`
`Modulation of Glycerol and Ethanol Yields During
`Alcoholic Fermentation in Saccharomyces cerevjsjae
`Strains Overexpressed or Disrupted for CPD] Encoding
`Glycerol 3—Phosphate Dehydrogenase
`
`SUMIO MICHNICKl, JEAN—LOUIS ROUSTAN‘, FABIENNE REMIZEI, PIERRE BARREl
`AND SYLVIE DEQUIN‘*
`
`1Laboratoire de Microbiologie er. Teclinologi'e des Ferment‘al‘z’ons, [NRA—IPV, 2 Place Viala, F~34060 Montpellier
`cedex 01, France
`
`Received 20 August 1996; accepted 16 December 1996
`
`The possibility of the diversion of carbon flux from ethanol towards glycerol in Saccharomyces cerevr’sr’ae during
`alcoholic fermentation was investigated. Variations in the glycerol 3—phosphate dehydrogenase (GPDH) level and
`similar trends for alcohol dehydrogenase (ADH), pyruvate decarboxylase and glycerol~3—phosphatase were found
`when low and high glycerol—forming wine yeast strains were compared. GPDH is thus a limiting enzyme for glycerol
`production. Wine yeast strains with modulated CPD] (encoding one of the two GPDH isoenzymes) expression were
`constructed and characterized during fermentation on glucose—rich medium. Engineered strains fermented glucose
`with a strongly modified [glycerol] : [ethanol] ratio. gdeA mutants exhibited a 50% decrease in glycerol production
`and increased ethanol yield. Overexpression of CPD] on synthetic must (200 g/l glucose) resulted in a substantial
`increase in glycerol production ( X 4) at the expense of ethanol. Acetaldehyde accumulated through the competitive
`regeneration of NADH via GPDH. Accumulation of by—products such as pyruvate, acetate, acetoin, 2,3 butane—diol
`and succinate was observed. with a marked increase in acetoin production. © 1997 by John Wiley & Sons. Ltd.
`
`Yeast 13: 783—793, 1997.
`No. of Figures: 5. No. of Tables: 2. Not of References: 38.
`
`KEY WORDS—Saccl1arorriyces cerevisr’ae', alcoholic fermentation; glycerol; glycerol 3—phosphate dehydrogenase;
`redox balance; metabolic engineering
`
`INTRODUCTION
`
`Glycerol, quantitatively the most important by—
`product of alcoholic fermentation,
`is synthesized
`by reduction of dihydroxyacetone phosphate to
`glycerol 3—phosphate (G3P) by NAB—dependent
`glycerol 3—phosphate dehydrogenase (GPDI—l) fol-
`lowed by dephosphorylation of GSIJ to glycerol
`by means of a specific glycerol 3—phosphatase
`(G3Pase).
`During alcoholic fermentation, the major role of
`glycerol formation is to maintain the redox bal-
`ance. Although ethanol production, which ensures
`reoxidation of the NADH formed during the oxi—
`
`*Correspondence to; Sylvie Dequin.
`Contract grant sponsor: French MRT.
`
`CCC 0749—503X/97/090783—11 $17.50
`(C) 1997 by John Wiley & Sons Ltd
`
`dation of glyceraldehyde 3-phosphate, is a redox—
`equilibrated process, excess NADH is produced
`during biomass formation. Glycerol is mainly pro-
`duced to counterbalance this surplus of NADH
`and may be considered to form a redox valve
`(Nordstom, 1968; Lagunas and Gancedo, 1973;
`Oura, 1977; Van Dijken and Schefiers, 1986). In
`addition, glycerol production may play a role in
`balancing the ratio of free to bound phosphate
`in the cytosol, as has been suggested recently
`(Luyten er 2]., 1995).
`Another aspect is the essential role of glycerol as
`a compatible solute during hyperosmotic stress.
`The yeast Saccharomyces cerevisiae responds to
`increased external osmolarity by enhanced produc—
`tion and intracellular accumulation of glycerol to
`
`BUTAMAX 1024
`
`BUTAMAX 1024
`
`
`
`784
`
`S. MlCHNlCK ET AL.
`
`counterbalance the osmotic pressure (Blomberg
`and Adler, 1989, 1992; Mager and Varela, 1993).
`The NAB—dependent GPDH is strongly induced in
`these conditions (Andre et a]., 1991; Albertyn
`er al., 1994a). GPDl, one of the two isogenic genes
`for GPDH,
`is directly involved in osmotic-stress
`response and its expression is partly controlled via
`the high-osmolarity glycerol response MAP kinase
`pathway (Larsson er a1., 1993; Albertyn er a].,
`1994b). On the other hand, GPDZ, recently char-
`acterized,
`is not subjected to osmotic regulation
`but assumed to be involved in redox balancing
`during anaerobiosis (Ericksson er a1, 1995; Ansell
`el‘a1.. 1995).
