Comparative transcriptome analysis linked to key volatiles reveals molecular mechanisms of aroma compound biosynthesis in Prunus mume

Abstract

Background: Mei (Prunus mume) is the only woody plant in the genus Prunus with a floral fragrance, but the underlying mechanisms of aroma compound biosynthesis are unclear despite being a matter of considerable interest.

Results: The volatile contents of the petals of two cultivars with significantly different aromas, Prunus mume 'Xiao Lve' and Prunus mume 'Xiangxue Gongfen', were characterised by GC-MS at different flowering periods, and a total of 44 volatile compounds were detected. Among these, the main substances forming the typical aroma of P. mume were identified as eugenol, cinnamyl acetate, hexyl acetate and benzyl acetate, with variations in their relative concentrations leading to sensory differences in the aroma of the two cultivars. We compiled a transcriptome database at key stages of floral fragrance formation in the two cultivars and used it in combination with differential analysis of floral volatiles to construct a regulatory network for the biosynthesis of key aroma compounds. The results indicated that PmPAL enzymes and PmMYB4 transcription factors play important roles in regulating the accumulation of key biosynthetic precursors to these compounds. Cytochrome P450s and short-chain dehydrogenases/reductases might also influence the biosynthesis of benzyl acetate by regulating production of key precursors such as benzaldehyde and benzyl alcohol. Furthermore, by analogy to genes with verified functions in Arabidopsis, we predicted that three PmCAD genes, two 4CL genes, three CCR genes and two IGS genes all make important contributions to the synthesis of cinnamyl acetate and eugenol in P. mume. This analysis also suggested that the downstream genes PmBGLU18-like, PmUGT71A16 and PmUGT73C6 participate in regulation of the matrix-bound and volatile states of P. mume aroma compounds.

Conclusions: These findings present potential new anchor points for further exploration of floral aroma compound biosynthesis pathways in P. mume, and provide new insights into aroma induction and regulation mechanisms in woody plants.

Keywords: Floral scent; Gene expression and regulation; Prunus mume; Transcription factors; Transcriptome.

Fig. 1

a Different developmental stages of P. mume ‘Xiao lve’ (LE) and P. mume ‘Xiangxuegongfen’ (GF) flowers: (S2) petal baring period; (S3) early flowering period; (S4) full flowering period. b Heatmap analysis of volatile compounds of two P. mume cultivars at three flowering development stages. The colour of the heatmap ranges from dark blue (value, − 2.5) to magenta (value, + 2.5) on a natural logarithmic scale. c Amounts of the six main floral aroma volatile compounds during three different flowering stages of GF and LE. d Score plot of OPLS-DA model of headspace volatiles of P. mume cultivars with the statistical parameters: R²X = 0.999, R²Y = 0.994, Q² = 0.967

Fig. 2

a Number of detected transcripts in each sample. b Principal component analysis of RNA-Seq data. The PC1 coordinate represents the first principal component, and the percentage in brackets represents the level of contribution of the first principal component to the sample difference. Similarly, the PC2 coordinate represents the second principal component, with the percentage in brackets indicating the value of the contribution of this principal component to the sample difference. The coloured points correspond to distinct samples, as indicated in the legend. c Venn diagram of differentially expressed transcripts at the three flowering stages of GF and LE

Fig. 3

Unigene expression profiles during flowering of two P. mume cultivars. Three expression profiles were selected by gene screening: a profile7 (group I; top row), b profile0 (group II; middle row) and c profile5 (group III; bottom row), where profiles 7 and 0 denote upregulated and downregulated unigenes, respectively, and profile 5 denotes upregulated (from S2 to S3) then downregulated (from S3 to S4) unigenes. The nine different coloured lines denote absolute expression levels over the three flowering periods, with FPKM values of 0–0.1, 0.1–0.7, 0.7–2, 2–4, 4–8, 8–20, 20–100, 100–1000 and 1000–12,500 indicated successively by colours 1–9

Fig. 4

Distribution of DEGs, at different flowering stages of GF and LE, in enrichment pathways associated with major floral aroma compound anabolism

Fig. 5

Heatmap plot of DEGs identified by enrichment analysis at the different flowering stages, showing functional categories of significantly over-represented DEGs at S2 (a), S3 (b), S4 (c); and in LE vs. GF (d). The name of the pathway associated with each KEGG term are in Supplementary Table 4. The colour of the heatmap ranges from green (value, − 2) to red (value, + 2) on a natural logarithmic scale

Fig. 6

Coexpression network analysis for the three flowering stages of GF and LE. a Hierarchical cluster tree showing coexpression modules identified by WGCNA. Each leaf in the tree represents one gene. The major tree branches constitute 15 modules, labelled with different colours. b Trait association correlation analysis. Each row corresponds to a module and is marked with the colour corresponding to panel a, with the number of genes in each module shown on the left. Each column corresponds to a specific floral volatile compound. The colour of each module at the row–column intersection indicates the correlation coefficient and p-value between the module and the volatiles. The legend (top right) shows a colour scale for module trait correlation, from − 1 to + 1

Fig. 7

Proteins corresponding to genes highly relevant to the biosynthesis of floral aroma compounds in Prunus were screened by nucleotide sequence alignment with the related Arabidopsis proteins

Fig. 8

The key genes screened out by combining GO enrichment, KEGG PATHWAY enrichment [28] and WGCNA analysis were further analysed using Cytoscape. Nine key genes and MYB TFs were significantly enriched in a total of 18 metabolic processes

Fig. 9

Putative regulatory model between key enes and TFs related to the biosynthesis of floral compounds in GF and LE petals. Enzyme names, unigene IDs and expression patterns are indicated for each step. The expression pattern of each unigene is represented by a grid of six squares. From left to right, the first three squares represent the relative log₂(expression ratio) at the S2, S3 and S4 stages for LE; and the latter three squares represent the equivalent values for GF. The colour scale corresponds to log₂(expression ratio), as indicated

Fig. 10

qRT-PCR validation of genes related to flower aroma compound biosynthesis in GF and LE petals at all three flowering stages