OUP user menu

Metabolic engineering of Saccharomyces cerevisiae for the synthesis of the wine-related antioxidant resveratrol

John V.W. Becker, Gareth O. Armstrong, Marthinus J. van der Merwe, Marius G. Lambrechts, Melané A. Vivier, Isak S. Pretorius
DOI: http://dx.doi.org/10.1016/S1567-1356(03)00157-0 79-85 First published online: 1 October 2003


The stilbene resveratrol is a stress metabolite produced by Vitis vinifera grapevines during fungal infection, wounding or UV radiation. Resveratrol is synthesised particularly in the skins of grape berries and only trace amounts are present in the fruit flesh. Red wine contains a much higher resveratrol concentration than white wine, due to skin contact during fermentation. Apart from its antifungal characteristics, resveratrol has also been shown to have cancer chemopreventive activity and to reduce the risk of coronary heart disease. It acts as an antioxidant and anti-mutagen and has the ability to induce specific enzymes that metabolise carcinogenic substances. The objective of this pilot study was to investigate the feasibility of developing wine yeasts with the ability to produce resveratrol during fermentation in both red and white wines, thereby increasing the wholesomeness of the product. To achieve this goal, the phenylpropanoid pathway in Saccharomyces cerevisiae would have to be introduced to produce p-coumaroyl-CoA, one of the substrates required for resveratrol synthesis. The other substrate for resveratrol synthase, malonyl-CoA, is already found in yeast and is involved in de novo fatty-acid biosynthesis. We hypothesised that production of p-coumaroyl-CoA and resveratrol can be achieved by co-expressing the coenzyme-A ligase-encoding gene (4CL216) from a hybrid poplar and the grapevine resveratrol synthase gene (vst1) in laboratory strains of S. cerevisiae. This yeast has the ability to metabolise p-coumaric acid, a substance already present in grape must. This compound was therefore added to the synthetic media used for the growth of laboratory cultures. Transformants expressing both the 4CL216 and vst1 genes were obtained and tested for production of resveratrol. Following β-glucosidase treatment of organic extracts for removal of glucose moieties that are typically bound to resveratrol, the results showed that the yeast transformants had produced the resveratrol β-glucoside, piceid. This is the first report of the reconstruction of a biochemical pathway in a heterologous host to produce resveratrol.

  • Antioxidant
  • Coenzyme-A ligase
  • β-glucosidase
  • Resveratrol synthase
  • Stilbene
  • Wine yeast

1 Introduction

The stress metabolite resveratrol has enjoyed considerable tenure as a research subject in diverse fields, ranging from disease resistance in plants to its contribution to human health. Both structural and biochemical mechanisms are involved in providing a plant with resistance to its pathogens [1]. The active defence mechanism in plants in response to infection involves the induced accumulation of antimicrobial, low-molecular-mass secondary metabolites known as phytoalexins [2]. Stilbene synthases synthesise the backbone of the stilbene phytoalexins that have antifungal properties and contribute to the pathogen defences of the plant [3]. The speed and intensity with which stilbene compounds are formed are indicators of the plant's ability to resist fungal infection.

During red wine vinification, the must is fermented with the grape skins. As the temperature and alcohol content increase, phenolic substances, including resveratrol, are extracted into the wine. The must used for the production of white wine is generally fermented in the absence of the grape skins and therefore contains very little of these compounds, including resveratrol, which is primarily synthesised in the fruit skins. Mattivi [4] has performed solid-phase extractions of resveratrol from red wines and determined the levels to be as high as 7 mg l−1 in Cabernet Sauvignon wines, while the levels in white wine were seldom above 0.1 mg l−1[5]. Due to this difference in vinification practices, only the consumption of red wine has been linked to the ‘French paradox’. This dietary anomaly suggests that, although the French generally follow a high-fat diet and low-exercise lifestyle, they have a remarkably low incidence of coronary heart disease.

The oxidative modification of low-density lipoproteins is recognised as an important factor in the development of atherosclerosis [6]. Resveratrol has been implicated in this beneficial action of red wine, mainly due to its ability to act as an antioxidant and as an inhibitor of platelet aggregation [7]. On the basis of the structural similarity between resveratrol and the synthetic estrogen diethylstilbesterol [8], it has been shown that resveratrol acts as a phytoestrogen. Given the known cardioprotective benefits of estrogens [9], these findings are extremely appealing to both wine producers and consumers.

