Comparative transcriptomics of Arabidopsis sperm cells

Abstract

In flowering plants, the two sperm cells are embedded within the cytoplasm of the growing pollen tube and as such are passively transported to the embryo sac, wherein double fertilization occurs upon their release. Understanding the mechanisms and conditions by which male gametes mature and take part in fertilization are crucial goals in the study of plant reproduction. Studies of gene expression in male gametes of maize (Zea mays) and Plumbago and in lily (Lilium longiflorum) generative cells already showed that the previously held view of transcriptionally inert male gametes was not true, but genome-wide studies were lacking. Analyses in the model plant Arabidopsis (Arabidopsis thaliana) were hindered, because no method to isolate sperm cells was available. Here, we used fluorescence-activated cell sorting to isolate sperm cells from Arabidopsis, allowing GeneChip analysis of their transcriptome at a genome-wide level. Comparative analysis of the sperm cell transcriptome with those of representative sporophytic tissues and of pollen showed that sperm has a distinct and diverse transcriptional profile. Functional classifications of genes with enriched expression in sperm cells showed that DNA repair, ubiquitin-mediated proteolysis, and cell cycle progression are overrepresented Gene Ontology categories. Moreover, analysis of the small RNA and DNA methylation pathways suggests that distinct mechanisms might be involved in regulating the epigenetic state of the paternal genome. We identified numerous candidate genes whose involvement in sperm cell development and fertilization can now be directly tested in Arabidopsis. These results provide a roadmap to decipher the role of sperm-expressed proteins.

Figure 1.

Fluorescence-activated sperm cell sorting based on cell size (FSC), intracellular complexity (SSC), GFP signal, and presence of intracellular DNA, via DRAQ5 staining. Low granulosity (low SSC) and GFP positive signals were used to identify the sperm cell population (R1) from the total population. To guarantee purity, a low FSC signal (small particles; R2) within the GFP/DRAQ5 double positive population (R3) were used to exclude other small particles. A displays total population, B shows cells within region R2, and C shows cells within region R3. [See online article for color version of this figure.]

Figure 2.

Visualization of FACS-purified sperm cells. Wide-field fluorescence microscopy was used to visualize Arabidopsis sperm cells expressing AtGEX2∷eGFP (Engel et al., 2005) before (B) and after FACS purification (D). Differential interference microscopy (DIC) microscopy confirmed that the debris in the filtrate before FACS (A) was removed after sorting (C). For DIC imaging of FACS-purified sperm cells (C), we captured and merged several images along the optical axis. A higher magnification of a sorted sperm cell shows GFP fluorescence (E), cell-shape integrity by DIC microscopy (F), and cell viability using fluorescein diacetate staining (G). The bars represent 5 microns and the arrowheads are pointing to sperm cells.

Figure 3.

Venn diagram, depicting the overlap between genes whose expression was called present in sperm cells (5,829), pollen (7,177), seedlings (14,464), and additionally the Arabidopsis genes that are represented on the ATH1 array (1,239) and that match those of maize sperm cells ESTs (Engel et al., 2003). Intersection of common genes between Arabidopsis sperm cells and pollen, maize sperm cells, and Arabidopsis seedlings is 3,813, 594, and 4,757, respectively. [See online article for color version of this figure.]

Figure 4.

RT-PCR analysis. A and B, Gel figures presenting confirmatory RT-PCR analysis for a gene highly expressed in pollen but not detected in sperm cells in our microarray analysis, encoding a carbonic anhydrase family protein (At5g44340; A); and genes involved in the sRNAs and RdDM pathways, which are not detectable in sperm cells, DCL3 (At3g43920), CMT3 (At1g69770), and NRPD2a (At3g23780; B). C, RT-PCR on total RNA from two samples of FACS-isolated sperm cells (SC#1, SC#2), pollen, and leaf, showing enrichment of MGH3 (At1g19890) transcripts and absence of VEX1 (At5g62850) transcripts in both sperm cell replicates. D, Expression of several genes presented in Table II was tested by RT-PCR in sperm cells, pollen, seedling, ovule, and silique cDNA samples. TUB4 (At5g04180) was used as positive control. SC, Sperm cells; G, genomic DNA.

Figure 5.

Hierarchical clustering and PCA of tissue-dependent gene expression patterns. Transcriptome data of seedlings, pollen, and sperm cells were compared with those of leaf, flower, silique, ovule, and unpollinated pistil from previous studies (Pina et al., 2005; L.C. Boavida, F. Borges, J.D. Becker, and J.A. Feijó, unpublished data). A, Hierarchical clustering dendrogram, using Pearson's dissimilarity to calculate row dissimilarity and Ward's method for row clustering. B, PCA. [See online article for color version of this figure.]