Timing of the maternal-to-zygotic transition during early seed development in maize

Abstract

In animals, early embryonic development is largely dependent on maternal transcripts synthesized during gametogenesis. Recent data in plants also suggest maternal control over early seed development, but the actual timing of zygotic genome activation is unclear. Here, we analyzed the timing of the maternal-to-zygotic transition during early Zea mays seed development. We show that for 16 genes expressed during early seed development, only maternally inherited alleles are detected during 3 d after fertilization in both the embryo and the endosperm. Microarray analyses of precocious embryonic development in apomictic hybrids between maize and its wild relative, Tripsacum, demonstrate that early embryo development occurs without significant quantitative changes to the transcript population in the ovule before fertilization. Precocious embryo development is also correlated with a higher proportion of polyadenylated mRNA in the ovules. Our data suggest that the maternal-to-zygotic transition occurs several days after fertilization. By contrast, novel transcription accompanies early endosperm development, indicating that different mechanisms are involved in the initiation of endosperm and embryo development.

Figure 1.

RT-PCR Analyses of Seed-Expressed Genes in Maize. (A) RT-PCR assay of genes expressed during early seed development in maize. Genomic DNA was extracted from a bulk of leaf, stem, and root samples; Zmfie1 and Zmfie2 provide controls for genomic DNA contamination. All PCR reactions consisted of 40 cycles of amplification. MW, molecular weight. (B) SSR-based strategy for the identification of allele-specific PCR conditions, illustrated here for the β-zein sequence. A single primer pair designed from the DNA flanking the repeats can be used to amplify both alleles. (C) Allele-specific expression of three endosperm-expressed genes at 3 and 7 DAP. In all three cases, a paternal CML72-specific allele is detected either after amplification of genomic DNA or amplification of cDNA 3 DAP from selfed plants. Only the maternal alleles are detected at 3 DAP in a cross between both lines. Both alleles are detected 7 DAP. Note the presence at 7 DAP of an extra transcript for the α-globulin not found in fertilized ovules from selfed plants 3 DAP. (D) Detection threshold for biallelic amplification of CML216 and CML72 alleles. Lane 1, cDNAs from ovules collected at 3 DAP from CML216; lane 2, cDNAs from ovules collected at 3 DAP from CML72; lanes 3 to 7, controlled dilutions of the same cDNAs mixed in the following proportions: 1:2, 1:10, 1:25, 1:50, and 1:100; lane 8, water negative control.

Figure 2.

Strategy for Microarray Analysis of Early Seed Development. In apomictic hybrid plants between maize and Tripsacum, as in Tripsacum apomicts, the developmental courses of the embryo and the endosperm are desynchronized. Unfertilized mature gametophytes in apomicts (0 DAP) contain a proembryo (PE) that is arrested after four to five divisions (A). Micrographs of two consecutive sections of the same ovule at 1 DAP show the simultaneous occurrence of a proembryo (top section) and unfused polar nuclei (PN; bottom section). At 3 to 4 DAP, the development of the endosperm (EN) after fertilization occurs in the absence of further embryo growth, which resumes only later during development, after endosperm cellularization. Therefore, profiling mRNA from mature sexual gametophytes against mRNA from mature apomictic gametophytes (B) captures transcriptional differences related to early embryogenesis. Similarly, transcription specific to early endosperm development is revealed by profiling unfertilized ovules of maize against fertilized ovules in apomicts collected at 3 DAP. E, endosperm; EC, egg cell; SYN, synergids.

Figure 3.

Microarray Analysis of Early Endosperm Development. (A) Frequency distribution of intensity ratio between samples at 3 and 0 DAP. Each gene is represented by the mean of the adjusted intensity values of all replicates. (B) Scatter plot of spatially adjusted fluorescence intensities between RNA extracted 3 DAP (y axis) and RNA from nonfertilized ovules from the apomictic genotype 38C (x axis). The spots indicate the localization of genes with significantly different expression at P < 0.001. Each gene is represented by the mean of the adjusted intensity values of all replicates. The group of spots indicated by the arrow indicates those genes where either x value was not significantly different from 0 but y value was not equal to 0 (and thus likely repressed after fertilization), or y value was not significantly different from 0, but x value was not equal to 0 (and thus likely induced after fertilization).

Figure 4.

Validation of Microarray Data Using RT-PCR. Somatic tissues are represented by a bulk containing RNA from leaves, stem, husk, and roots. The examples illustrate three situations of downregulation (AI833477) or upregulation (AI820386 and AI714918) of genes after fertilization. AI820386 illustrates a case of more complex regulation. The optimum number of PCR cycles was chosen for each gene so that the signal would fall within the log phase of PCR amplification.

Figure 5.

Relative Amounts of Polyadenylated mRNA in Ovules of Unfertilized Apomictic and Sexual Plants. (A) Dilution series of total RNA extracted from ovules of apomictic genotype 38C shows that dig-dUTP incorporation is linearly proportional to the amount of total RNA, with both random-priming and oligo(dT) reverse transcription. The actual dot blots of the dilution series [RP, random primers; OdT, oligo(dT) primers] are shown. The graphical representation of the same information shows the linearity of the signal intensity. The numbers represent arbitrary intensity units (a.u.). (B) Various independent reverse transcription reactions (replicates) were performed for both materials to estimate relative polyadenylated RNA; the graph summarizes the results for eight repetitions each. The diamonds indicate the mean value of the signal units; the bars indicate the minimum–maximum range. Note that the signal values fall within the linear part of the figure in (A), suggesting that the differences are not due to a possible saturation of the reverse transcription reaction.