Construction of floral transcriptome and mirna databases for apomictic and sexual genotypes of eragrostis curvula

Abstract

Eragrostis curvula is a perennial grass native to Southern Africa and widely distributed in semiarid regions of Argentina. Some cultivars are polyploids (from 4x to 8x) and reproduce by diplosporous apomixis, mode ofreproduction strongly affected by ploidy. Diploids (2n = 2x = 20) are always sexual and very infrequent. Here we report the construction of floral transcriptome and miRNA databases for apomictic and sexual genotypes of this grass as important tools to study the complex nature of this intriguing reproductive mode. Total RNA was extracted from flowers at different developmental stages from the tetraploid genotypes Tanganyika USDA (apomictic) and OTA USDA (sexual). Two samples from each genotype were collected (two different plants, biological replicas constituted by a mix of different developmental stages) and the same RNA was divided in order to sequence the transcripts and the miRNAs. Sequencing was carried out at INDEAR (Instituto de Agrobiotecnología de Rosario, Rosario, Argentina). Each library (sexual or apomictic) by using the 454 GS FLX+ Roche method, according to the protocol provided by the manufacturer. miRNA were sequenced using the Illumina platform at GenXpro (Germany). The total number of reads from the transcriptome was 2,617,197, with a total number of952,693,285 bp and an average read length of 364.01 bp. Transcripts de novo assembly (Newbler software package) gave a total of 49,568 contigs (~80,000,000 bp) and 133,782 singletons (~40,000,000 bp). Annotation was done by using the software BLAST2GO, which assigns Gene Ontology terms, being the most represented categories biological, biosynthetic and metabolic processes, reproduction, response to stress, small molecule metabolic process and transport. From these genes, 27, 806 were shared for both, apomictic and sexual genotypes and 3,384 and 6, 366 transcripts were detected only in the sexual and apomictic genotype, respectively. From these, ~45% could not be annotated (~5,000 genes). The use of the software MEGAN allowed us to detect approximately 5000 genes that are not represented in other plant species and that are excellent candidates to analyze. Relative to miRNA, 1,670,483 and 2,274,914 reads were detected in the sexual and apomictic genotype respectively. The role ofthese elements in the reproductive mode is being investigated.