Isolation, cloning and expression of a multifunctional O-methyltransferase capable of forming 2,5-dimethyl-4-methoxy-3(2H)-furanone, one of the key aroma compounds in strawberry fruits


* For correspondence (fax +49(0)931 888/5484; e-mail


Strawberry fruits contain an uncommon group of key aroma compounds with a 2,5-dimethyl-3(2H)-furanone structure. Here, we report on the methylation of 2,5-dimethyl-4-hydroxy-3(2H)-furanone (DMHF) to 2,5-dimethyl-4-methoxy-3(2H)-furanone (DMMF) by a S-adenosyl-L-methionine dependent O-methyltransferase, the cloning of the corresponding cDNA and characterization of the encoded protein. Northern-hybridization indicated that the Strawberry-OMT specific transcripts accumulated during ripening in strawberry fruits and were absent in root, petiole, leaf and flower. The protein was functionally expressed in E. coli and exhibited a substrate specificity for catechol, caffeic acid, protocatechuic aldehyde, caffeoyl CoA and DMHF. A common structural feature of the accepted substrates was a o-diphenolic structure also present in DMHF in its dienolic tautomer. FaOMT is active as a homodimer and the native enzyme shows optimum activity at pH 8.5 and 37°C. It does not require a cofactor for enzymatic activity. Due to the expression pattern of FaOMT and the enzymatic activity in the different stages of fruit ripening we suppose that FaOMT is involved in lignification of the achenes and the vascular bundles in the expanding fruit. In addition, it is concluded that the Strawberry-OMT plays an important role in the biosynthesis of strawberry volatiles such as vanillin and DMMF.


Strawberries, one of the most popular and important fruits, are cultivated almost worldwide and are consumed fresh, preserved or manufactured in various products (Hancock, 1999; Manning, 1993). Strawberry flavour consists of a huge variety of volatile compounds and was intensively studied (Honkanen and Hirvi, 1990). To date more than 300 substances have been identified (Ulrich et al., 1997). The relative abundance of individual volatiles is a ‘fingerprint’ of a particular cultivar and species. Since volatiles differ in their organoleptic properties only a relatively few of these are likely to contribute significantly to flavour. Recently, the ‘aroma value’ concept, that is the ratio of concentration to odour threshold has been applied to strawberry flavour (Schieberle and Hofmann, 1997). The studies showed that in particular the ‘fruity’ smelling esters ethyl butanoate, ethyl hexanoate and methyl 2-methylbutanoate as well as (Z)-3-hexenal (freshly cut grass), 2-methylbutanoic acid (sweaty), γ-decalactone (peach-like) and the ‘caramel-like’ smelling odourants 2,5-dimethyl-4-hydroxy-3(2H)-furanone (Furaneol®, DMHF) and 2,5-dimethyl-4-methoxy-3(2H)-furanone (mesifurane, DMMF) are likely to contribute significantly to strawberry flavor (Larsen et al., 1992; Pyysalo et al., 1979). Among these DMHF is most important because of its high concentration (up to 55 mg kg−1 strawberry fruit FW) (Larsen et al., 1992) and low odour threshold (10 p.p.b) (Schieberle and Hofmann, 1997). DMHF was first isolated from pineapples (Rodin et al., 1965) and some years later from strawberries (Ohloff, 1969). The methyl ether of DMHF – DMMF, first reported by Willhalm et al. (1965), was also identified as an aroma component of different fruits such as overripe strawberries, mango fruits and arctic brambles (Schwab and Roscher, 1997). Until now the furanones were isolated only from fruits but not from roots, stems, leaves, flowers or other plant parts (Schwab and Roscher, 1997). The β-glucopyranoside of DMHF was identified as a natural ingredient of strawberry and tomato fruits (Mayerl et al., 1989) and the malonylated derivative of DMHF-glucoside (Roscher et al., 1996) was detected in strawberries. DMHF and DMMF content in strawberries varies remarkably in the different cultivars and varieties (Douillard and Guichard, 1989; 1990). Enantiomeric analyses showed that DMHF and DMMF occur as racemates in the different fruits (Bruche et al., 1991). Several efforts have been made to clarify the biosynthesis of DMHF and its derivatives. Even though the biogenesis in fruits is yet unknown, all studies indicate that DMHF is deriving from sugar metabolism (Schwab and Roscher, 1997).

The quantification of DMHF and DMMF during fruit ripening indicated a rapid conversion of DMHF into DMMF as well as DMHF-glucoside (Pérez et al., 1996). By in vivo feeding experiments Roscher et al. (1997) showed the incorporation of 14C-label into DMMF after the application of S-[methyl-14C]-adenosyl-L-methionine (14C-SAM) and 14C-DMHF, respectively. The data supported the hypothesis that SAM is the natural source of the methyl group in the 4-methoxy compound DMMF.

