OUP user menu

Monoterpenoid biosynthesis in Saccharomyces cerevisiae

Marilyne Oswald, Marc Fischer, Nicole Dirninger, Francis Karst
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00172.x 413-421 First published online: 1 May 2007

Abstract

Plant monoterpenoids belong to a large family of plant secondary metabolites with valuable applications in cosmetics and medicine. Their usual low levels and difficult purification justify the need for alternative fermentative processes for large-scale production. Geranyl diphosphate is the universal precursor of monoterpenoids. In yeast it occurs exclusively as an intermediate of farnesyl diphosphate synthesis. In the present study we investigated the potential use of Saccharomyces cerevisiae as an alternative engineering tool. The expression of geraniol synthase of Ocimum basilicum in yeast allowed a strong and specific excretion of geraniol to the growth medium, in contrast to mutants defective in farnesyl diphosphate synthase which excreted geraniol and linalool in similar amounts. A further increase of geraniol synthesis was obtained using yeast mutants defective in farnesyl diphosphate synthase. We also showed that geraniol synthase expression affects the general ergosterol pathway, but in a manner dependent on the genetic background of the strain.

Keywords
  • monoterpenoids
  • Saccharomyces cerevisiae
  • geranyl diphosphate
  • farnesyl diphosphate synthase
  • geraniol synthase
  • ergosterol biosynthesis

Introduction

Terpenoids are a large family of metabolites and have been the subject of numerous studies, especially as plant metabolites. Among them, monoterpenes and their corresponding alcohols share useful properties, such as fragrances in essential oils and perfumes or variety aroma in wines. Some others exhibit antimicrobial (Pattnaik, 1997) or cancer chemopreventive properties (Zheng, 1992; Hohl 1996; Carnesecchi, 2001) or can be used for insect control (Lee, 1997). Geranyl diphosphate (GPP) is the precursor of monoterpenes as well as of a number of secondary plant metabolites, such as antimitotic alkaloids (Burlat, 2004). The relatively small number of terpenoids available commercially has prompted considerable interest in development of microbial fermentative processes as an alternative approach for industrial-scale production of these metabolites (Chotani, 2000). However, such production requires modifying microbial internal metabolism in order to shift production towards the desired terpenoid.

Terpenoid production has been investigated in the prokaryote Escherichia coli. Carter (2003) have evaluated the ability of E. coli cells to produce carvone. For this purpose, E. coli was transformed with plasmids carrying cDNAs encoding the four enzymes required for carvone synthesis from dimethylallyl diphosphate (DMAPP) and isopentenyl diphosphate (IPP). However, the low level of production attained was attributed to the limiting availability of IPP and DMAPP in E. coli. Installation of the yeast mevalonate pathway in the bacteria allowed a 20-fold increase in terpenoid production (Martin, 2003).

Yeast is an alternative eukaryotic model, widely used in peptide production and metabolic engineering. It also has the ability to produce isoprenoids, as described by Szczebara (2003), with the biosynthesis of hydrocortisone from the sterol pathway. The most basic requirement for monoterpenoid production in yeast is the in vivo availability of GPP. No specific cellular function for GPP has been described in yeast, besides being an intermediate compound of farnesyl diphosphate (FPP) synthesis (Fig. 1). FPP is the precursor of key products such as sterols, dolichols and geranylgeranyl diphosphate (Szkopinska, 1997) and therefore it contributes to membrane structure, cell-wall synthesis, protein prenylation or ubiquinone synthesis. FPP is a basic product of the yeast cell, but its cellular relevance makes it widely present among eukaryotic and prokaryotic cells. The tight binding of GPP to the farnesyl diphosphate synthase (FPPS) catalytic site might explain why generally in animals and microorganisms no GPP is released and made available for the biosynthesis of C10 byproducts. However, yeast mutants excreting geraniol and linalool have been characterized previously (Chambon, 1990, 1991). It has been shown that they carry a specific mutation in the ERG20 gene encoding FPPS (Blanchard & Karst, 1993). The enzyme activity involved in GPP dephosphorylation has not yet been identified. However, Faulkner (1999) showed that the LPP1 and DPP1 genes encoding diacylglycerol phosphate phosphatases accept isoprenoid pyrophosphates as substrates in vitro.