Certain phytochemicals in fruit, vegetables, spices, beverages and foods that are obtained as part of the dietary intake have been identified as potential cancer chemopreventive agents [10]. The antioxidant and anti-inflammatory activities of resveratrol enable it to inhibit biochemical changes involved in tumour initiation, promotion and progression [10,11]. From the wealth of information that is available, it has been shown that resveratrol inhibits the growth of several cancerous cell lines [1215], or has the ability to cause apoptosis in these lines [1622]. Recently, the possibility of managing liver cancer by means of resveratrol administration was highlighted after it was shown that resveratrol was widely distributed in mice liver cells after oral dosage [23]. Given these health benefits, the aim of this study was to investigate the feasibility of developing wine yeast strains for the production of resveratrol during fermentation in both red and white wines. This required the introduction of the phenylpropanoid pathway in Saccharomyces cerevisiae. Malonyl-CoA, one of the substrates required for the production of resveratrol, is present in the yeast and is actively involved in fatty-acid biosynthesis [24]. The other substrate for resveratrol synthase, p-coumaroyl-CoA, can be produced from p-coumaric acid, which is found in small quantities in grape must [25] and has been shown to be accumulated by yeast [26]. However, the amount present in the grape must, as well as the amount accumulated by yeast, could be a limiting factor in the production of resveratrol. The pathway of resveratrol production is shown in Fig. 1.

Figure 1

The biosynthesis of resveratrol from phenylalanine (adapted from Schröder and Schröder [27]).

In this study, the coenzyme-A ligase-encoding gene (4CL216) from a hybrid poplar and the grapevine resveratrol synthase gene (vst1) were inserted into yeast expression cassettes and expressed in a laboratory strain of S. cerevisiae. These recombinant yeast strains were tested for the expression of these heterologous genes and for their ability to produce resveratrol.

2 Materials and methods

2.1 Microbial strains and culture conditions

Escherichia coli DH5α cells were used for bacterial transformations and were grown at 37°C in Luria–Bertani (LB) broth containing 1.2% tryptone, 1.2% sodium chloride and 0.6% yeast extract [28]. Selection was done on solid media (LB agar) containing 100 μg ml−1 ampicillin (Roche, Mannheim, Germany).

S. cerevisiae FY23 cells [29] were grown at 30°C on a rotary shaker in synthetic SCDL medium containing 0.67% yeast nitrogen base without amino acids (Difco, Sparks, MD, USA), supplemented with the required growth factors and 0.8% glucose, as well as in YPD medium (containing 1.2% yeast extract, 2.5% peptone and 1.2% glucose). Solid media contained 2% agar.

2.2 DNA manipulations and plasmid construction

Standard procedures were used for the isolation and manipulation of DNA and for yeast and bacterial transformations [28,30]. Restriction enzymes, T4 DNA ligase and Expand Hi-Fidelity DNA polymerase were used in the enzymatic manipulation of DNA according to the specifications of the supplier (Roche, Mannheim, Germany). The coenzyme-A ligase 4CL216 gene from hybrid poplar was amplified from plasmid p4CL216 [31] using the polymerase chain reaction (PCR) method and the following primer pair: for the forward primer, 5′ GATCAGATCTATGGAGGCAAAAATGATCA 3′, and the reverse primer, 5′ GTACCGGGCCCCCTCGAG 3′. The underlined sequences denote BglII and XhoI restriction enzyme sites, respectively. The PCR product was subcloned into the pGEM-T vector system (Promega, Madison, WI, USA). The 4CL216 gene was then excised from the latter plasmid construct and subcloned into pDLG3 [32] at the BglII–XhoI sites, thereby placing 4CL216 under the control of the yeast alcohol dehydrogenase II gene (ADH2) promoter and terminator, yielding pDLG34CL216. The ADH2P-4CL216-ADH2T gene was named CAL1. Plasmid pDLG3 contains the yeast URA3 auxotrophic marker. The grapevine (Vitis vinifera) vst1 gene was excised from plasmid pVST1 [33] and inserted into the EcoRI site of a YEp352-based plasmid (URA3), thereby placing the vst1 structural gene under the control of the yeast enolase gene (ENO2) promoter (ENO2P) and terminator (ENO2T). The resulting gene construct (ENO2P-vst1-ENO2T) was designated VST1 and the plasmid was designated YEpeno2Vst1. The correct orientation of the vst1 gene in the ENO2P-ENO2T expression cassette was confirmed by digestion with KpnI. The entire expression cassette was cut out at the BamHI–XbaI sites and cloned into the corresponding sites of YEplac181 [34] containing the LEU2 auxotrophic marker, generating YEplac181Vst1. This strategy enabled co-transformations of S. cerevisiae strain FY23 with the recombinant plasmids pDLG34CL216 and YEplac181Vst1. The Leu+ and Ura+ co-transformants were selected on SCDL medium without the growth factors leucine and uracil [30].

