Full-length transcriptome analysis of Phytolacca americana and its congener P. icosandra and gene expression normalization in three Phytolaccaceae species

Danfeng Liu,Qingjun Li,Yi Wang,Xingkui An,Chao Chen,Yongjun Zhang,Li Chen

doi:10.1186/s12870-020-02608-9

Abstract

BackgroundPhytolaccaceae species in China are not only ornamental plants but also perennial herbs that are closely related to human health. However, both large-scale full-length cDNA sequencing and reference gene validation of Phytolaccaceae members are still lacking. Therefore, single-molecule real-time sequencing technology was employed to generate full-length transcriptome in invasive Phytolacca americana and non-invasive exotic P. icosandra. Based on the transcriptome data, RT-qPCR was employed to evaluate the gene expression stability in the two plant species and another indigenous congener P. acinosa.ResultsTotal of 19.96 Gb and 19.75 Gb clean reads of P. americana and P. icosandra were generated, including 200,857 and 208,865 full length non-chimeric (FLNC) reads, respectively. Transcript clustering analysis of FLNC reads identified 89,082 and 98,448 consensus isoforms, including 86,989 and 96,764 high-quality ones. After removing redundant reads, 46,369 and 50,220 transcripts were obtained. Based on structure analysis, total 1675 and 1908 alternative splicing variants, 25,641 and 31,800 simple sequence repeats (SSR) as well as 34,971 and 36,841 complete coding sequences were detected separately. Furthermore, 3574 and 3833 lncRNA were predicted and 41,676 and 45,050 transcripts were annotated respectively. Subsequently, seven reference genes in the two plant species and a native species P. acinosa were selected and evaluated by RT-qPCR for gene expression analysis. When tested in different tissues (leaves, stems, roots and flowers), 18S rRNA showed the highest stability in P. americana, whether infested by Spodoptera litura or not. EF2 had the most stable expression in P. icosandra, while EF1-α was the most appropriate one when attacked by S. litura. EF1-α showed the highest stability in P.acinosa, whereas GAPDH was recommended when infested by S. litura. Moreover, EF1-α was the most stable one among the three plant species whenever germinating seeds or flowers only were considered.ConclusionFull-length transcriptome of P. americana and P. icosandra were produced individually. Based on the transcriptome data, the expression stability of seven candidate reference genes under different experimental conditions was evaluated. These results would facilitate further exploration of functional and comparative genomic studies in Phytolaccaceae and provide insights into invasion success of P. americana.

Highlights

Phytolaccaceae species in China are ornamental plants and perennial herbs that are closely related to human health
Compared to non-invasive exotic congener P. icosandra and native congener P. acinosa, P. americana is of interest because it exhibits multiple biological activities, such as plant pesticides, antimicrobial property, heavy metal accumulation capacity [2,3,4]
To classify the plant species, these three Phytolaccaceae members, P. americana, P. icosandra and P.acinosa were identified by PCR and followed by sequences alignment based on sequences of second internal transcribed spacer (ITS2) and the intergenic spacer of photosystem II protein D1 gene and tRNA-His gene of chloroplast genome (Table S1)

Summary

Introduction

Phytolaccaceae species in China are ornamental plants and perennial herbs that are closely related to human health. Both large-scale full-length cDNA sequencing and reference gene validation of Phytolaccaceae members are still lacking. Reports have showed that jasmonic acid-induced and cadmium-treated transcriptome data of P. americana have been obtained by Illumina HiSeq 2500 and Illumina HiSeq 2000 platform, respectively [5, 6]. These data were both achieved by second generation sequencing (SGS), which could not produce full-length transcripts. Genomic data of P. americana was available at the SRA under project PRJNA544344, but it’s raw reads without coding sequences prediction and functional annotation [7]

Methods

Results

Discussion

Conclusion