`Since formation of both glycerol and ethanol
`plays a role in balancing the redox state of the cell,
`it might be assumed that limitation or amplifi—
`cation of the steps specifically involved in the
`production of
`these compounds
`(catalysed by
`pyruvate decarboxylase, PDC; alcohol dehydroge~
`nase, ADH; GPDI—l and GSPase) would result in a
`change in the [glycerol] : [ethanol] ratio. In agree—
`ment with this hypothesis, the amount of glycerol
`increases, for example, if acetaldehyde is trapped
`with sulfite. Furthermore, strains with low ADH
`specific activity or mutants deficient
`in ADHI
`produce more
`glycerol
`during
`fermentation
`(Ciriacy, 1975; Wills and Phelps, 1975; Johansson
`and Sjostrom, 1984). On the other hand, a rela—
`tionship between GPDI—I activity and the ability of
`yeast strains to produce glycerol had been put
`forward (Radler and Schiitz, 1982). Finally, a
`strain of Saccharomyces diastaticus disrupted for
`GPDl (previously named DARl) displays a 75%
`decrease in glycerol production on 8% glucose
`medium (Wang et ai., 1994).
`ratio
`Modulation of the [glycerol] : [ethanol]
`during glucose fermentation would be interesting
`for
`industrial purposes.
`Improvement of wine
`yeast strains for glycerol production would be
`advantageous in the case of wines that are lacking
`in body. Moreover, glycerol contributes to the
`taste of wine by providing sweetness (Hinreiner
`er a/., 1955; Noble and Bursick, 1984). In addition,
`utilization of a yeast that overproduces glycerol at
`the expense of ethanol for the elaboration of
`beverages with a low ethanol content, for instance
`wines and half—fermented beverages such as ‘pétil—
`[ant de raisin’ would be an alternative to physical
`techniques for alcohol removal that do not always
`conserve the organoleptic characteristics of the
`product. On the other hand. a sugar—fermenting
`yeast that produces a small quantity of glycerol
`
`but a higher alcohol yield would be of great value
`for the distilling industry.
`In this study, we have investigated the possibility
`of constructing S. cerevisiae strains exhibiting
`modified glycerol and ethanol yields during alco—
`holic fermentation. Evidence that GPDH is a
`
`rate-limiting enzyme in glycerol production under
`enological
`fermentation conditions
`is
`reported
`and sustained by demonstration that disruption
`and overexpression of CPD], one of the two iso-
`genic genes for GPDH, strongly alter the [gly—
`cerol] : [ethanol] ratio. The strains overexpressed
`for GPDI exhibited a four-fold increase in glycerol
`production under conditions similar to those of
`wine—making. The study of the metabolic modifi-
`cations caused by the strains overproducing gly-
`cerol
`to cope with carbon diversion and redox
`imbalance is presented.
`
`MATERIALS AND METHODS
`
`Strains and culture conditions
`
`Escherichia coli DH50. was used for cloning
`experiments. The S. cerevisiae strains used in this
`study were V5 (ScVSM, MA Ta, ura3) derived from
`a Champagne wine strain (variety bayanus) and the
`enological strains 861 and 8267 (variety cerevi—
`siae). E. coli cultivation and media were as de—
`scribed previously (Sambrook et a[., 1989). S.
`cerevis/ae was maintained and grown in YPD
`medium (1% bacto yeast extract, 2% bactopep-
`tone, 2% glucose).
`Batch fermentation experiments were carried
`out in minimal synthetic medium YNB (967%
`yeast nitrogen base without amino acid, 10% glu~
`cose, pH 3-3 with hydrochloric acid and supple-
`mented with (15% casamino acids) or
`in MS
`synthetic must medium simulating a standard
`grape juice containing 18—20% glucose, as de—
`scribed by Bely er a1. (1990), but without proline.
`Nitrogen was in the form of 80 mg/l ammoniacal—
`nitrogen (NH4C1) and 120 mg/l
`at amino acid—
`nitrogen.
`The inoculum was grown for approximately
`36h at 28°C in 50 ml flasks without agitation.
`Fermentations were realized by inoculation of
`precultured cells at a density of 1 x 10" per ml in
`fermentors with a working volume of 200 ml or 1,1
`1
`(no differences in the results due to difierent
`cultivation volumes were observed) equipped with
`fermentation locks. Fermentations were carried
`
`out at 28°C with permanent stirring (500 rpm).
`
`YEAST
`
`vor. 13: 783—793 (1997)
`
`© 1997 by John Wllcy & Sons, Ltd
`
`
`
`GLYCEROL AND ETHANOL YIELDS DURING ALCOHOLIC FERMEN'I‘A'I‘ION
`
`785
`
`automatic
`by
`determined
`release was
`CO2
`measurement of fermentor weight loss each 20 min
`(Bely et al, 1990). Fermentation was charac—
`terized by fermentation progress expressed as
`l-S/S0
`(S=glucose
`concentration,
`So=initial
`concentration).
