INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Saccharomyces cerevisiae has been used for centuries in the production of bread and alcoholic beverages. Along with ethanol and carbon dioxide, fermenting cultures of this yeast produce a variety of low-molecular-weight flavor compounds (including alcohols, diacetyl, esters, organic acids, organic sulfides, and carbonyl compounds). The compounds 3-methyl-1-butanol, 2-methyl-1-butanol, and 2-methyl-1-propanol, commonly known as fusel alcohols, and their esters make an important contribution to the flavor of alcoholic beverages and bread (1, 14).

A metabolic pathway for production of fusel alcohols by yeast was first proposed by Ehrlich (6). The Ehrlich pathway starts with the enzyme-catalyzed decarboxylation of branched-chain 2-oxo acids to the corresponding aldehydes. Subsequently, the aldehyde is reduced to the corresponding fusel alcohol by an alcohol dehydrogenase (11, 16, 24). The branched-chain 2-oxo acid substrates for the Ehrlich pathway can be produced by the deamination of L-leucine, L-isoleucine, or L-valine. Growth of S. cerevisiae with any of these three amino acids as the nitrogen source results in the accumulation of the corresponding fusel alcohol (2, 3, 21). Alternatively, branched-chain 2-oxo acids may be synthesized de novo from carbohydrates as intermediates of branched-chain amino acid synthesis (13).

The conversion of branched-chain oxo acids into their respective aldehydes and alcohols via the Ehrlich pathway resembles the fermentative metabolism of pyruvate, which yields ethanol and carbon dioxide. In both cases, the decarboxylation of a 2-oxo acid is followed by the reduction of the resulting aldehyde. Partially purified preparations of yeast pyruvate decarboxylase have been shown to catalyze the decarboxylation of various 2-oxo acids, including the putative intermediates of the Ehrlich pathway (8, 12, 16, 21). However, it has not been conclusively proven that pyruvate decarboxylase is essential for or even involved in fusel alcohol production by S. cerevisiae.

Dickinson and Dawes (4) have reported that, at least under some conditions, oxidative decarboxylation by a mitochondrial branched-chain oxo acid dehydrogenase complex (17) is involved in the catabolism of branched-chain 2-oxo acids. Mutants that did not express the lipoamide dehydrogenase subunit of this enzyme complex accumulated branched-chain oxo acids in batch cultures grown on media containing leucine, isoleucine, or valine (4), thus casting some doubt on the exclusive role of pyruvate decarboxylase in the decarboxylation of branched-chain oxo acids.

The aim of this study was to reinvestigate the role of pyruvate decarboxylase in the production of fusel alcohols by S. cerevisiae. The S. cerevisiae genome harbors three structural genes (PDC1, PDC5, and PDC6) that can each encode an active pyruvate decarboxylase (9). In wild-type yeast strains, PDC6 expression is either very low or absent (7, 9). However, revertants of pdc1-pdc5 double mutants, in which a recombination event has caused a fusion of the PDC1 promoter and the PDC6 open reading frame, express a functional enzyme (10). Therefore, studies on the physiological effects of pyruvate decarboxylase deficiency are most easily interpreted when they are performed with strains in which all three PDC genes are disrupted.

In the present study, the decarboxylation of branched-chain 2-oxo acids was studied in cell extracts of wild-type S. cerevisiae and in extracts of an isogenic pyruvate decarboxylase-negative mutant. Furthermore, conversion of branched-chain amino acids to the corresponding fusel alcohols by intact cells was analyzed in ethanol-grown cultures of a wild-type S. cerevisiae strain and in those of the Pdc mutant.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and growth conditions. The yeast strains used in this study are listed in Table 1. S. cerevisiae T2-3D and the isogenic, prototrophic pyruvate decarboxylase-negative strain GG570 were grown in aerobic carbon-limited chemostat cultures (dilution rate, [D] = 0.10 h) on a mineral medium containing 7.125 g of glucose/liter. To meet the requirement of the Pdc mutant for cytosolic acetyl-coenzyme A, 0.375 g of acetate/liter was also added to the reservoir medium (for further details on growth conditions, see reference 7). Strains GG567 and GG569 were grown overnight at 30°C in shake flask cultures on complex medium (Difco yeast extract [10 g/liter], Difco peptone [20 g/liter], glucose [20 g/liter]) prior to preparation of cell extracts.

