Abstract Background The medicinal plant, Catharanthus roseus (C. roseus), accumulates a wide range of terpenoid indole alkaloids (TIAs). Ethylene (ET) and methyl-jasmonate (MeJA) were previously reported as effective elicitors for the production of various valuable secondary metabolites of C. roseus, while a few ET or MeJA induced transcriptomic research is yet reported on this species. In this study, the de-novo transcriptome assembly of C. roseus is performed by using the next-generation sequencing technology. Results The result shows that phenolic biosynthesis genes respond specifically to ET in leaves, monoterpenoid biosynthesis genes respond specifically to MeJA in roots. By screening the database, 23 ATP-binding cassette (ABC) transporter partial sequences are identified in C. roseus. On this basis, more than 80 key genes that encode key enzymes (namely TIA pathway, transcriptional factor (TF) and candidate ABC transporter) of alkaloid synthesis in TIA biosynthetic pathways are chosen to explore the integrative responses to ET and MeJA at the transcriptional level. Our data indicated that TIA accumulation is strictly regulated by the TF ethylene responsive factor (ERF) and bHLH iridoid synthesis 1 (BIS1). The heatmap, combined with principal component analysis (PCA) of C. roseus, shows that ERF co-expression with ABC2 and ABC8 specific expression in roots affect the root-specific accumulation of vinblastine in C. roseus. On the contrast, BIS1 activities follow a similar pattern of ABC3 and CrTPT2 specific expression in leaves, which affects the leaf-specific accumulation of vindoline in C. roseus. Conclusions Results presented above illustrate that ethylene has a stronger effect than MeJA on TIA induction at both transcriptional and metabolite level. Furthermore, meta-analysis reveals that ERF and BIS1 form a positive feedback loop connecting two ABC transporters respectively and are actively involved in TIAs responding to ET and MeJA in C. roseus. Electronic supplementary material The online version of this article (10.1186/s12864-018-4879-3) contains supplementary material, which is available to authorized users. Keywords: ABC transporter, Catharanthus roseus, Ethylene, Methyl-jasmonate, Next-generation sequencing, Terpenoid indole alkaloids Background Alkaloids are diverse groups of low-molecular-weight nitrogen-containing compounds found in about 20% of plant species and approximate 12,000 alkaloids have been elucidated for their chemical structures [[35]1]. Due to their strong and divergent biological activities, many alkaloids have been extensively applied to clinical therapy [[36]2]. Catharanthus roseus, a Terpenoid Indole Alkaloids (TIAs) producing plant, is the sole source of the anti-cancer compounds, namely vinblastine and vincristine [[37]3–[38]10]. However, the accumulation of these anti-cancer components is exceedingly limited, which hardly meet the demand worldwide. Studies on the elucidation of the biosynthesis pathway and pathway regulations for these compounds production in Catharanthus roseus can be beneficial for promoting the production by metabolic engineering manipulation. During the past several decades, efforts were made to the discovery of TIAs biosynthesis genes, regulation on TIAs biosynthesis pathway and elucidation of pathway intermediates transport [[39]8, [40]11–[41]15]. Although great progress was achieved, there are still many undefined pathway genes, pathway regulators and TIA transporters that require further identification for a better understanding of the production of anti-cancer drug components by C. roseus. Next-generation sequencing (NGS) technologies provide a rapid, cost-efficient way for analyzing genome and transcriptome in non-model species. Recently, many transcriptomic types of research over C. roseus were carried out, which dramatically boosted the discovery of TIAs biosynthesis genes and pathway elucidation [[42]2, [43]11, [44]14, [45]16]. In 2015, Kellner et al. generated a genome assembly for C.roseus that greatly promoted the discovery of monoterpenoid indole alkaloid (MIA) biosynthesis and identified many putative missing pathway genes, transcription factors (TFs) and intermediates transporters [[46]17]. According to the previous research, it is showed that the expression of pathway genes for TIAs biosynthesis is strictly regulated by development-, tissue- and cell-specific controls in response to various environmental stimuli, both biotic and abiotic, in C. roseus [[47]12, [48]17–[49]20]. At the transcription level, the biosynthesis of these TIAs is also extensively regulated by many transcriptional activators and repressors. Apart from pathway genes and transcription regulators, transporters