`
`Preparation ofpolyclonal serum against Gpdlp
`protein, protein and Western blot analyses
`
`To produce antibodies against Gpdlp, GPDH
`protein was purified as described (Michnick, 1995)
`from the enological strain 861. A major protein
`exhibiting GPDH activity had a molecular weight
`of 43 kDa. This protein was
`isolated from
`Coomassie—stained
`sodium dodecyl
`sulphate
`(SDS)—polyacrylamide gel and the dried acryla—
`mide band was used for a rabbit immunization,
`performed by Eurogentec S. A. For Western blot
`studies, proteins from whole cell extracts were
`analysed by SDS—polyacrylamide
`gel
`electro~
`phoreses (PAGE; Sambrook et al., 1989). Gels
`were stained with 025% Coomassie blue R—ZSO
`
`in 40% methanol and 7% acetic acid.
`(Sigma)
`Analyses were carried out using the polyclonal
`serum and the alkaline phosphatase detection
`method (Sambrook et al, 1989). The protein con—
`centration in crude extracts was determined with
`
`the BCA Protein Assay Reagent (Pierce).
`
`DNA manipulation, cloning techniques and
`transformation methods
`
`Restriction and modification enzymes were used
`according to the manufacturer’s instructions. E.
`coli plasmid DNA was prepared using standard
`protocols (Sambrook et al, 1989). Oligonucleo~
`tides were synthesized by Eurogentec. E.
`coli
`transformation was carried out by the CaClZ/
`RbCl2 method (Hanahan, 1985). Transformation
`of S. cerevisiae was performed using the LiAc
`procedure (Schiestl and Gietz, 1989).
`
`OvereXpression and disruption of GPDI
`
`Polymerase chain reaction (PCR) cloning of the
`CPD] gene and introduction of Xhol and BamI-ll
`sites respectively at the 5’ and 3’ ends of the coding
`region was achieved using the oligonucleotides
`CGCTCGAGCCCCTCCACAAACACA,
`comp—
`lementary to a
`region upstream of the ATG
`start codon (nucleotides —34 to ~18) and GC
`GGATCCGGGGAAGTATGATATGTT,
`corre—
`
`sponding to a region located 17 nucleotides down—
`stream of
`the
`stop codon (Albertyn et al,
`
`(including
`1994b). The 1260 bp DNA fragment
`sites) defined by the two primers was amplified
`with total DNA from V5 strain as a template and
`directly cloned into the pGEM—T vector (Promega)
`to give the pGEM-T-GPDJ vector. The insert was
`partially sequenced to verify the identity of the
`gene. For construction of the CPD] expression
`vector, the PCR fragment was cut with XhoI and
`Baml-II and cloned into the Xhol and BamHI sites
`
`of the yeast expression vector pVTlOO—U (Vernet
`et al, 1987)
`to give the pVT100»U-GPDI vector
`used to transform V5 strain.
`
`Disruption of CPD] was obtained by internal
`deletion of the BsmI~XbaI 800 bp fragment of
`pGEM—T—GPDI,
`giving pGEM-T-gpdlA. The
`URA3 BglII 12 kb fragment from pUT332 (Cayla)
`was then inserted in the BglII site of pGEM-T-
`gdeA. The gpdiA gene was isolated by digestion
`of pGEM—T—gpdlA by XhoI and BamHI. The
`resulting 16 kb XhoI—BamHI fragment was used to
`transform V5 strain. Replacement of the chromo—
`somic CPD] allele by the inactivated copy was
`confirmed by PCR and southern blot analysis of
`genomic DNA.
`
`Enzyme assays
`
`(EC
`dehydrogenase
`3—phosphate
`Glycerol
`1.1.1.8) and glycerol 3—phosphatase (EC 3.1.3.21)
`were assayed according to the method of Gancedo
`et a]. (1968), alcohol dehydrogenase (EC 1.1.1.1)
`as described by Millan et al. (1987) and pyruvate
`decarboxylase (EC 4.1.1.1) according to Schmitt
`and Zimmermann (1982)
`in crude extracts ob—
`tained by vortexing yeast cells with glass beads
`(0-5 mm in diameter) for 4 min at 4°C. Specific
`activities were expressed as umol or nmol substrate
`degraded per min and mg protein.
`
`Analytical methods
`Yeast cells were counted using an electronic
`particle counter (ZM, Coultronics). Glucose, gly-
`cerol, ethanol, pyruvate, acetate and succinate
`were analysed by HPLC on an HPX—87H Aminex
`column (BioRad). Elution was performed at 45°C
`with 8 mM-HZSO4 at a flow rate of 0-6 ml/min.
`Detection was performed by means of dual detec—
`tion: refractometer (Shimadzu) and UV detector
`(Shimadzu SDD—ZA, X2214 nm). Quantification
`was performed using external standards prepared
`from pure compounds (Sigma) and an HP 3365
`integration system. Acetaldehyde was
`deter—
`mined enzymatically according to the method of
`
`© 1997 by John Wiley & Sons, Ltd.