Enzymatic O-methylation is catalysed by O-methyltransferases (OMTs) [EC 2.1.1.x], and involves the transfer of the SAM methyl group to the hydroxyl group of an acceptor molecule, with the formation of its methyl ether derivative and S-adenosyl-L-homocysteine (SAH) as products (Figure 1). Up to now no SAM-dependent OMT has been characterized that is able to catalyse 4-O-methylation of DMHF.

Figure 1.

Chemical structures of substrates and products of strawberry O-methyltransferase (FaOMT).

Although the chemical mechanism of methyl transfer reactions are identical, OMTs differ in selectivity with respect to the stereochemistry of the methyl acceptor molecules, as well as the substitution pattern of their phenolic hydroxyl groups (Ibrahim et al., 1998). Despite a particular specificity for phenolic substrates, plant OMTs share highly conserved domains (Ibrahim et al., 1998; Joshi and Chiang, 1998). To date, more than 10 distinct groups of SAM-OMTs that utilize SAM and a variety of substrates were described in higher plants. Individual plant OMTs accepting phenolic compounds such as o-diphenols, caffeic acid and caffeoyl-CoA as substrates were analysed with regard to lignin production. In contrast, no biochemical or molecular information is available concerning the methylation of 4-hydroxy-3(2H)-furanones.

In this study, we provide biochemical evidence for a SAM:DMHF O-methyltransferase and report on the cloning and characterization of a fruit ripening induced gene encoding an OMT, which catalyses the formation of DMMF, one of the key aroma compounds in strawberry fruits. Elaborated analysis of expression pattern and DMMF formation clearly demonstrated the correlation of OMT activity and DMMF production.


Determination of SAM:DMHF O-methyltransferase activity in strawberry fruits

Incubation of strawberry cell-free extracts with 14C- or 3H-labelled S-adenosyl-L-methionine (SAM) and DMHF gave rise to ethyl acetate-soluble radioactive product(s). The label was not present when the protein extracts were not added or previously boiled for 5 min. To confirm the identity of the formed radiolabelled compounds, ethyl acetate extracts were evaporated and analysed by TLC-autoradiography. Only one radioactive substance originating in DMHF and 14C-SAM fed extracts was detected and its Rf coincide with that of authentic DMMF. These results indicate that a methyltransferase activity, present in strawberry fruits is able to O-methylate DMHF and release DMMF.

The activity levels increased linearly with incubation time and were directly dependent on protein concentration up to 25 µg protein. Optimum activity temperature was around 37°C and the pH optimum was at about 8.5; the levels at 30°C were about 90%. The activity was irreversibly lost at incubation temperatures over 45°C, probably due to protein denaturation. Enzymatic activity was undetectable below pH 6 or above pH 10. Some methyltransferases require the presence of metal cofactors. To test whether this is the case for the DMHF O-methyltransferase from strawberry, 1 or 10 mm CaCl2, MgCl2, MnCl2, CoCl2, ZnSO4 or FeSO4 were, respectively, added to the assays. Regardless of the added metal-compound and already at 1 mm concentration, a diminution by 10% was observed indicating that a metal cofactor is not required. By gel permeation chromatography a native molecular weight of approximately 80 kDa was determined on a calibrated Superdex HR 75 column, which is within the same range of OMT's from other sources (Ibrahim et al., 1998).

Employing Lineweaver-Burk and Eadie-Hofstee equations apparent Km-values were 15 µm for SAM and 5 mm for DMHF. Keeping a constant 10 mm of the acceptor substrate DMHF, the Km for SAM was 7.5 µm. Noticeably, this is not within the range of Km-values obtained for substrates of O-methyltransferases from other sources, that were usually orders of magnitude lower (Ibrahim et al., 1998).

To test substrate-specificity of the enzyme(s), we chose several natural alcohols, in two concentrations (1 and 10 mm, Figure 2). Interestingly, only ortho-dihydroxyphenols (catechol, caffeic acid, protocatechuic aldehyde, pyrogallol) or similar ortho-dihydroxy compounds (DMHF, DTT) could serve as substrate. However, the enzyme preparation had a much lower Km for catechol (120 µm) and caffeic acid (145 µm) than for DMHF (5 mm). Substrates that contained only one OH group, such as p-coumaric- and (E)-ferulic acid, vanillic acid, p-anol, chavicol, coniferyl alcohol, phenol, hydrochinone, and o-cresol were not accepted.

