LAUSR

Abstract Plant adaptation to drought stress is essential for plant survival and crop yield. Recently, harnessing drought memory, which is induced by repeated stress and recovery cycles, was suggested as a means to improve drought resistance at the transcriptional level. However, the genetic mechanism underlying drought memory is unclear. Here, we carried out a quantitative analysis of alternative splicing (AS) events under drought memory stress conditions in rice, generating 12 transcriptome datasets. Notably, we identified exon skipping (ES) as the predominant AS type (>80%) in differential alternative splicing (DAS) in response to drought stress. Applying our analysis pipeline to investigate DAS events following drought stress in six other plant species revealed variable ES frequencies ranging from 9.94% to 60.70% depending on the species, and indicated that our method is free from systematic software bias and that ES frequencies are species specific. To validate our observation that ES was the main DAS type in rice, we re-analyzed two publicly available drought stress datasets from rice and obtained similarly high ES frequencies of 70%. The dinucleotide sequence at AS splice sites in rice following drought stress was preferentially GC-AG and AT-AC. Since U12-type splicing uses the AT-AC site, this suggests that drought stress may increase U12-type splicing, and thus increase ES frequency. We hypothesize that multiple isoforms derived from exon skipping may be induced by drought stress in rice. We also identified 20 transcription factors and three highly connected hub genes with potential roles in drought memory that may be good targets for plant breeding.