Figure 1

Biosynthesis pathway of geraniol and linalool from primary precursors dimethylallyl diphosphate (DMAPP) and isopentenyl diphosphate (IPP). The indicated enzymes are: 1, farnesyl diphosphate synthase (FPPS); 2, geraniol synthase (GES); 3, linalool synthase (LIS); 4, geranyl diphosphate synthase (GPPS).

In aromatic plants, specific terpene synthases using GPP as substrate have been isolated (Fig. 1), such as linalool synthase (LIS) (Dudareva, 1996) and more recently geraniol synthase (GES) (Iijima, 2004).

With the objective of specific monoterpenol production in yeast, we overexpressed Clarkia breweri LIS and Ocimum basilicum (sweet basil) GES in wild-type (WT) and mutant strains to evaluate monoterpenol production levels in strains with normal or defective FPP biosynthesis. The impact of this production on the sterol pathway was also analysed.

Materials and methods

Chemicals

Linalool (97%) and geraniol (98%) were obtained from Sigma, amorolfine was a gift from Hoffmann-La Roche, simvastatin was kindly supplied by Merck, and citronellol (98%), eugenol (>99%), m-cresol (>99%) and methylene chloride puriss (>99.9%) were purchased from Fluka. Methanol (LiChrosolv quality) and n-pentane (>99%) were purchased from Merck.

Strains and plasmids

The Saccharomyces cerevisiae strains, vectors and their origin are listed in Table 1. Strain FY1679-28C was derived from S. cerevisiae S288C (Thierry, 1995). Both CC25 and 5247 were derived from FL100 (ATCC 28383) (Chambon, 1990; Soustre, 1996). MO56 and MO57 were obtained from diploid strain Y21258 (EUROSCARF) bearing a chromosomal ERG20 gene disruption respectively complemented by a plasmid carrying the ERG20 or erg20-2 gene. Strains YCM30, YCM33, YCM45, YCM51 and YCM62, isogenic to FL100, were kindly provided by Dr C. Marcireau (Sanofi-Aventis, Vitry sur Seine, France).

View this table:
Table 1

Yeast strains and plasmids

Strains and plasmidsCharacteristicsSource
Strains
FY1679-28CMATa, ura3-52, trp1Δ63, leu2Δ1, his3Δ200Thierry (1995)
5247MATa, ura 3-1, trp1-1Soustre (1996)
CC25MATα, ura 3-1, trp1-1, erg 20-2, erg 12-2Delourme (1994)
MO56MATα, his3Δ1, leu2Δ, ura3Δ0, trp1Δ63, erg20::kanMX4, [pBS]This study
MO57MATα, his3Δ1, leu2Δ, ura3Δ0, trp1Δ63, erg20::kanMX4, [pKS]This study
YCM30MATa, ura 3-1, trp1-1, pdr1ΔMarcireau C.
YCM33MATa, ura 3-1, trp1-1, pdr5ΔMarcireau C.
YCM62MATa, ura 3-1, trp1-1, snq2ΔMarcireau C.
YCM45MATa, ura 3-1, trp1-1, pdr1Δ, pdr5ΔMarcireau C.
YCM51MATa, ura 3-1, trp1-1, pdr1Δ, pdr5Δ, snq2ΔMarcireau C.
Plasmids
pNEV- NPMA1 expression cassette in 2 μ-based vector, URA3 markerSauer & Stolz (1994)
pMAGTpNEV-N, yEGFP3-CYC1term, TRP1markerThis study
pBSERG20 in pNEV-NThis study
pKSerg20-2 in pNEV-NThis study
pMO5GES-GFP in pMAGTThis study
pMO6LIS-GFP in pMAGTThis study

Plasmid pMAGT was constructed by integrating the NotI-PvuII fragment (yEGFP3 coding sequence with CYC1 terminator but without ATG) from pGB7 (Sagot, 1999) in the pNEV-N replicative vector (Sauer & Stolz, 1994), NotI-SmaI-digested. The URA3 marker was disrupted by TRP1 by homologous recombination.