2.3 RNA isolation and Northern-blot analysis

The FY23 recombinant yeast cells were grown in 10 ml of selective media for 48 h. Total RNA was isolated using the FastRNA RED kit (BIO 101, Vista, CA, USA). RNA (10 μg) from each culture was subjected to formaldehyde gel electrophoresis. The RNA was transferred to Hybond-N nylon membranes (Amersham Pharmacia Biotech, Buckinghamshire, UK) and individually hybridised to radioactively labelled probes according to standard Northern-blot procedures [30]. A 1178-bp vst1 PCR product was generated using the primer pair 5′ GATCAAGCTTCAATGGCTTCAGTCGAGGAA 3′ and 5′ GATCAAGCTTTTAATTTGTCACCATAGGAA 3′ for the forward and reverse primers, respectively. The 1673-bp 4CL216 (primer pair previously described for the 4CL216 gene) and 1178-bp vst1 PCR products were used as probes. All probes were labelled with [32P]dATP, using the Prime-It II random labelling kit (Stratagene, La Jolla, CA, USA).

2.4 Resveratrol assays

Strain FY23 transformed with pDLG34CL216 and YEplac181Vst1 and the FY23 control (untransformed) strain were inoculated from an overnight pre-culture into 200 ml of SCDL medium containing 10 mg l−1p-coumaric acid. Cultures were generated using individual transformants. The cells were grown at 30°C for 48 h. The cultures were then centrifuged at 5000 rpm for 5 min and the supernatant was removed. The remaining cells were resuspended in 100% ice-cold methanol, after which the cells were broken with glass beads. Extraction was allowed to proceed at 4°C for 2 days and for an additional hour at 37°C. The cell debris were subsequently removed by centrifugation at 5000 rpm for 5 min. The supernatants were dried under nitrogen and dissolved in 500 μl of 25 mM citrate–phosphate buffer containing 0.5 mg ml−1β-glucosidase from almonds (Sigma, St Louis, MO, USA) to liberate glucose molecules linked to the resveratrol. The latter step would liberate free resveratrol that would be analytically measured in subsequent steps. A duplicate culture in this initial analysis was not subjected to β-glucosidase treatment to ascertain whether free resveratrol was being produced by the yeast. After incubation at 37°C for 1 h with β-glucosidase, the free resveratrol was extracted three times with ethyl acetate and dried under nitrogen. The pellet was dissolved in 50% acetonitrile. Mass-spectrometric analysis was developed in conjunction with the Central Analytical Facility (Stellenbosch University) and performed on a Micromass (Manchester, UK) Quattro triple quadropole mass spectrometer fitted with an electrospray ionisation source. Samples were injected by a Waters (Milford, MA, USA) 717 Plus autosampler and transported to the ionisation source in a carrier stream of solvent A [acetonitrile/water: 1/1 (v/v)], pumped by a Pharmacia (New York, NY, USA) LKB 2249 gradient pump.

Ions were detected in the negative mode and the ionisation was optimal at a capillary voltage of 3.5 kV, a cone voltage of 45 V and a source temperature of 120°C. The nebuliser gas used was nitrogen at a gas flow of 40 l h−1. For detection of the molecular ion of resveratrol, the first analyser was scanned through a range of m/z=100–300 at a scan rate of 150 amu s−1. A representative scan was produced by the addition of scans across the elution peak and by subtracting the background. For the fragmentation analysis, the molecular ion was selected by the first analyser and passed into the fragmentation cell, where collisionally induced dissociation was accomplished by the addition of argon at a pressure of 2×10−3 mbar and the application of a collision energy of 30 eV. The fragmentation pattern was generated by scanning the second analyser in the range of m/z=10–240 at 150 amu s−1.

