Molecular mechanisms of high levels of L-ascorbic acid accumulation in chestnut rose fruits

Abstract

The fruit of chestnut rose (Rosa roxburghii Tratt.) contains exceptionally high levels of L-ascorbic acid (AsA) (∼1762 mg/100 g fresh weight), approximately 40-fold higher than those found in sweet orange (Citrus sinensis), which is well known for its high AsA content. However, the molecular mechanisms driving such high accumulation in chestnut rose remain unclear. Here, we report that the genes R. roxburghiiPECTIN METHYLESTERASE (RroxPME), D-GALACTURONATE REDUCTASE (RroxGalUR), and DEHYDROASCORBATE REDUCTASE 2 (RroxDHAR2) play crucial roles in AsA accumulation in chestnut rose fruit. By comparing R. roxburghii with the closely related Rosamultiflora, which has low AsA concentrations, we identified a 545-bp insertion in the promoter of RroxGalUR. We found that TRANSPARENT TESTA GLABRA 2 (RroxTTG2), a well-known key regulator of trichome development, binds to the W-box-containing inserted region of the RroxGalUR promoter as well as the promoters of RroxPME and RroxDHAR2. In contrast, in sweet orange, CsTTG2 can bind only to CsPME. Furthermore, RroxTTG2 retains its conserved role in the regulation of trichome development during early fruit development, suggesting its spatiotemporal specificity in regulating both trichome development and AsA biosynthesis. To evaluate the application value of this pathway in other species, we heterologously expressed RroxTTG2, RroxPME, RroxGalUR, and RroxDHAR2 in lettuce (Lactuca sativa L.), which increased AsA concentrations in the transgenic lines by up to 355% (an increase from approximately 2 to 10 mg/100 g fresh weight). This study provides insights into mechanisms underlying AsA accumulation in chestnut rose and the spatiotemporal transcriptional regulation of AsA biosynthesis and trichome development.

Keywords: L-ascorbic acid; Rosa roxburghii; fruit; transcriptional regulation; trichomes.

Figure 1

AsA concentrations in fruit from Rosaceae and Rutaceae species and expression analysis of AsA biosynthesis genes in R. roxburghii and R. multiflora. (A) AsA concentrations in mature fruit of various Rosaceae and Rutaceae species. R. rox, R. roxburghii (AsA concentrations are averaged over 3 years); R. ch, R. chinensis Old Blush; R. mul, R. multiflora var. multiflora; C. si, C. sinensis cv. Tarocco; A. bu, A. buxifolia. Error bars represent ±SD (n = 3). Scale bars, 1 cm. (B) Trichome phenotypes of the R. roxburghii fruit surface at 30 days after flowering (DAF). Scale bars, 1 mm and 1 cm. (C) Expression of AsA metabolism-related genes in R. roxburghii and R. multiflora. I, L-galactose pathway; II, D-galacturonic acid pathway; III, L-gulose pathway; IV, myo-inositol pathway. Heatmaps show Fragments Per Kilobase of exon per Million mapped fragments (FPKM) values for gene expression in R. roxburghii and R. multiflora at four stages of fruit development (30, 50, 60, and 80 DAF; S1–S4) . Each heatmap tile shows the mean of three replicates. PG, polygalacturonase; PME, pectin methylesterase; GalUR, D-galacturonic acid reductase; GMP, GDP-D-mannose pyrophosphorylase; GME, GDP-D-mannose-3′,5′-epimerase; GGP, GDP-L-galactose phosphorylase; GalDH, L-galactose dehydrogenase; GalLDH, L-galactose-1,4-lactone dehydrogenase; AO, ascorbate oxidase; APX, ascorbate peroxidase; MDHAR, monodehydroascorbate reductase; DHAR, dehydroascorbate reductase; IPS, L-myo-inositol 1-phosphate synthase; MIOX, myo-inositol oxygenase; Alase, ATPase family AAA domain-containing protein.

Figure 2

Correlation analysis and functional assay of the candidate genes PME, GalUR, and DHAR2. (A and B) Correlation analysis of qPCR results and AsA concentrations for the PME, GalUR, and DHAR2 genes in R. roxburghii (A) and R. multiflora (B) at four stages of fruit development (Rrox_S1–Rrox_S4 and Rmul_S1–Rmul_S4). r and p, correlation and significance between gene expression levels and AsA concentrations during fruit development. p values were calculated using an unpaired two-sided t-test. Expression levels of PME, GalUR, and DHAR2 during R. roxburghii and R. multiflora fruit development and the exact p values for (A) and (B) are presented in Supplemental Table 8. (C) Overexpression of 35S::RroxPME, 35S::RroxGalUR, and 35S::RroxDHAR2 in transgenic lettuce. WT, Italian lettuce (L. sativa L. var. ramosa Hort.). Scale bars, 5 cm. (D) Measurement of AsA concentrations in transgenic lettuce. WT, Italian lettuce (L. sativa L. var. ramosa Hort.); 35S::RroxPME, 35S::RroxGalUR, and 35S::RroxDHAR2, transgenic lines overexpressing the respective genes; OE#, three independent overexpression lines per transgene. p values indicate significance. Error bars represent ±SD (n = 3).

