IC4R011-RNA-Seq-2010-21122150

Project Title

Massive parallel sequencing of mRNA in identification of unannotated salinity stress-inducible transcripts in rice (Oryza sativa L.)

Microarray technology is limited to monitoring the expression of previously annotated genes that have corresponding probes on the array. Computationally annotated genes have not fully been validated, because ESTs and full-length cDNAs cannot cover entire transcribed regions. Here, mRNA-Seq (an Illumina cDNA sequencing application) was used to monitor whole mRNAs of salinity stress-treated rice tissues.

Seeds of rice (Oryza sativa L.‘Nipponbare’) were germinated in the dark at 28°C on a sterilized germination tray. Germinated seeds were evenly distributed on 96-well PCR plates supported by a plastic container. Seeds were grown in a growth chamber at 28°C, as previously described. After the seedlings had been grown for 7 days, they were transferred on their 96-well plates into containers filled with 150 mM NaCl solution, or with control solution, and placed at 28°C in a growth chamber for 1 h. Four kinds of tissue (normal shoot, normal root, shoot with 1-h salinity stress, or root with 1-h salinity stress) were collected and immediately frozen in liquid nitrogen. For RNA extraction from each treatment group, 10 plants were collected and mixed, to minimize the effect of transcriptome unevenness among plants.

As the number of reads increased, the fraction of highly expressed genes(RPKM ≥ 300) close to their final RPKM was almost unchanged, whereas those of genes with relatively low expression (RPKM 3-30) converged more slowly (Figure 1b). With four technical replicates (corresponding to about 27 to 35 million reads), 81.2% of genes with relatively low expression levels (RPKM 3-30) reached to within ± 5% of their final RPKM (Figure 1b).

Figure 1 Accumulation of 36-bp reads to cover whole transcripts.

Rice transcriptome analysis was based on response to salinity stress. mRNAs were prepared from the tissues of normal rice shoots and roots and from those subjected to 1 h of salinity stress. Of the 27 to 35 million quality-evaluated reads (Table 1; Total filtered reads), 72.0% to 75.2% were mapped uniquely to the rice genome (Table 1; Unique-genome); 5.0% to 5.7% of the reads bridged flanking exons (Table 1; Unique-bridged); 6.0% to 11.2% of the reads were repetitive sequences (Table 1; Multiple); and 10.1% to 16.7% had no match in the genome (Table 1; Unmapped). Thus, a total of 76.9% to 80.9% of the reads were mapped uniquely to the rice genome or to exon-exon junctions (Table 1; Unique-total).

Tabel 1 Numbers of mapped reads.

The researchers comprehensively compared the RPKM of each gene in response to salinity stress (r = 0.95 in shoot and 0.94 in root; Figure 2). The researchers used the G-test with a 1% false discovery rate (FDR) and identified 6,469 (in shoot) and 10,321 (in root) differentially expressed RAP2 genes. Of these, 3,050 (up, 1,651; down, 1,399) genes were commonly differentially expressed. The number of highly differentially expressed genes (>32×) was greater in the root (58 genes) than in the shoot (5 genes). Expression of genes previously identified under salinity stress was induced in the root (> 2×). The distribution of mapped reads on the rice genome was graphed on a GBrowse(Figure 3).

Figure 2 Comparison of RPKM of each gene after salinity stress.

Figure 3 Differential expression of genes in a 100-kb region on chromosome 2.

In the shoot, 51,301 transcripts were predicted (RPKM ≥ 2, length ≥ 100 bp) (Table 2); 94.6% (48,506/51,301) of the predicted transcripts were mapped on previously annotated loci in RAP2; thus, the remaining 2,795 predicted transcripts were unannotated in RAP-DB (Table 2). In the root, 3,082 of the 54,491 predicted transcripts were mapped on unannotated regions(Table 2). The reseachers attempted to predict the functions of unannotated transcripts by BLASTX search and longest-ORF search. In a BLASTX search against the UniProt and RefSeq sequences, of the predicted transcripts, 995 (shoot) and 1,052 (root) had ORFs similar to those encoding the amino acid sequences of functional proteins (Table 2). Of the remaining unannotated transcripts, 1,670 (shoot) and 1,873 (root) had ORFs encoding at least 20 amino acids by longest-ORF search (Table 2). Amino acid length was widely distributed: the mean and median were 125 and 77 amino acids in the shoot, and 123 and 74 in the root (Figure 4).

Table 2 Numbers and ORF predictions of transcripts.

Figure 4 ORF length and number of transcripts.

National Institute of Agrobiological Sciences (NIAS), Division of Genome and Biodiversity Research, 1-2, Kannondai 2-chome, Tsukuba, Ibaraki 305-8602, Japan.
Institute of the Society for Techno-innovation of Agriculture, Forestry and Fisheries, 446-1, Ippaizuka, Kamiyokoba, Tsukuba, Ibaraki 305-0854, Japan.
Hitachi Government & Public Corporation System Engineering Ltd., Solution Division 4, Research & Development Center Bioinformatics Group, 2-4-18 Toyo, Koto-ku, Tokyo 135-8633, Japan.