GES-green fluorescent protein (GFP) and LIS-GFP fusions were obtained as follows. Primers recGESfw (5′-GAAAGAAAAAAAATATACCCCAGCGGCCGCATGCCTCTAAGTTCAACTCC-3′), recGESrv (5′-CCTGCAGCCCGGGGGATCCACTAGTTTGAGTGAAGAAGAGGGCATC-3′) and recLISfw (5′-GAAAGAAAAAAAATATACCCCAGCGGCCGCATGCAGCTCATAACAAATTT-3′), recLISrv (5′-ATCCCTGCAGCCCGGGGGATCCACTAGTACTGAAACATAGTTTGATGT-3′) were designed to amplify the GES and LIS coding sequences (including ATG but without stop codon), respectively. Both PCR products, bordered by the homologous sequences of the end of PMA1prom and the start of yGFP3, were mixed with a pMAGT NotI opened vector and used to transform FY1679-28C. Trp+ colonies were selected, then fluorescent clones were isolated and the integration was checked by sequencing.

To construct the pBS and pKS plasmids, WT and mutant strains were used to amplify the ERG20 and erg20-2 genes with the following primers: fwERG20 (5′-AAGGCTCGAGATGGCTTCAGAAAAAGAAATTAGG-3′) and rvERG20 (5′-AAGGACGCGTCTATTTGCTTCTCTTGTAAACTTTGT-3′). Both PCR products were subcloned in pGEM-T plasmid (Promega). These intermediate constructions were digested with NotI and the ERG20 and erg20-2 genes were inserted at the NotI site of pNEV-N to yield pBS and pKS, respectively.

Strains MO56 and MO57 were obtained by transformation of Y21258 (Mat a/Matα, his 3Δ1/his 3Δ1, leu 2Δ0/leu 2Δ0, lys 2Δ0/LYS 2, MET 15/met 15Δ0, ura 3Δ0/ura 3Δ0, erg20 : : kan MX4/ERG20) with pBS and pKS, sporulation and selection of haploid strains that were able to grow in the presence of G418 at 200 μg mL−1. Trp auxotropy was introduced by crossing with FY1679-28C.

Culture conditions

Yeast cells were grown in a shaking incubator at 30°C in 500-mL Erlenmeyer flasks containing 150 mL of liquid minimum medium [1.7 g L−1 Yeast Nitrogen Base (YNB, Difco), 5 g L−1 ammonium sulphate] supplemented with 1% glucose as carbon source. Auxotrophic requirements were supplied as required at 50 μg mL−1.

Western blot analysis

Proteins were extracted from 50 mL yeast culture in exponential growth phase. Cell pellets (5000 g, 5 min) were crushed in 200 μL Tris-HCl 0.5 M, pH 6.8, by vortexing for 5 min with glass beads. Cell debris was eliminated by centrifugation (12 000 g, 15 min) and protein concentration was measured with the Bio-Rad ‘Protein Assay’ reagent (Bradford). Laemmli was added to solubilize the proteins and this preparation was boiled for 5 min. Proteins (40 μg) were separated by electrophoresis on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gels and electrotransferred onto nitrocellulose membrane. After transfer, the membrane was soaked in 1 × TBS buffer (Tris 100 mM, NaCl 140 mM, pH 7.8) with 2% Tween 20 for 1 h. The nitrocellulose membrane was then incubated for 2 h with mouse monoclonal anti-GFP antibodies (Roche) diluted 1 : 3000 in TBS buffer. After several washings in the same buffer, the membrane was incubated for 2 h with alkaline phosphatase-conjugated goat antimouse IgG (Sigma) diluted 1 : 5000. Bound antibodies were detected by incubation of the membrane in Fast Blue RR salt, 0.6 mM α-naphtyl phosphate, 0.1 M Tris-HCl, pH 8.6.

Monoterpenoid analysis

The cells from a stationary phase culture were harvested by centrifugation (5000 g for 5 min). The medium was adjusted to pH 8.5 with KOH and eugenol (40 μg) was added to evaluate extraction efficiency. The mixture (40 mL) was passed trough a Bond Elut C18 (5 g) SPE cartridge (Varian) conditioned with methanol (3 × 10 mL) and water (3 × 10 mL). After washing with water (10 mL), terpenols were eluted with methylene chloride (10 mL). The sample was dried with anhydrous Na2SO4, concentrated to about 500 μL under a nitrogen stream, m-cresol (20 μg) was added as internal standard and the sample was stored at −20°C prior to GC-MS analysis.

