Genome-Wide Analysis of Long Noncoding RNAs and Their Responses to Drought Stress in Cotton (Gossypium hirsutum L.)

Abstract

Recent researches on long noncoding RNAs (lncRNAs) have expanded our horizon of gene regulation and the cellular complexity. However, the number, characteristics and expression patterns of lncRNAs remain poorly characterized and how these lncRNAs biogenesis are regulated in response to drought stress in cotton are still largely unclear. In the study, using a reproducibility-based RNA-sequencing and bioinformatics strategy to analyze the lncRNAs of 9 samples under three different environment stresses (control, drought stress and re-watering, three replications), we totally identified 10,820 lncRNAs of high-confidence through five strict steps filtration, of which 9,989 were lincRNAs, 153 were inronic lncRNAs, 678 were anti-sense lncRNAs. Coding function analysis showed 6,470 lncRNAs may have the ability to code proteins. Small RNAs precursor analysis revealed that 196 lncRNAs may be the precursors to small RNAs, most of which (35.7%, 70) were miRNAs. Expression patterns analysis showed that most of lncRNAs were expressed at a low level and most inronic lncRNAs (75.95%) had a consistent expression pattern with their adjacent protein-coding genes. Further analysis of transcriptome data uncovered that lncRNAs XLOC_063105 and XLOC_115463 probably function in regulating two adjacent coding genes CotAD_37096 and CotAD_12502, respectively. Investigations of the content of plant hormones and proteomics analysis under drought stress also complemented the prediction. We analyzed the characteristics and the expression patterns of lncRNAs under drought stress and re-watering treatment, and found lncRNAs may be likely to involve in regulating plant hormones pathway in response to drought stress.

Fig 1. Morphological changes of seedlings after treatments and identification of long noncoding RNAs (lncRNAs).

(a) Morphological changes of cotton seedlings after drought and re-watering treatments at trefoil stage. (b) The classification of transcripts by RNA-seq. (c) The Pipeline for the identification of lncRNAs in Gossypium hirsutum L.

Fig 2. Characterization analysis of lncRNAs.

(a) The length compassion analysis of lncRNA and mRNA. X axis indicates the length of lncRNAs and mRNAs (bp), while Y axis represents the frequent count. (b) The features of different types of lncRNAs, including lincRNAs, anti-sense_lncRNAs and intronic_lncRNAs. (c) The ORF compasion analysis of lncRNA and mRNA. X axis indicates the number of ORFs, while Y axis represents the frequent count. (d) Coding potential prediction of lncRNAs by CPC and PFAM software. Venn diagrams show number of lncRNAs with coding potential.

Fig 3. Transposable elements (TEs) prediction and expression patterns analysis of lncRNAs.

(a) The results of repeat sequences and transposable elements (TEs) prediction. GhAt, GhDt and GhUn represent A subgroup, D subgroup and ungrouped, respectively. (b) The faction of lncRNAs with different characteristics. (c) Overall expression levels difference between lncRNAs and mRNAs. (d) Expression patterns analysis of different lncRNAs in response to the drought stress. D-C, Re_W-D means drought vs control, Re-Watering vs drought, respectively and +, - represents up-regulated and down-regulated, respectively.

Fig 4. Differentially expressed lncRNAs and expression validation of lncRNAs in cotton with qRT-PCR.

(a) Expression pattern analysis of lncRNAs-PCgene pairs. (b) Heatmap showed the real-time (RT)-PCR validation of the expression of 32 lncRNAs, including 16 lincRNAs, 10 anti-sense lncRNAs and 6 intronic lncRNAs. Histogram showed the FPKM value of 32 lncRNAs from the result of RNA-seq. (c) Differentially expressed genes (DEGs) analysis under drought and re-watering stress.

Fig 5. Gene ontology (GO) enrichment of lncRNAs and KEGG enrichment of lncRNA-targets.

(a) The results of GO analysis based on the differently expressed lncRNAs. (b) Main enrichments of differentially expressed lncRNAs based on Gene ontology (GO) analysis. (c) Statistical KEGG enrichment of lncRNA-targets genes using KOBAS software.

Fig 6. Regulation mechanism prediction of functional lncRNAs predicted by cis-acting and trans-acting and relative analysis.

(a) Regulation mechanism prediction of functional lncRNA-targets predicted by cis-acting and trans-acting. (b) The expression analysis of 30 randomly selected lncRNAs associated with each kind of hormones. (c) The content variations of plant hormones under drought stress and re-watering treatment. (d) The 2-DE image of total protein in leaves of ZhongH177 at trefoil stage and several differentially expressed proteins were enlarged. A represents proteins in ZhongH177-CK, B represents ZhongH177-Drought and C represents the enlargement of differentially expressed proteins.