Transcriptome Analysis of Aconitum carmichaelii and Exploration of the Salsolinol Biosynthetic Pathway

Aconitum carmichaelii has been used in traditional Chinese medicine for treating various diseases for several thousand years. The biosynthetic pathway of some alkaloids such as C19-diterpenoid alkaloids has been reported, but pathways in different varieties of A. carmichaelii remain unknown. Herein, we performed transcriptome analysis of varieties A. carmichaelii and characterized the biosynthetic pathway of salsolinol. The results expand our knowledge of alkaloids biosynthesis, and provide a theoretical basis for analysing differences in alkaloids biosynthesis patterns in different varieties. A total of 56 million raw reads (8.28 G) and 55 million clean reads (8.24 G) were obtained from two varieties (Z175 and R184) leaf transcriptomes, respectively, and 176,793 unigenes were annotated using six protein databases. This yielded 6,873 differentially expressed genes (DEGs) in the two varieties, of which 281 are involved in the salsolinol biosynthetic pathway, including 158 and 75 related to glycolysis and the shikimate pathway, respectively. Furthermore, 843 DEGs were found to be involved in the formation of C19-diterpenoid alkaloids, with 34 differed between the two varieties. These target genes were analysed to explore differences in C19-diterpenoid alkaloid biosynthesis in Z175 and R184. In addition, 322 DEGs encoding transcription factors potentially related to alkaloid accumulation were identified. 2-C-methyl-D-erythritol 4-phosphate cytidylyltransferase, ISPE: 4-(cytidine- 50-diphospho)-2-C-methyl-D-erythritol 2-C- methyl-D-erythritol2,4 -cyclodiphosphate ISPG: (E)-4-hydroxy-3- methylbut-2-enyldiphosphate (E)-4-hydroxy-3-methylbut-2-enyl diphosphate IPPI: isopentenyl diphosphate 3-hydroxy-3-methylglutaryl-

gated sodium channels, the release of neurotransmitters and changes in receptors, and the promotion of lipid peroxidation and cell apoptosis in the heart, liver and other tissues [9]. The principal toxic and pharmacological ingredients of fuzi are aconitine-type C19-DAs [5,9].
In Aconitum, DAs and their derivatives, including C18, C19 and C20 classes, are the principal constituents responsible for biological activity [11]. C19-DAs in processed roots of Aconitum following hydrolysis are possibly the most beneficial bioactive constituents of fuzi [12]. The majority of C19-type diterpene alkaloid skeletons, which account for most of the alkaloid content of A. carmichaelii, are derived from various substitutions on the aconitine skeleton, leading to 75 different known alkaloids [9]. Mevalonate (MVA) and methylerythritol (MEP) pathways play important roles in the biosynthesis of terpenoids [11], and biosynthesis of C19-DAs in A. carmichaelii begins with MVA and MEP biosynthetic pathways, involving geranylgeranyl pyrophosphate synthase (GGPPS) that produce the precursor of DAs [13]. Furthermore, ent-CPP synthases (CPS), ent-kaurene synthases (KS), kaurene oxidases (KOX), cyclases, aminotransferases, monooxygenases, methyltransferase and BAHD acyltransferases act together to synthesis C19-DAs in fuzi, and transcription factors (TFs) appear to regulate the accumulation of DAs.
In A. carmichaelii, C19-type diterpenes are the most abundant DAs, followed by C20-type and other diterpenes [9]. High-performance liquid chromatography (HPLC) analysis with an evaporative scattering light detector (ELSD) carried out on roots from Z175 and Z184 varieties in 2016 and 2017 identified C19-diterpenoid alkaloids including benzoylaconine, hypaconitine, neoline, mesaconitine, benzoylmesaconine, fuziline, aconitine, talatizamine, mesaconine, hypaconine and aconine, as well as the C20-diterpenoid alkaloid songorine, and one other alkaloid, salsolinol (unpublished data). Analysis of expression patterns of unigenes involved in C19-type diterpene biosynthesis in different tissues showed that flowers and buds are the most transcriptionally active tissues, while roots and leaves are characterized by a distinct set of transcriptionally active unigenes [13]. Herein, differences in alkaloid content in different A. carmichaelii varieties and expression patterns in different tissues were investigated. Transcriptome data were acquired from two varieties, and salsolinol biosynthesis pathway-related genes were identified. Finally, differences in diterpenoid alkaloid content and enzymes mediating biosynthesis of C19-DAs were compared.

