Bioinformatic tools for understanding apomixis in Eragrostis curvula (Schrad) Nees

ECHENIQUE V; GARBUS I; SELVA JP; ROMERO J; RODRIGO JM; ZAPPACOSTA D

Bahía Blanca

Congreso; VI Argentinian Conference on Bioinformatics and Computational Biology; 2015

Universidad Nacional del Sur

Apomixis, clonal reproduction by seeds, is a puzzling process focused by the scientific community for many years that has interesting evolutionary implications as well as great economical and agricultural potentials, that could significantly change agricultural practices. It probably arose from a few alterations in genes controlling plant sexuality. The switch from sex to apomixis is hypothesized to result from deregulation of developmental pathways leading to sexual seed development, and the trigger for deregulation involves the global genomic effects of hybridization and polyploidy. Our research group has been studying the molecular mechanisms involved in the reproductive behavior of Eragrostis curvula, an apomictic perennial grass widely naturalized in semiarid regions or Argentina. Among the strategies that have been chosen to dissect apomixis regulation are transcriptome sequencing and the use bioinformatic tools . Thus, initial studies were conducted on 12,295 expressed sequence tags (ESTs) generated from near-isogenic lines with different ploidy levels and reproductive modes, leading to genome coverage of 22%. ESTs were clustered and assembled in 8,864 unigenes, 79% of which were functionally categorized by BLASTX analysis against public databases, but only 38% could be classified according to Gene Ontology. A comparative expression analysis between libraries revealed that 112 unigenes were modulated by the reproductive mode or by ploidy. These genes were in silico mapped onto maize genome to identify candidates mapping to the region syntenic to the diplospory locus. As next generation sequencing (NGS) technologies became available, a reference transcriptome for this grass was obtained through a deeper sequencing by 454 GS FLX+ Roche (INDEAR, Rosario, Argentina), leading to 2,617,197 reads (952 Mbp) with an average length of 364 bp. Due to a lack of E. curvula whole-genome sequence, a de novo assembly of the high quality clean reads was conducted. Using the Newbler assembler software (v 2.6; Roche, IN, USA), 63,763 contiguous sequences were obtained and further assembled into 49,568 isotigs, the bioinformatic equivalent to unique RNA transcripts. Annotation was done by the software BLAST2GO. An in silico differential expression analysis revealed that ~10% of the transcripts were expressed by the sexual genotype and ~20% by the apomictic one, constituting a starting point to the in vitro validation of genes involved in apomixis regulation. The use of the software MEGAN allowed us to detect ~5000 genes that are not represented in other plant species and that are excellent candidates to analyze. The E. curvula reference transcriptome has been also useful to develop probes to microarray design (Agilent) to identify genes expressed according reproductive mode though hybridization with mRNA from ovules. Complementary to such strategies, four small RNAs libraries were constructed though NGS from apomictic and sexual genotypes, to elucidate their relevance in apomixis expression. A total of 163.728.389 reads were obtained with a size average of 23 nt, being overrepresented as expected the sRNAs of 21 and 24nt. The software CLC Genomics Workbench (6.0.3) set to a minimum sampling count of 3, lead to 1,869,631 and 2,596,410 different tags for the sexual and apomictic libraries, respectively. The BLAST analysis of such sequences against miRNA databases revealed that 0.4% of the tags correspond to miRNAs being identified matches with only the 10% of the sequences deposited in such database. Bioinformatic tools are very useful tools for our research and will help to understand the nature of apomixis in plants.