An algal enzyme required for biosynthesis of the most abundant marine carotenoids

Abstract

Fucoxanthin and its derivatives are the main light-harvesting pigments in the photosynthetic apparatus of many chromalveolate algae and represent the most abundant carotenoids in the world's oceans, thus being major facilitators of marine primary production. A central step in fucoxanthin biosynthesis that has been elusive so far is the conversion of violaxanthin to neoxanthin. Here, we show that in chromalveolates, this reaction is catalyzed by violaxanthin de-epoxidase-like (VDL) proteins and that VDL is also involved in the formation of other light-harvesting carotenoids such as peridinin or vaucheriaxanthin. VDL is closely related to the photoprotective enzyme violaxanthin de-epoxidase that operates in plants and most algae, revealing that in major phyla of marine algae, an ancient gene duplication triggered the evolution of carotenoid functions beyond photoprotection toward light harvesting.

Fig. 1. A vdl knockout mutant of N. oceanica no longer synthesizes the allenic vaucheriaxanthin acyl esters.

(A) HPLC analyses (system IIb) of pigment extracts from wild type, the vdl mutant, and a vdl mutant complemented with its native VDL gene (vdl + VDL, clone #4) demonstrate that knockout of the VDL gene results in loss of vaucheriaxanthin acyl esters (peaks 3 and 4) and a concomitant increase in violaxanthin (2). Chromatograms were normalized to chlorophyll a (5); other peaks were identified as latoxanthin (1) and β-carotene (6). For detailed pigment stoichiometries in the wild type, the vdl mutant, and two VDL-complemented strains of the vdl mutant, see table S1. (B) Scheme showing the insertion site of the 1.8-kb zeocin resistance cassette (Zeo^R) within the second exon of the VDL gene in the vdl mutant. Binding sites of the primers used for differentiation of wild-type (WT) and mutated VDL gene are indicated by black arrows. (C) Agarose gel showing 1.8-kb size difference of polymerase chain reaction (PCR) products of WT or the VDL-deficient mutant (vdl) using PCR primers specified in (B). (D) Agarose gel showing additional band of WT VDL fragment at 1.4 kb for PCR products from successfully complemented transformants of the vdl mutant; clones #2 and #4 were used for pigment analyses. Other experimental details are described in Materials and Methods.

Fig. 2. Transient expression of VDL1 from P. tricornutum in tobacco leaves results in accumulation of neoxanthin.

HPLC analysis (system Ib) of pigment extracts from untreated leaves (control) and leaves after Agrobacterium-mediated transformation with VDL1 or VDL2 from P. tricornutum. Chromatograms were normalized to chlorophyll a (peak 8), and the parts showing neoxanthin (1), violaxanthin (2), 9′-cis-neoxanthin (3), deepoxyneoxanthin (4), and antheraxanthin (5) were magnified four times. Other peaks were identified as lutein (6), chlorophyll b (7), β-carotene (9), and 9-cis-β-carotene (10). For detailed pigment stoichiometries in leaf samples from the three treatments, see table S2.

Fig. 3. Violaxanthin-neoxanthin tautomerase activity of VDL proteins from further chromalveolate algae in tobacco leaves and in vitro.

(A) HPLC analyses (system Ib) of pigment extracts from tobacco leaves transiently expressing VDL from algae with diadinoxanthin cycle (brown-colored chromatograms) or from algae that use the violaxanthin cycle for photoprotection (green-colored chromatograms). Depicted are details of chromatograms normalized to the chlorophyll a peak showing trans-neoxanthin (Nx), violaxanthin (Vx), and 9′-cis-neoxanthin (cNx). Blue label indicates that trans-neoxanthin is only found in leaves transformed with the algal VDL genes. Note the stronger accumulation of trans-neoxanthin when leaves are transformed with VDL from algae with diadinoxanthin cycle. Detailed pigment stoichiometries in leaf samples from all treatments are given in table S3. (B) HPLC analyses (system II) of samples from in vitro assays with recombinant algal VDL proteins using violaxanthin as substrate and incubation times of 3 hours [chromatogram colors and peak labels as in (A), reaction products labeled blue; values below species names indicate VDL concentrations in assay; activity of VDL from P. minimum was measured using bacterial lysate]. For each VDL, at least two different preparations were examined with similar results. VDL concentrations in the assays varied because of different expression efficiencies of VDL proteins in E. coli that could not be overcome by changing expression conditions. VDL proteins from algae with diadinoxanthin cycle showed substantial in vitro activities that were correlated with protein concentrations in the assays, while VDL proteins from algae with violaxanthin cycle showed minor or no activity independent of the protein concentrations used. Other experimental details are described in Materials and Methods.

Fig. 4. Functional comparison of VDE and VDL1 from P. tricornutum.

(A) Proposed pathway of carotenoid biosynthesis in chromalveolate algae and reactions catalyzed by VDE (red arrows) and VDL (blue arrows). In addition, the two putative pathways to 9′-cis-neoxanthin—a neoxanthin isomer specific to land plants and green algae—are indicated by broken arrows (green-colored paths; enzymes unknown). (B to D) HPLC analyses (system II) of in vitro assays with recombinant VDE showing (B) enzymatic de-epoxidation of Vx to Ax and Zx and of Ddx to Dtx (reaction products labeled red); (C) pH dependence of VDE activity; and (D) dose-response curve of VDE inhibition by reductant DTT. (E to G) In vitro assays with recombinant VDL1 showing (E) enzymatic tautomerization of Vx to Nx and of Nx to Vx (reaction products labeled blue); (F) pH dependence of VDL1 activity; and (G) dose-response curve of VDL1 inhibition by DTT. The experimental details are described in Materials and Methods.

Fig. 5. Phylogeny and summary of enzymatic activities in tobacco or in vitro of VDL proteins from selected dinophyte, haptophyte, and heterokont algae.

Investigated VDL proteins are underlined, and their violaxanthin-neoxanthin tautomerase activities in tobacco or in vitro were indicated as strong (++), medium (+), weak (o), or not detectable (−) based on the results shown in Fig. 3. Also indicated are the type of xanthophyll cycle (Xan cycle) and the major allenic light-harvesting carotenoid (LH car) present in the respective algal species [for abbreviations of carotenoids, see Fig. 4]. Midpoint-rooted maximum likelihood tree of VDE family proteins from selected chromalveolate algae shows bootstrap supports (100 replicates) above 50% for major nodes. “VDL1” or “1” behind species name indicates additional presence of a VDL2 in that species. For VDE, VDL2, and VDR, values in brackets indicate the number of sequences included in the phylogenetic analysis. The full tree is shown in fig. S2.