Extraction of intracellular monoterpenoids was performed on the yeast pellet obtained after centrifugation. Cells were crushed with glass beads in 1 mL of water and after addition of eugenol (40 μg) the mixture was extracted three times with n-pentane. Organic solvent extracts were treated as described above.

Terpenol identification

Terpenol extracts were analysed by capillary GC-MS using a Varian 3300 chromatograph equipped with on-column injector coupled to an HP 5970 mass-selective detector in electron ionization mode at 70 eV. Mass spectra were recorded by means of HP ChemStation version A.03.00 and compared with those of reference compounds also used for quantification. Separation of 1 μL of extract was carried out on a 60-m, 0.32-mm ID, 0.5-μm DB-Wax (J&W Scientific) capillary column using helium as carrier gas at a constant pressure of 9 psi, 620 mbar. Separation conditions were: initial injector temperature of 70°C for 30 s, then increased to 220°C at 160°C min−1; oven temperature was programmed without initial hold time at a rate of 2.7 min−1, from 67 to 235°C (hold for 5 min).

Total amounts of linalool, citronellol and geraniol were determined using linear calibration curves with an R2 value >0.98 over the concentration range from 0 to 200 μg mL−1.

Sterol analysis

Sterol extraction after saponification of lyophilized cells and analysis were performed as previously described (Marcireau, 1990). Δ-5,7 sterols (ergosterol and ergosta-5,7-dienol) were quantified by measuring the UV absorbance at 281.5 nm (Servouse & Karst 1986). Sterol composition was determined by GC (Trace GC, Thermo Electron equipped with a on-column injector and a flame ionization detector) under the following conditions (capillary column SPB5 Supelco 15 m, 0.32-mm ID, 0.25-μm film thickness, carrier gas helium, 1.3 mL min−1): 60°C for 1 min, 30°C min−1 to 220°C, 3°C min−1 to 290°C, hold for 5 min. Cholesterol was used as internal standard.

Confocal microscopy

Exponential phase yeast cells were grown in minimum medium. Fluorescent cells were examined by confocal laser scanning microscopy using a Zeiss LSM 510 equipped with a Plan-Apochromat × 63 oil objective. Data were processed using lsm 5 image browser software (Zeiss).

Results

Expression of GES and LIS in yeast

The ORFs of the two cDNAs were cloned in yeast replicative vectors under the control of the strong yeast PMA1 promoter and as a GFP fusion. Yeast strains FY1679-28C and 5247 were transformed with these plasmids. The latter carries a gain-of-function pdr1-8 allele associated with multidrug resistance to prevent possible monoterpenol toxicity. Western blots of yeast cell extracts assayed with anti-GFP antibodies showed that GES-GFP was strongly expressed in WT and mutant strains in contrast to LIS-GFP which was only barely detected (Fig. 2a). Confocal microscopy confirmed these results and showed that GES-GFP was mainly localized in cytosol. LIS-GFP was detected in only about 10% of the cells, where it was present in small amounts mainly in aggregated structures in the cytosol (Fig. 2b).

Figure 2

Western blot analysis and localization of the GES-GFP and LIS-GFP fusion proteins expressed in Saccharomyces cerevisiae. (a) Total proteins (40 μg) extracted from FY1679-28C [pMO5] (lane 1), CC25 [pMO5] (lane 2), FY1679-28C [pMO6] (lane 3) and CC25 [pMO6] (lane 4) were hybridized with anti-GFP monoclonal antibodies. (b) Subcellular localization of the GES-GFP (1) and LIS-GFP (2) fusion proteins by confocal microscopy. Cells (FY1679-28C) presented were chosen from the most representative fluorescent cells. *For strain CC25 [pMO6], only about 10% of the cells were fluorescent and had the fluorescence pattern shown.

Monoterpenol production in WT yeast

Table 2 shows the level of geraniol and linalool excreted in minimal medium by stationary growth phase cells. As expected, no detectable terpenoid excretion was observed for the two WT yeasts used as controls. By contrast, expression of GES-GFP clearly gave rise to a significant release of monoterpenoids. Strains 5247 and FY1679-28C released about 500 μg L−1 monoterpenoid alcohols in culture medium, corresponding to about 1 μg mg−1 cell dry weight. Geraniol and linalool were the major compounds and only trace amounts of other terpenoids could be detected by GC-MS.

