of aleurone fate (Becraft et al., 1996, 2001; Jin et al., 2000). Recessive cr4 mutants show sporadic patches that lack aleurone, predominantly on the abgerminal face of the kernel. The phenotypes of cr4 mutants resemble those of dek1-D, a weak allele of dek1. Genetic interactions between cr4 and dek1 mutants strongly suggest that the two genes function in overlapping biological processes (Becraft et al., 2002). Immunological studies demonstrated that DEK1 and CR4 proteins co-localized in the plasma membrane and in endocytic vesicles (Tian et al., 2007). At this point it remains unclear whether one factor functions upstream of the other or whether their functions converge downstream. The weaker phenotype and intermediate cell fate of peripheral cells in cr4 mutants suggest that cr4 is more likely to be downstream of dek1 (Wisniewski and Rogowsky, 2004).
Mutants of the supernumerary aleurone1 (sal1) gene have multiple layers of aleurone cells instead of the normal single layer, indicating that the gene functions as a negative regulator of aleurone fate (Shen et al., 2003). The SAL1 protein resembles human CHMP1, a protein involved in vesicle trafficking. It has been proposed that SAL1 functions as a negative regulator of DEK1 and CR4 by directing their retrograde cycling off the plasma membrane and thus dampening their signalling activity. Co-localization of the SAL1 protein with DEK1 and CR4 in endocytic vesicles is consistent with this proposed function (Tian et al., 2007).
A barley mutant, des5, has pleiotropic effects on seed development, including aleurone defects (Bosnes et al., 1987; Olsen et al., 2008). Instead of the typical three-layered aleurone, des5 mutant seeds possess a single layer of cells with compromised aleurone characteristics. The cells are larger, the cytoplasmic contents less dense, and anticlinal cell walls are thinner. Although the molecular identity of des5 has not yet been reported, it provides some key insights into the regulation of aleurone development. There is a strong reduction in the level of cr4 transcripts in des5 mutants, while decreases in dek1 transcripts were not significant. The differential effect on expression suggests that the dek1 and cr4 genes may be regulated independently. Furthermore, sal1 transcripts also showed a decreased expression, indicating that the regulation of cell layer number might be more complex than the interplay between cr4 and sal1 functions.
The involvement of additional factors is further supported by quantitative trait locus (QTL) mapping of aleurone layer number in barley. Diverse parents with two- versus three-layered aleurones were crossed to generate a mapping population. Three major loci were identified which accounted for>50% of the trait variation in this population. These loci did not co-localize with Sal1, indicating that the aleurone layer number is regulated by variation in other, as yet unidentified, factors (Jestin et al., 2008).
In rice, two transcription factors that control the expression of seed storage proteins also influence aleurone cell fate (Kawakatsu et al., 2009). RISBZ1 and RPBF are bZIP and DOF zinc finger transcription factors, respectively. Together they function to promote the expression of most storage proteins of the rice endosperm (Yamamoto et al., 2006). Transgenic knockdown (KD) lines of each factor were generated, apparently by co-suppression (Kawakatsu et al., 2009). The RISBZ1 KD line had normal aleurone but the RPBF KD line showed sporadic multilayered aleurone, while the RISBZ1/RPBF double KD line produced aleurone consisting of multiple layers of large, disordered cells. Interestingly, OsCR4, OsDEK1, and OsSAL1 transcript levels were all decreased in each of the KD lines, reinforcing the notion that the control of aleurone cell layer number is more complex than just these three factors.
There appear to be multiple levels in the regulation of aleurone differentiation. This is evident in the phenotypes of different mutants that indicate that genes function at different levels in the specification of aleurone cell fate and in the subsequent differentiation process (Becraft and Asuncion-Crabb, 2000). It is also evident in the differential effect of several mutants on various molecular markers (Wisniewski and Rogowsky, 2004). In addition to the cues that pattern the endosperm by identifying the surface position, there appears to be communication that occurs at a more localized level. In normal endosperm, the aleurone is extremely regular and the patterns of cell division and subsequent behaviour of daughter cells is highly organized (Randolph, 1936; Kiesselbach, 1949; Morrison et al., 1975; Brown et al., 1994; Becraft and Asuncion-Crabb, 2000). Maize disorgal (dil1 and dil2) mutants have aleurone cells arranged at the periphery of the endosperm, but instead of the highly ordered single layer of cuboidal cells the mutant cells are irregularly shaped and sized. The number of layers is variable and internal cells lack the columnar arrangement that reflects lineage relationships in the wild type (Lid et al., 2004). The phenotype of the barley elo2 mutant shows similar disorganization and cellular irregularities of the aleurone layers (Lewis et al., 2009). These mutant phenotypes demonstrate that regulatory functions are required for aleurone cells to negotiate and coordinate their behaviours with neighbouring cells and generate the ordered arrangement seen in normal endosperm.
