Cytology, Taxonomy and Ecology of Grape and Wine …...Cytology,Taxonomy and Ecology of Grape and...

1

Cytology, Taxonomy and Ecologyof Grape and Wine Yeasts

1.1 Introduction 11.2 The cell wall 31.3 The plasmic membrane 71.4 The cytoplasm and its organelles 111.5 The nucleus 141.6 Reproduction and the yeast biological cycle 151.7 The killer phenomenon 191.8 Classification of yeast species 221.9 Identification of wine yeast strains 35

1.10 Ecology of grape and wine yeasts 40

1.1 INTRODUCTION

Man has been making bread and fermented bev-erages since the beginning of recorded history.Yet the role of yeasts in alcoholic fermentation,particularly in the transformation of grapes intowine, was only clearly established in the middleof the nineteenth century. The ancients explainedthe boiling during fermentation (from the Latinfervere, to boil) as a reaction between substances

that come into contact with each other duringcrushing. In 1680, a Dutch cloth merchant, Antonievan Leeuwenhoek, first observed yeasts in beerwort using a microscope that he designed andproduced. He did not, however, establish a rela-tionship between these corpuscles and alcoholicfermentation. It was not until the end of the eigh-teenth century that Lavoisier began the chemicalstudy of alcoholic fermentation. Gay-Lussac con-tinued Lavoisier’s research into the next century.

Handbook of Enology Volume 1 The Microbiology of Wine and Vinifications P. Ribereau-Gayon, D. Dubourdieu, B. Doneche and A. Lonvaud 2006 John Wiley & Sons, Ltd

2 Handbook of Enology: The Microbiology of Wine and Vinifications

As early as 1785, Fabroni, an Italian scientist, wasthe first to provide an interpretation of the chem-ical composition of the ferment responsible foralcoholic fermentation, which he described as aplant–animal substance. According to Fabroni, thismaterial, comparable to the gluten in flour, waslocated in special utricles, particularly on grapesand wheat, and alcoholic fermentation occurredwhen it came into contact with sugar in the must. In1837, a French physicist named Charles Cagnardde La Tour proved for the first time that the yeastwas a living organism. According to his findings,it was capable of multiplying and belonged to theplant kingdom; its vital activities were at the baseof the fermentation of sugar-containing liquids.The German naturalist Schwann confirmed his the-ory and demonstrated that heat and certain chem-ical products were capable of stopping alcoholicfermentation. He named the beer yeast zucker-pilz, which means sugar fungus—Saccharomycesin Latin. In 1838, Meyen used this nomenclaturefor the first time.

This vitalist or biological viewpoint of the roleof yeasts in alcoholic fermentation, obvious tous today, was not readily supported. Liebig andcertain other organic chemists were convinced thatchemical reactions, not living cellular activity,were responsible for the fermentation of sugar.In his famous studies on wine (1866) and beer(1876), Louis Pasteur gave definitive credibilityto the vitalist viewpoint of alcoholic fermentation.He demonstrated that the yeasts responsible forspontaneous fermentation of grape must or crushedgrapes came from the surface of the grape;he isolated several races and species. He evenconceived the notion that the nature of the yeastcarrying out the alcoholic fermentation couldinfluence the gustatory characteristics of wine. Healso demonstrated the effect of oxygen on theassimilation of sugar by yeasts. Louis Pasteurproved that the yeast produced secondary productssuch as glycerol in addition to alcohol and carbondioxide.

Since Pasteur, yeasts and alcoholic fermen-tation have incited a considerable amount ofresearch, making use of progress in microbiology,

biochemistry and now genetics and molecularbiology.

In taxonomy, scientists define yeasts as unicel-lular fungi that reproduce by budding and binaryfission. Certain pluricellular fungi have a unicellu-lar stage and are also grouped with yeasts. Yeastsform a complex and heterogeneous group foundin three classes of fungi, characterized by theirreproduction mode: the sac fungi (Ascomycetes),the club fungi (Basidiomycetes), and the imper-fect fungi (Deuteromycetes). The yeasts found onthe surface of the grape and in wine belong toAscomycetes and Deuteromycetes. The haploidspores or ascospores of the Ascomycetes class arecontained in the ascus, a type of sac made fromvegetative cells. Asporiferous yeasts, incapable ofsexual reproduction, are classified with the imper-fect fungi.

In this first chapter, the morphology, repro-duction, taxonomy and ecology of grape andwine yeasts will be discussed. Cytology is themorphological and functional study of the struc-tural components of the cell (Rose and Harrison,1991).

Fig. 1.1. A yeast cell (Gaillardin and Heslot, 1987)

Cytology, Taxonomy and Ecology of Grape and Wine Yeasts 3

Yeasts are the most simple of the eucaryotes.The yeast cell contains cellular envelopes, acytoplasm with various organelles, and a nucleussurrounded by a membrane and enclosing thechromosomes. (Figure 1.1). Like all plant cells,the yeast cell has two cellular envelopes: thecell wall and the membrane. The periplasmicspace is the space between the cell wall andthe membrane. The cytoplasm and the membranemake up the protoplasm. The term protoplastor sphaeroplast designates a cell whose cellwall has been artificially removed. Yeast cellularenvelopes play an essential role: they contributeto a successful alcoholic fermentation and releasecertain constituents which add to the resultingwine’s composition. In order to take advantage ofthese properties, the winemaker or enologist musthave a profound knowledge of these organelles.

1.2 THE CELL WALL

1.2.1 The General Roleof the Cell Wall

During the last 20 years, researchers (Fleet, 1991;Klis, 1994; Stratford, 1999; Klis et al., 2002) havegreatly expanded our knowledge of the yeast cellwall, which represents 15–25% of the dry weightof the cell. It essentially consists of polysaccha-rides. It is a rigid envelope, yet endowed with acertain elasticity.

Its first function is to protect the cell. Withoutits wall, the cell would burst under the internalosmotic pressure, determined by the compositionof the cell’s environment. Protoplasts placed inpure water are immediately lysed in this manner.Cell wall elasticity can be demonstrated by placingyeasts, taken during their log phase, in a hypertonic(NaCl) solution. Their cellular volume decreasesby approximately 50%. The cell wall appearsthicker and is almost in contact with the membrane.The cells regain their initial form after being placedback into an isotonic medium.

Yet the cell wall cannot be considered an inert,semi-rigid ‘armor’. On the contrary, it is a dynamicand multifunctional organelle. Its composition andfunctions evolve during the life of the cell, in

response to environmental factors. In addition toits protective role, the cell wall gives the cellits particular shape through its macromolecularorganization. It is also the site of moleculeswhich determine certain cellular interactions suchas sexual union, flocculation, and the killerfactor, which will be examined in detail later inthis chapter (Section 1.7). Finally, a number ofenzymes, generally hydrolases, are connected tothe cell wall or situated in the periplasmic space.Their substrates are nutritive substances of theenvironment and the macromolecules of the cellwall itself, which is constantly reshaped duringcellular morphogenesis.

1.2.2 The Chemical Structureand Function of the ParietalConstituents

The yeast cell wall is made up of two prin-cipal constituents: β-glucans and mannoproteins.Chitin represents a minute part of its composi-tion. The most detailed work on the yeast cellwall has been carried out on Saccharomyces cere-visiae —the principal yeast responsible for thealcoholic fermentation of grape must.

Glucan represents about 60% of the dry weightof the cell wall of S. cerevisiae. It can bechemically fractionated into three categories:

1. Fibrous β-1,3 glucan is insoluble in water,acetic acid and alkali. It has very few branches.The branch points involved are β-1,6 linkages.Its degree of polymerization is 1500. Underthe electron microscope, this glucan appearsfibrous. It ensures the shape and the rigidity ofthe cell wall. It is always connected to chitin.

2. Amorphous β-1,3 glucan, with about 1500glucose units, is insoluble in water but solublein alkalis. It has very few branches, like thepreceding glucan. In addition to these fewbranches, it is made up of a small number ofβ-1,6 glycosidic linkages. It has an amorphousaspect under the electron microscope. It givesthe cell wall its elasticity and acts as an anchorfor the mannoproteins. It can also constitute anextraprotoplasmic reserve substance.


3. The β-1,6 glucan is obtained from alkali-insoluble glucans by extraction in acetic acid.The resulting product is amorphous, water sol-uble, and extensively ramified by β-1,3 glyco-sidic linkages. Its degree of polymerization is140. It links the different constituents of thecell wall together. It is also a receptor site forthe killer factor.

The fibrous β-1,3 glucan (alkali-insoluble) proba-bly results from the incorporation of chitin on theamorphous β-1,3 glucan.

Mannoproteins constitute 25–50% of the cellwall of S. cerevisiae. They can be extracted fromthe whole cell or from the isolated cell wallby chemical and enzymatic methods. Chemicalmethods make use of autoclaving in the pres-ence of alkali or a citrate buffer solution atpH 7. The enzymatic method frees the manno-proteins by digesting the glucan. This methoddoes not denature the structure of the mannopro-teins as much as chemical methods. Zymolyase,obtained from the bacterium Arthrobacter luteus,is the enzymatic preparation most often used toextract the parietal mannoproteins of S. cerevisiae.This enzymatic complex is effective primarilybecause of its β-1,3 glucanase activity. The actionof protease contaminants in the zymolyase com-bine, with the aforementioned activity to liberatethe mannoproteins. Glucanex, another industrialpreparation of the β-glucanase, produced by a fun-gus (Trichoderma harzianum), has been recentlydemonstrated to possess endo- and exo-β-1,3 andendo-β-1,6-glucanase activities (Dubourdieu andMoine, 1995). These activities also facilitate theextraction of the cell wall mannoproteins of theS. cerevisiae cell.

The mannoproteins of S. cerevisiae have amolecular weight between 20 and 450 kDa. Theirdegree of glycosylation varies. Certain ones con-taining about 90% mannose and 10% peptides arehypermannosylated.

Four forms of glycosylation are described(Figure 1.2) but do not necessarily exist at thesame time in all of the mannoproteins.

The mannose of the mannoproteins can consti-tute short, linear chains with one to five residues.

They are linked to the peptide chain by O-glycosyllinkages on serine and threonine residues. Theseglycosidic side-chain linkages are α-1,2 and α-1,3.

The glucidic part of the mannoprotein can alsobe a polysaccharide. It is linked to an asparagineresidue of the peptide chain by an N -glycosyllinkage. This linkage consists of a double unit ofN -acetylglucosamine (chitin) linked in β-1,4. Themannan linked in this manner to the asparagineincludes an attachment region made up of a dozenmannose residues and a highly ramified outerchain consisting of 150 to 250 mannose units.The attachment region beyond the chitin residueconsists of a mannose skeleton linked in α-1,6with side branches possessing one, two or threemannose residues with α-1,2 and/or α-1,3 bonds.The outer chain is also made up of a skeleton ofmannose units linked in α-1,6. This chain bearsshort side-chains constituted of mannose residueslinked in α-1,2 and a terminal mannose in α-1,3. Some of these side-chains possess a branchattached by a phosphodiester bond.

A third type of glycosylation was describedmore recently. It can occur in mannoproteins,which make up the cell wall of the yeast. It consistsof a glucomannan chain containing essentiallymannose residues linked in α-1,6 and glucoseresidues linked in α-1,6. The nature of the glycan–peptide point of attachment is not yet clear, but itmay be an asparaginyl–glucose bond. This type ofglycosylation characterizes the proteins freed fromthe cell wall by the action of a β-1,3 glucanase.Therefore, in vivo, the glucomannan chain mayalso comprise glucose residues linked in β-1,3.

The fourth type of glycosylation of yeast manno-proteins is the glycosyl–phosphatidyl–inositolanchor (GPI). This attachment between the ter-minal carboxylic group of the peptide chain anda membrane phospholipid permits certain manno-proteins, which cross the cell wall, to anchorthemselves in the plasmic membrane. The regionof attachment is characterized by the followingsequence (Figure 1.2): ethanolamine-phosphate-6-mannose-α-1,2-mannose-α-1,6-mannose-α-1,4-glucosamine-α-1,6-inositol-phospholipid. A C-phospholipase specific to phosphatidyl inositoland therefore capable of realizing this cleavage


6M[M 6M 6M 6M ]n 6M 6M

2

M

2

M

2

M

2

M2

M

2

M

2

M

3

M

3

M

3

M

P

M

3

M

2

M3

M

3

M

2

M P2

M

3

2 3

M M

MP 6M6

Mβ 4 GNAcβ 4 GNAcβ NH Asn

3M 3M 2M 2M O Ser/Thr

(G,M) Xxx

lipid P Ins 6 GN 4 M 6 M 2 M 6 P (CH2)2 NH C O

Fig. 1.2. The four types of glucosylation of parietal yeast mannoproteins (Klis, 1994). M = mannose; G = glucose;GN = glucosamine; GNAc = N-acetylglucosamine; Ins = inositol; Ser = Serine; Thr = threonine; Asn = asparagine;Xxx = the nature of the bond is not known

was demonstrated in the S. cerevisiae (Flick andThorner, 1993). Several GPI-type anchor manno-proteins have been identified in the cell wall ofS. cerevisiae.

Chitin is a linear polymer of N -acetylglucos-amine linked in β-1,4 and is not generally found inlarge quantities in yeast cell walls. In S. cerevisiae,chitin constitutes 1–2% of the cell wall and isfound for the most part (but not exclusively) inbud scar zones. These zones are a type of raisedcrater easily seen on the mother cell under theelectron microscope (Figure 1.3). This chitinic scaris formed essentially to assure cell wall integrityand cell survival. Yeasts treated with D polyoxine,an antibiotic inhibiting the synthesis of chitin, arenot viable; they burst after budding.

The presence of lipids in the cell wall has notbeen clearly demonstrated. It is true that cell walls

Fig. 1.3. Scanning electron microscope photograph ofproliferating S. cerevisiae cells. The budding scars onthe mother cells can be observed


prepared in the laboratory contain some lipids(2–15% for S. cerevisiae) but it is most likelycontamination by the lipids of the cytoplasmicmembrane, adsorbed by the cell wall during theirisolation. The cell wall can also adsorb lipids fromits external environment, especially the differentfatty acids that activate and inhibit the fermentation(Chapter 3).

Chitin are connected to the cell wall or sit-uated in the periplasmic space. One of themost characteristic enzymes is the invertase (β-fructofuranosidase). This enzyme catalyzes thehydrolysis of saccharose into glucose and fruc-tose. It is a thermostable mannoprotein anchoredto a β-1,6 glucan of the cell wall. Its molecularweight is 270 000 Da. It contains approximately50% mannose and 50% protein. The periplasmicacid phosphatase is equally a mannoprotein.

Other periplasmic enzymes that have been notedare β-glucosidase, α-galactosidase, melibiase, tre-halase, aminopeptidase and esterase. Yeast cellwalls also contain endo- and exo-β-glucanases (β-1,3 and β-1,6). These enzymes are involved in thereshaping of the cell wall during the growth andbudding of cells. Their activity is at a maximumduring the exponential log phase of the populationand diminishes notably afterwards. Yet cells in thestationary phase and even dead yeasts containedin the lees still retain β-glucanases activity intheir cell walls several months after the completionof fermentation. These endogenous enzymes areinvolved in the autolysis of the cell wall during the

ageing of wines on lees. This ageing method willbe covered in the chapter on white winemaking(Chapter 13).

1.2.3 General Organization of the CellWall and Factors Affecting itsComposition

The cell wall of S. cerevisiae is made up of anouter layer of mannoproteins. These mannopro-teins are connected to a matrix of amorphous β-1,3glucan which covers an inner layer of fibrous β-1,3 glucan. The inner layer is connected to a smallquantity of chitin (Figure 1.4). The β-1,6 glucanprobably acts as a cement between the two lay-ers. The rigidity and the shape of the cell wallare due to the internal framework of the β-1,3fibrous glucan. Its elasticity is due to the outeramorphous layer. The intermolecular structure ofthe mannoproteins of the outer layer (hydrophobiclinkages and disulfur bonds) equally determinescell wall porosity and impermeability to macro-molecules (molecular weights less than 4500). Thisimpermeability can be affected by treating thecell wall with certain chemical agents, such asβ-mercaptoethanol. This substance provokes therupture of the disulfur bonds, thus destroying theintermolecular network between the mannoproteinchains.

The composition of the cell wall is stronglyinfluenced by nutritive conditions and cell age.The proportion of glucan in the cell wall increases

Cytoplasm

Cytoplasmic membrane

Mannoproteins and β-1,3 amorphous glucan

β - 1,3 fibrous glucan

Cell wall

Periplasmic space

External medium

Fig. 1.4. Cellular organization of the cell wall of S. cerevisiae


with respect to the amount of sugar in the cul-ture medium. Certain deficiencies (for example,in mesoinositol) also result in an increase in theproportion of glucan compared with mannopro-teins. The cell walls of older cells are richer inglucans and in chitin and less furnished in manno-proteins. For this reason, they are more resistantto physical and enzymatic agents used to degradethem. Finally, the composition of the cell wall isprofoundly modified by morphogenetic alterations(conjugation and sporulation).

1.3 THE PLASMIC MEMBRANE

1.3.1 Chemical Compositionand Organization

The plasmic membrane is a highly selective barriercontrolling exchanges between the living cell andits external environment. This organelle is essentialto the life of the yeast.

Like all biological membranes, the yeast plasmicmembrane is principally made up of lipids andproteins. The plasmic membrane of S. cerevisiaecontains about 40% lipids and 50% proteins.Glucans and mannans are only present in smallquantities (several per cent).

The lipids of the membrane are essentiallyphospholipids and sterols. They are amphiphilicmolecules, i.e. possessing a hydrophilic and ahydrophobic part.

The three principal phospholipids (Figure 1.5)of the plasmic membrane of yeast are phos-phatidylethanolamine (PE), phosphatidylcholine(PC) and phosphatidylinositol (PI) which repre-sent 70–85% of the total. Phosphatidylserine (PS)and diphosphatidylglycerol or cardiolipin (PG) areless prevalent. Free fatty acids and phosphatidicacid are frequently reported in plasmic membraneanalysis. They are probably extraction artifactscaused by the activity of certain lipid degradationenzymes.

The fatty acids of the membrane phospholipidscontain an even number (14 to 24) of carbon atoms.The most abundant are C16 and C18 acids. Theycan be saturated, such as palmitic acid (C16) andstearic acid (C18), or unsaturated, as with oleic

acid (C18, double bond in position 9), linoleic acid(C18, two double bonds in positions 9 and 12) andlinolenic acid (C18, three double bonds in positions9, 12 and 15). All membrane phospholipids sharea common characteristic: they possess a polar orhydrophilic part made up of a phosphorylatedalcohol and a non-polar or hydrophobic partcomprising two more or less parallel fatty acidchains (Figure 1.6). In an aqueous medium, thephospholipids spontaneously form bimolecularfilms or a lipid bilayer because of their amphiphiliccharacteristic (Figure 1.6). The lipid bilayers arecooperative but non-covalent structures. Theyare maintained in place by mutually reinforcedinteractions: hydrophobic interactions, van derWaals attractive forces between the hydrocarbontails, hydrostatic interactions and hydrogen bondsbetween the polar heads and water molecules.The examination of cross-sections of yeastplasmic membrane under the electron microscopereveals a classic lipid bilayer structure with athickness of about 7.5 nm. The membrane surfaceappears sculped with creases, especially duringthe stationary phase. However, the physiologicalmeaning of this anatomic character remainsunknown. The plasmic membrane also has anunderlying depression on the bud scar.

Ergosterol is the primary sterol of the yeast plas-mic membrane. In lesser quantities, 24 (28) dehy-droergosterol and zymosterol also exist (Figure1.7). Sterols are exclusively produced in the mito-chondria during the yeast log phase. As with phos-pholipids, membrane sterols are amphipathic. Thehydrophilic part is made up of hydroxyl groupsin C-3. The rest of the molecule is hydrophobic,especially the flexible hydrocarbon tail.