`
`YEAST
`
`VOL. 13: 783—793 (1997)
`
`
`
`786
`
`i
`
`)>
`
`Numberoicells/mlx106
`
`
`
`Glycerol(git)
`
`0.0
`
`0,2
`
` .,._fi_.A.T—l
`
`0.4
`
`0.6
`
`0.5
`
`1.0
`
`1.5/50
`
`Figure l. The kinetics 01’ glycerol production by .S'. (screws/ac
`strains 561 and 8267 during alcoholic fermentation.
`(A)
`Growth curve; (B) glycerol production; 861 (open circles): 8267
`(closed circles). Fermentation was carried out on MS medium
`containing 180 g/l glucose, pH 3-3.
`
`Lundquist (1974). Acetaldehyde-ammoniac trimer
`(Aldrich) was used as standard. Acetoin and 2,3
`butane—diol were determined by gas chromatog—
`raphy (GC). They were extracted as described by
`Hagenauer—Hener er a1.
`(1990). Samples
`(2 ml)
`containing an internal standard (hexanol 04% v/v
`in methanol) were saturated with 2-5 g KZCO3 and
`extracted by chloroform (2 ml). The organic phase
`was dried with NaZSO4. 1 ul of a dilution l/l (v/v)
`with methanol was injected on an HP 5890 appa—
`ratus fitted with a DBWAX megabore column
`(Jandel). Injector and detector temperatures were
`respectively 240°C and 250°C. Oven temperature
`was kept at 80°C for 2 min and then programmed
`from 80°C to 200°C at 10°C/min.
`(Meso)
`2,3—
`butane—diol was separated from optically active
`forms. Enantiomers of butane~diol and acetoin
`were not resolved.
`
`RESULTS
`
`Gpdlp is a limiting enzyme for glycerol production
`When grown anaerobically on synthetic must
`(180 g/l glucose), $61 and 8267 strains produced
`7-3 and 11 g/l of glycerol respectively (Figure 18).
`The glycerol production kinetics were biphasic in
`both strains, with a high production rate during
`
`S. MlCl-[NICK ET AL.
`
`the growth phase and a slowing after the cells
`entered the stationary phase (Figure 1A), until
`complete depletion of sugar. Differences in glyc—
`erol
`levels between the two strains were mainly
`achieved during this second phase of production.
`The activities of enzymes specifically involved in
`ethanol (PDC and ADH) or in glycerol (GPDH
`and GBPase) production were monitored during
`this second phase (Figure 2) in order to determine
`whether these differences may be related to a
`limiting step either in ethanol or glycerol path-
`ways. Although the low (861) and high (8267)
`glycerol-forming strains exhibited similar PDC,
`ADI-l and GBPase specific activities, significant
`variations between the two strains were found with
`
`GPDH. These data suggest that the level of GPDH
`activity may have some importance in the capacity
`of yeast to form glycerol during fermentation.
`In order to estimate the amount of GPDH
`
`protein in S61 and 8267, this protein was purified
`from strain 8267, which displayed the highest
`glycerol yield (Michnick,
`1995). The purified
`GPDH protein was shown to correspond to
`Gpdlp since its internal sequence was identical to
`the protein sequence deduced from the nucleotidic
`sequence of CPD] (Larsson er al, 1993; Albertyn
`et al., 1994b). Antibodies were raised against the
`purified protein. GPDH production was moni—
`tored during fermentation on YNB medium con—
`taining 180 g/l of glucose, pH 3-3 (Figure 3).
`Proteins from crude cell extracts were resolved by
`SDS~PAGE and analysed by Western blotting.
`For both strains, GPDl—l protein was detected at
`the expected molecular weight of 43 kDa. The
`maximum amount of GPDH protein was detected
`towards the end of fermentation for both strains
`
`(Figure 3), in general agreement with the increase
`in GPDH activity observed at this stage (Figure
`2D). GPDI—l protein in 861 strain was found to be
`induced a little later than expected from GPDI—l
`activity data, but the results obtained with 8267
`are in full agreement with the variation in GPDH
`activity. In addition,
`the amount of GPDH was
`generally smaller for the strain with a low glycerol
`yield (861) in comparison with strain 8267.
`The glycerol formation capacity may thus be
`related to variations in the amount of GPDH.
`
`Since two GPDH isoenzymes which exhibit 69%
`identity in amino acid sequence are found in S.
`cerevisiae, encoded by GPDI and GPD2 genes
`(Larsson et al, 1993; Albertyn et al, 1994b;
`Ericksson et al, 1995), both may have been
`detected. However.
`results obtained with the
`
`YEAST
`
`VOL. 13: 783—793 (1997)
`
`© 1997 by John Wiley & Sons, Ltd
`
`
`
`GLYCEROL AND ETHANOL YIELDS DURING ALCOHOLIC FERMEN'I‘A'I'ION
`
`787
`
`700
`
`A
`§ 600
`D.
