To assess the abilities of commercial wine lactic acid bacteria (LAB) to synthesize potentially flavour active fatty acid ethyl esters and determine mechanisms involved in their production.
To assess the abilities of commercial wine lactic acid bacteria (LAB) to synthesize potentially flavour active fatty acid ethyl esters and determine mechanisms involved in their production.
Oenococcus oeni AWRI B551 produced significant levels of ethyl hexanoate and ethyl octanoate following growth in an ethanolic test medium, and ester formation generally increased with increasing pH (4·5 > 3·5), anaerobiosis and precursor supplementation. Cell-free extracts of commercial O. oeni strains and Lactobacillus plantarum AWRI B740 were also tested for ester-synthesizing capabilities in a phosphate buffer via: (i) acyl coenzyme A: alcohol acyltransferase (AcoAAAT) activity and (ii) reverse esterase activity. For both ester-synthesizing activities, strain-dependent variation was observed, with AcoAAAT activity generally greater than reverse esterase. Reverse esterase in O. oeni AWRI B551 also esterified 1-propanol to produce propyl octanoate, and deuterated substrates ([2H6]ethanol and [2H15]octanoic acid) to produce the fully deuterated ester, [2H5]ethyl [2H15]octanoate.
Wine LAB exhibit ethyl ester-synthesizing capability and possess two different ester-synthesizing activities, one of which is associated with an acyl coenzyme A: alcohol acyltransferase.
This study demonstrates that wine LAB exhibit enzyme activities that can augment the ethyl ester content of wine. This knowledge will facilitate greater control over the impacts of malolactic fermentation on the fruity sensory properties and quality of wine.
Wine lactic acid bacteria (LAB) are important micro-organisms in grape vinification, particularly recognized for conducting malolactic fermentation (MLF), which is the enzymatic conversion of l-malic acid to l-lactic acid and CO2. Oenococcus oeni is the principal LAB species associated with MLF, largely due to its ability to survive and grow in the harsh wine environment of low pH and the presence of alcohol (Wibowo et al. 1985; Versari et al. 1999; Swiegers et al. 2005; Costantini et al. 2009). In addition to deacifidifying wine through MLF, wine LAB possess a diversity of other metabolic activities that potentially impact on wine sensory properties, including the formation of diacetyl and volatile sulfur compounds, the metabolism of polysaccharides and glycosidase and esterase activities (Davis et al. 1988; Henick-Kling 1993; Guilloux-Benatier et al. 2000; D'Incecco et al. 2004; Pripis-Nicolau et al. 2004; Swiegers et al. 2005; Matthews et al. 2006, 2007; Dols-Lafargue et al. 2007; Sumby et al. 2010; Knoll et al. 2011a, 2012).
Of the many fermentation-derived volatile compounds that contribute to wine aroma and flavour, the ester profile is of major importance for fruity sensory properties and overall quality of wine. In particular, certain ethyl and acetate esters have been shown to be key contributors to the berry fruit characteristics of red wine (Escudero et al. 2007; Pineau et al. 2009). Moreover, a number of studies have shown that MLF can modulate the wine ester profile (Sumby et al. 2010; López et al. 2011; Abrahamse and Bartowsky 2012; Knoll et al. 2012; Malherbe et al. 2012), and the potential for wine LAB to modulate wine ester content and fruity sensory properties has been demonstrated (Gámbaro et al. 2001; du Plessis et al. 2002; Costello et al. 2012).