The plasmic membrane also contains numerousproteins or glycoproteins presenting a wide rangeof molecular weights (from 10 000 to 120 000).The available information indicates that the orga-nization of the plasmic membrane of a yeast cellresembles the fluid mosaic model. This model,proposed for biological membranes by Singer andNicolson (1972), consists of two-dimensional solu-tions of proteins and oriented lipids. Certain pro-teins are embedded in the membrane; they arecalled integral proteins (Figure 1.6). They interact


R' C O

O

CH

H2C O P

O

O−

O CH2 CH2 NH3+

Phosphatidyl ethanolamine

R C

O

O

R' C

O

O

CH2

CH

H2C O P

O

O−

O CH2 C

H

COO−

NH3+

Phosphatidyl serine

OHOH

H H

O

H

OHH

H

HO

OH H

P

O

O

O−

CH2

HC

H2C

O

O C

C

O

O

R'

R

Phosphatidyl inositol

R C O

O

CH2

CHOCR'

O H2C O P

O

O−

O CH2 CH2 N+(CH3)3

Phosphatidyl choline

R C

O

O CH2

CHOCR'

O H2C O P

O

O−

O CH2 C CH2 O P

O

O−

O CH2

HC O

H2C O

C

C R

O

R'

O

Diphosphatidyl glycerol (cardiolipin)

R C

O

O CH2

Fig. 1.5. Yeast membrane phospholipids

strongly with the non-polar part of the lipid bilayer.The peripheral proteins are linked to the precedentby hydrogen bonds. Their location is asymmetrical,at either the inner or the outer side of the plasmicmembrane. The molecules of proteins and mem-brane lipids, constantly in lateral movement, arecapable of rapidly diffusing in the membrane.

Some of the yeast membrane proteins have beenstudied in greater depth. These include adenosinetriphosphatase (ATPase), solute (sugars and amino

acids) transport proteins, and enzymes involved inthe production of glucans and chitin of the cellwall.

The yeast possesses three ATPases: in the mito-chondria, the vacuole, and the plasmic membrane.The plasmic membrane ATPase is an integral pro-tein with a molecular weight of around 100 Da. Itcatalyzes the hydrolysis of ATP which furnishesthe necessary energy for the active transport ofsolutes across the membrane. (Note: an active


Polar head: phosphorylated alcohol

Hydrocarbon tails: fatty acid chains

a

b

Fig. 1.6. A membrane lipid bilayer. The integralproteins (a) are strongly associated to the non-polarregion of the bilayer. The peripheral proteins (b) arelinked to the integral proteins

transport moves a compound against the concen-tration gradient.) Simultaneously, the hydrolysis ofATP creates an efflux of protons towards the exte-rior of the cell.

The penetration of amino acids and sugarsinto the yeast activates membrane transport sys-tems called permeases. The general amino acid

permease (GAP) contains three membrane proteinsand ensures the transport of a number of neutralamino acids. The cultivation of yeasts in the pres-ence of an easily assimilated nitrogen-based nutri-ent such as ammonium represses this permease.

The membrane composition in fatty acids andits proportion in sterols control its fluidity. Thehydrocarbon chains of fatty acids of the membranephospholipid bilayer can be in a rigid and orderlystate or in a relatively disorderly and fluid state. Inthe rigid state, some or all of the carbon bondsof the fatty acids are trans. In the fluid state,some of the bonds become cis. The transitionfrom the rigid state to the fluid state takes placewhen the temperature rises beyond the fusiontemperature. This transition temperature dependson the length of the fatty acid chains and theirdegree of unsaturation. The rectilinear hydrocarbonchains of the saturated fatty acids interact strongly.These interactions intensify with their length. Thetransition temperature therefore increases as thefatty acid chains become longer. The doublebonds of the unsaturated fatty acids are generallycis, giving a curvature to the hydrocarbon chain(Figure 1.8). This curvature breaks the orderly

H3C

CH3

CH3

CH3

H3C

HO

H3C

H3C

CH3

CH3

CH2

H3C

HO

H3C

H3C

CH3

CH3

H3C

HO

H3C

H

Ergosterol (24) (28) Dehydroergosterol

Zymosterol

Fig. 1.7. Principal yeast membrane sterols


Stearic acid (C18, saturated)

Oleic acid (C18, unsaturated)

Fig. 1.8. Molecular models representing the three-di-mensional structure of stearic and oleic acid. The cisconfiguration of the double bond of oleic acid producesa curvature of the carbon chain

stacking of the fatty acid chains and lowers thetransition temperature. Like cholesterol in the cellsof mammals, ergosterol is also a fundamentalregulator of the membrane fluidity in yeasts.Ergosterol is inserted in the bilayer perpendicularlyto the membrane. Its hydroxyl group joins, byhydrogen bonds, with the polar head of thephospholipid and its hydrocarbon tail is insertedin the hydrophobic region of the bilayer. Themembrane sterols intercalate themselves betweenthe phospholipids. In this manner, they inhibitthe crystallization of the fatty acid chains at lowtemperatures. Inversely, in reducing the movementof these same chains by steric encumberment, theyregulate an excess of membrane fluidity when thetemperature rises.

1.3.2 Functions of the PlasmicMembrane

The plasmic membrane constitutes a stable,hydrophobic barrier between the cytoplasm andthe environment outside the cell, owing to its

phospholipids and sterols. This barrier presents acertain impermeability to solutes in function ofosmotic properties.

Furthermore, through its system of permeases,the plasmic membrane also controls the exchangesbetween the cell and the medium. The function-ing of these transport proteins is greatly influencedby its lipid composition, which affects membranefluidity. In a defined environmental model, thesupplementing of membrane phospholipids withunsaturated fatty acids (oleic and linoleic) pro-moted the penetration and accumulation of certainamino acids as well as the expression of the gen-eral amino acid permease (GAP), (Henschke andRose, 1991). On the other hand, membrane sterolsseem to have less influence on the transport ofamino acids than the degree of unsaturation ofthe phospholipids. The production of unsaturatedfatty acids is an oxidative process and requires theaeration of the culture medium at the beginningof alcoholic fermentation. In semi-anaerobic wine-making conditions, the amount of unsaturated fattyacids in the grape, or in the grape must, probablyfavor the membrane transport mechanisms of fattyacids.

The transport systems of sugars across the mem-brane are far from being completely elucidated.There exists, however, at least two kinds of trans-port systems: a high affinity and a low affinitysystem (ten times less important) (Bisson, 1991).The low affinity system is essential during the logphase and its activity decreases during the station-ary phase. The high affinity system is, on the con-trary, repressed by high concentrations of glucose,as in the case of grape must (Salmon et al., 1993)(Figure 1.9). The amount of sterols in the mem-brane, especially ergosterol, as well as the degreeof unsaturation of the membrane phospholipidsfavor the penetration of glucose in the cell. Thisis especially true during the stationary and declinephases. This phenomenon explains the determininginfluence of aeration on the successful completionof alcoholic fermentation during the yeast multi-plication phase.

The presence of ethanol, in a culture medium,slows the penetration speed of arginine and glucoseinto the cell and limits the efflux of protons


00

0

0

0

0.0 0.1 0.2 0.3 0.4 0.5 0.60

1

2

3

4

5

6

high affinity transport system activity

Length of the fermentation as a decimal of total time

Glu

cose

pen

etra

tion

spee

d (m

mol

/h/g

dry

wei

ght) low affinity

transport system activity

Fig. 1.9. Evolution of glucose transport system activityof S. cerevisiae fermenting a medium model (Salmonet al., 1993). LF = Length of the fermentation as adecimal of total time GP = Glucose penetration speed(mmol/h/g of dry weight) 0 = Low affinity transportsystem activity ∗ = High affinity transport systemactivity

resulting from membrane ATPase activity (Alexan-dre et al., 1994; Charpentier, 1995). Simulta-neously, the presence of ethanol increases thesynthesis of membrane phospholipids and theirpercentage in unsaturated fatty acids (especiallyoleic). Temperature and ethanol act in synergy toaffect membrane ATPase activity. The amount ofethanol required to slow the proton efflux decreasesas the temperature rises. However, this modifica-tion of membrane ATPase activity by ethanol maynot be the origin of the decrease in plasmic mem-brane permeability in an alcoholic medium. Therole of membrane ATPase in yeast resistance toethanol has not been clearly demonstrated.

The plasmic membrane also produces cellwall glucan and chitin. Two membrane enzymesare involved: β-1,3 glucanase and chitin syn-thetase. These two enzymes catalyze the poly-merization of glucose and N -acetyl-glucosamine,derived from their activated forms (uridinediphosphates—UDP). The mannoproteins areessentially produced in the endoplasmic reticulum

(Section 1.4.2). They are then transported by vesi-cles which fuse with the plasmic membraneand deposit their contents at the exterior of themembrane.

Finally, certain membrane proteins act as cel-lular specific receptors. They permit the yeast toreact to various external stimuli such as sexual hor-mones or changes in the concentration of externalnutrients. The activation of these membrane pro-teins triggers the liberation of compounds such ascyclic adenosine monophosphate (cAMP) in thecytoplasm. These compounds serve as secondarymessengers which set off other intercellular reac-tions. The consequences of these cellular mecha-nisms in the alcoholic fermentation process meritfurther study.

1.4 THE CYTOPLASM AND ITSORGANELLES

Between the plasmic membrane and the nuclearmembrane, the cytoplasm contains a basiccytoplasmic substance, or cytosol. The organelles(endoplasmic reticulum, Golgi apparatus, vacuoleand mitochondria) are isolated from the cytosol bymembranes.

1.4.1 Cytosol

The cytosol is a buffered solution, with a pHbetween 5 and 6, containing soluble enzymes,glycogen and ribosomes.

Glycolysis and alcoholic fermentation enzymes(Chapter 2) as well as trehalase (an enzyme cat-alyzing the hydrolysis of trehalose) are present.Trehalose, a reserve disaccharide, also cytoplas-mic, ensures yeast viability during the dehydrationand rehydration phases by maintaining membraneintegrity.

The lag phase precedes the log phase in asugar-containing medium. It is marked by a rapiddegradation of trehalose linked to an increase intrehalase activity. This activity is itself closelyrelated to an increase in the amount of cAMP inthe cytoplasm. This compound is produced by amembrane enzyme, adenylate cyclase, in response


to the stimulation of a membrane receptor by anenvironmental factor.

Glycogen is the principal yeast glucidic reservesubstance. Animal glycogen is similar in structure.It accumulates during the stationary phase in theform of spherical granules of about 40 µm indiameter.

When observed under the electron microscope,the yeast cytoplasm appears rich in ribosomes.These tiny granulations, made up of ribonucleicacids and proteins, are the center of proteinsynthesis. Joined to polysomes, several ribosomesmigrate the length of the messenger RNA. Theytranslate it simultaneously so that each oneproduces a complete polypeptide chain.

1.4.2 The Endoplasmic Reticulum,the Golgi Apparatusand the Vacuoles

The endoplasmic reticulum (ER) is a doublemembrane system partitioning the cytoplasm. It islinked to the cytoplasmic membrane and nuclearmembrane. It is, in a way, an extension of thelatter. Although less developed in yeasts than inexocrine cells of higher eucaryotes, the ER hasthe same function. It ensures the addressing ofthe proteins synthesized by the attached ribosomes.As a matter of fact, ribosomes can be either freein the cytosol or bound to the ER. The pro-teins synthesized by free ribosomes remain in thecytosol, as do the enzymes involved in glycolysis.Those produced in the ribosomes bound to the ERhave three possible destinations: the vacuole, theplasmic membrane, and the external environment(secretion). The presence of a signal sequence (aparticular chain of amino acids) at the N -terminalextremity of the newly formed protein determinesthe association of the initially free ribosomes inthe cytosol with the ER. The synthesized proteincrosses the ER membrane by an active transportprocess called translocation. This process requiresthe hydrolysis of an ATP molecule. Having reachedthe inner space of the ER, the proteins undergo cer-tain modifications including the necessary excisingof the signal peptide by the signal peptidase. Inmany cases, they also undergo a glycosylation.

The yeast glycoproteins, in particular the struc-tural, parietal or enzymatic mannoproteins, con-tain glucidic side chains (Section 1.2.2). Some ofthese are linked to asparagine by N -glycosidicbonds. This oligosaccharidic link is constructed inthe interior of the ER by the sequential additionof activated sugars (in the form of UDP deriva-tives) to a hydrophobic, lipidic transporter calleddolicholphosphate. The entire unit is transferred inone piece to an asparagine residue of the polypep-tide chain. The dolicholphosphate is regenerated.

The Golgi apparatus consists of a stack ofmembrane sacs and associated vesicles. It is anextension of the ER. Transfer vesicles transportthe proteins issued from the ER to the sacs of theGolgi apparatus. The Golgi apparatus has a dualfunction. It is responsible for the glycosylationof protein, then sorts so as to direct them viaspecialized vesicles either into the vacuole or intothe plasmic membrane. An N-terminal peptidicsequence determines the directing of proteinstowards the vacuole. This sequence is present inthe precursors of two vacuolar-orientated enzymesin the yeast: Y carboxypeptidase and A proteinase.The vesicles that transport the proteins of theplasmic membrane or the secretion granules, suchas those that transport the periplasmic invertase,are still the default destinations.

The vacuole is a spherical organelle, 0.3 to3 µm in diameter, surrounded by a single mem-brane. Depending on the stage of the cellularcycle, yeasts have one or several vacuoles. Beforebudding, a large vacuole splits into small vesi-cles. Some penetrate into the bud. Others gatherat the opposite extremity of the cell and fuseto form one or two large vacuoles. The vacuo-lar membrane or tonoplast has the same generalstructure (fluid mosaic) as the plasmic membranebut it is more elastic and its chemical com-position is somewhat different. It is less richin sterols and contains less protein and glyco-protein but more phospholipids with a higherdegree of unsaturation. The vacuole stocks someof the cell hydrolases, in particular Y carboxypep-tidase, A and B proteases, I aminopeptidase,X-propyl-dipeptidylaminopeptidase and alkalinephosphatase. In this respect, the yeast vacuole can


be compared to an animal cell lysosome. Vacuolarproteases play an essential role in the turn-overof cellular proteins. In addition, the A proteaseis indispensable in the maturation of other vacuo-lar hydrolases. It excises a small peptide sequenceand thus converts precursor forms (proenzymes)into active enzymes. The vacuolar proteases alsoautolyze the cell after its death. Autolysis, whileageing white wine on its lees, can affect wine qual-ity and should concern the winemaker.

Vacuoles also have a second principal function:they stock metabolites before their use. In fact,they contain a quarter of the pool of the aminoacids of the cell, including a lot of arginine as wellas S-adenosyl methionine. In this organelle, thereis also potassium, adenine, isoguanine, uric acidand polyphosphate crystals. These are involvedin the fixation of basic amino acids. Specificpermeases ensure the transport of these metabolitesacross the vacuolar membrane. An ATPase linkedto the tonoplast furnishes the necessary energyfor the movement of stocked compounds againstthe concentration gradient. It is different from theplasmic membrane ATPase, but also produces aproton efflux.

The ER, Golgi apparatus and vacuoles canbe considered as different components of aninternal system of membranes, called the vacuome,participating in the flux of glycoproteins to beexcreted or stocked.

1.4.3 The MitochondriaDistributed in the periphery of the cytoplasm, themitochondria (mt) are spherically or rod-shapedorganelles surrounded by two membranes. Theinner membrane is highly folded to form cristae.The general organization of mitochondria is thesame as in higher plants and animal cells. Themembranes delimit two compartments: the innermembrane space and the matrix. The mitochon-dria are true respiratory organelles for yeasts. Inaerobiosis, the S. cerevisiae cell contains about50 mitochondria. In anaerobiosis, these organellesdegenerate, their inner surface decreases, and thecristae disappear. Ergosterol and unsaturated fattyacids supplemented in culture media limit thedegeneration of mitochondria in anaerobiosis. In

any case, when cells formed in anaerobiosis areplaced in aerobiosis, the mitochondria regain theirnormal appearance. Even in aerated grape must,the high sugar concentration represses the synthe-sis of respiratory enzymes. As a result, the mito-chondria no longer function. This phenomenon,catabolic glucose repression, will be described inChapter 2.

The mitochondrial membranes are rich in phos-pholipids—principally phosphatidylcholine, phos-phatidylinositol and phosphatidylethanolamine(Figure 1.5). Cardiolipin (diphosphatidylglycerol),in minority in the plasmic membrane (Figure 1.4),is predominant in the inner mitochondrial mem-brane. The fatty acids of the mitochondrial phos-pholipids are in C16:0, C16:1, C18:0, C18:1.In aerobiosis, the unsaturated residues predomi-nate. When the cells are grown in anaerobiosis,without lipid supplements, the short-chain satu-rated residues become predominant; cardiolipinand phosphatidylethanolamine diminish whereasthe proportion of phosphatidylinositol increases. Inaerobiosis, the temperature during the log phase ofthe cell influences the degree of unsaturation of thephospholipids- more saturated as the temperaturedecreases.

The mitochondrial membranes also containsterols, as well as numerous proteins and enzymes(Guerin, 1991). The two membranes, inner andouter, contain enzymes involved in the synthesis ofphospholipids and sterols. The ability to synthesizesignificant amounts of lipids, characteristic of yeastmitochondria, is not limited by respiratory deficientmutations or catabolic glucose repression.

The outer membrane is permeable to mostsmall metabolites coming from the cytosol since itcontains porine, a 29 kDa transmembrane proteinpossessing a large pore. Porine is present inthe mitochondria of all the eucaryotes as wellas in the outer membrane of bacteria. Theintermembrane space contains adenylate kinase,which ensures interconversion of ATP, ADP andAMP. Oxidative phosphorylation takes place in theinner mitochondrial membrane. The matrix, on theother hand, is the center of the reactions of thetricarboxylic acids cycle and of the oxidation offatty acids.


The majority of mitochondria proteins are codedby the genes of the nucleus and are synthesized bythe free polysomes of the cytoplasm. The mito-chondria, however, also have their own machineryfor protein synthesis. In fact, each mitochon-drion possesses a circular 75 kb (kilobase pairs)molecule of double-stranded AND and ribosomes.The mtDNA is extremely rich in A (adenine) andT (thymine) bases. It contains a few dozen genes,which code in particular for the synthesis of cer-tain pigments and respiratory enzymes, such ascytochrome b, and several sub-units of cytochromeoxidase and of the ATP synthetase complex. Somemutations affecting these genes can result in theyeast becoming resistant to certain mitochondrialspecific inhibitors such as oligomycin. This prop-erty has been applied in the genetic marking ofwine yeast strains. Some mitochondrial mutantsare respiratory deficient and form small colonieson solid agar media. These ‘petit’ mutants are notused in winemaking because it is impossible toproduce them industrially by respiration.

1.5 THE NUCLEUS

The yeast nucleus is spherical. It has a diameterof 1–2 mm and is barely visible using a phasecontrast optical microscope. It is located near theprincipal vacuole in non-proliferating cells. Thenuclear envelope is made up of a double membraneattached to the ER. It contains many ephemeralpores, their locations continually changing. Thesepores permit the exchange of small proteinsbetween the nucleus and the cytoplasm. Contraryto what happens in higher eucaryotes, the yeastnuclear envelope is not dispersed during mitosis.In the basophilic part of the nucleus, the crescent-shaped nucleolus can be seen by using a nuclear-specific staining method. As in other eucaryotes, itis responsible for the synthesis of ribosomal RNA.During cellular division, the yeast nucleus alsocontains rudimentary spindle threads composed ofmicrotubules of tubulin, some discontinuous andothers continuous (Figure 1.10). The continuousmicrotubules are stretched between the twospindle pole bodies (SPB). These corpuscles arepermanently included in the nuclear membrane and

Discontinous tubules

Continuous tubules

Nucleolus

Cytoplasmic microtubules

Chromatin

Pore

Spindle pole body

Fig. 1.10. The yeast nucleus (Williamson, 1991). SPB =Spindle pole body; NUC = Nucleolus; P = Pore; CHR =Chromatin; CT = Continuous tubules; DCT = Discon-tinuous tubules; CTM = Cytoplasmic microtubules

correspond with the centrioles of higher organisms.The cytoplasmic microtubules depart from thespindle pole bodies towards the cytoplasm.

There is little nuclear DNA in yeasts comparedwith higher eucaryotes—about 14 000 kb in ahaploid strain. It has a genome almost three timeslarger than in Escherichia coli, but its geneticmaterial is organized into true chromosomes. Eachone contains a single molecule of linear double-stranded DNA associated with basic proteinsknown as histones. The histones form chromatinwhich contains repetitive units called nucleosomes.Yeast chromosomes are too small to be observedunder the microscope.