Positional regulation of aleurone fate
While it is clear there are positional cue(s) that induce aleurone cell fate in the peripheral layer of the endosperm, the nature and source of these cues remain elusive. Because the endosperm is normally embedded in maternal nucellar tissues, one obvious hypothesis is that these surrounding tissues signal to the surface cells of the endosperm. However, several lines of evidence suggest that the endosperm may be self-organizing. A maize mutant produces an endosperm composed of multiple spheroid masses of cells, each with a layer of aleurone cells (Olsen, 2004). Since some of these aleurone cells were not in contact with maternal tissues, Olsen proposed the ‘surface rule’ where an intrinsic property of the surface position induces aleurone identity. This hypothesis is supported by recent reports of aleurone differentiation on isolated endosperms grown in vitro in the absence of maternal tissues (Gruis et al., 2006; Reyes et al., 2010). This is also consistent with the observation that upon fusion of distinct endosperms, the internalized aleurone layers in the fusion plane redifferentiate as starchy endosperm (Geisler-Lee and Gallie, 2005). However, isolated foci of internalized aleurone cells persisted at sites along these fusion zones, and internal aleurone cells were observed in the disorganized endosperm of the maize globby1 (glo1) mutant (Costa et al., 2003). Thus, in exceptional circumstances, it appears feasible for internal endosperm cells to possess aleurone identity.
Involvement of hormones in aleurone development
The hormones ABA and GA are well known to act antagonistically to mediate late stages of aleurone development, with aleurone maturation promoted by ABA and germination promoted by GA (Bethke et al., 2006). Evidence suggests that earlier stages of aleurone differentiation are also influenced by hormones, in particular auxin and cytokinin. The cytokinin biosynthetic enzyme gene isopentenyl transferase (IPT) was placed under the control of the senescence-responsive SAG12 promoter (Geisler-Lee and Gallie, 2005). Expression of this transgene during maize kernel development resulted in the production of mosaic aleurone (interspersed patches of aleurone and starchy endosperm cells) on the crown region of kernels, suggesting that cytokinin has an inhibitory effect on aleurone fate. It remains unclear whether endogenous cytokinins function in controlling normal aleurone differentiation.
Conversely, a recent report suggests that auxin might have a positive influence on aleurone fate. Maize plants treated with the auxin transport inhibitor, N-1-naphthylphthalamic acid (NPA), produced kernels with multiple layered aleurones, compared with the single layer of normal maize (Forestan et al., 2010). Immunological indole acetic acid detection showed that NPA treatment caused a build-up of auxin in the periphery of the endosperm. There was also an expansion in the expression of the auxin transporter, ZmPIN1, from the single aleurone layer in untreated kernels to several layers in NPA-treated kernels. One possible interpretation of these results is that a high auxin concentration is a cue for aleurone cell fate and the multilayered aleurone seen in NPA-treated kernels is a result of the expanded region of auxin accumulation (Fig. 3).
Hypothetical auxin concentration gradient model for the control of aleurone cell fate. Auxin concentrations are highest at the endosperm periphery and decline internally. Above a threshold concentration (dotted line), aleurone fate is specified. In normal maize endosperm, the gradient is steep and only the outer cell layer is above threshold concentration. Upon NPA treatment, auxin accumulates above threshold levels in an increased number of cell layers, resulting in multilayered aleurone.
Mutants in the maize viviparous8 (vp8) gene show compromised anthocyanin accumulation in the germinal regions of the endosperm (Suzuki et al., 2008). This pattern is opposite to that observed in cr4 and dek1 mutants (Becraft et al., 1996, 2002; Becraft and Asuncion-Crabb, 2000), and vp8;cr4 double mutants showed anthocyanin defects on both sides of the kernel, supporting the idea that independent developmental domains exist in maize endosperm (Becraft and Asuncion-Crabb, 2000; Suzuki et al., 2008). The vp8 aleurone showed abnormal cellular morphology most pronounced in regions where anthocyanin defects were seen. ABA levels are decreased in vp8 mutants (Neill et al., 1986; Suzuki et al., 2008), which probably contributes to the precocious germination (vivipary), but the pleiotropic effects of this mutation, and the fact that other ABA-deficient mutants do not show similar aleurone phenotypes (personal observations), argue that an ABA deficiency is not causal in the aleurone defects.
