Trends in Plant Science 9(10): 507-514 (Oct 2004)
More than a yolk: the short life and complex times of the plant endosperm
Liliana M. Costa, José F. Gutièrrez-Marcos and Hugh G. Dickinson
Department of Plant Sciences, University of Oxford, South Parks Road, Oxford, UK OX1 3RB

In most angiosperms, the endosperm is limited to a short phase of the plant life-cycle where it forms part of the seed and plays a major role in its development. Our understanding of this terminally differentiated organ has accelerated in recent years, with discoveries in the fields of phylogeny and developmental genetics shedding light on its evolutionary origin, function and development. Here we explore various conserved and unique features of plant endosperms, including the establishment of distinct functional domains that regulate nutrient transfer from the maternal parent to the developing embryo. We also review data from Arabidopsis and maize that confirm the existence of complex genetic mechanisms operating during endosperm development. Importantly, these findings confirm that, in addition to nourishing the embryo, the endosperm fulfils several other key functions including that of fertilization ‘sensor’; detecting and aborting the fertilization products of incompatible or wide hybridizations between related species.

More than a yolk: endosperm function in embryogenesis, hybridization and speciation
In the majority of angiosperms, including most apomictic plants, the endosperm is required for successful embryogenesis, particularly during early developmental stages. In many cases, embryo development proceeds more rapidly once the endosperm has enlarged and differentiated. Thus, with reason, the principal function ascribed to the endosperm has been that of nutritive support during embryogenesis and, in the monocots, seedling germination. Other than playing a major role in providing resources for the developing embryo, there is increasing evidence that identifies the endosperm as playing a more fundamental part in reproductive development. One way in which this is achieved is by acting as an intermediary between the embryo and the surrounding maternal sporophytic tissue. Several mutations disrupting communication between the maternal integument and the endosperm have been shown consequently to affect development of the seed or surrounding maternal sporophyte [64–68]. Another way is through the possession of unique genetic systems that render the endosperm ‘sensitive’ to genomic balance. Thus, although embryogenesis will often commence following crosses between distant relatives or lines of differing ploidy, endosperm development will abort. Abortion of the endosperm has been mainly attributed to alterations in gene dosage and to an imbalance between the maternal and paternal populations of imprinted alleles [37,53,69–71]. This ‘genomic equilibrium’ is usually fixed, but in some species can be altered to promote the success of interspecific hybrids [53,72]. Significantly, the stringent requirement for correct parental genomic balance in the endosperm can also be overcome through mechanisms by which epigenetic patterns are established, such as DNA methylation [73]. Therefore, if we consider polyploidy and hybridization as major forces in the evolution of flowering plants and stimuli for the appearance of invasive new species [74,75], then epigenetic and dosage-dependent mechanisms operating in the endosperm, which regulate these processes, must have played, and continue to play, a crucial role in angiosperm evolution.

Conspectus
In recent years, a combination of developmental genetics, molecular analyses and evolutionary studies has elevated the flowering plant endosperm from a straightforward, terminally differentiated nurse tissue to an intricate and largely maternally regulated organ that exerts a profound influence on reproductive success. To further unravel the complexities of the endosperm, we must focus on the identification of genes that are necessary for its development. Equally important, attention must be given to identifying more female gametophytic genes that play a crucial role in regulating endosperm development. Success to date has been limited, not least because of the lethality of such mutations leading to poor or no transmission to the next (diploid sporophytic) generation. However, the development of more effective genetic screens, such as enhancer or gene trap lines expressing visible markers and novel mutagenesis strategies [76], should overcome this present constraint.

Another outstanding issue that needs to be tackled  concerns the molecular mechanisms involved in establishing and maintaining imprinting in the endosperm. Further efforts should be employed to decipher the specific changes in DNA methylation and chromatin configuration occurring at the parental alleles of imprinted genes. Intriguingly, non-coding RNAs are associated with the imprinted expression pattern of some mammalian genes, and hence future research should be directed at attempting to identify non-coding RNAs that might also regulate imprinting in the plant endosperm.

Finally, from the findings reviewed here, it is clear that much important information has been gleaned from Arabidopsis and maize compared to other, less-studied species. However, it is anticipated that comparative studies across the range of angiosperms will hold the key to gaining exciting and rewarding insights into the molecular and evolutionary basis of endosperm development.

Hetero-Fertilization / Endosperm Failure