Pulse-field electrophoresis (Carle and Olson,1984; Schwartz and Cantor, 1984) permits the sep-aration of the 16 chromosomes in S. cerevisiae,whose size range from 200 to 2000 kb. Thisspecies has a very large chromosomic polymor-phism. This characteristic has made karyotypeanalysis one of the principal criteria for the iden-tification of S. cerevisiae strains (Section 1.9.3).The scientific community has nearly establishedthe complete sequence of the chromosomic DNAof S. cerevisiae. In the future, this detailed knowl-edge of the yeast genome will constitute a powerfultool, as much for understanding its molecular phys-iology as for selecting and improving winemakingstrains.

The yeast chromosomes contain relatively fewrepeated sequences. Most genes are only present


in a single example in the haploid genome, but theribosomal RNA genes are highly repeated (about100 copies).

The genome of S. cerevisiae contains transpos-able elements, or transposons—specifically, Ty(transposon yeast) elements. These comprise a cen-tral ε region (5.6 kb) framed by a direct repeatedsequence called the δ sequence (0.25 kb). The δ

sequences have a tendency to recombine, resultingin the loss of the central region and a δ sequence.As a result, there are about 100 copies of the δ

sequence in the yeast genome. The Ty elementscode for non-infectious retrovirus particles. Thisretrovirus contains Ty messenger RNA as well asa reverse transcriptase capable of copying the RNAinto complementary DNA. The latter can reinsertitself into any site of the chromosome. The ran-dom excision and insertion of Ty elements in theyeast genome can modify the genes and play animportant role in strain evolution.

Only one plasmid, called the 2 µm plasmid, hasbeen identified in the yeast nucleus. It is a circularmolecule of DNA, containing 6 kb and there are50–100 copies per cell. Its biological function isnot known, but it is a very useful tool, used bymolecular biologists to construct artificial plasmidsand genetically transform yeast strains.

1.6 REPRODUCTION AND THEYEAST BIOLOGICAL CYCLE

Like other sporiferous yeasts belonging to theclass Ascomycetes, S. cerevisiae can multiplyeither asexually by vegetative multiplication orsexually by forming ascospores. By definition,yeasts belonging to the imperfect fungi can onlyreproduce by vegetative multiplication.

1.6.1 Vegetative MultiplicationMost yeasts undergo vegetative multiplication bya process called budding. Some yeasts, such asspecies belonging to the genus Schizosaccha-romyces, multiply by binary fission.

Figure 1.11 represents the life cycle of S. cerevi-siae divided into four phases: M, G1, S, and G2.M corresponds with mitosis, G1 is the period

G1

G2

SM

Fig. 1.11. S. cerevisiae cell cycle (vegetative mul-tiplication) (Tuite and Oliver, 1991). M = mitosis;G1 = period preceding DNA synthesis; S = DNA syn-thesis; G2 = period preceding mitosis

preceding S, which is the synthesis of DNA and G2is the period before cell division. As soon as thebud emerges, in the beginning of S, the splittingof the spindle pole bodies (SPB) can be observedin the nuclear membrane by electron microscopy.At the same time, the cytoplasmic microtubulesorient themselves toward the emerging bud. Thesemicrotubules seem to guide numerous vesicleswhich appear in the budding zone and are involvedin the reshaping of the cell wall. As the budgrows larger, discontinued nuclear microtubulesbegin to appear. The longest microtubules formthe mitotic spindle between the two SPB. At theend of G2, the nucleus begins to push and pullapart in order to penetrate the bud. Some of themitochondria also pass with some small vacuolesinto the bud, whereas a large vacuole is formedat the other pole of the cell. The expansion ofthe latter seems to push the nucleus into thebud. During mitosis, the nucleus stretches to itsmaximum and the mother cell separates from thedaughter cell. This separation takes place only afterthe construction of the separation cell wall and


the deposit of a ring of chitin on the bud scar ofthe mother cell. The movement of chromosomesduring mitosis is difficult to observe in yeasts,but a microtubule–centromere link must guidethe chromosomes. In grape must, the duration ofbudding is approximately 1–2 hours. As a result,the population of the cells double during the yeastlog phase during fermentation.

1.6.2 Sexual Reproduction

When sporiferous yeast diploid cells are in ahostile nutritive medium (for example, depletedof fermentable sugar, poor in nitrogen and veryaerated) they stop multiplying. Some transforminto a kind of sac with a thick cell wall. Thesesacs are called asci. Each one contains four haploidascospores issued from meiotic division of thenucleus. Grape must and wine are not propitiousto yeast sporulation and, in principal, it neveroccurs in this medium. Yet Mortimer et al. (1994)observed the sporulation of certain wine yeaststrains, even in sugar-rich media. Our researchershave often observed asci in old agar culture mediastored for several weeks in the refrigerator or atambient temperatures (Figure 1.12). The naturalconditions in which wild wine yeasts sporulate andthe frequency of this phenomenon are not known.In the laboratory, the agar or liquid medium

Fig. 1.12. Scanning electron microscope photograph ofS. cerevisiae cells placed on a sugar-agar mediumfor several weeks. Asci containing ascospores can beobserved

conventionally used to provoke sporulation hasa sodium acetate base (1%). In S. cerevisiae,sporulation aptitude varies greatly from strain tostrain. Wine yeasts, both wild and selected, donot sporulate easily, and when they do they oftenproduce non-viable spores.

Meiosis in yeasts and in higher eucaryotes(Figure 1.13) has some similarities. Several hoursafter the transfer of diploid vegetative cells toa sporulation medium, the SPB splits during theDNA replication of the S phase. A dense body(DB) appears simultaneously in the nucleus nearthe nucleolus. The DB evolves into synaptonemalcomplexes—structures permitting the coupling andrecombination of homologous chromosomes. After8–9 hours in the sporulation medium, the two SPBseparate and the spindle begins to form. This stageis called metaphase I of meiosis. At this stage, thechromosomes are not yet visible. Then, while thenuclear membrane remains intact, the SPB divides.At metaphase II, a second mitotic spindle stretchesitself while the ascospore cell wall begins to form.Nuclear buds, cytoplasm and organelles migrateinto the ascospores. At this point, edification of thecell wall is completed. The spindle then disappearswhen the division is achieved.

Placed in favorable conditions, i.e. nutritivesugar-enriched media, the ascospores germinate,breaking the cell wall of the ascus, and begin tomultiply. In S. cerevisiae, the haploid cells havetwo mating types: a and α. The ascus contains twoa ascospores and two α ascospores (Figure 1.14).Sign a (MATa) cells produce a sexual pheromonea. This peptide made up of 12 amino acids iscalled sexual factor a. In the same manner, sign α

cells produce the sexual factor α, a peptide madeup of 13 amino acids. The a factor, emitted bythe MATa cells, stops the multiplication of MATα

cells in G1. Reciprocally, the α factor producedby α cells stops the biological cycle of a cells.Sexual coupling occurs between two cells of theopposite sexual sign. Their agglutination permitscellular and nuclear fusion and makes use ofparietal glycoproteins and a and α agglutinins.The vegetative diploid cell that is formed (a/α)can no longer produce sexual pheromones and isinsensitive to their action; it multiplies by budding.


(a) (b) (c)

(e) (f)

Spindle pole body

Dense bodySynaptonemalcomplexes

(d)

(g) (h)

Fig. 1.13. Meiosis in S. cerevisiae (Tuite and Oliver, 1991). SPB = spindle pole body; DB = dense body;SC = synaptonemal complexes. (a) Cell before meiosis; (b) dividing of SPB; (c) synaptonemal complexes appear;(d) separation of the SPB; (e) constitution of spindle (metaphase I of meiosis); (f) dividing of the SPB; (g) metaphase.II of meiosis; (h) end of meiosis; formation of ascospores

Ascospores

Ascus

Type a haploid cells

Conjugation(sexual coupling)

Zygote

Vegetative diploid cell

Type a haploid cells

α

a/α

a/α

α

αα

αα

α αα

α

·aa

aa

aa

a

a

Fig. 1.14. Reproduction cycle of a heterothallic yeaststrain (a, α: spore sexual signs)

Some strains, from a monosporic culture, can bemaintained in a stable haploid state. Their sexualsign remains constant during many generations.They are heterothallic. Others change sexual sign

during a cellular division. Diploid cells appearin the descendants of an ascospore. They arehomothallic and have an HO gene which inversessexual sign at an elevated frequency duringvegetative division. This changeover (Figure 1.15)occurs in the mother cell at the G1 stage of the

S

S F1

S F2 F1 F1.1

α

α

α∗

αα∗aa*

αα∗ aa* αα∗aa*

Fig. 1.15. Sexual sign commutation model of haploidyeast cells in a homothallic strain (Herskowitz et al.,1992) (∗ designates cells capable of changing sexualsign at the next cell division, or cells already havingundergone budding). S = initial cell carrying the HOgene; F1, F2 = daughter cells of S; F1.1. = daughtercell of F1


biological cycle, after the first budding but beforethe DNA replication phase. In this manner, a signα ascospore S divides to produce two α cells(S and the first daughter cell, F1). During thefollowing cellular division, S produces two cells(S and F2) that have become a cells. In the samemanner, the F1 cell produces two α cells after thefirst division and two a cells during its secondbudding. Laboratory strains that are deficient ormissing the HO gene have a stable sexual sign.Heterothallism can therefore be considered theresult of a mutation of the HO gene or of genes thatcontrol its functioning (Herskowitz et al., 1992).

Most wild and selected winemaking strains thatbelong to the S. cerevisiae species are diploidand homothallic. It is also true of almost all ofthe strains that have been isolated in vineyardsof the Bordeaux region. Moreover, recent studiescarried out by Mortimer et al. (1994) in Californianand Italian vineyards have shown that the majorityof strains (80%) are homozygous for the HOcharacter (HO /HO); heterozygosis (HO /ho) isin minority. Heterothallic strains (ho/ho) arerare (less than 10%). We have made the same

observations for yeast strains isolated in theBordeaux region. For example, the F10 strain fairlyprevalent in spontaneous fermentations in certainBordeaux growths is HO /HO . In other words, thefour spores issued from an ascus give monoparentdiploids, capable of forming asci when placed ina pure culture. This generalized homozygosis forthe HO character of wild winemaking strains isprobably an important factor in their evolution,according to the genome renewal phenomenonproposed by Mortimer et al. (1994) (Figure 1.16),in which the continuous multiplication of a yeaststrain in its natural environment accumulatesheterozygotic damage to the DNA. Certainslow-growth or functional loss mutations of certaingenes decrease strain vigor in the heterozygousstate. Sporulation, however, produces haploidcells containing different combinations of theseheterozygotic characters. All of these sporesbecome homozygous diploid cells with a seriesof genotypes because of the homozygosity of theHO character. Certain diploids which prove to bemore vigorous than others will in time supplantthe parents and less vigorous ones. This very

Initialpopulation

Mutation andvegetative growth

Meiosis

Homothallisme

Vegetative growth

Finalpopulation

___++

___++

+___a

+___+

+______a

b

Spores+___+

+___+

+ +

+___+

b___b

+ b

a___a

+___+a +

___ ___aa

bb

a b ___ ___++

bb

___ ___+a

b+

___ ___++

bb

+

Fig. 1.16. Genome renewal of a homozygote yeast strain for the HO gene of homothallism, having accumulatedrecessive mutations during vegetative multiplication (Mortimer et al., 1994) (a and b = mutation of certain genes)


tempting model is reaffirmed by the characteristicsof the wild winemaking strains analyzed. In these,the spore viability rate is the inverse functionof the heterozygosis rate for a certain numberof mutations. The completely homozygous strainspresent the highest spore viability and vigor.

In conclusion, sporulation of strains in naturalconditions seems indispensable. It assures theirgrowth and fermentation performance. With this inmind, the conservation of selected strains of activedry yeasts as yeast starters should be questioned. Itmay be necessary to regenerate them periodicallyto eliminate possible mutations from their genomewhich could diminish their vigor.

1.7 THE KILLER PHENOMENON

1.7.1 Introduction

Certain yeast strains, known as killer strains (K),secrete proteinic toxins into their environment thatare capable of killing other, sensitive strains (S).The killer strains are not sensitive to their toxin butcan be killed by a toxin that they do not produce.Neutral strains (N) do not produce a toxin but areresistant. The action of a killer strain on a sensitivestrain is easy to demonstrate in the laboratory on anagar culture medium at pH 4.2–4.7 at 20◦C. Thesensitive strain is inoculated into the mass of agarbefore it solidifies; then the strain to be tested isinoculated in streaks on the solidified medium. If itis a killer strain, a clear zone in which the sensitivestrain cannot grow encircles the inoculum streaks(Figure 1.17).

This phenomenon, the killer factor, was dis-covered in S. cerevisiae but killer strains alsoexist in other yeast genera such as Hansenula,Candida, Kloeckera, Hanseniaspora, Pichia, Toru-lopsis, Kluyveromyces and Debaryomyces. Killeryeasts have been classified into 11 groups accord-ing to the sensitivity reaction between strains aswell as the nature and properties of the toxinsinvolved. The killer factor is a cellular interactionmodel mediated by the proteinic toxin excreted.It has given rise to much fundamental research(Tipper and Bostian, 1984; Young, 1987). Barre(1984, 1992), Radler (1988) and Van Vuuren and

Fig. 1.17. Identification of the K2 killer phenotype inS. cerevisiae. The presence of a halo around the twostreaks of the killer strain is due to the death of thesensitive strain cultivated on the medium

Jacobs (1992) have described the technologicalimplications of this phenomenon for wine yeastsand the fermentation process.

1.7.2 Physiology and Geneticsof the Killer Phenomenon

The determinants of the killer factor are bothcytoplasmic and nuclear. In S. cerevisiae, the killerphenomenon is associated with the presence ofdouble-stranded RNA particles, virus-like particles(VLP), in the cytoplasm. They are in the samecategory as non-infectious mycovirus. There aretwo kinds of VLP: M and L. The M genome(1.3–1.9 kb) codes for the K toxin and for theimmunity factor (R). The L genome (4.5 kb) codesfor an RNA polymerase and the proteinic capsidthat encapsulates the two genomes. Killer strains(K+R+) secrete the toxin and are immune to it.The sensitive cells (K−R−) do not possess M VLPbut most of them have L VLP. The two types ofviral particles are necessary for the yeast cell toexpress the killer phenotype (K+R+), since the Lmycovirus is necessary for the maintenance of theM type.


There are three kinds of killer activities inS. cerevisiae strains. They correspond with the K1,K2 and K3 toxins coded, respectively, by M1, M2and M3 VLPs (1.9, 1.5 and 1.3 kb, respectively).According to Wingfield et al. (1990), the K2 andK3 types are very similar; M3 VLP results from amutation of M2 VLP. The K2 strains are by farthe most widespread in the S. cerevisiae strainsencountered in wine. Neutral strains (K−R+) areinsensitive to a given toxin without being capableof producing it. They possess M VLPs of normaldimensions that code only for the immunityfactor. They either do not produce toxins or areinactive because of mutations affecting the M-typeRNA.

Many chromosomic genes are involved in themaintenance and replication of L and M RNAparticles as well as in the maturation and transportof the toxin produced.

The K1 toxin is a small protein made up oftwo sub-units (9 and 9.5 kDa). It is active andstable in a very narrow pH range (4.2–4.6) andis therefore inactive in grape must. The K2 toxin,a 16 kDa glycoprotein, produced by homothallicstrains of S. cerevisiae encountered in wine, isactive at between pH 2.8 and 4.8 with a maximumactivity between 4.2 and 4.4. It is therefore activeat the pH of grape must and wine.

K1 and K2 toxins attack sensitive cells byattaching themselves to a receptor located in thecell wall—a β-1,6 glucan. Two chromosomicgenes, KRE1 and KRE2 (Killer resistant), deter-mine the possibility of this linkage. The kre1 geneproduces a parietal glycoprotein which has a β-1,6 glucan synthetase activity. The kre1 mutantsare resistant to K1 and K2 toxins because theyare deficient in this enzyme and devoid of a β-1,6glucan receptor. The KRE2 gene is also involvedin the fixation of toxins to the parietal recep-tor; the kre2 mutants are also resistant. The toxinlinked to a glucan receptor is then transferred to amembrane receptor site by a mechanism needingenergy. Cells in the log phase are, therefore, moresensitive to the killer effect than cells in the station-ary phase. When the sensitive cell plasmic mem-brane is exposed to the toxin, it manifests seriousfunctional alterations after a lag phase of about

40 minutes. These alterations include the interrup-tion of the coupled transport of amino acids andprotons, the acidification of the cellular contents,and potassium and ATP leakage. The cell dies in2–3 hours after contact with the toxin because ofthe above damage, due to the formation of poresin the plasmic membrane.

The killer effect exerts itself exclusively onyeasts and has no effect on humans and animals.

1.7.3 The Role of the KillerPhenomenon in Winemaking

Depending on the authors and viticultural regionsstudied, the frequency of the killer character variesa lot among wild winemaking strains isolated ongrapes or in fermenting grape must. In a work byBarre (1978) studying 908 wild strains, 504 man-ifested the K2 killer character, 299 were sensitiveand 95 neutral. Cuinier and Gros (1983) reported ahigh frequency (65–90%) of K2 strains in Mediter-ranean and Beaujolais region vineyards, whereasnone of the strains analyzed in Tourraine mani-fested the killer effect. In the Bordeaux region, theK2 killer character is extremely widespread. In astudy carried out in 1989 and 1990 on the ecol-ogy of indigenous strains of S. cerevisiae in severaltanks of red must in a Pessac-Leognan vineyard, allof the isolated strains manifested K2 killer activity,about 30 differentiated by their karyotype (Frezier,1992). Rossini et al. (1982) reported an extremelyvaried frequency (12–80%) of K2 killer strainsin spontaneous fermentations in Italian wineries.Some K2 killer strains were also isolated in thesouthern hemisphere (Australia, South Africa andBrazil). On the other hand, most of the killerstrains isolated in Japan presented the K1 char-acteristic. Most research on the killer character ofwine yeasts concerns the species S. cerevisiae. Lit-tle information exists on the killer effect of thealcohol-sensitive species which essentially makeup grape microflora. Heard and Fleet (1987) con-firmed Barre’s (1980) observations and did notestablish the existence of the killer effect in Can-dida, Hanseniaspora, Hansenula and Torulaspora.However, some killer strains of Hanseniasporauvarum and Pichia kluyveri have been identifiedby Zorg et al. (1988).


Barre (1992) studied the activity and stabilityof the K2 killer toxin in enological conditions(Figure 1.18). The killer toxin only manifested apronounced activity on cells in the log phase. Cellsin the stationary phase were relatively insensitive.The amount of ethanol or SO2 in the wine has

practically no effect on the killer toxin activity.On the other hand, it is quickly destroyed by heat,since its half-life is around 30 minutes at 32◦C.It is also quickly inactivated by the presence ofphenolic compounds and is easily adsorbed bybentonite.

103

104

105

1

Num

ber

of v

iabl

e ce

lls m

l−1

Num

ber

of v

iabl

e ce

lls m

l−1

Num

ber

of v

iabl

e ce

lls m

l−1

*

o*o

2

Time (hours)

*

o

3

*

o

4

o

*

103

104

105

*o

*o

*

o

o

*

1 2 3 4

103

104

105

1 2 3 4

*o

o

*o *

oo*

o

(a)

Time (hours)(c)

Time (hours)(b)

Fig. 1.18. Yeast growth and survival curves in a grape juice medium containing killer toxin (Barre, 1992): ∗, 10%K2 strain active culture supernatant; Ž, 10% supernatant inactivated by heat treatment. (a) White juice, pH 3.4; cellsin exponential phase introduced at time = 0. (b) Same juice, cells in stationary phase introduced at time = 0. (c) Redjuice extracted by heated maceration, pH 3.4; cells in exponential phase introduced at time = 0


Scientific literature has reported a diversity offindings on the role of the killer factor in the com-petition between strains during grape must fermen-tation. In an example given by Barre (1992), killercells inoculated at 2% can completely supplant thesensitive strain during the alcoholic fermentationof must. In other works, the killer yeast/sensitiveyeast ratio able to affect the implantation of sensi-tive yeasts in winemaking varies between 1/1000and 100/1, depending on the author. This con-siderable discrepancy can probably be attributedto implantation and fermentation speed of thestrains present. The killer phenomenon seems moreimportant to interstrain competition when the killerstrain implants itself quickly and the sensitivestrain slowly. In the opposite situation, an elevatedpercentage of killer yeasts would be necessary toeliminate the sensitive population. Some authorshave observed spontaneous fermentations domi-nated by sensitive strains despite a non-negligibleproportion of killer strains (2–25%). In Bordeaux,we have always observed that certain sensitivestrains implant themselves in red wine fermenta-tion, despite a strong presence of killer yeasts inthe wild microflora (for example, 522M, an activedry yeast starter). In white winemaking, the neu-tral yeast VL1 and sensitive strains such as EG8,a slow-growth strain, also successfully implantthemselves. The wild killer population does notappear to compete with a sensitive yeast starter andtherefore is not an important cause of fermentationdifficulties in real-life applications.

