Production of Flavours and Fragrances via Bioreduction of (4R)-(-)-Carvone and (1R)-(-)-Myrtenal by Non-Conventional Yeast Whole-Cells

As part of a program aiming at the selection of yeast strains which might be of interest as sources of natural flavours and fragrances, the bioreduction of (4R)-(−)-carvone and (1R)-(−)-myrtenal by whole-cells of non-conventional yeasts (NCYs) belonging to the genera Candida, Cryptococcus, Debaryomyces, Hanseniaspora, Kazachstania, Kluyveromyces, Lindnera, Nakaseomyces, Vanderwaltozyma and Wickerhamomyces was studied. Volatiles produced were sampled by means of headspace solid-phase microextraction (SPME) and the compounds were analysed and identified by gas chromatography–mass spectroscopy (GC-MS). Yields (expressed as % of biotransformation) varied in dependence of the strain. The reduction of both (4R)-(−)-carvone and (1R)-(−)-myrtenal were catalyzed by some ene-reductases (ERs) and/or carbonyl reductases (CRs), which determined the formation of (1R,4R)-dihydrocarvone and (1R)-myrtenol respectively, as main flavouring products. The potential of NCYs as novel whole-cell biocatalysts for selective biotransformation of electron-poor alkenes for producing flavours and fragrances of industrial interest is discussed.


Introduction
Biocatalysis represents an effective and sometimes preferable alternative to the standard synthesis of fine and/or optically active chemicals [1][2][3][4][5][6][7]. Overall, reactions catalysed by biological systems frequently exhibit high selectivity (chemo-, regio-, and stereo-selectivity) and can be considered environmentally acceptable because they typically occur under mild conditions. Both isolated enzymes and whole-cells can be used of as biocatalysts, but whole-cell biocatalysts are often preferable to the former because they are more convenient and stable sources of enzymes, with no need for costly enzyme purification and coenzyme addition. Moreover, because of the enzymes are kept within the natural environment of living cells, usually less enzyme inactivation occurs.
Flavours play a very important role in the quality perception of food and beverages, whereas fragrances represent an important part of soap and perfume industry [8][9][10]. Consumers have a strong preference for natural food additives over chemically synthesized ones. Both United States (US) [11] and European (EU) [12] laws have already labelled as "natural flavour" all those obtained from living cells, including Generally Regarded As Safe (GRAS) microorganisms [8]. Thus, products obtained by microorganisms and enzymes can be considered natural as long as natural raw materials are used. As a result, the "natural" label, allocated by EU and US food legislation, represents a strong marketing advantage [9,10].
Monoterpenes are one of the largest classes of flavouring compounds (over 400 different naturally occurring structures), and represent a valuable resource for the flavour and fragrance industry. The consumer requests for natural flavours and fragrances have encouraged a growing part of scientific community to study and develop novel biocatalysts for producing this class of molecules. Thus, the microbial and enzymatic biotransformation of some monoterpenoids, in particular a few ketones and aldehydes (e.g., carvone, menthol, citronellol, myrtenal and geraniol) into highly valuable flavouring derivatives is becoming of increasing interest because of their economic potential for the perfume, soap, food, and beverage industry [13][14][15][16][17][18][19][20][21][22]. Carvone is produced by over 70 different plants. It is found basically in two distinct stereoisomeric forms, which differ between them for their flavouring attributes: (i) (4R)-(−)-carvone, which is the principal constituent in spearmint (Mentha spicata) oil, and (ii) S-(+)-enantiomer, which is present in oils extracted from caraway (Carum carvi) seeds and from dill (Anethum graveolens) seeds. Biocatalytic transformation of carvone has recently been the focus of several studies, reporting that some enzymes may catalyze the reduction of C=C and C=O double bonds competitively, affording a mixture of saturated ketones, saturated alcohol and, more rarely, the allylic alcohol [14,20,[23][24][25][26][27][28][29][30][31][32]. From an industrial point of view, carvone and related compounds are important flavours and fragrances of industrial interest [33]. In particular, due their high volatility, dihydrocarvones are potent inhibitors of bacteria and filamentous fungi, as well as prospective insect repellents [34], and have been used as chiral starting compounds in the synthesis of natural products (e.g., striatenic acid, pechueloic acid) [35][36][37], antimalarial drugs [38] and valuable chiral synthons [39,40]. Dihydrocarveols are valuable fragrance ingredients currently used in decorative cosmetics, fine fragrances, shampoos, soaps and other toiletries as well as in household products such as cleaners and detergents [41].
Although the use of yeast whole-cells as biocatalysts is a well-established practice, and a few yeast-catalysed processes have been even successfully scaled up from laboratory to the industrial level [5,[42][43][44][45], if compared with other microbial domains (e.g., bacteria, filamentous fungi), the number of studies reporting the use of yeast whole-cells to catalyse the biotransformation of monoterpenes represents only a little percentage of the literature published so far [46]. To most people, yeasts are exemplified by the common baker's yeast (taxonomically defined as belonging to the ascomycetous species Saccharomyces cerevisiae). This is in spite of the fact that this species represents only a small fragment of the huge taxonomic and metabolic diversity occurring in the yeast world. In fact, in recent decades, biotech-oriented research had paid its attention to the so-called non-conventional yeasts (NCYs), which demonstrated sometimes a superior biocatalytic aptitude than S. cerevisiae [43,44,47].

