An Independent Evolutionary Origin for Insect Deterrent Cucurbitacins in Iberis amara

Abstract

Pieris rapae and Phyllotreta nemorum are Brassicaceae specialists, but do not feed on Iberis amara spp. that contain cucurbitacins. The cucurbitacins are highly oxygenated triterpenoid, occurring widespread in cucurbitaceous species and in a few other plant families. Using de novo assembled transcriptomics from I. amara, gene co-expression analysis and comparative genomics, we unraveled the evolutionary origin of the insect deterrent cucurbitacins in I. amara. Phylogenetic analysis of five oxidosqualene cyclases and heterologous expression allowed us to identify the first committed enzyme in cucurbitacin biosynthesis in I. amara, cucurbitadienol synthase (IaCPQ). In addition, two species-specific cytochrome P450s (CYP708A16 and CYP708A15) were identified that catalyze the unique C16 and C22 hydroxylation of the cucurbitadienol backbone, enzymatic steps that have not been reported before. Furthermore, the draft genome assembly of I. amara showed that the IaCPQ was localized to the same scaffold together with CYP708A15 but spanning over 100 kb, this contrasts with the highly organized cucurbitacin gene cluster in the cucurbits. These results reveal that cucurbitacin biosynthesis has evolved convergently via different biosynthetic routes in different families rather than through divergence from an ancestral pathway. This study thus provides new insight into the mechanism of recurrent evolution and diversification of a plant defensive chemical.

Keywords: biosynthesis; cucurbitacin; cytochrome P450; evolution; triterpenoid.

Fig. 1.

Distribution of cucurbitacins across families in the Eukaryota and their accumulation pattern in a brassicaceous and a cucurbitaceous species. (A) Phylogenetic relationships in Eukaryota indicating multiple origins of cucurbitacin biosynthesis. The picture was adapted according to Stevens, P. F. (2001 onwards). Angiosperm Phylogeny Website. Version July 14, 2017 (http://www.mobot.org/MOBOT/research/APweb/). Note, only the major cucurbitacins in each order are shown color-coded. (B) LC-ESI-MS extracted ion chromatograms (EIC) of extracts from roots, leaves, stems, flower buds, sepals, petals, pistils, and stamen of 8-week-old Iberis amara plants. (C) LC-ESI-MS extracted ion chromatograms (EIC) of extracts from Cucurbita pepo roots, leaves, stems, male flowers, stamen, female flowers, pistils, and fruits. The extracted ions at m/z 539, 537, 581, and 579 are representative of sodium adducts of cucurbitacin D, I, B, E, respectively. Peaks indicated by blue, gray, and black arrows were verified by comparing the corresponding mass spectra to those of authentic cucurbitacin I, B, and E standards, respectively.

Fig. 2.

Identification and characterization of Iberis amara cucurbitadienol synthase. (A) Total ion current (TIC) GC-MS chromatograms of extracts from an empty vector (EV) control yeast strain (black) and a strain expressing IaCPQ (red), and of purified cucurbitadienol as a standard (blue). The peak indicated with a black arrow at 21.9 min corresponds to cucurbitadienol. (B) TIC GC-MS chromatograms of extracts from Nicotiana benthamiana leaves agro-infiltrated with an EV control, IaCPQ, ClCPQ, CpCPQ, and CpSE2 with IaCPQ, and the chromatogram of the cucurbitadienol standard. Black arrows indicate the cucurbitadienol peak. (C) TIC GC-MS chromatograms of extracts from an EV control yeast strain (black) and a strain expressing IaCAS (red), and the chromatogram of a cycloartenol standard (blue). The peak indicated with a black arrow eluting at 26.2 min corresponds to cycloartenol. The second peak indicated with a green arrow eluting at 24.5 min likely corresponds to 31-norcycloartenol. Note, the GC-MS analysis program for yeast samples differs slightly from the one used for GC-MS analysis for plant extracts, giving rise to differences in retention times.

Fig. 3.

Iberis amara cucurbitadienol synthase evolved from I. amara cycloartenol synthase. The evolutionary history of OSCs was inferred by using the maximum likelihood method to build a rooted tree. About 67 OSCs nucleotide sequences were selected so they broadly represent OSCs from species with known product profiles. The coding sequences were aligned based on codons. Branches of CASs and CPQs were labeled with ω values. Branches were color coded corresponding to their bootstrap value. Predicted gene duplications were identified by searching for all branching points in the topology with at least one species that is present in both subtrees of the branching point. D1–D17 indicate duplicate events. The scale bar represents the number of substitution sites corresponding to the branch length. Evolutionary analyses were conducted in MEGA X. Note, some species pictures on the tree represent the cucurbitaceous species, brassicaceous specious, fabaceous species, Arabidopsis species, monocot species, lower plants. Only pictures for I. amara, Citrullus lanatus, Ricinus communi, Aster tataricus represent single species. Gene names from I. amara were coded with light red background. CAS names from cucurbitacecous species were colored with white background. Lj, Lotus japonicus; Ps, Pisum sativum; As, Avena strigose; Rc, Ricinus communi; Ast, Aster tataricus; Rhs, Rhizophora stylosa; Mt, Medicago truncatula; Sg, Siraitia grosvenorii; Mc, Momordica charantia; Cp, Cucurbita pepo; Cl, Citrullus lanatus; Cs, Cucumis sativus; Cm, Cucumis melo; Cr, Capsella rubella; Rs, Raphanus sativus; At, Arabidopsis thaliana; Al, Arabidopsis lyrate; Br, Brassica rapa; Bo, Brassica oleracea; Es, Eutrema salsugineum; Cs, Camelina sativa; Ia, I. amara; Os, Oryza sativa; Sb, Sorghum bicolor; Zm, Zea mays; Pp, Physcomitrella patens; Cr, Chlamydomonas reinhardtii. IAS1, isoarborinol synthase 1, ABAS, mixed α- and β-amyrin synthase; ABS, achilleol B synthase; MSS, mixed simiarenol synthase; PKS, parkeol synthase; ABS, arabidiol synthase; SAS, seco-amyrin synthase; TRS, Tirucalladienol synthase; OSC, oxidosqualene synthase; PEN1, pentacyclic triterpene synthase 1; PEN6, pentacyclic triterpene synthase 6; PEN3, pentacyclic triterpene synthase 3; THAS1, thalianol synthase 1; BARS1, baruol synthase 1; MRN1, marneral synthase 1; LUP, Lupeol synthase; CPR, putative oxidosqualene cyclase; CPX, cycloartenol synthase; CPQ, cucurbitadienol synthase; CAS, cycloartenol synthase; LAS, lanosterol synthase; AMY2, mixed β-amyrin synthase; MbAS, multifunctional β-amyrin synthase; bAS, β-amyrin synthase.