The high frequency of killer strains amongthe indigenous yeasts in many viticultural regionsconfers little advantage to the strain in termsof implantation capacity. In other words, thischaracter is not sufficient to guarantee the implan-tation of a certain strain during fermentation overa wild strain equally equipped. On the other hand,under certain conditions, inoculating with a sensi-tive strain will fail because of the killer effect ofa wild population. Therefore, the resistance to theK2 toxin (killer or neutral phenotype) should beincluded among the selection criteria of enologi-cal strains. The high frequency of the K2 killercharacter in indigenous wine yeasts facilitates thisstrategy.

A medium that contains the toxin exerts aselection pressure on a sensitive enological strain.Stable variants survive this selection pressure andcan be obtained in this manner (Barre, 1984).This is the most simple strategy for obtaining akiller enological strain. However, the developmentof molecular genetics and biotechnology permitsscientists to construct enological strains modifiedto contain one or several killer characters.Cytoduction can achieve these modifications.This method introduces cytoplasmic determinants(mitochondria, plasmids) issued from a killer straininto a sensitive enological strain without alteringthe karyotype of the initial enological strain.Seki et al. (1985) used this method to makethe 522M strain K2 killer. By another strategy,new yeasts can be constructed by integrating thetoxin gene into their chromosomes. Boone et al.(1990) were able to introduce the K1 characterinto K2 winemaking strains in this manner.The K1 killer character among wine yeasts israre, and so the enological interest of this lastapplication is limited. The acquiring of multi-killer character strains presents little enologicaladvantage. Sensitive selected strains and currentK2 killer strains can already be implanted withouta problem. On the other hand, the disseminationof these newly obtained multi-killer strains innature could present a non-negligible risk. Thesestrains could adversely affect the natural microflorapopulation, although we have barely begun toinventory its diversity and exploit its technologicalpotentials. It would be detrimental to be no longerable to select wild yeasts because they havebeen supplanted in their natural environment bygenetically modified strains—a transformation thathas no enological interest.

1.8 CLASSIFICATION OF YEASTSPECIES

1.8.1 General RemarksAs mentioned in the introduction to this chapter,yeasts constitute a vast group of unicellularfungi—taxonomically heterogeneous and verycomplex. Hansen’s first classification at the


beginning of this century only distinguishedbetween sporiferous and asporiferous yeasts. Sincethen, yeast taxonomy has incited considerableresearch. This research has been regroupedin successive works progressively creating theclassification known today. The last enologicaltreaty of the University of Bordeaux (Ribereau-Gayon et al., 1975) was based on Lodder’s (1970)classification. Between this monograph and theprevious classification (Lodder and Kregger-VanRij, 1952), the designation and classification ofyeasts had already changed profoundly. In thisbook, the last two classifications by Kregger-VanRij (1984) and Barnett et al. (2000) are of interest.These contain even more significant changes in thedelimitation of species and genus with respect toearlier systematics.

According to the current classification, yeastsbelonging to Ascomycetes, Basidiomycetes andimperfect fungi (Deuteromycetes) are divided into81 genera, to which 590 species belong. Takinginto account synonymy and physiological races(varieties of the same species), at least 4000 namesfor yeasts have been used since the nineteenth cen-tury. Fortunately, only 15 yeast species exist ongrapes, are involved as an alcoholic fermentationagent in wine, and are responsible for wine dis-eases. Table 1.1 lists the two families to which eno-logical yeasts belong: Saccharomycetaceae in theAscomycetes (sporiferous) and Cryptococcaceaein the Deuteromycetes (asporiferous). Fourteen

genera to which one or several species of grapeor wine yeasts belong are not listed in Table 1.1.

1.8.2 Evolution of the GeneralPrinciples of Yeast Taxonomyand Species Delimitation

Yeast taxonomy (from the Greek taxis: putting inorder), and the taxonomy of other microorganismsfor that matter, includes classification andidentification. Classification groups organisms intotaxa according to their similarities and/or theirties to a common ancestor. The basic taxon isspecies. A species can be defined as a collection ofstrains having a certain number of morphological,physiological and genetic characters in common.This group of characters constitutes the standarddescription of the species. Identification comparesan unknown organism to individuals alreadyclassed and named that have similar characteristics.

Taxonomists first delimited yeast species usingmorphological and physiological criteria. The firstclassifications were based on the phenotypic dif-ferences between yeasts: cell shape and size, sporeformation, cultural characters, fermentation andassimilation of different sugars, assimilation ofnitrates, growth-factor needs, resistance to cyclo-heximide. The treaty on enology by Ribereau-Gayon et al. (1975) described the use of thesemethods on wine yeasts in detail. Since then,many rapid, ready-to-use diagnostic kits have been

Table 1.1. Classification of grape and wine sporogeneous and asporogeneous yeast genere (Kregger-Van Rij, 1984)

Saccharomycetaceae family Spermophtoracae Cryptococcaceae(sporogeneous) family family

(asporogeneous) (asporogeneous)

Sub-family Sub-family Sub-family —Schizosaccharomycetoideae Nadsonioideae Saccharomycetoideae

Genus Genus Genus Genus GenusSchizosaccharomyces Saccharomycodes Saccharomyces Metschnikowia Brettanomyces

Hanseniaspora Debaryomyces CandidaDekker KloeckeraHansenula RhodotorulaKluyveromycesPichiaZygosaccharomycesTorulaspora


proposed to determine yeast response to differentphysiological tests. Lafon-Lafourcade and Joyeux(1979) and Cuinier and Leveau (1979) designedthe API 20 C system for the identification of eno-logical yeasts. It contains eight fermentation tests,10 assimilation tests and a cycloheximide resis-tance test. For a more complete identification, theAPI 50 CH system contains 50 substrates for fer-mentation (under paraffin) and assimilation tests.Heard and Fleet (1990) developed a system thatuses the different tests listed in the work of Barnettet al. (1990).

Due to the relatively limited number of yeastspecies significantly present on grapes and in wine,these phenotypic tests identify enological yeastspecies in certain genera without difficulty. Certainspecies can be identified by observing growingcells under the microscope. Small apiculatedcells, having small lemon-like shapes, designatethe species Hanseniaspora uvarum and itsimperfect form Kloeckera apiculata (Figure 1.19).Saccharomycodes ludwigii is characterized bymuch larger (10–20 µm) apiculated cells. Sincemost yeasts multiply by budding, the genusSchizosaccharomyces can be recognized becauseof its vegetative reproduction by binary fission(Figure 1.20). In modern taxonomy, this genusonly contains the species Schizosacch. pombe.Finally, the budding of Candida stellata (formerlyknown as Torulopsis stellata) occurs in the shapeof a star.

According to Barnett et al. (1990), the physio-logical characteristics listed in Table 1.2 can beused to distinguish between the principal grapeand wine yeasts. Yet some of these characters(for example, fermentation profiles of sugars) varywithin the species and are even unstable for agiven strain during vegetative multiplication. Tax-onomists realized that they could not differentiatespecies based solely on phenotypic discontinuitycriteria. They progressively established a delimita-tion founded on the biological and genetic defini-tion of a species.

In theory, a species can be defined as a collectionof interfertile strains whose hybrids are themselvesfertile—capable of producing viable spores. Thisbiological definition runs into several problems

(a)

(b)

Fig. 1.19. Observation of two enological yeast specieshaving an apiculated form. (a) Hanseniaspora uvarum.(b) Saccharomycodes ludwigii

Fig. 1.20. Binary fission of Schizosaccharomycespombe


Table 1.2. Physiological characteristics of the principal grape and wine yeasts (Barnett et al., 1990)

F1

D-G

luco

sefe

rmen

tatio

nF

2D-G

alac

tose

ferm

enta

tion

F3M

alto

sefe

rmen

tatio

nF

4M

eα

-D-g

luco

side

ferm

enta

tion

F5Su

cros

efe

rmen

tati

onF

5α

,α-T

reha

lose

ferm

enta

tion

F7M

elib

iose

ferm

enta

tion

F8la

ctos

eF

erm

enta

tion

F9C

ello

bios

efe

rmen

tatio

nF1

0M

elez

itose

ferm

enta

tion

F11

Raf

finos

efe

rmen

tatio

nF1

2In

ulin

ferm

enta

tion

F13

Sta

rch

ferm

enta

tion

C1

D-G

luco

segr

owth

C2

D-G

alac

tose

grow

thC

3L-S

orbo

segr

owth

C4

D-G

luco

sam

ine

grow

thC

5D-R

ibos

egr

owth

C6

D-X

ylos

egr

owth

C10

Suc

rose

grow

thC

11M

alto

segr

owth

C12

α,α

-Tre

halo

segr

owth

C13

Me

α- D

-Glu

cosi

degr

owth

C14

Cel

lobi

ose

grow

thC

15S

alic

ingr

owth

C16

Arb

utin

grow

thC

17M

elib

iose

grow

thC

18L

acto

segr

owth

C19

Raf

finos

egr

owth

C20

Mel

ezito

segr

owth

C21

Inul

ingr

owth

C22

Sta

rch

grow

thC

23G

lyce

rol

grow

thC

24E

ryth

rito

lgr

owth

C25

Rib

itol

grow

thC

26X

ylit

olgr

owth

C28

D-G

luci

tol

grow

thC

29D-M

anni

tol

grow

th

Candida stellata + − − − + − − − − − v − − + − v − − − + − − − − − − − − + − − − − − − − − −Candida vini v − − − − − − − − − − − − + − − − − − − − − − − − − − − − − − − v − v v + +Candida famata v v v − v v v − v v v − − + + v v v + + + + + v v v v v + + v v + v + + + +Dekkera anomala + v v v + v − v v v v − − v v − v − − v v v v v v v − v v v − − v − − − − −Dekkera bruxellensis v v v + + + − − v v − − − v v − v v − v v v v v v v − − v v v − v − − − − −Hanseniaspora uvarum + − − − − − − − v − − − − + − − − − − − − v − + + + − − − − − − − − − − v −Metschnikowia pulcherrima + v − − − − − − − − − − − + + v v v v + + + + + + + − − − + − − + − v + + +Pichia anomala + v v v + v − − v − v − v + v − − v v + + v + v + + − − v v − + + v v v + +Pichia fermentans + − − − − − − − − − − − − + − − v − + − − − − − − − − − − − − − + − − − − −Pichia membranefaciens v − − − − − − − − − − − − + − v v − v − − − − − − − − − − − − − v − − v − −∗Saccharomyces cerevisiae + v v v v v v − − v v − v + v − − − − v v v v − − − v − v v − v v − − − v vSaccharomycodes ludwigii + − − − + − − − − − v − − + − − − − − + − − − + + + − − v − − − v − − − − −Kluyveromyces thermolerens + v v v + v − − − v + v − + v + − − − + + + + − − − − − + + v − + − v + v vSchizosaccharomyces pombe + v v v + − v − − − v v v + v − − − − + + v v − − − v − + − v v − − − − − −Zygosaccharomyces bailii + − − − v v − − − − v − − + v v − − − v − v − − − − − − v − − − v − v v v v

C30

Cal

actit

olgr

owth

C31

myo

-Ino

sito

lgr

owth

C32

D-G

luco

no-1

,5-l

acto

negr

owth

C33

2-K

eto-

D-g

luco

nate

grow

thC

345-

Ket

o-D-g

luco

nate

grow

thC

35D-G

luco

nate

grow

th

C38

DL-L

acta

tegr

owth

C39

Suc

cina

tegr

owth

C40

Cit

rate

grow

thC

42E

than

olgr

owth

N1

Nit

rate

grow

thN

2N

itri

tegr

owth

N3

Eth

ylam

ine

grow

thN

4L-L

ysin

egr

owth

N5

Cad

aver

ine

grow

thN

6C

reat

ine

grow

thN

7C

reat

inin

egr

owth

N8

Glu

cosa

min

egr

owth

V1

Gro

wth

W/O

vita

min

sV

2G

row

thW

/Om

yo-I

nosi

tol

V3

Gro

wth

W/O

Pan

toth

enat

eV

4G

row

thW

/OB

ioti

nV

5G

row

thW

/OT

hiam

inV

6G

row

thW

/OB

ioti

n&

Thi

amin

V7

Gro

wth

W/O

Pyr

idox

ine

V8

Gro

wth

W/O

Pyr

idox

ine

&T

hiam

inV

9G

row

thW

/ON

iaci

n

T2

Gro

wth

at30

◦ CT

3G

row

thW

/Oat

35◦ C

T4

Gro

wth

W/O

at37

◦ CT

5G

row

thW

/Oat

40◦ C

O1

0.01

%C

yclo

hexi

mid

egr

owth

O3

1%A

cetic

acid

grow

thO

450

%D-G

luco

segr

owth

M2

Ace

ticac

idpr

oduc

tion

M3

Ure

ahy

drol

ysis

Candida stellata − − − − − − − − − − − − − + − − − − − − v − − − + − v + v − − − − v − −Candida vini − − − v − v + − + − − + + + − − − − + + v v v + + v − − − − − − − −Candida famata v − v + v v + + + − v + + + v − − v + + v + v + + + v v − v − + − −Dekkera anomala − − v − − v v v − v v + + + + − − − − v + − − − + − + + + v − + − − + −Dekkera bruxellensis − − v v − − v − − v v v + + + − − − − + + − − − + − + + v v v + v − + −Hanseniaspora uvarum − − + + − v − − − − − − + + + − − − − − − v v − − − − + − − − + − v − −Metschnikowia pulcherrima − − + + + v + + + − − + + + − − − − + + − + − + + v v − − − − v − −Pichia anomala − − + − − v + + + − + + + + + − − − + + + + + + + + + + v v − − − v − −Pichia fermentans − − v − − − + + + − − − + + + − − + − + + + − − + − + + + + − − − v − −Pichia membranefaciens − − v − − − v v − + − − + + + − − v v + + v v v + v + + v v − − − v − −∗Saccharomyces cerevisiae − − v − − − v v − v − − − − − − − − v v v v v v v v + + v v v − − v − −Saccharomycodes ludwigii − − v − − − v − − v − − + + + − − + − + − − v − − − − + + v − − − − − −Kluyveromyces thermolerens − − v v − v − v − v − − + + + − − − − − v v v v v v + + v − − − + − −Schizosaccharomyces pombe − − v v − v − − − − − − v v v − − − − − v v v − v − − + + + v v + + − +Zygosaccharomyces bailii − − v v − − − − − + − − + + + − − − − + + − + − + + + v v − − + + − −+: test positive; −: test negative; v: variable result.∗With these tests they cannot be differentiated from S. bayanus, S. paradoxus and S. pastorianus.


when applied to yeasts. First of all, a large num-ber of yeasts (Deuteromycetes) are not capable ofsexual reproduction. Secondly, a lot of Ascomy-cetes yeasts are homothallic; hybridization tests areespecially fastidious and difficult for routine iden-tification. Finally, certain wine yeast strains havelittle or no sporulation aptitude, which makes theuse of strain infertility criteria even more difficult.

To overcome these difficulties, researchers havedeveloped a molecular taxonomy over the last15 years based on the following tests: DNA recom-bination; the similarity of DNA base composition;the similarity of enzymes; ultrastructure charac-teristics; and cell wall composition. The DNArecombination tests have proven to be effectivefor delimiting yeast species. They measure therecombination percentages of denatured nuclearDNA (mono-stranded) of different strains. Anelevated recombination rate between two strains(80–100%) indicates that they belong to the samespecies. A low recombination percentage (less than20% of the sequences in common) signifies thatthe strains belong to different and very distantspecies. Combination rates between these extremesare more difficult to interpret.

1.8.3 Successive Classificationsof the Genus Saccharomycesand the Position of Wine Yeastsin the Current Classification

Due to many changes in yeast classification andnomenclature since the beginning of taxonomicstudies, enological yeast names and their positionsin the classification have often changed. Thishas inevitably resulted in some confusion forenologists and winemakers. Even the most recentenological works (Fleet, 1993; Delfini, 1995; Boul-ton et al., 1995) use a number of different epithets(cerevisiae, bayanus, uvarum, etc.) attached to thegenus name Saccharomyces to designate yeastsresponsible for alcoholic fermentation. Althoughstill in use, this enological terminology is nolonger accurate to designate the species currentlydelimited by taxonomists.

The evolution of species classification forthe genus Saccharomyces since the early 1950s

(Table 1.3) has created this difference between thedesignation of wine yeasts and current taxonomy.By taking a closer look at this evolution, the originof the differences may be understood.

In Lodder and Kregger-Van Rij (1952), thenames cerevisiae, oviformis, bayanus, uvarum, etc.referred to a number of the 30 species of thegenus Saccharomyces. Ribereau-Gayon and Pey-naud (1960) in the Treatise of Œnology consid-ered that two principal fermentation species werefound in wine: S. cerevisiae (formerly called ellip-soideus) and S. oviformis. The latter was encoun-tered especially towards the end of fermentationand was considered more ethanol resistant. Thedifference in their ability to ferment galactose dis-tinguished the two species. S. cerevisiae (Gal+)fermented galactose, whereas S. oviformis (Gal−)did not. According to the same authors, the speciesS. bayanus was rarely found in wines. Althoughit possessed the same physiological fermentationand sugar assimilation characters as S. oviformis,its cells were more elongated, its fermentation wasslower, and it had a particular behavior towardsgrowth factors. The species S. uvarum was iden-tified in wine by many authors. It differed fromcerevisiae, oviformis and bayanus because it couldferment melibiose.

In Lodder’s following edition (Lodder 1970), thenumber of species of the genus Saccharomycesincreased from 30 to 41. Some species formerlygrouped with other genera were integrated into thegenus Saccharomyces. Moreover, several speciesnames were considered to be synonyms anddisappeared altogether. Such was the case ofS. oviformis, which was moved to the speciesbayanus. The treatise of Ribereau-Gayon et al.(1975) considered, however, that the distinctionbetween oviformis and bayanus was of enologicalinterest because of the different technologicalcharacteristics of these two yeasts. Nevertheless,by the beginning of the 1980s most enologicalwork had abandoned the name oviformis andreplaced it with bayanus. This name change beganthe confusion that currently exists.

The new classification by Kregger-Van Rij(1984), based on Yarrow’s work on base per-centages of guanine and cytosine in yeast DNA,

Cytology, Taxonomy and Ecology of Grape and Wine Yeasts 27Ta

ble

1.3.

