Transcriptome analysis provides insights into the molecular mechanism of GhSAMDC1 involving in rapid vegetative growth and early flowering in tobacco

In previous study, ectopic expression of GhSAMDC1 improved vegetative growth and early flowering in tobacco, which had been explained through changes of polyamine content, polyamines and flowering relate genes expression. To further disclose the transcript changes of ectopic expression of GhSAMDC1 in tobacco, the leaves from wild type and two transgenic lines at seedling (30 days old), bolting (60 days old) and flowering (90 days old) stages were performed for transcriptome analysis. Compared to wild type, a total of 938 differentially expressed genes (DEGs) were found to be up- or down-regulated in the two transgenic plants. GO and KEGG analysis revealed that tobacco of wild-type and transgenic lines were controlled by a complex gene network, which regulated multiple metabolic pathways. Phytohormone detection indicate GhSAMDC1 affect endogenous phytohormone content, ABA and JA content are remarkably increased in transgenic plants. Furthermore, transcript factor analysis indicated 18 transcript factor families, including stress response, development and flowering related transcript factor families, especially AP2-EREBP, WRKY, HSF and Tify are the most over-represented in those transcript factor families. In conclusion, transcriptome analysis provides insights into the molecular mechanism of GhSAMDC1 involving rapid vegetative growth and early flowering in tobacco.


Results
Phenotype identification of transgenic plants. In previous study, transgenic plants (3-1, 3-2, and 4-3) from three independent transformation events had been analyzed and identified 22 . Compared to wild type, transgenic plants showed rapid vegetative growth in early bolting and flowering stage ( Fig. 1A-C). Based on the previous expression analysis, we selected seedling stage ( Primary analysis of transcriptome data. Transcriptome analysis indicated sequencing produced libraries averaging 57.83 million clean reads per library, and average of Q20 and Q30 value were 97.24% and 92.11%, respectively (Table S1). Heat map of Pearson correlation coefficient between different samples was shown in Fig. 2A, the coefficient were 0.915-0.982 between twice replicates, and details were described in supplementary. Principal Component Analysis (PCA) show that the clusters in different periods have large differences (Figure S1). Fragments Per Kilobase of transcript sequence per Millions base pairs sequenced (FPKM) distribution analysis suggested the percentage of FPKM interval 0-1, 1-3, 3-15, 15-60, > 60 are 54.52%, 14.96%, 21.76%, 6.69% and 2.08%. The violin diagram of gene expression level is shown in Fig. 2B, and details were described in supplementary Table S2. Based on the above data analysis, a cluster analysis of the differentially expressed transcripts was described in Fig. 2C. Com-parisons of gene expression between WT and the two transgenic lines were performed at different stages. Overall, there were 938 genes identified that were differentially expressed in the transgenic lines compared to the WT (|log 2 (fold change) |≥ 2 and p adj < 0.05). Among these genes, 82, 241 and 615 DEGs were acquired at seedling, bolting and flowering stages, respectively (Table S3). Compared to WT, transgenic line 3-2 had 181 DEGs (131 up-and 50 down-regulated), and transgenic line 4-3 had 757 DEGs (438 up-and 319 down-regulated), the details are described in Fig. 3A. As was shown in Fig. 3B, two three-way Venn diagrams of 3-2 vs WT and 4-3 vs WT at different stages were displayed, respectively. 120 were unique ones with 15, 23 and 41 unique to seedling, bolting and flowering stage in the transgenic line 3-2 and WT comparison, respectively. Of the 676 non-redundant DEGs in the transgenic line 4-3 compared to WT, there were 31 DEGs shared at different stages, there were 6, 133 and 487 genes to seedling, bolting and flowering stage in the transgenic line 4-3 and WT.