`E’
`V
`2 500
`a 400
`n.
`
`300
`
`3.5
`30
`
`A
`g 2.5
`m
`g
`2.0
`a 1.5
`I
`2
`
`1.0
`0.5
`0'0
`
`..
`
`A
`
`‘
`
`0.2
`
`0.4
`
`0.6
`
`0.5
`
`1.0
`
`'
`
`C
`
`0.2
`
`0.4
`
`_r”"'—|
`0.8
`1.0
`
`0.6
`
`250
`
`200
`
`A
`*5
`B.
`U)
`g 150
`100
`E
`i
`a
`
`-
`
`B
`
`.
`
`50
`0 """”‘”7—l—r—’1
`0.2
`0.4
`0.6
`0.5
`1.0
`
`r:
`2
`a
`E’
`5
`15,
`I
`E
`o
`
`
`
`0.2
`
`0,4
`
`0.6
`
`0.8
`
`1.0
`
`1-3/8o
`
`1.518.,
`
`Figure 2. Variations in GPDH activity between $61 and 8267 strains. Specific activities of (A)
`PDC, (B) G3Pase, (C) ADH and (D) GPDH were monitored during the stationary phase. 861
`(open circles); 8267 (closed circles). Three individual determinations were performed. Fermen~
`ration was carried out under the same conditions as described in Figure 1.
`
`MW w New. WWWMM
`
`W
`
`
`
`Y“:
`
`:1“ i
`
`m
`r
`' 1m "-
`3:
`
`a;
`23""
`..
`
`‘
`I
`r j
`i
`c
`I:
`d E,
`~
`r“
`tum-w» «iv ”an E ..
`
`t
`‘
`
`s-
`t it}!
`.‘i
`i’
`.g
`ii
`3r
`
`it“
`
`3
`
`t"
`ii
`>
`
`N"
`
`We“
`
`.~*
`
`WT...
`CH!
`
`v‘
`V
`;
`
`’3 5‘
`9
`g .2
`~35
`
`. E
`WM .. L
`! il
`
`
`
`
`
`2"
`N
`-'
`at
`a
`
`3
`Q?)
`'5
`s
`g
`5
`.3);
`
`3 51$.
`
`$3311.
`
`a
`
`l:
`
`I;
`
`u
`
`1"
`
`f:
`
`’0‘
`
`ti“
`
`w»
`
`emww ~--——-—~--~- W O“- Qkfia «my. WW
`
`Figure 3. Expression level of Gpdi protein in 861 and 8267 strains during alcoholic
`fermentation.
`(A) Growth (circles) and glycerol production (triangles).
`(B) Western blot
`analysis. 10 pg of protein extracted from 861 and $267 at the times indicated was anaiysed by
`immunoblotting using anti-Gpdip polyclonal serum. Fermentation was carried out on YNB
`medium containing 180 g/l of glucose. pH 3-3.
`
`(C) 1997 by ,lohn Wiley 8: Sons. Ltd.
`
`YEAST
`
`VOL. 13: 783493 (1997)
`
`
`
`788
`
`S. MICIINICK 13'1" AL.
`
`gpdlA mutant (this study) strongly suggest that
`under the conditions used for the determination of
`
`GPDH activity, Gdep activity was not detected
`and that the antibodies directed against Gpdlp
`were specific for this isoform. This suggests that
`most of the GPDH activity and most of the
`protein detected corresponded to Gpdlp. Surpris-
`ingly, the in vitro Gpdlp activity level and amount
`were highest during the stationary phase when
`glycerol production is the lowest. This indicates
`that GPDH activity is regulated in viva. Neverthe—
`less, these results strongly suggest that the amount
`of Gpdlp is one of the limiting factors for gly—
`cerol production in S, cerevislae under enological
`conditions, with the possibility of a diversion of
`the carbon flux towards glycerol or ethanol by
`modulation of the level of expression of the CPD]
`gene.
`
`Overexpression and disruption of GPDI strongly
`modifi/ the yield of glycerol and ethanol
`The consequences of overexpression and dele—
`tion of CPD] on glycerol production were studied
`on minimal medium YNB with 100 g/l of glucose,
`pH 33. Transformants carrying CPD] under the
`control of the ADH] promoter and terminator on
`a multicopy plasmid (VS/GPDI) displayed high
`GPDH specific activity during the growth phase
`with a maximum level observed when the cells
`
`reached the stationary phase. corresponding to a
`15—fold increase compared to the control strain
`(Figure 4A). In contrast, GPDH specific activity
`for the control strain (VS/pVTU) was low and
`constant during the growth phase; it then increased
`and reached a peak of 80 mU/mg protein during
`the second part of fermentation (LS/80:06, Fig—
`ure 4A). Western blot analysis was performed on
`protein from whole cell extracts (Figure 4B). (3de
`protein was
`shown to be produced in large
`amounts in VS/GPDI in agreement with the high
`
`
`
`w «m r...» W . ms
`
`Figure 4. GPDH activity and Western blot analysis in S,
`cerevisiae overexpressing CPD]. (A) Specific GPDH activity in
`crude extracts of VS/GPDl (closed circles) and in V5/pVTU
`(open circles) cells.
