Identification of genes associated with ricinoleic acid accumulation in Hiptage benghalensis via transcriptome analysis

Abstract

Background: Ricinoleic acid is a high-value hydroxy fatty acid with broad industrial applications. Hiptage benghalensis seed oil contains a high amount of ricinoleic acid (~ 80%) and represents an emerging source of this unusual fatty acid. However, the mechanism of ricinoleic acid accumulation in H. benghalensis is yet to be explored at the molecular level, which hampers the exploration of its potential in ricinoleic acid production.

Results: To explore the molecular mechanism of ricinoleic acid biosynthesis and regulation, H. benghalensis seeds were harvested at five developing stages (13, 16, 19, 22, and 25 days after pollination) for lipid analysis. The results revealed that the rapid accumulation of ricinoleic acid occurred at the early-mid-seed development stages (16-22 days after pollination). Subsequently, the gene transcription profiles of the developing seeds were characterized via a comprehensive transcriptome analysis with second-generation sequencing and single-molecule real-time sequencing. Differential expression patterns were identified in 12,555 transcripts, including 71 enzymes in lipid metabolic pathways, 246 putative transcription factors (TFs) and 124 long noncoding RNAs (lncRNAs). Twelve genes involved in diverse lipid metabolism pathways, including fatty acid biosynthesis and modification (hydroxylation), lipid traffic, triacylglycerol assembly, acyl editing and oil-body formation, displayed high expression levels and consistent expression patterns with ricinoleic acid accumulation in the developing seeds, suggesting their primary roles in ricinoleic acid production. Subsequent co-expression network analysis identified 57 TFs and 35 lncRNAs, which are putatively involved in the regulation of ricinoleic acid biosynthesis. The transcriptome data were further validated by analyzing the expression profiles of key enzyme-encoding genes, TFs and lncRNAs with quantitative real-time PCR. Finally, a network of genes associated with ricinoleic acid accumulation in H. benghalensis was established.

Conclusions: This study was the first step toward the understating of the molecular mechanisms of ricinoleic acid biosynthesis and oil accumulation in H. benghalensis seeds and identified a pool of novel genes regulating ricinoleic acid accumulation. The results set a foundation for developing H. benghalensis into a novel ricinoleic acid feedstock at the transcriptomic level and provided valuable candidate genes for improving ricinoleic acid production in other plants.

Keywords: Co-expression network analysis; Hiptage benghalensis; Industrial oils; Lipid biosynthesis; Long noncoding RNA; Oilseed; RNA-Seq; Ricinoleic acid; Transcription factor; Transcriptomics.

Fig. 1

H. benghalensis seed development and lipid accumulation. a The developmental progress of H. benghalensis seeds (stages S1–S6). The samaras were harvested at 13 days after pollination (S1), and then every 3 days until 28 days after pollination (S6, mature seeds). b The oil content of H. benghalensis developing and mature seeds. c The composition of the six major fatty acids in H. benghalensis developing and mature seeds (mean ± SD, n = 3)

Fig. 2

Functional annotation of differentially expressed genes (DEGs) at different seed development stages. a GO enrichment analysis of DEGs. Genes were assigned into three main categories: biological processes, cellular components or molecular functions. The y-axis indicates the number of genes in a given category. b Scatterplot of GO biological process involved in lipid metabolism. c Histogram of cluster of KEGG pathways of DEGs. The results were summarized in five main categories (black words). d Scatterplot of KEGG pathway involved in lipid metabolism. Rich Factor refers to the ratio of the differentially expressed gene number and the number of genes annotated in this pathway and large Rich Factor indicates high degree of enrichment. The area of each colored circle is proportional to the number of genes involved in each pathway, the color indicated the p value, and the x-axis is the Rich Factor

Fig. 3

Cluster analysis of the differentially expressed genes (DEGs) in H. benghalensis seeds. a Hierarchical clustering dendrogram of the DEGs. The red highlight indicates genes were highly expressed whereas the blue highlight indicates genes were low expressed. The z score indicates genes expression values. b The three cluster groups of different gene expression patterns

Fig. 4

Co-expression networks. a The co-expression network between lipid-related genes and transcript factors (TFs). b The co-expression network between lipid-related genes and long noncoding RNAs (lncRNAs). The blue circle nodes indicate the lipid-related genes, the brown arrow nodes indicate the TFs, and the brown triangle nodes indicate the lncRNAs. The symbol size indicates the degree of nodes. The gray lines indicate positive co-expression, and the red lines indicate negative co-expression

Fig. 5

Validation of the transcriptomic data with quantitative RT-PCR. Eighteen genes were used in this validation. Refer to Tables 1 and 2 for the detail information on the selected genes. The comparative FPKM and 2^−ΔΔCt at stage S1 were used as the control for normalization. Results represent the mean of three biological replicates (mean ± SD, n = 3)

Fig. 6

Proposed gene networks involved in hydroxy fatty acids and triacylglycerol biosynthesis in H. benghalensis seeds. The expression levels (represented by Log2FPKM) of the possible candidates are highlighted in color scales (blue to red scale) in H. benghalensis developing seeds at different development stages (S1–S5). ABI3 abscisic acid insensitive 3, ACBP acyl CoA-binding protein, ACP acyl carrier protein, ACCase-α acetyl-CoA carboxylase α-carboxyltransferase, bZIP basic region/leucine zipper motif, CoA coenzyme A, DAG diacylglycerol, DGAT diacylglycerol acyltransferase, EAR enoyl-ACP reductase, ER endoplasmic reticulum, FAD fatty acid desaturase, FAH12 oleate-12-hydroxylase, FAT acyl-ACP thioesterase, FFA free fatty acid, FPKM Fragments per Kilobase of transcript per Million mapped reads, HFA hydroxy fatty acid, GPAT9 sn-glycerol-3-phosphate acyltransferase, G3P sn-glycerol-3-phosphate, KAR ketoacyl-ACP reductase, KAS ketoacyl-ACP synthase, LACS long-chain acyl-CoA synthase, LEC1 leafy cotyledon1, lncRNA long noncoding RNA, LPA lysophosphatidic acid, LPAAT acyl-CoA:lysophosphatidic acid acyltransferase, LPC lysophosphatidylcholine, LPCAT lysophosphatidylcholine acyltransferase, NPC non-specific phospholipase C, OLE oleosin, PA phosphatidic acid, PAP phosphatidic acid phosphatase, PC phosphatidylcholine, PDAT phospholipid:diacylglycerol acyltransferase, PDCT phosphatidylcholine:diacylglycerol cholinephosphotransferase, PLA₂ phospholipase A₂, PLC phospholipase C, PLD phospholipase D, PXG peroxygenase, SAD stearoyl-ACP desaturase, TAG triacylglycerol, TF transcription factor, TZF tandem CCCH zinc finger, WRI1 wrinkled1. This model was development based on the transcriptome data of this study and information from Block and Jouhet [64], Du et al. [65], Li-Beisson et al. [66], and Xu et al. [51]