In order to assess global changes in gene expression patterns in stress-induced tissues, we conducted large-scale analysis of expressed sequence tags (ESTs) in common wheat. Twenty-one cDNA libraries derived from stress-induced tissues, such as callus, as well as liquid cultures and abiotic stress conditions (temperature treatment, desiccation, photoperiod, moisture and ABA) were constructed. Several thousand colonies were randomly selected from each of these 21 cDNA libraries and sequenced from both the 5' and 3' ends. By computing abundantly expressed ESTs, correlated expression patterns of genes across the tissues were monitored. Furthermore, the relationships between gene expression profiles among the stress-induced tissues were inferred from the gene expression patterns. Multi-dimensional analysis of EST data is analogous to microarray experiments. As an example, genes specifically induced and/or suppressed by cold acclimation and heat-shock treatments were selected in silico. Four hundred and ninety genes showing fivefold induction or 218 genes for suppression in comparison to the control expression level were selected. These selected genes were annotated with the BLAST search. Furthermore, gene ontology was conducted for these genes with the InterPro search. Because genes regulated in response to temperature treatment were successfully selected, this method can be applied to other stress-treated tissues. Then, the method was applied to screen genes in response to abiotic stresses such as drought and ABA treatments. In silico selection of screened genes from virtual display should provide a powerful tool for functional plant genomics.

Stress-induced abscisic acid (ABA) is mainly catabolized by ABA 8′-hydroxylase (ABA8ox), which also strictly regulates endogenous ABA levels. Although three members of the ABA8ox gene family are conserved in rice, it is not clear which stressors induce expression of these genes. Here, we found that OsABA8ox1 was induced by cold stress within 24 h and that OsABA8ox2 and OsABA8ox3 were not. In contrast, OsABA8ox2 and OsABA8ox3 were ABA-inducible, but OsABA8ox1 was not. OsABA8ox1, OsABA8ox2, and OsABA8ox3 restored germination of a cyp707a1/a2/a3 triple mutant of Arabidopsis to rates comparable to those of the wild type, indicating that OsABA8ox1, OsABA8ox2, and OsABA8ox3 function as ABA-catabolic genes in vivo. Transgenic rice lines overexpressing OsABA8ox1 showed decreased levels of ABA and increased seedling vigor at 15 °C. These results indicate that sustained low levels of ABA lead to increased seedling vigor during cold stress. On the other hand, excessively low endogenous ABA levels caused reduced drought and cold tolerance, although some of the transgenic rice lines expressing OsABA8ox1 at moderate levels did not show these harmful effects. Adequate regulation of endogenous ABA levels is thought to be crucial for maintaining seedling vigor under cold stress and for cold and drought tolerance in rice.

Galactinol synthase (GolS) is considered to be a key regulator of the biosynthesis of Raffinose family oligosaccharides (RFOs). Accumulation of RFOs has been reported to play a role in protection against abiotic stresses. We identified two cDNAs encoding galactinol synthase from wheat (Triticum aestivum L.), which we designated as TaGolS1 and TaGolS2. Expression of the two TaGolS genes was induced by cold stress but not by drought, heat stress or ABA treatment in wheat. We generated transgenic lines of rice (Oryza sativa L.) constitutively overexpressing TaGolS1 or TaGolS2. These transgenic plants accumulated significantly higher levels of galactinol and raffinose than did wild-type plants and exhibited enhanced cold-stress tolerance. The results demonstrate the involvement of galactinol and raffinose in the development of chilling stress in rice and indicate that the genetic modification of the biosynthesis of RFOs by transformation with GolS genes could be an effective method for enhancing chilling-stress tolerance in rice.

Potato (Solanum tuberosum L.) and sweetpotato (Ipomoea batatas L.), which are nutritionally and commercially important tuberous crops, possess a perplexing heredity because of their autopolyploid genomes. To reduce cross-breeding efforts for selecting superior cultivars from progenies with innumerable combinations of traits, DNA markers tightly linked to agronomical traits are required. To develop DNA markers, we developed a method for quantitative trait loci (QTL) mapping using whole-genome next-generation sequencing (NGS) in autopolyploid crops. To apply the NGS-based bulked segregant method, QTL-seq was modified. (1) Single parentspecific simplex (unique for one homologous chromosome) single-nucleotide polymorphisms (SNPs), which present a simple segregation ratio in the progenies, were exploited by filtering SNPs by SNP index (allele frequency). (2) Clusters of SNPs, which were inherited unevenly between bulked progenies with opposite phenotypes, especially those with an SNP index of 0 for the bulk that did not display the phenotypes of interest, were explored. These modifications allowed for separate tracking of alleles located on each of the multiple homologous chromosomes. By applying this method, clusters of SNPs linked to the potato cyst nematode resistance H1 gene and storage root anthocyanin (AN) content were identified in tetraploid potato and hexaploid sweetpotato, respectively, and completely linked DNA markers were developed at the site of the presented SNPs. Thus, polyploid QTL-seq is a versatile method that is free from specialized manipulation for sequencing and construction of elaborate linkage maps and facilitates rapid development of tightly linked DNA markers in autopolyploid crops, such as potato and sweetpotato.