View this table:
Table 2

Monoterpenoids produced by various WT strains bearing a plasmid expressing the GES gene. Terpenoids were extracted from minimal medium at stationary growth phase. Data are mean values of three independent experiments

StrainGeraniol [μg L−1 OD600 nm−1 μg (mg dry weight)−1]Linalool [μg L−1 OD600 nm−1 μg L−1 (mg dry weight)−1]Ratio (geraniol/linalool)
5247
/pMAGTNDNDNDND
/pMO5559 ± 1451.1 ± 0.343 ± 410.09 ± 0.0213/1
FY1679-28C
/pMAGTNDNDNDND
/pMO5447 ± 940.91 ± 0.229 ± 100.06 ± 0.0215/1
  • ND, not detected.

  • * Assuming 490 mg cell dry weight L−1 OD600 nm−1.

  • ± confidence interval.

These results clearly show that WT yeast cells contain free GPP that can be used as a substrate by appropriate enzymes, such as GES. The lack of monoterpenoid alcohol formation in WT yeasts is therefore clearly linked to the absence of such activity, and not to the lack of GPP availability.

By contrast, in strains expressing LIS, no terpenoid formation could be detected. It is very likely that this was caused both by the low amount of protein synthesized in yeast and by the presence of the enzyme in aggregated structures, as shown in (Fig. 2a and b). Consequently, for the remainder of the study we focused only on the effect of GES.

Expression of GES in FPPS-defective yeasts

In haploid yeast, FPPS is encoded by a single gene copy of the essential ERG20 gene. Temperature-sensitive yeast mutant strains defective in FPPS have been isolated. They exhibit an additional characteristic feature – the ability to excrete geraniol and linalool (Chambon, 1991), as with the WT yeast strains bearing GES described in our study. It has been shown that this property was directly linked to a K197 to E substitution (erg20-2) in FPPS (Blanchard & Karst, 1993). Indeed, overexpression of the erg20-2 gene in a strain deleted for its chromosomal ERG20 copy led to geraniol excretion (Blanchard & Karst, 1993). The measurement of FPPS activity in vitro showed that for E197 FPPS the reaction products were 25% FPP and 75% GPP instead of 75% FPP and 25% GPP for the WT enzyme. Therefore, the excess of GPP produced by the defective FPPS is probably dephosphorylated and excreted. However, the mechanism of dephosphorylation has yet to be identified.

GES was then expressed in mutant strain CC25 bearing an erg20-2 mutation in order to test if GPP dephosphorylation is a limiting factor in geraniol formation. Table 3 (rows 1 and 2) shows a 10-fold increase of geraniol production in the strain expressing GES-GFP, which confirms that GPP dephosphorylation is indeed a limiting step in terpenoid formation in yeast.

View this table:
Table 3

Monoterpenoids produced by WT and FPPS-defective yeast strains bearing plasmids expressing the GES, ERG20 or erg20-2 genes. Terpenoids were extracted from minimal medium at stationary growth phase. Data are mean values of three independent experiments

StrainGeraniol (μg L−1 OD600 nm−1)Linalool (μg L−1 OD600 nm−1)Ratio (geraniol/linalool)
CC25
/pMAGT70 ± 1747 ± 123/2
/pMO5920 ± 25660 ± 3915/1
FY1679-28C
/pBSNDND
/pBS, pMO5447 ± 15917 ± 3326/1
/pKSNDND
/pKS, pMO5476 ± 16722 ± 2122/1
MO56
/pMAGTNDND
/pMO5797 ± 6996 ± 14319/1
MO57
/pMAGT23 ± 246 ± 101/2
/pMO5989 ± 3717 ± 13140/1
  • ND, not detected.

To increase terpenoid formation further, we overexpressed WT and E197 FPPS, in WT and ERG20-deleted strains. WT FY1679-28C did not excrete terpenoids, regardless of the source of FPPS overexpressed (Table 3, rows 3 and 5). Coexpression of GES allowed excretion of monoterpenoids (Table 3, rows 4 and 6), but at a level not increased compared with the WT strain (Table 2). By contrast, in ERG20-deleted strains, overexpression of the ERG20 or erg20-2 genes resulted in a roughly twofold increase in terpenoid formation when GES was simultaneously expressed (Table 3, rows 8 and 10). It is noteworthy that FPPS functions as a dimeric enzyme which can use either IPP and DMAPP, or IPP and GPP as substrates. In our study, FPPS can function either as a homodimer (Erg20p/Erg20p or Erg20-2p/Erg20-2p) or as an heterodimer (Erg20p/Erg20-2p). The Erg20-2p/Erg20-2p homodimer releases an increased amount of GPP, whereas both Erg20p/Erg20p homodimer and Erg20p/Erg20-2p heterodimer do not increase the available GPP pool that is used for FPP synthesis. Some of the observed differences could therefore be due to the ratio of Erg20-2p/Erg20-2p homodimers vs. Erg20p/Erg20p and Erg20p/Erg20-2p dimers.