Results and Discussion
The biotransformations of the α,β-unsaturated ketone (4R)-(−)-carvone (1) catalyzed by whole-cells of NCYs in aqueous media were investigated. The possible reaction pathway is illustrated in Scheme 1. According to the proposed scheme, the biotransformation resulted in the reduction of the α,β-unsaturated C=C bond of the cyclic ketone, catalyzed by ene-reductases (ERs) associated to the yeast cells, to give two dihydrocarvones 2a,b. The ER-catalysed reduction was thus followed by the subsequent reduction of the carbonyl group of both dihydrocarvone isomers, catalyzed by carbonyl reductases (CRs), which determined the formation of a mixture of four dihydrocarveols 3a-d (Scheme 1).
Although the use of purified ERs needs an accompanying regeneration system for the nicotinamide cofactor [NAD(P)H] to close the catalytic cycle and improve the bioreduction efficiency, we assumed that lyophilized cells contain the needed recycling system except for the co-substrate. Accordingly, glucose was added because we found that this compound was the best co-substrate for co-factor recycling system [23]. In fact, from a quantitative point of view, the presence of glucose in the reaction mixture [acting as auxiliary substrate for NAD(P)H regeneration] resulted critical for enhance the % of conversion of precursor, in close agreement with current literature [23,48]. With no glucose addition, whole-cells of NCYs showed only a little ability to reduce (4R)-(−)-carvone. Only three strains gave acceptable results: Hanseniaspora guilliermondii DBVPG 6790 (conversion about 14%), Lindnera amylophila DBVPG 6346 (about 10%) and Vanderwaltozyma polyspora DBVPG 6243 (about 8%). The prevalent catalytic activity of whole-cells of NCYs was the ER-catalysed reduction of (4R)-(−)-carvone into a mixture of (1R,4R)-and (1S,4R)-dihydrocarvone, with a clear-cut preference towards the production of (1R,4R)-diastereomer. Only traces of dihydrocarveols 3a-d, derived from the subsequent CR-catalysed reduction of the carbonyl group of (4R)-(−)-carvone were found. As expected, the addition of glucose to the reaction mixture, visibly increased the aptitude of some strains to reduce (4R)-(−)-carvone 1 ( Table 1). Table 1. Bioconversion of (4R)-(−)-carvone 1 into derivative products after 120 h by whole-cells of NCYs in the presence of glucose. In particular, whole-cells of H. guilliermondii DBVPG 6790 exhibited a superior level of conversion (about 63%). Interestingly, the ERs associated with whole-cells of this strain exhibited a high chemoselectivity, thus resulting in the preferential reduction of the α,β-unsaturated C=C bond of (4R)-(−)-carvone to give almost exclusively (1R,4R)-dihydrocarvone 2b (about 62%). No formation of dihydrocarveols was observed ( Table 2). The high diasteroisomeric excess observed for this reaction (d.e. = 95%) suggested that the ERs of H. guilliermondii may attack hydrogen at the conjugated C=C double bond in a stereoselective way from the si-face at C-1 and re-face at C-6 by anti-addition. Similarly to the biotransformation yields observed when H. guilliermondii was used, the addition of glucose to the reaction mixture increased the aptitude of some strains (otherwise practically incapable of catalysing the reaction) to reduce (4R)-(−)-carvone to levels of bioconversion > 10%: Candida shehatae DBVPG 6850 (17%), Candida oregonensis DBVPG 6149 (15%), Candida maltosa DBVPG 6021 and V. polyspora DBVPG 6243 (13%), Kazachstania naganishii DBVPG 7133 (12%) and Kazachstania spencerorum DBVPG 6746 (11%) ( Table 1).