`(B) 10 ug of proteins extracted from V5/
`CPD] cells was analysed by immunoblotting using anti~Gpd1p
`antiserum. Fermentation was carried out on minimal medium
`YNB containing 100 g/l glucose. pH 38.
`
`level of activity, whereas it was only detected in the
`sample displaying the highest GPDH activity (l—S/
`80:06) in V5/pVTU (data not shown). Although
`the activity was found to decrease during the
`stationary phase, the protein was always detected
`in large amounts, suggesting in Vivo regulation
`mechanisms. GPDH activity (related to Gpdl
`isoenzyme) was not detected in the gpdlA mutant
`when grown under the same conditions as those
`described in Figure 4, and Gpdl protein was not
`detected by Western blot analysis
`(data not
`shown).
`As a result of overproduction of (3de protein,
`the increase in the glycerol yield was more than
`three times greater in VS/GPDI in comparison
`with the control strain (Table 1). The large increase
`
`Table 1.
`
`Glycerol and ethanol yields in V5/pVTU and Vii/CPD]. The cells were grown on
`YNB medium containing 100 g/l glucose, pH 33. Metabolites were analysed after complete
`glucose depletion.
`
`Molar ratio of
`
`
`
` Strain Glycerol (g/l) Ethanol (g/l) Glycerol/glucose Ethanol/glucose
`
`
`
`
`
`
`
`VS/pVTU
`VS/GF’DJ
`vsgpo/,A
`
`43
`14
`2.4
`
`46-8
`366
`49.2
`
`0-08
`0-27
`005
`
`183
`143
`1.92
`
`YEAST
`
`vor. 13: 783—793 (1997)
`
`© 1997 by John Wiley & Sons, Ltd
`
`
`
`GLYCIEROL AND ETHANOL YIIELDS DURING ALCOHOLIC I’ERMEN'I‘A'I'ION
`
`789
`
`A
`
`1000
`
`100
`
`10
`
`.. 0
`H
`
`
`
`0.0
`
`0.2
`
`0.4
`
`0.6
`
`0.6
`
`1.0
`
`1-S/Sa
`
`
`
`Ethanol(9/1)
`
`
`
`Pyruvate(9/!)
`
`20
`
`0.8
`0.6
`
`0.4 .
`
`0.2
`
`0.0
`
`E
`
`0.0
`
`0.2
`
`CA
`
`0.6
`
`0.8
`
`1,0
`
`1 6/80
`
`u3:,
`><
`g
`
`8Ea
`
`;.0
`
`E2|
`2
`
`e3E82
`
`‘
`o
`
`a3O
`
`J2.
`
`C(B
`1:.
`.‘E
`8<(
`
`The effects of CPD] overexpression on the kinetics of glycerol, ethanol,
`Figure 5.
`acetaldehyde and pyruvate production under enological conditions.
`(A) Growth curves;
`(B—E) production of glycerol, ethanol, acetaldehyde and pyruvate. V5/pVTU (open
`circles); VS/CPDI (closed circles) Fermentation was carried out on MS medium contain~
`ing 200 g/l glucose, p113-3.
`
`instead of 4-3 g/l)
`in glycerol production (14 g/l
`took place simultaneously with a decrease in the
`production of ethanol, the molar ratio of ethanol
`to glucose decreasing by 20% for the transformant
`overproducing glycerol. In contrast,
`the glycerol
`production level was reduced for the strain dis—
`rupted for CPD] whereas the ethanol yield was
`slightly enhanced (Table 1). Residual glycerol pro—
`duction was 50% for the gdeA strain, suggesting
`that both CPD] and GPDZ genes may contribute
`to glycerol formation under enological conditions.
`
`Consequences ofgfycerol ave/production under
`etiological—like conditions
`The kinetics of formation of glycerol, ethanol
`and intermediate metabolites were monitored in
`
`the transformant overexpressing CPD] during fer—
`
`in
`mentation on synthetic must (glucose 200 g/l)
`order to examine how S. cerevisiae reacts to a
`
`glycolytic flux diverted to glycerol. Growth and
`production of glycerol, ethanol, acetaldehyde and
`pyruvate are shown in Figure 5. A strong increase
`in the glycerol rate was observed, leading to the
`production of 28 g/l compared to 7 g/l for the con-
`trol strain (Figure SB), while ethanol was produced
`at a lower rate, resulting in a decrease in the final
`concentration of about 15 g/l (Figure 5C). Glycerol
`yield was about four times higher and ethanol yield
`was reduced about 20% for VS/GPDI compared
`with V5 (Table 2). In consequence, acetaldehyde
`and pyruvate were transiently accumulated in the
`strain that overproduced glycerol (Figure 5D,E)i
`Peak level was
`reached for both compounds
`when the cells entered the stationary phase;
`the
`concentration in these compounds subsequently
`
`thJ 1997 by John Wiley & Sons, Ltd,
`
`YEAST
`
`VOL. 13: 783493 (1997)
`
`
`
`790
`
`S. MICHNICK 111‘ AL.