Fig. 4.

Iberis amara cucurbitadienol synthase is expressed in all plant organs tested and is a monotopic membrane protein. (A) Relative expression of IaCPQ in different I. amara organs quantified by qPCR. (B–R) Representative confocal images of the epidermal cell space ([B], [C], [F], [G], [J], [K], [O], [P]) and the nucleus ([D], [E], [H], [I], [M], [N], [Q], [R]) of Nicotiana benthamiana leaves expressing C-terminally tagged proteins. mRFP1 and CYP98A1 (P450) are used as controls for cytosolic and ER localization, respectively. Soluble mRFP1 is found in the nucleus and fills in gaps between plant cell compartments. ER-localized proteins (CpCPQ and CYP98A1) are restrained to nuclear membranes and a well-defined ER membrane network. IaCPQ is observed on the nuclear membrane as a monotopic membrane protein, however, excess protein appears to permeate the nucleus. Similar images were obtained with N-terminal fluorescent fusion constructs for IaCPQ and CpCPQ (supplementary fig. 6B, Supplementary Material online). Scale bars=10 µm.

Fig. 5.

CYP708A16 belongs to the CYP708 family and catalyzes the conversion of cucurbitadienol to 16β-hydroxy-cucurbitadienol. (A) Maximum likelihood phylogenetic tree of Iberis amara and selected Arabidopsis thaliana P450 amino acid sequences. Bootstrap values in % are shown at the branch points. The tree was constructed with MEGA X and depicted using the online tool iTOL v5. (https://itol.embl.de/). The assembled 25 full-length I. amara P450s (MW514544, MW514545, MW514546, MW514547, MW514552, MW514553, MW514554, MW514555, MW514557, MW514558, MW514559, MW514560, MW514561, MW514562, MW514563, MW514564, MW514565, MW514566, MW514567, MW514568, MW514569, MW514570, MW514571, MW514572) from the Miseq data and additional full-length CYP708As (MW514548, MW514549, MW514550, MW514551, MW514556) assembled from the Hiseq data are labeled with red dots. Cucurbitacin biosynthetic related P450 genes from Citrullus lanatus (blue dot), Cucumis sativus (green dot), Cucumis melo (orange dot) (Zhou et al. 2016), and selected P450 sequences from A. thaliana are also included in the tree. (B) GC-MS chromatograms of the cucurbitadienol standard and extracts from Nicotiana benthamiana leaves agro-infiltrated with an EV control and vectors expressing IaCPQ, CYP708A16, and IaCPQ together with CYP708A16. Peaks in the gray box indicate cucurbitadienol. The red arrow indicates the unique peak in the extract of the IaCPQ and CYP708A16 co-expressing leaves. (C) TIC GC-MS of extracts from yeast strains expressing McCPQ, and McCPQ with CYP708A16. Cucurbitadienol is indicated with a black arrow. The peak of 16β-hydroxy-cucurbitadienol is indicated with a red arrow. Note, the GC-MS analysis program for yeast samples differs slightly from the one used for GC-MS analysis for plant extracts, giving rise to the observed differences in retention times.

Fig. 6.

IaCYP708A15v2 catalyzes 22-hydroxylation in the cucurbitacin biosynthesis. (A) Hierarchical cluster heat map of expression patterns of the 287 P450 contigs and five oxidosqualene cyclases expressed in the roots, petals, and stems of Iberis amara. Putative biosynthetic P450 genes involved in cucurbitacin biosynthesis are marked by a yellow box. (B) Doughnut chart was drawn based on the exact number of genes from each P450 subfamily of I. amara (outer doughnut) and Arabidopsis thaliana (inner doughnut). Each subfamily was color coded. In particular, the CYP72A, CYP708A, CYP81D, and CYP707A subfamilies were expanded in I. amara as compared with A. thaliana. (C) GC-MS chromatograms of extracts from Nicotiana benthamiana leaves agro-infiltrated with EV control and vectors expressing IaCPQ, IaCPQ+CYP708A16, IaCPQ+CYP708A15v2, IaCPQ+CYP708A16+CYP708A15v2, and CYP708A16+CYP708A15v2. Cucurbitadienol and 16β-hydroxy-cucurbitadienol are indicated with black arrows. The new peak that appeared only in the extract from IaCPQ+CYP708A16+CYP708A15v2 expressing leaves is indicated by a red arrow.