Analysis of fusel alcohol production by growing cells. Precultures were grown at 30°C in shake flask cultures containing mineral medium with vitamins (20) supplemented with 1% (vol/vol) ethanol, with an initial pH of 6.0. The ammonium sulfate concentration in the mineral medium was decreased to 0.5 g/liter to obtain a nitrogen-depleted inoculum culture. After 24 h of incubation, the cells were centrifuged at 20,000 × g for 5 min and aseptically transferred to a 1,000-ml shake flask containing 400 ml of mineral medium supplemented with 1% (vol/vol) ethanol, with an initial pH of 6.0. The mineral medium lacked ammonium sulfate and contained instead a mixture of leucine, isoleucine, and valine (15 mM each) as a nitrogen source. The flasks were shaken (200 rpm) at 30°C. Samples were taken at appropriate time intervals and analyzed for optical density at 660 nm (23). Culture supernatants, obtained by centrifugation at 20,000 × g for 5 min, were analyzed for amino acids and fusel alcohols.

Analytical procedures. Biomass concentrations were determined as described previously (7). Concentrations of 3-methyl-butanol, 2-methyl-butanol, and 2-methyl-propanol were analyzed by gas chromatography on a Perkin-Elmer 5800 gas chromatograph fitted with a Chrompack CP-SIL5CB column (length, 50 m; internal diameter, 0.32 mm; film thickness, 1.2 µm). The identity of the peaks was confirmed by high-pressure liquid chromatography (HPLC) analysis of the same samples on a Rezex ROA organic acid column at 60°C. The HPLC column was eluted with 0.5 g of H₂SO₄/liter; detection was by means of an ERMA ERC-7515A refractive index detector coupled with a Hewlett-Packard 3390A integrator. Concentrations of leucine, isoleucine, and valine were analyzed after derivatization with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) with the Waters AccQ.Fluor kit. Derivatized samples were then analyzed by HPLC on a Waters Nova-Pak C18 column. The two mobile phases were (i) 60 mM ammonium acetate (pH 5.00) and (ii) 50% of mobile phase A plus 50% (vol/vol) acetonitrile. Gradient conditions were as follows: 27 min linearly from 97% of mobile phase A to 89% of mobile phase A and thereafter in 22 min linearly to 54% mobile phase A. Column temperature was 25°C; flow rate was 1 ml/min.

Preparation of cell extracts and enzyme assays. Cell extracts were prepared as described previously (7). Pyruvate decarboxylase was assayed spectrophotometrically at 30°C. The assay mixture consisted of 40 mM imidazole-HCl buffer (pH 6.5), 0.2 mM thiamine pyrophosphate, 5 mM MgCl₂, 150 µM NADH, yeast alcohol dehydrogenase (88 U/ml; Boehringer Mannheim) 0.05% Triton X-100, and cell extract. The reaction was started by the addition of 10 mM pyruvate. Essentially the same coupled assay was used to measure decarboxylation of -ketoisocaproate, -keto--methyl valerate, or -keto isovalerate by cell extracts. However, to measure decarboxylation of these substrates, horse liver alcohol dehydrogenase (2 U/ml in assay mixture; Sigma) was used instead of yeast alcohol dehydrogenase (see Results). Alcohol dehydrogenase (EC 1.1.1.1) was assayed in a reaction mixture (1 ml) containing glycine-KOH buffer (pH 9.0), 50 mM; NAD (lithium salt), 1 mM; and cell extract. The reaction was started by the addition of 10 mM of either ethanol, 2-methylbutanol, 3-methylbutanol, or 2-methylpropanol.

Protein determination. Protein concentrations in cell-free extracts were determined by the Lowry method (12a). Bovine serum albumin (fatty acid free; Sigma Chemical Co.) was used as a standard.

Biochemicals. Partially purified preparations of pyruvate decarboxylase, yeast alcohol dehydrogenase, and horse liver alcohol dehydrogenase were obtained from Sigma.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Decarboxylation of 2-oxo acids by cell extracts of wild-type S. cerevisiae. Cell extracts of the wild-type strain S. cerevisiae T2-3D, pregrown in aerobic carbon-limited chemostat cultures, exhibited high activities of pyruvate decarboxylase (Table 2) in a coupled spectrophotometric assay with yeast NAD-dependent alcohol dehydrogenase. Although the extracts exhibited some activity when pyruvate was replaced by the branched-chain 2-oxo acid -keto--methyl valerate, -keto isovalerate, or -ketoisocaproate, only very low activities were observed (ca. 1% of those found with pyruvate). Furthermore, these low rates were not constant over time and not always linearly proportional to the amount of cell extract added to the assays. Control experiments showed that, even at an alcohol concentration of 10 mM, the activity of the coupling enzyme yeast alcohol dehydrogenase with fusel alcohols was much lower than when ethanol was used as the substrate (Table 3).