Figure 2.

Substrate specificity of OMT from strawberry fruits (a).

Assays were performed as described in Experimental procedures. p-Coumaric acid (E)-cinnamic acid (E)-ferulic acid, vanillic acid, p-anol, chavicol, coniferyl alcohol, phenol, hydroquinone, and o-cresol were ineffective substrates for the OMT activity. OMT activity towards caffeic acid (white bars), catechol (grey bars) and DMHF (black bars) in three strawberry varieties during fruit ripening (b).

DMHF, DMMF and OMT activity in different varieties and maturation stages

Three strawberry varieties differing in their aroma properties were analysed and the corresponding content of DMHF and DMMF was assessed during ripening. The furanone-concentration considerably increased along fruit ripening with maximum values at the ripe stage, which is in agreement with the results obtained by Sanz et al. (1995). ‘Malach’ is an aromatic variety, which accumulated high levels of DMHF-specific OMT at the ripe stage no matter which substrate was used (Figure 2). Tamar′ and ‘Yael’ varieties showed lower levels of total furanones and thus were less aromatic. However, their cell-free extracts showed significant OMT activities towards DMHF, catechol and caffeic acid (Figure 2). OMT activities were found only in red fruits, and protein extracts from unripe fruits of any of the three varieties were devoid of OMT activity. The ratio of OMT activity towards caffeic acid, catechol and DMHF was similar in extracts from the three cultivars. Thus, the concentration ratios of the three OMTs are comparable or only one enzyme is capable of performing the methylation of the different substrates.

Identification of an OMT-coding gene from strawberry fruits

The fruit colour is a clear indication of ripening, and in most cases it is associated with flavour and aroma accumulation. Poly(A)+ RNA was isolated from Fragaria x ananassa fruits in the pink ripening stage, where DMMF formation is highest (Pérez et al., 1996) and the O-methyltransferase-transcripts were expected to reach high levels. SAM-OMT fragments were isolated by PCR using the pink ripening stage (turning) cDNA-library as a template. The amplification-primers were based on highly conserved sequences in plant OMTs corresponding to the SAM binding site. A 106-bp PCR product was obtained, which was used as a probe to screen 3 × 105 PFU of the cDNA library. The obtained 24 clones were sequenced and it became evident that all clones contain the same complete open reading frame as well as identical sequences in the 5′ untranslated region (UTR). The majority of the cDNA-inserts was about 1.6 kb in length while a few exhibited a prolonged 3′ UTR resulting in a total length of approximately 1.9 kb. Northern analysis revealed that only the 1.9 kb messenger is represented in ripening or fully ripe strawberry tissue. The shorter clones are either artifacts, formed during the first-strand cDNA synthesis or correspond to a low abundance mRNA-species. Consequently, further investigations focused on the longer cDNA. The corresponding gene was designated FaOMT (Fragaria x ananassa O-methyltransferase). The cDNA was 1876 bp in length with an open reading frame of 1098 bp, encoding a deduced polypeptide of 365 amino acids, with a calculated molecular weight of 39.817 kDa. A partial sequence of FaOMT was already published by Manning (1998) and was characterized as a ripening-related gene.

Sequence alignments with already published OMTs reveal 87% homology on amino acid level to an aspen caffeic acid O-methtyltransferase (PTOMT1), which has been tested for its substrate specificity (Figure 3) (Bugos et al., 1991). Caffeic acid and 5-hydroxyferulic acid were accepted as substrates. The highest sequence identity (91% identical amino acids) was found for a putative OMT clone isolated from Prunus dulcis, however, the substrate specificity was not reported (Swissprot Accession Q43609) (Garcia-Mas et al., 1995). The FaOMT sequence includes the consensus SAM-binding site-motifs A, B and C (Joshi and Chiang, 1998).

Figure 3.

Sequence alignment of strawberry OMT (FaOMT), OMT clone from Prunus dulcis (PRUDU) and aspen caffeic acid O-methyltransferase (PTOMT1).