Evo

lutio

nof

the

nom

encl

atur

efo

rth

eSa

ccha

rom

yces

genu

s,19

52–

1990

1952

:T

heYe

asts

,19

70:

The

Yeas

ts,

1984

:T

heYe

asts

,20

00:

Yeas

ts(B

arne

ttet

al.)

aTa

xono

mic

Stud

y—I

aTa

xono

mic

Stud

y—II

aTa

xono

mic

Stud

y—II

I(L

odde

ran

dK

regg

er-V

anR

ij)(L

odde

r)(Y

arro

w)

Sacc

haro

myc

esce

revi

siae

⇒Sa

ccha

rom

yces

cere

visi

aeSa

ccha

rom

yces

past

oria

nus

Sacc

haro

myc

esac

eti

Sacc

haro

myc

esce

revi

siae

Sacc

haro

myc

esba

yanu

s⇒

Sacc

haro

myc

esba

yanu

sSa

ccha

rom

yces

ovif

orm

isSa

ccha

rom

yces

cape

nsis

Sacc

haro

myc

eslo

gos

Sacc

haro

myc

espr

osto

serd

ovii

Sacc

haro

myc

esba

yanu

sSa

ccha

rom

yces

chev

alie

ri⇒

Sacc

haro

myc

esch

eval

ieri

Sacc

haro

myc

esfr

uctu

umSa

ccha

rom

yces

core

anus

Gro

upI

Sacc

haro

myc

esla

ctis

Sacc

haro

myc

esdi

asta

ticus

Sacc

haro

myc

esSa

ccha

rom

yces

eleg

ans

Sacc

haro

myc

esgl

obos

usSa

ccha

rom

yces

cere

visi

aeSa

ccha

rom

yces

para

doxu

sse

nsu

stri

cto

Sacc

haro

myc

eshe

tero

geni

cus

Sacc

haro

myc

eshe

tero

geni

cus

Sacc

haro

myc

esfe

rmen

tati

Sacc

haro

myc

eshi

enip

iens

isSa

ccha

rom

yces

mel

lisSa

ccha

rom

yces

inus

itatu

sSa

ccha

rom

yces

italic

us⇒

Sacc

haro

myc

esita

licus

Sacc

haro

myc

esst

eine

riSa

ccha

rom

yces

norb

ensi

sSa

ccha

rom

yces

past

oria

nus

Sacc

haro

myc

espa

stor

iSa

ccha

rom

yces

olea

ceus

Sacc

haro

myc

esca

rlsb

erge

nsis

Sacc

haro

myc

esol

eagi

nosu

sSa

ccha

rom

yces

uvar

um⇒

Sacc

haro

myc

esuv

arum

Sacc

haro

myc

esin

com

pspi

cuus

Sacc

haro

myc

esac

idif

acie

nsSa

ccha

rom

yces

amur

cae

Sacc

haro

myc

esba

ilii⇒

Sacc

haro

myc

esba

ilii

Sacc

haro

myc

esfr

agili

sSa

ccha

rom

yces

cidr

iSa

ccha

rom

yces

dair

ensi

s⇒

Sacc

haro

myc

esda

iren

sis

Sacc

haro

myc

esda

iren

sis

Sacc

haro

myc

esde

lbru

ecki

i⇒

Sacc

haro

myc

esde

lbru

ecki

iSa

ccha

rom

yces

mar

xian

usSa

ccha

rom

yces

eupa

gycu

sSa

ccha

rom

yces

exig

uus

⇒Sa

ccha

rom

yces

exig

uus

⇒Sa

ccha

rom

yces

exig

uus

Sacc

haro

myc

esex

iguu

sSa

ccha

rom

yces

vero

nae

Sacc

haro

myc

esfe

rmen

tati

Sacc

haro

myc

esflo

rent

inus

⇒Sa

ccha

rom

yces

flore

ntin

usSa

ccha

rom

yces

unis

poru

s

Gro

upII

Sacc

haro

myc

esbi

spor

us⇒

Sacc

haro

myc

esbi

spor

usSa

ccha

rom

yces

Sacc

haro

myc

esw

illia

nus

Sacc

haro

myc

eskl

oeck

eria

nus

sens

ula

toSa

ccha

rom

yces

kluy

veri

⇒Sa

ccha

rom

yces

serv

azzi

Sacc

haro

myc

esm

icro

ellip

sode

s⇒

Sacc

haro

myc

esm

icro

ellip

sode

sSa

ccha

rom

yces

mon

tanu

sSa

ccha

rom

yces

mra

kii

Sacc

haro

myc

esca

stel

liSa

ccha

rom

yces

pret

orie

nsis

Sacc

haro

myc

esro

sei⇒

Sacc

haro

myc

esro

sei

Sacc

haro

myc

esro

uxii

⇒Sa

ccha

rom

yces

roux

iiSa

ccha

rom

yces

saito

anus

Sacc

haro

myc

eskl

uyve

ri

G

roup

III

Sacc

haro

myc

este

lluri

s⇒

Sacc

haro

myc

este

lluri

sSa

ccha

rom

yces

tran

svaa

lens

isSa

ccha

rom

yces

unis

poru

s⇒

Sacc

haro

myc

esun

ispo

rus

Sacc

haro

myc

esva

fer

Sacc

haro

myc

esse

rvaz

zi


brought forth another important change in the des-ignation of the Saccharomyces species. Only sevenspecies continued to exist, while 17 names becamesynonyms of S. cerevisiae. Certain authors con-sidered them to be races or physiological vari-eties of the species S. cerevisiae. As with the pre-ceding classification, these races of S. cerevisiaewere differentiated by their sugar utilization profile(Table 1.4). However, this method of classifica-tion was nothing more than an artificial taxonomywithout biological significance. Enologists took tothe habit of adding the varietal name to S. cere-visiae to designate wine yeasts: S. cerevisiae var.cerevisiae, var. bayanus, var. uvarum, var. cheva-lieri, etc. In addition, two species, bailii and rosei,were removed from the genus Saccharomyces andintegrated into another genus to become Zygosac-charomyces bailii and Torulaspora delbrueckii,respectively.

The latest yeast classification (Barnett et al.,2000) is based on recent advances in genet-ics and molecular taxonomy—in particular, DNArecombination tests reported by Vaughan Martiniand Martini (1987) and hybridization experiments

between strains (Naumov, 1987). It has againthrown the species delimitation of the genusSaccharomyces into confusion. The species nownumber 10 and are divided into three groups(Table 1.3). The species S. cerevisiae, S. bayanus,S. paradoxus and S. pastorianus cannot be differ-entiated from one another by physiological testsbut can be delimited by measuring the degree ofhomology of their DNA (Table 1.5). They form thegroup Saccharomyces sensu stricto. S. pastorianusreplaced the name S. carlsbergensis, which wasgiven to brewer’s yeast strains used for bot-tom fermentation (lager) and until now includedin the species cerevisiae. The recently delimitedS. paradoxus species includes strains initially iso-lated from tree exudates, insects, and soil (Naumovet al., 1998). It might constitute the natural com-mon ancestor of three other yeast species involvedin the fermentation process. Recent genomic anal-ysis (Redzepovic et al., 2002) identified a highpercentage of S. paradoxus in Croatian grapemicroflora. The occurrence of this species in othervineyards around the world and its winemakingproperties certainly deserve further investigation.

Table 1.4. Physiological Races of Saccharomyces cerevisiae regrouped under asingle species Saccharomyces cerevisiae by Yarrow and Nakase (1975)

Fermentation

Ga Su Ma Ra Me St

Saccharomycesaceti − − − − − −bayanus − + + + − −capensis − + − + − −cerevisiae + + + + − −chevalieri + + − + − −coreanus + + − + + −diastaticus + + + + − +globosus + − − − − −heterogenicus − + + − − −hienipiensis − − + − + −inusitatus − + + + + −norbensis − − − − + −oleaceus + − − + + −oleaginosus + − + + + −prostoserdovii − − + − − −steineri + + + − − −uvarum + + + + + −

Ga = D-galactose; Su = saccharose; Ma = maltose; Ra = raffinose; Me = melibiose; St =soluble starch.


Table 1.5. DNA/DNA reassociation percentages between the four species belonging togenus Saccharomyces sensu stricto (Vaughan Matini and Martini, 1987)

S. cerevisiae S. bayanus S. pastorianus S. paradoxus

S. cerevisiae 100S. bayanus 20 100S. pastorianus 58 70 100S. paradoxus 53 24 24 100

A second group, Saccharomyces sensu largo, ismade up of the species exiguus, castelli, servazziand unisporus. The third group consists only ofthe species kluyveri. Only the first group com-prises species of enological interest: S. cerevisiae,S. bayanus, and, possibly, S. paradoxus, if its suit-ability for winemaking is demonstrated. This newclassification has created a lot of confusion inthe language pertaining to the epithet bayanus.For taxonomists, S. bayanus is a species distinctfrom S. cerevisiae. For enologists and winemakers,bayanus (ex oviformis) designates a physiologicalrace of S. cerevisiae that does not ferment galac-tose and possesses a stronger resistance to ethanolthan Saccharomyces cerevisiae var. cerevisiae.

By evaluating the infertility of strains (a basicspecies delimitation criterion), Naumov et al.(1993) demonstrated that most strains fermentingmelibiose (Mel+) isolated in wine, and until nowclassed as S. cerevisiae var. uvarum, belong tothe species S. bayanus. Some strains, however,can be crossed with a reference S. cerevisiae toproduce fertile descendants. They are thereforeattached to S. cerevisiae. These results confirm, butnevertheless put into perspective, the past worksof Rossini et al. (1982) and Bicknell and Douglas(1982), which were based on DNA recombinationtests. The DNA recombination percentages arelow between the uvarum and cerevisiae strainstested, but they are elevated between these sameuvarum strains and the S. bayanus strain (CBS380). In other words, most enological strainsformerly called uvarum belong to the speciesS. bayanus. This relationship, however, is notcomplete. Certain Saccharomyces Mel+ found inthe spontaneous fermentations of grapes belong tocerevisiae. The yeasts that enologists commonlycalled S. cerevisiae var. bayanus , formerly S. ovi-formis, were studied to determine if they belong to

the species bayanus or to the species cerevisiae, asthe majority of uvarum strains. In this case, theirdesignation only leads to confusion.

All of the results of molecular taxonomy pre-sented above show that the former phenotypicclassifications, based on physiological identifica-tion criteria, are not even suitable for delimiting thesmall number of fermentative species of the genusSaccharomyces found in winemaking. Moreover,specialists have long known about the instability ofphysiological properties of Saccharomyces strains.Rossini et al. (1982) reclassified a thousand strainsfrom the yeast collection of the Microbiology Insti-tute of Agriculture at the University of Perouse.During this research, they observed that 23 out of591 S. cerevisiae strains conserved on malt agarlost the ability to ferment galactose. Twenty threestrains ‘became’ bayanus, according to Lodder’s(1970) classification. They found even more fre-quently that, over time, strains acquired the abilityto ferment certain sugars. For example, 29 out of113 strains of Saccharomyces oviformis becamecapable of fermenting galactose, thus ‘becoming’cerevisiae. According to these authors, this physi-ological instability is a specific property of strainsfrom the Saccharomyces group sensu stricto. In thesame collection, no noticeable change in fermen-tation profiles was observed in 150 strains of Sac-charomyces rosei (today Torulaspora delbrueckii )or in 300 strains of Kloeckera apiculata. Geneticmethods are therefore indispensable for identifyingwine yeasts. Yet DNA recombination percentagemeasures or infertility tests between homothallicstrains, a long and fastidious technique, are notpractical for routine microbiological controls. Theamplification of genome segments by polymeriza-tion chain reaction (PCR) is a quicker and easiermethod which has recently proved to be an excel-lent tool for discrimination of wine yeast species.


1.8.4 Delimitation of WinemakingSpecies of S. cerevisiae andS. bayanus by PCR

Since its discovery by Saiki et al. (1985), PCRhas often been used to identify different plant andbacteria species. This technique consists of enzy-matically amplifying one or several gene fragmentsin vitro. The reaction is based on the hybridiza-tion of two oligonucleotides which frame a targetregion on a double-stranded DNA or template.These oligonucleotides have different sequences

and are complementary to the DNA sequenceswhich frame the strand to amplify. Figure 1.21illustrates the different stages of the amplificationprocess. The DNA is first denatured at a high tem-perature (95◦C). The reactional mixture is thencooled to a temperature between 37 and 55◦C, per-mitting the hybridization of these oligonucleotideson the denatured strands. The strands serve asprimers from which a DNA polymerase permits thestage-by-stage addition of desoxyribonucleotidicunits in the 5′ –3′ direction. The DNA polymerase(Figure 1.22) requires four deoxyribonucleoside-5′

First step: denaturation of DNA at 95°C

Second step: primer hybridization

Third step: elongation at 72°C

End of first cycle

Amplified zone

3’OH

3’OH

Fig. 1.21. Principle of the polymerization chain reaction (PCR)


OH2C

H

Base

H

O

Primer chain

OH2c

H

Base

H

HO H

HH

−O P O

O

O−

P O PO

O−

O

O− O

H H

HHO

••

Simple strand D

NA

matrix

Fig. 1.22. Mode of action of a DNA polymerase

triphosphates (dATP, dGTP, dTTP, dCTP). Aphosphodiester bond is formed between the 3′-OHend of the primer and the innermost phosphorus ofthe activated deoxyribonucleoside; pyrophosphateis thus liberated. The newly synthesized strand isformed on the template model. A thermoresistantenzyme, the TAQ DNA polymerase, is derivedfrom the thermoresistant bacteria Thermus aquati-cus. It permits a large number of amplificationcycles (25–40) in vitro without having to add theDNA polymerase after each denaturation. In thismanner, the DNA fragment amplified during thefirst cycle serves as the template for the follow-ing cycles. In consequence, each successive cycledoubles the target DNA fragment—amplified bya factor of 105 to 106 during 25–30 amplificationcycles.

Hansen and Kielland-Brandt (1994) proposedMET 2 gene PCR amplification to differentiatebetween S. cerevisiae and S. bayanus, while work-ing on strain types of the species cerevisiaeand bayanus and on a strain of the varietyS. uvarum. This gene, which codes for the synthe-sis of the homoserine acetyltransferase, has differ-ent sequences in the two species. Part of the geneis initially amplified by using two complementaryoligonucleotides of the sequences which border the

fragment to be amplified. The amplificat obtained,about 600 b.p, is the same size for the strains ofthe species cerevisiae and bayanus tested, as wellas for the isolate designated S. uvarum. Differentrestriction endonucleases, which recognize certainspecific DNA sequences, then digest the ampli-fied fragment. Figure 1.23 gives an example of themode of action of the EcoRI restriction endonucle-ase. This enzyme recognizes the base sequencesGAATTC and cuts at the location indicated bythe arrows. Electrophoresis is used to separatethe obtained fragments. As a result, the restric-tion fragment length polymorphism (RFLP) canbe appreciated. The restriction profiles obtaineddiffer between cerevisiae and bayanus. They areidentical for the strain types bayanus and uvarumtested.

This PCR–RFLP technique associated to theMET 2 gene has been developed and adapted forrapid analysis. The whole cells are simply heatedin water (95◦C), 10 minutes before amplification.Only two restriction enzymes are used: EcoRI andPstI (Masneuf et al., 1996a,b). The MET 2 ampli-ficat (580 bp) is cut into two fragments (369 and211 bp) by EcoRI in S. cerevisiae. PstI restrictioncreates two fragments for strain type S. bayanus.EcoRI does not cut the MET 2 amplificat ofS. bayanus, nor does PstI cut the S. cerevisiaeamplificat (Figure 1.24). Masneuf (1996) demon-strated that S. paradoxus can be identified by thismethod. Its MET 2 gene amplificat produces onefragment of the same size as with the two otherspecies. This one, however, is not cleaved byEcoRI or PstI, but rather by Mae III.

G A A T T C

C T T A A G

G

C T T A A5′p G

5′

5′

3′

5′

5′

3′

3′

3′

5′pA A T T C

Fig. 1.23. Recognition site and cutting mode of anECOR1 restriction endonuclease


S. cerevisiae S. bayanus S. bayanusS. cerevisiae

EcoRI PstI

580 pb

369 pb

219 pb

365 pb

215 pb

Fig. 1.24. Identification principles for the speciesS. cerevisiae and S. bayanus by the MET2 genePCR-RFLP technique, after cutting the amplificat byEcoRI and PstI restriction enzymes

1 2 3 4 5 6 7 8 9 M1 2 3 4 5 6 7 8 9 M

692 pb404 pb242 pb

(a) (b)

Fig. 1.25. Agar gel electrophoresis (1.8%) of (a) EcoR1and PstI digestions of the MET2 gene amplificats of theMel+ strains studied by Naumov et al. (1993). Band 1:S. bayanus SCU 11; band 2: S. bayanus SCU 13; band3: S. bayanus SCU 73; band 4: S. bayanus L19; band5: S. bayanus L490: band 6: S. cerevisiae L 579; band7: S. cerevisiae L 1425; band 8: S. cerevisiae VKM-Y502; band 9: S. bayanus VKM-Y 1146. M = molecularweight markers

By applying this relatively simple and quicktechnique to different enological strains of S. uva-rum studied by Naumov et al. (1993), strainsattached to the species bayanus by hybridizationtests have been clearly demonstrated to presentthe same profile characteristic as bayanus (twobands after restriction with PstI, no restriction withEcoRI). On the other hand, the uvarum strainsincluded in the species cerevisiae, according tohybridization tests, effectively have a restrictionprofile characteristic of S. cerevisiae (Figure 1.25and Table 1.6). The delimitation of the speciescerevisiae and bayanus by these two methodsproduced identical results for the 12 strainsanalyzed.

This type of PCR–RFLP analysis of the MET 2gene has been extended to different selected yeaststrains available in the trade and currently used

in winemaking. Depending on their ability toferment galactose, wine professionals in the entireworld still call these strains cerevisiae or bayanus(Table 1.7). For all of these strains, restrictionprofile characteristics of the species S. cerevisiaehave been obtained.

In the same manner, the species of 82 indige-nous Saccharomyces strains isolated in wines infermentation and on grapes has been determined(Table 1.8). For the eight Gal+Mel− strains andthe 47 Gal−Mel− strains analyzed, called cere-visiae and bayanus respectively by enologists, therestriction profiles of the MET2 gene amplificatare characteristic of the species S. cerevisiae. Sim-ilar results were obtained for 2 chevalieri strainsfermenting galactose but not maltose (Ma−), aswell as for the capensis strain (Gal−Ma−). Mostof the Mel+ strains, called uvarum until now,(11 out of 12 for the isolates from Sauternesand 11 out of 11 for the isolates from Sancerre),belong to the species S. bayanus. Yet certain Mel+are S. cerevisiae (one strain from Sauternes andtwo strains from the Lallemand collection). Tosummarize, the classification of the main wine-making yeasts (Section 1.8.3) has gone throughthree stages. Initially, several separate specieswere envisaged: S. cerevisiae, S. bayanus and/orS. oviformis, and S. uvarum. Subsequently, all ofthese were thought to belong to a single species:S. cerevisiae. The current classification identifiesthree distinct species on the basis of molecu-lar biological data: S. cerevisiae, S. bayanus, andS. paradoxus. As the strains of S. bayanus used inwinemaking belong exclusively to the S. uvarumvariety (or sub-species), the remainder of thisHandbook will consider just two species of wine-making yeasts, S. cerevisiae and S. uvarum. Theinvolvement of S. paradoxus in grape fermentationmicroflora has yet to be confirmed.

Finally, PCR–RFLP associated with the MET 2gene can be used to demonstrate the existenceof hybrids between the species S. cerevisiae andS. bayanus. This method has been used to provethe existence (Masneuf et al., 1998) of such anatural hybrid (strain S6U var. uvarum) amongdry yeasts commercialized by Lallemand Inc.(Montreal, Canada). Ciolfi (1992, 1994) isolated


Table 1.6. Characterization by PCR–RFLP of the MET 2 gene of 12 S. uvarum (Mel+) reclarified, after hybridizationtesty by Naumov et al. (1993), as the species S. cerevisiae and S. bayanus (Masneuf, 1996)

Strain CLIB number Origin Author Hybridization test PCR–RFLP of theMET 2 gene

VKM Y-502 219 Russia Naumov S. cerevisiae (control) S. cerevisiaeVKM Y-1146 218 Russia grape Naumov S. bayanus (control) S. bayanus

58 l — FŒB must Sapis-Domercq S. bayanus S. bayanusSCU 11 101 ITVN wine Poulard S. bayanus S. bayanusSCU 13 102 ITVN wine Poulard S. bayanus S. bayanusSCU 74 103 ITVN wine Poulard S. bayanus S. bayanusL 19 108 ITVT wine Cuinier S. bayanus S. bayanusL 99 109 ITVT wine Cuinier S. bayanus S. bayanusL 490 110 ITVT wine Cuinier S. bayanus S. bayanusDBVPG 1642 113 UPG grape Vaughan Martini S. bayanus S. bayanusDBVPG 1643 114 UPG grape Vaughan Martini S. bayanus S. bayanusDBVPG 1689 115 UPG grape Vaughan Martini S. bayanus S. bayanus

L 579 94 ITVT wine Cuinier S. cerevisiae S. cerevisiaeL 1425 95 ITVT wine Cuinier S. cerevisiae S. cerevisiae

CLIB: Collection de Levures d’Interet Biotechnologique (collection of yeasts of Biotechnological Interest) INA-PG, Grignon, France.FŒB: Faculte Œnologie de I’Universite de Bordeaux II, Talence, France.ITVN: Institut Technique du Vin (Institut of Wine Technology), Centre d’experimentation de Nantes, France.ITVT: Institut Technique du Vin (Institut of Wine Technology), Centre d’experimentation de Tours, France.UPG: Univeraita degli Studi de Perugia, Italy.