involved in alkaloid translocation also attracted many concerns for TIAs production engineering by manipulating intermediates transport at both intra- and inter- cellular levels [[50]8, [51]14, [52]16, [53]21]. Biosynthetic pathway of TIAs exists in different types of membranes both inside and between cells in C. roseus. The ATP-binding cassette (ABC) proteins can be classified as a large, widespread family; a big part of ABC family is membrane-associated primary transporters. According to past studies, it is discovered that ABC transporters in plant participate in various vital physiological processes [[54]14, [55]22]. It is widely perceived that ABC transporters can bring out vincristine and vinblastine from human cancer cells. Furthermore, ABC transporters are recently discovered to participate in the process of transporting different monoterpenoid alkaloids in C. roseus cells [[56]14]. Our previous works showed that there are at least 12 steps of intermediates transport in the TIAs biosynthesis pathway and an ABC transporter, CrTPT2, is identified as a specific MIA transporter controlling the secretion of catharanthine from leaf epidermis to leaf surface [[57]14]. In C. roseus, two important plant hormones, ethylene (ET) and methyl-jasmonate (MeJA), were confirmed to be involved in regulating TIAs biosynthesis. Many pathway genes and transcriptional regulators were reported to be regulated by these plant hormones, which indicated that Catharanthus TIAs biosynthesis was involved in MeJA and ET signaling transduction pathway [[58]18, [59]23–[60]26]. Investigations of the way these plant hormones regulating TIAs biosynthesis and genes responding to these plant hormones are very helpful for understanding these compounds biosynthesis under different environmental, developmental, and stress conditions. Recently, the application of NGS technique provided plenty of important information on how MeJA regulated genes expression and of which genes (pathway genes, transcription factors, and putative intermediate transporters) possibly involved in TIAs production [[61]18, [62]27]. In comparison with most methyl-jasmonate (MeJA) elicited transcriptomic information in studies about C.roseus, there was a limitation on deep-investigation of ET-induced TIAs production in this species. In order to get a comprehensive understanding of the regulation mechanism of ET and MeJA on TIAs production at both transcriptional and metabolic levels, we sequenced transcriptomes of Catharanthus plants, treated with ET (30 μM), MeJA (50 μM) and control (CK). By conducting blast and qRT-PCR analysis, 23 candidate ABC family transporter genes were screened out that the expression levels are distinct from each other in response to ET and MeJA. Afterward, more than 80 key genes encoding key enzymes (TIA pathway, transcriptional factor (TF), and candidate ABC transporter) of alkaloid synthesis in TIA biosynthetic pathways were chosen to demonstrate that the dose-dependent response of TIA production to MeJA and ET was correlated with the levels of these TFs and transporters. The integration of these genes expression and TIA profiles suggested a mechanism by which TIA production was controlled at the transcriptional level with MeJA and ET dosage. Results De novo assembly and quantitative assessment of Illumina sequence cDNA libraries from Catharanthus seedlings with 4 h treatment of ET (30 μM) and MeJA (50 μM) along with control (CK) were prepared for high throughput sequencing by Illumina Hiseq 2000 system. Three biological repeats were prepared for each condition. 32,891,528,340 nt bases were generated totally. After removing the adapter sequences, empty reads and low-quality sequences, 38,018,320 (CK1), 43,830,696 (CK2), and 40,406,978 (CK3) clean reads from CK, 40,601,918 (ET1), 44,225,374 (ET2), and 40,215,966 (ET3) clean reads from ET, and 40,689,468 (MJ1), 40,044,736 (MJ2), and 37,427,970 (MJ3) clean reads from MeJA were obtained. An overview of sequencing and assembled results have been summarized in Additional file [63]1: Table S1. We then search unigene sequences against protein databases (Nr, SwissProt, KEGG, COG) using blastx (e-value < 0.00001). Protein function information can be predicted from annotation of the most similar protein in those databases (Additional file [64]2: Table S2). Differential gene expression The differentially expressed genes (DEGs) were defined by cuffdiff as those with a false discovery rate (FDR) ≤ 0.001 and fold change equal or larger than 2, as well as requiring Fragment Per Kilobase of exon model per Million mapped reads (FPKMs) > 2. Based on this criterion, we identified 304 (ET vs CK), 910 (MeJA vs CK), and 783 (MeJA vs ET) up-regulated