GO and KEGG pathway analysis of DEGs between 4-3 and WT at different stages.
To gain a better understanding of the molecular mechanism of GhSAMDC 1 involving in rapid vegetative growth and early flowering in tobacco, GO annotations of the DEGs and enrichment of KEGG pathway were performed in comparison 3-2 vs WT and 4-3 vs WT at different stages, it found that due to the small number of DEGs between 3 and 2 and WT, GO were not enriched at different stages. Therefore, we only compared between 4 and 3 and WT at different stages. GO annotations of the DEGs between WT plants and transgenic line 4-3 at different stages www.nature.com/scientificreports/ were presented in Fig. 4A-C, Using the criterion of corrected P Value ≤ 0.05, a total of 49 GO terms were enriched at bolting stage, There were 41 GO terms involved in biological process level, mainly including regulation of cellular process, regulation of biological process, primary metabolic process, transcription, DNA-templated, nucleic acid-templated transcription, RNA biosynthetic process, cellular macromolecule biosynthetic process, macromolecule metabolic process, RNA metabolic process, nitrogen compound metabolic process, nucleobasecontaining compound metabolic process. In cellular component level, DEGs were over-represented in proteinaceous extracellular matrix, extracellular matrix, extracellular region part, extracellular region. For molecular function level, over-represented DEGs were associated with nucleic acid binding transcription factor activity, sequence-specific DNA binding, inositol-3-phosphate synthase activity, intramolecular lyase activity (Fig. 4B, Table S4). 15 GO terms were enriched at flowering stage, 8 biological process terms were enriched were related to coenzyme M biosynthetic process, coenzyme M metabolic process, cell morphogenesis, cellular component morphogenesis, cellular developmental process, protein folding, oxidation-reduction process, anatomical structure morphogenesis, in molecular function level mainly involved terpene synthase activity, carbon-oxygen lyase activity, magnesium ion binding, lyase activity, oxidoreductase activity, unfolded protein binding. (Fig. 4C, Table S5). In order to further analyze the DEGs, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis ( Fig. 4D-F), using the criterion of corrected P Value ≤ 0.05, KEGG enrichment analysis indicated histidine metabolism between 3 and 2 and WT at bolting stage (Table S6). At the bolting stage in 4-3 vs WT, pathways related to plant-pathogen interaction (Fig. 4E, Table S6), at the flowering stage between 4 and 3 and WT, they were involved protein processing in endoplasmic reticulum, alpha-Linolenic acid metabolism, spliceosome, sesquiterpenoid and triterpenoid biosynthesis, linoleic acid metabolism (Fig. 4F, Table S6). These results indicate that tobacco of wild-type and transgenic are controlled by a complex gene network, which regulates multiple metabolic pathways.

Protein-protein interaction (PPI) analysis.
To further understand the relationship of DEGs, PPI network analyses of DEGs with FDR ≤ 0.05 and |Log 2 FC|≥ 2 was performed using STRING database. Through  Quantitative RT-PCR validation of RNA-sequencing data. To confirm the accuracy of the transcriptome analysis results, real-time quantitative reverse transcription-PCR (qRT-PCR) was used to analyze DEGs identified in the RNA-Seq results. 6 DEGs between wild type and transgenic lines at different stages were selected for qRT-PCR analysis. There was a general agreement between the RNA-seq (Table S7) and qRT-PCR ( Fig. 6) results for all genes analyzed validating the RNA-seq library.

Discussion
Interaction of polyamine and hormone affected plant growth and development. Based on our previous study, ectopic expression of GhSAMDC 1 resulted in Spd content decrease and then increase accompany with Put decrease at flowering stage in transgenic lines, which indicates Spd plays an important role in floral induction 22 . Changes of PA content could alter gene expression level of genes associated with plant hormone signaling transduction in Arabidopsis 31 , and Spd involved floral induction by decreasing GA3 and increasing MdNCED1 and MdNCED3 through ABA enrichment in apple 32 . Recently, plant CuAOs involving in polyamine terminal oxidation were induced by stress-related hormones, including methyl-jasmonate (MeJA), abscisic acid (ABA) and salicylic acid (SA), which indicated polyamine homeostasis was affected by plant hormone 33 . In thus, crosstalk between polyamine and hormone plays a crucial role in plant growth and development. www.nature.com/scientificreports/ Phytohormone detection was performed in wild type and transgenic lines at different stages (Fig. 7). IAA, BR and ZT were not detected in all samples, GA3 was only detected in transgenic line 4-3 at flowering stage. GA are closely associated with DELLA involving in flowering pathways 34,35 , and a highly conserved family of R2R3 MYB transcription factors had been reported involving in the GA signaling pathway and flowering [36][37][38] . SA content was detected in all stages and decreased accompany with growth in wild type plants, whereas only was detected in later stages and showed reverse tendency in transgenic lines. Increase of Spd enhanced SA content in the leaves had been reported in wheat 39 , and vice versa SA treatment affects the PAs metabolism in plants 27,[40][41][42] . In Arabidopsis, exogenous SA appears to be a repressor of the expression of flowering related gene FLC 43 . Remarkably, ABA and JA were detected in all samples, and the content of them in transgenic lines were higher than that in wild type plants, it was consistent with exogenous Spm induced JA accumulation in lima bean 24 . Furthermore, a positive feedback loop has also been found between ABA and PAs, ABA activates the PA metabolism as well as PA induces ABA synthesis 25,44,45 , and various miRNAs have also been documented to be involved in flowering by means of ABA signaling and regulation [46][47][48] . Further indicated a complicated relationship exist between PAs

