Elsevier

Phytochemistry

Volume 139, July 2017, Pages 33-46
Phytochemistry

A re-evaluation of the final step of vanillin biosynthesis in the orchid Vanilla planifolia

https://doi.org/10.1016/j.phytochem.2017.04.003Get rights and content

Highlights

  • Vanilla planifolia extracts have 4-hydroxybenzaldehyde synthase activity.

  • This activity is associated with a cysteine protease-like protein.

  • This protein could not be shown to possess vanillin synthase activity.

  • This protein is constitutively expressed in all V. planifolia tissues tested.

Abstract

A recent publication describes an enzyme from the vanilla orchid Vanilla planifolia with the ability to convert ferulic acid directly to vanillin. The authors propose that this represents the final step in the biosynthesis of vanillin, which is then converted to its storage form, glucovanillin, by glycosylation. The existence of such a “vanillin synthase” could enable biotechnological production of vanillin from ferulic acid using a “natural” vanilla enzyme. The proposed vanillin synthase exhibits high identity to cysteine proteases, and is identical at the protein sequence level to a protein identified in 2003 as being associated with the conversion of 4-coumaric acid to 4-hydroxybenzaldehyde. We here demonstrate that the recombinant cysteine protease-like protein, whether expressed in an in vitro transcription-translation system, E. coli, yeast, or plants, is unable to convert ferulic acid to vanillin. Rather, the protein is a component of an enzyme complex that preferentially converts 4-coumaric acid to 4-hydroxybenzaldehyde, as demonstrated by the purification of this complex and peptide sequencing. Furthermore, RNA sequencing provides evidence that this protein is expressed in many tissues of V. planifolia irrespective of whether or not they produce vanillin. On the basis of our results, V. planifolia does not appear to contain a cysteine protease-like “vanillin synthase” that can, by itself, directly convert ferulic acid to vanillin. The pathway to vanillin in V. planifolia is yet to be conclusively determined.

Introduction

Vanillin (4-hydroxy-3-methoxybenzaldehyde), probably the world's most popular flavor compound and a weak antimicrobial agent, accumulates in the pods of the vanilla orchid, Vanilla planifolia, where it is stored as its 4-O-glucoside, glucovanillin. Vanillin is also found in other Vanilla species, but is not widely found among the Orchidaceae. Vanillin is believed to be made in specialized cells within the inner mesocarp area of the pod (Brillouet et al., 2014, Joel et al., 2003), although exactly which cell types produce vanillin has been a matter of debate (Joel et al., 2003, Odoux and Brillouet, 2009). Accumulation of vanillin has recently been proposed to occur in re-differentiated plastids termed “phenyloplasts” (Brillouet et al., 2014). Whether vanillin is synthesized in plastids, or transported to them after glycosylation, is not clear. Work on vanillin biosynthesis has been hampered by the lack of good genetic systems in V. planifolia, although recent transcriptomic analyses of this species (Gallage et al., 2014, Rao et al., 2014) should facilitate gene discovery for flavor compound synthesis.

Vanillin is formed via the phenylpropanoid pathway from trans-cinnamic acid (Fig. S1) (Supporting Information) (Dignum et al., 2001, Dixon, 2011, Walton et al., 2003, Zenk, 1965). The conversion of coumaric acid (4-hydroxycinnamate) to vanillin essentially requires four steps: 1) shortening of the side chain by 2 carbons; 2) introduction of the aldehyde function to the side chain (in some models this may occur as an integral part of chain shortening; 3) introduction of the 3-hydroxyl group; 4) 3-O-methylation (Fig. 1). The actual order of these reactions has proven difficult to determine, and a number of conflicting models have been presented, based mainly on analysis of in vitro enzyme activities or precursor incorporation (Dixon, 2011, Funk and Brodelius, 1990, Funk and Brodelius, 1992, Negishi et al., 2009, Podstolski et al., 2002, Rao and Ravishankar, 2000, Zenk, 1965). Other than the fact that the O-methylation reaction has to occur after the 3-hydroxylation, these reactions could theoretically occur in any order. Furthermore, there is more than one mechanism for chain shortening of hydroxycinnamic acids, and these lead to products with different oxidation states of the terminal group of the side-chain (Dixon, 2011). Labeling and genetic studies have demonstrated that the formation of benzoic acids in petunia flowers occurs by multiple pathways involving both oxidative and non-oxidative chain shortening (Boatright et al., 2004, Orlova et al., 2006).