and 79 (ET vs CK), 185 (MeJA vs CK), and 442 (MeJA vs ET) down-regulated genes, respectively (Fig. [65]1, Additional file [66]3: Table S3). DEGs were identified in ET and MeJA treated sample cDNA libraries in comparison with that in CK sample cDNA library via pairwise comparisons. A total of 383 and 1095 genes were differentially expressed with ET or MeJA treatment, respectively, which indicated that the number of DEGs with MeJA treatment was considerably bigger than that with ET treatment. With Nr annotation, we use Blast2GO program to get GO annotation of Unigenes. Then we use WEGO software to do GO functional classification for all Unigenes [[67]28]. According to GO enrichment, some important metabolic activities occurred in C. roseus in response to MeJA and ET. As shown in Additional file [68]4: Figure S1, treatment with MeJA specifically enhanced nutrient reservoir comparing with ET. In contrast, translation regulator and viral reproduction were significantly enriched in response to ET. Meanwhile, treatment with MeJA also enhanced biological adhesion, death, immune system process, and rhythmic process comparing to the treatment with ET by GO assignment system (Additional file [69]4: Figure S1, Additional file [70]5: Table S4). Fig. 1. Fig. 1 [71]Open in a new tab Differential expression analysis of C. roseus transcriptome. Statistical chart of DEGs of transcriptomes in response to hormone stress. Compared with the transcriptional level of control (CK, non-treatment), ET and MeJA, 304, 910, and 783 genes were up-regulated, and 79, 185, and 442 genes were down-regulated, respectively. The number of up-regulated genes was shown in blue, while down-regulated genes were shown in yellow Pathway annotation The KEGG pathway database records networks of molecular interactions in the cells, and variants of them specific to particular organisms. Pathway-based analysis contributed to further understand biological functions of genes in response to ET or MeJA. In total, 30,061 genes (30,061 transcripts) were found to be involved in one of the 401 different KEGG metabolic pathways in total. In order to compare the specific metabolic pathways in which DEGs participated between ET and MJ treatment, KEGG pathway enrichment analysis was conducted. Comparing Fig. [72]2a, b, it was found that Phenylpropanoid biosynthesis (9.9%), Phenylalanine metabolism (5.0%) genes responded specifically to ET, Monoterpenoid biosynthesis (9.0%) and Terpenoid backbone biosynthesis (7.0%) genes responded specifically to MeJA (Additional file [73]6: Table S5). This indicated that the secondary metabolites of C.roseus showed special sensitivity when stimulated by exogenous hormones; with respect to transcriptional level, phenolic compounds were more sensitive to ET, while MeJA had a stronger influence over MIA. Among the 64 different KEGG metabolic pathways under ET treatment, 197 genes were selected involving 3 different KEGG metabolic pathways, which were Isoflavonoid biosynthesis (97 genes), Indole alkaloid biosynthesis (41 genes) and Monoterpenoid biosynthesis (59 genes) pathway (Additional file [74]6: Table S5). Our results showed that 33 of 197 genes in the pathways were up-regulated, while 35 of 197 genes were down-regulated following ET treatment. The results indicated that the ET up-regulated genes mostly associated with two GO terms, metal ion binding (7%) and lyase activity (7%), whereas ET down-regulated genes gathered in the methylation-enzyme-activity group (27.27%, include methylation indole-3-acetate esterase activity, methylation salicylate esterase activity, and methylation jasmonate esterase activity). An interesting finding was that the activity of other hormones metabolic process (include jasmonic acid metabolic process 9.09% and salicylic acid metabolic process 9.09%) was peculiarly decreased (Additional file [75]7: Figure S2, Additional file [76]8: Table S6). Fig. 2. [77]Fig. 2 [78]Open in a new tab KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway annotation. In total, 30,061 genes (30,061 transcripts) were found to be involved in one of the 401 different KEGG metabolic pathways. We used R Package gplot ([79]https://cran.r-project.org/web/packages/gplots/gplots.pdf) to create the graph. a The top 20 statistics of pathway enrichment were represented by comparing ET with control (CK). b The top 20 statistics of pathway enrichment were represented by comparing MeJA and control (CK) The statistical analysis of TIA pathway genes in response to ET or MeJA by RNA-seq Under pathway annotation, it was known that secondary metabolic pathways had the largest number of DEGs when treated with ET or MeJA. Based on the references in recent 10 years, 37 key genes were