For quantitative analysis, the samples were subjected to separation by high-performance liquid chromatography (HPLC) prior to mass spectrometry. A Phenomenex (Torrance, CA, USA) Luna 3μ C18, 2×150 mm column was used and the mobile phase was 50/50: acetonitrile/water (v/v), at a flow rate of 100 μl min−1. Resveratrol was quantified in yeast extracts by multiple reaction monitoring, using the molecular anion at m/z=227 as the precursor and the fragment at m/z=142 as the product ion. Standards were analysed to construct a calibration curve, from which the concentration of resveratrol in the extract samples was calculated.

3 Results and discussion

3.1 Construction of expression cassettes, yeast transformation and confirmation of gene expression

The production of resveratrol by yeast was attempted by the introduction of two key enzymes that are not present in S. cerevisiae, namely coenzyme-A ligase and resveratrol synthase. The recombinant plasmids pDLG34CL216 and YEplac181Vst1, harbouring the coenzyme-A ligase (CAL1) and resveratrol synthase (VST1) gene constructs, respectively, were co-transformed into the laboratory yeast strain FY23. Northern-blot analysis of the recombinant yeast strains showed active transcription of the 4CL216 and vst1 genes (Fig. 2), whereas the control strain did not show any hybridisation signal. The detected signals in the lanes containing RNA from the transformed strains corresponded to the expected sizes of the various genes when compared to a Bio-Rad Laboratories RNA molecular-size marker.

Figure 2

Northern-blot analysis on total RNA from an untransformed laboratory strain of S. cerevisiae (as a control host strain) and strains transformed with the hybrid poplar coenzyme-A ligase (CAL1) gene construct and the grapevine resveratrol synthase (VST1) gene construct. CAL1 consisted of the ADH2P-4CL216-ADH2T gene construct, whereas VST1 consisted of ENO2P-vst1-ENO2T. Therefore, the transcription of the 4CL216 and vst1 heterologous genes was effected by the yeast alcohol dehydrogenase II (ADH2) and enolase II (ENO2) regulatory elements, respectively. The RNA of S. cerevisiae FY23 laboratory strain transformed with the CAL1 and VST1 gene constructs were probed with the radioactively labelled vst1 (A) or 4CL216 gene sequences (B). Hybridisation signals corresponded to transcripts of the correct size. FY23 was used as a control. Lanes 1–4 correspond to four individual yeast transformants in the case of each hybridisation. Lane C corresponds to the untransformed FY23 host yeast strain.

3.2 Production of resveratrol by yeast transformants

Dual mass-spectrometric assays on seven recombinant S. cerevisiae FY23 strains harbouring the CAL1 coenzyme-A ligase and VST1 resveratrol synthase gene constructs confirmed the presence of trans-resveratrol in the yeast cells (Table 1). Detectable levels of this compound were only obtained after the samples were treated with β-glucosidase to release the glucose moieties that are typically bound to resveratrol (Fig. 3). Isolations were performed on yeast cultures grown to the stationary phase and an optical density of 6.7. These isolations indicated that yeast cultures (200 ml) with the addition of p-coumaric acid as the only precursor were able to produce up to 291 ng of resveratrol, in the form of its β-glucoside piceid, as evidenced upon β-glucosidase treatment (Table 1). Levels varied mostly between 100 and 300 ng per 200 ml culture, with the control showing no production of this compound. These results are very promising, since they confirmed that the CAL1 and VST1 gene constructs were not only effectively transcribed under the control of the ADH2 and ENO2 regulatory elements, but that their encoded products (coenzyme-A ligase and resveratrol synthase) were biologically active in the heterologous yeast system. It is, moreover, the first report of the construction of a pathway to produce resveratrol in a heterologous system by introducing some of the biosynthetic genes into the producer organism. Although the levels of resveratrol that were obtained are low in comparison to those found in commercial red [4] and even white wines [5], several optimisations might still be implemented to increase these levels. More resveratrol could be obtained when optimal extraction procedures have been elucidated. In addition, optimal culture conditions, the time of harvest of the cells, as well as the stability of resveratrol in the heterologous yeast environment need to be evaluated. At present, it seems that all the piceid produced remains within the yeast cell (result not shown). Insertion of the 4CL216 and vst1 genes into constitutive expression–secretion cassettes and co-expression with an appropriate heterologous β-glucosidase gene are of particular interest if the long-term goal of developing wine yeast strains capable of resveratrol production is to be realised.