Species and strain
The increase of the bioconversion yield caused by glucose addition described above is not surprising. ERs are NAD(P)H-dependent flavoproteins and need an accompanying co-factor acting as regeneration system for the nicotinamide cofactor [NAD(P)H]. ERs often show a marked specificities for NADH or NADPH as co-factor, which allows to choose the best co-substrate for cofactor-recycling system on a case-to-case basis. A few authors [48] reported that yeast whole-cells contain only catalytic amounts of NAD(P)H, so its regeneration must take place by means of metabolism of an electron donor. In this framework, glucose could act as co-substrate for cofactor-recycling systems, thus exhibiting a significant efficacy in increasing the ER-catalysed reduction of (4S)-(−)-carvone. This result is consistent with current literature [44] that postulate that the presence of auxiliary substrates (e.g., glucose, formate, isopropanol, etc.) for cofactor-recycling system could reduce the problem of co-factor consumption and, therefore, greatly enhance the bioconversion yield.
Surprisingly, C. shehatae reduced the α,β-unsaturated C=C bond of (4R)-(−)-carvone in an chemoselective way divergent from that discussed above for H. guilliermondii, to give a slightly excess of (1S,4R)-dihydrocarvone (d.e. = 10%) ( Table 2). This evidence could apparently support the hypothesis that the ERs associated with whole-cells of C. shehatae may attack hydrogen at the conjugated C=C double bond from both the si-and re-face at C-1. This could be due to the possibility of (4R)-(−)-carvone to flip into the enzymatic pocket of ER, or, alternatively, to the presence of a second ER (associated to whole-cells too) showing an opposite stereo-preference.
The biotransformations of the α,β-unsaturated aldehyde (1R)-(−)-myrtenal (4) catalyzed by whole-cells of NCYs were investigated. The possible reaction pathway is reported in Scheme 2. Based on the proposed scheme, ERs associated to whole-cells of NCYs catalysed the reduction of the α,β-unsaturated C=C bond of (1R)-(−)-myrtenal to give dihydromyrtenals 6. The carbonyl group of the aldehyde was further reduced by CRs to give dihydromyrtenols 7. A concurrent reaction, giving the production of (1R)-(-)-myrtenol 5 was also observed in some strains (Scheme 2).