In wine, beer and other food fermentations, ester concentration is the product of net synthesis and hydrolysis reactions through the activity of several enzymes, including esterases, lipases and alcohol acyl transferases (Cristiani and Monnet 2001; Liu et al. 2004b; Saerens et al. 2010; Sumby et al. 2010). Enzyme activity associated with the biosynthesis of esters has been well documented for a number of fermentation micro-organisms. For example, the formation of medium-chain fatty acid ethyl esters by Saccharomyces cerevisiae yeast is catalysed by two acyl-CoA: ethanol O-acyltransferases, Eeb1 (ethyl ester biosynthesis gene 1) and Eht1 (ethanol hexanoyl transferase) (Saerens et al. 2006, 2010). Further, dairy LAB exhibit two enzyme activities in the biosynthesis of ethyl esters, an alcohol acyltransferase (alcoholysis) reaction involving the transfer of fatty acyl groups directly from acylglycerol substrates to alcohols and esterification. Both enzyme activities have been shown to be catalysed by esterases, with the former alcohol acyltransferase reaction considered the major mechanism of ethyl ester synthesis in these bacteria (Liu et al. 2003, 2004a,b). In contrast, wine LAB have been shown to exhibit only hydrolysing esterase activities (Davis et al. 1988; Matthews et al. 2006, 2007). Recently, an esterase gene (EstB28) has been cloned from O. oeni, and this was shown to retain hydrolysing activity under winemaking conditions (10–20°C, pH 3·5, 14% v/v ethanol) (Sumby et al. 2009).
Although changes to wine ester content during MLF are indicative of ester turnover, there is a lack of knowledge concerning the ester-synthesizing capabilities of wine LAB. Further understanding of such capability in wine LAB, as well as the mechanisms and factors governing this metabolism, is critical to gaining greater control over the impact of MLF on the profile of esters and fruity sensory properties in wine.
This study investigated O. oeni-driven biosynthesis of medium-chain fatty acid ethyl esters following growth in a complex test medium and in varying conditions such as oxygen availability, pH and precursor supplementation. Potential mechanisms of ester biosynthesis by wine LAB strains via alcohol acyltransferase and reverse esterase activities were also investigated. For the purposes of this study, reverse esterase activity is defined as the enzymatic capability of ester biosynthesis by direct esterification of carboxylic acid and alcohol substrates, without the involvement of intermediates.
Commercial malolactic starter culture strains (lodged into the Australian Wine Research Institute culture collection) are listed in Table 1. All strains were maintained on MRSA agar medium [de Man, Rogosa and Sharpe (MRS) medium (Amyl, Dandenong, Australia) supplemented with 20% preservative-free apple juice] and stored at 4°C. Several of these strains were also maintained on Protect bacterial preserver beads (Technical Service Consultants, Heywood, UK) and stored at −20°C.
|LAB strain||Ester synthesis by AcoAAAT activity (μg l−1 mg−1 protein)b|
|Oenococcus oeni AWRI B548||859·0 ± 51·8|
|O. oeni AWRI B549||285·2 ± 25·8|
|O. oeni AWRI B550||143·8 ± 3·9|
|O. oeni AWRI B551||169·2 ± 6·1|
|O. oeni AWRI B552||418·3 ± 18·2|
|O. oeni AWRI B553||513·8 ± 31·6|
|O. oeni AWRI B584||305·6 ± 0·9|
|O. oeni AWRI B585||171·7 ± 3·6|
|Lactobacillus plantarum AWRI B740||234·2 ± 3·4|
LAB strains were precultured in a modified MRS medium (MRS-FMEt3%) for approximately 7 days at 27°C. MRS-FMEt3% medium comprised MRS supplemented with d-fructose (10 g l−1), dl-malic acid (10 g l−1) and ethanol (3% v/v), at pH 4·5. In experiments utilizing cell-free extracts (CFE), wine LAB strains were further cultured for approximately 7 days at 27°C in a second preculture medium (MRS-FMEt5%). The latter medium had the same composition as MRS-FMEt3% medium, with the exception that the ethanol concentration was 5% v/v. Both preculture media were sterilized by filtration with a sterile membrane filter (0·22-μm pore size).