Developmental expression of FaOMT

Northern analysis was performed on eight different strawberry tissues: root, petiole, leaf, flower, green fruit, white fruit, turning fruit (from white to red) and ripe fruit to demonstrate the spatial and developmental expression of FaOMT. It was barely expressed in root, petiole, leaf and flower, whereas increased expression was observed in the different developmental stages of the fruits. As expected, green and white fruits showed weak expression levels and maximum levels were reached in ripening fruits. The spatial and developmental expression pattern and nucleotide sequence of FaOMT is identical to clone FAN R1 recently published by Manning (1998). Accordingly, FaOMT is a fruit specific and ripening related gene

Heterologous expression of FaOMT

The entire coding region of the FaOMT cDNA was cloned in frame into the expression vector pGEX for production of a GST-fusion protein (glutathion S-transferase) in E. coli. The recombinant protein was isolated by GST-affinity chromatography and the FaOMT-protein was released by thrombin cleavage. Successful expression of the fusion protein was monitored by SDS-PAGE confirming the calculated molecular weights of both the fusion protein (66 kDa) and the free O-methyltransferase (40 kDa, Figure 4). Due to the fact that the native protein had a molecular weight of 80 kDa and assuming that FaOMT is its cloned counterpart, it must be concluded that the functional enzyme forms a dimer.

Figure 4.

SDS PAGE of GST-fusion protein expressed in E. coli after purification with GST affinity column (a) (E1 to E3: fraction 1–3, M: marker proteins) and purified FaOMT after cleavage of the fusion protein with thrombin (b) (M: marker proteins, P: FaOMT released by thrombin). Bands were visualized by silver staining.

Purification and characterization of the recombinant FaOMT

For functional characterization of recombinant FaOMT aliquots of the thrombin-eluate were assayed for methylation activity. The tested substrates included compounds that are natural constituents of strawberry fruits such as caffeic acid, caffeoyl CoA and DMHF, as well as those which have not been detected in strawberry fruits until now, for example catechol and protocatechuic aldehyde (Table 1). Catechol, caffeic acid and caffeoyl CoA were used as substrates that represent the major classes of compounds commonly transformed by OMTs. All substrates share an aromatic system substituted by two adjacent hydroxy groups. DMHF may also be considered a heterocyclic aromatic compound as one of its tautomeric structures is formally a dienolic furan (Rodin et al., 1965) (Figure 1) and resembles to an o-diphenol structure. The kinetic constants were measured at a saturated and constant concentration of SAM and evaluated according to Hanes (Cornish-Bowden, 1995). All tested compounds were accepted as substrate by the FaOMT-enzyme. The apparent Km values of the phenolic substrates ranged from 66 to 150 µm comparable with the results obtained with the crude strawberry enzyme. However, DMHF exhibited an Km of 440 µm, which is one order of magnitude lower than for the native enzyme (Table 1). As DMHF is a relatively unstable molecule, it could be subjected to oxidative degradation in the partially purified extract, which causes the comparable high Km-value.

Table 1.  Substrate specificity of partially purified native OMT isolated from strawberry fruits and recombinant enzyme FaOMT
SubstrateKm (µM)
Km (µM)
Vmax (pkat*mg−1)
(µM−1 min−1)
  1. Enzyme assays were carried out using 1.7–3.8 µg of the purified recombinant protein or 25 µg of the strawberry protein, 1 µM−1 mM phenolic substrates or 0.1–10 mm DMHF and 0.05 µCi 14C-SAM. Kinetic parameters were determined according Hanes (Cornish-Bowden, 1995). nd not determined

Protocatechuic aldehydend99.532.578.12 * 10−3Vanillin
Caffeic acid14514619.213.06 * 10−3Ferulic acid
Catechol120169.51.332.00 * 10−4Guaiacol
Caffeoyl CoAnd660.712.58 * 10−4Feruloyl CoA
DMHF50004400.532.95 * 10−5DMMF

If it is assumed that the dienolic tautomer of DMHF (Figure 1) fulfills the ideal conformation for successful conversion by FaOMT-activity, then the Km-value of DMHF has to be much higher in comparison with other o-diphenolic substrates. Moreover, the o-diphenolic compounds showed the typical course of a Michealis-Menten kinetic, while a different behaviour was observed for DMHF. Maximum enzyme-activity was observed at 10 mm DMHF but higher concentrations inhibited FaOMT. The enzyme methylated rapidly protocatechuic aldehyde and caffeic acid, which was demonstrated by the specific activities of 32.6 and 19.2 pkat*mg−1, respectively. The other substrates were converted less efficiently to their methylated counterparts, reflected in specific activities of 0.5–1.3 pkat*mg−1. Catechol seems to be rather well accepted, while caffeoyl-CoA and DMHF are methylated to a lesser extend. Radio-HPLC analysis of the formed radioactively labelled compounds showed only mono-methylated products (Figure 5). The broad substrate specificity was rather unexpected and observed for the first time for an enzyme involved in phenylpropanoid formation and strawberry fruit flavour biosynthesis.

Figure 5.

