Transcriptome profiling and comparative analysis of Panax ginseng adventitious roots

Abstract

Background: Panax ginseng Meyer is a traditional medicinal plant famous for its strong therapeutic effects and serves as an important herbal medicine. To understand and manipulate genes involved in secondary metabolic pathways including ginsenosides, transcriptome profiling of P. ginseng is essential.

Methods: RNA-seq analysis of adventitious roots of two P. ginseng cultivars, Chunpoong (CP) and Cheongsun (CS), was performed using the Illumina HiSeq platform. After transcripts were assembled, expression profiling was performed.

Results: Assemblies were generated from ∼85 million and ∼77 million high-quality reads from CP and CS cultivars, respectively. A total of 35,527 and 27,716 transcripts were obtained from the CP and CS assemblies, respectively. Annotation of the transcriptomes showed that approximately 90% of the transcripts had significant matches in public databases. We identified several candidate genes involved in ginsenoside biosynthesis. In addition, a large number of transcripts (17%) with different gene ontology designations were uniquely detected in adventitious roots compared to normal ginseng roots.

Conclusion: This study will provide a comprehensive insight into the transcriptome of ginseng adventitious roots, and a way for successful transcriptome analysis and profiling of resource plants with less genomic information. The transcriptome profiling data generated in this study are available in our newly created adventitious root transcriptome database (http://im-crop.snu.ac.kr/transdb/index.php) for public use.

Keywords: Panax ginseng; adventitious root; de novo assembly; next-generation sequencing; transcriptome.

Fig. 1

Adventitious roots of CP and CS at 2 weeks after inoculation in 2 L BTBBs. (A, B) Adventitious roots of CP are dark yellow, with callus-like morphology, (C, D) whereas those of CS show bright yellow, root hair-like morphology. (E) The rate of adventitious root induction from cotyledons. Data are represented as means with standard deviation from four independent experiments. BTBB, balloon-type bubble bioreactor; CP, Chunpoong; CS, Cheongsun.

Fig. 2

GO analysis of transcripts expressed in CP and CS adventitious roots. Terms in the (A) molecular function, (B) biological process, and (C) cellular component categories are shown. A total of 26,423 CP and 21,096 CS transcripts were assigned to GO terms using Blast2GO with whole assembled transcripts of CP (35,527) and CS (27,716). CP, Chunpoong; CS, Cheongsun; GO, gene ontology.

Fig. 3

RPKM values of transcripts expressed in CP and CS adventitious roots. RPKM values were calculated after reads of both cultivars were mapped onto the respective assembled transcripts. (A) Distribution of RPKM values. The number of transcripts belonging to the RPKM range defined on the X axis is shown at the bottom of each bar. Transcripts showing RPKM values of < 1.0 were included. (B) RPKM value comparison calculated by mapping reads of both cultivars onto CP transcripts as references. (C) RPKM value comparison calculated by mapping reads of both cultivars onto CS transcripts as references. Red spots indicate transcripts expressed in both cultivars. Regression lines [r² = 0.92 in (B) and r² = 0.94 in (C) at 99% confidential level] were calculated and depicted using Sigmaplot s/w. X and Y axes are in log-scale. CP, Chunpoong; CS, Cheongsun; RPKM, reads per kilobase per million.

Fig. 4

Functional distribution of the 100 most highly expressed transcripts in CP and CS adventitious roots. The top 100 transcripts were chosen on the basis of their expression level determined by RPKM values, and then biological processes in which those transcripts are putatively involved were predicted based on GO assignment and BLASTX searches. CP, Chunpoong; CS, Cheongsun; GO, gene ontology; RPKM, reads per kilobase per million.

Fig. 5

Putative ginsenoside biosynthesis pathway and expression level of ginsenoside biosynthesis genes found in transcript datasets of CP and CS adventitious roots. (A) Candidate genes identified in this study, which are shown in bold. (B) Expression levels of the candidate genes in both cultivars. Several isoforms of the candidate genes were identified, and RPKM values were averaged and represented with the color scale shown at the bottom. The expression map was generated using MeV s/w (http://www.tm4.org/mev/) with log₂ (RPKM) values. β-AS, beta-amyrin synthase; CP, Chunpoong; CS, Cheongsun; CYP716A47, cytochrome P450 for protopanaxadiol synthase; CYP716A53v2, cytochrome P450 for protopanaxatriol synthase; DDS, dammarenediol synthase; FPP, farnesyl diphosphate; FPS, farnesyl diphosphate synthase, GPP, geranyl diphosphate; RPKM, reads per kilobase per million; SQE, squalene epoxidase; SQS, squalene synthase; UGT, UDP-glycosyltransferase.

Fig. 6

Phylogenetic analysis of putative UGT proteins and comparison of their expression levels. (A) Twenty-one transcripts predicted to encode proteins with high similarity to UGT proteins involved in ginsenoside biosynthesis, SvUGT74M1 (ABK76266), MtUGT73K1 (AAW56091), and MtUGT71G1 (AAW56092), were identified in CP transcript dataset (Table 3). The deduced amino acid sequences were aligned using ClustalW, and the phylogenetic tree was generated using Poisson correction and the neighbor-joining method in MEGA5. Bootstrap values calculated for 1,000 replicates are shown on the branches; values < 50% are not shown. Average RPKM values of transcripts are shown in parentheses next to the transcript name. An Medicago truncatula homolog of SvUGT74M1, Medtr5g035580, was identified in the M. truncatula genome database (http://medicago.jcvi.org/cgi-bin/medicago/overview.cgi) and then included in the tree. (B) Expression comparison of the UGT genes in both cultivars. The RPKM values of transcripts closely grouped in the phylogenetic tree were compared and depicted by the expression map using MeV s/w (http://www.tm4.org/mev/) with log₂ (RPKM) values. CP, Chunpoong; CS, Cheongsun; RPKM, reads per kilobase per million; UGT, UDP-glycosyltransferase.

Fig. 7

GO analysis of transcripts common and unique to CP adventitious roots and normal roots. Terms are presented based on (A) molecular function, (B) biological process, and (C) cellular component. A total of 3,241 CP-unique and 2,378 root-unique transcripts were assigned GO terms using Blast2GO. GO assignment of transcripts common to CP adventitious roots and normal roots was analyzed for 23,675 common transcripts from the CP dataset. CP, Chunpoong; GO, gene ontology.