Multifunctional oxidosqualene cyclases and cytochrome P450 involved in the biosynthesis of apple fruit triterpenic acids

Abstract

Apple (Malus × domestica) accumulates bioactive ursane-, oleanane-, and lupane-type triterpenes in its fruit cuticle, but their biosynthetic pathway is still poorly understood. We used a homology-based approach to identify and functionally characterize two new oxidosqualene cyclases (MdOSC4 and MdOSC5) and one cytochrome P450 (CYP716A175). The gene expression patterns of these enzymes and of previously described oxidosqualene cyclases were further studied in 20 apple cultivars with contrasting triterpene profiles. MdOSC4 encodes a multifunctional oxidosqualene cyclase producing an oleanane-type triterpene, putatively identified as germanicol, as well as β-amyrin and lupeol, in the proportion 82 : 14 : 4. MdOSC5 cyclizes 2,3-oxidosqualene into lupeol and β-amyrin at a ratio of 95 : 5. CYP716A175 catalyses the C-28 oxidation of α-amyrin, β-amyrin, lupeol and germanicol, producing ursolic acid, oleanolic acid, betulinic acid, and putatively morolic acid. The gene expression of MdOSC1 was linked to the concentrations of ursolic and oleanolic acid, whereas the expression of MdOSC5 was correlated with the concentrations of betulinic acid and its caffeate derivatives. Two new multifuntional triterpene synthases as well as a multifunctional triterpene C-28 oxidase were identified in Malus × domestica. This study also suggests that MdOSC1 and MdOSC5 are key genes in apple fruit triterpene biosynthesis.

Keywords: Malus × domestica; apple; betulinic acid; cytochrome P450; germanicol; oxydosqualene cyclase; triterpene; ursolic acid.

Figure 1

Biosynthetic pathway of triterpenic acids in apple (Malus × domestica). Triterpenes are synthesized via the mevalonic acid (MVA) pathway. The enzymes that catalyse the various steps are indicated in blue. IPP, isopentenyl diphosphate; DMAPP, dimethylallyl diphosphate; FPP, farnesyl diphosphate; OSC, oxidosqualene cyclase; CYP450, cytochrome P450. MdOSCs are underlined when they catalyse the production of the specific triterpene considered as a major component. CYP716A175 catalyses the three‐step oxidation of triterpene backbones at the C‐28 position.

Figure 2

Alignment of oxidosqualene cyclase (OSC) predicted amino acid sequences from apple (Malus × domestica). The conserved DCTAE as well as the M(W/Y)CY(C/S)R sequences are shown in bold (residues 485 onwards and 256 onwards, respectively). Six of the QW motifs (Q‐X3‐G‐X‐W) (Siedenburg & Jendrossek, 2011) are underlined. The text/background colour code refers to the degree of similarity between the different sequences: red/yellow, identical; dark blue/turquoise, conservative; black/green, a block of similar sequences; black/white, nonsimilar.

Figure 3

Phylogenetic tree of a wide range of triterpene synthase amino acid sequences, built using the neighbour‐joining method. Sequences were selected from GenBank based on their authentication in the literature (unless otherwise indicated). The scale bar indicates 0.2 amino acid substitutions per site. The GenBank or The Arabidopsis Information Resource (TAIR) identifier is included in the gene name. The two‐letter prefix for each name in the tree identifies the species name as follows: Ab, Abies magnifica; Ae, Aralia elata; As, Avena strigosa; Bg, Bruguiera gymnorhiza; Bp, Betula platyphylla; Ca, Centella asiatica; Cp, Cucurbita pepo; Cs, Costus speciosus; Et, Euphorbia tirucalli; Gg, Glycyrrhiza glabra; Kc, Kandelia candel; Kd, Kalanchoe daigremontiana; Lc, Luffa cylindrical; Lj, Lotus japonicas; Mt, Medicago truncatula; Oe, Olea europaea; Pg, Panax ginseng; Rc, Ricinus communis; Rs, Rhizophora stylosa; To, Taraxacum officinale. Measured triterpene activities are encoded in the suffix for each entry name as follows: aAS, alpha amyrin synthase; AraS, arabidiol synthase; bAS, beta‐amyrin synthase; BS, baruol synthase; CAMS, camelliol c synthase; CAS, cycloartenol synthase; CDS, cucurbitadienol synthase; DAS, dammarenediol‐II synthase; IMFS, isomultiflorenol synthase; LAS, lanosterol synthase; LUS, lupeol synthase; mAS, mixed amyrin synthase (both alpha and beta amyrin); MS, marneral synthase; mTS or TS, multifunctional terpene synthase; pbAS, putative bAS; ThS, thalianol synthase. Multifunctional triterpene synthases within monofunctional triterpene synthase groups are marked with a bullet.