Figure 3

The insertion in the RroxGalUR promoter is responsible for high AsA levels. (A) Schematic representation of variation in GalUR promoters among Rosaceae species closely related to R. roxburghii. In R. roxburghii, the GalUR promoter contains a 545-bp insertion, including a W-box and an LTR element. (B) Correlation analysis of RroxTTG2 expression levels and AsA levels during R. roxburghii fruit development. r values indicate Pearson correlation coefficients, and p values indicate significance. Error bars represent ± SD (n = 3). p values for RroxTTG2 expression levels during R. roxburghii fruit development were calculated using an unpaired two-sided t-test. Exact p values for (B) are presented in Supplemental Table 8. (C) Transient expression of the LUC/Renilla reporter gene in tobacco leaves with live-imaging confirming the ability of RroxTTG2 to activate the RroxGalUR promoter (RroxGalURpro). EV served as the negative control. Error bars represent ±SD (n = 3). p values were calculated using an unpaired two-sided t-test. (D) Yeast one-hybrid assay confirming the binding of RroxTTG2 to the RroxGalUR promoter (RroxGalURpro). AD-EV served as the negative control. 3-AT, 0 and 50 mmol/l. (E) EMSA confirming the binding of RroxTTG2 to the W-box element within the insertion of the RroxGalUR promoter. RroxTTG2-His, RroxTTG2-histidine (His) fusion protein; FAM-Probe, 5′6-FAM labeled probe; “+” indicates presence and “−" indicates absence of the added component.

Figure 4

Spatiotemporal expression of RroxTTG2 and functional assays of RroxTTG2 and RroxGL3. (A) Correlation analysis of RroxTTG2 and RroxGL3 gene expression with AsA concentrations at different stages of R. roxburghii fruit development. B_S1–B_S3, different developmental stages of flower buds (correlation analysis; r₁ and p₁). S1–S4, different stages of R. roxburghii fruit development (correlation analysis; r₂ and p₂). Scale bars, 5 mm. r values indicate Pearson correlation coefficients, and p values indicate significance. p values were calculated using an unpaired two-sided t-test for the expression levels of RroxGalUR, RroxTTG2, and RroxGL3 during R. roxburghii fruit development. Exact p values are presented in Supplemental Table 8. (B) Virus-induced gene silencing (VIGS) of RroxTTG2 reduces AsA concentrations in R. roxburghii fruit. EV, negative control; TRV, RroxTTG2 VIGS. Error bars represent ±SD (n = 3). (C) Observation of trichomes in leaves from transgenic A. thaliana overexpressing RroxTTG2 and RroxGL3. Scale bars, 1 cm. (D and E) Analysis of leaf AsA concentrations in transgenic A. thaliana overexpressing RroxTTG2 and RroxGL3. Error bars represent ±SD (n = 3). (F and G) Analysis of leaf trichomes in transgenic A. thaliana overexpressing RroxTTG2 and RroxGL3. Error bars represent ±SD (n = 5). (C–G) Col-0, WT A. thaliana; RroxTTG2/Col-0, Col-0 overexpressing RroxTTG2; Atttg2, A. thaliana ttg2 mutant; RroxTTG2/Atttg2, ttg2 mutant complementation line overexpressing RroxTTG2; RroxGL3/Col-0, Col-0 overexpressing RroxGL3; Atgl3, A. thaliana gl3 mutant; RroxGL3/Atgl3, gl3 mutant complementation line overexpressing RroxGL3. OE#, three independent lines.

Figure 5

Heterologous overexpression of RroxTTG2-RroxPME-RroxGalUR-RroxDHAR2 in lettuce. (A) Construction of multi-gene aggregation vectors and generation of transgenic lettuce. WT, wild-type lettuce control. “+,” positive plasmid. #1, #6, and #7 represent three independent positive transgenic lines expressing all four genes (RroxTTG2-RroxPME-RroxGalUR-RroxDHAR2) from the 4TU construct. 4TU, multi-gene construct driven by the Cauliflower Mosaic Virus 35S promoter (35S) and the Arabidopsis polyubiquitin 10 (UBQ10) promoter. Scale bars, 5 cm. (B) Gene expression levels in transgenic lettuce. OE#, three independent lines overexpressing the transgene. Error bars represent ±SD (n = 3). (C) Analysis of AsA concentrations in lettuce leaves. p values were calculated using a two-tailed Student’s t-test. Error bars represent ±SD (n = 3). OE#, three independent lines. p values were calculated using an unpaired two-sided t-test. Expression levels of RroxPME, RroxGalUR, RroxDHAR2, and RroxTTG2 during R. roxburghii fruit development and exact p values from (B) are presented in Supplemental Table 8.

Figure 6

Molecular regulatory mechanism underlying the high AsA accumulation in R. roxburghii and sweet orange. (A) Spatiotemporal differential expression of TTG2 is involved in the regulation of both trichome development and AsA accumulation, with a number of differentially regulated downstream genes contributing to AsA accumulation. (B) Schematic diagram showing how the variation in the upstream promoter of the RroxGalUR gene in R. roxburghii (insertion of a W-box-binding element) influences differential AsA accumulation in various species.