Table 1.7. Characterization by PCR–RFLP of the MET2 gene of various selected commercial strains used inwinemaking

Strains Commercial brand Origin Enological designation Species

VL1 Zymaflore VL1 FŒB S. cerevisiae S. cerevisiaeVL3c Zymaflore VL3 FŒB S. cerevisiae S. cerevisiaeWET 136 Siha levactif 3 Dormstadt S. cerevisiae S. cerevisiae71B Actiflore primeur INRA Narbonne S. cerevisiae S. cerevisiaeF10 Zymaflore F10 FŒB S. bayanus S. cerevisiaeR2 Vitlevure KD n-a S. bayanus S. cerevisiaeBO213 Actiflore bayanus Institut Pasteur S. bayanus S. cerevisiaeCH158 Siha levactif 4 n-a S. bayanus S. cerevisiaeQA23 Lalvin QA23 UTM S. bayanus S. cerevisiaeIOC182007 IOC 182007 IŒC S. bayanus S. cerevisiaeDV10 Vitlevure DV10 CIVC S. bayanus S. cerevisiaeO16 Lalvin O16 UB S. bayanus S. cerevisiaeEpemay2 Uvaferm CEG n-a S. bayanus S. cerevisiae

CLIB: Collection de Levures d’Interet Biotechnologique (Collection of yeast of biotechnological interest), INA-PG, Grignon, France.FŒB: Faculte d’ Œnologie de I’Universite de Bordeeux II, Talence, France.UTM: Universite de Trasos Montes, Portugal.ŒC: Institut Œnologique de Champagne, France.CIVC: Comite Interprofessionnel des vins de Champagne (Interprofessional Champagne Committee), Epernay, France.UB: Universite de Bourgogne, Dijon, France.na: not available.


Table 1.8. Characterization by PCR–RFLP of the MET 2 gene of various species of wild Saccharomyces isolated onthe grape and in wine (Masneuf, 1996)

Number of different Origin Collection Enological Speciesstrain analyzed designation

8 Sauternes wines FŒB cerevisiae S. cerevisiae2 Dry white Bordeaux wines FŒB bayanus S. cerevisiae9 Sauternes wines FŒB bayanus S. cerevisiae2 Dry white Bordeaux wines FŒB chevalieri S. cerevisiae1 Sauternes wines FŒB capensis S. cerevisiae

36 Unknown Lallemand bayanus S. cerevisiae11 Sauternes wines FŒB uvarum S. bayanus

1 Sauternes wines FŒB uvarum S. cerevisiae10 Sancerre and FŒB uvarum S. bayanus

Pouilly/Loire Valley wines1 Sancerre grapes FŒB uvarum S. bayanus2 Unknown Lallemand uvarum S. cerevisiae

FŒB: Faculte dŒnologie de I’Universite de Bordeaux II, Talence, France. Lallemand: Lallemand Inc. Montreal, Quebec, Canada.

1 2 3 4 5 6 7 8 9 M 1 2 3 4 5 6 7 8 9 M

692 pb404 pb242 pb

(a) (b)

Fig. 1.26. Electrophoresis in agarose gel (1.8%) of(a) EcoRI and (b) PstI digestions of the amplificats ofthe MET2 gene of the strain hybrid. Bands 1, 2, 3:sub-clones of the hybrid strain; band 4: hybrid strain;band 5: S. cerevisiae VKM-Y 502; band 6: S. bayanusVKM-Y 1146. M = molecular weight marker

this yeast in an Italian winery. It was selectedfor certain enological properties, in particular itsaptitude to ferment at low temperatures, its lowproduction of acetic acid, and its ability to preservemust acidity. The MET 2 gene restriction profilesof this strain by EcoRI and PstI, constitutedby three bands, are identical (Figure 1.26). Inaddition to the amplified fragment, two bandscharacteristic of S. cerevisiae with EcoRI and twobands characteristic of the species S. bayanus withPstI are obtained. The bands are not artefacts due toan impurity in the strain, because the amplificationof the MET 2 gene carried out on subclones(obtained from the multiplication of unique cellsisolated by a micromanipulator) produces identical

results. Furthermore, after sporulation of the strainin the laboratory, 10 tetrads were equally isolated bya micromanipulator after the digestion of the ascuscell wall. None of the 40 ascospores analyzed couldgerminate. The non-viability of the ascosporesconcurs with the hypothesis that this strain isan interspecific hybrid. Hansen of the Carlsberglaboratory (Denmark) sequenced two of the MET2gene alleles from this strain. The sequence of oneof the alleles is identical to that of the S. cerevisiaeMET2 gene, with the exception of one nucleotide.The sequence of the other allele is 98.5% similarto that of S. bayanus. The presence of this allele isthus probably due to an interspecific cross.

Recent research (Naumov et al., 2000b) hasshown that the S6U strain is, in fact, a tetraploidinterspecific hybrid. Indeed, the percentage germi-nation of spores from 24 tetrads, isolated using amicromanipulator, was very high (94%), whereasit would have been very low for a “normal” diploidinterspecific hybrid. The monospore clones in thisfirst generation (F1) were all capable of sporulat-ing, while none of the ascospores of the second-generation tetrads were viable. The hybrid natureof the monospore clones produced by F1 was con-firmed by the presence of the S. cerevisiae andS. uvarum MET2 gene, identified by PCR/RFLP.Finally, measuring the DNA content per cellusing flux cytometry estimation confirmed thatthe descendants of S6U were interspecific diploids


and that S6U itself was an allotetraploid. Natu-ral S. cerevisiae/S. uvarum hybrids have been iso-lated on grapes and in spontaneously fermentingmusts in Alsace (Lejeune and Masneuf, unpub-lished results).

Several other methods using PCR/RFLP havebeen applied to typing Saccharomyces itself.The fragments amplified were ribosomal DNAsequences (DNAr) (Guillamon et al., 1998; Nguyenet al., 2000).

1.9 IDENTIFICATION OF WINEYEAST STRAINS

1.9.1 General PrinciplesThe principal yeast species involved in grapemust fermentation, particularly S. cerevisiae, com-prise a very large number of strains with variedtechnological properties. The yeast strains whichare involved during winemaking influence fermen-tation speed, the nature and quantity of secondaryproducts formed during alcoholic fermentation,and the aromatic characters of the wine. The abil-ity to differentiate between the different strains ofS. cerevisiae is required for the following fields:the ecological study of wild yeasts responsible forthe spontaneous fermentation of grape must; theselection of strains presenting the best enologicalqualities; production and marketing controls; theverification of the implantation of selected yeastsused as yeast starter; and the constitution and main-tenance of wild or selected yeast collections.

Bouix et al. (1981) (cited in Van Vuuren andVan Der Meer, 1987) conducted the initial researchon infraspecific differentiation within S. cerevisiae.They attempted to distinguish strains by elec-trophoretic analysis of their exocellular proteinsand later (1987) used the separation of intra-cellular proteins. Other teams proposed identi-fying the strains by the analysis of long-chainfatty acids using gas chromatography (Tredouxet al., 1987; Augustyn et al., 1991; Bendova et al.,1991; Rozes et al., 1992). Although these dif-ferent techniques differentiate between certainstrains, they are irrefutably less discriminatingthan genetic differentiation methods. They also

present the major inconvenience of dependingon the physiological state of the strains andthe cultural conditions, which must always beidentical.

In the late 1980s, owing to the development ofgenetics, certain techniques of molecular biologywere successfully applied to characterize wineyeast strains. They are based on the clonalpolymorphism of the mitochondrial and genomicDNA of S. cerevisiae. These genetic methodsare independent of the physiological state of theyeast, unlike the previous techniques based on theanalysis of metabolism byproducts.

1.9.2 Mitochondrial DNA Analysis

The mtDNA of S. cerevisiae has two remarkableproperties: it is extremely polymorphous, depend-ing on the strain; and its is stable (it mutates verylittle) during vegetative multiplication. Restrictionendonucleases (such as EcoR5) cut this DNA atspecific sites. This process generates fragments ofvariable size which are few in number and can beseparated by electrophoresis on agarose gel.

Aigle et al. (1984) first applied this techniqueto brewer’s yeasts. Since 1987, it has beenused for the characterization of enological strainsof S. cerevisiae (Dubourdieu et al., 1987; Halletet al., 1988).

The extraction of mtDNA comprises severalstages. The protoplasts obtained by enzymaticdigestion of the cell walls are lysed in a hypo-tonic buffer. The mtDNA is then separated fromthe chromosomic DNA by ultracentrifugation in acesium chloride gradient, in the presence of bis-benzimide which acts as a fluorescent intercalat-ing agent. This agent amplifies the difference indensity between chromosomic and mtDNA. ThemtDNA has an elevated amount of adenine andthymine base pairs for which the bisbenzimide hasa strong affinity. Finally, the mtDNA is purifiedby a phenolchloroform-based extraction and anethanol-based precipitation.

Defontaine et al. (1991) and Querol et al. (1992)simplified this protocol by separating the mito-chondria from the other cellular constituents beforeextracting the DNA. In this manner, they avoided


the ultracentrifugation step. The coarse cellulardebris is eliminated from the yeast lysate by cen-trifuging at 1000 g. The supernatant is recen-trifuged at 1500 g to obtain the mitochondria. Themitochondria are then lysed in a suitable buffer toliberate the DNA.

Unlike industrial brewer strains analyzed byAigle et al. (1984), which have the same mtDNArestriction profile, implying that they are ofcommon origin, the enological yeast strains have alarge mtDNA diversity. This method differentiatesbetween most of the selected yeasts used inwinemaking as well as wild strains of S. cerevisiaefound in spontaneous fermentations (Figure 1.27).

This technique is very discriminating and nottoo expensive, but it is long and requires severalcomplex manipulations. It is useful for the subtlecharacterization of a small number of strains.Inoculation effectiveness can also be verified bythis method. To verify an inoculation, a sampleis taken during or towards the end of alcoholicfermentation. In the laboratory, the lees are placedin a liquid medium culture. The mtDNA restrictionprofile of this total biomass and of the yeast starterstrain are compared. If the restriction profile of thesample has no supernumerary bands with respectto the yeast starter strain profile, the yeast starterhas been properly implanted, with an accuracy of90%. In fact, in the case of a binary mixture,the minority strain must represent around 10% ofthe total population to be detected (Hallet et al.,1989).

1 2 3 4 5 6M

Fig. 1.27. Restriction profile by EcoR5 of mtDNA ofdifferent strains of S. cerevisiae. Band 1: F10; band 2:BO213; band 3: VLI; band 4: 522; band 5: Sita 3; band6: VL3c. M = marker

1.9.3 Karyotype Analysis

S. cerevisiae has 16 chromosomes with a sizerange between 250 and 2500 Kb. Its genomicDNA is very polymorphic; thus it is possible todifferentiate strains of the species according to thesize distribution of their chromosomes. Pulse-fieldelectrophoresis is used to separate S. cerevisiaechromosomes and permits the comparison of thekaryotypes of the strains. This technique usestwo electric fields oriented differently (90 to 120degrees). The electrodes placed on the sides of theapparatus apply the fields alternately (Figure 1.28).

+ + + + +

+ + + + +

+ + + + +

+ + + + +

+

+

+

+

+

+

+

++

++

+

−−

−−

+

−

+

−

+

−

+

−+

++

+

++

+Electrodes

Direction ofDNA migration

Gel wells

Direction ofElectrical current

Fig. 1.28. CHEF pulsed field electrophoresis device (contour clamped electrophoresis field)


Instant of first pulse After first pulse Instant of second pulse After second pulsePosition of fragments

Fig. 1.29. Principle of DNA molecule separation by pulsed field electrophoresis

The user can define the duration of the electriccurrent that will be applied in each direction(pulse). With each change in direction of theelectric field, the DNA molecules reorientatethemselves. The smaller chromosomes reorientatethemselves more quickly than the larger ones(Figure 1.29).

Blondin and Vezhinet (1988), Edersen et al.(1988) and Dubourdieu and Frezier (1990) appliedthis technique to identify enological yeast strains.Sample preparation is relatively easy. The yeastsare cultivated in a liquid medium, collected duringthe log phase, and then placed in suspension ina warm agarose solution that is poured into apartitioned mold to form small plugs.

Figure 1.30 gives an example of the iden-tification of S. cerevisiae strains isolated froma grape must in spontaneous fermentation by

Fig. 1.30. Example of electrophoretic (pulsed field)profile of S. cerevisiae strain caryotypes

this method. Vezhinet et al. (1990) have shownthat karyotype analysis can distinguish betweenstrains of S. cerevisiae as well or better thanthe use of mtDNA restriction profiles. Further-more, karyotype analysis is much quicker andeasier to use than mtDNA analysis. In the caseof ecological studies of spontaneous fermentationmicroflora, pulse-field electrophoresis of chromo-somes is extensively used today to characterizestrains of S. cerevisiae (Frezier and Dubourdieu,1992; Versavaud et al., 1993, 1995).

Very little research on the chromosomicpolymorphism in other species of grape andwine yeasts is currently available. Naumov et al.(1993) suggested that S. bayanus and S. cerevisiaekaryotypes can be easily distinguished. Otherauthors (Vaughan-Martini and Martini, 1993;Masneuf, 1996) have confirmed his results. Infact, a specific chromosomic band systematicallyappears in S. bayanus. Furthermore, there are onlytwo chromosomes whose sizes are less than 400 kbin S. bayanus but generally more in S. cerevisiae,in all of the strains that we have analyzed.

Species other than Saccharomyces, in partic-ular apiculated yeasts (Hanseniaspora uvarum,Kloeckera apiculata), are present on the grapeand are sometimes found at the beginning of fer-mentations. These species have fewer polymor-phous karyotypes and fewer bands than in Sac-charomyces. Versavaud et al. (1993) differentiatedbetween strains of apiculated yeast species andCandida famata by using restriction endonucleasesat rare sites (Not I and Sfi I). The endonucleasescut the chromosomes into a limited number of frag-ments, which were then separated by pulse-fieldelectrophoresis.


1.9.4 Genomic DNA Restriction ProfileAnalysis Associated with DNAHybridization by Specific Probes(Fingerprinting)

The yeast genome contains DNA sequences whichrepeat from 10 up to 100 times, such as the δ

or Y1 sequences of the chromosome telomeres.The distribution, or more specifically, the numberand location of these elements, has a certainintraspecific variability. This genetic fingerprint isused to identify strains (Pedersen, 1986; Degreet al., 1989).

The yeasts are cultivated in a liquid medium.Samples are taken during the log phase, as inthe preceding techniques. The entire DNA is iso-lated and digested by restriction endonucleases.The generated fragments are separated by elec-trophoresis on agarose gel and then transferredto a nylon membrane (Southern, 1975). Comple-mentary radioactive probes (nucleotide sequencestaken from δ and Y1 elements) are used tohybridize with fragments having homologoussequences. The result gives a hybridization profilecontaining several bands.

Genetic fingerprinting is a more complicatedand involved method than mtDNA or karyotypeanalysis. It is, however, without doubt the mostdiscriminating strain identification method andmay even discriminate too well. It has correctlyindicated minor differences between very closelyrelated strains. In fact, in the Bordeaux region,S. cerevisiae clones isolated from spontaneousfermentations in different wineries have beenencountered which have the same karyotype andthe same mtDNA restriction profile. Yet theirhybridization profiles differ according to sampleorigin (Frezier, 1992). These strains, probablydescendants of the same mother strain, havetherefore undergone minor random modifications,maintained during vegetative multiplication.

1.9.5 Polymerization Chain Reaction(PCR) Associated to δ Sequences

This method consists of using PCR to amplify cer-tain sequences of the yeast genome (Section 1.8.4),

occurring between the repeated δ elements, whoseseparation distance does not exceed a certainvalue (1 kb). Ness et al. (1992) and Masneuf andDubourdieu (1994) developed this method to char-acterize S. cerevisiae strains. The amplification iscarried out directly on whole cells. They are simplyheated to make the cellular envelopes permeable.The resulting amplification fragments are sepa-rated according to their size by electrophoresis inagarose gel and viewed using ultraviolet fluores-cence (Figure 1.31).

PCR profile analysis associated with δ sequencescan distinguish between most S. cerevisiae activedry yeast strains (ADY) used in winemaking(Figure 1.32): 25 out of the 26 selected commer-cial yeast strains analyzed. Lavallee et al. (1994)also observed excellent discriminating power withthis method while analyzing industrially producedcommercial strains from Lallemand Inc. (Mon-treal, Canada). In addition, this method permitsthe identification of 25 to 50 strains per day; itis the quickest of the different strain identifica-tion techniques currently available. When used for

A B C D

Strain A

Strain B

Strain C

Strain DAmplified fragmentsDelta sequences

Direction ofmigration

Analysis of amplified fragments by electrophoresis in agar gel

Fig. 1.31. Principle of identification of S. cerevisiaestrains by PCR associated with delta elements


1 2 3 4 5 6 7 8 9 M T

692 pb404 pb242 pb

Fig. 1.32. Electrophoresis in agar gel (at 1.8%) ofamplified fragments obtained from various commercialyeast strains. Band 1: F10; band 2: BO213; band 3:VL3c; band 4: UP30Y5; band 5: 522 D; band 6: EG8;band 7: L-1597; band 8: WET 136. M = molecularweight marker; T = negative control

indigenous strain identification in a given viticul-tural region, however, it seems to be less dis-criminating than karyotype analysis. PCR profilesof wild yeasts isolated in a given location oftenappear similar. They have several constant bandsand only a small number of variable discriminatingbands. Certain strains have the same PCR amplifi-cation profile while having different karyotypes. Ina given location, the polymorphism witnessed byPCR associated with δ sequences is less importantthan that of the karyotypes. This method is there-fore complementary to other methods for charac-terizing winemaking strains. PCR permits a rapidprimary sort of an indigenous population. Kary-otype analysis refines this discrimination.

S. bayanus strains cannot be distinguished bythis technique because their genome contains onlya few Ty elements.

Finally, because of its convenience and rapidityPCR associated with δ sequences facilitates verifi-cation of the implantation of yeast starters used inwinemaking. The analyses are effectuated on theentire biomass derived from lees, placed before-hand in a liquid medium in a laboratory culture.The amplification profiles obtained are comparedwith inoculated yeast strain profiles. They are iden-tical with a successful implantation, and differentif the inoculation fails. Figure 1.33 gives examplesof successful (yeasts B and C) and unsuccessful

1 2 3 4 5 6 7 8 9 10 11 M

692 pb

404 pb

242 pb

Fig. 1.33. Electrophoresis in agar gel (1.8%) ofamplified fragments illustrating examples of verifyingyeast implantation (successful: yeasts B and C;unsuccessful: yeasts A, D and E). Band 1: negativecontrol; band 2: Lees A; band 3: ADY A; band 4: LeesB; band 5: ADY B; band 6: Lees C; band 7: ADY C;band 8: Lees D; band 9: ADY D; band 10: Lees E; band11: ADY E. M = molecular weight marker

T 1 2 3 4 5 6 7

692 pb

404 pb

242 pb

Fig. 1.34. Determination of the detection threshold ofa contaminating strain. Band 1: strain A 70%, strainB 30%; band 2: strain A 80%, strain B 20%; band 3:strain A 90%, strain B 10%; band 4: strain A 99%, bandB 1%; band 5: strain A 99.9%, strain B 0.1%; band6: strain A; band 7: strain B. M = molecular weightmarker; T = negative control

(yeasts A, D, and E) implantations. Contaminatingstrains have a different amplification profile thanthe yeast starter. The detection threshold of a con-taminating strain was studied in the laboratory byanalyzing a mixture of two strains in variable pro-portions. In the example given in Figure 1.34, thecontaminating strain is easily detected at 1%. Inwinery fermentations, however, several minorityindigenous strains can coexist with the inoculated


strain. When must in fermentation or lees is ana-lyzed by PCR, the yeast implantation rate is at least90% when the amplification profiles of the lees andthe yeast starter are identical.

In light of various research, different DNAanalysis methods should be combined to identifywine yeast strains.