HPLC-RP18 separation of a reference mixture consisting of DMHF (tR = 9 min) and DMMF (tR = 13 min), UV detection at 284 nm (a); as well as HPLC-RP18 separation of the diethyl ether extract obtained from a FaOMT assay with 10 mm DMHF (b).

Fractions of 1 ml were collected and determined by LSC. Panel b was background substracted.

Correlation of FaOMT expression, FaOMT activity and levels of DMHF and DMMF during strawberry fruit development

FaOMT enzyme activity was only detected in red ripe strawberry fruits, whereas protein extracts isolated from green immature fruits contain negligible OMT activity. Comparison of FaOMT expression and DMMF formation showed that the concentration of the transcript hybridising to FaOMT increased at the turning stage (Figure 6). Shortly after the appearance of FaOMT DMMF was formed in the fruits. These experiments clearly demonstrate the direct correlation between FaOMT expression, FaOMT activity and DMMF formation in ripening strawberry fruits.

Figure 6.

Comparison of furanone formation and FaOMT expression pattern in different strawberry tissues: green fruit (G), white fruit (W), turning fruit (early) (T1), and turning fruit (late) (T2).

Concentration of DMHF (♦), DMMF (•) and glycosylated DMHF (▴) during strawberry fruit development (a). Northern blot analysis of strawberry total RNA (10 µg per lane) probed with 32P-labelled FaOMT (b). Equal amounts of RNA were run in each lane as indicated by the intensities of rRNA bands prior to transfer (c).


In contrast to the data obtained with plant extracts, the results from heterologous expression studies of plant secondary metabolite enzymes have shown that their activity covers a surprisingly broad substrate spectrum. The strawberry alcohol acyl-CoA transferase (SAAT), for example, was capable of utilizing short- and medium-chain, branched, and aromatic acyl-CoA molecules as cosubstrates (Aharoni et al., 2000). Work on Thalictrum tuberosum OMTs suggested that some biosynthetic enzymes species were common to both phenylpropanoid and alkaloid anabolism (Frick and Kutchan, 1999).

A comparison of the amino acid sequence of 56 SAM-OMTs from different plants showed that plant OMTs fall into two distinct groups: Pl-OMT I and Pl-OMT II. The length of Pl-OMT I proteins vary from 231 to 248 amino acids, whereas the length of Pl-OMT II enzymes is 344–383. Unlike Pl-OMT I members that are known to utilize only a pair of substrates, the members of the Pl-OMT II group can accept a variety of substrates and are multifunctional enzymes (Li et al., 1997). FaOMT encodes a sequence of 365 amino acids and accepts a substrate spectrum for compounds containing o-diphenol structures. Accordingly, it was assigned to the Pl-OMT II group.

Anthocyanin biosynthesis takes place concomitantly with fruit ripening and DMMF formation in strawberry. Ripening-related gene sequences that code for proteins involved in key metabolic events including anthocyanins biosynthesis have already been isolated from strawberry (Manning, 1998), and were not found active in green fruits. Cyanidin, an anthocyanin precursor in strawberry contains the same ortho-di-hydroxyphenol structure typical for this OMT activity and peonidin-glucoside, the O-methyl derivative of cyanidin glucoside has been found in strawberry fruits and strawberry cell suspensions (Nakamura et al., 1998). We assume that DMHF and cyanidin methylation occurs as a non-directed side-effect action of caffeic acid methylation, which increases during fruit ripening. Unfortunately, cyanidin was not available at the time of the investigation.

Nam et al. (1999) characterized mRNAs differentially expressed during ripening of wild strawberry (Fragaria vesca L). One of the sequences encoded a putative caffeoyl-CoA 3-O-methyltransferase (CCOMT). As ripe strawberry fruits are rich in phenolic compounds but produce low level of lignin it was concluded that the wild strawberry enzyme is involved in the synthesis of phenols and their derivatives. A sequence comparison revealed no significant homology between FaOMT and wild strawberry CCOMT indicating different functions for both O-methyltransferases during ripening.

The enzyme FaOMT considerably catalyses the transformation of caffeic acid to ferulic acid. Ferulic acid is a precursor of lignin, and in grasses ferulate moieties play an important role in other cell wall polymerization processes, such as cross-linking between polysaccharides. Many published caffeic acid OMTs were related to lignin biosynthesis (Joshi and Chiang, 1998) but there are only a few examples of plant SAM-OMTs with distinct substrate specificities. Together with the results from Northern hybridizations, which clearly show a fruit-maturation specific FaOMT-expression, it is suggested that the enzyme contributes to formation of fibres and tracheary elements. In the receptacle of strawberries occurrence of lignins is limited to the vascular bundles, while the achenes are lignified strongly (Suutarinen et al., 1998). In accordance with other plant caffeic acid OMTs, the strawberry enzyme prefers the free acid compared with the CoA ester (Inoue et al., 2000; Maury et al., 1999). Due to the expression pattern of FaOMT in the different stages of fruit ripening we assume that in the beginning of fruit development FaOMT could be involved in lignification of the vascular bundles in the expanding fruit. In later stages (turning and ripe fruit) FaOMT activity probably provides the precursors for lignification of the achenes. It is interesting to note that PAL (Phenylalanine ammonia-lyase), the key enzyme of phenylpropanoid metabolism providing caffeic acid shows a similar activity pattern in ripening strawberries. (Given et al., 1988a, 1988b).