If the 3-O-methylation reaction occurs before chain shortening, then vanillin would be formed directly from ferulic acid (Fig. 1). Labeling experiments have supported this model (Negishi et al., 2009, Zenk, 1965). Although some labeling studies have also suggested that 4-hydroxybenzaldehyde (4HBA) is not a good precursor of vanillin (Negishi et al., 2009), there is also evidence for chain shortening prior to O-methylation; at least in vitro, 4-coumarate can be converted to 4HBA through a non-oxidative process by a 4HBA synthase (4HBS) chain shortening enzyme requiring the presence of a thiol reagent but no other cofactor (Podstolski et al., 2002) (Fig. 1), and V. planifolia pod tissues contain high levels of 4HBA, mainly in the form of its glucoside (Negishi et al., 2009). Conversion to vanillin would then require 3-hydroxylation (presumably by a cytochrome P450 enzyme) and O-methylation. Enzymes capable of efficiently catalyzing 3-O-methylation of 4HBA have been shown to function in parallel reactions in lignin biosynthesis (Kota et al., 2004), transcripts encoding several such enzymes are expressed in V. planifolia (Rao et al., 2014), and an O-methyltransferase with preferential activity for flavonoids but nevertheless active with 3,4-dihydroxybenzaldehyde, is found in the plastids of the placental hair cells where vanillin may be synthesized, but its transcripts are more abundant in leaves than in hair cells (Widiez et al., 2011).

It is noteworthy that conversion of coumaric acid to 4HBA occurs much more efficiently than conversion of ferulic acid to vanillin in crude and partially purified extracts from V. planifolia (Negishi and Negishi, 2014, Podstolski et al., 2002). The 4HBS activity of V. planifolia is, however, unusual in that it requires a high concentration of reductant such as dithiothreitol (Negishi and Negishi, 2014, Podstolski et al., 2002), and does not exhibit Michaelis-Menten kinetics, failing to saturate as substrate concentration is increased (Podstolski et al., 2002). An apparent iron-dependent 2,3-dioxygenase was partially purified from pods of V. planifolia and demonstrated to catalyze the chain shortening of both coumaric and ferulic acids (Negishi and Negishi, 2015). It has also been suggested that vanillin may originate from coumaric acid glucoside (Palama et al., 2010), although enzymes for the subsequent interconversions at the glucoside level have yet to be definitely identified.

In view of the controversies surrounding the chain-shortening reaction of vanillin biosynthesis, the suggestion that V. planifolia possesses a vanillin synthase that converts ferulic acid directly to vanillin (Gallage et al., 2014), and that this protein has high identity to cysteine protease enzymes, has drawn much attention. The fact that this vanillin synthase was reported to be unable to convert coumaric acid to 4HBA was surprising to us because our earlier analysis of the V. planifolia 4HBS (based on assay of 4HBA formation from coumarate) (Havkin-Frenkel et al., 2003) resulted in the cloning of a gene encoding exactly the same protein as reported subsequently by Gallage et al. (2014). The recent claims for ferulic acid as the direct precursor of vanillin through the action of vanillin synthase (Gallage et al., 2014) do not provide evidence for the enzyme's activity in vivo, provide no quantitative data for enzyme turnover rates or kinetic constants, and, although reporting labeling studies, themselves provide no direct evidence for incorporation of ferulic acid into vanillin in V. planifolia tissues. Because Gallage et al. (2014) ascribe a different activity to a protein that we had identified as being associated with a 4HBS, we here re-examine, using a combination of transcriptomic, biochemical, and histochemical approaches, the evidence that the cysteine-protease-like protein (CPLP) of V. planifolia is a vanillin synthase.

Section snippets

A 4-Hydroxybenzaldehyde synthase activity of V. planifolia is associated with a cysteine protease-like protein

In 2003 (Havkin-Frenkel et al., 2003), we reported the cloning of a gene encoding a 28 kDa protein identical to that subsequently reported by Gallage et al. (2014). The gene was cloned based on protein purification and peptide sequencing. Essentially, proteins with 4HBS activity were partially purified from crude V. planifolia embryo culture extracts by a four step procedure involving ammonium sulfate fractionation, hydrophobic interaction chromatography (HIC), ion exchange chromatography and

Conclusions

Vanilla planifolia possess a protein with high sequence identity to cysteine proteases that is expressed throughout the plant and co-purifies with at least one of several enzymes or enzyme complexes that catalyze or facilitate, at least in vitro, conversion of 4-coumaric acid to 4HBA. This protein alone has weak catalytic activity in the chain shortening of coumaric acid when expressed in an in vitro transcription/translation system. However, we were unable to demonstrate activity of this

Plant materials

Freshly harvested stems, leaves, roots, and beans of V. planifolia were from plants grown in the greenhouse and were frozen immediately in liquid nitrogen and stored at −80 °C. Beans were further dissected into dark green mesocarp and light inner mesophyll/placental tissues, V-shaped papillae (hairs), placental laminae and seeds prior to freezing (Rao et al., 2014).

Embryo cultures

Embryo cultures, containing differentiated cell aggregates, were established as described previously (Herz, 2000). Cultures were

Acknowledgements

This work was supported by David Michael and Company, the Samuel Roberts Noble Foundation, and the University of North Texas.

References (37)

  • J. Boatright et al.

    Understanding in vivo benzenoid metabolism in petunia petal tissue

    Plant Physiol.

    (2004)
  • J.-M. Brillouet et al.

    Phenol homeostasis is ensured in vanilla fruit by storage under solid form in a new chloroplast-derived organelle, the phenyloplast

    J. Exp. Bot.

    (2014)
  • C.D. Broeckling et al.