1.9.6 PCR with Microsatellites

Microsatellites are tandem repeat units of shortDNA sequences (1–10 nucleotides), i.e. in thesame direction and dispersed throughout theeukaryote genome (Field and Wills, 1998). Thenumber of motif repetitions is extremely vari-able from one individual to another, making thesesequences highly polymorphous in size. Theseregions are easily identified, thanks to the fullsequence of the S. cerevisiae genome, availableon the Internet in the Saccharomyces GenomeDatabase. Approximately 275 sequences have beenlisted, mainly AT dinucleotides and AAT and AACtrinucleotides (Perez et al., 2001). Furthermore,these sequences are allelic markers, transmittedto the offspring in a Mendelian fashion. Conse-quently, these are ideal genetic markers for iden-tifying specific yeast strains, making it possiblenot only to distinguish between strains but also toarrange them in related groups. This technique hasmany applications in man: paternity tests, forensicmedicine, etc. In viticulture, this molecular identi-fication method has already been applied to Vitisvinifera grape varieties (Bowers et al., 1999).

The technique consists of amplifying the regionof the genome containing these microsatellites,then analyzing the size of the amplified portion to alevel of detail of one nucleotide by electrophoresison acrylamide gel. This varies by a certain numberof base pairs (approximately 8–40) from one strainto another, depending on the number of times themotif is repeated. A yeast strain may be heterozy-gous for a given locus, giving two different-sizedamplified DNA fragments. Using 6 microsatellites,Perez et al. (2001) were able to identify 44 differ-ent genotypes within a population of 51 strains ofS. cerevisiae used in winemaking. Other authors(Gonzalez et al., 2001; Hennequin et al., 2001)

have shown that the strains of S. cerevisiae usedin winemaking are weakly heterozygous for theloci studied. However, interstrain variability ofthe microsatellites is very high. The results areexpressed in numerical values for the size of themicrosatellite in base pairs or the number of rep-etitions of the motifs on each allele. These digitaldata are easy to interpret, unlike the karyotypeimages on agarose gel, which are not really compa-rable from one laboratory to another. Microsatel-lite analysis has also been used to identify thestrains of S. uvarum used in winemaking (Masneufand Lejeune, unpublished). As the S. uvarum andS. cerevisiae microsatellites have different amplifi-cation primers, this method provides an additionalmeans of distinguishing between these species andtheir hybrids.

In future, this molecular typing method will cer-tainly be a useful tool in identifying winemakingyeast strains, ecological surveys, and quality con-trol of industrial production batches.

Finally, another technique has recently beenproposed for identifying Saccharomyces strainswith PCR by amplifying introns of the COX1mitochondrial DNA gene, which varies in numberand position in different strains. It is possibleto amplify either purified DNA or fermentingmust. This technique has been used to monitoryeast development during fermentation (Lopezet al., 2003).

1.10 ECOLOGY OF GRAPEAND WINE YEASTS

1.10.1 Succession of Grape and WineYeast Species

Until recently, a large amount of research focusedon the description and ecology of wine yeasts.It concerned the distribution and succession ofspecies found on the grape and then in wineduring fermentation and conservation (Ribereau-Gayon et al. 1975; Lafon-Lafourcade 1983).

The ecological study of grape and wine yeastspecies represents a considerable amount ofresearch. De Rossi began his research in the 1930s(De Rossi, 1935). Castelli (1955, 1967) pursued


yeast ecology in Italian vineyards. Peynaud andDomercq (1953) and Domercq (1956) publishedthe first results on the ecology of enologicalyeasts in France. They described not only thespecies found on the grape and during alcoholicfermentation, but also contaminating yeasts anddiseases. Among the many publications on thistheme since the 1960s in viticultural regionsaround the world, the following works are worthnoting: Brechot et al. (1962), Minarik (1971),Barnett et al. (1972), Park (1975), Cuinier andGuerineau (1976), Soufleros (1978), Belin (1979,1981), Poulard et al. (1980), Poulard and Lecocq(1981), Bureau et al. (1982), Rossini et al. (1982).

Yeasts are widespread in nature and are found insoils, on the surface of vegetables and in the diges-tive tract of animals. Wind and insects disseminatethem. They are distributed irregularly on the sur-face of the grape vine; found in small quantitieson leaves, the stem and unripe grapes, they col-onize the grape skin during maturation. Observa-tions under the scanning electron microscope haveidentified the location of yeasts on the grape. Theyare rarely found on the bloom, but multiply pref-erentially on exudates released from microlesionsin zones situated around the stomatal apparatus.Botrytis cinerea and lactic acid and acetic acid bac-teria spores also develop in the proximity of theseperistomatic fractures (Figure 1.35).

The number of yeasts on the grape berry, justbefore the harvest, is between 103 and 105, depend-ing on the geographical situation of the vineyard,

Fig. 1.35. Grape surface under scanning electron micro-scope, with detail of yeast peristomatic zones. Depart-ment of Electronic Microscopy, University of Bor-deaux I

climatic conditions during maturation, the sani-tary state of the harvest, and pesticide treatmentsapplied to the vine. The most abundant yeast pop-ulations are obtained in warm climatic conditions(lower latitudes, elevated temperatures). Insecti-cide treatments and certain fungicidal treatmentscan contribute to the rarefaction of indigenousgrape microflora. Quantitative results available onthis subject, however, are few. After the harvest,transport and crushing of the crop, the numberof cells capable of forming colonies on an agarmedium generally attains 106 cells/ml of must.

The number of yeast species significantlypresent on the grape is limited. Strictly oxida-tive metabolism yeasts, which belong to the genusRhodotorula and a few alcohol-sensitive species,are essentially found there. Among the latter,the apiculated species (Kloeckera apiculata andits sporiferous form Hanseniaspora uvarum) arethe most common. They comprise up to 99%of the yeasts isolated in certain grape samples.The following are generally found but in lesserproportions: Metschnikowia pulcherrima, Candidafamata, Candida stellata, Pichia membranefaciens,Pichia fermentans, Hansenula anomala.

All research confirms the extreme rarity ofS. cerevisiae on grapes. Yet these yeasts arenot totally absent. Their existence cannot beproven by spreading out diluted must on a solidmedium prepared in aseptic conditions, but theirpresence on grapes can be proven by analyzingthe spontaneous fermentative microflora of grapesamples placed in sterile bags, then asepticallycrushed and vinified in the laboratory in theabsence of all contamination. Red and white grapesfrom the Bordeaux region were treated in thismanner. At mid-fermentation in the majority ofcases, S. cerevisiae represented almost all of theyeasts isolated. In some rare cases, no yeast of thisspecies developed and apiculated yeasts began thefermentation.

Ecological surveys carried out at the BordeauxFaculty of Enology from 1992 to 1999 (Nau-mov et al., 2000a) demonstrated the presence ofS. uvarum yeasts on grapes and in spontaneouslyfermenting white musts from the Loire Valley,Jurancon, and Sauternes. The frequency of the


presence of this species alongside S. cerevisiaevaries from 4–100%. On one estate in Alsace,strains of S. uvarum were identified on grapes,in the press, and in vats, where they representedup to 90% of the yeasts involved throughoutfermentation in two consecutive years (Lejeune,unpublished work). More recently, Naumov et al.(2002) showed that S. uvarum, identified on grapesand in fermenting must, was involved in makingTokay wine.

The adaptation of S. uvarum to relatively lowtemperatures (6–10◦C) certainly explains its pres-ence in certain ecological niches: northerly vine-yards, late harvests, and spontaneous “cool” fer-mentation of white wines. In contrast, this strainis sensitive to high temperatures and has not beenfound in spontaneous fermentations of red Bor-deaux wines.

Recent observations also report the presence ofnatural S. cerevisiae/S. uvarum hybrids on grapesand in wineries where both species are present(Lejeune, unpublished work).

Between two harvests, the walls, the floors,the equipment and sometimes even the winerybuilding are colonized by the alcohol-sensitivespecies previously cited. Winemakers believe,however, that spontaneous fermentations are moredifficult to initiate in new tanks than in tanks whichhave already been used. This empirical observationleads to the supposition that S. cerevisiae canalso survive in the winery between two harvests.Moreover, this species was found in non-negligibleproportions in the wooden fermenters of some ofthe best vineyards in Bordeaux during the harvest,just before they were filled.

In the first hours of spontaneous fermentations,the first tanks filled have a very similar microflorato that of the grapes. There is a large proportionof apiculated yeasts and M. pulcherrima. Afterabout 20 hours, S. cerevisiae develops and coexistswith the grape yeasts. The latter quickly disappearat the start of spontaneous fermentation. In redwinemaking in the Bordeaux region, as soon asmust density drops below 1.070–1.060, the colonysamples obtained by spreading out diluted muston a solid medium generally isolate exclusivelyS. cerevisiae (107 to 108 cells/ml). This species

plays an essential role in the alcoholic fermentationprocess. Environmental conditions influence itsselection. This selection pressure is exhibited byfour principal parameters: anaerobiosis; must orgrape sulfiting; the sugar concentration; and theincreasing presence of ethanol. In winemakingwhere no sulfur dioxide is used, such as whitewines for the production of spirits, the dominantgrape microflora can still be found. It is largelypresent at the beginning of alcoholic fermentation(Figure 1.36). Even in this type of winemaking,the presence of apiculated yeasts is almost non-existent at mid alcoholic fermentation.

During dry white winemaking, the separation ofthe marc after pressing combined with clarificationby racking strongly reduces yeast populations,at least in the first days of the harvest. Theyeast population of a severely racked must rarelyexceeds 104 to 105 cells/ml.

A few days into the harvest, the alcogeneousS. cerevisiae yeasts contaminate the harvestmaterial, grape transport machinery and especiallythe harvest receiving equipment, the crusher-stemmer, and the wine press. For this reason, itis already largely present at the time of filling thetanks (around 50% of yeasts isolated during thefirst homogenization pumping-over of a red-grapetank). Fermentations are initiated more rapidly inthe course of the winemaking campaign because

BA0

5

10

15

20

25

30 Sacch. cerevisiae

Candida famata

Kloeckera apiculata

Num

ber

of c

olon

ies

isol

ated

Fig. 1.36. Comparison of yeast species present at thestart of alcoholic fermentation (d = 1.06). (A) in a tankof sulfited red grapes in Bordeaux (Frezier, 1992); (B) ina tank of unsulfited white must, for the elaboration ofCognac (Versavaud, 1994)


of this increased percentage of S. cerevisiae. Infact, the last tanks filled often complete theirfermentations before the first ones. Similarly, staticracking in dry white winemaking is becomingmore and more difficult to achieve, even atlow temperatures, from the second week of theharvest onward, especially in hot years. The entireinstallation inoculates the must with a sizeablealcogeneous yeast population. General weeklydisinfection of the pumps, the piping, the winepresses, the racking tanks, etc. is therefore stronglyrecommended.

During the final part of alcoholic fermenta-tion (the yeast decline phase), the population ofS. cerevisiae progressively decreases while stillremaining greater than 106 cells/ml. In favorablewinemaking conditions, characterized by a rapidand complete exhaustion of sugars, no other yeastspecies significantly appears at the end of fermen-tation. In poor conditions, spoilage yeasts can con-taminate the wine. One of the most frequent andmost dangerous contaminations is due to the devel-opment of Brettanomyces intermedius, which isresponsible for serious olfactive flaws (Volume 2,Section 8.4.5).

In the weeks that follow the completion ofalcoholic fermentation, the viable populations ofS. cerevisiae drop rapidly, falling below a few hun-dred cells/ml. In many cases, other yeast species(spoilage yeasts) can develop in wines during age-ing or bottle storage. Some yeasts have an oxida-tive metabolism of ethanol and form a veil on thesurface of the wine, such as Pichia or Candida, oreven certain strains of S. cerevisiae—sought afterin the production of specialty wines. By toppingoff regularly, the development of these respiratorymetabolism yeasts can be prevented. Some otheryeasts, such as Brettanomyces or Dekkera, candevelop in anaerobiosis, consuming trace amountsof sugars that have been incompletely or notfermented by S. cerevisiae. Their population canattain 104 to 105 cells/ml in a contaminated redwine in which alcoholic fermentation is other-wise completed normally. These contaminationscan also occur in the bottle. Refermentation yeastscan develop significantly in sweet or botrytizedsweet wines during ageing or bottle storage; the

principal species found are Saccharomycodes lud-wigii, Zygosaccharomyces bailii, and also somestrains of S. cerevisiae that are particularly resis-tant to ethanol and sulfur dioxide.

1.10.2 Recent Advances in the Studyof the Ecology of S. cerevisiaeStrains

The ecological study of the clonal diversity ofyeasts, and in particular of S. cerevisiae dur-ing winemaking, was inconceivable for a longtime because of a lack of means to distin-guish yeast strains from one another. Suchresearch has become possible with the develop-ment of molecular yeast strain identification meth-ods (Section 1.9). This Section focuses on recentadvances in this domain.

The alcoholic fermentation of grape must orgrapes is essentially carried out by a single yeastspecies, S. cerevisiae. Therefore, an understandingof the clonal diversity within this species ismuch more important for the winemaker thaninvestigations on the partially or non-fermentativegrape microflora.

The analysis in this Section of S. cerevisiaestrains in practical winemaking conditions in par-ticular intends to answer the following questions:

• Is spontaneous fermentation carried out by adominant strain, a small number or a very largenumber of strains?

• Can the existence of a succession of strainsduring alcoholic fermentation be proven? If so,what is their origin: grape, harvest material, orwinery equipment?

• During winemaking and from one year toanother in the same winery or even the samevineyard, is spontaneous alcoholic fermentationcarried out by the same strains?

• Can the practice of inoculating with selectedstrains modify the wild microflora of a vineyard?

During recent research (Dubourdieu and Frezier1990; Frezier 1992; Masneuf 1996), many samplesof yeast microflora were taken at the vineyard and


the winery from batches of white and red winesspontaneously fermenting or inoculated with activedry yeasts. Several conclusions can be drawn fromthis research, carried out on several thousand wildstrains of S. cerevisiae.

In the majority of cases, a small numberof major strains (one to three) representing upto 70–80% of the colonies isolated, carry outthe spontaneous fermentations of red and drywhite wines. These dominant strains are found incomparable proportions in all of the fermentorsfrom the same winery from start to end of alcoholic

fermentation. This phenomenon is illustrated bythe example given in Figure 1.37, describing theindigenous microflora of several tanks of red mustin a Pessac-Leognan vineyard in 1989. The strainsof S. cerevisiae, possessing different karyotypes,are identified by an alphanumeric code comprisingthe initial of the vineyard, the tank number, thetime of the sampling, the isolated colony numberand the year of the sample. Two strains, FzIb1(1989) and FzIb2 (1989), are encountered in all ofthe tanks during the entire alcoholic fermentationprocess.

12

10

8

6

4

2

0b c d

Samples

Tank I

Samples

Tank II

Samples

Tank III

Samples

Tank IV

12

10

8

6

4

2

0b c d

12

10

8

6

4

2

0b c d

12

10

8

6

4

2

0b c d

Other different yeast strains.

Fzlb1

Fzlb2

Fig. 1.37. Breakdown of S. cerevisiae caryotypes during alcoholic fermentation in red grape tanks in Fz vineyard(Pessac-Leognan, France) in 1989 (Frezier, 1992) (b, c, and d designate the start, middle, and end of alcoholicfermentation, respectively). Tanks I and II (Merlot) and III and IV (Cabernet-Sauvignon) are filled on the 1st, 3rd, 7thand 23rd day of the harvest, respectively


The spontaneous fermentation of dry whitewines in the same vineyard is also carried outby the same dominant yeast strains in all of thebarrels.

The tank filling order and the grape varietyhave little effect on the clonal composition of thepopulations of S. cerevisiae spontaneously foundin the winery. The daily practice of pumping-over the red grape must with pumping equipmentused for all of the tanks probably ensures thedissemination of the same strains in the winery.In white winemaking, the wine press installationplays the same role as an inoculator.

In addition, in Figure 1.37, all of the strains ana-lyzed are K2 killer. The two dominant strains donot ferment galactose (phenotype Gal−). Their for-mer denomination was therefore S. oviformis orS. cerevisiae (race bayanus) in previous classifi-cations. Domercq (1956) observed a lesser propor-tion of S. oviformis in the spontaneous microfloraof Bordeaux region fermentations in the 1950s(one-fifth at the beginning of fermentation toone-third at the end). In the indigenous fermenta-tive microflora of Bordeaux musts, certain strainsof S. cerevisiae Gal− which dominate from thestart of alcoholic fermentation were selected overtime. The causes of this change in the microflora,remain unknown. On the other hand, a systematicincrease in the proportion of Gal− strains duringred or dry white wine fermentation has not beenobserved (Table 1.9). In botrytized sweet winesfrom Sauternes, the succession of strains is moredistinct.

The same major strain is frequently encounteredfor several consecutive vintages in the samevineyard in spontaneous-fermentation red-grapemust tanks. In 1990, one of the major strains wasthe same as the previous year in the red grape musttank of the Fz vineyard. Other strains appeared,however, which had not been isolated in 1989.

When sterile grape samples are taken, pressedsterilely, sulfited at winemaking levels andfermented in the laboratory in sterile containers,one or several dominant strains responsible forspontaneous fermentations in the winery exist insome samples. These strains are therefore presentat the vineyard. In practice, they probably begin

Table 1.9. Example of physiological race breakdown(%) of Saccharomyces cerevisiae during spontaneousalcoholic fermentation (Frezier, 1992)

Stage of Physiological racefermentation cerevisiae oviformis capensis chevalieri

Red winea

start 23 77 — —middle 10 90 — —end 14 84 2 —

White wineb

start 23 62 — 14middle 35 60 — 4end 32 62 — 6

Sweet winec

start 37 51 4 8middle 40 56 4 3end 23 73 4 4

aIsolation of 100 colonies from six tanks of a Pessac-Leognanvineyard (four tanks in 1989 and two tanks in 1990).bIsolation of 100 colonies in three barrels of a Pessac-Leognanvineyard.cIsolation of 100 colonies in two barrels of a Sauternes vineyardin 1990.

Table 1.10. Rate of occurrence of the dominant FZIB2-89 caryotype in microvinifications carried out on sterilegrape samples (I, II, III) in the FZ vineyard

Year Number Sample I Sample II Sample IIIof clonesanalyzed

1990 30 — 70% —1991 60 — — —1992 85 25% 31% 3%1993 74 — — —1994 79 87% — 40%

to multiply as soon as the grapes arrive at thewinery. A few days into the harvest, they infestthe winery equipment which in turn ensures asystematic inoculation of the fresh grape crop.

The presence each year of the same dom-inant strain in the vineyard is not systematic(Table 1.10). In the Fz vineyard, the FzIb2-89strain could not be isolated in 1991 although itwas present in certain vineyard samples in 1990,1992 and 1994. In 1993, another strain proved tobe dominant in spontaneous fermentations of ster-ile grape samples.


The spontaneous microflora of S. cerevisiaeseems to fluctuate. At present, the factors involvedin this fluctuation have not been identified. In agiven vineyard, spontaneous fermentation is notsystematically carried out by the same strains eachyear; strain specificity does not exist and thereforedoes not participate in vineyard characteristics.Ecological observations do not confirm the notionof a vineyard-specific yeast. Furthermore, someindigenous strains, dominant in a given vineyard,have been found in other nearby or distantvineyards. For example, the FzIb2-89 strain,isolated for the first time in a vineyard in Pessac-Leognan, was later identified not only in thespontaneous fermentation of dry white and redwines of other vineyards in the same appellation,but also in relatively distant wineries as far away asthe Medoc. This strain has since been selected andcommercialized under the name Zymaflore F10.

In some cases (Figure 1.38), S. cerevisiae popu-lations with a large clonal diversity carry out spon-taneous must fermentation. Many strains coexist.Their proportions differ from the start to the endof fermentation and from one winery to another.In the Bordeaux region, this diversity causes slow

fermentations and sometimes even stuck fermen-tations. No strain is capable of asserting itself. Onthe other hand, the presence of a small numberof dominant strains generally characterizes com-plete and rapid spontaneous fermentations. Thesedominant strains are found from the start to theend of the fermentation.

