Transcription profiling of fertilization and early seed development events in a solanaceous species using a 7.7 K cDNA microarray from Solanum chacoense ovules

Abstract

Background: To provide a broad analysis of gene expression changes in developing embryos from a solanaceous species, we produced amplicon-derived microarrays with 7741 ESTs isolated from Solanum chacoense ovules bearing embryos from all developmental stages. Our aims were to: 1) identify genes expressed in a tissue-specific and temporal-specific manner; 2) define clusters of genes showing similar patterns of spatial and temporal expression; and 3) identify stage-specific or transition-specific candidate genes for further functional genomic analyses.

Results: We analyzed gene expression during S. chacoense embryogenesis in a series of experiments with probes derived from ovules isolated before and after fertilization (from 0 to 22 days after pollination), and from leaves, anthers, and styles. From the 6374 unigenes present in our array, 1024 genes were differentially expressed (>or= +/- 2 fold change, p value <or= 0.01) in fertilized ovules compared to unfertilized ovules and only limited expression overlap was observed between these genes and the genes expressed in the other tissues tested, with the vast majority of the fertilization-regulated genes specifically or predominantly expressed in ovules (955 genes). During embryogenesis three major expression profiles corresponding to early, middle and late stages of embryo development were identified. From the early and middle stages, a large number of genes corresponding to cell cycle, DNA processing, signal transduction, and transcriptional regulation were found. Defense and stress response-related genes were found in all stages of embryo development. Protein biosynthesis genes, genes coding for ribosomal proteins and other components of the translation machinery were highly expressed in embryos during the early stage. Genes for protein degradation were overrepresented later in the middle and late stages of embryo development. As expected, storage protein transcripts accumulated predominantly in the late stage of embryo development.

Conclusion: Our analysis provides the first study in a solanaceous species of the transcriptional program that takes place during the early phases of plant reproductive development, including all embryogenesis steps during a comprehensive time-course. Our comparative expression profiling strategy between fertilized and unfertilized ovules identified a subset of genes specifically or predominantly expressed in ovules while a closer analysis between each consecutive time point allowed the identification of a subset of stage-specific and transition-specific genes.

Figure 1

Gene expression profiling of unfertilized and fertilized ovule samples. Scatter plots comparing the average fluorescence intensities in microarray experiments measuring the change in transcript abundance between A, pooled control and individual control; or B, 16 DAP pollinated ovules versus unfertilized ovules. The individual control samples were different from the pooled one that is used as the universal control in all our experiments. Black transversal lines show the ±2 fold change.

Figure 2

Transcriptional changes in S. chacoense ovules across an embryo developmental time course. A, Each line represents a probe that was spotted on the microarray while each time points of embryo developmental stage are represented in X-axis. The Y-axis shows the log of the normalized fluorescence ratios. Indicated in red are the genes expressed to levels higher than the median value. Genes indicated in blue are expressed to lower levels than the median value. The corresponding embryo developmental stages for each time point was ascertain by ovule clearings and observed by DIC microscopy and represented with a schematic figure. B, Cluster analysis using unsupervised hierarchical clustering of 1024 genes that exhibit a statistically significant change between all samples (P < 0.01). The analysis was performed using condition tree clustering on all samples. Each row represents a different gene, and each column displays gene expressions at each time point (0 to 22 DAP). Each experimental data point is colored according to the change in fluorescence ratio at that time point: data values displayed as red and blue represent increased and reduced expression, respectively while data values in yellow are not differentially expressed when compared to unfertilized ovules.

Figure 3

Tissue-specific comparison between ovule, style, anther and leaf tissues. A, Cluster analysis of the transcriptional profiles obtained from leaves, styles, and anthers. The clustering was performed using condition tree clustering on all samples. Each row represents a different gene, and each column displays gene expressions at different tissues (UO, style, leaf and anther). Data values displayed as red and blue represent elevated and reduced expression, respectively. A total of 601 genes with a statistically significant change in transcript abundance of ≥2-fold in at least one tissue compared to unfertilized ovules (p value < 0.01) are found. B, Venn diagram showing overlap between lists of of genes differentially expressed in anther, style and leaf tissues. C, Principal Components Analysis (PCA) of various tissue types (A: Anthers; S: Styles; Le: Leaves; C: unfertilized ovules. E, M, L: early, middle and late stages of embryo development, respectively) based on the the expression profile of 1376 genes including 558 that are differentienlly expressed in at least one tissue from styles, leaves and anthers and 1024 transcripts that are regulated in ovule which resulted in a clear differentiation distributed between 7 classifications.

Figure 4

Venn diagram showing overlap between gene lists that show significant changes in transcript abundance during the three major embryo development stages (early, middle and late stages).

Figure 5

Functional annotations of the 955 genes (passing the significance filter: P < 0.01 and 2 fold cut-off) during the three major embryo developmental stages. The color keys show the designated gene class annotations.

Figure 6

Graphical representation of the number of stage-specific genes identified at each time point of embryo development (Volcano plot using Student's t-test analysis taking p value < 0.01 and 2 fold cut off). To find specific genes in each time, a Venn diagram was used between each time point and the total of genes remaining in the other stages. Numbers in the inner circle are genes shared by at least another time point. Numbers in the outer circle are genes considered stage-specific by the Volcano analysis.

Figure 7

Distribution of 558 differentially expressed (by 2 fold-change, P < 0.01) genes in anthers styles and leaves in various biological process categories. The color key shows the designated gene class annotations.

Figure 8

Quantitative real-time PCR (qRT-PCR) of selected genes representing the early, middle and late stage of embryo development. qRT-PCR was performed with two independent samples at each point.

Figure 9

K-means clustering using Pearson Correlation (GeneSpring version 7.2) during an embryo development time course. A, Models of dynamic gene expression patterns representing all the nine different patterns found based on the three point time (early, middle and late). B, Clustering of specific differentially expressed genes in the ovules based on 15 clusters representing all the stages of embryonic development from 0 to 22 DAP. The black discontinues line represent the average of gene profile.