Transcription factor genes affect rapid vegetative growth and early flowering in transgenic plants overexpressing GhSAMDC 1 . Analyzing GO annotation of the DEGs between 4 and 3 and WT
at bolting stage found that there were enriched transcription factor activity, therefore, the transcription factor was analyzed. As is shown in Table S8, 18 transcription factor families were identified in those DEGs, and AP2-EREBP, WRKY, HSF and Tify were the most over represented families. In these families, AP2/EREBP including 14 genes was the most abundant transcription factor family. As we known, AP2/EREBP comprise AP2-like genes (with two AP2 domains) and EREBP-like or ERF-like genes (with one AP2 domain) 49 , and they are involved in plant development, especially reproductive growth through regulating ABA content 50,51 . In this study, compared to wild type, ABA content was remarkably increased in transgenic lines at different stages, which indicated increase of ABA content might be regulated by those AP2/EREBP genes. Furthermore, 9 WRKY transcription factor genes were identified and over-represented. Except for responding to biotic and abiotic stress 52,53 , WRKY genes are also involved in several physiological and developmental processes 54,55 . Especially, many studies have demonstrated that WRKY genes play crucial roles in flowering [56][57][58][59] . In addition, both HSF and Tify transcription factor family contained 6 genes, and they were also over-represented in those DEGs. HSF had been reported that they not only involve in abiotic stress response [60][61][62] , but also regulate plant development 63 and flowering 29,30 . Tify are plant-specific transcription factors characterized by the presence of a highly conserved motif (TIF(F/Y) XG) in the TIFY domain 64 . Many studies had been reported they encodes proteins involve in multiple biological processes through modulating jasmonate signaling pathway 65 , including leaf development 66 , abiotic stress  (Table S9). Twice biological replicates were performed.
Library preparation and Transcriptome sequencing. Total RNA was extracted using Trizol reagent following the manufacturer's protocol. RNA purity was checked using the NanoPhotometer ® spectrophotometer (IMPLEN www.nature.com/scientificreports/ sidered significantly enriched by differential expressed genes. KOBAS software was used to test the statistical enrichment of differential expression genes in KEGG pathways (www. kegg. jp/ kegg/ kegg1. html). PPI analysis of differentially expressed genes was based on the STRING database (https:// cn. string-db. org/), which known and predicted Protein-Protein Interactions 73 , Nodes represent the DEGs enriched in the STRING database, while edges reflect the interactions between differentially expressed genes, the interaction data was imported into Cytoscape software v3.2.0 (http:// cytos cape. org) to realize the visualization of the interaction network. Transcript factors analysis was performed using iTAK software online (http:// itak. feilab. net/ cgi-bin/ itak/ index. cgi).
RNA isolation and quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) analysis. Total RNA was extracted from the samples using the modified CTAB method 22 . qRT-PCR was performed using Power SYBR Green Master (Roche, Basel, Switzerland) on a Roche Light Cycler 480 system (Roche), as described previously. The reaction was run as follows: pre-incubation at 94 °C for 2 min, 40 cycles of 94 °C for 20 s, 58 °C for 20 s and 72 °C for 20 s. The actin gene (Tac9; X69885) was used as the reference gene, and the relative 2 −∆ct quantification method was used to evaluate quantitative variation. Three biological replicates and three technical repeats were run. The qRT-PCR primers are listed in Table S10.
Phytohormone detection. Leaves from wild type and transgenic lines 3-2 and 4-3 were sampled for phytohormone detection at seedling (30 days old), bolting (60 days old) and flowering (90 days old) stages, Three biological replicates were performed for each sample. Samples were rapidly frozen by liquid nitrogen and transported with dry ice for analysis. Phytohormone detection was performed by Shiseido SP HPLC-Thermo TSQ Quantum Ulta MS/Ms in sci-tech innovation company in Qingdao. All images are combined through Photoshop.
Ethics statement.

Data availability
The data and material that support the fndings of this study are available from the corresponding author on reasonable request. www.nature.com/scientificreports/