Ester synthesis associated with wine LAB growth was determined for O. oeni AWRI B551 in MRS-FMEt5% medium. The composition of this medium was further modified to assess the effects of supplementation with precursor substrates (butanoic acid, 5 mg l−1; hexanoic acid, 5 mg l−1; octanoic acid 5 mg l−1; and 3-methylbutanol, 250 mg l−1) (Sigma-Aldrich, St Louis, MO, USA), at pH values 3·5 and 4·5. In addition, the effects of anaerobic culture conditions (anaerobic chamber; Coy, Grasslake, MI, USA) were investigated, such that the effects of the following variables on ester biosynthesis were assessed in a full factorial design experiment: (i) with or without precursor supplementation, (ii) at pH 3·5 and 4·5 and (iii) in aerobic or anaerobic cultures. Aliquots (50 ml) of MRS-FMEt5% test medium were dispensed into 100-ml sterile glass Schott bottles and inoculated (0·5% v/v) with O. oeni AWRI B551 precultured in MRS-FMEt3% as described above. Duplicate inoculated and uninoculated media were incubated at 22–23°C, either aerobically or anaerobically as required. Uninoculated test media served as controls to correct for background ester levels. Cultures were grown to stationary phase for 9 and 14 days at pH 4·5 and 3·5, respectively. Cells were subsequently removed by centrifugation (approximately 2600 g, 15 min, 4°C), and each assay was stored at −20°C prior to GC-MS analysis.
Investigation of ester synthesis via alcohol acyltransferase and reverse esterase activities was performed using CFE of wine LAB strains and utilized procedures based upon methods described by Hosono et al. (1974), Yamauchi et al. (1989), Liu et al.(1998, 2003) and Abeijón-Mukdsi et al. (2009). For CFE preparations, cells from final preculture in MRS-FMEt5% were harvested by centrifugation (approximately 14 000 g, 30 min, 4°C), washed twice in 0·1 mol l−1 phosphate-buffered saline (PBS; comprising 0·1 mol l−1 potassium phosphate buffer, 0·1 mol l−1 NaCl, pH 5·9) and resuspended (20 × concentration) in lysis buffer (PBS containing 10% glycerol, pH 5·8–5·9). In some cases, cells were washed in lysis buffer. Wet weight of washed cells was approximately 100–200 mg ml−1. CFE were subsequently prepared from washed cells by addition of lysozyme (1·0 mg ml−1) and cell sonication (Hielscher UP200S probe, Teltow, Germany) for approximately 18–24 min. Cell suspensions were maintained on ice during sonication. Clarified CFE sublots were snap-frozen in liquid nitrogen and stored at −80°C for subsequent use in CFE assays.
Screening LAB strains for ester biosynthesis via AcoAAAT activity was conducted in an assay medium (1·0-ml final volume) containing 770 μl of 0·1 mol l−1 PBS (pH 5·9), 100 μl of CFE (0·1–5 mg ml−1 protein), 30 μl of 5·5 mmol l−1 hexanoyl-CoA and 100 μl of 5 mol l−1 ethanol. The final concentrations of hexanoyl-CoA and ethanol substrates in the reaction mix were 0·165 mmol l−1 and 0·5 mol l−1, respectively. All assays were conducted in glass screwcap vials (20 ml, septa in cap) in duplicate and incubated statically at 26–28°C for approximately 17 h, after which 9 ml of Milli-Q (Millipore Corporation, Billerica, MA, USA) purified water was added to the reaction and vials stored at −20°C prior to GC-MS analysis. Controls utilizing heat-inactivated (70°C for 10 min) CFE of respective strains and also lacking CFE were included.
Assays for reverse esterase activity used a similar protocol as for AcoAAAT activity, except that hexanoyl-CoA was replaced with a fatty acid (butanoic acid, hexanoic acid or octanoic acid) substrate (0·2 mmol l−1 final concentration), and the reaction was undertaken in 0·1 mol l−1 PBS (pH 5·9). In addition, substrate specificity of the reverse esterase activity was examined by the replacement of: (i) ethanol with 1-propanol and (ii) octanoic acid and ethanol with the respective polydeuterated substrates, [2H15]octanoic acid (Sigma-Aldrich) and [2H6]ethanol (Cambridge Isotope Laboratories, Andover, MA, USA). Assays utilizing respective heat-inactivated CFE and omission of CFE served as controls. In the case of assays investigating substrate specificity of reverse esterase, diluted reactions were stored at 4°C prior to immediate GC-MS analysis.
