• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Dec 13, 2005; 102(50): 17887–17888.
Published online Dec 5, 2005. doi:  10.1073/pnas.0509021102
PMCID: PMC1312409

Control of seed size in plants

Reproductive success in seed plants depends on a healthy seed set. The viability of the embryo is enhanced if a seed contains substantial reserves of starch and protein to nourish the seedling when it germinates months or years later in uncertain conditions. Increased reserves will generally result in an increased seed size, but large seeds are less efficiently dispersed, unless there is human or other intervention. Since the beginning of agriculture, food grains have been subjected to selection and breeding for size as well as for other qualities, and most of the grains consumed today have seeds far larger than their wild relatives. Although grain size has been much analyzed and used by plant breeders over the past century, it is only in the past decade that we have at last begun to identify the molecular regulators of seed size in plants, mainly from studies in the model plant Arabidopsis (reviewed in ref. 1). The article by Luo et al. (2) in a recent issue of PNAS marks another step toward understanding the nature of these controls.

The life cycle of plants involves an alternation of generations between the haploid gametophyte and the diploid sporophyte. In angiosperms (flowering plants), seed development begins with double fertilization. Pollen grains (male gametophytes) carry two haploid sperm cells, which fertilize the egg cell and the central cell of the haploid embryo sac (female gametophyte) contained within the maternal tissues of the ovule. This event results in the formation of the diploid embryo and the triploid endosperm, respectively, the latter arising from the central cell that contains two identical haploid sets of chromosomes. Seed development is marked by the rapid growth of the endosperm and the embryo, until seed maturation, which is accompanied by desiccation. Simultaneously, the maternal ovule also undergoes regulated growth to accommodate the growing embryo and endosperm, and the integuments of the ovule ultimately constitute the coat of the mature seed (Fig. 1).

Fig. 1.
Schematic sketch of seed development in the model plant Arabidopsis. (Left) The unfertilized ovule contains the haploid egg cell and the double haploid central cell, enclosed within the maternal tissue of the integuments. (Center) Seed development follows ...

The endosperm grows much more rapidly than the embryo, growing initially through nuclear divisions as a syncitium for several mitotic cycles and subsequently cellularizing followed by decreased rate of growth. The growth of the seed is coupled with the growth of the endosperm, with the major increase in seed volume occurring in concordance with the rapid growth of the endosperm. In monocots and some dicots, the endosperm constitutes the major contribution to the volume of the mature seed. In Arabidopsis and many other dicots, the endosperm is eventually consumed, being replaced by the growing embryo, which then constitutes most of the mature seed. However, in all cases the growth of the seed is primarily associated with the initial growth of the endosperm, and not with the later growth of the embryo. Thus, the size of the seed is the result of three different growth programs: those of the diploid embryo, the triploid endosperm, and the diploid maternal ovule. The control and coordination of these growth programs are under genetic regulation as described below.

Differences in the contributions of maternal and paternal genomes to seed size are evident from experiments that alter gene dosage through crosses between diploid and tetraploid plants (3, 4). Such crosses show that when the paternal genome is in excess, seed growth is promoted, and conversely, excess of the maternal genome results in smaller seeds. Models to account for these “parent-of-origin” effects have been proposed, based either on theories of “genome conflict” (3) or on differential dosage (4). These parent-of-origin effects on seed development appear to act primarily through regulation of endosperm growth (1). If seed growth involves coordinated growth control of ovule and seed, we may predict that mutants causing reduced seed size would result in reduction of integument growth. It has long been known that embryo lethal mutants can result in smaller seeds with reduced seed coats, but the arrested growth of the integuments could be a general response to embryo arrest and death, rather than a specific control of integument growth rate. Recently, a class of reduced seed size mutants called haiku mutants has been characterized (5, 6), in which embryo and endosperm growth are reduced but not arrested, and the seeds remain viable. Plants grown from homozygous haiku mutant seeds are normal, indicating that these genes might specifically function in seed growth. Detailed analysis of the haiku mutants showed that they appear to be primarily restricting endosperm growth, and that integument cell elongation is reduced as a result. Interestingly, in the same study it was found that mutation of the TTG2 (Transparent Testa Glabrous 2) gene, which results in reduction of integument cell elongation, also reduces endosperm growth (6). The TTG2 gene encodes a WRKY transcription factor, previously shown to be required for pigmentation of the seed coat as well as formation of leaf trichomes, and the reduction of integument cell elongation in ttg2 mutants might be an indirect effect of the reduction in pigment synthesis in the integument cells. Another example of maternal control of seed size is provided by the floral homeotic gene AP2 (APE-TELA2). Mutations in AP2 result in an increase in seed size, and this effect is independent of the genotype of the embryos, consistent with a maternal function for this gene (7, 8). AP2 encodes a transcription factor, and the mechanism by which it exerts its effect on seed size is still unclear, but it has been proposed to act through alteration of sugar metabolism (7). Integument growth and endosperm growth are likely regulated through independent pathways, because the seed size of ttg2 haiku double mutants is smaller than those of either mutant alone (6).

Now, Luo et al. (2) describe the molecular characterization of IKU2 (HAIKU2) and a second gene called MINI3 (MINISEED3) identified through an independent screen for small seed mutants (2). Mutants in either gene result in a reduced seed size, which depends on the genotype of the embryo/endosperm and not on that of the maternal ovule. A map-based approach was used to clone the genes. MINI3 was found to encode a transcription factor of the WRKY family. MINI3 is not expressed in the unfertilized ovule but is expressed after fertilization in both the endosperm and the embryo. IKU2 was found to encode an LRR (leucine-rich repeat) receptor kinase, one of a large gene family of plants known to be involved in many different cell signaling processes. Importantly, IKU2 expression was detected in the endosperm but not in the embryo or elsewhere in the plant. This finding provides direct molecular evidence for a regulator of seed size that acts solely through control of endosperm proliferation.