On the other hand, FaOMT also plays an important role in the biosynthesis of strawberry volatiles, because it efficiently converts protocatechuic aldehyde, a compound that has not yet been found in strawberry, to vanillin. Vanillin has been reported to contribute to strawberry flavour in wild strawberries and in few commercial cultivars (Hirvi and Honkanen, 1982). In addition, FaOMT also catalyses the conversion of DMHF to DMMF and this is the first report on an OMT capable in methylating a 4-hydroxy-3(2H)-furanone. Although the substrate homology with o-diphenols is not obvious at first glance, a structure comparison of the dienol tautomer with o-diphenols uncovers a certain relationship (Figure 1). DMHF and DMMF are key components of the strawberry flavour (Larsen et al., 1992; Pyysalo et al., 1979) as they show the highest aroma values in strawberry extracts (Schieberle and Hofmann, 1997). Strawberry fruits contain on average 0.7–55 mg DMHF kg-1 FW corresponding to 5.5–430 µm, respectively (Douillard and Guichard, 1989; Larson et al., 1992). These values are in the same order of magnitude as the Km value of DMHF determined for FaOMT although compartmentalisation, as suggested by Sen et al. (1991), was not yet taken into account. The Km values obtained for SAAT, another enzyme with relatively low substrate specificity are even higher (e.g. 46 mm for butanol) (Aharoni et al., 2000). Furthermore, DMHF formation always precedes DMMF formation in ripening strawberry fruits (Sanz et al., 1995). Consequently, we propose an in planta methylation of DMHF by FaOMT. In contrast to caffeic acid or protocatechuic aldehyde, the methylation of DMHF proceeds quite slowly, but unequivocally the methoxy derivative descends from DMHF by OMT activity (Figure 5). This finding and the fact that DMHF and DMMF were extracted as racemates from natural sources support the hypothesis of Roscher et al. (1997) who proposed the presence of an O-methyltransferase in strawberry fruits based on in vivo feeding studies. The assumption that DMHF and DMMF are generated by two independent biosynthetic pathways (Bruche et al., 1995) seems unlikely.

The broad substrate specificity of FaOMT was somewhat surprising (Pichersky and Gang, 2000). Due to the huge number of flavour compounds found in strawberry Zabetakis and Holden (1997) suggested that it is not possible for each substance to have its ‘own’ enzymes. Accordingly, the side activity of the caffeic acid OMT involved in phenylpropanoid metabolism of strawberry fruits is responsible for the formation of DMMF. Consistent with this view, it has been reported that a single OMT from Chrysosplenium americanum can methylate both flavonoid and phenylpropanoid compounds (Gauthier et al., 1998).

The result of this work suggests that evolution of secondary metabolites does not proceed step by step. Due to the presence of multifunctional enzymes new plant metabolites can be converted by several already existing biocatalysts. Formation of DMHF probably led to the immediate production of DMMF and the corresponding glucoside. Thus, at a specific moment in evolution three new products evolved simultaneously.

Although strawberry aroma consists of a variety of attractive flavour compounds such as esters of aliphatic alcohols and acids, γ-lactones, hexenals, and 3(2H)-furanones, knowledge about their formation pathways is limited. Only recently, biochemical evidence for involvement of a strawberry alcohol acyl-CoA transferase (SAAT) gene in the formation of fruity esters was provided by characterising the recombinant protein expressed in E. coli (Aharoni et al., 2000). This paper presents biochemical information of a second enzyme involved in the formation of strawberry fruit volatiles.

Experimental procedures

Plant material

Plant tissue was collected from strawberry plants (Fragaria × ananassa) cv. Elsanta, Calypso grown under glasshouse conditions. The material was used either fresh or frozen in liquid nitrogen and stored at − 80°C until use. Fresh or frozen fruits of Fragaria x ananassa cv. Tamar, Yael and Malach were obtained from the Volcani Institute, Israel.