Ethyl esters produced by LAB in culture medium and CFE assays were quantified using a stable isotope dilution analysis method similar to that described by Siebert et al. (2005), employing headspace solid-phase microextraction coupled with gas chromatography–mass spectrometry (HS-SPME-GCMS). An Agilent Technologies 6890 gas chromatograph was equipped with Gerstel MPS2 multi-purpose sampler and coupled to an Agilent 5973N mass selective detector. Instrument control and data analysis utilized Agilent G1701DA ChemStation software. The GC was fitted with a 30 m × 0·25 mm Agilent fused silica capillary column DB-Wax, 0·25-μm film thickness with a 2 m × 0·25 mm retention gap. The carrier gas was helium (Ultra High Purity; BOC, Adelaide, Australia), linear velocity 38 cm s−1 and flow rate 1·0 ml min−1 in constant flow mode. The oven temperature was started at 40°C, held at this temperature for 1 min, increased to 130°C at 5°C min−1, increased at 40–240°C and held at this temperature for 5 min. The inlet was fitted with a deactivated borosilicate glass solid-phase microextraction (SPME) inlet liner (0·75 mm I.D.; Supelco, Sigma-Aldrich) and held at 220°C. The SPME fibre was desorbed in splitless mode, and the splitter, at 35 : 1, was opened after 36 s. The fibre was allowed to bake in the inlet for 10 min. The temperatures of the mass spectrometer quadrupole, source and transfer line were set at 106, 230 and 250°C, respectively. Positive-ion electron impact spectra at 70 eV were recorded in selective ion monitoring (SIM) with 25-ms dwell time (Table 2). For SPME of supernatant MRS-FMEt5% media following growth of O. oeni AWRI B551, 1·0-ml aliquot of sample was added to a 20-ml glass vial containing 9·0 ml of Milli-Q water, and the vial was immediately crimp-capped (magnetic, Teflon lined silicone septum, Grace, Deerfield, IL, USA). For SPME of ester biosynthesis assays using CFE described above, prediluted reaction mixtures were utilized in situ in 20-ml glass vials immediately upon thawing. To each vial, a combined internal standard solution (100 μl) comprising [2H5]ethyl butanoate, [2H5]ethyl hexanoate and [2H5]ethyl octanoate (each at approximately 100 μg l−1 in the SPME vial) was subsequently injected through the septum, and the contents were mixed. The vial and contents were then heated to 35°C, and a polydimethylsiloxane (PDMS, red; Supelco) 100-μm fibre was exposed to the headspace for 10 min.
|Internal standard||RT (min)||Target Ion m/z||Qualifier Ions (%)||Analyte||RT (min)||Target Ion m/z||Qualifier Ions (%)|
|[2H5]ethyl butanoate||3·82||93||106 (17) 121 (6)||Ethyl butanoate||3·87||88||101 (17) 116 (9)|
|[2H5]ethyl hexanoate||7·47||93||106 (31) 120 (10)||Ethyl hexanoate||7·55||115||99 (586) 88 (997)|
|[2H5]ethyl octanoate||12·40||106||177 (5) 93 (259)||Ethyl octanoate||12·48||101||172 (5) 88 (246)|
In experiments investigating substrate specificity of reverse esterase activity, end-product esters were identified in accordance with reference mass spectra and anticipated retention times for propyl octanoate (target ion m/z 145; qualifier ions m/z 127, 186; ret. time 14·54 min) and [2H5]ethyl [2H15]octanoate (target ion m/z 96; qualifier ions m/z 192, 142; ret. time 12·01 min). The concentrations of ethyl octanoate, propyl octanoate and [2H5]ethyl [2H15]octanoate synthesized in these experiments were determined from respective target ion responses relative to [2H5]ethyl octanoate (100 μg l−1) as the internal standard.
Protein concentration of the CFE preparations was determined using the methodology based upon that described by Bradford (1976), using dye reagent (Bio-Rad Laboratories, Hercules, CA, USA) with bovine serum albumin as standard and absorbance (595 nm) measured in microplate format. Protein concentrations were corrected for lysozyme (1·0 mg ml−1) added during CFE preparation.