View this table:
Table 1

Amount of liberated resveratrol produced by different cultures of recombinant yeast strains following β-glucosidase treatment

Yeast cultureResveratrol produced (μg l−1 culture volume)
S. cerevisiae FY23 (control strain)0
Recombinant strain 10.61
Recombinant strain 20.84
Recombinant strain 31.30
Recombinant strain 40.32
Recombinant strain 51.45
Recombinant strain 60.83
Recombinant strain 71.38
  • The recombinant strains refer to S. cerevisiae FY23 transformed with the CL216 gene under control of the ADH2 promoter and with the vst1 gene under control of the ENO2 promoter.

Figure 3

Dual mass-spectrometrical analysis of organic extracts from untransformed and recombinant FY23 yeast strains following separation on a C18 column by HPLC analysis for resveratrol content. The chromatograms illustrate the resveratrol standard peak at 50 ng ml−1 (A), the untransformed S. cerevisiae FY23 host strain after β-glucosidase treatment, showing no resveratrol production (B), resveratrol extraction of a recombinant yeast (transformed with CAL1 and VST1) without β-glucosidase treatment, showing no resveratrol detection (C), resveratrol extraction of a recombinant yeast (expressing CAL1 and VST1), after β-glucosidase treatment, yielding liberated resveratrol with a retention time similar to that of the standard (D).

Another important aspect that is still under investigation is the availability of the necessary precursors at adequate levels in the culture medium for optimal resveratrol production. Although the precursors do occur in the yeast, they are also utilised in the fatty-acid biosynthetic pathway, which might cause competition for these precursors. However, initial experiments aimed at clarifying precursor necessity have shown that cultures provided with all the necessary precursors do not produce significantly more resveratrol (results not shown). The single addition of the precursor p-coumaric acid to the synthetic medium led to the observed production of resveratrol in laboratory yeast. This compound is present in wine must prior to fermentation, suggesting that no additions would be necessary to produce resveratrol via yeast fermentation.

In conclusion, by successfully co-expressing the coenzyme-A ligase-encoding gene (4CL216) from a hybrid poplar and the grapevine resveratrol synthase gene (vst1) in a laboratory strain of S. cerevisiae, this pilot study has resulted in progress towards the possible development of wine yeast starter culture strains for the production of wine with increased levels of resveratrol. The preliminary results obtained indicate the possibility of increasing the amounts of resveratrol in both white and red wine by producing this compound in the wine yeast during the fermentation. Such a resveratrol-producing yeast might offer a viable way to meet the demands of a growing number of health-conscious and prudent wine consumers [3539]. However, further research is essential to optimise the levels of yeast-derived resveratrol, and to ensure that such a yeast would not compromise the safety and sensory quality of the wine. Furthermore, it is important to realise that no commercial wine is currently being produced by a genetically modified yeast and that the wine industry will not use such strains as starter cultures unless both the industry and the consumers are satisfied that they are safe and beneficial. Thus, several obstacles relating to scientific, technical, economic, marketing, safety, regulatory, legal and ethical issues remain to be overcome in the short to medium term before a genetically modified yeast could be harnessed for the production of wine at a commercial scale. These issues and challenges have recently been reviewed extensively to point out that, despite the current vocal opposition to genetically modified organisms and products, it will be at the peril of both the wine producer and the wine consumer should gene technology be ignored by the international wine industry [3641]. A well-constructed resveratrol-producing wine yeast that fully complies with international biosafety regulation and legislation, and that increases the cancer chemopreventive and cardioprotective benefits of moderate wine consumption would undoubtedly support this viewpoint.


The authors would like to thank Carl Douglas, for the coenzyme-A ligase genes, and Bayer, for the resveratrol synthase gene. We express our sincere gratitude to the National Research Foundation (NRF) and the South African wine industry (Winetech) for financial support.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
View Abstract