However, the terpenoid level for ERG20-deleted strains was similar to CC25 expressing GES-GFP. Thus, FPPS activity is not the limiting factor in geraniol formation. One possible explanation is that terpenoid synthesis may be limited by the availability of IPP and DMAPP. In addition it cannot be excluded that the GES activity is limiting, despite a high level of protein observed in yeast.

Role of ATP binding cassette transporters on geraniol excretion

Monoterpenoids are generally reported to be cytotoxic compounds and have been used as antifungal drugs since ancient times. A common mechanism underlying yeast resistance to toxic compounds is the amplification of extrusion proteins belonging to the ATP binding cassette (ABC) superfamily. It was of interest to ascertain whether these transporters are also involved in geraniol excretion and/or resistance. Indeed, the 25% higher amount of monoterpenols produced in strain 5247 compared with FY1679-28C (Table 2) could be linked to the pdr1-8 gain-of-function allele known to direct a high level of resistance to cycloheximide, oligomycine, ketoconazole or 4-NQO (Carvajal, 1997). We investigated the effect of PDR1, PDR5 and SNQ2 gene deletions on strain sensitivity to exogenously added geraniol. Pdr5p and Snq2p are well-known yeast ABC transporters, whereas Pdr1p is a transcription activating factor for PDR5, SNQ2 and YOR1. The results (Fig. 3) showed that strains YCM62 bearing an SNQ2 deletion and YCM45 and YCM51 carrying both PDR5 and PDR1 deletions presented the highest sensitivity to exogenous supplied geraniol. However, it is of note that the inhibitory concentration was about 200 mg L−1, which is more than 200-fold higher that the concentrations obtained in growth medium of strains carrying GES. Nevertheless, we determined monoterpenoid production in strains disrupted for PDR1, PDR5 and SNQ2. The results obtained (Table 4) were similar to those obtained for WT strains. We also investigated geraniol and linalool intracellular contents for 5247 and PDR mutants carrying GES, but the amounts were below the resolution of the GC-MS system (data not shown). It is therefore likely that the monoterpenols produced intracellularly are excreted by single diffusion.

Figure 3

Effect of inactivation of PDR1, PDR5 and SNQ2 genes on geraniol sensitivity. Cells were serially diluted and 107–105 cells were sequentially spotted on YNB supplemented with uracil and tryptophan. Results were obtained after 60 h of incubation at 28°C. Geraniol was added from a 20 mg mL−1 ethanol stock solution.

View this table:
Table 4

Effect of pleiotropic drug transporters Pdr5p and Snq2p on terpenoid excretion by yeast expressing GES. Data are mean values of two independent experiments

StrainGeraniol (μg L−1 OD600 nm−1)Linalool (μg L−1 OD600 nm−1)Ratio (geraniol/linalool)
YCM30632 ± 37120 ± 375 : 1
YCM33517 ± 5835 ± 1215 : 1
YCM62407 ± 5768 ± 236 : 1
YCM45570 ± 2019 ± 930 : 1
YCM51676 ± 3732 ± 1121 : 1

Effect of GES expression on ergosterol level

The ergosterol content of WT yeast is about 10 μg mg−1 cell dry weight. GES expression allowed about 1–2 μg mg−1 cell dry weight monoterpenoid production in addition to sterols (Table 3). It could not be ruled out that in spite of a low isoprenoid precursor loss, GPP metabolization in GES-GFP-expressing cells could affect ergosterol biosynthesis. Therefore, we checked the effect of GES-GFP on the sterol pathway. Table 5 shows that in strain 5247 the level of ergosterol decreased by c. 30% when GES was expressed. This was associated with a dramatic change in the overall sterol composition as ergosterol precursors such as ergosta-5,7 and lanosterol disappeared, probably as a consequence of inhibition upstream from squalene synthesis. By contrast, in strain FY 1679-28C the ergosterol level remained unchanged and the sterol composition was only slightly modified. The different behaviour of these two strains may be linked to distinct genetic backgrounds (FL100- or S288C-derived strains). Indeed, the ergosterol content and the sterol composition of the control strains (Table 5, rows 1 and 3) are not the same, which could explain a different response to GPP formation.