Extraction of volatiles

Fresh strawberry fruits (30 g) were homogenized in a food processor (Braun, Marktheidenfeld, Germany) and extracted with 50 ml methyl tert-butyl ether containing 10 µg internal standard (isobutylbenzene) by shaking for 2 h. The organic phase was dried on anhydrous Na2SO4 and evaporated under nitrogen to 1 ml.

GC-MS analysis

Volatile compounds were analysed on a HP-GCD apparatus equipped with an HP-5 (30 m × 0.25 mm) fused-silica capillary column. Helium (1 ml min−1) was used as a carrier gas. The injector temperature was 250°C, set for splitless injection. The oven was set to 50°C for 1 min, then the temperature was increased to 200°C at a rate of 4°C min−1. The detector temperature was 280°C. Mass range was recorded from 45 to 450 m z−1, with electron energy of 70 eV. Identification of the main components was done by comparison of mass spectra and retention time data with those of authentic samples and supplemented with a Wiley GC-MS library. The quantitative analyses were determined using isobutylbenzene as an internal standard.

Preparation of crude cell-free extracts

Fresh or frozen strawberry fruits were cut into slices (4–5 g), weighed and placed in a chilled mortar. The fruits were then ground with a pestle in the presence of 0.5 g PVPP to spell-out phenolic materials, and 15 ml extraction buffer A (100 mm Tris–HCl pH 8.5 containing 10% glycerol, 5 mm Na2S2O5, 10 mm 2-mercaptoethanol, 1% PVP-10) was added. The slurry was centrifuged at 20 000 g for 10 min at 4°C. The supernatant (crude extract) was used for further purification steps and enzymatic assays. Inclusion of PVP and PVPP during the extraction was crucial for stabilizing the enzymatic activity, which could be kept for more than 3 months at − 20°C without an apparent loss of activity.

Partial purification

Control assays conducted without added DMHF also contained soluble radiolabelled product (10–30% in crude cell-free extracts), indicating the presence of some endogenous substrates. In order to remove these internal substrates, partial purification was conducted using adsorbance-, gel filtration- and ion exchange chromatography. This effectively removed almost all the contaminants, and the activity was stable after the purification.

All procedures were performed at 0–4°C using buffer B (50 mm Tris–HCl pH 8.5 containing 10% glycerol, 5 mm Na2S2O5, 10 mm 2-mercaptoethanol, 1 mm EDTA). Proteins were purified as follows:

Four ml of the crude extracts were loaded to XAD-2 columns (1 × 15 cm). Column equilibration and protein elution were done with buffer B. Fractions were tested for enzyme activities and active fractions were combined. Three ml of the XAD-active fractions was loaded to Bio-gel P-6 column (1 × 10 cm). Column equilibration and protein elution were done with buffer B. Fractions were tested for enzyme activity and active fractions were combined. Three ml of the P-6-active fractions were loaded to DE52 column (1 × 3 cm). Column equilibration was done with buffer B. Protein elution was done with 10 ml of buffer B in gradient of 0–1 m KCl. Fractions were tested for enzyme activity and active fractions were combined and used for further tests.

Assays of enzyme activity in strawberry protein extracts

The standard assay mixture consisted of 30 µl buffer B, 50 µl enzyme solution, 10 mm DMHF, and 10 µl 3H-SAM (S-[methyl-3H]-adenosyl-L-methionine, 15 Ci mmol−1), in a total volume of 100 µl. The mixture was incubated at 30°C for 1–2 h. The reaction was stopped by adding 10 µl 2 N HCl. Then 1 ml ethyl-acetate was added to each tube, mixed and centrifuged for 1 min at 20 000 g to separate the phases. The upper ethyl acetate phase layers containing the radioactive labelled enzyme products were added to scintillation fluid (4 g l−1 2,5-diphenyloxazol (PPO) and 0.05 g l−1 2,2′-p-phenylen-bis(5-phenyloxazol (POPOP) in toluene) and the radioactivity was quantified by scintillation. To confirm the identity of the biosynthetic products, similar incubations were performed, except that 14C-labelled SAM (S-[methyl-14C] adenosyl-L-methionine, 55 mCi mmol−1) was used. In this case the ethyl acetate layer was evaporated to a volume of 20 µl using a gentle stream of N2, and analysed by TLC-autoradiography using Silica gel 60 F254 plates developed with pentane: diethyl ether (5 : 1). Spots were visualized by UV light and radioactive spots detected by autoradiography.

Gel permeation chromatography

The molecular mass of the native enzyme was determined by gel filtration chromatography through a Superdex 75 Hiload Prep 16/60 (FPLC, Amersham Pharmacia Biotech, Freiburg, Germany), using buffer B at flow rate of 1 ml min−1, and compared with the molecular mass of known proteins.