In normal red winemaking conditions, the inoc-ulation of the first tanks in a winery influencesthe wild microflora of non-inoculated tanks. Thestrain(s) used for inoculating the first tanks are fre-quently found in majority in the latter. Figure 1.39provides an example comparing the microflora ofa tank of Merlot from Pomerol, inoculated withan active dry yeast strain (522M) on the first dayof the harvest, with a non-inoculated tank filledlater. From the start of alcoholic fermentation,the selected strain is successfully implanted in theinoculated tank. Even in the non-inoculated tank,the same strain is equally implanted throughout thefermentation. It is therefore difficult to select thedominant wild strains in red winemaking tanks,when some of the tanks have been inoculated.An early and massive inoculation of the must,however, permits the successful implantation of

b c d0

2

4

6

8

10

LGld5 LGld1 LGlc8 LGlc6 LGlb10 LGlb3

LGld9 LGld3 LGlc9 LGlc7 LGlc2 LGlb5 LGlb1

Fig. 1.38. Breakdown of S. cerevisiae caryotypes in tank I of red grapes at LG vineyard (Pomerol, France) in 1989(Frezier, 1992) (b, c, and d designate the start, middle, and end of alcoholic fermentation, respectively)


Samples

Samples

Tank I

Tank II

522M

12

10

8

6

4

2

0

12

10

8

6

4

2

0

b c d

b c d


Fig. 1.39. Breakdown of S. cerevisiae caryotypes intank I and tank II P vineyard in 1990 (b, c, andd designate the start, middle, and end of alcoholicfermentation, respectively). Tank I is inoculated with522M dry yeasts and tank II underwent spontaneousfermentation (b, c and d designate the start, middle andend of alcoholic fermentation, respectively)

different selected yeasts in several tanks at thesame winery (Figure 1.40).

In white winemaking, inoculating rarely influ-ences the microflora of spontaneous fermentationsin wineries. For the most part, dominant indige-nous strains in non-inoculated barrels of ferment-ing dry white wine are observed, whereas in thesame wine cellar, other batches were inoculated

12

10

8

6

4

2

0b c d

10

8

6

4

2

0b c d

12

10

8

6

4

2

0b c d

522M

F10

F5

Samples

Samples

Samples

Tank I, inoculated with F5

Tank II, inoculated with F10

Tank III, inoculated with 522M


Fig. 1.40. Breakdown of S. cerevisiae caryotypes intanks I, II and III in F vineyard in 1990, with massiveearly inoculation with F5, F10 and 522M, respectively(Frezier, 1992) (b, c, and d designate the start, middle,and end of alcoholic fermentation, respectively)


with different selected yeasts. The absence ofpumping-overs probably hinders the disseminationof the same yeasts in all of the fermenting barrels.This situation permits the fermentative behaviorand enological interest of different selected strainsto be compared with each other and with indige-nous strains in a given vineyard. The barrels arefilled with the same must; some are inoculated withthe yeast to be compared. A sample of the biomassis taken at mid fermentation. The desired implan-tation is then verified by PCR associated with δ

sequences. Due to the ease of use of this method,information on characteristics of selected strainsand their influence on wine quality can be gatheredat the winery.

Vezhinet et al. (1992) and Versavaud et al.(1995) have also studied the clonal diversity ofyeast microflora in other vineyards. Their resultsconfirm the polyclonal character of fermentativepopulations of S. cerevisiae. The notion ofdominant strains (one to two per fermentation) isobvious in the work carried out in the Charentesregion. As in Champagne and the Loire Valley,some Charentes region strains are found forseveral years in a row in the same winery. Thepresence of these dominant strains on the grapehas been confirmed before any contact with wineryequipment during several harvests.

Why do some S. cerevisiae strains issued froma very heterogeneous population become domi-nant during spontaneous fermentation? Why canthey be found several years in a row at the samevineyard and wine cellar? Despite their practi-cal interest, these questions have not often beenstudied and there are no definitive responses. Itseems that these strains rapidly start and com-plete alcoholic fermentation and have a goodresistance to sulfur dioxide (up to 10 g/hl). Fur-thermore, during mixed inoculations in the lab-oratory of either 8% ethanol or non-fermentedmusts, these strains rapidly become dominantwhen placed in the presence of other wildnon-dominant strains of S. cerevisiae isolated atthe start and end of fermentation. This sub-ject merits further research. Without a doubt,it would be interesting to compare the geneticcharacteristics of dominant and non-dominant

yeasts and their degree of heterozygosity. Con-sidering the genome renewal theory of Mortimeret al. (1994) (Section 1.6.2), dominant strains arepossibly more homozygous than non-dominantstrains.

REFERENCES

Aigle M., Erbs D. and Moll M. (1984) Am. Soc. ofBrewing Chemists, 42 (1), 1–7.

Alexandre E.H., Rousseaux I. and Charpentier C.(1994) Biotechnol. Appl. Biochem., 19.

Augustyn O.P.H., Kock J.L.F. and Ferreira D. (1991)System. Appl. Microbiol., 15, 105–115.

Barnett J.A., Delaney M.A., Jones E., Magson A.B. andWinch B. (1972) Arch. Microbiol., 83, 52–55.

Barnett J.A., Payne R.W. and Yarrow D. (2000) Yeasts:Characteristics and Identification, 3rd edn. Cam-bridge University Press Cambridge, UK.

Barre P. (1978) Killer factor activity under vinificationconditions. Vth International Symposium on Yeasts,Montpellier.

Barre P. (1980) Bull. OIV, 53, 560–567.Barre P. (1984) Bull. OIV, 57, 635–643.Barre P. (1992) Le facteur Killer. In Les acquisitions

recentes de la microbiologie du vin (ed. B. Doneche),pp. 63–69. Tec & Doc Lavoisier, Paris.

Belin J.M. (1979) Mycopathologia, 67 (2), 67–81.Belin J.M. (1981) Biologie des levures liees a la vigne

et au vin. These Docteur es Sciences Naturelles,Universite de Dijon.

Bendova O., Richter V., Janderova B. and Hausler J.(1991) Appl. Microbiol. Biotechnol., 35, 810–812.

Bicknell J.N. and Douglas H.C. (1982) J. Bacteriol.,101, 505–512.

Bisson L.F. (1991) Yeasts–metabolism of sugars. InWine Microbiology and Biotechnology, pp. 55–75.Harwood Academic Publishers.

Blondin B. and Vezhinet F. (1988) Rev. Fr. Oenol., 115,7–11.

Boone C., Sdicu A.M., Wagner J., Degre R., Sanchez C.and Bussey H. (1990) Am. J. Enol. Vitic., 41, 37–42.

Bouix M., Leveau J.Y. and Cuinier C. (1981) Conn.Vigne Vin, 15, 41–52.

Boulton R.B., Singleton V.L., Bisson L.F. and KunkeeR. (1995) Principles and Practices of Winemaking.Chapman & Hall Enology Library, New York.

Bowers J., Boursiquot J.M., This P., Chu K., Johan-sson H. and Meredith C. (1999) Science, 285, 1562–1565.

Brechot P., Chauvet J. and Girard H. (1962) Ann. Techn.Agric., 11 (3), 235–244.

Bureau G., Brun O., Vigues A., Maujean A., VesselleG. and Feuillat A. (1982) Conn. Vigne Vin, 16 (1),15–32.


Carle G.F. and Olson M.V. (1984) Nucl. Acids Res., 12,5647–5664.

Castelli T. (1955) Am. J. Enol. Vitic., 6, 18–20.Castelli T. (1967) Ecologie et systematique des levures

du vin. IIeme Symposium International d’Enologie,Bordeaux-Cognac, INRA Ed. Paris.

Charpentier Cl. (1995) Rev. Enolo. 73 S, 25–28.Ciolfi G. (1992) L’enotechnico, XXVIII, (11) 87.Ciolfi G. (1994) L’enotechnico, Nov., 71–76.Cuinier C. and Gros C. (1983) Vignes Vins, 318,

25–27.Cuinier C. and Guerineau L. (1976) Evolution de la

microflore au cours de la vinification des vins deChinon. Vignes Vins, 269, 29–41.

Cuinier C. and Leveau J.Y. (1979) Vignes Vins, 283,44–49.

Defontaine A., Lecocq F.M. and Hallet J.N. (1991) Nucl.Acids Res., 19 (1), 185.

Degre R., Thomas D.Y., Frenette J. and Mailhot K.(1989). Rev. Fr. Oenol., 119, 23–26.

Delfini C. (1995) Scienza e technica di microbioloiaenologica. Il Lievito, Asti.

De Rossi G. (1935) Il lieviti della fermentazione nellaregione umbra. IVeme Congres International de laVigne et du Vin, Lausanne.

Domercq S. (1956) Etude et classification des levuresde vin de la Gironde. These de Docteur-Ingenieur del’Universite de Bordeaux.

Dubourdieu D. and Frezier V. (1990) Rev. Fr. Oenol.,30, 37–40.

Dubourdieu D. and Moine V. (1995) Role des conditionsd’elevage sur la stabilisation proteique des vinsblancs. In Actualites enologiques 95, Compte-rendusdu 5eme Symposium International d’Enologie deBordeaux.

Dubourdieu D., Sokol A., Zucca J., Thalouarn P., DateeA. and Aigle M. (1987) Conn. Vigne Vin, 4, 267–278.

Field D. and Wills C. (1998) Proc. Natl. Acad. Sci., 95,1647–1652.

Fleet G.H. (1991) Cell wall. In The Yeasts, Vol. 4: YeastsOrganelles, pp. 199–277. Academic Press, London.

Fleet G.H. (1993) Wine Microbiology and Biotech-nology. Harwood Academic Publishers, Chur, Swit-zerland.

Flick J.S. and Thorner J. (1993) Mol. Cell. Biol. 13,5861–5876.

Frezier V. (1992) Recherche sur l’ecologie des souchesde Saccharomyces cerevisiae au cours des vinifica-tions bordelaises. These de Doctorat de l’Universitede Bordeaux II.

Frezier V. and Dubourdieu D. (1992) Am. J. Enol. Vitic.43, 375–380.

Gaillardin C. and Heslot H. (1987) La levure. La Recher-che, 188 (18) 586–596.

Gonzalez Techera A., Jubany S., Carrau F.M. andGaggero C. (2001) Lett. Appl. Microbiol., 33, 71–75.

Guerin B. (1991) Mitochondria. In The Yeasts, Vol. 4:Yeasts Organelles, pp. 541–589. Academic Press,London.

Guillamon J.M., Sabate J., Barrio E., Cano J. and QuerolA. (1998) Arch. Microbiol., 169 (5), 387–392.

Hallet J.N., Craneguy B., Zucca J. and Poulard A. (1988)Prog. Agric. Vitic., 105, 328–333.

Hallet J.N., Craneguy B., Daniel P. and Poulard A.(1989) Caracterisation des souches levuriennes desmouts et des lies par le polymorphism de restrictionde leur ADN mitochondrial. In Actualites Enologiques89, Comptes rendus du 4eme Symposium d’Enologiede Bordeaux. Dunod, Paris.

Hansen J. and Kielland-Brandt M.C. (1994) Gene, 140,33–40.

Heard G.M. and Fleet G.H. (1987) Appl. Environ.Microbiol. 51, 539–45.

Heard G.M. and Fleet G. (1990) J. Appl. Bacteriol., 68,445–447.

Hennequin C., Thierry A., Richard G.F., Lecointre G.,Nguyen H.V., Gaillardin Cl., Dujon B. (2001), J. Clin.Microbiol., 39 (2), 551–559.

Henschke P.A. and Rose A.H. (1991) Plasma membrane.In The Yeasts, Vol. 4: Yeasts Organelles, pp. 297–345.Academic Press, London.

Herskowitz I., Rine J. and Strathern J.N. (1992) Mating-type determination and mating-type interconversionin Saccharomyces cerevisiae. In Jones, (eds E.W.J.N. Strathern and J.R. Broach). Broach J.R. (eds)The Molecular Biology of the Yeast Saccharomyces:Gene Expression (eds E.W. Jones, J.N. Strathernand J.R. Broach), pp. 583–656. Cold Spring Harbor,Laboratory Press, New York.

Klis F.M. (1994) Yeast, 10, 851–869.Kregger-Van Rij N.J.W. (1984) The Yeasts, a Taxonomic

Study. Elsevier Science Publishers, B.V. Amsterdam.Lafon-Lafourcade S. (1983) Wine and brandy. In

Biotechnology Vol. V: Food and Feed Productionwith Microorganisms (eds H.J. Rehm and G. Reed).Verlag Chemie, Weinheim.

Lafon-Lafourcade S. and Joyeux A. (1979) Conn. VigneVin, 4, 295–310.

Lavallee F., Salves S., Lamy S., Thomas D.Y., Degre R.and Dulau L. (1994) Am. J. Enol. Vitic. 45, 86–91.

Lodder J. (1970) The Yeasts, a Taxonomic Study, 2ndedn. Elsevier Science Publishers, B.V. Amsterdam.

Lodder J. and Kregger-Van Rij N.J.W. (1952) TheYeasts, a Taxonomic Study. Elsevier Science Pub-lishers, B.V. Amsterdam.

Lopez V., Fernandez-Espinar M.T., Barrio E., Ramon D.and Querol A. (2003) Int. J. Food Microbiol., 81 (1),63–71.

Masneuf I. (1996) Recherches sur l’identificationgenetique des levures de vinification. Applicationsenologiques. These de Doctorat de l’Universite deBordeaux II.


Masneuf I., Aigle M. and Dubourdieu D. (1996a) FEMSMicrobiol. Letters,.

Masneuf I., Aigle M. and Dubourdieu D. (1996b) J. Int.Sci. Vigne Vin, 30 (1), 15–21.

Masneuf I. and Dubourdieu D. (1994) J. Int. Sci. VigneVin, 28 (2), 153–160.

Masneuf I., Hansen J., Groth C., Piskpur J. andDubourdieu D. (1998) Appl. Environ. Microbiol.,64 (10), 3887–3892.

Minarik E. (1971) Conn. Vigne Vin, 2, 185–198.Mortimer R.K., Romano P., Suzzi G. and Polsinelli M.

(1994) Yeast, 10, 1543–1552.Naumov G.I. (1987) Genetic basis for classification and

identification of the ascomycetous yeasts. Studies inMycology, 30, 469–475.

Naumov G.I., Masneuf I., Naumova E.S., Aigle M.and Dubourdieu D. (2000a) Res. Microbiol., 151,683–691.

Naumov G.I., Naumova E.S., Masneuf I., Aigle M.,Kondratieva V.I. and Dubourdieu D. (2000b) Syst.Appl. Microbiol., 23, 442–449.

Naumov G.I., Naumova E.S., Antunovics Z. andSipiczki M. (2002) Appl. Microbiol. Biotechnol.,59 (6), 727–30.

Naumov G., Naumova E. and Gaillardin C. (1993)Genetic and karyotypic identification of wine Saccha-romyces bayanus yeasts isolated in France and Italy.System. Appl. Microbiol., 16, 274–279.

Naumov G.I., Naumova E.S. and Sniegowski P.D.(1998) Can. J. Microbiol., 44, 1045–1050.

Nguyen H.V., Lepingle A. and Gaillardin C.A. (2000)Syst. Appl. Microbiol., 23 (1), 71–85.

Ness F., Lavallee F., Dubourdieu D., Aigle M. andDulau L. (1992) J. Sci. Food Agric., 62, 89–94.

Park Y.H. (1975) Conn. Vigne Vin, 3, 253–278.Pasteur L. (1866) Etudes sur le vin. Imprimerie Impe-

riale, Paris.Pasteur L. (1876) Etudes sur la biere. Gauthier-

Villars, Paris.Pedersen M.B. (1986) Carlsberg Res. Commun. 51,

163–183.Perez M.A., Gallego F.J., Martinez I. and Hidalgo P.

(2001) Lett. Appl. Microbiol., 33, 461–466.Peynaud E. and Domercq S. (1953) Ann. Technol. Agric.,

4, 265–300.Poulard A. and Lecocq M. (1981) Rev. Fr. Oenol., 82,

31–35.Poulard A., Simon L. and Cuinier Cl. (1980) Conn.

Vigne Vin, 14, 219–238.Querol A., Barrio E. and Ramon D. (1992) System. Appl.

Microbiol., 15, 439–446.Radler F. (1988) in Application a l’oenologie des

progres recents en microbiologie et en fermentation.Office international de la vigne et du vin Ed., Paris.pp. 273–282.

Redzepovic S., Orlic S., Sikora S., Madjak A. andPretorius I.S. (2002) Lett. Appl. Microbiol., 35,305–310.

Ribereau-Gayon J., Peynaud E., Ribereau-Gayon P. andSudraud P. (1975) Traite d’Enologie. Sciences ettechniques du vin, Vol 2. Dunod Paris.

Rose A.H. and Harrison J.S. (1969) The Yeasts, Vol. 1:Biology of Yeasts. Academic Press, London.

Rose A.H. and Harrison J.S. (1991) The Yeasts, Vol. 4:Yeasts Organelles. Academic Press, London.

Rossini G., Federici F. and Martini A. (1982) Microbiol.Ecol., 8, 83–89.

Rozes N., Garcia-Jares C., Larue F. and Lonvaud-FunelA. (1992) J. Sci. Food Agric., 59, 351–357.

Saiki R.K., Sharf S., Falcona F. and Lonvaud-Funel A.(1992) J. Sci. Food Agric., 59, 315–357.

Salmon J.M., Vincent O., Mauricio J.C. and Bely M.(1993) Am. J. Enol. Vitic., 44 (1), 56–64.

Schwartz D.C. and Cantor C.R. (1984) Cell, 37, 67–75.Seki T., Choi E.H and Ryu D. (1985) Appl. Environ.

Microbiol. 49, 1211.Singer S.J. and Nicolson G.L. (1972) Science, 175,

720–731.Soufleros E. (1978) Les levures de la region viticole

de Naoussa (Grece). Identification et classification.Etude des produits volatils formes au cours de lafermentation. These Docteur-Ingenieur Universite deBordeaux II.

Southern E. (1975) J. Biol. Chem., 98, 503–517.Stratford M. (1994) Yeast, 10, 1741–1752.Tipper D.J. and Bostian K.A. (1984) Double stranded

ribonucleic acid killer system in Yeasts. Microbiol.Rev. 48 (2), 125–136.

Tredoux H.G., Kock J.L.F., Lategan P.L. and MullerH.B. (1987) Am. J. Enol. Vitic., 38, 161–164.

Tuite M.F. and Oliver S.G. (1991) Saccharomyces Bio-technology Handbooks. Plenum Press, New York.

Van Vuuren H.J.J. and Jacobs C.J. (1992) Am. J. Enol.Vitic., 43 (2), 119–128.

Van Vuuren H.J.J. and Van Der Meer L. (1987) Am. J.Enol. Vitic., 38, 49–53.

Vaughan Martini A. and Martini A. (1987) Three newlydelimited species of Saccharomyces sensu stricto.Antonie van Leeuwenhoek, 53, 77–84.

Vaughan Martini A. and Martini A. (1993) System. Appl.Microbiol., 16, 113–119.

Versavaud A., Dulau L. and Hallet J.N. (1993) Rev. Fr.Oenol., 142, 20–28.

Versavaud A., Courcoux Ph., Roulland Cl., Dulau L. andHallet J.N. (1995) Appl. Environ. Microbiol., 61 (10),3521–3529.

Vezhinet F., Blondin B. and Hallet J.N. (1990) Appl.Micro-Biotecn., 32, 568–571.

Vezhinet F., Hallet J.N., Valade M. and Poulard A.(1992) Am. J. Enol. Vitic., 43 (1), 83–86.

Williamson D.H. (1991) Nucleus, chromosomes andplasmids. In The Yeasts, Vol. 4: Yeasts organelles, (eds


A.H. Rose and J.S. Harison), pp. 433–482. AcademicPress, London.

Wingfield B.D., Van Der Meer L.J., Pretorius S. and VanVuuren H.J.J. (1990) Mycol. Res., 94, 901–906.

Yarrow D. and Nakase T. (1975) Antonie van leeuwen-hoek, 41 81–88.

Young T.W. (1987) Killer yeasts. In The Yeasts, Vol.2, Yeasts and the Environment (eds A.H. Rose andJ.S. Harrison), pp. 131–164. Academic Press, NewYork.

Zorg J. and Kilian J. and Radler F. (1988) Archiv.Microbiol., 149, 261–267.

Cytology, Taxonomy and Ecology of Grape and Wine …...Cytology,Taxonomy and Ecology of Grape and...

Documents

Transcript of Cytology, Taxonomy and Ecology of Grape and Wine …...Cytology,Taxonomy and Ecology of Grape and...