MOLECULAR CONTROL OF REPRODUCTIVE ORGAN DEVELOPMENT IN RICE (ORYZA SATIVA L.)

Several articles published in literature show that the genes belonging to the MADS-box transcription factors family play important roles during flower development. In particular the genes belonging to the AGAMOUS subfamily, are involved in the development of stamens, carpels, and ovules but also have a function in floral meristem determinacy. Despite the extensive knowledge that has become available about these MADS domain transcription factors in the model plant Arabidopsis thaliana, little was known about these genes in rice (Oryza sativa L.), a model species for monocots. In rice, the AGAMOUS subfamily, recently studied in detail, is composed of the genes OsMADS3 and OsMADS58, phylogenetically placed in the AG-lineage, and OsMADS13 and OsMADS21, belonging to the STK-lineage. Therefore, we decided to investigate the functional conservation of these genes between Arabidopsis thaliana and rice, two species that are quite distant from an evolutionary point of view. First, we performed in situ hybridization experiments and found that OsMADS3 and OsMADS58 displayed an expression pattern very similar to AGAMOUS. Then, we started to analyse single and higher order knock-out and knock-down mutants of osmads3 and osmads58, and found different phenotypic alteration affecting reproductive organs. In the osmads3 mutant the formation of stamens and carpel is partially affected, whereas in the osmads58 single mutant we didn’t find any visible alteration in the morphology of the flower. When we investigated the osmads3 osmads58 double mutant, we observed a severer phenotype than the osmads3 single mutants indicating a certain level of redundancy between these genes. The identity of the third and fourth whorls was completely lost, and stamens and carpels were replaced, respectively, by lodicule and ectopic palea-like structures. In addition, in the osmads3 osmads58 double mutant the floral meristem was indeterminate. Recently, it was shown that the rice STK-lineage gene OsMADS13 also plays a role in meristem determinacy (Dreni et al., 2007). To test the functionally redundancy of AGAMOUS subfamily genes in floral meristem determinacy, we crossed osmads13 with osmads3 and osmads58 to obtain the double mutants. In both cases we observed a complete loss of floral meristem determinacy. We also generated and analysed the triple mutant osmads3 omads13 osmads58, in which we found an increased development of the palea-like organ which develops in place of the carpel. OsMADS13 is specifically expressed in the ovary during all developmental stages, and the single mutant is female sterile because the ovules are converted into carpelloid structures, moreover as mentioned above the floral meristem is partially indeterminate (Dreni et al., 2007). From these results obtained in this first part of my PhD project we concluded that OsMADS3, OsMADS58 and OsMADS13 redundantly control floral meristem determinacy, and that OsMADS3 and OsMADS58 are essential for the development of stamens and carpel in rice. In conclusion we confirmed our hypothesis that the AGAMOUS-like genes are highly conserved between monocots and dicots plants (Dreni et al., 2011). In the second part of my PhD project, we focused our attention on the rice’s ovule identity gene OsMADS13 and we tried to identify its direct target genes. In order to find genes that are mis-regulated in the female sterile osmads13 mutant, we compared the transcriptomic changes between WT and osmads13 inflorescences by Next Generation Sequencing (RNA-seq). The output of the analysis conducted by the bioinformatics team of Dott. David Horner, was a list of 475 differentially expressed genes with a “False Discovery Rate” (FDR) cut-off value lower than 0.05. Furthermore, we looked to the Gene Ontology to select the ones with an annotated function, we selected genes that are mainly expressed during reproductive stages and also looked for putative Arabidopsis orthologs involved in flower development. Since MADS domain proteins recognize and bind CArG boxes [CC(A/T)6GG], we further reduced our list selecting genes with putative MADS-domain binding sequences in their regulatory regions. From this selection we obtained 45 interesting genes that we validate using high-throughput RT-qPCR approach (Fluidigm technology). Subsequently, using Laser Micro-Dissection (LMD), we collected cells of the superficial layer of the floral meristem and ovule primordia from wild-type and mutant flowers to further analyse the expression levels of selected genes in the OsMADS13 expression domain. This is especially important since in the osmads13 mutant ovules are homeotically converted into carpels. Profiling expression differences between wild-type and mutant inflorescences gives the risk that we are comparing ovules with carpels. Therefore, to further reduce our gene list of candidate target genes involved in early phases of ovule development we used LMD to dissect early stage ovule primordia. Using this dissected material a high-throughput RT-qPCR approach was used (publication in preparation). The aim of the last part of my PhD project was to produce a transgenic line carrying the construct pOsMADS13::OsMADS13(genomic)::GFP in order to use commercial α-GFP for ChIP assay, to confirm the direct binding of OsMADS13 to putative targets. For this reason, we transformed rice calli carrying the osmads13 mutation in homozygous state, and at the moment we are analysing the transformed plants for GFP expression. So far we have analysed several positive plants but we found GFP expression in the ovary only in a few lines. When the construct complements the osmads13 phenotype we will soon collect seeds from these positive T0 lines, and then we can start with the ChIP analysis using T1 young panicles.