CDNA library

Poly(A)+ RNA isolated from ripening strawberries was used to construct a cDNA library in the Uni-ZAP® XR vector (Stratagene, La Jolla, CA, USA) according to the manufacture's instructions.

Cloning and expression of FaOMT

Two degenerate primers (MTsI 5′-GTI-GAC/T-GTI-GGI-GGI-GGI-ACI-GGI-GC-3′; MTasII 5′-GGI-GCA/G-TCC/T-TCI-ATI-ACA/G-TGI-GG-3′) based on two highly conserved regions of plant O-methyltransferases (Frick and Kutchan, 1999) were synthesized and used for PCR amplification. The resulting PCR product of 106 bp was cloned into pCR®2.1 vector (Invitrogen, Karlsruhe, Germany). This fragment was used as probe for an initial screening of the cDNA providing 24 OMT-related clones with identical sequences but different sizes. The clone with the longest cDNA-insertion, designated FaOMT (Fragaria×ananassa O-methyltransferase) was chosen for further investigations.

FaOMT-cDNA was cloned into a pGEX-4T-2 fusion vector (Pharmacia, Freiburg, Germany) and protein synthesis was induced following the supplier's protocol. After lysis of the bacteria by sonification pre-equilibrated GST-affinity resins (Stratagene) were added to the supernatant. After 1.5 h of incubation at 4°C the resins were spun down at 500 g for 5 min and washed five times with EB (140 mm NaCl, 2.7 mm KCl, 10 mm Na2HPO4, 1.8 mm KH2PO4, 5 mmβ-mercaptoethanol, 5% glycerol, pH 7.2–7.4). The beads were resuspended in EB including thrombin (0.2 NIH Units thrombin/100 µl resin). After incubation overnight at 4°C with gentle agitation the supernatant was collected and used for SDS-PAGE or enzyme activity assays. Protein concentration was determined by the Bradford method (Bradford, 1976) using BSA as a standard.

Enzyme assay

Enzyme activity was tested in EB buffer containing 1 µm−1 mm substrate, 9.1 µm adenosyl-L-[methyl-14C]-methionine (55 m Ci mmol−1) and 1.8–3.7 µg protein in a final volume of 100 µl. After incubation at 30°C for 15 min/30 min the reaction was stopped by adding 6 µl of 6 N HCl. If caffeoyl-CoA was used as substrate, the assays were treated as described by Meng and Campbell (1998). The reaction mixtures were extracted twice with 600 µl of ethyl acetate, respectively. Organic layers were combined and analysed by liquid scintillation counting LSC (Packard Tri-Carb Liquid Scintillation Analyser, Meriden, CT, USA).

Identification of the methylated products by HPLC

Enzyme assays were stopped and extracted twice with 700 µl of diethyl ether. The organic layers were pooled, dried, concentrated and transferred into 100 µl water and analysed by high performance liquid chromatography (HPLC). HPLC separations were carried out on an Eurospher 100 C-18 column (25 cm × 4.0 mm i.d., particle size 5 µm, Knauer, Berlin, Germany) using a linear gradient with a flow rate of 1 ml min−1. The gradient proceeded from 95% water acidified with 0.05% formic acid and 5% acetonitrile to 100% acetonitrile in 30 min. Fractions of 1 ml each were collected, scintillation cocktail Emulsifier-Safe (Packard BioScience, Groningen, the Netherlands) was added and analysed by LSC (LKB Rackbeta 1214, Pharmacia).

Isolation of RNA and Northern-blot hybridization

For Northern analysis, RNA was isolated from different strawberry tissues as described by Manning (1991). Total RNA (10 µg) was separated on a formaldehyde gel and transferred by capillary transfer to a nylon membrane (Hybond N, Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany). The FaOMT cDNA probe was generated by random labelling oligonucleotide priming (Feinberg and Vogelstein, 1984). Hybridisation was carried out at 68°C overnight in Roti-Hybri-Quick solution (Roth Chemikalien, Karlsruhe, Germany). The membrane was washed consecutively in 1 × SSC, 0.1% SDS and 0.1 × SSC, 0.1% SDS at 68°C and exposed to X-ray retina film (XBD) at − 80°C.


We thank Till Beuerle and Eran Pichersky for providing caffeoyl-CoA as well as Asaph Aharoni for helpful discussions. We are indebted to Beate Otto and Martin Eckert for help in heterologous expression techniques and gene isolation. This work was supported by GIF and SFB 567.