To assess whether increases in ester concentration following bacterial growth in MRS-FMEt5% medium were significant, two-way analysis of variance (anova) and means comparison tests [Fisher's least significant difference test (P < 0·05)] of pH 3·5 and pH 4·5 data sets were undertaken. JMP 5.0.1 (SAS Institute, Cary, NC, USA) was used for data analysis.
To study ethyl ester-synthesizing ability of O. oeni, ethyl ester concentrations were quantified for O. oeni AWRI B551 at stationary phase in MRS-FMEt5% test medium. The effects of several culture parameters (pH, aerobic or anaerobic conditions and precursor supplementation) on this ability were also investigated. As shown in Fig. 1, O. oeni AWRI B551 produced significant (P < 0·05) concentrations of ethyl octanoate and also ethyl hexanoate in several treatments. In contrast, there were no significant increases in ethyl butanoate, and a small but significant increase in 3-methylbutyl acetate occurred in one treatment combination (pH 4·5, anaerobic, with precursor supplementation) (data not shown).
The ability of O. oeni AWRI B551 to produce ethyl octanoate and ethyl hexanoate was highly dependent upon the culture parameters, particularly pH. At low pH (3·5), only a small quantity (33 μg l−1) of ethyl octanoate was produced in one treatment (pH 3·5, anaerobiosis and precursor supplementation), and no ethyl hexanoate was produced. In contrast, ester production was more evident at pH 4·5, whereby ethyl octanoate was produced in each treatment, and ethyl hexanoate was also produced in treatments supplemented with precursors. Furthermore, compared with respective controls, the production of ethyl octanoate at this pH was increased approximately 3–5-fold with precursor supplementation and approximately 2–3-fold by anaerobiosis. Overall, production of ethyl hexanoate and the highest concentration of ethyl octanoate occurred with the combined parameters of pH 4·5, anaerobiosis and precursor supplementation. Production of another ethyl ester, ethyl acetate, was also observed at pH 4·5, particularly under anaerobiosis (data not shown).
Initial studies investigating the ability of O. oeni to synthesize fruity ethyl esters from fatty acids (e.g. butanoic, hexanoic and octanoic acids) and ethanol were conducted using whole-cell suspensions in a phosphate buffer (pH 3·0–4·0). Using whole cells in a simple buffer system was unsuccessful in demonstrating ester synthesis in O. oeni (data not shown). However, experiments using CFE of wine LAB and activated hexanoic acid (hexanoyl-CoA) enabled ester biosynthesis via AcoAAAT activity. Preliminary observations using the AcoAAAT assay revealed that synthesis of ethyl hexanoate by O. oeni was highly dependent upon the concentration of hexanoyl-CoA substrate, was generally greatest after prolonged (approximately 18 h) rather than short (1·5–2 h) incubation periods (data not shown) and did not require addition of protease inhibitors to stabilize the assay (including phenylmethanesulfonyl fluoride, benzamidine, p-aminobenzamidine and ε-amino-n-caproic acid) (data not shown).
CFE of nine commercial malolactic bacteria strains (eight O. oeni strains and one Lb. plantarum strain) were screened for the ability to synthesize ethyl hexanoate from hexanoyl-CoA and ethanol via AcoAAAT activity (Table 1). All strains exhibited enzyme activity in the formation of ethyl hexanoate, ranging from 143·8 to 859·0 μg l−1 mg−1 protein with O. oeni strains AWRI B550 and AWRI B548, respectively, and 234·2 μg l−1 mg−1 protein with Lb. plantarum AWRI B740. Ethyl hexanoate was generally not detectable with heat-inactivated CFE (≤0·5 μg l−1 mg−1 protein) and without CFE (≤0·2 μg l−1) (data not shown). The esters ethyl butanoate and ethyl octanoate were not detected for any of the strains tested (Table 1).