Relationship of the aleurone to leaf epidermis
There appears to be a functional connection between the aleurone layer of the endosperm and the epidermis of leaves. This first became evident with the maize cr4 mutant, which disrupts aleurone specification and perturbs the leaf epidermis in various ways; cells are often irregularly shaped with poorly developed cuticles and the epidermis sometimes contains multiple cell layers (Becraft et al., 1996). Similarly, weak alleles of dek1 have a pronounced effect on the leaf epidermis of maize, rice, and Arabidopsis (Becraft et al., 2002; Johnson et al., 2005; Lid et al., 2005; Hibara et al., 2009). In addition to causing irregular epidermal cell morphology, dek1 mutants affect leaf axialization, causing normally adaxial cell types such as bulliform cells to form on the abaxial surface. The dil mutants of maize and elo2 of barley similarly cause cellular irregularities in both the aleurone and leaf epidermis (Lid et al., 2004; Lewis et al., 2009). The maize Extra cell layer (Xcl) mutant was identified because it causes multilayered leaf epidermis (in addition to abaxial bulliform-like cells) and was found also to produce a double aleurone layer (Kessler et al., 2002). With their obvious overlaps in genetic programmes and analogous positions in their respective organs, it is tempting to speculate that the aleurone and leaf epidermis might represent homologous tissues. Whatever the case, the two tissue types clearly have distinct functions and this is reflected in sets of mutants that are specific to each. For example, four other barley elo mutants (elo1, elo3, elo4, and elo5) cause leaf epidermal irregularities similar to elo2 but have no obvious effect on the aleurone (Lewis et al., 2009). Similarly, several aleurone mutants such as naked endosperm show normal phenotypes in the leaf epidermis (Becraft and Asuncion-Crabb, 2000; Becraft, 2007; unpublished observations).
Future prospects
Our current state of knowledge on cereal aleurone development is summarized in Fig. 4. Analysis of mutant phenotypes makes it clear that multiple levels of regulation and poorly understood complexities exist. Understanding these processes will provide valuable information about the fundamental mechanisms of how plant cells communicate to achieve higher order tissue organization. This information also has the potential to lead to improved nutritional quality of cereal grains. The molecular underpinnings of these processes can only be deciphered when additional components of this system are identified. Identification of genes for key mutants should be a top priority.
Model synthesizing the relationships of regulatory factors to the aleurone developmental process. Genes are represented in italics and their functions defined by loss-of-function phenotypes. Positive regulators of aleurone cell fate show loss of aleurone cell identity in the mutant, while negative regulators show the formation of extra aleurone layers. The phytohormones auxin and cytokinin have also been implicated in the control of aleurone cell fate. Differentiation entails the establishment of aleurone morphology, accumulation of aleurone-specific compounds such as anthocyanin (in particular maize genotypes), and the regulation of the maturation and germination processes. Well established relationships are depicted with solid lines and more hypothetical ones with dotted lines. See the text for further details.
The hormonal regulation of aleurone differentiation should also be further explored. Auxin clearly regulates cell fate in the embryo sac (Pagnussat et al., 2009), thus it is reasonable that it might also regulate post-fertilization development of embryo sac derivatives. Experiments should be designed to test specifically the function of auxin in aleurone development. It will also be informative to test the relationships of auxin transport and signalling to other genes that affect aleurone development and cell layer number. Also, while artificially increased cytokinin synthesis can apparently inhibit aleurone differentiation, it remains to be seen whether endogenous cytokinins have a role in regulating endosperm development.
Recent technical advances promise to provide exciting opportunities for conducting transient gene expression studies on endosperms grown in culture. Maize endosperms isolated from kernels 6d after pollination can be cultured on a high sucrose (15%) medium and will differentiate aleurone and starchy endosperm cells, but not a transfer layer (Gruis et al., 2006). These cultured endosperms appear to express a surprisingly normal developmental programme, albeit at an accelerated rate. They express marker transgenes for aleurone and starchy endosperm, as well as endogenous markers such as zein storage proteins (Gruis et al., 2006; Reyes et al., 2010). Most exciting is the demonstration that cultured endosperms of particular genotypes are amenable to Agrobacterium tumefaciens-mediated genetic transformation (Reyes et al., 2010). Using standard laboratory strains and binary vectors, a green fluorescent protein marker gene as well as epitope-tagged endosperm genes (under their native promoters) were delivered and expressed in cells of the cultured endosperms. This system holds tremendous promise for evaluating the functions of genes in developing endosperm as well as studying cell biological problems such as protein trafficking.
Finally, recent advances in genomics promise to greatly facilitate these efforts. Gene expression profiling technologies and systems-level analyses provide opportunities to understand the effects of gene mutations and how genes interact to regulate developmental programmes. Comparative genomics also offers exciting possibilities to study variation in endosperm patterning among related cereal species such as maize and barley.
We thank members of the Becraft lab who provided useful discussions. Research in the authors' lab was supported by USDA-NRI grant 2006-01163.
Comments