Increased maternal genome dosage bypasses the requirement of the FIS polycomb repressive complex 2 in Arabidopsis seed development

Abstract

Seed development in flowering plants is initiated after a double fertilization event with two sperm cells fertilizing two female gametes, the egg cell and the central cell, leading to the formation of embryo and endosperm, respectively. In most species the endosperm is a polyploid tissue inheriting two maternal genomes and one paternal genome. As a consequence of this particular genomic configuration the endosperm is a dosage sensitive tissue, and changes in the ratio of maternal to paternal contributions strongly impact on endosperm development. The fertilization independent seed (FIS) Polycomb Repressive Complex 2 (PRC2) is essential for endosperm development; however, the underlying forces that led to the evolution of the FIS-PRC2 remained unknown. Here, we show that the functional requirement of the FIS-PRC2 can be bypassed by increasing the ratio of maternal to paternal genomes in the endosperm, suggesting that the main functional requirement of the FIS-PRC2 is to balance parental genome contributions and to reduce genetic conflict. We furthermore reveal that the AGAMOUS LIKE (AGL) gene AGL62 acts as a dosage-sensitive seed size regulator and that reduced expression of AGL62 might be responsible for reduced size of seeds with increased maternal genome dosage.

Figure 1. Seeds Derived from Crosses osd1×2n Mimic the Effect of Maternal Excess Interploidy Hybridizations.

(A) Seeds from crosses of Col 2n×2n (upper panel) and osd1×2n (diploid and triploid seeds, middle and lower panel, respectively. Ploidy refers to ploidy of the embryo). Bar = 0.5 mm. (B) Average seed size (upper panel) and seed weight (lower panel) of different crosses. Numbers on top of bars correspond to number of analyzed seeds. For seed weight the average of 100 seeds was calculated in triplicates. Error bars indicate SD. The reference line corresponds to wild-type seed size and weight. (C) Number of endosperm nuclei in Col 2n×2n and osd1×2n crosses. n = 5. Error bars indicate SD. (D) Sections of seeds from Col 2n×2n and osd1×2n crosses. Bar = 100 µm.

Figure 2. FIS-PRC2 Target Genes Are Deregulated in Response to Maternal Excess Interploidy Hybridizations.

Venn diagrams showing overlap of genes being deregulated in seeds derived from osd1×2n crosses with genes deregulated in seeds of 4n×2n crosses (upper panel); overlap of deregulated genes in seeds derived from osd1×2n crosses with H3K27me3 target genes in the endosperm (middle panel), and genes that are upregulated in fis2 mutant seeds (lower panel). In the upper panel only deregulated genes that are present on the ATH1 array have been used to calculate the overlap.

Figure 3. Expression Level of FIS2 but Not of MEA Is Increased in osd1×2n Crosses.

Quantitative RT-PCR analysis of MEA and FIS2 expression in seeds derived from osd1×2n crosses at 1–4 days after pollination (DAP). Error bars indicate s.e.m.

Figure 4. Maternal Excess Triploid mea and fis2 Seeds Are Strongly Enlarged.

(A) Distribution of seed sizes from different crosses. A minimum of 250 seeds was analyzed for each cross. Inset shows percentage of seeds larger than 0.14 mm². (B) Percentages of normal, enlarged and collapsed seeds from different crosses. Numbers in parenthesis correspond to numbers of analyzed seeds.

Figure 5. Endosperm Cellularization Is Restored in Triploid mea Seeds.

Sections of seeds from 2n×2n (left panels), mea/MEA×2n (middle panels) and mea/MEA; osd1/osd1×2n (right panels) crosses at 6–12 days after pollination (DAP). Bar = 100 µm. Images of mature seeds are shown in bottom panels. Bar = 0.5 mm.

Figure 6. Expression of AGL MADS Box Genes Is Decreased in osd1×2n Crosses.

Quantitative RT-PCR analysis of AGL62, PHE1, AGL90, AGL36, AGL40 and AGL28. Error bars indicate s.e.m.

Figure 7. Expression of AGL MADS Box Genes Is Normalized in Triploid mea and fis2 Seeds.

Quantitative RT-PCR analysis of AGL62, PHE1, AGL90, AGL40, AGL28 and AGL36 at 8 DAP in mea and fis2 seeds (upper and lower panel, respectively). Enlarged seeds with a fis phenotype were selected as 3n mea and fis2 seeds from mea/MEA; osd1/osd1×2n and fis2/FIS2; osd1/osd1×2n crosses, respectively. Diploid mea seeds are a mixture of wild-type and mea seeds. Error bars indicate s.e.m.

Figure 8. Maternal Loss of NRPD1a Does Not Restore Wild-Type Levels of AGL Expression and Wild-Type Seed Size.

(A) Quantitative RT-PCR analysis of PHE1, AGL62, AGL90, AGL40, AGL28 and AGL36 in seeds derived from crosses 2n×2n, nrpd1a/nrpd1a×2n, osd1/osd1×2n, and nrpd1a/nrpd1a; osd1/osd1×2n at 3 DAP. Error bars indicate s.e.m. (B) Distribution of seed sizes from different crosses. A minimum of 250 seeds was analyzed for each cross.

Figure 9. Size of Triploid Seeds Is Dependent on the Dosage of AGL62.

(A) Size of seeds derived from crosses agl62/AGL62; osd1/osd1×wild type and diploid seeds from crosses agl62/AGL62×wild type. Average seed size of different genotypes is shown. Wild-type (+) and mutant (−) allele frequencies of AGL62 in the endosperm are shown below the x axis. Numbers on top of bars refer to numbers of analyzed seeds. Error bars indicate s.e.m. (B) Images of triploid seeds from the three categories shown in (A). Bar = 0.1 mm.