CFE from two O. oeni strains (AWRI B551 and AWRI B553) and Lb. plantarum AWRI B740 were used to further investigate the synthesis of ethyl butanoate, ethyl hexanoate and ethyl octanoate from their respective free fatty acids (butanoic, hexanoic and octanoic acids) and ethanol by reverse esterase activity. The latter O. oeni strains were selected as they had previously been characterized in studies of the effects of MLF on the chemical and sensory properties of wine (Costello et al. 2012), in which strains AWRI B551 and AWRI B553 were identified as R1105 and R1118, respectively. Each strain exhibited the capacity to synthesize one or more ethyl esters by reverse esterase activity, the extent of which varied considerably depending upon the LAB strain and fatty acid substrate involved (Table 3). The esters were not detected in heat-inactivated and non-CFE controls (data not shown). All three LAB strains exhibited reverse esterase activity, synthesizing varying concentrations of ethyl hexanoate from hexanoic acid and, in the case of the two O. oeni strains, comparatively high concentrations of ethyl octanoate from octanoic acid. Both O. oeni strains exhibited greater activity than Lb. plantarum AWRI B740 in the production of ethyl hexanoate from hexanoic acid. However, none of the strains synthesized ethyl butanoate from butanoic or other fatty acids. When octanoic acid was used as a substrate, O. oeni AWRI B553 also exhibited a low activity in synthesizing the C6 ethyl ester, ethyl hexanoate (trace amount).
|Fatty acid substrate||LAB strain||Ester synthesis by reverse esterase activity (μg l−1 mg−1 protein)b|
|Ethyl butanoate||Ethyl hexanoate||Ethyl octanoate|
|Butanoic acid||Oenococcus oeni AWRI B551||ndc||nd||nd|
|O. oeni AWRI B553||nd||Trd||nd|
|Lactobacillus plantarum AWRI B740||nd||nd||nd|
|Hexanoic acid||O. oeni AWRI B551||nd||1·6 ± 0·0||nd|
|O. oeni AWRI B553||nd||7·0 ± 0·6||nd|
|Lb. plantarum AWRI B740||nd||0·5 ± 0·0||nd|
|Octanoic acid||O. oeni AWRI B551||nd||nd||28·8 ± 0·6|
|O. oeni AWRI B553||nd||Tr||66·5 ± 1·4|
|Lb. plantarum AWRI B740||nd||nd||nd|
Substrate specificity of reverse esterase activity of O. oeni AWRI B551 was examined by substituting ethanol with 1-propanol and employing stable isotope substrates ([2H6]ethanol and [2H15]octanoic acid). The results of these experiments are outlined in Table 4 and demonstrate that replacement of: (i) ethanol with 1-propanol yielded the homologue propyl ester, propyl octanoate, and (ii) substituting ethanol and octanoic acid with [2H6]ethanol and [2H15]octanoic acid yielded the polydeuterated ester, [2H5]ethyl [2H15]octanoate. The latter products were yielded exclusively from their respective substrates and, with the exception of a trace amount of propyl octanoate detected in one replicate of the assay containing [2H6]ethanol and [2H15]octanoic acid, were not produced in other assays of this experiment.
|Assaya||Alcohol and fatty acid substratesb||Ester synthesis activity (μg l−1 mg−1 protein)c|
|Ethyl octanoate||Propyl octanoate||[2H5]Ethyl [2H15]octanoate|
|1||Ethanol, Octanoic acid (control)||48·7 ± 0·6||ndd||nd|
|2||1-Propanol, Octanoic acid||nd||61·6 ± 12·8||nd|
|3||[2H6]Ethanol, [2H15]Octanoic acid||nd||Tre||61·1 ± 2·4|
In this study, we have confirmed that wine malolactic bacteria, in particular O. oeni, are capable of synthesizing ethyl esters that are known to contribute to fruity sensory characteristics. We also demonstrate for the first time in O. oeni and Lb. plantarum that this activity is due, at least in part, to AcoAAAT activity and, to a lesser extent, reverse esterase activity. These findings are in support of other reports associating MLF-driven modulation of wine ester profiles with the activity of specific malolactic bacteria strains (see Sumby et al. 2010; López et al. 2011; Knoll et al. 2012; Malherbe et al. 2012). The current study has also demonstrated that several culture parameters (i.e. pH, oxygen availability and precursor content) have major impacts on ethyl ester production by O. oeni. In particular, compared with other treatments, the highest concentrations of ethyl esters were produced with the combination of ‘higher’ pH (4·5), anaerobiosis and precursor supplementation; under these conditions, the formation of ethyl octanoate was considerably enhanced and the production of ethyl hexanoate was enabled, with the production of ethyl octanoate approximately 10-fold that of ethyl hexanoate. It was also noted that higher pH (4·5) facilitated the production of some acetate esters including ethyl acetate and, in one treatment, 3-methylbutyl acetate. Ethyl ester accumulation was generally suppressed at lower pH (3·5), however, with only a small amount of ethyl octanoate produced with one treatment combination (pH 3·5, anaerobiosis and precursor supplementation) (Fig. 1).