Scheme 2. Bioconversion pathway of (1R)-(-)-myrtenal by whole-cells of NCYs
Overall, NCYs used in this study showed good (in some cases excellent) aptitudes to biotransform (1R)-(−)-myrtenal (4) into derivative compounds, even in the absence of co-substrates for NAD(P)H regeneration (i.e., glucose): about one third of strains gave percentage of conversion ≥ 95%. Among them, Candida freyschussii DBVPG 6208 and Kazachstania spencerorum DBVPG 6746 converted 100% of the precursor (Table 3). Interestingly, in almost all cases, biocatalytic ability was prevalently related to the activity of CRs associated to whole-cells, affording myrtenol 5 as the main product of the bioconversions. On the contrary, the results obtained apparently suggest that (1R)-(−)-myrtenal (4) is not a good substrate for the ER-activity. In fact, only a few strains showed prevalent ER-catalysed asymmetric C=C bioreduction of the α,β-unsaturated aldehyde: they produced dihydromyrtenals 6, which was then further reduced by CRs to give dihydromyrtenols 7, that resulted the main products when Cryptococcus albidus DBVPG 6237 and Lindnera amylophila DBVPG 6346 were used as biocatalysts (Table 2). Surprisingly, in some cases, (in particular when the biotransformation was promoted by Cryptococcus gastricus DBVPG 6057, L. amylophila DBVPG 6346, Kluyveromyces marxianus DBVPG 6854 and Nakaseomyces bacillisporus DBVPG 6962), the sum of (1R)-(−)-myrtenol 5 + dihydromyrtenals 6 + dihydromyrtenols 7 was apparently lower than % of conversion (Table 2). This could be due to the fact that GC-MS analysis did not reveal additional volatile by-products. This suggests that the actually lower biotransformation yield might be the result of the possible presence of side reactions. Among them, a special mention could be dedicated to myrtenoic acid, which was detected in little amounts (< 10%) in some strains and was revealed after extraction of samples with ethyl acetate. The production of this compound (not indicated in the bioconversion pathway reported in Scheme 2 may be due to a dismutation side reaction of aldehyde into the corresponding alcohol and carboxylic acid, promoted by an alcohol dehydrogenase (ADH), which was probably associated to whole-cells [49].
Among naturally occurring oxygen-containing monoterpene derivatives, (1R)-(−)-myrtenal is one of the most widespread flavouring compounds distributed in the Asteraceae (Compositae) family. It is a constituent of the cumin seed, juniper berry, pepper, peppermint, scotch spearmint and it is ubiquitous in the essential oils of flowers, stems, and leaves which possess a sweet and spicy cinnamon-like odour highly demanded as a key constituent of some fragrances [50]. Although only one study describing the biotransformation of myrtenal by algae has been so far published [51], the industrial importance of its structurally-related monoterpene alcohol derivative myrtenol is becoming greater and greater: this compound is a fragrance ingredient used in decorative cosmetics, fine fragrances, shampoos, toilet soaps and other toiletries as well as in non-cosmetic products such as household cleaners and detergents [52]. The following microbiological culture media were used: i) growth phase: YEPG: yeast extract 10 g L −1 , peptone 10 g L −1 , glucose 20 g L −1 , agar 15 g L −1 ); ii) biotransformation phase: Yeast medium (YM): yeast extract 3 g L −1 , malt extract 3 g L −1 , peptone 5 g L −1 , glucose 10 g L −1 , pH 6.5. Ingredients of culture media were from Difco (BD, Franklin Lakes, NJ, USA).

Yeast Strains
Twenty-five NCY strains belonging to the genera Candida, Cryptococcus, Debaryomyces, Hanseniaspora, Kazachstania, Kluyveromyces, Lindnera, Nakaseomyces, Vanderwaltozyma, and Wickerhamomyces, were used (Table 3). They were preliminarily selected from amongst over 200 strains of environmental origin (representative of about 80 ascomycetous and basidiomycetous yeast species) for their ability to catalyse the biotransformation of (4S)-(+)-carvone (which was preliminarily used as model target substrate) into reduced products (dihydrocarvones and dihydrocarveols) [23]. All strains are conserved at the Industrial Yeasts Collection DBVPG of the University of Perugia, Italy [53] and are publicly available upon request. NCY strains were routinely maintained in frozen form (−80 °C). Working cultures were grown on YEPG agar slants at 25 °C.