The study of Luo et al. (2) also provides, for the first time, a framework to assemble a genetic pathway for seed size control. Because IKU2 expression is down-regulated in mini3 mutants but not vice versa, MINI3 must act upstream of IKU2. Although it might seem contrary to expectation to have the transcription factor acting upstream of the kinase, the authors point out that there is a precedent in the transcriptional regulation of the RLK4 (Receptor-Like Kinase 4) gene in Arabidopsis by the WRKY18 transcription factor (9). This regulatory hierarchy is also consistent with the broader expression pattern of MINI3 vs. IKU2. MINI3 could have different sets of targets in the endosperm and embryo, with IKU2 being one of the targets in the endosperm. There are indications that MINI3 might also be partially autoregulatory in a negative-feedback loop, based on decreasing expression with increased gene doseage.

IKU1 appears to act upstream of both MINI3 and IKU2, because neither gene is expressed in an iku1 mutant. It had been previously shown that the reduced endosperm growth in iku mutants is correlated with premature cellularization of the endosperm. The regulation of endosperm size might be achieved through control of cellularization by the IKU and MINI3 genes, or it might be that they directly control endosperm nuclear proliferation and that the apparent premature cellularization is a secondary consequence of a slower rate of division in the mutants. The latter mechanism seems more plausible because in the maternal-acting ttg2 mutants, slower growth of the integuments, which results in slower divisions of endosperm nuclei, also results in early cellularization of the endosperm.

The study provides a framework to assemble a genetic pathway for seed size control.

The larger question of how the endosperm and the integuments communicate to coordinate growth remains open. Mutation of the Miniature1 gene of maize, which encodes a cell wall invertase 2, also results in a small seed phenotype (10). The protein appears to be localized to the proximal end of the endosperm near the point of attachment to the placenta, in a region of the seed that develops abnormally in miniature1 mutants. This region of the seed is adjacent to the conduit for nutrients from maternal tissue and may also be involved in communication with the integuments. The iku and mini3 mutant seeds do not appear to have similar abnormalities, and the study of Luo et al. (2) does not address how the endosperm signals to the integuments. The findings from this study might, however, shed some light on the question of how the integuments signal to the endosperm. The identification of IKU2 as an endospermspecific LRR receptor kinase raises the attractive hypothesis that signaling from the integuments to the endosperm occurs through the IKU2 kinase. However, the observation that the ttg2 iku2 double mutant is more severe than ttg2 or iku2 alone (6) is not easily explained by this model, because we might expect that mutation of the receptor would render the seed insensitive to respond to any reduction of integument growth by ttg2. One possible explanation might be a partial redundancy of IKU2 with other unidentified LRR kinases, which constitute a very large family in Arabidopsis. This study also raises the question of what the ligand might be for the IKU2 receptor kinase, whether it is synthesized by the integuments, and, if so, how it is translocated to the endosperm. Determining the localization of IKU2 protein within the developing seed is likely to provide significant clues toward uncovering this presumptive signaling network. The work of Luo et al. (2) and other recent advances in this field are important because they indicate that at last we have a foundation for asking the types of specific questions that will undoubtedly yield interesting insights in the future.


Author contributions: V.S. wrote the paper.

Conflict of interest statement: No conflicts declared.

See companion article on page 17531 in issue 48 of volume 102.


1. Chaudhury, A. M., Koltunow, A., Payne, T., Luo, M., Tucker, M. R., Ming, L., Dennis, E. S. & Peacock, W. J. (2001) Annu. Rev. Cell Dev. Biol. 17, 677–699. [PubMed]
2. Luo, M., Dennis, E. S., Berger, F., Peacock, W. J. & Chaudhury, A. (2005) Proc. Natl. Acad. Sci. USA 102, 17531–17536. [PMC free article] [PubMed]
3. Spielman, M., Vinkenoog, R., Dickinson, H. G. & Scott, R. J. (2001) Trends Genet. 17, 705–711. [PubMed]
4. Dilkes, B. P. & Comai, L. (2004) Plant Cell 16, 3174–3180. [PMC free article] [PubMed]
5. Garcia, D., Saingery, V., Chambrier, P., Mayer, U., Jurgens, G. & Berger, F. (2003) Plant Physiol. 131, 1661–1670. [PMC free article] [PubMed]
6. Garcia, D., Gerald, J. N. F. & Berger, F. (2005) Plant Cell 17, 52–60. [PMC free article] [PubMed]
7. Ohto, M. A., Fischer, R. L., Goldberg, R. B., Nakamura, K. & Harada, J. J. Proc. Natl. Acad. Sci. USA 102, 3117–3122. [PMC free article] [PubMed]
8. Jofuku, K. D., Omidyar, P. K., Gee, Z. & Okamuro, J. K. (2005) Proc. Natl. Acad. Sci. USA 102, 3123–3128. [PMC free article] [PubMed]
9. Du, L. & Chen, Z. (2000) Plant J. 24, 837–847. [PubMed]
10. Cheng, W., Taliercio, E. W. & Chourey, P. S. (1996) Plant Cell 8, 971–983. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

  • Control of seed size in plants
    Control of seed size in plants
    Proceedings of the National Academy of Sciences of the United States of America. Dec 13, 2005; 102(50)17887

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...