Transcriptomic analysis identifies potential regulators ofCpATFL1
To study the transcriptional changes associated with the induction of flowering in more depth, differential expression (DE) profiling, using RNA-seq, was performed. Leaf samples (January 2017) from 17Hot transplants located at UC which flowered heavily the next season were compared to the control plants (17Control; January 2017) that remained vegetative in the next season which was a non-masting year.
High-throughput 150 bp paired-end sequencing yielded 32 GB of raw data with 120 million average read counts for each replicate. Reads were assembled into a reference transcriptomic assembly using the Trinity pipeline. The generated de novo assembly yielded 140,826 transcripts, comprised of 383,092 contigs with an average length of 1428.92 bp and an N50 length of 2414 bp (Table S3). A total of 29,566 contigs (adj P < 0.01) were significantly differentially expressed (DE) with 14,514 and 15,052 transcripts significantly up and downregulated, respectively (Fig. 3).
Gene Ontology analysis: To gain insights into the function of the genes that were DE, contigs were functionally categorised on the basis of putative biological processes, molecular function, and cellular localisation. Out of 29,566 DE contigs, 15,974 (54.09 %) were annotated against the B. distachyon protein database with an E-value of 10-5. The DE genes were further categorised based on gene ontology using a hypergeometric test with a significance threshold of 0.05 to identify key correlations between the internal cellular activity and the phenotypic differences.
Upregulated genes in the leaf samples associated with tillers that flowered in the next season were significantly enriched in the cellular components belonging to the cytoplasm, organelle, and intracellular organelle as the top three categories. Proteins encoded by the upregulated genes were further clustered into separate functional categories belonging to protein binding, transferase activity and anion binding. These genes were found to be involved in biological processes enriched in response to stimuli, oxidation-reduction processes and cellular response to stimulus (Fig. 3d). Similarly, gene ontology analysis was also carried out for the proteins encoded by the downregulated transcripts. The downregulated transcripts were significantly enriched in the cellular components assigned to the cytoplasm, cell periphery and plasma membrane. These transcripts were then further clustered into distinct molecular functions with most of them belonging to transferase activity, anion binding, and small molecule binding. Finally, these proteins were assigned to the category of biological processes involved in organo-nitrogen compound metabolic process, protein metabolic process, and biological regulation as the top three classes (Fig. 3d).
2,786 downregulated transcripts when mapped to the KEGG database were enriched in biosynthesis of secondary metabolites, metabolic pathways and circadian rhythms with a false discovery rate of less than 0.05 (Fig. S5). About 3,788 (12.9%) transcripts of the upregulated genes were found to be associated with the KEGG pathways. The genes were significantly enriched in metabolic pathways (44.1%), biosynthesis of secondary metabolites (26.1%) and protein processing in endoplasmic reticulum (4.75%) as the top three categories.
Differentially expressed orthologues of floral genes in C. pallens: Out of the 29,567 DE contigs, 200 homologous floral protein sequences (from A. thaliana and B. distachyon ) were significantly differentially expressed in the leaves of the tillers that flowered in the next season compared with tillers that remained vegetative (Table S4). Floral integrator genes, including CpMADS1and CpATFL1 , were highly expressed in the tillers that flowered in the next season which also aligns with the quantitative PCR analysis performed earlier (Fig. 3c; Appendix S2). FRIGIDA (CpFRI ), a known floral repressor in A. thaliana (Choi et al., 2011) andB. distachyon and other CpFRI -interacting genes were downregulated in the tillers that flowered in the next season (Table S4).
Leaf samples from 17Hot transplants that flowered in the next season also showed an increase in the expression of thermosensory genes including CpPIF4 , CpPIF5 and CpbHLH80 relative to plants at the control site. These genes also act as floral promoters in response to high temperatures (Kumar et al., 2012). CpSPL15 a known floral promoter in perennial plants (Hyun et al., 2019), was also upregulated in the 17Hot transplants compared to the control plants.CpVRN2 , another temperature regulated floral repressor (Yan et al., 2004) was also downregulated in the tillers during the inductive summer period that subsequently flowered (Fig. 3c).
Two gibberellin catabolism gene family members (CpGA2ox1 andCpGA2ox8 ) were also downregulated, while genes involved in gibberellin synthesis, including CpKS (kaurene synthase) andCpGA20ox2 were upregulated in the tillers that subsequently flowered. Gibberellins have been shown to promote flowering in plants either by the activation of FT through SPL-family proteins or by activation of SOC1 , independent of FT (Yu et al., 2012).
Several epigenetic editor genes, known to deposit active methylation marks to activate the expression of flowering promoting genes, includingCpREF6 , CpMSI1 , CpFLD , and CpEBS (He, 2012), were upregulated in the tillers that flowered in the next season compared to the tillers from the plants at the control site that had remained vegetative (Table S4). Additionally, genes involved in the epigenetic repression of floral repressors such as CpFLK ,CpFY , CpFPA and CpVEL1 (Qüesta, Song, Geraldo, An, & Dean, 2016) were also upregulated in the tillers that subsequently flowered (Table S4).