Pyruvate decarboxylase is involved in decarboxylation of branched-chain 2-oxo acids by cell extracts. To determine the role of pyruvate decarboxylase in the decarboxylation of branched-chain 2-oxo acids by cell extracts, assays were performed with extracts of the isogenic Pdc strain S. cerevisiae GG570 (7). Extracts of this strain had no activity when pyruvate or the branched-chain 2-oxo acids were used as a substrate (Table 2). The addition of commercial, partially purified pyruvate decarboxylase to reaction mixtures containing Pdc cell extracts restored decarboxylation activities (data not shown). This result demonstrated that pyruvate decarboxylase was the sole component lacking for the conversion of -oxo acids to aldehydes by cell extracts.

Production of fusel alcohols by cell suspensions does not require an active pyruvate decarboxylase. We examined the role of pyruvate decarboxylase during the production of fusel alcohols from three branched-chain amino acids (leucine, isoleucine, and valine) by growing wild-type S. cerevisiae T2-3D and its isogenic Pdc strain GG570 with a mixture of these three amino acids as the nitrogen source. Ethanol was used as the carbon source to circumvent the inability of the Pdc strain to grow on glucose in batch cultures (7).

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Production of fusel alcohols by whole cells of S. cerevisiae T2-3D (wild type) and the isogenic pyruvate decarboxylase-negative S. cerevisiae GG570. Cells were grown on ethanol with a mixture of leucine, isoleucine, and valine (15 mM each) as the nitrogen source. In an independent duplicate experiment, optical density at 660 nm (OD660) and concentrations of fusel alcohols differed from those shown in the figure by less than 20%. Closed symbols represent wild-type cultures, and open symbols represent the Pdc strain. (A) Optical density at 660 nm. (B) Concentrations of amino acids ( and , leucine;  and , isoleucine;  and , valine). (C) Concentrations of fusel alcohols in wild-type culture (, 3-methylbutanol; , 2-methylbutanol; , 2-methylpropanol). (D) Concentrations of fusel alcohols in Pdc culture (, 3-methylbutanol; , 2-methylbutanol; , 2-methylpropanol).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Our results demonstrate that, although pyruvate decarboxylase is able to catalyze the decarboxylation of branched-chain 2-oxo acids, this enzyme is not essential for the production of fusel alcohols by S. cerevisiae. Consequently, an alternative mechanism for fusel alcohol production must exist in this yeast. Our data do not exclude involvement of the Ehrlich pathway (including pyruvate decarboxylase) in fusel alcohol production by wild-type S. cerevisiae. Indeed, the rate of fusel alcohol production in the Pdc strain was lower than that in the wild type, particularly in the case of 2-methylpropanol (Fig. 1C). The very poor activity of yeast NAD-dependent alcohol dehydrogenase with the putative intermediates of the Ehrlich pathway (Table 3) identifies the alcohol dehydrogenase reaction as an interesting target for attempts to improve productivity of fusel alcohols via this pathway. Expression of the horse liver enzyme in S. cerevisiae is an interesting option to test the viability of such an approach.

Discussion of the nature of the pathway for fusel alcohol production by the Pdc strain is speculative. One possibility is that a decarboxylase other than pyruvate decarboxylase is present in S. cerevisiae and is not detected by the coupled enzyme assay used in this study. Dickinson and Dawes (4) have reported the involvement of a mitochondrial branched-chain 2-oxo acid dehydrogenase complex in the catabolism of branched-chain 2-oxo acids. To act as an intermediate in the Ehrlich pathway, the coenzyme A derivative formed by this enzyme complex (25) must be reduced to the corresponding aldehyde. So far, no enzyme activities have been identified in S. cerevisiae that catalyze this conversion. The only subunit of the branched-chain 2-oxo-acid dehydrogenase complex whose structural gene has been cloned is the lipoamide-dehydrogenase subunit (encoded by the LPD1 gene [15]). This subunit is also an essential part of the pyruvate-dehydrogenase and -ketoglutarate-dehydrogenase complexes (15, 25) and of glycine decarboxylase (18), thereby complicating physiological studies of gene disruption mutants. For example, it is impossible to study the effect of an lpd1 null mutation in a Pdc strain, since the resulting double mutant is not viable (respiratory growth requires LPD1, and fermentative growth requires a functional PDC gene). Identification of the structural gene encoding the E1 subunit of the branched-chain 2-oxo acid dehydrogenase complex, which is probably specific for this complex (4), seems to be a prerequisite for elucidating its possible role in fusel alcohol production.

Biochemical research to elucidate the pyruvate decarboxylase-independent formation of fusel alcohol production is not only of fundamental scientific interest; identification of the enzymes and genes involved may ultimately enable the optimization of fusel alcohol formation independent of the fermentative production of ethanol. Such a development might be applicable in processes such as the production of low-alcohol beers and high-gravity brewing.