View this table:
Table 5

Influence of GES over-expression on the sterol level in WT yeast. Sterols were extracted from lyophilized cells after growth in minimum medium. Data are mean values of two independent experiments

StrainΔ-5,7 sterols [μg (mg dry weight)−1]Percentage of individual sterols
SqualeneZymosterolErgosterolErgosta-5,7Lanosterol
5247
/pMAGT7.2 ± 1.111462177
/pMO54.7 ± 0.7ND990NDND
FY1679-28C
/pMAGT4.3 ± 0.313450525
/pMO53.9 ± 0.611364318
  • * Δ-5,7 sterol content expressed as μg (mg dry weight)−1 is determined by UV spectrum analysis.

  • Individual sterols (mean values) are determined by GC and expressed as a percentage of the total sterols.

To analyse further the effect of terpenoid production on the sterol pathway we checked two inhibitors of sterol biosynthesis in strain 5247, the FL100 derivative, which showed an effect of GES-GFP expression on sterol composition. Simvastatin inhibited hydroxymethylglutaryl co-enzyme A (HMG-CoA) reductase upstream of GPP synthesis and, as expected, strongly blocked both ergosterol and terpenoid formation (Fig. 4). Amorolfine, an inhibitor of sterol Δ8→Δ7-isomerase, acts downstream in the pathway. At 0.3 μM in the growth medium, ergosterol formation was strongly inhibited but no effect on terpenoid formation was observed (data not shown). This last result showed that geraniol formation cannot be further increased by blocking the synthesis of ergosterol, the end-product of the sterol pathway in yeast.

Figure 4

Effect of simvastatin on terpenoid and sterol levels in strain 5247 [pMO5]. Cells were cultivated in minimal medium at 30°C. Levels of 5–7 sterols were determined by measuring the UV absorbance at 281.5 nm.

Endogenous isoprenoid phosphatases in yeast

Strains defective in FPPS such as CC25 exhibited a ratio of geraniol/linalool of about 3 : 2 (Table 3). GES expression strongly and specifically increased geraniol level as the geraniol/linalool ratio increased to at least 15 : 1. This result is in good agreement with a specific synthesis of geraniol by GES; no farnesol or other sesquiterpenic alcohols could be detected in our experiments.

The natural formation of geraniol and linalool in strains exhibiting a defect in FPP synthesis is of interest as it may be linked to an endogenous enzyme activity. Faulkner (1999) reported that the LPP1 and DPP1 genes encode diacylglycerol phosphate phosphatases with isoprenoid diphosphate phosphatase activity in vitro. Moreover, it was shown that Dpp1p exhibits the highest activity and that DPP1 gene expression is induced by Zn2+ starvation (Han, 2001). It was therefore of interest to see if Dpp1p is also implicated in dephosphorylation of GPP in strains defective in FPPS such as CC25. DPP1 was overexpressed in WT and MO57 strains in place of GES. No significant increase in terpenoid formation was observed. Zn2+ starvation was equally inefficient (data not shown). These results strongly suggest that proteins other than Lpp1p and Dpp1p are involved in vivo in monoterpenoid formation.

It is also noteworthy that allylic isoprenoid diphosphates are known to be unstable at acidic pH (Rittersdorf & Cramer, 1967) and isomerization can be linked to a physicochemical process. In yeast, the low vacuolar pH (4–5) environment generated by vacuolar H+ pumping ATPase is essential for a number of cellular processes (Rashid, 2004). We observed that GPP incubated in vitro at pH 4 at 30°C decomposes into 80% linalool, 19% geraniol and 1%α-terpineol (Table 6). It is therefore possible that in strain CC25 the GPP dephosphorylation and isomerization took place in the vacuole. By contrast, in strains expressing GES-GFP in cytoplasm the GPP dephosphorylation catalysed by GES may have occurred in the cytoplasm at neutral pH, thus mainly providing geraniol.

