De novo assembly, functional annotation and comparative analysis of Withania somnifera leaf and root transcriptomes to identify putative genes involved in the withanolides biosynthesis

Abstract

Withania somnifera is one of the most valuable medicinal plants used in Ayurvedic and other indigenous medicine systems due to bioactive molecules known as withanolides. As genomic information regarding this plant is very limited, little information is available about biosynthesis of withanolides. To facilitate the basic understanding about the withanolide biosynthesis pathways, we performed transcriptome sequencing for Withania leaf (101L) and root (101R) which specifically synthesize withaferin A and withanolide A, respectively. Pyrosequencing yielded 8,34,068 and 7,21,755 reads which got assembled into 89,548 and 1,14,814 unique sequences from 101L and 101R, respectively. A total of 47,885 (101L) and 54,123 (101R) could be annotated using TAIR10, NR, tomato and potato databases. Gene Ontology and KEGG analyses provided a detailed view of all the enzymes involved in withanolide backbone synthesis. Our analysis identified members of cytochrome P450, glycosyltransferase and methyltransferase gene families with unique presence or differential expression in leaf and root and might be involved in synthesis of tissue-specific withanolides. We also detected simple sequence repeats (SSRs) in transcriptome data for use in future genetic studies. Comprehensive sequence resource developed for Withania, in this study, will help to elucidate biosynthetic pathway for tissue-specific synthesis of secondary plant products in non-model plant organisms as well as will be helpful in developing strategies for enhanced biosynthesis of withanolides through biotechnological approaches.

Figure 1. Size distribution of 454 high-quality reads (A) and assembled contigs (B).

Black and white bars represents reads and contigs from leaf (101L) and root (101R) transcriptome data respectively.

Figure 2. The number of unigenes of 101L (A) and 101R (B) transcriptomes annotated using TAIR10, NR, tomato and potato databases.

Figure 3. Histogram of gene ontology classification.

The results are summarized in three main categories: biological process, cellular component and molecular function. Bars represent the percent number of assignments of leaf (101L; white bars) and root (101R; black bars) unigenes with BLAST matches in the TAIR10 database to each GO term.

Figure 4. Size distribution of assembled contigs (A) and annotation (B).

Raw reads from 101L and 101R were tagged and assembled together. Contigs and singletons generated from combined assembly were annotated using TAIR10, NR, tomato and potato databases.

Figure 5. Putative pathway for withanolide biosynthesis in Withania somnifera.

All the enzymes found in this study related to different steps are shown between the reactions catalyzed. Expression of different contigs related to these enzymes in leaf and root is shown by heatmap. Broken arrow in Step 3 represents putative steps involving CYP450s, GTs and MTs. Names of the enzymes identified are provided in Table 4.

Figure 6. Clusters containing cytochrome P450 with their differential expression in leaf and root.

Two columns represent leaf and root, while each row represents contigs encoding different members of CYP450 gene family (Table S12). Clustering was carried out with log₂tpm value of each contig in leaf and root transcriptome to visualize differential expression.

Figure 7. Validation of differentially expressed CYP450 genes in 101L and 101R.

Quantitative Real time PCR (A) of selected differentially expressed CYP450s was carried out using total RNA isolated from leaf and root tissues. Digital gene expression (B) of CYP450s used for validating through quantitative Real time PCR.