Preparation of Lyophilized Whole-Cells Biocatalysts
Lyophilized NCY cells were obtained as previously described [23], with some modifications. Aliquots (200 μL) of 24 h cell suspensions, calibrated to A 580 = 0.5 (average cell concentration = 10 6 cells mL −1 ), were used to inoculate 5 mL of YM. Once the NCY biomass was reached following 48 h of incubation, cells and supernatants were separately harvested. NCY cells were washed three times by using 50 mM phosphate buffer (pH 6.5), centrifuged each time for 15 min at 4,000 rpm, snap frozen (−80 °C) and lyophilized for 48 h in a Lyophilizer Modulyo (Edwards, Sanborn, NY, USA).

Bioconversion Reactions
Thirty mg of lyophilized NCY cells were resuspended in 25 mL sterile vials containing 5 mL of 50 mM phosphate buffer (pH 6.5) + 50 mM glucose [acting as cofactor-recycling system for NAD(P)H]. Each substrate (final concentration 10 mM) was added and the vials were incubated on an orbital shaker (120 rpm) at 25 °C for 120 h. After incubation vials were sealed and frozen (−30 °C) until SPME + GC-MS analysis. All the reactions were carried out in triplicate. In order to determine whether substrates were spontaneously reduced in the absence of NCY cells, blank (cell-free) vials containing 50 mM phosphate buffer + 50 mM glucose and each substrate were analyzed at 120 h. As with previous samples, following incubation vials were sealed and frozen (−30 °C) until SPME + GC-MS analysis.

SPME and GC-MS Analyses
(4R)-(−)-Carvone and (1R)-(−)-myrtenal derivatives occurring in vial headspace after bioreduction reactions were detected by GC-MS using the solid-phase microextraction (SPME) sampling technique. Sealed vials containing the yeast suspensions were thawed by immersion in a silicon oil bath at 35 °C for 15 minutes. Headspace was analyzed using a 1-cm needle containing a fiber coated with 75 μM Carboxen/polydimethylsiloxane bonded to a flexible fused silica core (Supelco, Sigma-Aldrich Co, St. Louis, MO, USA). The needle was inserted into the vial through the septum and the fiber was exposed to headspace volatiles for 5 min at 30 °C. After direct desorption into the injector port at 280 °C for 5 min, the products were analyzed using a Hewlett Packard (Palo Alto, CA, USA) G1800C Series II gas chromatograph-mass spectrometer equipped with a HP-5 column (25 m × 0.2 mm, 0.5 μM film thickness) coated with (5%)-diphenyl-(95%)-dimethylpolysiloxane copolymer.
Compounds derived from biotransformation of (4R)-(−)-carvone and (1R)-(−)-myrtenal were identified by comparing their respective mass fragmentation patterns (EI, 70 eV) with the database library NIST05 (Varian MS Library Software, Palo Alto, CA, USA). A concentration of monoterpenes was measured in the vial headspace by an internal standard method in which thawing vial contents were spiked with 50 μL of a freshly prepared chlorobenzene solution (0.05 mg/mL in deionized water).
All the results were expressed as biotransformation yield, i.e., % of substrate converted to a given derivative. The total ene-reductase (ER) or carbonyl reductase (CR) activities of NCY whole-cells were calculated as the sum of products obtained by reduction of a given substrate (catalyzed by ERs or CRs, respectively). All the results represented the average of three independent experiments, and the statistical significance of these average data was assessed via ANOVA.

Conclusions
Although prokaryotic and eukaryotic microorganisms (including yeasts) are recognized to produce flavouring compounds potentially attractive for industry, the number of processes that can be considered truly competitive with chemical synthesis is at present limited, primarily because of their low yields [43]. Accordingly, as underlined by previous authors [46,53], the potential offered by yeast diversity, still far from being fully explored, could represent a noteworthy source of novel biocatalysts for the synthesis of flavours and fragrances. Moreover, the use of whole-cells is nowadays adequately accepted, not only as laboratory curiosity, but also in the industrial scale, because they are considered much more economical and easy-to-handle than purified enzymes [23,54,55].