View this table:
Table 6

Effect of pH on GPP stability

IncubationGeraniolα-TerpineolLinalool
ApH 6NDNDND
pH 418%5%77%
BpH 6100%NDND
pH 442%ND58%
  • ND, not detected.

  • Panel A: GPP (20 μg) was incubated for 36 h at 30°C at pH 4 and 6 in citrate/HCl buffer. After extraction by methylene chloride the organic phase was analysed by GC-MS.

  • Panel B: the remaining aqueous phase was adjusted at pH 10 and incubated in the presence of bovine alkaline phosphatase (110 U) for 15 h at 20°C. Terpenoids were extracted and analysed as described above.

Discussion

To determine the availability of GPP for terpenoid production in yeast, we expressed GES from Ocimum basilicum and LIS from Clarkia breweri in fusion with GFP under control of the strong PMA1 promoter.

No linalool or terpenoid formation was detected in vivo with LIS, which could be linked to both a low level of fluorescent cells and aggregated protein structures in the cytosol.

By contrast, the expression of GES allowed strong monoterpenol excretion in the growth medium (500 μg L−1). In contrast to results for previously described mutants defective in FPPS, which excreted linalool and geraniol in similar amounts, expression of GES leads to a specific excretion of geraniol. This specificity of geraniol biosynthesis in GES-GFP-expressing cells was not further investigated. It is possible that in FPPS-defective mutants, GPP is concentrated in the vacuole at acidic pH where it is spontaneously transformed into geraniol and linalool; on the other hand, in strains expressing GES-GFP, localized in cytoplasm at neutral pH, GPP is specifically metabolized to geraniol.

The absence of an effect of amorolfine, a specific inhibitor of ergosterol biosynthesis, on terpenoid formation shows that ergosterol starvation does not up-regulate the sterol pathway upstream in strain 5247, an FL100 derivative. We showed, however, that GES expression in strain 5247 clearly affects sterol biosynthesis. It is especially noteworthy that this GES effect on sterol biosynthesis was not observed for strain FY1679, an S288C derivative. The exact nature of the difference of reaction between both strains remains unknown. It may either be linked to a loss of GPP, no longer available for ergosterol biosynthesis, or to a feedback effect of geraniol produced by GES.

The potential role of known ABC multidrug transporters in cell detoxication from geraniol was also investigated. PDR1, PDR5 and SNQ2 are apparently not involved in geraniol excretion, which could indicate that monoterpenoids diffuse freely from the cell, as has been shown for short-chain alcohols.

The main finding from the present study is that in yeast, GPP is not simply an intermediate metabolite of FPP synthesis remaining tightly bound to the enzyme. Our data indicate the presence of an intracellular pool of free GPP, which opens the possibility of using yeast as an efficient and flexible engineering tool for the production of monoterpene-derived compounds.

It is noteworthy that monoterpenoids were obtained in our study in both WT and FPPS-defective mutant strains. Moreover, the maximal level of geraniol was obtained in FPPS-defective strains, confirming that in WT yeast, the absence of monoterpenoid formation is linked to the absence of GES or a similar enzyme activity and not only to the high specificity of FPPS for its GPP substrate.

This observation leads us to speculate that high enzyme activity of specific terpene synthases in aromatic plants could in a similar way shift part of the metabolic flux toward monoterpenoid production.

Acknowledgements

We are grateful to Prof. Eran Pichersky and Dr Yoko Iijima (University of Michigan, Ann Arbor, MI) for the kind gift of GES and LIS cDNAs. We acknowledge Prof. Marc Bonneu (University of Bordeaux II, France) for the pBG7 plasmid. We thank Dr Jérôme Mutterer (IBMP CNRS, Strasbourg, France) for assistance in confocal microscopy and Prof. Thomas Bach (IBMP CNRS, Strasbourg, France) for help in editing this manuscript. We thank Dr Christophe Marcireau (Sanofi-Aventis, Vitry sur Seine, France) for the YCM yeast strains carrying deletions in PDR genes. We thank the excellent technical assistance of G. Riveill and P. Claudel for terpenoid analysis and S. Meyer for pBS and pKS constructions.

Footnotes

  • Editor: Guenther Daum

References

View Abstract