`
`Table 2. Concentrations and yields of fermentation products and balances in V5/pVTU
`(control) and V5/GPDI, after anaerobic growth in MS medium containing 200 g/l glucose.
`pH 33.
`
`Value8 for the strain:
`
`V5/pVTU
`V5/GPDI
`
`
`Ethanol (g/l)
`(mmol/moi glucose)
`Carbon dioxide (g/l)
`(mmol/mol glucose)
`Glycerol (g/I)
`(mmol/mol glucose)
`Pyruvic acid" (g/l)
`(mmol/mol glucose)
`AcetaldehydeC (g/l)
`(mmol/mol glucose)
`Acetate (g/l)
`(mmol/mol glucose)
`Acetoin (g/l)
`(mmol/mol glucose)
`2,3 Butane-diol (g/l)
`(mmol/mol glucose)
`Succinate (g/l)
`(mmol/mol glucose)
`Biomass formation‘J
`dry weight (g/l)
`Carbon recovery (%)
`Redox balancee (‘36)
`
`
`
`8942 :t 3-6
`1746 :: 70
`94-5 :: 0-07
`1934 i 1
`7-1 :t 0-7
`69 :: 6
`0-13
`14
`<O-1
`
`0-52 :1: 0-06
`7-7 5: 1-0
`<01
`
`0-9 :: 0-4
`89 :: 4-4
`025 i 006
`19 i 05
`
`
`
`4 2 :: 0-8
`97-6 i 2-4
`95-3 i 36
`
`72-5 51: 2-1
`1417 :1: 42
`85-5 3: 0-02
`1748 :1: 58
`28-6 21:1-4
`279 i 13
`0-32
`3-6
`0-22 :1: 0-01
`45 :t 0-3
`1-6 :: 0-1
`
`23 :: 1
`61 :1: 0-8
`62 i 9
`13 :1: 0-1
`13 i 1
`054 i 004
`4-1 :1: 0-3
`
`2-55 a: 007
`98-2 1- 0-3
`96-4 i 06
`
`“Except as noted. all values are expressed as means derived from two independent experiments.
`including standard deviations.
`I’Single experiment.
`“Acetaldehyde loss was always less than 100 mg throughout the fermentation
`“’l'he
`carbon-molar mass
`of
`biomass was
`estimated
`using
`the
`elemental
`(C4l’l7.3202,24N0.5880.W) used by van Dijken and Schefiers (1986).
`l'The redox balance represents the ratio between the reductance degree of fermentation products
`(including biomass) and the reductance degree of glucose. expressed as a percentage.
`
`composition
`
`decreased until glucose was completely depleted
`and remained slightly higher than for the control
`strain at the end of fermentation (Table 2).
`Variation in the production of by—products of
`acetaldehyde and/or pyruvate was investigated by
`HPLC analysis (Table 2). The acetate level
`in—
`creased (three~fold) for the transformed strain in
`comparison with the control strain. Smaller but
`significant accumulation (about two—fold) was also
`observed for succinate. Large amounts of two
`other compounds eluted at retention times corre—
`sponding to those of acetoin and 2,3 butane—diol
`were also detected for the CPD] strain. The iden-
`
`tity of these compounds was further confirmed by
`GC analysis. Substantial accumulation of acetoin
`was observed for the CPD] strain (6-1g/l). The
`
`the product of acetoin
`level of 2,3 butaneediol,
`reduction, was 115~fold that of the V5/pVTU
`strain. The carbon and redox balances were equili-
`brated satisfactorily for both strains.
`The consequences of CPD] overexpression on
`growth are shown in Figure 5A and Table 2.
`VS/GPDJ cells reached the stationary phase earlier
`than the control strain (30 h compared with 45 h).
`and the final quantity of biomass for the strain
`overexpressing GPDJ was 16 times lower than
`that of the V5/pVTU strain.
`
`DISCUSSION
`
`There have been few previous investigations of the
`regulation of glycerol production under enological
`
`YEAST vor. 13: 783—793 (1997)
`
`© 1997 by John Wiley & Sons. Ltd
`
`
`
`GLYCIL‘ROL AND ETHANOL YIELDS DURING ALCOHOLIC FERMEN'I'A'l'ION
`
`791
`
`conditions. In the present study, it was shown that
`in vitro GPDl—l specific activity and the amount of
`Gpdlp isoenzyme monitored during fermentation
`were higher in a high glycerol—forming strain than
`in a low glycerol-forming one, whereas activities of
`other enzymes involved in glycerol or ethanol
`formation were similar for both strains, suggesting
`that GPDH is a limiting step for glycerol pro—
`duction under enological conditions. The strong
`modification of glycerol yield for the strains over—
`expressed or disrupted for CPD] supports this
`assumption.