Figure 4

Triterpene synthesis in tobacco leaves by heterologous expression of candidate triterpene synthases. (a) Chromatograms of typical LC‐MS analysis of the products of (iii) Malus × domestica OXIDOSQUALENE CYCLASE4 (MdOSC4), (iv) MdOSC5, and (v) MdOSC1 after transient expression in Nicotiana benthamiana. p19 was used as a negative control (ii) and mixed with authentic standards at 20 μg ml⁻¹ (1, lupeol; 2, β‐amyrin; 3, α‐amyrin) (i). Peak 4 on trace (iii) is the putative identification of germanicol. Compounds were identified and quantified on the basis of their mass spectral data (Supporting Information Figs S3, S4). Chromatograms are presented as selected ion plots of the m/z 409.8 [MH‐H₂O]⁺ ion. (b) Quantification of triterpene production by the various triterpene synthase genes in tobacco. Data show nmoles of triterpenes produced from 100 mg of extracted tissue (mean with SE; n = 4). Germanicol was quantified as β‐amyrin equivalents.

Figure 5

Neighbour‐joining tree of a wide range of triterpene‐related P450 amino acid sequences. Note that the apple cytochrome P450 (CYP716A175) is close to the Medicago truncatula authenticated triterpene C‐28 oxidase (CYP716A12). The prefix is the GenBank or database identifier. The scale bar indicates 2 amino acid substitutions per site. P450s are clustered according to the P450 clan they belong to.

Figure 6

Triterpene acid synthesis in tobacco leaves by heterologous coexpression of a cytochrome P450 (CYP) gene with triterpene synthases. (a) Chromatograms of typical LC‐MS traces of the products of CYP716A175 (P450) coexpressed with (v) Malus × domestica OXIDOSQUALENE CYCLASE4 (MdOSC4), (vi) MdOSC5, and (vii) MdOSC1. p19 was used as a negative control (iv) and compared with authentic standards of (i) betulinic acid (1) at 100 μg ml⁻¹; (ii) oleanolic acid (2) at 100 μg ml⁻¹, and (iii) ursolic acid (3) at 100 μg ml⁻¹. (4) Putative identification of morolic acid. Compounds were identified and quantified on the basis of their mass spectral data (Supporting Information Fig. S6). Chromatograms are presented as selected ion plots of the sum of three ions: m/z 474.6 [M+NH ₄]⁺, m/z 456.6 [M]⁺ and m/z 439.8 [MH‐H₂O]⁺. (b) Quantification of the amount of triterpene acids made by the addition of a P450 gene to the triterpene synthases. Morolic acid was quantified as oleanolic acid equivalent. Data are nmoles of triterpenes produced from 100 mg of extracted tissue (mean ± SE; n = 4).

Figure 7

Principal component analysis (PCA) performed on 20 individuals (apple cultivars) and six triterpene concentrations. (a) Score plot of the 20 cultivars separating into three russeting groups: russeted, semi‐russeted, and waxy. (b) Loading plot showing the relationships among phytochemical data. UA, ursolic acid; BA, betulinic acid; BA‐cisC, betulinic acid‐3‐cis‐caffeate; BA‐transC, betulinic acide‐3‐trans‐caffeate.

Figure 8

Expression analysis of triterpene biosynthetic genes (a, Malus × domestica OXIDOSQUALENE CYCLASE1 (MdOSC1); b, MdOSC3; c, MdOSC4; d, MdOSC5; e, CYP716A175) by real‐time quantitative (q)PCR on RNA extracted from apple skin tissues. The normalized relative expression was rescaled to the sample with the lowest relative expression, that is, the expression level was set to 1 in the cultivar ‘Topaz’ for MdOSC4. Error bars show SD (n = 3).