Results
Transcriptome sequencing and sequence assembly In total, 56 million raw reads (8.28G) and 55 million clean reads (8.24G) were obtained from six transcriptomes Z175 and R184 varieties in three biological replicates, respectively. In Z175, the percentage of reads with q20 and q30 quality scores was 97.32% and 92.61%, compared with 97.61% and 93.26% for R184 (Fig.1 Characterization and expression analysis of unigenes involved in C19-type diterpene biosynthesis in Z175 and R184. Based on the results of previous studies [ 13,21], putative unigenes involved in the aconitine biosynthesis pathway were identified, including 49 unigenes encoding 1-deoxy-  have been reported, one exploring expression in flower, bud, leaf and root tissues [13], and the other analysing rootstock and leaf tissues [21]. Both studies investigated C19and C20-diterpenoid alkaloid biosynthetic pathways. In the study, we obtained leaf transcriptome data from two varieties of A. carmichaelii (Z175 and R184) to investigate potential differences in alkaloid biosynthesis and probe the physiological significance.
Leaf transcriptome data showed that the average length of unigenes was 923 bp, considerably longer than 595 bp [21] and 642 bp [13] reported previously. From the six leaf transcriptome datasets collected in triplicate in the present work, 108,477 transcripts were annotated, compared with 128,183 transcripts from transcriptome data from four tissues [13] and 95,812 transcripts from two tissues [21]. The comparative results show that our present transcriptome sequencing and annotation analysis is significantly deeper than these previous studies. According to the latest plant TF categories [23], 4,866 transcripts were assigned to 80 TF families, compared with only 644 transcripts assigned to 39 branches in previous work [21], but IF, MYB and C3H TFs were essentially the same.
A. carmichaelii is a medicinal plant of high therapeutic value due to the presence of alkaloids, among which C19-and C20-diterpenoid alkaloids are the predominant groups.
The active compounds in A. carmichaelii endow broad biological activities, including effects on the cardiovascular system, anti-inflammation and analgesic action, anti-tumor activity, immune system effects, hypoglycemic and hypolipidemic effects, anti-aging, kidney protection and energy metabolism [9]. In the biosynthesis of aconitine, both MVA and MEP biosynthetic pathways contribute to the biosynthesis of DAs in plants [24]. A total of 843 unigenes are involved in the biosynthesis of C19-diterpenoid alkaloids, including 22 unigenes associated with the MEP pathway and 34 unigenes related to the MVA pathway, considerably more than the number identified in a previous four-tissue study [13].
In the aconitine biosynthesis pathway, unigenes related to the isopentenyl diphosphate pathway were the most highly expressed in A. carmichaelii leaves, followed by those in the MEP and MVA pathways, similar to previous results [ 13]. Several studies have reported the tissue compartmentalization of precursors, intermediates and final products in highly specialized cells, as well as in different plant tissues [25,26], and A. carmichaelii may adopt a similar strategy for synthesizing aconitine-type DAs, which may explain the observed differences in the expression patterns of unigenes involved in different In the isopentenyl diphosphate pathway, seven aminotransferases, seven cyclases, two methyltransferase, two BAHD acyltransferases, one kaurene synthase and one ent-kaurene oxidase were among the top 20 most highly expressed, but the significance for regulating the biosynthesis of aconitine remains unknown. Additionally, unigenes annotated as GGPPS and CDPS were highly expressed in A. carmichaelii leaf tissue [13], consistent with GGPPS but not CDPS in the present work. Several studies report that CDPS is involved in tissuespecific accumulation of DAs [30, 31], but the reason for the low expression of CDPS observed in the present work needs to be further explored.
Salsolinol was detected in both Z175 and R184, and at a higher concentration in Z175.
Salsolinol biosynthesis in A. carmichaelii is derived from the biosynthesis of alkaloids generated by glycolysis (map00010), the shikimate pathway, phenylalanine, tyrosine, and tryptophan biosynthesis (map00400) and isoquinoline alkaloid biosynthesis (map00950). In the salsolinol biosynthesis pathway, genes encoding enzymes related to glycolysis were the most highly expressed, followed by shikimate and chorismate pathway genes.
The glycolysis pathway and its enzymes convert glucose to pyruvic acid using the oxidative potential of NAD+, and this pathway is among the most ancient molecular metabolic  [34]. Glycolytic enzymes have acquired additional non-glycolytic functions, as exemplified by HK, GAPHD and ENO, which function in transcriptional regulation, while glucose-6-phosphate isomerase (GPI) functions in the stimulation of cell motility, and HK and GAPHD regulate apoptosis [35].
The shikimate pathway is the key process by which aromatic amino acids are synthesized, and it is also important for salsolinol biosynthesis. Aromatic amino acids are synthesized via the shikimate pathway followed by the branched aromatic amino acid metabolic pathway, with chorismate serving as a major branch point intermediate metabolite [36].  [36]. Thus, high expression of aroF may be conducive to successful salsolinol biosynthesis. Indeed, aroF genes were abundant among the identified unigenes (22), and all four unigenes in Z175 were highly expressed, consistent with higher salsolinol content in Z175 than R184, and any of these four DEGs can be selected as target genes. AROB, the second enzyme in this pathway, converts 3-deoxy-d-arabino-heptulosonate-7phosphate into 3-dehydroquinate, both of which are highly expressed in A. carmichaelii.
The actions of AROD/AROE lead to shikimate, bifunctional activity has been characterized in Solanum lycopersicum [37], and the crystal structure of the Arabidopsis enzyme with shikimate bound at the AROE site and tartrate at the AROD site has recently been elucidated [38]. The enzyme encoded by AROK converts shikimate to shikimate 3phosphate. Interestingly, Oryza sativa AROK genes are differentially expressed in specific developmental stages and in response to biotic stress [39], and plant AROK activity is sensitive to changes in cellular ATP, suggesting that plant AROK acts as a regulatory factor in the shikimate pathway, facilitating metabolic flux toward specific pools of secondary metabolites [36]. The activity of AROA leads to the synthesis of enolpyruvylshikimate-3-phosphate (EPSP), and AROC converts EPSP to chorismate.
In chorismate pathway, there are seven CHMU, five PAT, seven TYRC, 23 POX and six TYDC enzymes that complete the biosynthesis route of salsolinol (Fig. 5)

Consent for publication
It is not applicable.

Availability of data and material
All supporting data is included within the article and its additional files.

Competing interests
No competing interests and financial competing interests. Tables Table 1 Unigenes involved in shikimat pathway of salsolinol biosynthesis pathway in A. carmichaelii.