`The rate of glycerol production was shown to be
`high during the growth phase, in agreement with
`the role of glycerol in the reoxidation of NADH
`generated by anabolic processes Surprisingly, the
`GPDH synthesis level was shown to be low during
`this phase and high during the stationary phase
`when glycerol production slowed down (Figures
`2D, 3 and 4A). These results are rather unexpected
`since a 30—60% variation in GPDH activity during
`the stationary phase results in a significant differ—
`ence in glycerol production, as observed for strains
`861 and 3267. However,
`regulation at GPDH
`synthesis level and in viva regulation of its activity
`are both involved in the control of glycerol pro—
`duction. Enzymatic studies on purified prepara—
`tions of GPDH (Nader et a], 1979; Albertyn er a],
`1992) and of Gpdlp isoform (Michnick, 1995)
`from Saccharomyces strains have shown that
`GPDH activity is modulated by physiological con—
`centrations of ATP, ADP. NAD and fructose 1,6
`biphosphate.
`An intriguing finding is the induction of GPDH
`during the phase of low glycerol production.
`Might this synthesis meet a physiological require—
`ment? The existence of NADH—generating metab—
`olism (i.e. via the oxidative branch of the TCA
`cycle) unconnected with biomass formation at this
`stage of fermentation cannot be ruled out. How—
`ever, other mechanisms
`such as
`response to
`environmental conditions might be
`involved.
`Since osmotic stress is moderated under enological
`conditions (Jones and Greenfield, 1985), its partici—
`pation in the Gpdlp induction observed seems
`rather unlikely. On the other hand,
`involvement
`of regulation mechanisms related to stationary
`phase (which mainly result from nitrogen and
`micronutrient starvation under enological condi-
`tions; Bely er a], 1990) and to different stresses
`(high ethanol concentration,
`low pH) cannot be
`ruled out. This aspect clearly requires further
`attention, especially as evidence has been found
`
`that the expression of some stress-induced genes is
`controlled differently under enological conditions
`(glucose 20%, pH 36) in comparison with stan—
`dard laboratory conditions (glucose 2%, pH 6;
`Riou er a], 1997).
`The contribution and the role of the two GPDH
`
`isoenzymes in glycerol production during alcoholic
`fermentation is not known. Under the conditions
`
`used here, gdeA displayed a residual glycerol
`production of 50%, suggesting that both CPD]
`and GPDZ genes might be involved in glycerol
`production during anaerobiosis. Although no
`GPDI—l activity could be detected for this mutant
`under our experimental conditions, glycerol was
`certainly produced by means of Gdep, since most
`of the CPD] coding region was deleted and only
`two GPD isogenes are found in S. cerevjsjae (ge-
`nome sequence data). However, the possibility that
`GPDZ is only expressed in the gpd] deletion mu-
`tant cannot be ruled out. Expression studies of
`CPD] and GPDZ under enological conditions
`might show whether both genes are expressed in a
`wild—type yeast under enological conditions (in
`progress).
`Strains engineered for CPD] presenting strongly
`modified yields of glycerol and ethanol were
`obtained. Besides their potential interest for tech—
`nological processes,
`these strains provide direct
`genetic evidence for a main role of GPDI—I in glyc—
`erol formation. The S. cerevjsiae gdeA strain pre-
`sented half the glycerol yield and a small increase in
`ethanol production, in agreement with the results
`reported for S. diastarjcus gdeA strain (Wang
`et a], 1994), Although no increase in glycerol pro—
`duction has been shown previously in S. cerevisiae
`strains amplified for CPD] during fermentation on
`2% glucose (Albertyn et .21.. 1994b), amplification
`of CPD] in S. cerevisfae under the control of the
`
`ADH] promoter triggered substantial overproduc—
`tion of extracellular glycerol during growth on
`sugar-rich medium. Simple dilfusion and facilitated
`diffusion by means of the channel protein Fspip
`were suggested as being involved in glycerol efliux
`(Luyten er a], 1995). This efiiux seems not to be a
`limiting factor since VS/GPD] did not accumulate
`more intracellular glycerol than the control strain
`during the fermentation (data not shown).
`Furthermore, study of the strain overexpressed
`for GPDl provides strong evidence for the role of
`glycerol production in adjusting the redox balance.
`As a result of (i) diversion of the carbon flux
`towards glycerol and (ii) enhanced utilization
`of NADH through the glycerol pathway due to the
`
`(C) 1997 by John Wiley & Sons, Ltd.
`
`YEAST
`
`VOL. 13: 783—793 (1997)
`
`
`
`792
`
`S. MICI’INICK ET AL.
`
`15—fold increase in GPDH specific activity, ethanol
`formation was limited, resulting in transient accu—
`mulation of acetaldehyde, Acetaldehyde is a toxic
`compound with pleiotropic effect