Knoll et al. (2011b) and Costello et al. (2012) also reported that wine pH has a major impact on the modulation of wine ester profiles (including ethyl hexanoate and ethyl octanoate content) following MLF with O. oeni. However, such effects were highly dependent upon factors including ethanol concentration, the variety or source of grapes, as well as the particular ester and bacterial strain involved. Differences in the effects of pH on ester production by malolactic bacteria observed in the latter studies compared with those of the current work could reflect compositional differences between the MRSFMEt5% medium used here and the highly complex wine matrix.
The role of oxygen availability and other compositional parameters also significantly influenced the ability of O. oeni to produce ethyl esters during growth in MRSFM-Et5% medium. Consistent with the findings of the current work that ethyl ester formation (particularly ethyl octanoate) by O. oeni was enhanced by anaerobiosis, the presence of oxygen is well known to inhibit ester formation in yeast (Dufour et al. 2003; Verstrepen et al. 2003; Saerens et al. 2006, 2008). Further, ethanol availability and substrate concentration are known to impact on ester formation in dairy LAB (Holland et al. 2005) and in yeast (Verstrepen et al. 2003; Saerens et al. 2006, 2010), respectively. Further investigation into the effects of growth and MLF-related factors on ester synthesis by O. oeni and other wine LAB is required.
This manuscript describes, for the first time, that enzyme activities involved in the synthesis of ethyl esters by O. oeni and other wine LAB include both AcoAAAT and reverse esterase. As AcoAAAT activity was generally higher than reverse esterase in the formation of ethyl hexanoate, the former could be considered a major mechanism of medium-chain ethyl ester synthesis in wine LAB. Similar conclusions have also been made regarding alcohol acyltransferase activity as the major mechanism of ester synthesis in dairy LAB (Liu et al. 2003, 2004a,b). Yamauchi et al. (1989) also reported much stronger alcohol acyltransferase activity than reverse esterase activity in ethyl hexanoate synthesis by Neurospora sp. In contrast, Abeijón-Mukdsi et al. (2009) report that ester-synthesizing activity among LAB strains isolated from goat's and ewe's milk and cheeses was generally greater by esterification; this was highly dependent upon the genus, species and strain, with alcoholysis (transferase activity) also involved in ester synthesis among enterococci strains.
The detection of AcoAAAT activity in each of nine commercial malolactic starter culture strains (eight O. oeni strains and one Lb. plantarum) suggests this may be common among wine LAB. Further, the variation in AcoAAAT activity (approximately sixfold) observed among these strains supports other reports of strain variability in ester-synthesizing ability via alcoholysis among dairy LAB (Liu et al. 2003; Abeijón-Mukdsi et al. 2009). However, in contrast to the specificity for ethyl hexanoate formation via AcoAAAT activity in wine LAB reported here, variability in the types of esters synthesized via alcoholysis has been reported for dairy LAB (Liu et al. 2003; Abeijón-Mukdsi et al. 2009). Such differences could reflect the different substrates utilized in the respective assays in this study (hexanoyl-CoA) and with dairy LAB (acylglycerols).
The reverse esterase activities of several wine LAB examined in this study were highly genus- and strain dependent, with the activities as well as the type and range of esters produced by the two O. oeni strains generally greater than Lb. plantarum AWRI B740. Ethyl ester-synthesizing capability by esterification has also been reported in dairy LAB and other bacteria (Hosono and Elliott 1974; Hosono et al. 1974; Liu et al. 1998; Abeijón-Mukdsi et al. 2009) and in Neurospora sp. (Yamauchi et al. 1989).