    Metabolic profiling of Medicago truncatula cell cultures reveals effects of biotic and abiotic elicitors on primary metabolism

    J. Exp. Bot.

    (2005)
  • S.J. Clough et al.

    Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana

    Plant J.

    (1998)
  • M.J.W. Dignum et al.

    Vanilla production: technological, chemical, and biosynthetic aspects

    Food Rev. Int.

    (2001)
  • R.A. Dixon

    Vanillin biosynthesis- not as simple as it seems?

  • C. Funk et al.

    Phenylpropanoid metabolism in suspension cultures of Vanilla planifolia Andr. III. Conversion of 4-methoxycinnamic acids into 4-hydroxybenzoic acids

    Plant Physiol.

    (1990)
  • C. Funk et al.

    Phenylpropanoid metabolism in suspension cultures of Vanilla planifolia Andr. IV. Induction of vanillic acid formation

    Plant Physiol.

    (1992)
  • Cited by (34)

    • Physicochemical, microbiological, and structural relationship of vanilla beans (Vanilla planifolia, Andrews) during traditional curing process and use of its waste

      2023, Journal of Applied Research on Medicinal and Aromatic Plants
      Citation Excerpt :

      Further, Podstolski et al. (2002) reported an enzyme involved in the reaction that converts 4-coumaric acid to 4-hydroxybenzaldehyde in vanilla tissue cultures (4-benzaldehyde synthetase), while Negishi et al. (2009) and Gallage et al. (2014) reported the vanillin synthesis from ferulic acid through vanillin synthetase enzyme. However, although said studies have been focused to elucidate about possible pathways for vanillin synthesis, reports of a fully disclosed metabolic route for biosynthesis of this compound as well as its location at microstructural level are still unavailable since the participation of key enzymes in reported metabolic routes has not been demonstrated; however, a key intermediate in most of the proposed routes and modifications is glucovanillin (Negishi et al., 2009; Brillouet et al., 2014; Yang et al., 2017). On the other hand, following the vanillin synthesis during the curing process (Tapia-Ochoategui et al., 2011) the β-glucosidase activity and the vanillin concentration were determined in different steps during curing (Fig. 8), finding that at 10 drying-sweating cycles the enzyme activity and vanillin concentration are at maximum level representing a close relationship with water activity (0.973) in this assessed point, conditions to which the activity of a number of enzymes is favored; however, the humidity content is still high (60.08 ± 0.35) at the end stage of curing process, which supports microorganism contamination, therefore in spite of showing a significantly lower vanillin content and enzymatic activity in the beans in the conditioning stage, the physicochemical conditions of humidity and water activity are needed (14.55 ± 0.47 and 0.717) to perform the lipid oxidation, enzymatic darkening reactions (Labuza et al., 1972), as well as microstructural changes that together contribute to synthesis of aromatic compounds that complete the aromatic profile of vanilla beans.

    • Metabolite transformation and β-D-glucosidase activity during the high hydrostatic pressure assisted curing of vanilla beans (Vanilla planifolia) to improve phenolic compounds formation

      2022, Food Chemistry
      Citation Excerpt :

      This is a complex mixture of > 200 flavor components, and among these odor-active compounds, vanillin (4-hydroxy-3-methoxy-benzaldehyde), is the main phenolic compound responsible for flavor (Cai et al., 2019). Other odor-active components are vanillic acid, 4-hydroxybenzaldehyde, p-hydroxybenzoic acid, and vanillyl alcohol, which are present in lower concentrations (Cai et al., 2019; Yang et al., 2017). Natural vanillin is obtained during the curing of green mature beans from Vanilla planifolia Jacks.

    • Next-generation metabolic engineering approaches towards development of plant cell suspension cultures as specialized metabolite producing biofactories

      2020, Biotechnology Advances
      Citation Excerpt :

      We assume that the abundance of ferulic acid in various plant species can be used to produce vanillin by just introducing the VpVAN in heterologous PCSC (Kumar and Pruthi, 2014). This was disputed by Yang et al. (2017) who re-evaluated the catalytic efficiency of VpVAN in E. coli and yeast and concluded that VpVAN is unable to produce vanillin from ferulic acid but is instead involved in the conversion of p-coumaric acid to p-hydroxybenzaldehyde. However, it was subsequently reported that to achieve efficient gene expression and desired catalytic activity of the VpVAN gene, codon optimization and modulation of the precursor pathway are essential in heterologous hosts (Gallage et al., 2017).

    View all citing articles on Scopus
    1

    Present address: Department of Biochemistry and Molecular Biology, College of Biological Sciences, China Agricultural University, Beijing 100193, China.

    2

    Present address: 64 Booream Avenue, Milltown, NJ 08850, USA.

    3

    Present address: Chromatin Inc., 1301 East 50th St, Lubbock, TX 79404, USA.

    4

    Present address: Faculty of Biology and Environmental Science, Cardinal Stefan Wysznski University, Warsaw, Poland.

    View full text