Examination of substrate specificity of reverse esterase activity in O. oeni AWRI B551 revealed that alcohols other than ethanol could serve as substrates, as evidenced by the formation of propyl octanoate from 1-propanol and octanoic acid. Cross-specificity for a range of alcohols has also been reported for reverse esterase of Neurospora sp. (Yamauchi et al. 1989). Moreover, the production of the fully deuterated ester, [2H5]ethyl [2H15]octanoate, from the deuterated substrates [2H6]ethanol and [2H15]octanoic acid in the current work confirmed that the carboxylic acid and alcohol substrates are directly esterified by reverse esterase activity in O. oeni AWRI B551. Abeijón-Mukdsi et al. (2009) found that the type and range of esters produced via esterase activity in LAB isolated from goat's and ewe's milk and cheese was dependent upon the carboxylic acid substrate; ethyl hexanoate was predominantly synthesized from butanoic acid, whereas esters from C4 to C10 were produced from hexanoic acid. In this study, O. oeni AWRI B553 also synthesized a trace amount of the C6 ester, ethyl hexanoate, from butanoic (C4) and octanoic (C8) acid substrates, supporting the view of Abeijón-Mukdsi et al. (2009) that synthesis mechanisms other than esterification contribute to the production of other ethyl ester end products by LAB. Further study is required to more fully ascertain both the reverse esterase and AcoAAAT capabilities among other species and strains of wine LAB, as well as the range of esters that can be synthesized by these enzyme activities.
Although the ethyl ester-synthesizing capabilities of wine LAB have been clearly demonstrated, it remains uncertain as to the nature and identity of the enzyme(s) involved. In an attempt to identify such enzyme(s), blast searches conducted in the O. oeni gene database (GenBank) revealed no significant homology with the two genes EEB1 and EHT1 encoding acyl coenzyme A: ethanol O-acyltransferase in S. cerevisiae, thus suggesting such yeast-associated acyltransferase does not occur in O. oeni. On the other hand, consistent with the alcohol acyltransferase and reverse esterase activities demonstrated in wine LAB here, two similar mechanisms of ester synthesis have also been reported in dairy LAB, in which both acyltransferase (alcoholysis) (Liu et al. 2003, 2004a,b) and esterase (Nardi et al. 2002) activities are catalysed by esterases. While the role of esterase activity in catalysing ester hydrolysis is well characterized in wine LAB (Matthews et al. 2006, 2007; Sumby et al. 2009, 2010), further investigation of the potential involvement of esterases in ester synthesis in these bacteria is required. Furthermore, as there are at least four potential esterase genes in O. oeni (Sumby et al. 2010), it is possible that ester synthesis activities observed in wine LAB in this work may be associated with more than one esterase.
In conclusion, this study has demonstrated that O. oeni exhibits growth-associated ethyl ester-synthesizing capability and that several factors including pH, level of oxygen and supplementation with precursors can influence this capability. Importantly, wine strains of O. oeni and also Lb. plantarum have also been shown to possess two different enzyme activities capable of ethyl ester synthesis, AcoAAAT and reverse esterase. Of these, AcoAAAT exhibited the highest ester-synthesizing capability and could be a major mechanism of ester biosynthesis among wine LAB. Further study is required to characterize the enzyme(s) catalysing this reaction in greater detail and factors that can be modulated to gain greater control over the biosynthesis and hydrolysis of esters by wine LAB during the winemaking process.
The authors thank Lallemand for financial support and supply of wine LAB strains used in this study. Sylvester Holt and Dr Toni Cordente are kindly thanked for advice with CFE preparation and ester synthesis assays, and Dr Anthony Borneman is also thanked for assistance in conducting blast searches. The Australian Wine Research Institute (AWRI) is supported by Australian grapegrowers and winemakers through their investment agency the Grape and Wine Research and Development Corporation, with matching funds from the Australian Government. The AWRI is a member of the Wine Innovation Cluster in Adelaide, located at the Waite Precinct.